Risk and Exposure Assessment to Support the
Review of the SC>2 Primary National Ambient Air
Quality Standards: Final Report
-------
EPA-452/R-09-007
July 2009
Risk and Exposure Assessment to Support the
Review of the SC>2 Primary National Ambient Air
Quality Standards
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina
11
-------
Disclaimer
This document has been prepared by staff from the Ambient Standards Group, Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency. Any opinions,
findings, conclusions, or recommendations are those of the authors and do not necessarily reflect
the views of the EPA. For questions concerning this document, please contact Dr. Stephen
Graham (919-541-4344; graham.stephen@epa.gov), Mr. Harvey Richmond (919-541-5271;
richmond.harvey@epa.gov), or Dr. Michael Stewart (919-541-7524; stewart.michael@epa.gov)
in
-------
Table of Contents
LIST OF TABLES IX
LIST OF FIGURES XIII
LIST OF ACRONYMS/ABBREVIATIONS XX
1. INTRODUCTION 1
1.1 History 5
1.1.1 History of the SO2NAAQS 5
1.1.2 Health Evidence from the Previous Review 6
1.1.3 Assessment from Previous Review 8
1.2 Scope of the Risk and Exposure Assessment for the Current Review 9
1.2.1 Overview of the Risk and Exposure Assessment 9
1.2.2 Species of Sulfur Oxides Included in Analyses 11
2. OVERVIEW OF HUMAN EXPOSURE 12
2.1 Background 13
2.2 Sources of SO2 13
2.3 Background on the SO2 Monitoring Network 14
2.4 Ambient levels of SO2 18
2.5 Relationship of Personal Exposure to Ambient Concentrations 20
2.6 Key Observations 22
3. AT RISK POPULATIONS 23
3.1 Overview 23
3.2 Pre-existing Respiratory Disease 24
3.3 Genetics 24
3.4 Age 25
3.5 Time Spent Outdoors 26
3.6 Ventillation Rate 26
3.7 Socioeconomic Status 26
July 2009 iv
-------
3.8 Number of At Risk Individuals 26
3.9 Key Observations 27
4. INTEGRATION OF HEALTH EVIDENCE 28
4.1 Introduction 28
4.2 Respiratory Morbidity Following Short-term SO2 Exposure 31
4.2.1 Overview 31
4.2.2 Integration of Respiratory Morbidity Health Evidence 31
4.2.3 Medication as an Effect Modifier 34
4.3 What Constitutes an Adverse Health Impact from SO2 Exposure? 35
4.4 Key Observations 38
5. SELECTION OF POTENTIAL ALTERNATIVE STANDARDS FOR ANALYSIS 39
5.1 Introduction 39
5.2 Indicator 39
5.3 Averaging Time 39
5.4 Form 42
5.5 Level 43
5.6 Key Observations 52
6. OVERVIEW OF RISK CHARACTERIZATION AND EXPOSURE ASSESSMENT.. 53
6.1 Introduction 53
6.2 Potential Health Effect Benchmark Levels 54
6.3 Approaches for Assessing Exposure and Risk Associated with 5-Minute Peak SO2 Exposures 56
6.4 Approach for Estimating 5-Minute Peak SO2 Concentrations 59
6.5 Approach for Simulating the Current and Alternative Air Quality Standard Scenarios 62
6.6 Approaches for characterizing Variability and Uncertainty 64
6.6.1 Characterization of Variability 65
6.6.2 Characterization of Uncertainty 65
6.7 Key Observations 68
7. AMBIENT AIR QUALITY AND BENCHMARK HEALTH RISK
CHARACTERIZATION FOR 5-MINUTE PEAK SO2 EXPOSURES 69
July 2009 v
-------
7.1 Overview 69
7.2 Approach 73
7.2.1 Screening of Air Quality Data 73
7.2.2 Site Characteristics of Ambient SO2 Monitors 78
7.2.3 Statistical model to estimate 5-minute maximum SO2 concentrations 90
7.2.4 Adjustment of Ambient Concentrations to Evaluate the Current and Potential Alternative Air Quality
Scenarios 104
7.2.5 Air Quality Concentration Metrics 114
7.3 Results 120
7.3.1 Measured 5-minute Maximum and Measured 1-Hour SO2 Concentrations at Ambient Monitors -As Is Air
Quality 120
7.3.2 Measured 1-Hour and Modeled 5-minute Maximum SO2 Concentrations at All Ambient Monitors -As Is
Air Quality 127
7.3.3 Modeled 1-Hour and Modeled 5-minute Maximum SO2 Concentrations at Ambient Monitors in 40 Counties
-Air Quality Adjusted to Just Meet the Current and Potential Alternative Standards 133
7.4 Variability Analysis and Uncertainty Characterization 151
7.4.1 Variability Analysis 151
7.4.2 Uncertainty Characterization 151
7.5 Key Observations 181
8. EXPOSURE ANALYSIS 184
8.1 Overview 184
8.2 Overview of Human Exposure Modeling using APEX 185
8.3 Characterization of study areas 188
8.3.1 Study Area Selection 188
8.3.2 Study Area Descriptions 190
8.3.3 Time Period of Analysis 193
8.3.4 Populations Analyzed 193
8.4 Characterization of Ambient Hourly Air Quality Data Using AERMOD 193
8.4.1 Overview 193
8.4.2 General Model Inputs 194
8.4.3 Stationary Sources Emissions Preparation 196
8.4.4 Receptor Locations 206
8.4.5 Modeled Air Quality Evaluation 207
8.5 Simulated Population 221
8.5.1 Population Counts and Employment Probabilities 221
8.5.2 Asthma Prevalence 222
8.5.3 Commuting Database 223
8.5.4 Body Surface Area 224
8.5.5 Activity-Specific Ventilation Rates 224
8.6 Construction of Longitudinal Activity Sequences 227
8.7 Calculating Microenvironmental Concentrations 228
8.7.1 Approach for Estimating 5-Minute Maximum SO2 Concentrations 229
8.7.2 Microenvironments Modeled 231
July 2009 vi
-------
8.7.3 Microenvironment Descriptions 231
8.8 Exposure Measures and Health Risk Characterization 235
8.8.1 Estimation of Exposure 235
8.8.2 Estimation of Target Ventilation Rates 236
8.8.3 Adjustment for Just Meeting the Current and Alternative Standards 237
8.9 Exposure Modeling and Health Risk Characterization Results 242
8.9.1 Asthmatic Exposures to 5-minute SO2 Concentrations in Greene County 242
8.9.2 Asthmatic Exposures to 5-minute SO2 in St. Louis 248
8.10 Representativeness of Exposure Results 255
8.10.1 Introduction 255
8.10.2 Time spent outdoors 256
8.10.3 SO2 Emissions and Ambient Concentrations 258
8.10.4 American Housing Survey (AHS)Data 265
8.10.5 Asthma Prevalence 266
8.11 Variability Analysis and Uncertainty Characterization 268
8.11.1 Variability Analysis 268
8.11.2 Uncertainty Characterization 269
8.12 Key Observations 311
9. HEALTH RISK ASSESSMENT FOR LUNG FUNCTION RESPONSES IN
ASTHMATICS ASSOCIATED WITH 5-MINUTE PEAK EXPOSURES 314
9.1 Introduction 314
9.2 Development of Approach for 5-minute Lung Function Risk Assessment 315
9.2.1 General Approach 316
9.2.2 Exposure Estimates 321
9.2.3 Exposure-Response Functions 322
9.3 Lung Function Risk Estimates 332
9.4 Characterizing Uncertainty and Variability 345
9.5 Key Observations 350
10. EVIDENCE-AND EXPOSURE/RISK-BASED CONSIDERATIONS RELATED TO
THE PRIMARY SO2 NAAQS 353
10.1 Introduction 353
10.2 General Approach 354
10.3 Adequacy of the Current 24-hour Standard 356
10.3.1 Introduction 356
10.3.2 Evidence-based considerations 358
10.3.3 Air Quality, exposure and risk-based considerations 359
10.4 Adequacy of the Current Annual Standard 367
10.4.1 Introduction 367
July 2009 vii
-------
10.4.2 Evidence-based considerations 368
10.4.3 Risk-based considerations 369
10.4.4 Conclusions regarding the adequacy of the current annual standard 370
10.5 Potential Alternative Standards 370
10.5.1 Indicator 370
10.5.2 Averaging Time 371
10.5.3 Form 381
10.5.4 Level 387
10.6 Key Observations 397
REFERENCES 398
APPENDICES
Appendix A- Supplement to the SOi Air Quality Characterization
Appendix B- Supplement to the SOi Exposure Assessment
Appendix C- Sulfur Dioxide Health Risk Assessment
Appendix D- Supplement to the Policy Assessment
July 2009 viii
-------
LIST OF TABLES
Number Page
Table 2-1. SC>2 network monitoring objective distribution 17
Table 2-2. SC>2 network distribution across measurement scales 18
Table 4-1. Weight of evidence for causal determination 29
Table 7-1. Summary of all available 5-minute and 1-hour 862 ambient monitoring data, years
1997-2007, pre-screened 74
Table 7-2. Analytical data sets generated using the continuous-5, max-5, and 1-hour ambient
SO2 monitoring data, following screening 75
Table 7-3. Counts of complete and incomplete site-years of 1-hour 862 ambient monitoring data
for 1997-2006 79
Table 7-4. Descriptive statistics of the population residing within a 5 km radius of ambient
monitors by monitoring objective: monitors reporting 5-minute maximum SC>2
concentrations and the broader 862 monitoring network 89
Table 7-5. Comparison of prediction errors and model variance parameters for the four models
evaluated 103
Table 7-6. Prediction errors of the statistical model used in estimating 5-minute maximum SO2
concentrations above benchmark levels 104
Table 7-7. Counties selected for evaluation of air quality adjusted to just meeting the current and
potential alternative SC>2 standards and the number of monitors in each COV bin... 112
Table 7-8. The co-occurrence of daily 5-minute maximum and 1-hour daily maximum SC>2
concentrations using measured ambient monitoring data 115
Table 7-9. Example of how the probability of exceeding a 400 ppb 5-minute benchmark would
be calculated given 1 -hour daily maximum SC>2 concentration bins 118
Table 7-10. Percent of days having a modeled daily 5-minute maximum SC>2 concentration
above the potential health effect benchmark levels given air quality as is and air
quality adjusted to just meeting the current and each of the potential alternative
standards 138
Table 7-11. Modeled mean number of days per year with 5-minute maximum concentrations
above 100 ppb in 40 selected counties given 2001-2006 air quality as is and air quality
adjusted to just meet the current and alternative standards 147
Table 7-12. Mean number of modeled days per year with 5-minute maximum concentrations
above 200 ppb in 40 selected counties given 2001-2006 air quality as is and that
adjusted to just meet the current and alternative standards 148
Table 7-13. Mean number of modeled days per year with 5-minute maximum concentrations
above 300 ppb in 40 selected counties given 2001-2006 air quality as is and that
adjusted to just meet the current and alternative standards 149
Table 7-14. Mean number of modeled days per year with 5-minute maximum concentrations
above 400 ppb in 40 selected counties given 2001-2006 air quality as is and that
adjusted to just meet the current and alternative standards 150
Table 7-15. Summary of how variability was incorporated into the air quality characterization.
151
Table 7-16. Summary of qualitative uncertainty analysis for the air quality and health risk
characterization 153
July 2009 ix
-------
Table 7-17. Summary of descriptive statistics for the data removed using peak-to-mean ratio
criterion and the final 1-hour and 5-minute maximum SC>2 data set used to develop
PMRs 169
Table 7-18. The number and percent of days having multiple benchmark exceedances occurring
in the same day, using monitors reporting the 5-minute maximum SC>2 concentrations.
181
Table 8-1. Surface stations for the SO2 study areas 195
Table 8-2. Upper air stations for the 862 study areas 195
Table 8-3. Seasonal monthly assignments 196
Table 8-4. NLCD2001 land use characterization 199
Table 8-5. Summary of NEI emission estimates and total emissions used for dispersion
modeling in Greene County and St. Louis modeling domains 205
Table 8-6. Measured and modeled number of days in year 2002 with at least one 5-minute SC>2
benchmark exceedance at ambient monitors in Greene County 211
Table 8-7. Asthma prevalence rates by age and gender used in Greene County and St. Louis
modeling domains 223
Table 8-8. Population modeled in Greene County and St. Louis modeling domains 223
Table 8-9. Ventilation coefficient parameter estimates (&,) and residuals distributions (e\) from
Graham and McCurdy (2005) 227
Table 8-10. List of microenvironments modeled and calculation methods used 231
Table 8-11. Geometric means (GM) and standard deviations (GSD) for air exchange rates by
A/C type and temperature range 232
Table 8-12. Final parameter estimates of SO2 deposition distributions in several indoor
microenvironments modeled in APEX 233
Table 8-13. Comparison of benchmark levels, adjusted benchmark levels to just meet the current
standard, the benchmark level distribution percentiles, and the number of 5-minute
SC>2 benchmark exceedances at monitor 390350045 in Cuyahoga County for year
2002 240
Table 8-14. Exposure concentrations and adjusted potential health effect benchmark levels used
by APEX to simulate just meeting the current and potential alternative standards in the
Greene County and St Louis modeling domains 241
Table 8-15. Absolute difference in APEX exposure estimates for Greene County using either a
98th or 99th percentile form potential alternative standard at a 1-hour daily maximum
level of 200 ppb 244
Table 8-16. Absolute difference in APEX exposure estimates for St. Louis using either a 98th or
99th percentile form potential alternative standard at a 1-hour daily maximum level of
200 ppb 255
Table 8-17. States used to define five regions of the U.S. and characterize CHAD data diaries.
256
Table 8-18. Time spent outdoors by geographic region for children ages 5-17 based on CHAD
time-location-activity diaries 257
Table 8-19. Ranking of selected exposure locations using the modeled number of days with 5-
minute benchmark exceedances and the total emissions within 20 km of ambient
monitors 259
Table 8-20. Total SC>2 emissions and total port SC>2 emissions in the St. Louis and the 40
Counties used in the air quality characterization 261
July 2009
-------
Table 8-21. The top 40 counties with the greatest total port 862 emissions, including 862
emissions from ports in the St. Louis modeling domain 262
Table 8-22. SC>2 emission density the two exposure modeling domains and several counties
within selected U.S. Cities 263
Table 8-23. Residential A/C prevalence for housing units in several metropolitan locations in the
U.S. (AHS, 2008) 266
Table 8-24. Asthma prevalence rates for children in four regions of the U.S 267
Table 8-25. Asthma prevalence rates for adults in five regions of the U.S 267
Table 8-26. Summary of how variability was incorporated into the exposure assessment 268
Table 8-27. Summary of qualitative uncertainty analysis for the exposure assessment 270
Table 8-28. Distribution of APEX estimated annual average SC>2 exposures for simulated
individuals in the Greene County and St. Louis modeling domains 282
Table 8-29. Personal 862 exposure measurement data from the extant literature 283
Table 8-30. Percent of waking hours spent outdoors at an elevated activity level 292
Table 8-31. Number of multiple exceedances of potential health effect benchmark levels within
an hour 303
Table 9-1. Example calculation of the number of asthmatics in st. louis engaged in moderate or
greater exertion estimated to experience at least one lung function response (defined as
an increase in sRaw > 100%) associated with exposure to SC>2 concentrations just
meeting a 99th percentile, 1-hour 100 ppb standard 320
Table 9-2. Example calculation of number of occurrences of lung function response (defined as
an increase in sRaw > 100%), among asthmatics in St. Louis engaged in moderate or
greater exertion associated with exposure to SO2 concentrations that just meet a 99th
percentile 1-hour, 100 ppb standard 321
Table 9-3. Percentage of asthmatic individuals in controlled human exposure studies
experiencing SO2-induced decrements in lung function 324
Table 9-4. Number of asthmatics engaged in moderate or greater exertion estimated to
experience at least one lung function response associated with exposure to 862
concentrations under alternative air quality scenarios in a year 334
Table 9-5. Percent of asthmatics engaged in moderate or greater exertion estimated to
experience at least one lung function response associated with exposure to SO2
concentrations under alternative air quality scenarios in a year 335
Table 9-6. Number of occurrences (in hundreds) of a lung function response among asthmatics
engaged in moderate or greater exertion associated with exposure to SO2
concentrations under alternative air quality scenarios in a year 336
Table 9-7. number of asthmatic children engaged in moderate or greater exertion estimated to
experience at least one lung function response associated with exposure to SC>2
concentrations under alternative air quality scenarios in a year 337
Table 9-8. Percent of asthmatic children engaged in moderate or greater exertion estimated to
experience at least one lung function response associated with exposure to 862
concentrations under alternative air quality scenarios in a year 338
Table 9-9. number of occurrences (in hundreds) of a lung function response among asthmatic
children engaged in moderate or greater exertion associated with exposure to SO2
concentrations under alternative air quality scenarios in a year 339
Table 9-10. Explanation of labels on the x-axis of Figures 9-7 and 9-8 341
Table 9-11. Characterization of key uncertainties in the lung function response health risk
assessment for St. Louis and Greene County, Missouri 347
July 2009 xi
-------
Table 10-1 Ratios of 99th percentile 5-minute daily maximums to 99th percentile 24-hour average
and 1-hour daily maximum SC>2 concentrations for monitors reporting measured 5-
minute data from years 2004-2006 374
Table 10-2. 99th percentile 24-hour average SC>2 concentrations for 2004 given just meeting the
alternative 1-hour daily maximum 99th and 98th percentile standards analyzed in the air
quality assessment (note: concentrations in ppb) 376
Table 10-3. 2nd highest 24-hour average SO2 concentrations (i.e., the current 24-hour standard)
for 2004 given just meeting the alternative 1-hour daily maximum 99th and 98th
percentile standards analyzed in the air quality assessment (note: concentrations in
ppb) 378
Table 10-4. Annual average SO2 concentrations for 2004 given just meeting the alternative 99th
and 98th percentile 1-hour daily maximum standards analyzed in the air quality
assessment (note: concentrations in ppb) 379
Table 10-5. SO2 concentrations (ppb) corresponding to the 2nd-9th daily maximum and 98th/99th
percentile forms for alternative 1-hour daily maximum standards (2004-2006) 383
Table 10-6. Percent of counties that may be above the level of alternative standards (based on
years 2004-2006) 388
July 2009 xii
-------
LIST OF FIGURES
Number Page
Figure 1-1. Overview of the analyses described in this document and their interconnections 4
Figure 5-1. Effect estimates for U.S. all respiratory ED visit studies and associated 98th and 99th
percentile 1-hour daily maximum SC>2 levels 45
Figure 5-2. 24-hour effect estimates for U.S. asthma ED visit studies and associated 98th and
99th percentile 1-hour daily maximum SC>2 levels 46
Figure 5-3. 1-hour effect estimates for U.S. asthma ED visit studies and associated 98th and 99th
percentile 1-hour daily maximum SC>2 levels 47
Figure 5-4. 24-hour effect estimates for U.S. hospitalization studies and associated 98th and 99th
percentile 1-hour daily maximum SC>2 levels 48
Figure 5-5. Effect estimates for Canadian ED visits and hospitalization studies and associated
98th and 99th percentile 1-hour daily maximum SC>2 levels 49
Figure 6-1. Overview of analyses addressing exposures and risks associated with 5-minute peak
SO2 exposures. All three outputs are calculated considering current air quality, air
quality just meeting the current standards, and air quality just meeting potential
alternative standards. Note: this schematic was modified from Figure 1-1 54
Figure 6-2. Example of an hourly time-series of measured 1-hour and measured 5-minute
maximum SO2 concentrations 60
Figure 7-1. Location of the 98 monitors that reported 5-minute maximum SO2 concentrations
and comprising the first data analysis group 77
Figure 7-2. Location of the 809 monitors comprising the broader SO2 ambient monitoring
network (i.e., the second data analysis group) 80
Figure 7-3. Distribution of site-years of data considering monitoring objectives and scale:
monitors that reported 5-minute maximum SO2 concentrations (top) and the broader
SO2 monitoring network (bottom) 84
Figure 7-4. Distribution of site-years of data considering land-use and setting: monitors that
reported 5-minute maximum SO2 concentrations (top) and the broader SO2 monitoring
network (bottom) 85
Figure 7-5. The percent of total SO2 emissions of sources located within 20 km of ambient
monitors: monitors reporting 5-minute maximum SO2 concentrations (top) and the
broader SO2 monitoring network (bottom) 87
Figure 7-6. Distribution of the population residing within a 5 km radius of ambient monitors:
monitors reporting 5-minute maximum SO2 concentrations and the broader SO2
monitoring network 89
Figure 7-7. Comparison of hourly and 5-minute concentration COVs and GSDs at sixteen
monitors reporting all twelve 5-minute SO2 concentrations over multiple years of
monitoring 92
Figure 7-8. Cumulative density functions (CDFs) of hourly COVs (top) and GSDs (bottom) at
ambient monitors: monitors reporting 5-minute maximum SO2 concentrations and the
broader SO2 monitoring network 94
Figure 7-9. Peak-to-mean ratio (PMR) distributions for three COV and GSD variability bins and
seven 1-hour SO2 concentration stratifications 97
July 2009 xiii
-------
Figure 7-10. Distribution of total SO2 emissions (tpy) within 20 km of monitors by COV (left)
and GSD (right) concentration variability bins: monitors reporting 5-minute maximum
SO2 concentrations (top) and the broader SO2 monitoring network (bottom) 98
Figure 7-11. Percent of monitors within each concentration variability bin where the monitoring
objective was source-oriented: monitors reporting 5-minute maximum SO2
concentrations (solid) and the broader SO2 monitoring network (slotted) 99
Figure 7-12. Distribution of the measured number of daily 5-minute maximum SO2
concentrations above 200 ppb (left) and 400 ppb (right) in a year by hourly
concentration COV (top) and GSD (bottom) variability bins. Data were from the 98
ambient monitors reporting 5-minute maximum concentrations (471 site-years) 100
Figure 7-13. Comparison of measured daily maximum SO2 concentration percentiles in Beaver
County, PA for a high concentration year (1992) versus a low concentration year
(2007) at two ambient monitors (from Rizzo, 2009) 106
Figure 7-14. Distributions of hourly SO2 concentrations at five ambient monitors in Cuyahoga
County, as is (top) and air quality adjusted to just meet the current 24-hour SO2
standard (bottom), Year 2001 109
Figure 7-15. Locations of the 128 ambient monitors comprising the 40 County data set (i.e., the
third data analysis group) 113
Figure 7-16. Percent of monitors in each COV bin for the three data analysis groups: monitors
reporting 5-minute maximum SO2 concentrations, the broader SO2 monitoring
network, and SO2 monitors selected for detailed analysis in 40 counties 113
Figure 7-17. Example of empirically-based probability curves. The probability of a 5-minute
SO2 benchmark exceedance (P) was estimated by dividing the number of days with an
exceedance by the total number of days within each 1-hour daily maximum SO2
concentration bin 116
Figure 7-18. Example of logistic-modeled probability curves. The data used to generate these
modeled curves were the same used in generating the empirically-based curves in
Figure 7-17 119
Figure 7-19. The number of days per year with measured 5-minute maximum SO2
concentrations above potential health effect benchmark levels at 98 monitors given the
annual average SO2 concentration, 1997-2007 air quality as is. The level of the annual
average SO2NAAQS of 30 ppb is indicated by the dashed line 121
Figure 7-20. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 24-hour average SO2 concentration, using empirical
data (left) and a fitted log-probit model (right), 1997-2007 air quality as is. Both the
5-minute maximum and 24-hour SO2 concentrations were from measurements
collected at 98 ambient monitors and separated by population density 124
Figure 7-21. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SO2 concentration, using
empirical data (left) and a fitted log-probit model (right), 1997-2007 air quality as is.
Both the 5-minute maximum and 1-hour SO2 concentrations were from measurements
collected at 98 ambient monitors and separated by population density 126
Figure 7-22. The number of days per year with modeled daily 5-minute maximum SO2
concentrations above potential health effect benchmark levels at 809 ambient monitors
given the annual average SO2 concentration, 1997-2006 air quality as is. The level of
the annual average SO2 NAAQS of 30 ppb is indicated by the dashed line 129
July 2009 xiv
-------
Figure 7-23. Probability of daily 5-minute maximum SC>2 concentrations above potential health
effect benchmark levels given 24-hour average SC>2 concentrations, using empirical
data (left) and a fitted log-probit model (right), 1997-2006 air quality as is. The 5-
minute maximum SC>2 concentrations were modeled from 1-hour measurements
collected at 809 ambient monitors and then separated by population density 131
Figure 7-24. Probability of daily 5-minute maximum SC>2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SC>2 concentrations, using
empirical data (left) and a fitted log-probit model (right), 1997-2006 air quality as is.
The 5-minute maximum SC>2 concentrations were modeled from 1-hour measurements
collected at 809 ambient monitors and then separated by population density 132
Figure 7-25. The number of days per year with modeled 5-minute maximum SC>2 concentrations
above potential health effect benchmark levels per year at 128 ambient monitors in 40
selected counties given the annual average SC>2 concentration, 2001-2006 air quality
adjusted to just meet the current NAAQS. The level of the annual average SC>2
NAAQS of 30 ppb is indicated by the dashed line 135
Figure 7-26. The number of modeled daily 5-minute maximum SC>2 concentrations above 200
ppb per year at 128 ambient monitors in 40 selected counties given the annual average
SC>2 concentration, 2001-2006 air quality as is and that adjusted to just the current and
four potential alternative standards (text in graph indicate standard evaluated). The
level of the annual average 862 NAAQS of 30 ppb is indicated by the dashed line. 137
Figure 7-27. The number of days per year with modeled 5-minute maximum SC>2 concentrations
above benchmark levels given the 99th and 98th percentile forms, using the 40-county
air quality data set adjusted to just meet a 1-hour daily maximum level of 200 ppb. 139
Figure 7-28. The number of days per year with modeled 5-minute maximum SC>2 concentrations
above benchmark levels given the 99th and 98th percentile forms, using the 40-county
air quality data set adjusted to just meet a 1-hour daily maximum level of 100 ppb. 140
Figure 7-29. Probability of daily 5-minute maximum 862 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SC>2 concentrations, 2001-2006
air quality as is and that adjusted to just meet the current NAAQS. The 5-minute
maximum concentrations were modeled from 1-hour measurements collected at 128
ambient monitors from 40 selected counties and then separated by population density
within 5 km of monitors 143
Figure 7-30. Probability of daily 5-minute maximum SC>2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SC>2 concentrations, 2001-2006
air quality adjusted to just meet the current and each of the potential alternative
standards (99th percentile form). The 5-minute maximum concentrations were
modeled from 1-hour measurements collected at 128 ambient monitors from 40
selected counties, high-population density monitors 144
Figure 7-31. Temporal trends in the number of ambient monitors in operation per year for
monitors reporting both 5-minute and 1 -hour 862 concentrations 161
Figure 7-32. Temporal trends in the coefficient of variability (COV) for 5-minute maximum and
1-hour concentrations at the monitors that reported both 5-minute and 1-hour SC>2
concentrations. The number of monitors operating in each year is depicted in Figure
7-31 162
Figure 7-33. Comparison of measured daily maximum SC>2 concentration percentiles in
Allegheny County PA for one high concentration year (1998) versus a low
concentration years (2007) at five ambient monitors 167
July 2009 xv
-------
Figure 7-34. Distributions of annual average peak-to-mean ratios (PMRs) derived from the 98
monitors reporting both 5-minute maximum and 1-hour SC>2 concentrations, Years
1997 through 2007 170
Figure 7-35. Example histogram of peak-to-mean ratios (PMRs) compared with four fitted
distributions derived from monitors reporting the 5-minute maximum and 1-hour SC>2
concentrations (left) and the same PMRs compared with expected lognormal
percentiles (right). PMRs were derived from monitors with medium level variability
(COVbin = b) and 1-hour concentrations between 75 and 150 ppb (COVconcbin = 4).
171
Figure 7-36. Example of a measured peak-to-mean ratio (PMRs) distribution with the
percentiles of a fitted lognormal distribution. PMRs were derived from monitors with
high COV (COVbin = c) and 1-hour concentrations between 5 and 10 ppb
(COVconcbin = 2) 172
Figure 7-37. Comparison of observed and predicted number of daily benchmark exceedances in
a year at the 98 monitors reporting 5-minute maximum SO2 concentrations 175
Figure 7-38. Distributions of the maximum difference in the estimated mean number of
exceedances per site-year given 10 independent model runs (with 20 simulations per
run). Data used are from 40 county as is air quality (610 site-years). Box represents
the inner quartile range (IQR, or the 25th to 75th percentile), + indicates the mean,
whiskers are 1.5 times the IQR 177
Figure 8-1. General process flow used for SO2 exposure assessment 186
Figure 8-2. Modeling domain for Greene County Mo., along with identified emissions sources,
air quality receptors, ambient monitors, and meteorological station 191
Figure 8-3. Three county modeling domain for St. Louis, Mo., along with identified emissions
sources, air quality receptors, ambient monitors, and meteorological station 192
Figure 8-4. Derived best-fit non-point area source diurnal emission profile for the St. Louis
domain, compared to other possible profiles 205
Figure 8-5. Derived best-fit non-point area source diurnal emission profile for the Greene
County domain, compared to other possible profiles 206
Figure 8-6. Comparison of measured ambient monitor SO2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitors 290770026 and 29077032 in Greene County, Mo. Maximum 1-hour
concentration 212
Figure 8-7. Comparison of measured ambient monitor SO2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitors 290770040 and 29077041 in Greene County, Mo. Maximum 1-hour
concentration percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
213
Figure 8-8. Comparison of measured ambient monitor SO2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitor 290770037 in Greene County, Mo. Maximum 1-hour concentration
percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined 214
Figure 8-9. Comparison of measured ambient monitor SO2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitors 291890004 and 291890006 in St Louis, Mo. Maximum 1-hour
concentration percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
217
July 2009 xvi
-------
Figure 8-10. Comparison of measured ambient monitor SC>2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitors 291893001 and 291895001 in St Louis, Mo. Maximum 1-hour
concentration percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
218
Figure 8-11. Comparison of measured ambient monitor SC>2 concentration distribution and
diurnal profile with the modeled monitor receptor an d receptors within 4 km of
monitors 291897003 and 295100007 in St Louis, Mo. Maximum 1-hour
concentration percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
219
Figure 8-12. Comparison of measured ambient monitor SO2 concentration distribution and
diurnal profile with the modeled monitor receptor and receptors within 4 km of
monitor 295100086 in St Louis, Mo. Maximum 1-hour concentration percentile is
defined as 0.01 (or 100-99.99) because log(O) is undefined 220
Figure 8-13. Comparison of adjusted ambient monitoring concentrations or adjusted benchmark
level (dashed line) to simulate just meeting the current annual average standard at one
ambient monitor in Cuyahoga County for year 2002 239
Figure 8-14. Comparison of the upper percentile modeled daily 5-minute maximum SC>2
concentrations using either adjusted 1-hour ambient SO2 concentrations or an adjusted
benchmark level (with as is air quality) to simulate just meeting the current annual
standard at monitor 390350045 in Cuyahoga County for year 2002. Complete
distributions are provided in Figure 8-13 239
Figure 8-15. Number of all asthmatics (top) and asthmatic children (bottom) experiencing at
least one day with a 5-minute 862 exposure above selected benchmark levels in
Greene County, year 2002 air quality as is and adjusted to just meeting the current and
potential alternative standards 245
Figure 8-16. Percent of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute 862 exposure above selected benchmark levels in Greene
County, year 2002 air quality as is and adjusted to just meeting the current and
potential alternative standards 246
Figure 8-17. Number person days all asthmatics (top) and asthmatic children (bottom)
experience a 5-minute SO2 exposure above selected benchmark levels in Greene
County, year 2002 air quality as is and adjusted to just meeting the current and
potential alternative standards 247
Figure 8-18. Number of all asthmatics (top) and asthmatic children (bottom) experiencing at
least one day with a 5-minute 862 exposure above selected benchmark levels in St.
Louis, year 2002 air quality as is and adjusted to just meeting the current and potential
alternative standards 251
Figure 8-19. Percent of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute 862 exposure above selected benchmark levels in St. Louis,
year 2002 air quality as is and adjusted to just meeting the current and potential
alternative standards 252
Figure 8-20. Number person days all asthmatics (top) and asthmatic children (bottom)
experience a 5-minute 862 exposure above selected benchmark levels in St. Louis,
year 2002 air quality as is and adjusted to just meeting the current and potential
alternative standards 253
July 2009 xvii
-------
Figure 8-21. The frequency of estimated exposure level exceedances in indoor, outdoor, and
vehicle microenvironments given as is air quality (top), air quality adjusted to just
meeting the current standard (middle) and that adjusted to just meeting a 99th
percentile 1-hour daily maximum standard level of 150 ppb (bottom) in St. Louis.. 254
Figure 8-22. Means of weekly average personal Os exposures, measured and modeled (APEX),
Upland Ca. Figure obtained from EPA (2007d) 284
Figure 8-23. Daily average personal NO2 exposures, measured and modeled (APEX), Atlanta
Ga. Figure obtained from EPA (2008d) 285
Figure 8-24. Example comparison of estimated geometric mean and geometric standard
deviations of AER (h"1) for homes with air conditioning in several cities 298
Figure 8-25. Example of boot strap simulation results used in evaluating random sampling
variation of AER (h"1) distributions (data from cities outside California). Parameters
of the original distribution are given by the intersection of the two inner grid lines. 299
Figure 8-26. Example of boot strap simulation results used in evaluating random sampling
variation of AER (h"1) distributions (data from cities outside California). Parameters
of the original distribution are given by the intersection of the two inner grid lines. 300
Figure 8-27. Duration of time spent outdoors (in minutes) using all CHAD events 306
Figure 8-28. Percent of asthmatic children above given exposure level for two APEX
simulations: one using multiple peak concentrations in an hour, the other assuming a
single peak concentration. Continuous 5-minute monitoring data (ID 42007005, year
2002) were used as the air quality input 309
Figure 8-29. Percent of asthmatic children above given exposure level for two APEX
simulations: one using multiple peak concentrations in an hour, the other assuming a
single peak concentration. Continuous 5-minute monitoring data (ID 42007005, year
2005) were used as the air quality input 309
Figure 8-30. Frequency of exposure exceedances indoors for two APEX simulations: one using
multiple peak concentrations in an hour, the other assuming a single peak
concentration. Continuous 5-minute monitoring data (ID 42007005, year 2002) were
used as the air quality input 310
Figure 9-1. Major components of 5-minute peak lung function health risk assessment based on
controlled human exposure studies 318
Figure 9-2. Bayesian-estimated median exposure-response functions: increase in sRaw > 100%
for 5-Minute exposures of asthmatics under moderate or greater exertion 330
Figure 9-3. Bayesian-estimated median exposure-response functions: increase in sRaw > 200%
for 5-minute exposures of asthmatics under moderate or greater exertion 330
Figure 9-4. Bayesian-estimated median exposure-response functions: decrease in FEV1 331
Figure 9-5. Bayesian-estimated median exposure-response functions: decrease in FEV1 > 20%
for 5-minute exposures of asthmatics under moderate or greater exertion 331
Figure 9-6. Legend for Figures 9-7 and 9-8 showing total and contribution of risk attributable to
862 exposure ranges 341
Figure 9-7. Estimated percent of asthmatics experiencing one or more lung function responses
(defined as > 100% increase in sRaw) per year associated with short-term (5-minute)
exposures to 862 concentrations associated with alternative air quality scenarios -
total and contribution of 5-minute 862 exposure ranges (see Figure 9-6 for legend and
Table 9-10 for description of air quality scenarios included onx-axis) 342
Figure 9-8. Estimated percent of asthmatic children experiencing one or more lung function
responses (defined as > 100% increase in sRaw) per year associated with short-term
July 2009 xviii
-------
(5-minute) exposures to SC>2 concentrations associated with alternative air quality
scenarios - total and contribution of 5-minute SC>2 exposure ranges (see Figure 9-6 for
legend and Table 9-10 for description of air quality scenarios included on x-axis).. 343
Figure 10-1. Design value trends from 4 of the 54 sites analyzed in Thompson 2009 385
Figure 10-2. Boxplots of the distributions of standard deviations for alternative air quality
standard forms 386
July 2009 xix
-------
LIST OF ACRONYMS/ABBREVIATIONS
A/C Air conditioning
AER Air exchange rate
AERMOD American Meteorological Society (AMS)/EPA Regulatory Model
AHS American Housing Survey
APEX EPA's Air Pollutants Exposure model, version 4
ANOVA One-way analysis of variance
AQI Air Quality Index
ATL Atlanta Hartsfield airport
AQS EPA's Air Quality System
AQCD Air Quality Criteria Document
BRFSS Behavioral Risk Factor Surveillance System
CAA Clean Air Act
CAMD EPA's Clean Air Markets Division
CASAC Clean Air Scientific Advisory Committee
CDC Centers for Disease Control
CDF Cumulative density function
CFR Code of Federal Regulations
CHAD EPA's Consolidated Human Activity Database
Clev/Cinn Cleveland and Cincinnati, Ohio
CMSA Consolidated metropolitan statistical area
CO Carbon monoxide
COPD Chronic Obstructive Pulmonary Disease
COV Coefficient of Variation
C-R Concentration-Response
CTPP Census Transportation Planning Package
ED Emergency Department
EPA Environmental Protection Agency
EMS-HAP Emissions Modeling System for Hazardous Pollutants model
ER Emergency room
EOC Exposure of Concern
FEM Federal Equivalent Method
FEVi Forced expiratory volume in the first second
GM Geometric mean
GSD Geometric standard deviation
GST Glutathione S-transferase
ISCST Industrial Source Complex - Short Term dispersion model
ID Identification
ISA Integrated Science Assessment
ISH Integrated Surface Hourly Database
1ST Jefferson Street SEARCH monitor near Georgia Tech
km Kilometer
L95 Lower limit of the 95th confidence interval
July 2009
xx
-------
LOEL Lowest Observed Effect Level
m Meter
max Maximum
ME Microenvironment
med Median
min Minimum
MSA Metropolitan statistical area
NAAQS National Ambient Air Quality Standards
NAICS North American Industrial Classification System
NAMS National Ambient Monitoring Stations
NCEA National Center for Environmental Assessment
NEI National Emissions Inventory
NEM NAAQS Exposure Model
NCDC National Climatic Data Center
NHAPS National Human Activity Pattern Study
NHIS National Health Interview Survey
NC>2 Nitrogen dioxide
NOX Oxides of nitrogen
NWS National Weather Service
NYC New York City
NYDOH New York Department of Health
Os Ozone
OAQPS Office of Air Quality Planning and Standards
OR Odds ratio
ORD Office of Research and Development
ORIS Office of Regulatory Information Systems identification code
PAMS Photochemical Assessment Monitoring Stations
POC Parameter occurrence code
ppb Parts per billion
PEM Personal exposure measurements
PEN Penetration factor
PM Particulate matter
PMR Peak-to-mean ratio
ppm Parts per million
PRB Policy-Relevant Background
PROX Proximity factor
R2 R-square or the coefficient of determination
RE A Risk and Exposure Assessment
RECS Residential Energy Consumption Survey
RIU Rescue inhaler use
RR Relative risk
SAP Spatial allocation factors
SAS Statistical Analysis Software
SB Shortness of breath
SEARCH Southeast Aerosol Research and Characterization study (SEARCH) monitoring
SES Social-economic status
SIC Standard Industrial Code
July 2009
xxi
-------
SD
Se
SLAMS
SO2
SO3
SO4"
SOX
sRaw
tpy
TRIM
U95
UA
UC
UARG
USGS
Vs
Standard deviation
Standard error
State and Local Ambient Monitoring Stations
Sulfur dioxide
Sulfur trioxide
Sulfate
Oxides of Sulfur
Specific Airway Resistance
Tons per year
EPA's Total Risk Integrated Methodology
Upper limit of the 95th confidence interval
Urbanized area
Urban cluster
Utility Air Regulatory Group
United States Geological Survey
Ventilation rate
July 2009
xxn
-------
1. INTRODUCTION
The U.S. Environmental Protection Agency (EPA) is presently conducting a review of
the primary, health-based national ambient air quality standards (NAAQS) for sulfur dioxide
(802). Sections 108 and 109 of the Clean Air Act (The Act) govern the establishment and
periodic review of the NAAQS. These standards are established for pollutants that may
reasonably be anticipated to endanger public health and welfare, and whose presence in the
ambient air results from numerous or diverse mobile or stationary sources. The NAAQS are to
be based on air quality criteria, which are to accurately reflect the latest scientific knowledge
useful in indicating the kind and extent of identifiable effects on public health or welfare that
may be expected from the presence of the pollutant in ambient air. The EPA Administrator is to
promulgate and periodically review, at five-year intervals, primary (health-based) and secondary
(welfare-based) NAAQS for such pollutants. Based on periodic reviews of the air quality criteria
and standards, the Administrator is to make revisions in the criteria and standards and
promulgate any new standards as may be appropriate. The Act also requires that an independent
scientific review committee advise the Administrator as part of this NAAQS review process, a
function performed by the Clean Air Scientific Advisory Committee (CASAC).
The first step in the SC>2 NAAQS review was the development of an integrated review
plan. This plan presented the schedule for the review, the process for conducting the review, and
the key policy-relevant science issues that would guide the review. The final integrated review
plan was informed by input from CASAC, outside scientists, and the public. This plan was
presented in the Integrated Review Plan for the Primary National Ambient Air Quality Standards
for Sulfur Oxides (EPA, 2007a). It was made available to the public in October 2007 and can be
found at: http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_pd.html.
The second step in this review was a science assessment. A concise synthesis of the most
policy-relevant science was compiled into an Integrated Science Assessment (ISA). The ISA
was supported by a series of annexes that contained more detailed information about the
scientific literature. The final ISA to support this review of the 862 primary NAAQS was
presented in the Integrated Science Assessment for Oxides of Sulfur - Health Criteria, henceforth
referred to as the ISA (EPA, 2008a). This document was made available to the public in
July 2009
-------
September 2008 and can be found at:
http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_pd.html.
The third step in the primary SC>2 NAAQS review is a risk and exposure assessment
(REA) that describes exposures and characterizes risks associated with 862 emissions from
anthropogenic sources. The plan for conducting the risk and exposure assessment to support the
SC>2 primary NAAQS review was presented in the Sulfur Dioxide Health Assessment Plan:
Scope and Methods for Exposure and Risk Assessment, henceforth referred to as the Health
Assessment Plan (EPA, 2008b). This document was made available to the public in November
2007 and can be found at: http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_pd.html. The
first draft SC>2 REA was informed by comments from the public and CASAC on the Health
Assessment Plan, as well as the first and second drafts of the ISA for SOX. The first draft 862
REA developed estimates of human exposures and risks associated with recent ambient levels of
SO2 and levels that just met the current SO2 standards. The first draft REA was presented in the
Risk and Exposure Assessment to Support the Review of the SO2 Primary National Ambient Air
Quality Standards: First Draft. It was made available to the public in July 2008 and can be
found at: http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_rea.html
The second draft SC>2 REA was informed by comments from CASAC and the public on
the first draft REA, as well as findings and conclusions contained in the final ISA. This
document developed estimates of human exposures and risks associated with: (1) recent ambient
levels of 862, (2) levels that just met the current 862 standards, and (3) levels that just met
potential alternative standards: defined in terms of indicator, averaging time, form, and level.
This document also contained a draft policy assessment that addressed the adequacy of the
current SC>2 NAAQS and potential alternative standards. More specifically, the policy
assessment considered epidemiologic, human exposure, and animal toxicological evidence
presented in the ISA (EPA, 2008a), as well as the air quality, exposure, and risk characterization
results presented in the first draft REA, as they related to the adequacy of the current SC>2
NAAQS and potential alternative primary 862 standards (see Figure 1-1). The second draft REA
was presented in the Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standards: Second Draft. It was made available to the public in
March 2009 and can be found at: http://www.epa.gOv/ttn/naaqs/standards/so2/s so2 cr rea.html.
July 2009
-------
The final REA is this document, and has been informed by comments from CASAC and
the public on the second draft REA, as well as findings and conclusions contained in the final
ISA. The final REA further develops estimates of human exposures and risks associated with:
(1) recent ambient levels of 862, (2) levels that just meet the current 862 standards, and (3)
levels that just meet potential alternative standards. This document also contains a final policy
assessment (see Chapter 10). The final policy assessment will consider epidemiologic,
controlled human exposure, and animal toxicological evidence presented in the final ISA (EPA,
2008a), as well as the air quality, exposure, and risk characterization results presented in this
document, as they related to the adequacy of the current SC>2 NAAQS and potential alternative
primary 862 standards (Figure 1-1).
The final step in the review of the SC>2 NAAQS will be the rulemaking process. This
process will be informed by the risk and exposure information contained in the final REA, as
well the scientific evidence described in the final ISA. The rulemaking process will also take
into account CASAC advice and recommendations, as well as public comment on any policy
options under consideration. Notably, EPA is now under a consent decree to complete its review
of the SC>2 primary NAAQS by issuing a proposed rule no later than November 16, 2009 and a
final rule by June 2, 2010.
July 2009
-------
Evaluate Health Evidence in ISA
5-10 minute exposures
Identification of
potential 5-minute
health benchmark
values
Air Quality
analysis
Output 1: Number of days
per year 5-minute daily
maximum SO2
concentrations at
ambient monitors
exceed 5-minute potential
health benchmark values
Output 2: The probability
of a daily 5-minute
exceedance of
benchmark values given
a particular 24-hour
average, or 1 -hour daily
maximum SO2
concentration.
1
Characterization of human
clinical results
i
e
k
i >
*
i
Identify p
alternative ;
1
Characterizati
epidemioloc
3Y
i> 1-hour I
concentrations I
Dtential 4
Estimation of
exposure-response
function
Quantitative risk analysis
'fr I
/
Output. Estimates the
number and percent of
asthmatics at elevated
ventilation rates exposed
to 5-minute daily maximum
SO2 concentrations that
exceed 5-minute potential
health benchmark values
<•
— Air Quality
characterization
based on
epidemiology
Output: Number and
percentage of exposed
asthmatics that would
experience moderate or greater
lung function responses
/
]
Characterization
of toxicology
' i
exposures <
minutes to h
Risk-based considerations to
inform standard setting
Evidence based
considerations to inform
standard setting
Figure 1-1. Overview of the analyses described in this document and their interconnections
As mentioned above, an initial step in the review process was the development of an
integrated review plan. This plan identified policy relevant questions that would guide the
review of the SC>2 NAAQS. These questions are particularly important for the REA because they
provide a context for both evaluating health effects evidence presented in the ISA, as well as for
selecting the appropriate analyses for assessing exposure and risks associated with current
ambient 862 levels, 862 levels that just meet the current standards, and 862 levels that just meet
potential alternative standards. These policy relevant questions are:
• Has new information altered/substantiated the scientific support for the occurrence of
health effects following short- and/or long-term exposure to levels of SOX found in the
ambient air?
July 2009
-------
• Does new information impact conclusions from the previous review regarding the effects
of SOX on susceptible populations?
• At what levels of SOX exposure do health effects of concern occur?
• Has new information altered conclusions from previous reviews regarding the plausibility
of adverse health effects caused by SOX exposure?
• To what extent have important uncertainties identified in the last review been reduced
and/or have new uncertainties emerged?
• What are the air quality relationships between short-term and longer-term exposures
to SOX?
Additional questions will become relevant if the evidence suggests that revision of the
current standard might be appropriate. These questions are:
• Is there evidence for the occurrence of adverse health effects at levels of SOX different
than those observed previously? If so, at what levels and what are the important
uncertainties associated with that evidence?
• Do exposure estimates suggest that levels of concern for SOx-induced health effects will
occur with current ambient levels of SC>2, or with levels that just meet the current, or
potential alternative standards? If so, are these exposures of sufficient magnitude such
that the health effects might reasonably be judged to be important from a public health
perspective? What are the important uncertainties associated with these exposure
estimates?
• Do the evidence, the air quality assessment, and the risk/exposure assessment provide
support for considering different standard indicators, averaging times, or forms?
• What range of levels is supported by the evidence, the air quality assessment, and
risk/exposure assessment? What are the uncertainties and limitations in the evidence and
assessments?
1.1 HISTORY
1.1.1 History of the SO2 NAAQS
The first SO2 NAAQS was established in 1971. At that time, a 24-hour standard of 0.14
ppm, not to be exceeded more than one time per year, and an annual standard of 0.03 ppm were
judged to be both adequate and necessary to protect public health. The most recent review of the
SC>2 NAAQS was completed in 1996 and focused on the question of whether an additional short-
term standard (e.g., 5-minute) was necessary to protect against short-term, peak exposures.
Based on the scientific evidence, the Administrator judged that repeated exposures to 5-minute
peak SC>2 levels (> 600 ppb) could pose a risk of significant health effects for asthmatic
July 2009
-------
individuals at elevated ventilation rates. The Administrator also concluded that the likely
frequency of such effects should be a consideration in assessing the overall public health risks.
Based upon an exposure analysis conducted by EPA, the Administrator concluded that exposure
of asthmatics to 862 levels that could reliably elicit adverse health effects was likely to be a rare
event when viewed in the context of the entire population of asthmatics and therefore, did not
pose a broad public health problem for which a NAAQS would be appropriate. On May 22,
1996, EPA's final decision not to promulgate a 5-minute standard and to retain the existing 24-
hour and annual standards was announced in the Federal Register (61 FR 25566).
The American Lung Association and the Environmental Defense Fund challenged EPA's
decision not to establish a 5-minute standard. On January 30, 1998, the Court of Appeals for the
District of Columbia found that EPA had failed to adequately explain its determination that no
revision to the SO2 NAAQS was appropriate and remanded the decision back to EPA for further
explanation. Specifically, the court gave EPA the opportunity to provide additional rationale to
support the Agency judgment that 5-minute peaks of SO2 do not pose a public health problem
from a national perspective even though those peaks would likely cause adverse health impacts
in a subset of asthmatics. In response, EPA has collected and analyzed additional air quality data
focused on 5-minute concentrations of SO2. These air quality analyses conducted since the last
review will help inform the current review, which will answer the issues raised in the Court's
remand of the Agency's last decision.
1.1.2 Health Evidence from the Previous Review
The 1982 Air Quality Criteria Document (AQCD) for Particulate Matter and Sulfur
Oxides (EPA, 1982), and its subsequent addenda and supplement (EPA, 1986b, 1994a) presented
an evaluation of SO2 associated health effects primarily drawn from epidemiologic and human
clinical studies. In general, these documents identified adverse health effects that were likely
associated with both short- (generally hours to days), and long-term (months to years) exposures
to SO2 at concentrations present in the ambient mixture of air pollutants. Moreover, these
documents presented evidence for bronchoconstriction and respiratory symptoms in exercising
asthmatics following controlled exposures to 5-10 minute peak concentrations of SO2.
Evidence drawn from epidemiologic studies supported a likely association between 24-
hour average SO2 concentrations and daily mortality, aggravation of bronchitis, and small,
July 2009
-------
reversible declines in children's lung function (EPA 1982, 1994a). In addition, a few
epidemiologic studies found an association between respiratory symptoms and illnesses and
annual average 862 concentrations (EPA 1982, 1994a). However, it was noted that most of these
epidemiologic studies were conducted in years and cities where particulate matter (PM) counts
were also quite high, thus making it difficult to quantitatively determine whether the observed
associations were the result of SO2, PM, or a combination of both pollutants.
Evidence drawn from clinical studies exposing exercising asthmatics to <1000 ppb SC>2
for 5-10 minutes found that these types of SC>2 exposures evoked health effects that were similar
to those asthmatics would experience from other commonly encountered stimuli (e.g., exercise,
cold/dry air, psychological stress, etc. (EPA, 1994a). That is, there was an acute-phase response
characterized by bronchoconstriction and/or respiratory symptoms that occurred within 5-10
minutes of exposure but then subsided on its own within 1 to 2 hours. This acute-phase response
was followed by a short refractory period where the individual was relatively insensitive to
additional SC>2 challenges. Notably, the SC>2-induced acute-phase response was found to be
ameliorated by the inhalation of beta-agonist aerosol medications, and to occur without an
additional, often more severe, late-phase inflammatory response.
The 1994 supplement to the AQCD noted that of particular concern was the subset of
asthmatics in these clinical studies that appeared to be hyperresponsive (i.e., those experiencing
greater-than-average bronchoconstriction or respiratory symptoms at a given 862 concentration).
Thus, for a given concentration of 862, EPA estimated the number of asthmatics likely to
experience bronchoconstriction (and/or symptoms) of a sufficient magnitude to be considered a
health concern. At 600 to 1000 ppb SO2, EPA estimated that more than 25% of mild to moderate
exercising asthmatics would likely experience decrements in lung function distinctly exceeding
typical daily variations in lung function, or the response to commonly encountered stimuli (EPA,
1994a). Furthermore, the AQCD concluded that the severity of effects experienced at 600-1000
ppb was likely to be of sufficient concern to cause a cessation of activity, medication use, and/or
the possible seeking of medical attention. In contrast, at 200 - 500 ppb 862, it was estimated
that at most 10 - 20% of mild to moderate exercising asthmatics were likely to experience lung
function decrements larger than those associated with typical daily activity, or the response to
commonly encountered stimuli (EPA, 1994a).
July 2009
-------
1.1.3 Assessment from Previous Review
The risk and exposure assessment from the previous review of the SC>2 NAAQS
qualitatively evaluated both the existing 24-hour (0.14 ppm) and annual standards (0.03 ppm),
but primarily focused on whether an additional standard was necessary to protect against short-
term (e.g., 5-minute) peak exposures. Based on the human clinical data mentioned above, it was
judged that exposures to 5-minute SO2 levels at or above 600 ppb could pose an immediate
significant health risk for a substantial proportion of asthmatics at elevated ventilation rates (e.g.,
while exercising). Thus, EPA analyzed existing ambient monitoring data to estimate the
frequency of 5-minute peak concentrations above 500, 600, and 700 ppb, the number of repeated
exceedances of these concentrations, and the sequential occurrences of peak concentrations
within a given day (SAI, 1996). The results of this analysis indicated that in the vicinity of local
sources, several locations in the U.S. had a substantial number of 5-minute peak concentrations
at or above 600 ppb.
In addition to the ambient air quality analysis, the previous review also included several
annual exposure analyses that in general, combined SC>2 emission estimates from utility and non-
utility sources with exposure modeling to estimate the probability of exposure to short-term peak
SC>2 concentrations. The first such analysis conducted by the Agency estimated the number of 5-
minute exposures > 500 ppb associated with four selected coal-fired power utilities (EPA,
1986a). An expanded analysis sponsored by the Utility Air Regulatory Group (UARG)
considered the frequency of short-term exposure events that might result from the nationwide
operation of all power utility boilers (Burton et al., 1987). Additionally, the probability of peak
concentrations surrounding non-utility sources was the focus of another study conducted by the
Agency (Stoeckenius et al., 1990). The resultant combined exposure estimates based on these
early analyses indicated that between 0.7 and 1.8% of the total asthmatic population potentially
could be exposed one or more times annually, while outdoors at exercise, to 5-minute SC>2
concentrations > 500 ppb. It also was noted that the frequency of 5-minute exposures above the
health effect benchmark of 600 ppb, while not part of the analysis, would be anticipated to be
lower.
In addition to the early analyses mentioned above, two other analyses were considered in
the prior review. The first was an exposure assessment sponsored by the UARG (Rosenbaum et
al., 1992) that focused on emissions from fossil-fueled power plants. That study accounted for
July 2009
-------
the anticipated reductions in SC>2 emissions after implementation of the acid deposition
provisions (Title IV) of the 1990 Clean Air Act Amendments. This UARG-sponsored analysis
predicted that these emission reductions would result in a 42% reduction in the number of 5-
minute exposures to 500 ppb for asthmatic individuals (reducing the number of asthmatics
exposed from 68,000 down to 40,000) in comparison with the earlier Burton et al. (1987)
analysis. The second was a new exposure analysis submitted by the National Mining
Association (Sciences International, Inc. 1995) that reevaluated non-utility sources. In this
analysis, revised exposure estimates were provided for four of the seven non-utility source
categories by incorporating new emissions data and using less conservative modeling
assumptions in comparison to those used for the earlier Stoeckenius et al. (1990) non-utility
analysis. Significantly fewer exposure events (i.e., occurrence of 5-minute 500 ppb or greater
exposures) were estimated in this industry-sponsored revised analysis, decreasing the range of
estimated exposures for these four sources by an order of magnitude (i.e., from 73,000-259,000
short-term exposure events in the original analysis to 7,900-23,100 in the revised analysis).
1.2 SCOPE OF THE RISK AND EXPOSURE ASSESSMENT FOR THE
CURRENT REVIEW
1.2.1 Overview of the Risk and Exposure Assessment
The REA describes exposure and risks associated with recent ambient levels of SC>2,
levels that just meet the current SC>2 standards, and levels that just meet potential alternative
standards. This REA also contains a policy discussion regarding the adequacy of the current SC>2
NAAQS, and potential alternative primary standards. A concise overview of the information,
analyses, and policy discussion contained in this document is presented below.
Chapters 2-4 evaluate information presented in the ISA that is relevant for conducting an
exposure and risk assessment. This includes information on 1) human exposure to SO2; 2) at-risk
populations; and 3) health effects associated with short- and long-term exposures to SC>2.
Chapter 5 presents the rationale for the selection of the indicator, averaging time, forms, and
levels for the potential alternative standards that were assessed in the exposure and risk chapters
of the document. Specifically, these potential alternative standards are 99th percentile 1-hour
daily maximum SO2 levels of 50, 100, 150, 200, and 250 ppb, and 98th percentile 1-hour daily
maximum 862 levels of 200 ppb, and in some instances in the air quality analysis, 100 ppb. In
July 2009
-------
brief, the rationale takes into consideration both human exposure and epidemiologic evidence
from the ISA, as well as a qualitative analysis conducted by staff characterizing 98th and 99th
percentile 1-hour daily maximum 862 levels in cities and time periods corresponding to key U.S.
and Canadian hospitalization and ED visit studies for all respiratory causes and asthma (key
studies are identified in Table 5-5 of the ISA). Chapter 6 is an overview of the technical
analyses that are presented in the subsequent chapters of this document. This chapter also
presents the rationale for the selection of specific potential health benchmark values1 derived
from the human exposure literature.
Chapters 7-9 present the analytical portion of the document. Staff considered both
evidence of bronchoconstriction and respiratory symptoms from human exposure studies, as well
as CASAC advice on the first and second draft REA, and judged it appropriate to conduct a
series of three analyses to estimate risks associated with 5-minute 862 exposures ranging from
100-400 ppb in exercising asthmatics (see Figure 1-1 and Chapter 6). Chapter 7 presents an air
quality characterization that uses monitored and statistically estimated 5-minute ambient SC>2
concentrations as a surrogate for exposure. This analysis estimates the number of days per year
measured or statistically estimated 5-minute daily maximum SC>2 concentrations meet or exceed
the potential health benchmark values of 100, 200, 300 and 400 ppb This air quality analysis is
done under scenarios reflecting current air quality, air quality simulated to just meet the current
standards, and air quality simulated to just meet the potential alternative standards (i.e., 99th
percentile 1-hour daily maximum SO2 levels of 50, 100, 150, 200 and 250 ppb and an 98th
percentile 1-hour daily maximum 862 level of 200 ppb). Chapter 8 presents results from
exposure analysis case studies conducted in the St. Louis modeling domain (henceforth referred
to as St. Louis) and Greene County Missouri (MO). These analyses provide estimates of the
number and percent of asthmatics residing within 20 kilometers (km) of major SC>2 sources
experiencing 5-minute exposures to 100, 200, 300, and 400 ppb SC>2, while at elevated
ventilation rates under the air quality scenarios mentioned above (i.e., recent air quality, and air
quality adjusted to just meet the current and potential alternative standards). Chapter 9 is a
quantitative risk assessment that produces health risk estimates for the number and percent of
1 In general, potential health benchmark values are pollutant exposure levels that have consistently been shown to
induce adverse health effects in individuals participating in free-breathing human chamber studies.
July 2009 10
-------
exposed asthmatics (as determined by the exposure analysis; see Figure 1-1) that would
experience moderate or greater lung function responses under the air quality scenarios previously
described.
In addition to the technical analyses presented in Chapters 7-9, Chapter 10 integrates the
scientific evidence and the air quality, exposure, and risk information as they pertain to
informing decisions about the primary SC>2 NAAQS. More specifically, Chapter 10 considers
the epidemiologic, controlled human exposure, and animal toxicological evidence presented in
the ISA (EPA, 2008a), as well as the air quality, exposure, and risk characterization results
presented in this document, as they relate to the adequacy of the current SC>2 NAAQS and
potential alternative primary 862 standards.
1.2.2 Species of Sulfur Oxides Included in Analyses
The sulfur oxides include multiple gaseous (e.g., 862, 863) and particulate (e.g., sulfate)
species. In considering what species of sulfur oxides are relevant to the current review of the
SO2 NAAQS, we note that the health effects associated with particulate species of sulfur oxides
have been considered within the context of the Agency's review of the primary NAAQS for
particulate matter (PM). In the most recent review of the NAAQS for PM, it was determined
that size-fractionated particle mass, rather than particle composition, remains the most
appropriate approach for addressing ambient PM. This conclusion will be re-assessed in the
parallel review of the PM NAAQS; however, at present it would be redundant to also consider
effects of particulate sulfate in this review. Therefore, the current review of the 862 NAAQS
will focus on gaseous species of sulfur oxides and will not consider health effects directly
associated with particulate sulfur oxide species. Additionally, of the gaseous species, EPA has
historically determined it appropriate to specify the indicator of the standard in terms of SO2
because other gaseous sulfur oxides (e.g., SOs) are likely to be found at concentrations many
orders of magnitude lower than SC>2 in the atmosphere, and because most all of the health effects
and exposure information is for SC>2. The ISA has again found this to be the case, and therefore
this REA will use SC>2 as a surrogate for all gaseous sulfur oxides.
July 2009 11
-------
2. OVERVIEW OF HUMAN EXPOSURE
In order to help inform the air quality, exposure, and risk analyses presented in Chapters
7-9, staff has briefly summarized relevant human exposure information from the ISA. After
defining the concept of "integrated exposure," this chapter discusses major sources of SC>2
emissions. Characterizing these SC>2 sources helps identify the most relevant locations for
conducting air quality, exposure, and health risk analyses. This chapter then presents a
description of the SC>2 monitoring network, and discusses ambient levels of SC>2 associated with
1-hour, 24-hour, and annual averaging times. SC>2 concentrations associated with these
averaging times are relevant to the air quality, exposure, and risk analyses because the current
SC>2 standards have 24-hour and annual averaging times, and EPA is considering potential
alternative 1-hour averaging time standards (see section 5.3). Next, this chapter describes the
small subset of SO2 monitors that report 5-minute SO2 concentrations, as well as a broad
characterization of ambient 5-minute SC>2 levels (a more thorough discussion of these topics can
be found in Chapters 6 and 7). This discussion is particularly relevant to the analyses described
in this document because the potential health effect benchmarks and the outputs of the air
quality, exposure, and risk assessments are presented with respect to a 5-minute averaging time
(see section 6.2). More specifically, as previously described in section 1.2.1, an output of the air
quality analysis presented in Chapter 7 is the number of days per year measured, or statistically
estimated (see Chapter 6) 5-minute daily maximum 862 concentrations exceed 5-minute
potential health effect benchmark levels. Similarly, the output of the exposure analysis in
Chapter 8 is the number of exercising asthmatics exposed to 5-minute SC>2 concentrations above
benchmark levels. Outputs of the exposure analysis (i.e., the number of exercising asthmatics
exposed to 5-minute SC>2 concentrations above benchmark levels) are then used as inputs into the
quantitative risk assessment in Chapter 9 to estimate the number and percent of exposed
exercising asthmatics expected to experience a moderate or greater lung function response (see
Figure 6-1).
In addition to providing information relevant to the air quality, exposure, and risk
analyses, this Chapter also provides information relevant to the Chapter 4 health discussion and
the Chapter 10 policy assessment. That is, the current chapter highlights uncertainties involved
July 2009 12
-------
with using ambient SC>2 concentrations as a surrogate for personal exposure in epidemiologic
studies, as well as the ISA's conclusions on this topic.
2.1 BACKGROUND
The integrated exposure of a person to a given pollutant is the sum of the exposures over
all time intervals for all environments in which the individual spends time. People spend
different amounts of time in different microenvironments and each microenvironment is
characterized by different pollutant concentrations. There is a large amount of variability in the
time that individuals spend in different microenvironments, but on average people spend the
majority of their time (about 87%) indoors. Most of this time is spent at home with less time
spent in an office/workplace or other indoor locations (ISA, Figure 2-36). In addition, people
spend on average about 8% of their time outdoors and 6% of their time in vehicles. A potential
consequence of multiple sources of exposure or microenvironments is the exposure
misclassification that may result when total human exposure is not disaggregated between these
various microenvironments. In epidemiologic studies that rely on ambient pollutant levels as a
surrogate for exposure to ambient 862, such misclassification may obscure the true relationship
between ambient air pollutant exposures and health outcomes.
In addition to accounting for the times spent in different microenvironments, it is also
important to note the duration of exposure experienced. This is important because health effects
caused by long-term, low-level exposures may differ from those caused by relatively higher
shorter-term exposures.
2.2 SOURCES OF SO2
In order to estimate risks associated with SC>2 exposure, principle sources of the pollutant
must first be characterized because the majority of human exposures are likely to result from the
release of emissions from these sources. Anthropogenic SC>2 emissions originate chiefly from
point sources, with fossil fuel combustion at electric utilities (-66%) and other industrial
facilities (-29%) accounting for the majority of total emissions (ISA, section 2.1). Other
anthropogenic sources of SO2 include both the extraction of metal from ore as well as the
burning of high sulfur containing fuels by locomotives, large ships, and non-road diesel
equipment. Notably, almost the entire sulfur content of fuel is released as SC>2 or SOs during
July 2009 13
-------
combustion. Thus, based on the sulfur content in fuel stocks, oxides of sulfur emissions can be
calculated to a higher degree of accuracy than can emissions for other pollutants such as PM and
NO2 (ISA, section 2.1).
The largest natural sources of 862 are volcanoes and wildfires. Although 862 constitutes
a relatively minor fraction (0.005% by volume) of total volcanic emissions, concentrations in
volcanic plumes can be in the range of several to tens of ppm (thousands of ppb). Volcanic
sources of SC>2 in the U.S. are limited to the Pacific Northwest, Alaska, and Hawaii. Emissions
of SC>2 can also result from burning vegetation. The amount of SC>2 released from burning
vegetation is generally in the range of 1 to 2% of the biomass burned and is the result of sulfur
from amino acids being released as 862 during combustion.
2.3 BACKGROUND ON THE SO2 MONITORING NETWORK
The following section provides general background on the SC>2 monitoring network. A
more detailed description of this network can be found in Watkins (2009). The SC>2 monitoring
network was originally deployed to support implementation of the 862 NAAQS established in
1971. Despite the establishment of an SC>2 standard, uniform minimum monitoring requirements
for SC>2 monitoring did not appear until May 1979. From the time of the implementation of the
1979 monitoring rule through 2008, the SC>2 network has steadily decreased in size from
approximately 1496 sites in 1980 to the approximately 488 sites operating in 2008.
The 1979 monitoring rule established two categories of SC>2 monitoring sites: State and
Local Ambient Monitoring Stations (SLAMS) and the smaller set of National Ambient
Monitoring Stations (NAMS). No minimum requirements were established for SLAMS.
Minimum requirements (described below) were established for NAMS. The 1979 rule also
required that 862 only be monitored using Federal Reference Methods (FRMs) or Federal
Equivalent Methods (FEMs). The 1979 monitoring rule called for a range of number of sites in a
metropolitan statistical area (MSA) based both on population size and known concentrations
relative to the NAAQS (at that point in time; see Watkins, 2009).
In October 2006, EPA revised the monitoring requirements for SC>2 in light of the fact
that there was not an SC>2 non-attainment problem (Watkins, 2009). The 2006 rule eliminated
the minimum requirements for the number of SC>2 monitoring sites. The current SC>2 monitoring
rule, 40 CFR Part 58, Appendix D, section 4.4 states:
July 2009 14
-------
Sulfur Dioxide (SOi) Design Criteria.
(a) There are no minimum requirements for the number of SCh monitoring sites.
Continued operation of existing SLAMS SC>2 sites using FRM or FEM is required until
discontinuation is approved by the EPA Regional Administrator. Where SLAMS SC>2
monitoring is ongoing, at least one of the SLAMS 862 sites must be a maximum
concentration site for that specific area.
(b) The appropriate spatial scales for SC>2 SLAMS monitoring are the microscale, middle,
and possibly neighborhood scales. The multi-pollutant NCore sites can provide for
metropolitan area trends analyses and general control strategy progress tracking. Other
SLAMS sties are expected to provide data that are useful in specific compliance actions,
for maintenance plan agreements, or for measuring near specific stationary sources of
S02.
(1) Micro and middle scale - Some data uses associated with microscale and
middle scale measurements for 862 include assessing the effects of control strategies to
reduce concentrations (especially for the 3-hour and 24-hour averaging times) and
monitoring air pollution episodes.
(2) Neighborhood scale - This scale applies where there is a need to collect air
quality data as part of an ongoing 862 stationary source impact investigation. Typical
locations might include suburban areas adjacent to SO2 stationary sources for example, or
for determining background concentrations as part of these studies of population
responses to exposure to 862.
(c) Technical guidance in reference 1 of this appendix should be used to evaluate the
adequacy of each existing SO2 site, to relocate an existing site, or to locate new sites.
To ascertain what the current 862 network is addressing or characterizing, and in light of
the relatively recent removal of a specific SCh monitoring requirement, EPA reviewed some of
the SO2 network meta-data (Watkins, 2009). The data reviewed are those available from AQS
for calendar year 2008, for any monitors reporting data at any point during the year. The meta-
data fields are usually created by state and locals whenever a monitor or site is opened, moved,
or has a certain characteristic re-characterized. Often, EPA Regions consult with states and
locals on some of these metadata characteristics, but it is the responsibility of the state or local to
classify their own sites. With that, it should be noted that EPA must caveat such a review due to
the fact the AQS meta-data may have missing or 'old' meta-data field entries, as states and locals
do not have a routine or enforced process by which they must update or correct meta-data fields
(Watkins, 2009).
Monitoring Objective:
The monitoring objective meta-data field describes what the data from the monitor are
intended to characterize. The focus of the data presented is to show the nature of the network in
July 2009 15
-------
terms of its attempt to generally characterize health effects, source impacts, transport, or welfare
effects. In 2008, there were 488 862 monitors reporting data to AQS at some point during the
year. Any particular monitor can have multiple monitor objectives, however for this analysis
(see Watkins, 2009) we have selected one reported objective based on a hierarchy to represent an
individual monitor. The hierarchy used was to select, in order of priority: 1) source oriented, 2)
high concentration, 3) population exposure, or 4) general background, if they existed at a site
with multiple monitoring objectives. Table 2-1 presents the monitor objective distribution across
all SC>2 sites from the available AQS data. There are 12 categories of monitor objective for any
pollutant monitor within AQS. The "other" category is for sites likely addressing a state or local
need outside of the routine objectives, and the "unknown" category represents missing meta-
data. The six primary categories appropriate for use with SCh monitoring efforts stem directly
from categorizations of site types within the CFR. In 40 CFR Part 58 Appendix D, they are
defined as:
1. Sites located to determine the highest concentration expected to occur in the area
covered by the network (Highest Concentration).
2. Sites located to measure typical concentrations in areas of high population
(Population Exposure).
3. Sites located to determine the impact of significant sources or source categories
on air quality (Source Oriented).
4. Sites located to determine general background concentration levels (General
Background).
5. Sites located to determine the extent of regional pollutant transport among
populated areas; and in support of secondary standards (Regional Transport).
6. Sites located to measure air pollution impacts on visibility, vegetation damage, or
other welfare-based impacts (Welfare Related Impacts).
The remaining four categories available are a result of updating the AQS database. In the
more recent upgrade to AQS, the data handlers inserted the available site types for
Photochemical Assessment Monitoring Stations (PAMS) network as options for monitoring site
objectives. In our metadata review, three SO2 monitors have a listed monitoring objective that
EPA intended to be applied only to NOx or Os sites. As a result these three sites are presumably
co-located with a NOx or Os monitor with the same objective.
July 2009 16
-------
Measurement Scales
The spatial (measurement) scales are laid out in 40 CFR Part 58, Appendix D, Section 1
"Monitoring Objectives and Spatial Scales." This part of the regulation spells out what data
from a monitor can represent in terms of air volumes associated with area dimensions:
Microscale - 0 to 100 meters
Middle Scale - 100 to 500 meters
Neighborhood Scale - 500 meters to 4 kilometers
Urban Scale - 4 to 50 kilometers
Regional Scale - 50 kilometers up to 1000km
There are meta-data records for the 862 network to indicate what the measurement scale
of a particular monitor represents. In addition to the scales presented above, "industrial" scale
sites are an available option for characterizing 862 monitor sites in AQS. These "industrial"
scale sites are typically operated by industry, and are likely representative of the same scales that
are associated with sites having source oriented and high concentration monitoring objectives,
but we are unable to determine what spatial scale these monitors actually represent through AQS.
It is also noted that a monitor can only have one measurement scale, as opposed to the possibility
of a single monitor having multiple monitor objectives. Table 2-2 shows the measurement scale
distribution across all SC>2 sites from the available data in AQS of monitors reporting data in
2008.
Table 2-1. SO2 network monitoring objective distribution.
SO2 Monitoring
Objective
Population Exposure
Source Oriented
Highest Concentration
General Background
Regional Transport
Other
Max Precursor Impact (PAMS
Type 2 Site)
Welfare Related Impacts
Unknown
Totals:
Number of Monitoring
Objective Records
208
88
83
55
12
5
3
1
33
488
Percent Distribution
42.6 %
18.0%
17.0%
1 1 .3 %
2.5 %
1 .0 %
0.6 %
0.2 %
6.8 %
100%
July 2009
17
-------
Table 2-2. SO2 network distribution across measurement scales.
Measurement Scale
Microscale
Middle Scale
Neighborhood
Urban Scale
Regional Scale
Industrial Scale
Unknown
Totals:
Number of Measurement
Scale Records
1
35
309
61
41
6
35
488
Percent Distribution
0.2 %
7.2 %
63.3 %
12.5%
8.4 %
1 .2 %
7.2 %
100%
Urban/Rural Location Analysis
The US Census Bureau (http://www.census.gov/geo/www/ua/ua_2k.html) defines the
term "urban" as all territory, population, and housing units located within an urbanized area (UA) or an
urban cluster (UC). The Census bureau uses UA and UC boundaries to encompass densely settled
territory, which consists of:
• core census block groups or blocks that have a population density of at least 1,000 people
per square mile and
• surrounding census blocks that have an overall density of at least 500 people per square
mile
• Conversely, the Census Bureau's classification of "rural" consists of all territory,
population, and housing units located outside of UAs and UCs. Counties, metropolitan
areas, and the territory outside metropolitan areas, often are "split" between "urban" and
"rural" territory. A spatial analysis of the SC>2 monitors against the Census Bureau's
defined UAs and UCs shows that 63% of SC>2 monitors are in an "urban" setting and 37%
are in a "rural" setting.
2.4 AMBIENT LEVELS OF SO2
Since the integrated exposure to a pollutant is the sum of the exposures over all time
intervals for all environments in which the individual spends time, understanding the temporal
and spatial patterns of 862 levels across the U.S is an important component of conducting air
quality, exposure, and risk analyses. SO2 emissions and ambient concentrations follow a strong
east to west gradient due to the large numbers of coal-fired electric generating units in the Ohio
River Valley and upper Southeast regions. In the 12 CMSAs that had at least 4 SC>2 regulatory
monitors from 2003-2005, 24-hour average concentrations in the continental U.S. ranged from a
reported low of ~1 ppb in Riverside, CA and San Francisco, CA to a high of-12 ppb in
July 2009
18
-------
Pittsburgh, PA and Steubenville, OH (ISA, section 2.4.4). In addition, inside CMSAs from
2003-2005, the annual average 862 concentration was 4 ppb (ISA, Table 2-8). However, spikes
in hourly concentrations occurred; the mean 1-hour maximum concentration was 130 ppb, with a
maximum value of greater than 700 ppb (ISA, Table 2-8).
In addition to considering 1-hour, 24-hour, and annual SO2 levels in this document,
examining the temporal and spatial patterns of 5-minute peaks of SCh is also important given
that human clinical studies have demonstrated exposure to these peaks can result in adverse
respiratory effects in exercising asthmatics (see Chapter 4). Although the total number of SC>2
monitors across the continuous U.S. can vary from year to year, in 2006 there were
approximately 500 SC>2 monitors in the NAAQS monitoring network (ISA, section 2.5.2). State
and local agencies responsible for these monitors are required to report 1-hour average 862
concentrations to the EPA Air Quality System (AQS). However, a small number of sites, only
98 total from 1997 to 2007, and not the same sites in all years, voluntarily reported 5-minute
block average data to AQS (ISA, section 2.5.2). Of these, 16 reported all twelve 5-minute
averages in each hour for at least part of the time between 1997 and 2007. The remainder
reported only the maximum 5-minute average in each hour. When maximum 5-minute
concentrations were reported, the absolute highest concentration over the ten-year period
exceeded 4000 ppb, but for all individual monitors, the 99th percentile was below 200 ppb (ISA,
section 2.5.2). Medians from these monitors reporting data ranged from 1 ppb to 8 ppb, and the
average for each maximum 5-minute level ranged from 3 ppb to 17 ppb. Delaware,
Pennsylvania, Louisiana, and West Virginia had mean values for maximum 5-minute data
exceeding 10 ppb (ISA, section 2.5.2). Among aggregated within-state data for the 16 monitors
from which all 5-minute average intervals were reported, the median values ranged from 1 ppb to
5 ppb, and the means ranged from 3 ppb to 11 ppb (ISA, section 2.5.2). The highest reported
concentration was 921 ppb, but the 99th percentile values for aggregated within-state data were
all below 90 ppb (ISA, section 2.5.2).
EPA has generally conducted NAAQS risk assessments that focus on the risks associated
with levels of a pollutant that are in excess of policy relevant background (PRB). Policy relevant
background levels are defined as concentrations of a pollutant that would occur in the U.S. in the
absence of anthropogenic emissions in continental North America (defined here as the United
July 2009 19
-------
States, Canada, and Mexico). However, throughout much of the United States, SC>2 PRB levels
are estimated to be at most 30 parts per trillion and contribute less than 1% to present day 862
concentrations (ISA, section 2.5.3). We note that in the Pacific Northwest and Hawaii, PRB
concentrations can be considerably higher due to geogenic activity (e.g., volcanoes); in these
areas, PRB can account for 70-80% of total SO2 concentrations (ISA, section 2.5.3). Since we
do not plan on conducting SC>2 risk assessments in areas with high background SC>2 levels due to
natural sources, and the contribution of PRB is negligible in all other areas, EPA is addressing
the risks associated with monitored and/or modeled ambient SC>2 levels without regard to PRB
levels.
2.5 RELATIONSHIP OF PERSONAL EXPOSURE TO AMBIENT
CONCENTRATIONS
To help inform the evaluation of the epidemiologic evidence in Chapter 4 and the
evidence-based considerations presented in Chapter 10, this section discusses the relationship of
personal SC>2 exposure to ambient SC>2 concentrations. Many epidemiologic studies rely on
measures of ambient SC>2 concentrations as surrogates for personal exposure to ambient SC>2.
Thus, it is important to consider the potential sources of error that are associated with using SC>2
measured by ambient monitors as a surrogate for personal exposure to ambient SC>2. Key aspects
related to this issue include: (1) ambient and personal sampling issues, (2) the spatial variability
of ambient 862 concentrations, and (3) the relationship between ambient concentrations and
personal exposures as influenced by exposure factors (e.g., indoor sources).
Only a limited number of studies have focused on the relationship between personal
exposure and ambient concentrations of SC>2, in part because ambient SC>2 levels have declined
markedly over the past few decades. Indoor and outdoor SC>2 concentrations are often below
detection limits for personal samplers2 and in these situations, the ISA notes that associations
between ambient concentrations and personal exposures are inadequately characterized (ISA,
section 2.6.3.2). However, in studies with personal measurements above detection limits, the
ISA states that a reasonably strong association was observed between personal 862 exposure and
ambient concentrations (Brauer et al., 1989; Sarnat et al., 2006; described in ISA section 2.6.3.2).
2 The lower limit of detection of personal samplers is -60 ppb for 1-hour and ~5 ppb for 24-hour. A discussion of
personal sampler detection limits can be found in section 2.6.2 of the ISA.
July 2009 20
-------
In addition, the ISA notes that no study has examined the relationship between concentrations
measured at ambient monitors and the community average exposure: a relationship that is more
relevant than that of ambient concentration to personal exposure for community time-series
studies (ISA, section 5.3).
Because epidemiologic studies rely on ambient SO2 measurements at fixed site monitors,
there is concern about the extent to which instrument error could influence the results of these
studies. That is, the SC>2 monitoring network was designed and put into place when SC>2
concentrations were considerably higher, and thus, well within the standard monitor's limits of
detection. However, SC>2 concentrations have fallen considerably over the years and are
currently at, or very near these monitors' lower limit of detection (~3 ppb). As a result, greater
relative error is most often observed at lower ambient concentrations compared to less frequent
higher concentrations. Notably, the ISA states that it is unclear how instrument error will
influence the effect estimates of epidemiologic studies relying on these measurements (ISA,
section 2.6.4.1). As an additional matter, staff notes that the lower detection limit of these
monitors is not considered problematic with respect to determining attainment of SC>2 NAAQS
because the current 24-hour and annual standards, as well as the potential alternative 1-hour
daily maximum standards, are all well within the detection limits of the SC>2 monitoring network.
Uncertainty in epidemiologic studies is also associated with the spatial and temporal
variation of 862 across communities. The ISA finds that site-to-site correlations of 862
concentrations among monitors in U.S. cities ranges from very low to very high (ISA, section
2.6.4.1; ISA, Table 2-9). This suggests that at any given time, 862 concentrations at individual
monitoring sites may not highly correlate with the average SC>2 concentration in the community.
This could be the result of local sources (e.g., power plants) causing an uneven spatial
distribution of SC>2, monitors being sited to represent concentrations near local sources, or effects
related to terrain or weather (ISA, section 2.6.4.1). However, this type of error is not thought to
bias community time-series results in a positive direction because it generally tends to reduce,
rather than increase, effect estimates.
In epidemiologic studies, since people spend most of their time indoors, there is also
uncertainty in the relationship between ambient concentrations measured by local monitors and
actual personal exposure related to ambient sources. That is, the presence of indoor or
July 2009 21
-------
nonambient sources of SC>2 could complicate the interpretation of associations between personal
exposure and ambient 862 in exposure studies. Sources of indoor 862 are associated with the
use of sulfur-containing fuels, with higher levels expected when emissions are poorly vented. In
the U.S., the contribution of indoor sources is not thought to be a major contributor to overall
SO2 exposure because the only known indoor source is kerosene heaters and their use is not
thought to be widespread (ISA, section 2.6.4.1).
The ISA concludes that exposure error caused by using ambient concentrations of SC>2 as
a surrogate for exposure to ambient SC>2 is a source of uncertainty for epidemiologic studies.
However, in community time-series and short-term panel epidemiologic studies, exposure error
would tend to bias the effect estimate towards the null (ISA, section 2.6.4.4. and 5.3).
2.6 KEY OBSERVATIONS
• SC>2 emissions and ambient concentrations follow a strong east to west gradient due to the
large numbers of coal-fired electric generating units in the Ohio River Valley and upper
Southeast regions.
• In the 12 CMSAs that had at least 4 SO2 regulatory monitors from 2003-2005, 24-hour
average concentrations in the continental U.S. ranged from a reported low of ~1 ppb in
Riverside, CA and San Francisco, CA to a high of-12 ppb in Pittsburgh, PA and
Steubenville, OH.
• Inside CMSAs from 2003-2005, the annual average SO2 concentration was 4 ppb.
• Inside CMSAs from 2003-2005, the mean 1-hour maximum concentration was 130 ppb,
with a maximum value of greater than 700 ppb.
• A small number of sites, only 98 total from 1997 to 2007, and not the same sites in all
years—voluntarily reported 5-minute block average data to AQS. Of these, 16 reported
all twelve 5-minute averages in each hour, while the remainder reported only the
maximum 5-minute average in each hour.
• Throughout much of the United States, SO2 PRB levels are estimated to be at most 30
parts per trillion and contribute less than 1% to present day SO2 concentrations.
• The ISA concludes that exposure error caused by using ambient concentrations of SO2 as
a surrogate for exposure to ambient SO2 is a source of uncertainty for epidemiologic
studies. However, in community time-series and short-term panel epidemiologic studies,
exposure error would tend to bias the effect estimate towards the null. Thus, results of
these studies can be used, in part, to evaluate the adequacy of the current and potential
alternative SO2 standards (see Chapter 10)
July 2009 22
-------
3. AT RISK POPULATIONS
3.1 OVERVIEW
Interindividual variation in human responses to air pollutants indicates that some
subpopulations are at increased risk for the detrimental effects of ambient exposure to SC>2. The
NAAQS are intended to provide an adequate margin of safety for both general populations and
sensitive subpopulations, or those subgroups potentially at increased risk for health effects in
response to ambient air pollution. To facilitate the identification of subpopulations at the
greatest risk for SCVrelated health effects, studies have identified factors that contribute to the
susceptibility and/or vulnerability of an individual to 862. Susceptible individuals are broadly
defined as those with a greater likelihood of an adverse outcome given a specific exposure in
comparison with the general population (American Lung Association, 2001). The susceptibility
of an individual to SC>2 can encompass a multitude of factors which represent normal
developmental phases (e.g., age) or biologic attributes (e.g., gender); however, other factors (e.g.,
socioeconomic status (SES)) may influence the manifestation of disease and also increase an
individual's susceptibility (American Lung Association, 2001). In addition, subpopulations may
be vulnerable to 862 in response to an increase in their exposure during certain windows of life
(e.g., childhood or old age) or as a result of external factors (e.g., SES) that contribute to an
individual being disproportionately exposed to higher concentrations than the general population.
It should be noted that in some cases specific factors may affect both the susceptibility and
vulnerability of a subpopulation to SC>2. For example, a subpopulation that is characterized as
having low SES may have less access to healthcare resulting in the manifestation of a disease,
which increases their susceptibility to SC>2, but they may also reside in a location that results in
exposure to higher concentrations of SC>2, increasing their vulnerability to SC>2.
To examine whether 862 differentially affects certain subpopulations, stratified analyses
are often conducted in epidemiologic investigations to identify the presence or absence of effect
modification. A thorough evaluation of potential effect modifiers may help identify
subpopulations that are more susceptible and/or vulnerable to SO2. These analyses require the
proper identification of confounders and their subsequent adjustment in statistical models, which
helps separate a spurious, from a true causal association. Although the design of toxicological
and human clinical studies does not allow for an extensive examination of effect modifiers, the
July 2009 23
-------
use of animal models of disease and the study of individuals with underlying disease or genetic
polymorphisms do allow for comparisons between subgroups. Therefore, the results from these
studies, combined with those results obtained through stratified analyses in epidemiologic
studies, contribute to the overall weight of evidence for the increased susceptibility and
vulnerability of specific subpopulations to SO2. Those groups identified in the ISA to be
potentially at greater risk of experiencing an adverse health effect from SC>2 exposure are
described in more detail below.
3.2 PRE-EXISTING RESPIRATORY DISEASE
In human clinical studies, asthmatics have been shown to be more responsive to the
respiratory effects of SC>2 exposure than healthy non-asthmatics. While SO2-attributable
decrements in lung function have generally not been demonstrated at concentrations < 1000 ppb
in non-asthmatics, statistically significant increases in respiratory symptoms and decreases in
lung function have consistently been observed in exercising asthmatics following 5 to 10 minute
SC>2 exposures at concentrations ranging from 400-600 ppb (ISA, section 4.2.1.1). Moderate or
greater SCVinduced decrements in lung function have also consistently been observed at 862
concentrations ranging from 200-300 ppb in some asthmatics. The ISA also notes that a number
of epidemiologic studies have reported respiratory morbidity in asthmatics associated with SC>2
exposure (ISA 4.2.1.1). For example, numerous epidemiologic studies have observed positive
associations between ambient SC>2 concentrations and ED visits and hospitalizations for asthma
(ISA section 4.2.1.1). Overall, the ISA concludes that epidemiologic and controlled human
exposure studies indicate that individuals with pre-existing respiratory diseases, particularly
asthma, are at greater risk than the general population of experiencing SCh-associated health
effects (ISA, section 4.2.1.1).
3.3 GENETICS
The ISA notes that a consensus now exists among scientists that the potential for genetic
factors to increase the risk of experiencing adverse health effects due to ambient air pollution
merits serious consideration. Several criteria must be satisfied in selecting and establishing
useful links between polymorphisms in candidate genes and adverse respiratory effects. First,
the product of the candidate gene must be significantly involved in the pathogenesis of the effect
July 2009 24
-------
of interest, which is often a complex trait with many determinants. Second, polymorphisms in
the gene must produce a functional change in either the protein product or in the level of
expression of the protein. Third, in epidemiologic studies, the issue of effect modification by
other genes or environmental exposures must be carefully considered (ISA section 4.2.2).
While many studies have examined the association between genetic polymorphisms and
susceptibility to air pollution in general, only one study has specifically examined the effects of
SC>2 exposure on genetically distinct subpopulations. Winterton et al. (2001) found a significant
association between SO2-induced decrements in Forced Expiratory Volume in the first second
(FEVi) and the homozygous wild-type allele in the promoter region of Tumor Necrosis Factor-a
(TNF- a; AA, position -308). However, the ISA concluded that the overall body of evidence was
too limited to reach a conclusion regarding the effects of 862 exposure on genetically distinct
subpopulations at this time.
3.4 AGE
The ISA identifies children (i.e., <18 years of age) and older adults (i.e., >65 years of
age) as groups that are potentially at greater risk of experiencing SCVassociated adverse health
effects. In children, the developing lung is prone to damage from environmental toxicants as it
continues to develop through adolescence. The biological basis for increased risk in the elderly
is unknown, but one hypothesis is that it may be related to changes in antioxidant defenses in the
fluid lining the respiratory tract. The ISA found a number of epidemiologic studies that observed
increased respiratory symptoms in children associated with increasing SC>2 concentrations. In
addition, several studies have reported that the excess risk estimates for ED visits and
hospitalizations for all respiratory causes, and to a lesser extent asthma, associated with a 10-ppb
increase in 24-hour average 862 concentrations were higher for children and older adults than for
all ages together (ISA, section 4.2.3). However, the ISA also notes that the evidence from
controlled human exposure studies does not suggest that adolescents are either more or less at
risk than adults to the respiratory effects of SC>2, but rather adolescents may experience similar
respiratory effects at a given exposure concentration (ISA, sections 3.1.3.5 and 4.2.3). Overall,
the ISA finds that compared to the general population, there is limited evidence to suggest that
children and older adults are at greater risk of experiencing SCVassociated health effects (ISA,
section 4.2.3).
July 2009 25
-------
3.5 TIME SPENT OUTDOORS
Outdoor SC>2 concentrations are generally much higher than indoor concentrations. Thus,
the ISA notes that individuals who spend a significant amount of time outdoors are likely at
greater risk of experiencing SCVassociated health effects than those who spend most of their
time indoors (ISA section 4.2.5).
3.6 VENTILLATION RATE
Controlled human exposure studies have demonstrated that decrements in lung function
and respiratory symptoms occur at significantly lower 862 exposure levels in exercising subjects
compared to resting subjects. As ventilation rate increases, breathing shifts from nasal to
oronasal, thus resulting in greater uptake of 862 in the tracheobronchial airways due to the
diminished absorption of SO2 in the nasal passages. Therefore, individuals who spend a
significant amount of time at elevated ventilation rates (e.g. while playing, exercising, or
working) are expected to be at greater risk of experiencing SO2-associated health effects (ISA
section 4.2.5).
3.7 SOCIOECONOMIC STATUS
There is limited evidence that increased risk to SC>2 exposure is associated with lower
SES (ISA section 4.2.5). Finkelstein et al. (2003) found that among people with below-median
income, the relative risk for above-median exposure to 862 was 1.18(95%CI: 1.11, 1.26); the
corresponding relative risk among subjects with above-median income was 1.03 (95% CI: 0.83,
1.28). However, the ISA concludes that there is insufficient evidence to reach a conclusion
regarding SES and exposure to 862 at this time (ISA section 4.2.5).
3.8 NUMBER OF AT RISK INDIVIDUALS
Considering the size of the groups mentioned above, large proportions of the U.S.
population are likely to have a relatively high risk of experiencing SO2-related health effects. In
the United States, approximately 10% of adults and 13% of children have been diagnosed with
asthma. Notably, the prevalence and severity of asthma is higher among certain ethnic or racial
groups such as Puerto Ricans, American Indians, Alaskan Natives, and African Americans (ISA
for NOX, section 4.4). Furthermore, a higher prevalence of asthma among persons of lower SES
and an excess burden of asthma hospitalizations and mortality in minority and inner-city
July 2009 26
-------
communities have been observed. In addition, population groups based on age comprise
substantial segments of individuals that may be potentially at risk for SCVrelated health impacts.
Based on U.S. census data from 2000, about 72.3 million (26%) of the U.S. population are under
18 years of age, 18.3 million (7.4%) are under 5 years of age, and 35 million (12%) are 65 years
of age or older. There is also concern for the large segment of the population that is potentially
at risk to SO2-related health effects because of increased time spent outdoors at elevated
ventilation rates (those who work or play outdoors). Overall, the considerable size of the
population groups at risk indicates that exposure to ambient SC>2 could have a significant impact
on public health in the United States.
3.9 KEY OBSERVATIONS
• The susceptibility of an individual to SC>2 can encompass a multitude of factors which
represent normal developmental phases (e.g., age) or biologic attributes (e.g., gender);
however, other factors (e.g., SES) may influence the manifestation of disease and also
increase an individual's susceptibility.
• Subpopulations may be vulnerable to 862 in response to an increase in their exposure
during certain windows of life (e.g., childhood or old age) or as a result of external
factors (e.g., SES) that contribute to an individual being disproportionately exposed to
higher concentrations than the general population.
• In some cases specific factors may affect both the susceptibility and vulnerability of a
subpopulation to 862.
• The ISA concludes that individuals with pre-existing respiratory disease are likely at
greater risk than the general population of experiencing SCVassociated health effects.
• Epidemiologic studies suggest that children and older adults may be at greater risk of
experiencing SO2-associated health effects. However, the evidence from controlled
human exposure studies suggests that adolescents are neither more nor less at risk than
adults.
• People who spend extended periods of time outdoors and/or at elevated ventilation rates
are likely at increased risk of experiencing adverse health effects from 862 exposure.
• Large proportions of the U.S. population are likely to be at increased risk of experiencing
SO2-related health effects. Thus, exposure to ambient SC>2 could have a significant
impact on public health in the United States
July 2009 27
-------
4. INTEGRATION OF HEALTH EVIDENCE
4.1 INTRODUCTION
The ISA, along with its annexes, integrates newly available epidemiologic, human
clinical, and animal toxicological evidence with consideration of key findings and conclusions
from prior reviews to draw conclusions about the relationship between short- and long-term
exposure to SC>2 and numerous human health categories. For these health effects, the ISA
characterizes judgments about causality with a hierarchy (for discussion see ISA section 1.3.7)
that contains the following five levels:
• Sufficient to infer a causal relationship
• Sufficient to infer a likely causal relationship (i.e., more likely than not)
• Suggestive but not sufficient to infer a causal relationship
• Inadequate to infer the presence or absence of a causal relationship
• Suggestive of no causal relationship
The ISA notes that these judgments about causality are informed by a series of aspects of
causality that are based on those set forth by Sir Austin Bradford Hill in 1965 (ISA section
1.3.6). These aspects include strength of the observed association, availability of experimental
evidence, consistency of the observed association, biological plausibility, coherence of the
evidence, temporal relationship of the observed association, and the presence of an exposure-
response relationship. A summary of each of the five levels of the hierarchy is provided in Table
1-2 of the ISA, which has also been included below (Table 4-1).
July 2009 28
-------
Table 4-1. Weight of evidence for causal determination.
RELATIONSHIP
DESCRIPTION
Causal relationship
Evidence is sufficient to conclude that there is a causal relationship
between relevant pollutant exposures and the health outcome. That is, a
positive association has been observed between the pollutant and the
outcome in studies in which chance, bias, and confounding could be ruled
out with reasonable confidence. Evidence includes, for example, controlled
human exposure studies; or observational studies that cannot be explain
by plausible alternatives or are supported by other lines of evidence (e.g.
animal studies or mechanism of action information). Evidence includes
replicated and consistent high-quality studies by multiple investigators.
Likely to be a causal
relationship
Evidence is sufficient to conclude that a causal relationship is likely to exist
between relevant pollutant exposures and the health outcome but
important uncertainties remain. That is, a positive association has been
observed between the pollutant and the outcome in studies in which
chance and bias can be ruled out with reasonable confidence but potential
issues remain. For example: a) observational studies show positive
associations but copollutant exposures are difficult to address and/or other
lines of evidence (controlled human exposure, animal, or mechanism of
action information) are limited or inconsistent; or b) animal evidence from
multiple studies, sex, or species is positive but limited or no human data
are available. Evidence generally includes replicated and high-quality
studies by multiple investigators.
Suggestive of a causal
relationship
Evidence is suggestive of a causal relationship between relevant pollutant
exposures and the health outcome, but is limited because chance, bias
and confounding cannot be ruled out. For example, at least one high-
quality study shows a positive association but the results of other studies
are inconsistent.
Inadequate to infer a causal
relationship
Evidence is inadequate to determine that a causal relationship exists
between relevant pollutant exposures and the health outcome. The
available studies are of insufficient quantity, quality, consistency or
statistical power to permit a conclusion regarding the presence or absence
of an association between relevant pollutant exposure and the outcome.
Suggestive of no causal
relationship
Evidence is suggestive of no causal relationship between relevant
pollutant exposures and the health outcome Several adequate studies,
covering the full range of levels of exposure that human beings are known
to encounter and considering sensitive subpopulations, are mutually
consistent in not showing a positive association between exposure and the
outcome at any level of exposure. The possibility of a very small elevation
in risk at the levels of exposure studied can never be excluded.
Considering the framework presented in Table 4-1, the ISA concludes that there is
sufficient evidence to infer a causal relationship between respiratory morbidity and short-term
exposure to 862 (ISA, section 5.2). The ISA bases this conclusion on the consistency,
July 2009
29
-------
coherence, and plausibility of findings observed in controlled human exposure studies of 5-10
minutes, epidemiologic studies mostly using 24-hour average concentrations, and animal
toxicological studies using exposures of minutes to hours (ISA, section 5.2). The evidence of an
association between 862 exposure and other health categories is judged to be less convincing, at
most suggestive but not sufficient to infer a causal relationship. Key conclusions from the ISA
are summarized below and are described in greater detail in Table 5-3 of the ISA.
• Sufficient to infer a causal relationship:
o Short-Term Exposure to SC>2 and Respiratory Morbidity
• Suggestive but not sufficient to infer a causal relationship:
o Short-Term Exposure to SC>2 and Mortality
• Inadequate to infer the presence or absence of a causal relationship
o Short-Term Exposure to SC>2 and Cardiovascular Morbidity;
o Long-Term Exposure to SC>2 and Respiratory Morbidity;
o Long-Term Exposure to 862 and Other Morbidity;
o Long-Term Exposure to 862 and Mortality
The integrated health discussion in this chapter will focus on health effect categories for
which the ISA finds a causal or likely causal relationship, as these effect categories are the basis
for the potential health effect benchmarks and quantitative health risk assessment included in
Chapters 7 through 9 of this document. As a result, this chapter will present an integrated
discussion of the health evidence related to respiratory morbidity following short-term exposure
to SO2. This is because respiratory morbidity is the only health effect category found by the ISA
to have either a causal or likely causal association with 862. The focus on health effect
categories with the strongest evidence for purposes of the quantitative evaluation is consistent
with prior NAAQS reviews, including the recent NC>2 REA. However, we note that other health
endpoints will be considered as part of the policy discussion in Chapter 10 and during the
rulemaking process.
In addition to an integrated discussion of the respiratory morbidity health evidence,
section 4.3 of this chapter will discuss whether SO2-associated health effects can reasonably be
July 2009 30
-------
considered adverse. Briefly, this discussion will integrate: 1) respiratory morbidity health
evidence; 2) conclusions from previous NAAQS reviews regarding adversity of effect; 3) ATS
guidelines on what constitutes an adverse health effect of air pollution; and 4) CASAC views
regarding the impact of moderate decrements in lung function or respiratory symptoms on
individuals with pre-existing lung disease.
4.2 RESPIRATORY MORBIDITY FOLLOWING SHORT-TERM SO2
EXPOSURE
4.2.1 Overview
The ISA concludes that there is sufficient evidence to infer a causal relationship between
respiratory morbidity and short-term (5-minutes to 24-hours) exposure to SC>2 (ISA, section 5.2).
In large part, this determination is based on the results of controlled human exposure studies in
exercising asthmatics demonstrating a relationship between 5-10 minute peak SC>2 exposures and
decrements in lung function that are frequently accompanied by respiratory symptoms. In fact,
the ISA describes the controlled human exposure studies as being the "definitive evidence" for
its causal determination between short-term 862 exposure and respiratory morbidity (ISA,
section 5.2). In addition to the controlled human exposure evidence, the ISA finds supporting
evidence for its causal determination from a large body of epidemiologic studies observing
positive associations between ambient SC>2 levels and respiratory symptoms, as well as ED visits
and hospital admissions for all respiratory causes and asthma (ISA, section 5.2). An integrated
discussion of the controlled human exposure and epidemiologic evidence from the ISA is
presented below. In addition, section 4.2.3 discusses the effect of medication on SCVinduced
respiratory morbidity.
4.2.2 Integration of Respiratory Morbidity Health Evidence
As previously mentioned, the ISA's finding of a causal relationship between respiratory
morbidity and short-term SO2 exposure is based in large part on results from controlled human
exposure studies involving exercising asthmatics. In general, these studies demonstrate that
asthmatic individuals exposed to SC>2 concentrations as low as 200-300 ppb for 5-10 minutes
during exercise experience moderate or greater bronchoconstriction, measured as a decrease in
FEVi of > 15% or an increase in specific airway resistance (sRaw) of > 100% after correction for
exercise-induced responses in clean air (Bethel et al., 1983; Linn et al., 1983, 1984, 1987; 1988;
July 2009 31
-------
1990; Magnussen et al., 1990; Roger et al., 1985; Gong et al., 1995; Trenga et al., 1999). In
addition, the ISA finds that among asthmatics, both the percentage of individuals affected, and
the severity of the response increases with increasing 862 concentrations. That is, at
concentrations ranging from 200-300 ppb, the lowest levels tested in free breathing chamber
studies3, 5-30% percent of exercising asthmatics experience moderate or greater decrements in
lung function (ISA, Table 3-1). At concentrations > 400 ppb, moderate or greater decrements in
lung function occur in 20-60% of exercising asthmatics, and compared to exposures at 200-300
ppb, a larger percentage of asthmatics experience severe decrements in lung function (i.e., >
200% increase in sRaw, and/or a > 20% decrease in FEVi) (ISA, Table 3-1). Moreover, at SO2
concentrations > 400 ppb, moderate or greater decrements in lung function are frequently
accompanied by respiratory symptoms (e.g., cough, wheeze, chest tightness, shortness of breath)
(Balmes et al., 1987; Gong et al., 1995; Linn et al., 1983; 1987; 1988; 1990; ISA, Table 3-1).
Further analysis and discussion of the individual studies leading to the conclusions presented
above can be found in Sections 3.1.1 to 3.1.3.5 of the ISA.
Supporting the human clinical evidence is a relatively larger body of epidemiologic
studies published since the last review. In general, these studies observed positive associations
between ambient SC>2 concentrations and respiratory symptoms, as well as ED visits and
hospitalizations for all respiratory causes (particularly among children and older adults) and
asthma. Moreover, although copollutant adjustment had varying degrees of influence on the 862
effect estimate in ED visit and hospitalization studies, the effect of 862 appeared to be generally
robust and independent of gaseous copollutants, including NC>2 (Anderson et al., 1998; Lin et al.,
2004a; Sunyer et al., 1997) and O3 (Anderson et al., 1998; Hajat et al., 1999; Tsai et al., 2006;
Yang et al., 2003; 2005). With respect to potential confounding by PMio, the evidence of an
independent SC>2 effect on respiratory health was less consistent, with some positive associations
with ED visit and hospitalization results becoming negative (although the negative results were
not statistically significant) after inclusion of PMi0in regression models (Galan et al., 2003;
Schwartz, 1995 [in New Haven, CT]; Tsai et al., 2006). However, several other ED visit and
hospitalization studies found the 862 effect estimate to be generally robust after inclusion of
3 The ISA cites one chamber study with intermittent exercise where healthy and asthmatic children were exposed to
100 ppb SO2 in a mixture with ozone and sulfuric acid. The ISA notes that compared to exposure to filtered air,
exposure to the pollutant mix did not result in statistically significant changes in lung function or respiratory
symptoms (ISA section 3.1.3.4)
July 2009 32
-------
in regression models (Burnett et al., 1997; Hagen et al., 2000; Hajat et al., 1999; Schwartz,
1995 [in Tacoma, WA]). Furthermore, in most (Van der Zee et al., 1999; Mortimer et al., 2002
and Schildcrout et al., 2006), but not all (Schwartz et al., 1994) studies of respiratory symptoms,
the SC>2 effect estimate remained robust and relatively unchanged after inclusion of PMio in
multipolutant models (although the effect estimate may have lost statistical significance). In
addition, SO2-effect estimates generally remained robust in the limited number of studies that
included PM2.5 and/or PMio-2.5 in multipolutant models (Burnett et al., 1997; Ito et al., 2007; Lin
et al., 2003; NY DOH, 2006). Taken together, the ISA ultimately concludes that studies
employing multipollutant models suggest that SC>2 has an independent effect on respiratory
morbidity outcomes (ISA, section 5.2).
The ISA further characterizes the epidemiologic results of increases in respiratory
symptoms as well as increases in hospital admissions and ED visits as being consistent and
coherent. The evidence is consistent in that associations are reported in studies conducted in
numerous locations and with a variety of methodological approaches (ISA, section 5.2).
Epidemiologic results are coherent in that respiratory symptoms results from epidemiologic
studies with short-term (> 1-hour) exposures are generally in agreement with respiratory
symptom results from controlled human exposure studies of 5-10 minutes. However, the ISA
notes the differences in averaging times associated with respiratory effects in human exposure
and epidemiologic studies. That is, while adverse respiratory effects are observed following 5-
10 minute exposures in human clinical studies, the majority of positive respiratory results from
epidemiologic studies are associated with a 24-hour averaging time- the only averaging time
evaluated in the vast majority of these studies. As a potential explanation for the difference in
averaging times employed across study designs, the ISA suggests that it is possible that results
from epidemiologic studies are being driven, at least in part, by shorter-term peak SC>2
concentrations (ISA section 5.2). More specifically, with respect to epidemiologic studies of
respiratory symptoms, the ISA states "that it is possible that these associations are determined in
large part by peak exposures within a 24-hour period" (ISA, section 5.2). Similarly, the ISA
states that the respiratory effects following peak 862 exposures in controlled human exposure
studies provides a basis for a progression of respiratory morbidity that could result in increased
ED visits and hospital admissions (ISA, section 5.2). Also, it should be noted there is
July 2009 33
-------
epidemiologic evidence to suggest that shorter-term peak SC>2 concentrations can result in
adverse respiratory effects. That is, there are a relatively small number of epidemiologic studies
demonstrating positive associations between 1-hour daily maximum 862 concentrations and
respiratory symptoms, as well ED visits and hospitalizations (ISA, Tables 5-4 and 5-5). While
these studies are not limiting the exposure to a defined 1-hour period, they provide additional
evidence that the shorter term peaks result in adverse respiratory effects.
The ISA also finds that the respiratory effects of SC>2 are consistent with the mode of
action as it is currently understood from animal toxicological and human exposure studies (ISA,
section 5.2). The immediate effect of SC>2 on the respiratory system is bronchoconstriction. This
response is mediated by chemosensitive receptors in the tracheobronchial tree. Activation of
these receptors triggers central nervous system reflexes that result in bronchoconstriction and
respiratory symptoms that are often followed by rapid shallow breathing (ISA, section 5.2). The
ISA notes that asthmatics are likely more sensitive to the respiratory effects of SO2 due to
preexisting inflammation associated with the disease. For example, pre-existing inflammation
may lead to enhanced release of inflammatory mediators, and/or enhanced sensitization of the
chemosensitive receptors (ISA, section 5.2).
Taken together, the ISA concludes that the controlled human exposure, epidemiologic,
and toxicological evidence support its determination of a causal relationship between respiratory
morbidity and short-term (5-minutes to 24-hours) exposure to 862. Results from controlled
human exposure studies provide the definitive evidence for this conclusion, while supporting
evidence is found in numerous epidemiologic studies of respiratory symptoms and ED visits and
hospitalizations (ISA, section 5.2). The ISA further notes that both lines of evidence are
consistent with the SC>2 mode of action as it is currently understood (ISA, section 5.2).
4.2.3 Medication as an Effect Modifier
As mentioned above, the immediate effect of SC>2 on the respiratory system is
bronchoconstriction. Thus, we note that quick-relief and long-term-control asthma medications
have been shown to provide varying degrees of protection against SCVinduced
bronchoconstriction in mild and moderate asthmatics (ISA section 3.1.3.2 and Annex Table D-
1). More specifically, while no therapy has been shown to completely eliminate SCVinduced
respiratory effects in exercising asthmatics, some short- and long-acting asthma medications are
July 2009 34
-------
capable of significantly reducing SO2-induced bronchoconstriction (Gong et al., 1996; 2001;
Koenig et al., 1987; Linn et al., 1990). However, the ISA notes that asthma is often poorly
controlled even among severe asthmatics due to inadequate drug therapy or poor compliance
among those who are on regular medication (Rabe et al., 2004). Moreover, the ISA also notes
that mild asthmatics, who constitute the majority of asthmatic individuals, are much less likely to
use asthma medication than asthmatics with more severe disease (O'Byrne, 2007; Rabe et al.,
2004). Therefore, the ISA finds that it is reasonable to conclude that all asthmatics (i.e., mild,
moderate, and severe), are at high risk of experiencing adverse respiratory effects from SC>2
exposure (ISA section 3.1.3.2).
4.3 WHAT CONSTITUTES AN ADVERSE HEALTH IMPACT FROM SO2
EXPOSURE?
In making judgments as to when various 862 -related health effects become regarded as
adverse to the health of individuals, staff has relied upon the guidelines published by the
American Thoracic Society (ATS), conclusions from previous NAAQS reviews, and the advice
of CASAC. Taken together, staff concludes that for asthmatics, SC>2-induced respiratory effects
are adverse. The rationale for this conclusion is presented below.
The ATS has previously defined adverse respiratory health effects as "medically
significant physiologic changes generally evidenced by one or more of the following: (1)
interference with the normal activity of the affected person or persons, (2) episodic respiratory
illness, (3) incapacitating illness, (4) permanent respiratory injury, and/or (5) progressive
respiratory dysfunction" (ATS 1985). The ATS has also recommended that transient loss in lung
function with accompanying respiratory symptoms, or detectable effects of air pollution on
clinical measures (e.g., medication use) be considered adverse (ATS 1985). We also note that
during the last Os NAAQS review, the CD and Staff Paper indicated that for many people with
lung disease (e.g., asthma), even moderate decrements in lung function (e.g., FEVi decrements >
10% but < 20% and/or >100% increases in sRaw) or respiratory symptoms would likely interfere
with normal activities and result in additional and more frequent use of medication (EPA 2006,
EPA 2007e). In addition, CASAC has previously indicated that in the context of standard
setting, a focus on the lower end of the range of moderate functional responses is most
appropriate for estimating potentially adverse lung function decrements in people with lung
July 2009 35
-------
disease (73 FR16463). Finally, we note that in the current SC>2 NAAQS review, clinicians on the
CASAC Panel again advised that moderate or greater decrements in lung function can be
clinically significant in some individuals with respiratory disease (CASAC transcripts, July 30-
31 2008, pages 211-213)
Considering the advice and recommendations described above, as well as key
conclusions in the ISA, staff finds that for asthmatics, SO2-induced respiratory effects are
adverse. Human exposure studies are described in the ISA as being the "definitive evidence" for
a causal association between short-term SC>2 exposure and respiratory morbidity (ISA, section
5.2). These studies have consistently demonstrated that exposure to SC>2 concentrations as low
as 200-300 ppb for 5-10 minutes can result in moderate or greater decrements in lung function,
evidenced by a >15% decline in FEVi and/or > 100% increase in sRaw in a significant
percentage of exercising asthmatics (see section 4.2.2). It is highly likely that these decrements
in lung function will result in increased medication use and a disruption of normal activities for a
significant percentage of these asthmatics. This expectation is supported by a number of human
exposure studies reporting that some exercising asthmatics required the use of medication to treat
the respiratory effects that followed a 5-10 minute SC>2 exposure (EPA 1994a). It is also
supported by CASAC views during the previous Os review that moderate declines in FEVican be
clinically significant in some individuals (Henderson 2006). As an additional matter, we note
that human exposure studies have also reported that at 862 concentrations > 400 ppb, lung
function decrements (i.e., > 15% decline in FEVi and/or > 100% increase in sRaw) are
frequently accompanied by respiratory symptoms. Taken together, staff concludes that human
exposure studies demonstrate that adverse respiratory effects occur in exercising asthmatics
following 5-10 minute SC>2 exposures as low as 200 ppb. However, we also note that the
subjects participating in these exposure studies do not represent the most sensitive asthmatics
(i.e., severe asthmatics), and therefore, it is possible that adverse respiratory effects could occur
at lower SC>2 concentrations in these individuals.
Epidemiologic studies also indicate that adverse respiratory morbidity effects are
associated with 862. In reaching the conclusion of a causal relationship between respiratory
morbidity and short-term 862 exposure, the ISA generally found positive associations between
ambient SO2 concentrations and ED visits and hospitalizations for all respiratory causes and
July 2009 36
-------
asthma (see section 4.2.2). Notably, ED visits and hospitalizations attributable to air pollution
are considered adverse effects under ATS guidelines. These studies also indicate that 862 is
associated with episodic respiratory illness and aggravation of respiratory diseases, which under
ATS guidance, would also be considered adverse effects of air pollution.
In 2000, the ATS published updated guidelines on what constitutes an adverse health
effect of air pollution (ATS, 2000). These guidelines expanded those released in 1985 (ATS
1985). Among other considerations, the 2000 guidelines stated that measurable negative effects
of air pollution on quality of life should be considered adverse (ATS 2000). These updated
guidelines also indicated that exposure to air pollution that increases the risk of an adverse effect
to the entire population is adverse, even though it may not increase the risk of any individual to
an unacceptable level (ATS 2000). For example, a population of asthmatics could have a
distribution of lung function such that no individual has a level associated with significant
impairment. Exposure to air pollution could shift the distribution to lower levels that still do not
bring any individual to a level that is associated with clinically relevant effects. However, this
would be considered adverse because individuals within the population would have diminished
reserve function, and therefore would be at increased risk if affected by another agent (ATS
2000).
The 2000 ATS guidelines further strengthen the conclusion that SO2-induced respiratory
effects are adverse. As previously mentioned, human clinical studies have consistently
demonstrated that 862 exposure can result in moderate or greater decrements in FEVi and sRaw
at levels as low as 200-300 ppb in a significant percentage of exercising asthmatics. Staff finds
that these results could reasonably indicate an SO2-induced shift in these lung function
measurements for this population. As a result, a significant percentage of exercising asthmatics
exposed to SC>2 concentrations as low as 200 ppb would have diminished reserve lung function
and would be at greater risk if affected by another respiratory agent (e.g., viral infection).
Importantly, diminished reserve lung function in a population that is attributable to air pollution
is an adverse effect under ATS guidance.
Staff finds multiple lines of evidence indicating that exposure to 862 concentrations at
least as low as 200 ppb can result in adverse respiratory effects. We note that this is in
agreement with CASAC comments offered on the first draft SC>2 REA. The CASAC letter to the
July 2009 37
-------
Administrator states: "CASAC believes strongly that the weight of clinical and epidemiology
evidence indicates there are detectable clinically relevant health effects in sensitive
subpopulations down to a level at least as low as 0.2 ppm 862 (Henderson 2008)." Thus, when
examining the adequacy of the current and potential alternative standards (see Chapter 10), staff
finds it appropriate to consider the degree of protection these standards provide, or would
provide, against moderate or greater decrements in lung function and/or respiratory symptoms in
asthmatics at elevated breathing ventilation rates.
4.4 KEY OBSERVATIONS
• The ISA concludes that there is sufficient evidence from human exposure, epidemiologic,
and toxicological studies to infer a causal relationship between respiratory morbidity and
short-term exposure to 862
• The ISA characterizes no other health endpoints as having a causal or likely causal
association with short or long-term exposure to SC>2.
• Human exposure studies demonstrate that at 862 concentrations ranging from 200-300
ppb, the lowest levels tested in free breathing chamber studies, 5-30% percent of
exercising asthmatics experience moderate or greater decrements in lung function (i.e., >
100% increase in sRaw, and/or a > 15% decrease in FEVi). At concentrations > 400 ppb,
moderate or greater decrements in lung function occur in 20-60% of exercising
asthmatics, and compared to exposures at 200-300 ppb, a larger percentage of asthmatics
experience severe decrements in lung function (i.e., > 200% increase in sRaw, and/or a >
20% decrease in FEVi).
• At SC>2 concentrations > 400 ppb, moderate or greater decrements in lung function are
frequently accompanied by respiratory symptoms.
• In general, epidemiologic studies observed positive associations between ambient 862
concentrations and respiratory symptoms, as well as ED visits and hospitalizations for all
respiratory causes and asthma. In studies using multipollutant models, the effects of SC>2
were generally independent of effects of other ambient air pollutants
• No medication regimen has been shown to completely eliminate SO2-induced respiratory
effects in exercising asthmatics.
• Staff finds multiple lines of evidence indicating that 862 exposure can result in
respiratory effects that can reasonably be considered adverse to the health of asthmatics.
July 2009 38
-------
5. SELECTION OF POTENTIAL ALTERNATIVE STANDARDS
FOR ANALYSIS
5.1 INTRODUCTION
The primary goals of the SC>2 risk and exposure assessment described in this document
are to estimate short-term exposures and potential human health risks associated with 1) recent
levels of ambient 862; 2) 862 levels associated with just meeting the current standards; and 3)
862 levels associated with just meeting potential alternative standards. This section presents the
rationale for the selection of the potential alternative standards that are assessed in the
quantitative analyses discussed in Chapters 7 through 9. These potential alternative standards are
defined in terms of indicator, averaging time, form, and level.
5.2 INDICATOR
The SOX include multiple gaseous (e.g., SC>2, SOs) and particulate (e.g., sulfate) species.
In considering the appropriateness of different indicators, we note that the health effects
associated with particulate species of SOX have been considered within the context of the health
effects of ambient particles in the Agency's review of the PM NAAQS. Thus, as discussed in
the Integrated Review Plan (2007a), the current review of the 862 NAAQS is focused on the
gaseous species of SOX and will not consider health effects directly associated with particulate
species of SOX. Of the gaseous species, EPA has historically determined it appropriate to specify
the indicator of the standard in terms of SO2 because other gaseous sulfur oxides (e.g., SO3) are
likely to be found at concentrations many orders of magnitude lower than SC>2 in the atmosphere,
and because most all of the health effects evidence and exposure information is related to SC>2.
The final ISA has again found this to be the case. Therefore, staff concluded that SC>2 remains
the most appropriate indicator for the alternative standards that are analyzed in this document.
5.3 AVERAGING TIME
Staff concluded that the most robust evidence for SC>2-induced respiratory morbidity
exists for exposure durations < 1-hour. The strongest evidence for this conclusion comes from
controlled human exposure studies that have consistently demonstrated that exposure to 862 for
5-10 minutes can result in significant bronchoconstriction and/or respiratory symptoms in
exercising asthmatics (see section 4.2). In fact, the ISA describes the controlled human exposure
July 2009 39
-------
studies as being the "definitive evidence" for its causal determination between SC>2 exposure and
short-term respiratory morbidity (ISA, section 5.2). In addition to these controlled human
exposure studies, there is a relatively small body of epidemiologic evidence describing positive
associations between 1-hour maximum 862 levels and respiratory symptoms as well as hospital
admissions and ED visits for all respiratory causes and asthma (ISA, Tables 5.4 and 5.5). In
addition to the epidemiologic evidence for effects related to the 1-hour maximum concentration
in a 24-hour period, there is a considerably larger body of epidemiologic studies reporting
associations between 24-hour average SC>2 levels and respiratory symptoms, as well as
hospitalizations and ED visits; however, the ISA notes that it is possible that associations
observed in these 24-hour studies are being driven, at least in part, by short-term 862 peaks of
duration < 24-hours. More specifically, when describing epidemiologic studies observing
associations between ambient 862 and respiratory symptoms, the ISA states "that it is possible
that these associations are determined in large part by peak exposures within a 24-hour period"
(ISA, section 5.2). The ISA also states that the respiratory effects following peak SC>2 exposures
in controlled human exposure studies provides a basis for a progression of respiratory morbidity
that could result in increased ED visits and hospital admissions (ISA, section 5.2). It should also
be noted that epidemiologic studies conducted in Paris, France (Dab et al., 1996) and in
Manhattan and Bronx, NY (NY DOH, 2006) used both 24-hour average and 1-hour daily
maximum air quality levels and found similar effect estimates with regard to hospital admissions
for all respiratory causes (Dab et al., 1996) and asthma ED visits (NY DOH, 2006). Finally, in
addition to the controlled human exposure and epidemiologic evidence, the ISA describes key
toxicological studies with exposures ranging from minutes to hours resulting in decrements in
lung function, airway inflammation, and/or hyperresponsiveness in laboratory animals (ISA,
Table 5-2).
The scientific evidence described above suggests that at a minimum, averaging time(s)
selected for further risk and exposure analyses should address respiratory effects associated with
SC>2 exposures of < 1-hour. We note that analyses conducted in the ISA demonstrate that at
monitors measuring all twelve 5-minute SC>2 levels in an hour (n=16), there is a high Pearson
correlation between the 5-minute maximum level and the corresponding 1-hour average 862
concentration, with only one monitor observing a correlation < 0.9 (ISA, section 2.5.2; ISA,
July 2009 40
-------
Table 2-12). Thus, for the purpose of conducting quantitative exposure and risk analyses, staff
concluded that the focus should be on potential alternative 862 standards with an averaging time
of 1-hour. Staff believes that alternative standards with an averaging time of 1-hour will limit
both 5-minute peak concentrations within an hour, as well as other peak 862 concentrations (> 1-
hour) that are likely in part, driving the respiratory outcomes described in epidemiologic studies.
Staff also considered examining alternative 5-minute standards in the risk and exposure
assessment, but concluded for several reasons that such an analysis would be of questionable
utility in the decision-making process. We note that EPA historically conducts air quality,
exposure, and risk analyses of alternative standards by adjusting measured, not modeled air
quality data. This is an issue in evaluating alternative 5-minute standards for 862 because there
were, and continue to be relatively few locations reporting 5-minute 862 concentrations. As
described in Appendix A, from 1997-2007, there were a total of 98 monitors in 13 states and the
District of Columbia measuring maximum 5-minute SC>2 concentrations in an hour. In
comparison, there were 933 monitors in 49 states, the District of Columbia, Puerto Rico and the
Virgin Islands measuring 1-hour SC>2 concentrations. Moreover, it is important to consider that
those monitors reporting 5-minute concentrations do not represent data from a dedicated 5-
minute monitoring network, but rather a voluntary submission of 5-minute values from monitors
placed for the purpose of evaluating attainment of 24-hour and annual average SC>2 NAAQS.
Thus, staff has little confidence that this limited set of data, from monitors sited for a different
purpose, can provide the input required for a comprehensive air quality, exposure, and risk
analysis of a much shorter averaging time standard. In fact, given the spatial heterogeneity of 5-
minute peaks, and the aforementioned issues with monitor siting, staff is not confident (based on
5-minute monitoring data alone) that even in the 13 locations reporting 5-minute concentrations,
that those reported values adequately reflect the extent to which 5-minute peaks are occurring in
those areas.
While we have chosen to evaluate alternative 1-hour averaging time standards in the air
quality, exposure, and risk chapters of this document, this choice did not preclude the possibility
of considering 5-minute standards as part of the policy assessment discussion in Chapter 10, or
during the rulemaking process. Consideration of potential alternative 5-minute standards could
July 2009 41
-------
be based on evidence-based considerations, drawn from the discussion of the scientific evidence
related to 5-10 minute exposures from the ISA, and presented below in Chapter 10.
5.4 FORM
Staff recognizes that the adequacy of the public health protection provided by a 1-hour
daily maximum potential alternative standard will be dependent on the combination of form and
level (see section 5.5). It is therefore important that the particular form selected for a 1-hour
daily maximum potential alternative standard reflect the nature of the health risks posed by
increasing SO2 concentrations. That is, the form of the standard should reflect results from
human exposure studies demonstrating that the percentage of asthmatics affected, and the
severity of the respiratory response (i.e., decrements in lung function, respiratory symptoms)
increases as SC>2 concentrations increase (see section 4.2.2). Taking this into consideration, staff
concluded that a concentration-based form is more appropriate than an exceedance-based form.
This is because a concentration-based form averaged over three years (see below) would give
proportionally greater weight to 1-hour daily maximum 862 concentrations that are well above
the level of the standard, than to 1-hour daily maximum 862 concentrations that are just above
the level of the standard. In contrast, an expected exceedance form would give the same weight
to 1-hour daily maximum SC>2 concentrations that are just above the level of the standard, as to
1-hour daily maximum SC>2 concentrations that are well above the level of the standard.
Therefore, a concentration-based form better reflects the continuum of health risks posed by
increasing SC>2 concentrations (i.e., the percentage of asthmatics affected and the severity of the
response increases with increasing 862 concentrations). Concentration-based forms also provide
greater regulatory stability than a form based on allowing only a single expected exceedance.
Staff also recognizes that it is important to have a form that achieves a balance between
limiting the occurrence of peak concentrations and providing a stable and robust regulatory
target. The most recent review of the PM NAAQS (completed in 2006) judged that using a 98th
percentile form averaged over 3 years provides an appropriate balance between limiting the
occurrence of peak concentrations and providing a stable regulatory target (71 FR 61144). In
that review, staff also considered other forms within the range of the 95th to the 99th percentiles.
In making recommendations regarding the form, staff considered the impact on risk of different
forms, the year-to-year stability in the air quality statistic, and the extent to which different forms
July 2009 42
-------
of the standard would allow different numbers of days per year to be above the level of the
standard in areas that achieve the standard. Based on these considerations, staff recommended
either a 98th percentile form or a 99th percentile form. We have made similar judgments in
selecting appropriate forms for the potential alternative 1-hour daily maximum 862 standards
assessed in this REA. As a result of these judgments, we decided to consider both 98th and 99th
percentile SC>2 concentrations, averaged over 3 years. We have judged that the 98th and 99th
percentile, when combined with the selected range of alternative levels of a 1-hour daily
maximum standard (see below), will likely offer a sufficient range of options to balance the
objective of providing a stable regulatory target against the objective of limiting the occurrence
of peak 5-minute concentrations.
Notably, for a given 1-hour standard level, staffs initial judgment is that a 99th percentile
form will be appreciably more protective against 5-minute peaks than a 98th percentile form.
Staff finds this is likely the case because compared to a standard with a 98th percentile form, a
standard with a 99th percentile form (at the same level) will limit a greater number of peak 1-hour
concentrations, and thus, a greater number of peak 5-minute concentrations. Therefore, all
potential alternative standard levels (see section 5.5) were assessed with a 99th percentile form in
the air quality, exposure and risk analyses. However, as a comparison between forms, one
alternative standard level was examined with a 98th percentile form in the exposure and risk
analyses, and two alternative standard levels were examined with a 98th percentile form in the air
quality analysis.
5.5 LEVEL
When considering the appropriate range of levels for alternative 1-hour daily maximum
standards to analyze in the exposure and risk analyses, staff examined both the controlled human
exposure and epidemiologic evidence evaluated in the ISA. Controlled human exposure
evidence demonstrates that there is a continuum of SO2-related health effects following 5-10
minute peak SC>2 exposures in exercising asthmatics. That is, the ISA finds that the percentage
of asthmatics affected and the severity of the response increases with increasing SC>2
concentrations. At concentrations ranging from 200-300 ppb, approximately 5-30% percent of
exercising asthmatics are likely to experience moderate or greater bronchoconstriction (ISA,
Table 3-1). At concentrations > 400 ppb, moderate or greater bronchoconstriction occurs in
July 2009 43
-------
approximately 20-60% of exercising asthmatics, and compared to exposures at 200-300 ppb, a
larger percentage of subjects experience severe bronchoconstriction (ISA, Table 3-1). Moreover,
at concentrations > 400 ppb, moderate or greater bronchoconstriction was frequently
accompanied with respiratory symptoms (ISA, Table3-l).
In addition to the controlled human exposure evidence, we also considered the
epidemiologic evidence, as well as an air quality analysis conducted by staff characterizing 1-
hour daily maximum SC>2 air quality levels in cities and time periods corresponding to key U.S.
and Canadian ED visit and hospital admission studies for all respiratory causes and asthma4 (key
studies are identified in Table 5-5 of the ISA). Figures 5-1 to 5-5 show standardized effect
estimates and the 98th and 99th percentile 1-hour daily maximum 862 levels for locations and
time periods corresponding to these key U.S. (Figures 5-1 to 5-4) and Canadian5 (Figure 5-5)
studies. In general, staff concluded that the results presented in these figures demonstrate that
most of these epidemiologic studies show positive, although frequently not statistically
significant associations with SC>2. Furthermore, we concluded that Figures 5-1 to 5-5
demonstrate that positive effect estimates, including some that are statistically significant, are
found in locations that span a broad range of 98th and 99th percentile 1-hour daily maximum SC>2
concentrations (98th percentile range: 19- 401 ppb; 99th percentile range: 21-457 ppb). Thus,
staff decided to utilize the 1-hour daily maximum air quality data presented in these figures to
help inform both the upper and lower ranges of alternative 862 standards for analysis in this
RE A (see Chapters 7-9).
4 Authors of relevant U.S. and Canadian studies were contacted and air quality statistics from the study monitor that
recorded the highest SO2 levels were requested. In some cases, U.S. authors provided the AQS monitor IDs used in
their studies and the statistics from the highest reporting monitor were calculated by EPA. In cases where U.S.
authors were unable to provide the requested data (Schwartz 1995, Schwartz 1996, and Jaffe 2003), EPA identified
the maximum reporting monitor from all monitors located in the study area and calculated the 98th and 99th
percentile statistics (see Thompson and Stewart 2009).
5 The Canadian statistics presented in Figure 5-5 were calculated from a data set provided by Dr. Richard Burnett
and were used for all relevant single city studies on which he was an author. Note that air quality statistics presented
for Canadian studies are likely not directly comparable to those presented for U.S. studies. This is because SO2
concentrations presented for Canadian studies represent the 98th and 99th percentile 1-hour daily maximum SO2
concentrations across a given city, rather than concentrations from the single monitor that recorded the highest 98th
and 99th percentile SO2 levels in a given city (see Thompson and Stewart, 2009).
July 2009 44
-------
24-hour effect estimates
1 -hour effect estimates
30
25 --
20 --
15 --
Hi
&
10 f
e
£
EDRE
EDRC
EDRA
EDRW
EDRE^
EDRW
EDRA
EDRC
r
EDRA
EDRA
-5--
-10 --
-15
Legend:
EDRA: ED visits for all respiratory causes- all ages
EDRC: ED visits for all respiratory causes- children
EDRW: ED visits for all respiratory causes- ages 15-64
EDRE: ED visits for all respiratory causes-ages 65+
Figure 5-1. Effect estimates for U.S. all respiratory ED visit studies and associated 98th and 99th
percentile 1-hour daily maximum SO2 levels.
July 2009
45
-------
24-hour effect estimates
-7-
Wilson 2005
Manchester
1-hr99: 69
1-hr98: 59
EDAC
EDAA
EDAE
EDAW
"
Wilson 2005
Portland
1-hr99:47
1-hr98: 36
EDAA
EDAE
EDAC
EDAW
.
•
NYDOH 2006
Manhattan
1-hr99: 80
1-hr98: 62
EDI
EDAA
\A
\
Bronx
1-hr99: 78
1-hr98: 65
Jaffe 2003
Columbus
1-hr99:51
1-hr98: 42
Cincinnati
EDAJ
1-hr99: 457
1-hr98: 401
EDAJ
m
EDI
\J
Cleveland
1-hr99:211
1-hr98: 175
^=-^_
Ito 2007
New York
1-hr99:82
1-hr98: 71
EDAA
r
Legend:
EDAA: ED visits for asthma- all ages
EDAC: ED visits for asthma- children
EDAW: ED visits for asthma- ages 15-64
EDAE: ED visits for asthma- ages 65+
EDAJ: ED visits for asthma- ages 5-34
Figure 5-2. 24-hour effect estimates for U.S. asthma ED visit studies and associated 98th and 99th
percentile 1-hour daily maximum SO2 levels.
July 2009
46
-------
1-hour effect estimates
Peel 2005
EDAA
NYDOH 2006
EDAA
EDAA
Legend:
EDAA: ED visits for asthma- all ages
Figure 5-3. 1-hour effect estimates for U.S. asthma ED visit studies and associated 98th and 99th
percentile 1-hour daily maximum SO2 levels.
July 2009
47
-------
24-hour effect estimates
35 __«=:
30
25
20
1b
10
5
-5
in
Schwartz 1995
New Haven
1-hr 99: 150
1-hr 98: 126
Tacoma
1-hr99: 100
1-hr98:89
HARE
HARE
\
Schwartz 1996
Cleveland
1-hr98: 150
HARE
t
Sheppard 2003
Seattle
1-hr 99: 84
1-hr 98: 70
HAAS
==,1
Lin 2004
HA
AL
Bronx
1-hr98:93
Hi
s
Legend:
HARE: Hospital admissions for all respiratory causes- ages 65+
HAAS: Hospital admissions for asthma- ages <65
HAAL: Hospital admissions for asthma- ages 0-14
Figure 5-4. 24-hour effect estimates for U.S. hospitalization studies and associated 98th and 99
percentile 1-hour daily maximum SO2 levels.6
,th
' There were no key U.S. hospitalization studies with 1-hour effect estimates identified in Table 5-5 of the ISA
July 2009
48
-------
24-hour effect estimates
JL
1-hour effect estimate
24-hour effect estimates
60 -
40 -
20 -
-20 -
r
Yang 2003
Vancouver
1-hr 99: 41
1-hr98:35
HARY
HA
^
RE
Burnett 1997
Toronto
1-hr 99: 21
1-hr 98: 19
HARA
S~~
HA,
Lin 2003
Toronto
1-hr99:58
1-hr98:50
a,B
HA
>
AG
Legend:
HARY: Hospital admissions for all respiratory causes- ages <3
HARE: Hospital admissions for all respiratory causes- ages 65+
HARA: Hospital admissions for all respiratory causes- all ages
HAAB: Hospital admissions for asthma- boys ages 6-12
HAAG: Hospital admissions for asthma- girls ages 6-12
th
Figure 5-5. Effect estimates for Canadian ED visits and hospitalization studies and associated 98
and 99th percentile 1-hour daily maximum SO2 levels.
The highest 98th and 99th percentile 1-hour daily maximum air quality levels were found
in analyses conducted in the cities of Cincinnati (Figure 5-2), Cleveland (Figures 5-2 and 5-4)
and New Haven (Figure 5-4). These studies showed positive associations7 with respiratory-
related hospital admissions or ED visits during time periods when 98th and 99th percentile 1-hour
daily maximum SO2 concentrations ranged from 126 ppb to 457 ppb. Notably, this range of 1-
hour daily maximum SC>2 levels overlaps considerably with 5-10 minute SC>2 concentrations (>
200 ppb) that have consistently been shown in controlled human exposure studies to result in
lung function responses in exercising asthmatics. Of particular concern are the air quality levels
that were found in Cincinnati (Jaffe et al., 2003). The 98th and 99th percentile 1-hour daily
maximum SO2 concentrations were in excess of 400 ppb. Levels > 400 ppb have consistently
been shown in human exposure studies to result in moderate or greater bronchoconstriction in the
7 Results in Cincinnati (Jaffe et al., 2003) and New Haven (Schwartz et al., 1996) were statistically significant.
July 2009
49
-------
presence of respiratory symptoms in a considerable percentage of exercising asthmatics. As a
result, staff decided to analyze alternative standard levels up to 250 ppb. We concluded that a
98th or 99th percentile 1-hour daily maximum standard at this level had the potential to
substantially limit the number of days when the 1-hour daily maximum 862 concentration is >
200 ppb, while also potentially limiting the number of 5-10 minute SO2 peaks > 400 ppb.
In selecting the lower end of the range of alternative standards to be analyzed, staff again
considered controlled human exposure and epidemiologic evidence. However, with regard to the
controlled human exposure evidence, several additional factors were considered. First, we
considered that the subjects in human exposure studies do not represent the most SC>2 sensitive
asthmatics; that is, these studies included mild and moderate, but not severe asthmatics. Also,
while human clinical studies have been conducted in adolescents, younger children have not
been included in these exposure studies, and thus, it is possible asthmatic children represent a
population that is more sensitive to the respiratory effects of SO2 than the individuals who have
been examined to date. Moreover, we considered that approximately 5-30% of asthmatics who
engaged in moderate or greater exertion experienced bronchoconstriction following exposure to
200-300 ppb SC>2, which are the lowest levels tested in free breathing chamber studies (ISA,
Table 3-1). Thus, we concluded that it was highly likely that a subset of the asthmatic
population would also experience bronchoconstriction following exposure to levels lower than
200 ppb.
As an additional consideration, we noted that Figure 5-5 contains two epidemiologic
analyses observing positive associations between ambient 862 concentrations and hospital
admissions in Canadian cities when 99th percentile 1-hour daily maximum SC>2 levels were < 47
ppb. More specifically, positive associations between SC>2 and hospital admissions were found
in Toronto, (Burnett al., 1997) and Vancouver (Yang et. al., 2003) when 99th percentile 1-hour
daily maximum SC>2 levels were approximately 21 ppb and 41 ppb, respectively. However, as
previously noted, the 99th percentile 1-hour daily maximum SC>2 concentrations reported for
Canadian studies are not directly comparable to those reported for U.S. studies. That is, the
concentrations reported for Canadian studies represent the average 98th or 99th percentile 1-hour
daily maximum levels across multiple monitors in a given city (Figure 5-5), rather than 98th or
99th percentile concentrations from the single monitor that recorded the highest SC>2 levels
July 2009 50
-------
(Figures 5-1 to 5-4; see Thompson and Stewart, 2009). As a result, the SC>2 concentrations
presented in Figure 5-5 for Canadian studies would be relatively lower (potentially significantly
lower) than those levels presented in Figures 5-1 to 5-4 for U.S. epidemiologic studies. In
addition to these Canadian studies, we also noted that a U.S. study, Delfmo et al. (2003),
observed a statistically significant association between ambient SO2 and respiratory symptoms in
Hispanic children when the 1-hour daily maximum SC>2 concentration in Los Angeles was 26
ppb (ISA Table 5-4). However, this epidemiologic study was very small (n=22), and did not
examine potential confounding by co-pollutants. Thus, staff concluded that these three studies
alone do not provide sufficient evidence for considering alternative 1-hour daily maximum SC>2
standards below 50 ppb.
Staff noted that numerous studies reported positive associations between ambient 862
and hospital admissions and ED visits in cities and time frames when 98th and/or 99th percentile
1-hour daily maximum SO2 concentrations ranged from approximately 50 to 100 ppb (Figures 5-
1 to 5-5). Moreover, although most of these positive effect estimates were not statistically
significant, there were some statistically significant results in single pollutant models (Portland,
Wilson, 1995; Bronx, NYDOH, 2006; NYC, Ito, 2006; and Schwartz, 1995), as well as some
evidence of statistically significant associations in multi-pollutant models with PM8 (Bronx,
NYDOH, 2006 and NYC, Ito, 2007). Given these epidemiologic and air quality results, as well
as the considerations mentioned above regarding the controlled human exposure evidence, staff
concluded it was appropriate to examine a range of alternative standards in the air quality,
exposure, and risk analyses that include a level of 50 ppb as the lower bound. We judged that a
98th or 99th percentile 1-hour daily maximum standard at this level would both limit the number
of days when 1-hour daily maximum SC>2 levels are > 50 ppb, while also limiting 5-10 minute
peaks of SC>2 > 100 ppb. Moreover, we noted that a level of 50 ppb is substantially below the
98th and 99th percentile 1-hour daily maximum SC>2 levels observed in the Bronx during the
NYDOH analysis and in NYC during the period analyzed by Ito et al., (2006): two studies where
8 In the NYDOH study (2006), the Bronx positive effect estimate remained statistically significant in the presence of
PM2 5 In Ito et al., (2007), the NYC positive effect estimate was statistically significant in the presence of PM25
during the warm season. We also note that in Schwartz et al., (1995), the positive effect estimate in New Haven, but
not Tacoma remained statistically significant in the presence of PM10 when the 99th percentile 1-hour daily
maximum SO2 concentration in New Haven was 150 ppb.
July 2009 51
-------
the SC>2 effect estimate remained robust and statistically significant in multi-pollutant models
with PM2.5 (ISA, Table 5-5).
5.6 KEY OBSERVATIONS
• Staff concluded that 862 remains the most appropriate indicator for the potential
alternative standards to be analyzed in the air quality, exposure, and risk analyses
described in this document.
• For the purpose of conducting quantitative air quality, exposure, and risk analyses, staff
concluded that the focus should be on potential alternative 862 standards with an
averaging time of 1-hour.
• Staff also considered examining alternative 5-minute standards in the risk and exposure
assessment, but concluded that there was insufficient data to do so. However, this did not
preclude the possibility of considering 5-minute standards as part of the policy
assessment discussion in Chapter 10, or during the rulemaking process.
• With regard to the form of the potential alternative standards to be analyzed in the air
quality, exposure, and risk analyses, staff concluded that it was appropriate to consider
the annual 98th and 99th percentile SC>2 concentrations averaged over a 3 year period.
Staff found that a concentration-based form better reflected the continuum of health risks
posed by increasing 862 concentrations, and provided greater regulatory stability than a
form based on allowing only a single expected exceedance.
• Based on findings from controlled human exposure and epidemiologic studies, and
evaluation of air quality information from key U.S. and Canadian studies of ED visits and
hospitalizations, staff concluded that it was appropriate to examine alternative 1-hour
daily maximum standards in the air quality, exposure, and risk analyses in the range of
50-250 ppb.
July 2009 52
-------
6. OVERVIEW OF RISK CHARACTERIZATION AND
EXPOSURE ASSESSMENT
6.1 INTRODUCTION
The assessments presented in the subsequent chapters of this document characterize
short-term exposures (i.e., 5-minutes) and potential health risks associated with: (1) recent
ambient levels of SC>2, (2) levels associated with just meeting the current SC>2 NAAQS, and (3)
levels associated with just meeting several potential alternative standards (see Chapter 5 of this
document for the discussion of potential alternative standards). To characterize health risks, we
employed three approaches (Figure 6-1). With each approach, we characterize health risks
associated with the air quality scenarios mentioned above (i.e., recent air quality unadjusted, air
quality adjusted to simulate just meeting the current standards, and air quality adjusted to
simulate just meeting several potential alternative standards). In the first approach, SC>2 air
quality levels are compared to potential health effect benchmark values (see section 6.2) derived
from the controlled human exposure literature (Chapter 7). In the second approach, modeled
estimates of human exposure are compared to the same potential health effect benchmark values
derived from the human exposure literature (Chapter 8). In the third approach, outputs from the
exposure analysis are combined with exposure-response functions derived from the human
clinical literature to estimate the number and percent of exposed asthmatics that would
experience moderate or greater lung function responses under the different air quality scenarios
(Chapter 9). A more detailed overview of each of these approaches to characterizing health risks
is provided below (section 6.3), and each approach is described in more detail in their respective
chapters and associated appendices. In addition, this chapter also describes important
methodologies used throughout these analyses. This includes the approach used to estimate 5-
minute 862 concentrations from 1-hour data (section 6.4), how recent air quality was adjusted to
simulate alternative air quality standards scenarios (section 6.5), and an overview of how
uncertainty was characterized in each of the analyses performed (section 6.6).
July 2009 53
-------
Evaluation of Human Clinical
Evidence in the ISA
5-10 minute exposures in
exercising asthmatics
Identification of
potential 5-minute
health benchmark
values: 100- 400 ppb
Estimation of
exposure-response
function
Air Quality Analysis in
locations across U.S.
Informs location
selection
Exposure Analysis: modeling exercise that
assesses the likelihood an asthmatic will be
at moderate or greater exertion while
experiencing a 5-minute daily maximum SO2
concentration above a benchmark level
Outputl: Number of days per year 5-
minute daily maximum SO2
concentrations exceed 5-minute
potential health benchmark values
Output 2: The probability of a daily 5-
minute exceedance of benchmark
values given a particular 24-hour
average, or 1 -hour daily maximum SO2
concentration
Output: Estimates the number and
percent of asthmatics at elevated
ventilation rates exposed to 5-minute
daily maximum SO2 concentrations that
exceed 5-minute potential health
benchmark values
Quantitative Risk Analysis:
combines outputs of the exposure
analysis with estimated exposure-
response function
Output: Estimates the number and
percentage of exposed asthmatics that
would experience moderate or greater
lung function responses
Figure 6-1. Overview of analyses addressing exposures and risks associated with 5-minute peak
SO2 exposures. All three outputs are calculated considering current air quality, air
quality just meeting the current standards, and air quality just meeting potential
alternative standards. Note: this schematic was modified from Figure 1-1.
6.2 POTENTIAL HEALTH EFFECT BENCHMARK LEVELS
Potential health benchmark values used in the air quality, exposure, and risk analyses
were derived solely from the human exposure literature. This is primarily because
concentrations used in human clinical studies represent actual personal exposures rather than
concentrations measured at fixed site ambient monitors. In addition, human exposure studies can
examine the health effects of 862 in the absence of co-pollutants that can confound results in
epidemiological analyses; thus, health effects observed in clinical studies can confidently be
attributed to a defined exposure level of 862.
The ISA presents human exposure evidence demonstrating decrements in lung function
in approximately 5-30% of exercising asthmatics exposed to 200-300 ppb SC>2 for 5-10 minutes.
July 2009
54
-------
However, it is important to note: (1) subjects in human exposure studies do not include
individuals who may be most susceptible to the respiratory effects of 862, (e.g., severe
asthmatics and children) and (2) given that 5-30% of exercising asthmatics experienced
bronchoconstriction following exposure to 200-300 ppb SC>2 (the lowest levels tested in free-
breathing chamber studies), it is likely that a percentage of asthmatics would also experience
bronchoconstriction following exposure to levels lower than 200 ppb. That is, there is no
evidence to suggest that 200 ppb represents a threshold level below which no adverse respiratory
effects occur. We also noted that small SO2-induced lung function decrements have been
observed in asthmatics at concentrations as low as 100 ppb when SC>2 is administered via
mouthpiece9 (ISA, section 3.1.3). Considering this information, staff concluded it was
appropriate to examine potential 5-minute benchmark values in the range of 100-400 ppb. The
lower end of the range considers the factors mentioned above, while the upper end of the range
recognizes that 400 ppb represents the lowest concentration at which statistically significant
decrements in lung function are seen in conjunction with statistically significant respiratory
symptoms. Moreover, we note that this range of benchmark values is in general agreement with
consensus CASAC comments on earlier drafts of this document.
As an additional matter, we note that in the outputs of the air quality and exposure
analyses (see section 6.3), staff considered the number of days with a 5-minute maximum SC>2
concentration above benchmark levels rather than all 5-minute exceedances of benchmark levels
in a given day. This is because human exposure studies have suggested that after an initial 862
exposure, there is approximately a 5-hour period of time when asthmatics are less sensitive to
subsequent SO2 challenges (ISA, section 3.1.3.2). As a result, there is uncertainty as to whether
an additional SC>2 exposure(s) on a given day would be associated with an additional adverse
respiratory outcome(s) (i.e., moderate decrements in lung function and/or respiratory symptoms).
On the other hand, we recognize that not counting multiple exceedances in a day could possibly
9 Studies utilizing a mouthpiece exposure system cannot be directly compared to studies involving freely breathing
subjects, as nasal absorption of SO2 is bypassed during oral breathing, thus allowing a greater fraction of inhaled
SO2 to reach the tracheobronchial airways. As a result, individuals exposed to SO2 through a mouthpiece are likely
to experience greater respiratory effects from a given SO2 exposure. In addition, the two mouthpiece studies cited in
the ISA as exposing exercising asthmatics to 100 ppb SO2 (Koenig et al., 1990 and Sheppard et al., 1981) had a
small number of exposures at this concentration (e.g., Sheppard et al., exposed two subjects to 100 ppb SO2) and
observed very small changes in FEVi or sRaw. Nonetheless, these studies do provide very limited evidence for
SO2-induced respiratory effects at 100 ppb.
July 2009 55
-------
lead to an underestimate in the number of asthmatics experiencing an SC>2 concentration above a
benchmark level, and thus, an adverse respiratory outcome. Therefore, there is further
discussion and/or analysis of this topic and its relevance to uncertainty in each of the air quality,
exposure, and risk analysis outputs (see sections 7.4, 8.11 and 9.3).
6.3 APPROACHES FOR ASSESSING EXPOSURE AND RISK
ASSOCIATED WITH 5-MINUTE PEAK SO2 EXPOSURES
In the first approach (i.e., the air quality characterization), we have compared SC>2 air
quality with the potential health effect benchmark levels for SC>2. Scenario-driven air quality
analyses were performed using ambient SC>2 concentrations for the years 1997 though 2006. All
U.S. monitoring sites where 1-hour 862 data have been collected are represented by this analysis
and, as such, the results generated are considered a broad characterization of national air quality
and potential human exposures that might be associated with these concentrations.10 The output
of the air quality characterization is an estimate of the number of exceedances of the potential
health effect benchmark levels for several air quality scenarios. An advantage of this approach is
its relative simplicity; however, there is uncertainty associated with the assumption that SC>2 air
quality can adequately serve as an indicator of exposure to ambient SC>2. Actual exposures will
be influenced by factors not considered by this approach, such as the spatial and temporal
variability in human activities.
In the second approach (i.e., the exposure assessment), we have used an inhalation
exposure model to generate estimates of personal 862 exposures. The estimates of personal
exposure have also been compared to the potential health benchmark levels as was done in the
air quality characterization. This results in estimates of the number of individuals that are likely
to experience exposures exceeding these benchmark levels. For this exposure analysis, a
probabilistic approach was used to model individual exposures considering the time people
spend in different microenvironments and the variable SC>2 concentrations that occur within these
microenvironments across time, space, and microenvironment type. The exposure model also
accounts for activities that individuals perform within the microenvironments, allowing for
estimation of exposures that coincide with varying activity levels. As such, this approach to
10 Two additional subsets of the broader SO2 monitoring network were also used in detailed analyses, thus by
definition are not representative of the full set of monitors in the U.S.
July 2009 56
-------
assessing exposures was more resource intensive than evaluating ambient air quality; therefore,
staff has included the analysis of two specific locations in the U.S. (Greene County, MO. and St.
Louis, MO.)11 Although the geographic scope of this analysis is restricted, the approach
provides realistic estimates of SO2 exposures, particularly those exposures associated with
important emission sources of SO2 and serves to complement the broad air quality
characterization.
Staff used a range of short-term potential health effect benchmarks to characterize risk in
both the air quality and the exposure modeling analyses described above. The levels of potential
benchmarks are based on SO2 exposure levels that have been associated with respiratory
symptoms and decrements in lung function in exercising asthmatics during controlled human
exposure studies (ISA, section 5.2; see above section 6.2 for discussion). Benchmark values of
100, 200, 300, and 400 ppb have been compared to both SO2 air quality (measured and modeled
5-minute SO2 concentrations) and to estimates of SO2 exposures. In characterizing the SO2 air
quality using ambient monitors, the output of the analysis is an estimate of the number of days
per year specific locations experience 5-minute daily maximum levels of SO2 above a particular
benchmark. When personal exposures are simulated, the output of the analysis is an estimate of
the number of individuals at risk for experiencing daily maximum 5-minute levels of SO2 of
ambient origin that exceed a particular benchmark.
In the third approach (i.e., the quantitative risk assessment), we combine outputs from the
exposure analysis with exposure-response functions derived from controlled human exposure
studies. This analysis estimates the percentage and number of asthmatics likely to experience a
given decrement in lung function associated with recent air quality and SO2 levels adjusted to
simulate just meeting the current and potential alternative standards. Staff concluded that it was
appropriate to limit the scope of the quantitative risk assessment to lung function responses based
on findings from the controlled human exposure studies and the basis for this decision is
described below.
11 In the 1st draft REA, staff presented the results of an exposure analysis for Greene County (or Springfield, MO.)
and several other source-based modeling domains in Missouri. Based on CASAC comments received on that
exposure analysis, staff refined the modeling approach and applied those refinements to the Greene County analysis
presented in the 2nd draft REA and completed the exposure assessment in St. Louis which had been started at the
time of the 1st draft REA.
July 2009 57
-------
As discussed above in Chapter 4, the ISA concludes that the overall weight of the
evidence supports a causal relationship between short-term 862 exposure and respiratory
morbidity. The ISA states that the "definitive evidence" for its causal determination is from
controlled human exposure studies demonstrating decrements in lung function and/or respiratory
symptoms in exercising asthmatics exposed to > 200 ppb SO2 (ISA, section 5.2). The ISA
further notes that supporting this causal determination is a larger body of U.S and international
epidemiological studies examining respiratory symptoms and ED visits and hospitalizations for
all respiratory causes and asthma (ISA, section 5.2).
As previously described, staff is utilizing both the epidemiological evidence in the ISA,
and an air quality analysis based on U.S. and Canadian ED visit and hospitalization studies for
all respiratory causes and asthma (Figures 5-1 to 5-5), to qualitatively inform: (1) the selection of
potential 1-hour daily maximum alternative standards to be analyzed in the air quality, exposure,
and risk chapters of this document (see Chapter 5), and (2) the adequacy of the current standard
and consideration of potential alternative standards (Chapter 10). However, staff did not find the
overall breadth of the epidemiological evidence was robust enough to support a quantitative
assessment of risk.
We first note that for purposes of conducting a quantitative risk assessment for locations
in the U.S., staff concludes that only U.S. studies should be considered given differences in
monitoring networks, levels of co-pollutants, and other factors across different locations that may
well alter SCVconcentration-response relationships. Taking this into account, we reviewed the
available epidemiological literature and found relatively few studies that focused on these
endpoints were conducted in U.S. cities. In those U.S. cities where epidemiological studies had
been conducted, many of the SC>2 effect estimates were positive, but not statistically significant
in single pollutant models. Moreover, in the relatively few studies that employed multi-pollutant
models, inclusion of PMio in the model resulted in a loss of statistical significance for the SC>2
effect estimate in about half of these studies (although the effect estimate may have remained
positive). Overall, we conclude that these factors would make it particularly difficult to quantify
with confidence the magnitude of respiratory health effects related to 862 exposures and
therefore, we judge that the results of a quantitative risk assessment based on concentration-
response functions from epidemiological studies for these health outcomes would be of limited
July 2009 58
-------
utility in the decision-making process given the nature of the uncertainties associated with these
studies.
6.4 APPROACH FOR ESTIMATING 5-MINUTE PEAK SO2
CONCENTRATIONS
Health effects evaluated in this REA include those associated with 5-10 minute peak
concentrations of SC>2. While there are 98 ambient monitors that have reported 5-minute SC>2
concentrations some time during 1997-2007, the spatial and temporal representation is limited to
a few states and often only a few years of monitoring. Most of these monitors report the 5-
minute maximum SC>2 concentration occurring within an hour, though there were some that
reported all twelve continuous 5-minute 862 concentrations measured within the hour. The
ambient monitors reporting continuous 862 values are limited to fewer locations and number of
monitoring years, with sixteen monitors deployed within six US states and Washington DC, ten
of which operated only during one year. The overwhelming majority of the SC>2 ambient
monitoring data are for 1-hour average concentrations (upwards to 935 monitors), comprising a
broad monitoring network that includes most U.S. states and territories. Because the health
effects of greatest interest were associated with short-term exposures (5-10 minutes) and a
greater number of monitors and monitor-years were available for the 5-minute maximum SC>2
concentrations than 10-minute maximum concentrations, a model was developed to estimate 5-
minute maximum SC>2 concentrations from the comprehensive 1-hour SC>2 ambient monitoring
data.
Staff first reviewed the air quality characterization conducted in the prior SO2 NAAQS
review and supplementary analyses. In these prior analyses, relationships between maximum 5-
minute SC>2 concentrations and the 1-hour average SC>2 concentrations, or peak-to-mean ratios
(PMRs) were evaluated and used to approximate 5-minute maximum SC>2 concentrations from 1-
hour values (EPA, 1986a; EPA, 1994b; SAI, 1996; Thompson, 2000). While the relationship
between the two metrics is not expected to be linear, the temporal patterns in the two averaging
times are consistent. Five-minute maximum 862 concentrations are often much greater than that
of the corresponding 1-hour 862 concentrations, and observed increases in a given 1-hour 862
concentration often coincide with increases in the 5-minute maximum SC>2 concentration. As an
example of this pattern, the time-series of 1-hour average and 5-minute maximum SC>2
July 2009 59
-------
concentrations measured at an ambient monitor across a 3-day period in 2005 is illustrated in
Figure 6-2.
March 10-12, 2005 - ID 290770037
300
Q. 250
^
O 200
£ 150
O
tM
O
100
50
—•—5-minute maximum
-•e- 1-hour
—A..
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
m
o
o
CN
o
T
m
o
o
CN
o
oo
m
o
o
CN
o
CN
m
o
o
CN
CD
m
o
o
CN
o
CN
m
o
o
CM
o
m
o
o
CN
~~~
T
m
o
o
CN
~~~
oo
m
o
o
CN
~~~
CN
m
o
o
CNI
CD
m
o
o
CN
o
CN
m
o
o
CN
o
m
o
o
CN
CN
T
m
o
o
CN
CN
CO
m
o
o
CN
CN
CN
m
o
o
CN
CD
m
o
o
CN
o
CN
m
o
o
CN
CN
CN
CO CO
CO
CO CO
CO
CO CO
CO
CO CO
CO
CO CO
CO
CO CO
CN
CO
Hour of Day
Figure 6-2. Example of an hourly time-series of measured 1 -hour and measured 5-minute
maximum SO2 concentrations.
In general, PMRs were determined to be approximately two in some of the earlier studies
when used in estimating 5-minute peak SC>2 concentrations; though for the exposure analyses
conducted for the last NAAQS review, a distribution of PMRs was used with values of up to
eleven (EPA, 1994b). In each of the analyses conducted previously, estimates of PMRs were
derived using ambient monitoring data (i.e., where both 5-minute maximum and 1-hour average
SC>2 were measured) and then used to estimate the occurrence of peak 5-minute 862
concentrations given a 1-hour ambient SO2 concentration. The approach was generally as
follows:
=PMRxC
equation (6-1)
July 2009
60
-------
where,
Cmax-s = estimated 5-minute maximum 862 concentration (ppb)
PMR = peak-to-mean ratio (PMR)
C ' i-hour = measured 1-hour average 862 concentration
At the time of the last NAAQS review, there were very few monitors reporting 5-minute
SC>2 data. In fact, distributions of PMRs from ambient monitors surrounding a single coal-fired
power utility served as the primary source used in estimating 5-minute peak concentrations used
in the exposure analyses (EPA, 1994b). As mentioned above, the PMRs were determined to be
approximately two in these earlier studies; however, the ratio can vary depending on a several
factors. It has been shown that there can be increased variability in the ratio with decreasing 1-
hour average 862 concentrations, that is, there is a greater likelihood of values greater than two
at low hourly average concentrations than expected at high hourly average concentrations (EPA,
1986a). It has also been argued that the occurrence of short-term peak concentrations at ambient
monitors may be influenced by particular SC>2 emission sources (EPA, 1994b). Different sources
have variable emission amounts, temporal operating patterns (e.g., seasonal, time-of-day),
facility maintenance, and other physical parameters (e.g., stack height, area terrain) that likely
contribute to variability in 5-minute maximum SC>2 concentrations. In addition, a sensitivity
analysis conducted for copper-smelters determined that distance from the source was inversely
proportional to the PMR in all three of the 1-hour mean stratifications evaluated (i.e., < 0.04
ppm, 0.04 to < 0.15 ppm, and >0.15 ppm), with the highest 1-hour category having the lowest
range of PMR (Sciences International, 1995).12
There are some data available for the current SC>2 monitoring network regarding the type
of sources that may be near the ambient monitors, the magnitude of emissions, the temporal
variation in emissions, and distance from specific sources; however, staff determined that there
was no practical way to define every ambient monitor as being exclusively influenced by a single
source or a defined mix of sources. Given other conditions that may vary within a specific
source category (monitor-to-source distances, local meteorology, operating conditions, etc.), staff
also determined that there was no practical way to use such data quantitatively in the
12 In that analysis, normalized 1-hour SO2 concentrations were obtained by dividing by the maximum hourly
concentration.
July 2009 61
-------
construction of the PMR statistical model and apply such a model to the 1-hour SC>2 ambient
monitor data.
In recognizing the limited geographic span of the monitors reporting the 5-minute
maximum 862 concentrations and the overall uncertainty regarding the amount of influence of a
specific source on any given monitor, staff developed an approach based on hourly SO2
concentration levels and the variability observed at the monitors reporting both the 5-minute
maximum and 1-hour average SC>2 concentrations. The main assumption in the approach is that
the temporal and spatial pattern in SC>2 source emissions is influenced by the type of source(s)
present, its operating conditions, and that the emission pattern(s) is reflected in the ambient SC>2
concentration distribution measured at the monitor. Thus, measures of concentration level and
associated variability at each monitor were used as a surrogate for the variability in the source
characteristics that may impact concentrations at a particular monitor. Each monitor reporting 5-
minute maximum SO2 concentrations was categorized based on the coefficient of variation
(COV) of 1-hour average SC>2 concentrations and then used to estimate distribution of PMRs for
range of 1-hour SC>2 concentrations. This approach, that fully utilizes all of the available 5-
minute maximum SC>2 data, is detailed in section 7.2.3.
6.5 APPROACH FOR SIMULATING THE CURRENT AND
ALTERNATIVE AIR QUALITY STANDARD SCENARIOS
A primary goal of the risk and exposure assessments described in this document is to
evaluate the ability of the current 862 primary standards (30 ppb annual average, 140 ppb 24-
hour average)13 and potential alternative standards (99th percentile 1-hour daily maximum SO2
levels of 50, 100, 150, 200, and 250 ppb, and 98th percentile 1-hour daily maximum SO2 levels:
200 ppb; see Chapter 5 of this document) to protect public health. To evaluate the ability of a
specific standard to protect public health, ambient SC>2 concentrations need to be adjusted such
that they simulate levels of SC>2 that just meet that standard. Such adjustments allow for
comparison of the level of public health protection that could be associated with just meeting the
current and potential alternative standards.
13 For consistency, the concentration units in this chapter are reported as ppb, even though the SO2 NAAQS have
units of ppm.
July 2009 62
-------
All areas of the United States currently have ambient SC>2 levels below the current annual
standard (EPA, 2007c). One site in Northampton County, Pa., measured concentrations above
the level of the 24-hour standard in 2006. Therefore, to evaluate whether the current standards
adequately protect public health, nearly all 862 concentrations need to be adjusted upwards in all
areas included in our assessment to simulate levels of SO2 that would just meet the current
standard levels. Similarly, to simulate a potential air quality standard that is below current air
quality standards, those current levels must be adjusted downward.
Ambient SC>2 concentrations and exposures were characterized by considering as is air
quality (unadjusted concentrations) and several hypothetical air quality scenarios. Each of the
hypothetical air quality scenarios had an ambient concentration target, derived from the form and
level of the current NAAQS or from potential alternative standards. Staff chose a proportional
approach to adjust the 862 concentrations to simulate each of the current and alternative air
quality standard scenarios.14 A proportional approach was selected based on the mostly linear
relationship between older high concentration years of air quality when compared with recent
low concentration years at several locations (Rizzo, 2009). Briefly, for each location of interest
(/') and year (/), SC>2 concentration adjustment factors (F) were derived by the following
equation:
equation (6-2)
where,
FJJ = Adjustment factor derived from the air quality standard target
concentration in location /' and yeary (unitless)
S = concentration values allowed that would just meet the air quality
standard level (ppb)
Cmax,ij = maximum measured SC>2 concentration given particular form of
standard at a monitor in location /' and yeary (ppb)
14 The particular equation used to derive each of the adjustment factors is dependent on the form and level of the
standard considered, however the equations all share proportionality between the target level and ambient
concentration. To evaluate the current and alternative air quality scenarios in the exposure assessment (Chapter 8),
a mathematically equivalent proportional approach was used to adjust the benchmark levels rather than adjusting the
ambient concentrations as done for the air quality characterization (Chapter 7).
July 2009 63
-------
In these cases where staff simulated a proportional adjustment in ambient SC>2
concentrations using equation (6-2), it was assumed that the current temporal and spatial
distribution of air concentrations (as characterized by the current air quality data) is maintained
and increased 862 emissions contribute to increased 862 concentrations. All the hourly 862
concentrations in a location were multiplied by the same constant value F, whereas the highest
monitor (in terms of concentration) is adjusted such that it just meets the standard target level.
This procedure for adjusting either the ambient concentrations (i.e., in the air quality
characterization) or health effect benchmark levels (i.e., in the exposure assessment) was
necessary to provide insight into the degree of exposure and risk which would be associated with
an increase in ambient SC>2 levels such that the levels were just at the current standards in the
areas analyzed. Staff recognizes that it is extremely unlikely that SC>2 concentrations in any of
the selected areas where concentrations have been adjusted would rise to meet the current
NAAQS and that there is considerable uncertainty associated with the simulation of conditions
that would just meet the current standards. Nevertheless, this procedure was necessary to assess
the ability of the current standards, not current ambient SO2 concentrations, to protect public
health. This process of adjusting SC>2 concentrations to simulate just meeting a specific standard
is described in more detail in sections 7.2.4 and 8.8.1.
6.6 APPROACHES FOR CHARACTERIZING VARIABILITY AND
UNCERTAINTY
An important issue associated with any population exposure or risk assessment is the
characterization of variability and uncertainty. Variability refers to the inherent heterogeneity in
a population or variable of interest (e.g., residential air exchange rates) and cannot be reduced
through further research, only better characterized with additional measurement. Uncertainty
categorically refers to the lack of knowledge regarding the values of model input variables (i.e.,
parameter uncertainty), the physical systems or relationships used (i.e., use of input variables to
estimate exposure or risk or model uncertainty), and in specifying the scenario that is consistent
with purpose of the assessment (i.e., scenario uncertainty). Uncertainty is, ideally, reduced to
the maximum extent possible through improved measurement of key parameters and iterative
model refinement. The approaches used to assess variability and characterize uncertainty in this
REA are discussed in the following two sections.
July 2009 64
-------
6.6.1 Characterization of Variability
The purpose for addressing variability in this REA is to ensure that the characterization of
air quality and the estimates of exposure and risk reflect the variability of ambient 862
concentrations and associated 862 exposure and health risk across the study locations and
population. In this REA, there are several algorithms that account for variability of input data
when generating the number of estimated benchmark exceedances or health risk outputs. For
example, variability may result from the number of monitors operating in an area and their
associated temporal and spatial heterogeneity in ambient SC>2 concentrations. Variability may
also arise from differences in the population residing within a census block (e.g., age
distribution) and the activities that may affect SC>2 population exposure (e.g., time spent
outdoors), and/or the influential risk factors (e.g., the fraction of the population responding to an
SC>2 exposure). A complete range of potential exposure levels and associated risk estimates can
be generated when appropriately addressing variability in exposure and risk assessments; note
however that the range of values obtained would be within the constraints of the algorithm or
modeling system used, not the complete range of the true exposure or risk values.
Where possible, staff identified and incorporated any observed variability in input data
sets and estimated parameters within each of the analyses performed in Chapters 7-9 rather than
employing standard default assumptions and/or using point estimates to describe model inputs.
The details regarding variability distributions used in data inputs are described in the methods
sections of each assessment and summarized in sections 7.4, 8.11, and 9.3 for the air quality
characterization, the exposure assessment, and the risk characterization, respectively.
6.6.2 Characterization of Uncertainty
While it may be possible to capture a full range of exposure or risk values by accounting
for variability inherent to influential factors, the true exposure or risk for any given individual is
largely unknown. To characterize health risks, exposure and risk assessors commonly use an
iterative process of gathering data, developing models, and estimating exposures and risks, given
the goals of the assessment, scale of the assessment performed, and the limitations of the input
data available. However, significant uncertainty often remains and emphasis is then placed on
characterizing the nature of that uncertainty and its impact on exposure and risk estimates.
July 2009 65
-------
The characterization of uncertainty can include either qualitative or quantitative
evaluations, or a combination of both. The approach can also be tiered, that is, the analysis can
begin with a simple qualitative uncertainty characterization then progress to a complex
probabilistic analysis. This could follow when a lower tier analysis indicates a high degree of
uncertainty for certain identified sources, the sources are highly influential to exposure and risk
estimates, and sufficient information and resources are available to conduct a quantitative
uncertainty assessment. This is not to suggest that quantitative uncertainty analyses should
always be performed in all exposure and risk assessments. The decision regarding the type of
uncertainty characterization performed is also be informed by the intended scope and purpose of
the assessment, whether the selected analysis will provide additional information to the overall
decision regarding health protection, whether sufficient data are available to conduct a complex
quantitative analysis, and if time and resources are available for higher tier characterizations
(EPA, 2004b; WHO, 2008).
The primary purpose of the uncertainty characterization approach selected in this REA is
to identify and compare the relative impact important sources of uncertainty may have on the
potential health effect benchmarks and/or respiratory effects endpoints estimated in Chapters 7-9.
The approach used to evaluate uncertainty was adapted from guidelines outlining how to conduct
a qualitative uncertainty characterization (WHO, 2008), though staff also performed several
quantitative sensitivity analyses to iteratively inform both model development and the qualitative
uncertainty characterization, where possible. While it may be considered ideal to follow a tiered
approach in the REA to quantitatively characterize all identified uncertainties, staff selected the
mainly qualitative approach given the limited data available to inform probabilistic analyses, and
time and resource constraints.
The qualitative approach used in this REA varies from that of WHO (2008) in that a
greater focus of the characterization performed was placed on evaluating the direction and the
magnitude15 of the uncertainly; that is, qualitatively rating how the source of uncertainty, in the
presence of alternative information, may affect the estimated air quality, exposure, and health
risk assessment results. In addition and consistent with the WHO (2008) guidance, staff discuss
the uncertainty in the knowledge-base (e.g., the accuracy of the data used, acknowledgement of
' This is synonymous with the "level of uncertainty" discussed in WHO (2008), section 5.1.2.2.
July 2009 66
-------
data gaps) and decisions made (e.g., selection of particular model forms), though qualitative
ratings were assigned only to uncertainty regarding the knowledge-base.
First, staff identified the key sources of the assessment that may contribute to uncertainty
in the air quality, exposure, and risk estimates and provide the rationale for their inclusion.
Then, staff characterized the magnitude and direction each identified source of uncertainty
influences the assessment results. Consistent with the WHO (2008) guidance, staff subjectively
scaled the overall impact of the uncertainty by considering the degree of severity of the
uncertainty as implied by the relationship between the source of the uncertainty and the output of
the air quality characterization. Where the magnitude of uncertainty was rated low, it was judged
that large changes within the source of uncertainty would have only a small effect on the
assessment results. A designation of medium implies that a change within the source of
uncertainty would likely have a moderate (or proportional) effect on the results. A
characterization of high implies that a small change in the source would have a large effect on
results. Staff also included the direction of influence, indicating how the source of uncertainty
was judged to affect estimated benchmark exceedances or risk estimates; either the estimated
values were likely over- or under-estimated. In the instance where the component of uncertainty
can affect the assessment endpoint in either direction, the influence was judged as both. Staff
characterized the direction of influence as unknown when there was no evidence available to
judge the directional nature of uncertainty associated with the particular source. Staff also
subjectively scaled the knowledge-base uncertainty associated with each identified source using
a three level scale: low indicated significant confidence in the data used and its applicability to
the assessment endpoints, medium implied that there were some limitations regarding
consistency and completeness of the data used or scientific evidence presented, and high
indicated the knowledge-base was extremely limited.
The output of the uncertainty characterization was a summary describing, for each
identified source of uncertainty, the magnitude of the impact and the direction of influence the
uncertainty may have on the air quality, exposure, and risk characterization results. And finally,
an evaluation of the uncertainties presented in Chapters 7-9 is discussed in Chapter 10, providing
the overall implications in informing staffs evaluation of exposures and risks associated with
July 2009 67
-------
level, form, and averaging time related to judging the adequancy of the current standard and
consideration of potential alternative primary SCh standards.
6.7 KEY OBSERVATIONS
• Potential health effect benchmark values were derived from the controlled human
exposure literature.
• Staff concluded that there is no evidence from human exposure studies to suggest that
200 ppb represents a threshold level below which no adverse respiratory effects occur.
• Staff concluded that it was appropriate to consider 5-minute benchmark levels in the
range of 100 to 400 ppb in the air quality and exposure analyses.
July 2009 68
-------
7. AMBIENT AIR QUALITY AND BENCHMARK HEALTH RISK
CHARACTERIZATION FOR 5-MINUTE PEAK SO2 EXPOSURES
7.1 OVERVIEW
Ambient monitoring data for each of the years 1997 through 2007 were used in this
chapter to characterize 862 air quality across the U.S. The measured air quality, as well as
additional 862 concentrations derived from the measured air quality data, were used as an
indicator of potential human exposure. While an ambient monitor measures 862 concentrations
at a stationary location, the monitor may well represent the concentrations to which persons
residing nearby are exposed. The quality of the extrapolation of ambient monitor concentration
to personal exposure depends upon the spatial representativeness of the monitoring network, the
corresponding spatial distribution of important emission sources, local meteorological conditions
and geographical features, and a consideration of places that persons visit. Staff considers the
analyses presented in this chapter to be a broad characterization of national air quality and
potential human exposures that might be associated with a variety of scenario-driven
concentrations. This is because many of the 862 ambient monitoring sites used in this analysis
target public health monitoring objectives and some of the analysis results were separated by the
population density surrounding the ambient monitors.
As previously discussed in Chapter 4, the ISA finds the evidence for an association
between respiratory morbidity and SC>2 exposure to be "sufficient to infer a causal relationship"
(ISA, section 5.2). The ISA states that the "definitive evidence" for this conclusion comes from
the results of human exposure studies demonstrating decrements in lung function and/or
respiratory symptoms in exercising asthmatics following exposure to 862 levels as low as 200 to
300 ppb for 5-10 minutes (ISA, section 5.2). Accordingly, 5-minute potential health effect
benchmark levels ranging from 100-400 ppb were derived from the human exposure literature
(see section 6.2 for benchmark level rationale) and compared to measured and statistically
modeled 5-minute ambient concentrations. A broad analysis is first presented that evaluates the
potential health risk at all ambient monitors, and then for more detailed analyses, at monitors
located within selected U.S. counties (see section 7.2.4). Staff estimated the number of days in a
year with 5-minute benchmark exceedances and the probability of benchmark exceedances given
July 2009 69
-------
the occurrence of 1-hour daily maximum or 24-hour average SC>2 concentrations at ambient
monitors.
All ambient 862 monitors report hourly concentrations; a subset of those report 5-minute
maximum 862 concentrations as well, with a subset of these reporting continuous 5-minute 862
concentrations. Because there were two distinct sample averaging times reported for the
available ambient monitoring data (i.e., ambient monitors reporting 1-hour SC>2 concentration
measurements alone and monitors reporting both 5-minute and 1-hour average SC>2
concentrations), the data used in the analyses were separated by staff as follows.
The first set of ambient air quality data was from monitors reporting both 5-minute and 1-
hour 862 concentrations. Staff 1) analyzed the ambient monitoring data for trends in 1-hour and
5-minute 862 concentrations, 2) counted the number of measured daily 5-minute maximum 862
concentrations above the potential health effect benchmark levels given the annual average 862
concentrations, 3) estimated the probability of benchmark exceedances given the 24-hour
average and 1-hour daily maximum SC>2 concentrations, 4) developed a statistical model to
estimate 5-minute maximum SC>2 concentrations from 1-hour SC>2 concentrations, and 5)
evaluated the performance of the statistical model by comparing the model's predicted versus
measured numbers of exceedances (see section 7.2.3).
The second set of ambient data was comprised of 1-hour SC>2 concentrations from the
broader 862 monitoring network; therefore this set also included 1-hour 862 concentrations from
those monitors where 5-minute 862 data were reported, though the vast majority of the 1-hour
data were from monitors that did not report 5-minute concentration measurements. Staff applied
the statistical model that related 5-minute to 1-hour SO2 measurements to this second set of
ambient monitoring data to estimate 5-minute maximum SC>2 concentrations. As was done with
the 5-minute SC>2 ambient measurement data, staff 1) evaluated trends in SC>2 concentrations, 2)
counted the number of statistically modeled potential health effect benchmark exceedances in a
day using the same longer-term averaging times, and 3) estimated the probability of peak
concentrations associated with 1-hour daily maximum and 24-hour average 862 concentrations.
Staff considered three data analysis groups to characterize the ambient 862 air quality. In
the first group, we evaluated the combined 5-minute and 1-hour 862 measurement data as they
were reported, representing the conditions at the time of monitoring (termed in this assessment
July 2009 70
-------
"as is"). The second group also considered the as is air quality; however staff analyzed the
statistically modeled 5-minute 862 concentrations that were generated from as is 1-hour 862
measurements. This second data analysis group expanded the geographic scope of the 5-minute
air quality characterization by using the broader 862 monitoring network. The third data
analysis group considered 1-hour SO2 concentrations adjusted to just meeting the current
NAAQS16 and each of the potential alternative 1-hour daily maximum standard levels of 50, 100,
150, 200 and 250 ppb (see Chapter 5 for details). The data used to simulate the current and
alternative standard scenarios were limited to the most recent and comprehensive ambient
monitoring data available (i.e., 2001-2006) in forty selected U.S. counties.17 Due to the form of
the potential alternative standards considered here (98th and 99th percentiles of the 1-hour daily
maximum concentrations averaged over 3 years), the recent ambient monitoring data set was
evaluated using two three-year periods, 2001-2003 and 2004-2006.18 Whereas the first analysis
group used entirely 1-hour and 5-minute SO2 measurement data, the second and third analysis
groups used statistically modeled 5-minute SC>2 concentrations that were generated from 1-hour
SC>2 concentrations. The third data analysis group also included an adjustment of the 1-hour SC>2
concentrations to evaluate several air quality standard scenarios in 40 selected counties.
Staff expected that there would be variability in the number of persons living within close
proximity of each monitor (both the 5-minute and 1-hour SC>2 monitors) given the particular
siting characteristics of the ambient monitors (e.g., either source- or population-oriented
monitoring objectives). Therefore, we separated some of the air quality results within each
scenario by using the population density surrounding each ambient monitor. First, each monitor
was characterized by having one of three population densities (i.e., low, medium, and high),
groupings defined by the three characteristic regions of the population distribution generated
from the broader SC>2 monitoring network (section 7.2.2). Then, staff counted the number days
16 Just meeting the current NAAQS levels could either be meeting a 30 ppb annual average or the 140 ppb 24-hour
average concentration (one allowed exceedance), whichever is the controlling standard at that ambient monitor (see
section 7.2.4).
17 At the time of the initial data download from the AQS data mart, many of the monitors did not have complete
years of data available for 2007, therefore the most recent data for most monitors was from 2006. These complete
site-year data are a subset of the broader ambient monitoring data set available.
18 A number of 3-year groups are within 2001-2006 (e.g., 2001-2003, 2002-2004, etc.) and a number of years of
monitoring data are outside the 2001-2006 time frame that could have been used in an extended 3-year grouping of
2001-2006 air quality (e.g., 2000-2002). For convenience, the upper and lower groupings were chosen by staff to
represent 3-year air quality within the 6-year period when considering just meeting the potential alternative
standards.
July 2009 71
-------
with 5-minute benchmark exceedances per year at each monitor, either measured or estimated
depending on the data analysis group considered, and aggregated the results by the population
density group. Rather than count the total number of 5-minute 862 concentrations above a
particular benchmark, staff calculated the number of days in a year with a 5-minute 862
concentration above a potential health effect benchmark.19
One output of this air quality characterization is an estimate of the number of days per
year a monitor experienced 5-minute SC>2 concentration above those that may cause adverse
health effects in susceptible individuals (i.e., benchmark level exceedances). These counts are a
useful metric in comparing one ambient monitor or monitoring location to another and in
identifying where and when frequent benchmark exceedances could occur. However, earlier
analyses indicated that the relationship between the annual average 862 concentration and the
number of 5-minute benchmark exceedances was generally weak (1st draft 862 REA).
Therefore, a comparison of the number of days/year with benchmark exceedances to the annual
average SC>2 concentration is of limited use. This absence of a strong relationship highlights the
ineffectiveness of long-term averaged concentrations in controlling short-term peak
concentrations. Furthermore, while there was an improved relationship between the number of
5-minute maximum SC>2 concentrations and 24-hour average concentrations, it was also shown
that the number of benchmark exceedances in a day was variable given a specific 24-hour
average concentration.20 For example, there could be as many as five 5-minute maximum 862
concentrations above a selected benchmark levels at a particular 24-hour average 862
concentration, while in other instances there may be no benchmark exceedances at the same 24-
hour concentration.
Given that there is variability in the number of 5-minute peak SC>2 concentrations
associated with concentrations of longer-term averaging times, that a daily maximum 5-minute
SC>2 concentration was the metric of interest, and that the potential alternative standards
19 In the 1st draft SO2 REA, as well as the early draft NO2 REAs, all benchmark exceedances for any hour of the day
were reported. The use of the daily maximum exceedance was selected in the final NO2 REA as well in the 2nd draft
and final SO2 REA to improve the temporal perspective for the metric in the air quality analysis (i.e., the number of
daily maximum exceedances also gives the number of days in a year with an exceedance of a selected benchmark),
and to be consistent with the exposure and risk analyses. The implication of not counting multiple exceedances is
discussed further in sections 7.4, 8.11, and 10.3.3.1.
20 In the 1st draft SO2 REA, multiple exceedances within a day (if any) were counted. In the 2nd draft and final SO2
REA, there is only one counted maximum exceedance per day. Additional analysis of multiple exceedances within
the day is given in section 8.11.211.
July 2009 72
-------
investigated use 1-hour daily maximum SC>2 concentrations, staff decided that an appropriate
comparison would be between the frequency of peak 5-minute 862 concentrations given 1-hour
daily maximum 862 concentrations. Thus, the second output of this air quality characterization
is presented as the probability of a benchmark exceedance given a daily maximum 1-hour 862
concentration. In addition, the probability of a 5-minute benchmark exceedance given a 24-hour
average concentration is also provided to offer additional perspective on this averaging time.
7.2 APPROACH
There were five broad steps to characterize the SO2 air quality. The first step involved
compiling and screening the ambient air quality data collected since 1997 to ensure consistency
with the SC>2 NAAQS requirements and for usefulness in this air quality characterization. Next,
due to potential variable influence of SC>2 emission sources on ambient monitor concentrations,
the monitors from each of the two data sets (i.e., combined 5-minute and 1-hour, broader 1-hour
only) were categorized and evaluated according to their monitoring site attributes, including land
use characteristics, location type, monitoring objective, distance to emissions sources, and
population density. In addition, the variability in 5-minute and 1-hour 862 concentrations was
evaluated and used to categorize each ambient monitor. Staff used concentration variability in
the development and application of a statistical model used to estimate 5-minute maximum SC>2
concentrations. Then, a concentration adjustment approach was developed and applied in
selected locations to evaluate several air quality scenarios. And finally, air quality metrics of
interest (i.e., the number and probability of potential health effect benchmark exceedances) were
calculated using the air quality data from each scenario.
The following provides an overview of the five steps used to characterize air quality and
summarizes key portions of the analysis. Briefly, the five steps include: 1) screening of air
quality data; 2) evaluation of site characteristics of ambient SO2 monitors; 3) development of a
statistical model to estimate 5-minute maximum SC>2 concentrations; 4) adjustment of air quality;
and 5) generation of air quality metrics. Details regarding the ambient monitors used for
characterizing air quality and associated descriptive meta-data are provided in Appendix A. 1.
7.2.1 Screening of Air Quality Data
SC>2 air quality data and associated documentation from the years 1997 through 2007
were downloaded from EPA's Air Quality System for this analysis (EPA, 2007c, h). Data
July 2009 73
-------
obtained were used as reported by these sources; there were no substitutions performed for any
missing or zero concentration data. The total available 862 ambient monitoring data, reported
for either 5-minute or 1-hour averaging times, are summarized in Table 7-1. The 5-minute SC>2
monitoring data existed in either one of two forms; the single highest 5-minute concentration
occurring in a 1-hour period (referred to here as max-5 data set), or all twelve 5-minute
concentrations within a 1-hour period (referred to here as continuous-5 data set).
Table 7-1. Summary of all available 5-minute and 1-hour SO2 ambient monitoring data, years 1997-
2007, pre-screened.
Sample Type
Max-5
Continuous-5
1-hour
Number of
Monitors
104
16
935
Number of
States1
13 + DC
6 + DC
49 + DC, PR, VI
Years in
Operation
1997-2007
1999-2007
1997-2007
Number of
Measurements2
3,457,057
3,328,725
47,206,918
Notes:
1 DC=District of Columbia, PR=Puerto Rico, VI=Virgin Islands.
2 For the max-5 and 1 -hour data sets, this number represents the number of hours a sample was
collected/reported. The number for the continuous-5 data set is the number of 5-minute samples.
The total number of hours where measurements for the continuous-5 set were collected is
283,202 (see Table 7-2).
Staff evaluated the data for inconsistencies and duplication. The reported measurement
units varied within each of the data sets, therefore the staff converted all concentrations to parts
per billion (ppb). Next staff screened each of the three data sets listed in Table 7-1 for where
monitor IDs had multiple parameter occurrence codes (POCs) and identical monitoring times.
These duplicate measures could either result from co-location of ambient monitors (i.e., more
than one measurement instrument) or from duplicate reporting of ambient concentrations (i.e.,
the 5-minute maximum concentration in the max-5 data set is the same as the maximum 5-
minute concentration reported from the continuous-5 data set). As a result of this evaluation and
additional concentration level screening (see below), staff constructed several data sets for
analysis in this REA. These data sets are summarized in Table 7-2 and are described in detail
below.
July 2009
74
-------
Table 7-2. Analytical data sets generated using the continuous-5, max-5, and 1-hour ambient SO2
monitoring data, following screening.
Sample Type
Max-5
Continuous-5
with 1-hour1
1-hour
Within Set
Duplicates
(n)
300,438
0
0
Available
Data
(n)
3,156,619
283,2022
47,188,640
Combined Set
Dup
29,058
icates
n)
258,457
Final
Combined
Max-5 Data
(n)
3,410,763
Final
Combined
1-hour
(n)
47,21 3, 3853
Final
Combined
Max-5 Si-
hour
(n)
2,367,6864
Notes:
1 1-hour concentrations from continuous-5 data were calculated from all 5-minute values within the hour.
2 The number of 5-minute maximum SO2 samples.
3 There were a total of 24,745 unique 1 -hour values added from the continuous-5 monitors.
4 There were a total of 2,408,351 values where the 5-minute maximum and 1-hour measurements were
reported at the same time at the same monitor. Of these, a total of 40,665 were screened out for not
meeting the peak-to-mean (PMR) criterion.
Boxes spanning two rows are comprised of data from the two sample types. For example, there were
29,058 duplicate values when considering the max-5 and continuous-5 data sets. Therefore, in creating
the "Final Combined Max-5 Data" (n= 3,410,763), this was the sum of the max-5 (n=3, 156,619) and
continuous-5 (n=283,202) minus the duplicates (n=29,058).
1. Simultaneously reported/measured ambient SO2 data
Two separate data sets were constructed that had multiple 5-minute SC>2 measurements
collected at the same monitoring location and time for:
• max-5 duplicates (i.e., simultaneous measurements of 5-minute maximum 862
concentrations from co-located max-5 monitors; n=300,438)
• max-5 and continuous-5 duplicates (i.e., simultaneous 5-minute maximum 862
concentrations reported in max-5 and continuous-5 datasets; n=29,058)
A third data set was constructed that had simultaneous 1-hour SC>2 measurements
collected at the same monitoring location and time for:
• 1-hour duplicates (i.e., from 1-hour SC>2 monitors and from averaging the continuous-5
monitors; n=258,457)
Each of these duplicate data sets were used for quality assurance purposes only, the
evaluation of which is presented in Appendix A.2. The duplicate values were not used in the
statistical model development or for any other 5-minute or 1-hour SC>2 concentration analysis.
July 2009
75
-------
2. Combined 5-minute and 1-hour ambient SO2 data
A complete set of 5-minute maximum SC>2 concentrations,21 generated from the max-5
data set and from the maximum 5-minute concentrations reported by the continuous-5 monitors,
was then combined with their corresponding measured 1-hour 862 concentrations (see below).
Then, the combined data were screened for validity, recognizing that the combined max-5 and 1-
hour SO2 data set may have certain anomalies (e.g., 5-minute maximum SO2 concentrations < 1-
hour mean SC>2 concentration). A value of 1 was selected as the lower bound peak-to-mean ratio
(PMR),22 accepting the possibility that the 5-minute maximum concentrations (and all other 5-
minute concentrations within the same hour) may be identical to the 1-hour average
concentration. A PMR of <12 was selected as the upper bound since it would be a mathematical
impossibility to generate a value at or above 12 given there are twelve 5-minute measurements
within any 1-hour period.23 This screening resulted in a total of nearly 2.4 million values
comprising the combined 5-minute maximum and 1-hour 862 concentration dataset. The
locations of these 98 monitoring sites comprising this dataset are illustrated in Figure 7-1. Staff
used this data set to develop a statistical model (section 7.2.3) and to characterize the measured
5-minute maximum ambient air quality. Details on the monitors used and site attributes (e.g.,
latitude, longitude, operating years, monitoring objective) are provided in Appendix A.I.
21 A single 5 -minute and 1 -hour SO2 concentration was used in each of data set 2 and 3. The criteria for selection of
a particular value was first based on whether the 1-hour concentration was calculated from the continuous-5 data
(where present) followed by the monitor ID POC that had the greatest overall number of samples.
22 The peak-to-mean ratio is the maximum 5-minute SO2 concentration within an hour divided by the 1-hour average
SO2 concentration.
23 As the 5-minute maximum concentration approaches infinity, the other 11 concentrations measured in the hour
comparatively tend towards zero, giving a maximum PMR = Peak/Mean = Cmax/[(Cmax + (Cothers-> 0) x 11)/12] <
12.
July 2009 76
-------
Figure 7-1. Location of the 98 monitors that reported 5-minute maximum SO2 concentrations and
comprising the first data analysis group.
3. Broader 1-hour ambient SO2 data
This data set was comprised of all 1-hour SC>2 data, whether obtained from the 1-hour
ambient monitoring data set or from averaging 5-minute concentrations from the continuous-5
data set. The raw 1-hour data from a total of 935 ambient monitors were first screened for
negative concentrations (n=3,555) and for where concentrations were less than 0.1 ppb
(n=14,723). The screened data were not used in any analyses. The refined 1-hour data
(n=47,188,640) were then combined with the 1-hour average concentrations obtained from the
continuous 5-monitors. Staff retained the 1-hour average concentrations from the continuous-5
monitors where duplicate values existed. This was done to better maintain the relationship
between the 5-minute maximum and 1-hour SC>2 concentrations. As described above for data set
1, staff removed duplicate 1-hour values identified at each monitoring location originating from
the 1-hour and continuous-5 monitors for separate analysis (Appendix A-2). The remaining 1-
hour SC>2 data set (with duplicate 1-hour values removed) was then combined with the complete
5-minute maximum data set described above for data set 2 (with duplicate 5-minute maximum
July 2009 77
-------
SC>2 values removed). Staff used data set 2 in developing the statistical model to estimate 5-
minute maximum SC>2 concentrations (section 7.2.3).
Additional screening of the 1-hour 862 data set was performed using a 75%
completeness criterion. This monitoring data requirement is used in demonstrating attainment of
the SO2 NAAQS (61 FR 25579).24 For an ambient monitor to have a valid year of data, first,
valid days were selected as those with at least 18 hours of data. Then, each monitor was required
to have 75% of each calendar quarter with complete days (either 68 or 69 days per quartile).
This 75% completeness criterion was applied to the available monitoring data to generate a total
of 4,692 valid site-years of data obtained from 809 ambient monitors. The number of valid
monitoring site-years available as a result of this screening is presented in Table 7-3, effectively
encompassing ambient 862 monitoring in 48 US States, Washington DC, Puerto Rico and the
US Virgin Islands over years 1997 through 2006.25 The locations of the 809 monitors
comprising the broader SO2 monitoring network are illustrated in Figure 7-2. This data set was
used in the second data air quality characterization scenario that considered the measured as is 1-
hour SC>2 concentrations with statistically modeled 5-minute maximum concentrations. Details
on the monitors used and site attributes (e.g., latitude, longitude, operating years, monitoring
objective) are provided in Appendix A. 1.
7.2.2 Site Characteristics of Ambient SOi Monitors
The siting of the monitors is of particular importance, recognizing that proximity to local
sources could have an influence on the measured 862 concentration data and subsequent
interpretation of the air quality characterization. Staff evaluated the attributes of monitors within
each of the two data sets; the first data set was comprised of monitors that reported 5-minute
maximum SO2 concentrations, and the second was generated from monitors within the broader
SC>2 monitoring network and having valid 1-hour SC>2 concentrations. Two points are worth
mentioning for this analysis; the first being the number of monitors and the second being the
potential for differences in types of sources influencing each monitor. While there is overlap in
24 See http://www.epa.gov/air/oaqps/greenbook/40cfr50_2001.pdf
25 Based on the version date of the files downloaded from EPA's AQS data mart (6/20/2007), all 1-hour SO2 data
from 2007 were less than complete. In addition, two monitors located in Hawaii County, HI were identified in the
1st draft REA as having concentrations influenced by natural sources. Therefore, monitor IDs 150010005 and
150010007, while meeting the completeness criteria, were removed from the valid 1-hour SO2 data set due to the
influence of volcanic activity on measured SO2 concentrations at these locations. Alaska had no SO2 monitors
during the period of analysis.
July 2009 78
-------
Table 7-3. Counts of complete and incomplete site-years of 1-hour SO2 ambient monitoring data
for 1997-2006.
State
Abbr.
AL
AZ
AR
CA
CO
CT
DE
DC
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
Ml
MN
MS
MO
MT
NE
NV
NH
NJ
NM
NY
NC
ND
OH
OK
OR
PA
Rl
SC
SD
TN
Code
01
04
05
06
08
09
10
11
12
13
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
44
45
46
47
Number of Site-Years
Complete
36
44
17
308
33
69
27
10
223
65
31
17
235
276
110
28
104
57
25
10
102
84
74
25
166
121
9
16
63
117
56
229
61
155
309
59
0
398
21
90
7
175
Incomplete
15
24
14
136
13
18
16
1
76
34
19
10
30
80
33
27
42
11
18
7
33
28
23
11
40
50
13
6
26
21
24
72
29
45
74
32
4
97
2
34
4
70
Percent
Valid
71
65
55
69
72
79
63
91
75
66
62
63
89
78
77
51
71
84
58
59
76
75
76
69
81
71
41
73
71
85
70
76
68
78
81
65
0
80
91
73
64
71
Number of Valid
Monitors per year
Minimum
1
1
1
7
1
6
2
1
3
5
2
1
18
13
8
2
2
5
1
1
6
5
5
1
11
2
1
1
3
12
3
21
4
10
28
3
0
33
2
5
1
12
Maximum
5
6
2
41
6
12
4
1
28
9
4
3
30
34
14
4
13
6
7
3
15
15
12
4
21
18
2
4
11
14
9
24
9
18
35
9
0
51
3
11
3
23
July 2009
79
-------
State
Abbr.
TX
UT
VT
VA
WA
WV
Wl
WY
PR
VI
Code
48
49
50
51
53
54
55
56
72
78
Total or
Average1
Number of Site-Years
Complete
172
33
11
94
18
203
39
3
33
24
4692
Incomplete
71
14
4
28
24
28
18
8
32
23
1612
Percent
Valid
71
70
73
77
43
88
68
27
51
51
68
Number of Valid
Monitors per year
Minimum
10
3
1
8
1
14
2
1
1
1
6
Maximum
21
4
2
11
7
25
7
1
6
5
12
Notes:
1 Columns of complete and incomplete site years were summed. The percent
valid site-years and the monitors in operation per year with valid data were
averaged.
. . • .
• ••: * ... \'r
. „• *,.:• v f:.
Virain Islands
Puerto
Broader SO2 net
Figure 7-2. Location of the 809 monitors comprising the broader SO2 ambient monitoring network
(i.e., the second data analysis group).
July 2009
80
-------
the measurement of 5-minute maximum and its associated 1-hour SC>2 concentration at some
locations (n=98), the remainder of 862 monitors with valid data (n=711) are sited in other
locations where 5-minute 862 measurements have not been reported. Staff evaluated the
ambient monitor attributes within each data set because there may be influential attributes in the
subset of data used to develop the statistical model (i.e., monitors reporting 5-minute maximum
SC>2 concentrations) that are not applicable to the broader SC>2 monitoring network. Staff
acknowledges that the information available and the monitoring site characteristics considered
can limit how well the monitoring data serve as an indicator of human exposure.
First, staff evaluated the specific monitoring site characteristics provided in AQS,
including the monitoring objective, measurement scale, and predominant land-use. Additional
features such as proximity to 862 emission sources and the population residing within various
distances of each monitor were estimated using monitoring site and emission source geographic
coordinates and U.S Census data. Each of these attributes is summarized here to provide
perspective on the attributes of where 5-minute maximum SC>2 concentrations were reported
versus the attributes of the broader SC>2 monitoring network. A more thorough discussion of the
purpose of the existing ambient SC>2 monitoring network is provided in Chapter 2. Individual
monitor site characteristics are given in Appendix A.I.
The monitoring objective meta-data field describes the nature of the monitor in terms of
its attempt to generally characterize health effects, the presence of point sources, regional
transport, or welfare effects. In recognizing that there were variable numbers of ambient
monitors in operation and variation in the number of valid site-years available for each data set,
staff weighted the monitoring objectives by the number of site-years. This was done to provide
perspective on the air quality characterization results that are based on the total site-years of data
available, not just the number of ambient monitors. In addition, the monitors can have more than
one objective. Where multiple objectives were designated, staff selected a single objective to
characterize each monitor using the following order: population exposure, source-oriented, high
concentration, general/background, unknown.26 All other objectives (whether known or
indicated as "none") were grouped by staff into an "Other" category. Figure 7-3 summarizes the
26 This order was selected to characterize the monitors with a specific objective. Most of the time where there were
multiple objectives at a monitor, there was a specific objective (e.g., population exposure) and a non-specific
objective (e.g., unknown).
July 2009 81
-------
objectives for the monitors comprising each data set. Each of the data sets had a large proportion
of site-years that would target public health objectives through the population exposure and
highest concentration categories, though the monitors in the broader 862 monitoring network
had a greater percentage than the monitors reporting both 5-minute maximum and 1-hour 862
concentrations. The monitors reporting 5-minute concentrations had approximately twice the
percentage of site-years from source-oriented monitors when compared with the broader SC>2
monitoring network.
Similarly, the overall measurement scale of the monitors used for the air quality
characterization in each location was evaluated based on the weighting of valid site-years of
data. The measurement scale represents the air volumes associated with the monitoring area
dimensions. While a monitor can have multiple objectives, each monitor typically has only one
measurement scale. Figure 7-3 also summarizes the measurement scales for the monitoring site-
years comprising each data set. Both data sets had their greatest proportion of monitoring site-
years associated with neighborhood measurement scales (500 m to 4 km), though monitors
recording 1-hour concentrations had about 22 percentage points greater than the monitors
reporting 5-minute maximum concentrations. Furthermore, monitors reporting 5-minute values
had a larger proportion of site-years of data characterized at an urban (4 to 50 km) and regional
scale (50 km to 1,000 km) compared with the broader SC>2 monitoring network.
The land-use meta-data indicate the prevalent land-use within 1A mile of the monitoring
site. Figure 7-4 summarizes the land-use surrounding monitors that reported 5-minute
maximum concentrations and the monitors in the broader 1-hour SC>2 monitoring network. Over
half of the site-years are from residential and industrial areas and are of similar proportions for
both data sets considered. The greatest difference in the surrounding land-use was for the
percent of site-years associated with monitors sited in agricultural and commercial areas. The
monitors reporting 5-minute maximum SC>2 concentrations had about 10 percentage points more
site-years from monitors within agricultural areas and 10 percentage points less in commercial
areas when compared to the respective land use of the broader 862 monitoring network.
The setting is a general description of the environment within which the site is located.
Figure 7-4 also summarizes the setting of the monitors comprising each data set. For monitors
reporting 5-minute concentrations, the greatest proportion of site-years is from ambient monitors
July 2009 82
-------
with a rural setting (49%). Most of the site-years in the broader SC>2 monitoring network were
from monitors within a suburban setting (40%).
July 2009 83
-------
Objective
2
o
+j
'E
o
^
-------
B
o
+j
o
c
If)
Land-Use
UNKNOWN
2%
Setting
RESIDENTIAL
33%
AGRICULTURAL
27%
URBAN AND
CENTER CITY
21%
COMMERCIAL
13%
INDUSTRIAL
20%
DESERT
FOREST
5%
SUBURBAN
30%
RURAL
49%
0%
MOBILE
1%
UNKNOWN
1%
AGRICULTURAL
13%
o
RESIDENTIAL
36%
UNKNOWN
1%
URBAN AND
CENTER CITY
28%
COMMERCIAL
23%
FOREST
3%
RURAL
31%
INDUSTRIAL
23%
SUBURBAN
40%
Figure 7-4. Distribution of site-years of data considering land-use and setting: monitors that reported 5-minute maximum SO2
concentrations (top) and the broader SO2 monitoring network (bottom).
July 2009
85
-------
Stationary sources (in particular, power generating utilities using fossil fuels) are the
largest contributor to 862 emissions in the U.S. (ISA, section 2.1). First, staff determined the
distances, amounts of, and types of stationary source emissions associated with each of the
ambient 862 monitors. Then, staff selected the sources in close proximity of each monitor to
identify whether there are differences in the distribution of emission sources that could affect the
monitored concentrations. Stationary sources emitting > 5 tons per year (tpy) SC>2 and within 20
km of each monitor were identified using data from the 2002 National Emissions Inventory
(NEI).27 Details on the number of sources, the distribution of emissions, and the method for
determining the distances to each individual ambient monitor are provided in Appendix A. 1.
The total SC>2 source emissions within 20 km of every monitor were summed by their
source descriptions; the top eight source types were selected for evaluation followed by a
summing of all other remaining source types in a final source description group ("other").28
These emission results are presented in Figure 7-5 for the monitors reporting 5-minute maximum
SC>2 concentrations and for the broader SC>2 monitoring network. A comparison of the sources
located within 20 km of monitors comprising both data sets indicates strong similarity in the
types of sources present. Approximately 70% of the stationary source emissions local to
monitors comprising either data set originate from fossil fuel power generation.29 Similarity in
emission contributions from several other source categories is also evident (i.e., petroleum
refineries, iron and steel mills, cement manufacturing). One of the largest distinctions between
the sources surrounding the two data sets is the emission contribution from primary smelters.
There were greater source emissions from smelters located within 20 km of the monitors
reporting 5-minute maximum SC>2 concentrations (8.8%) than within 20 km of the broader SC>2
monitoring network (1.1%). A second difference between the two sets of data existed in the
emission contribution from a combined power generation, transmission and distribution
description; this source category contributes approximately 11% to emissions proximal monitors
27 2002 National Emissions Inventory Data & Documentation. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. Available at: http://www.epa.gov/ttn/chief/net/2002inventory.html.
28 Details for the number of sources and emissions surrounding each monitor are given in Appendix A. 1.1 and A. 1.2
29 This emission category was summed from fossil fuel power generation (NEI code 221112) and hydroelectric
utilities (NEI code 221111). Hydroelectric utility SO2 emissions arise from power generating facility operations that
require fossil fuel combustion (e.g., diesel-fueled backup generators).
July 2009 86
-------
in the broader SC>2 monitoring network compared with only 2% at monitors measuring 5-minute
SC>2 concentrations.
5-minute SO2 Monitors
Cement
Manufacturing
1%
Flour Milling
2% ~"^\
Power Generation,
Transmission and
Distribution
2%
Iron and Steel Mills_/
2%
Petroleum /
Refineries^
7%
Other
Petroleum/Coal
Products
Manufacturing
1%
Others
8%
Fossil Fuel Power
Generation
68%
Primary
Smelting/Refining
9%
All SO2 Monitors cement
Manufacturing
1%
Flour Milling
1%
Power Generation,
Transmission and
Distribution
11%
Iron and Steel
2%
Petroleum
Refineries^
Other
Petroleum/Coal
Products
Manufacturing
1%
/! Others
/-^ 9%
5%
Fossil Fuel Power
Generation
69%
Primary
Smelting/Refining
1%
Figure 7-5. The percent of total SO2 emissions of sources located within 20 km of ambient
monitors: monitors reporting 5-minute maximum SO2 concentrations (top) and the
broader SO2 monitoring network (bottom).
July 2009
87
-------
The population residing within four buffer distances of each ambient monitor was
estimated using Arc View. First, staff obtained block group population data from the US Census
and converted the location of each block group polygon to single central point. Then buffers
were created around each monitor location at progressive 5 km distances to a final buffer
distance of 20 km. The total population was estimated by summing the population of all block
group centroids that fell within the monitor buffers. We then created population distribution
functions (across monitors) for the monitors reporting 5-minuute maximum SC>2 concentrations
and for the broader SC>2 monitoring network. An example of the population distribution
represented by the monitors comprising each data set is given by Figure 7-6, with the population
within each of the buffer distances given in Appendix A. 1.30 In general, the shape of the
population distribution was similar for each data set, though as a whole, the monitors reporting
5-minute 862 concentrations tended to be sited in locations with lower population density when
considering any of the population buffers. Staff created population density groups of low, mid,
and high to categorize all ambient monitors using the population distribution within 5 km, by
apportioning each data set into three sample size groupings. The low-population density group
included those monitors with populations under 10,000 persons. Mid-population density
included those monitors with between 10,000 and 50,000 persons, while the high-population
density group was assigned to monitors with greater than 50,000 persons within a 5 km buffer.
These population density groups of low, mid, and high were used in separating some of the air
quality characterization results.
The population density surrounding each monitor was compared with its monitoring
objective. The descriptive statistics for each monitoring objective, separately considering those
monitors that reported 5-minute maximum SC>2 concentrations and the broader SC>2 monitoring
network, are provided in Table 7-4. The calculated population statistics generally support
expectations given the designated monitoring objectives. There are similarities in the population
density around monitors characterized as having highest concentration and population exposure
monitor objectives, both of which having the greatest number of persons residing within 5 km of
the monitors. Source-oriented monitors had consistently lower population densities, though
monitors assigned the general/background objective had the lowest population densities.
30
If the estimated population was zero, then the monitor value was not plotted in the figure.
July 2009 88
-------
100
90
80
70
60 -i
>
4-1
jo
D
E
D
O
• All SO2 Monitors (5 km)
o 5-minute SO2 Monitors (5 km)
1E+0 1E+1 1E+2 1E+3 1E+4 1E+5 1E+6
Population Residing Within 5 km of Ambient Monitors
1E+7
Figure 7-6. Distribution of the population residing within a 5 km radius of ambient monitors:
monitors reporting 5-minute maximum SO2 concentrations and the broader SO2
monitoring network.
Table 7-4. Descriptive statistics of the population residing within a 5 km radius of ambient
monitors by monitoring objective: monitors reporting 5-minute maximum SO2 concentrations and
the broader SO2 monitoring network.
Data
Source
5-
minute
monitors
All SO2
Monitors
Objective1
GEN
OTH
SRC
UNK
HIC
POP
GEN
SRC
UNK
OTH
HIC
POP
n
10
6
15
18
19
30
45
68
179
30
202
285
Population residing within 5 km of Ambient Monitor2
mean
8537
8881
9216
40177
59958
67886
18096
20594
58477
61878
86485
87406
max
28224
35872
42208
262592
316944
382995
378415
136288
1215989
1205886
1301071
1173879
p95
28224
35872
42208
262592
316944
216129
78376
76896
200253
320320
222716
276378
p75
17957
11967
17925
33774
90863
70221
7883
30070
59772
11205
94449
105796
p50
1330
2396
1103
20360
17963
49283
1947
9844
16676
4270
48179
54986
p25
0
655
0
4587
13314
21784
492
1112
3403
787
14142
21336
P5
0
0
0
0
0
3280
0
0
0
0
905
1865
min
0
0
0
0
0
2118
0
0
0
0
0
0
Notes:
1 Objectives are POP=Population Exposure; HIC=Highest Concentration; SRC=Source Oriented;
GEN=General/Background; OTH=Other; UNK=Unknown.
2 p5, p25, p50, p75, and p95 are the 5th, 25th, 50th, 75th, and 95th percentiles, respectively. The minimum (min),
maximum (max), and arithmetic average (mean) are also provided.
July 2009
89
-------
7.2.3 Statistical model to estimate 5-minute maximum SOi concentrations
As described earlier, staff noted there were a limited number of ambient monitors that
reported 5-minute maximum SC>2 concentrations. The majority of the SC>2 monitoring network
reports 1-hour average 862 concentrations. Staff developed a statistical model to extend the 5-
minute 862 air quality characterization to locations where 5-minute concentrations were not
reported. This statistical model was briefly introduced in section 6.4; this section details the
development of the statistical model designed to estimate 5-minute maximum SC>2
concentrations from 1-hour SC>2 concentrations, using the combined 5-minute maximum and 1-
hour SC>2 measurement data set (see section 7.2.1).
Fundamental to the statistical model are the peak-to-mean ratios or PMRs. Peak-to-mean
ratios are derived by dividing the 5-minute maximum 862 concentration by the 1-hour average
SC>2 concentration. These derived PMRs can be useful in estimating 5-minute maximum SC>2
concentrations when only the 1-hour SC>2 concentration is known. The values of PMRs derived
from the monitoring data can be variable and are likely dependent on local source emissions, site
meteorology, and other influential factors. Each of these factors will have variable influence on
the measured 1-hour and 5-minute SC>2 concentrations at the ambient monitors. Therefore, to
develop a useful tool for extrapolating from the measurement data, at a minimum, the approach
needed to account for variability in ambient concentrations. It is within this context that the
statistical model was developed.
Staff selected the variability in 862 concentrations at each individual ambient monitor as
a surrogate for source emissions, source types, and/or distance to sources to allow for a
purposeful application of the statistical model to the broader 1-hour 862 measurement data.
Many of the meta-data described earlier in section 7.2.2, while useful for qualitatively describing
characteristics of monitors in the SC>2 monitoring network, were not considered robust in
quantifying how sources might influence monitored concentrations. The utility of the meta-data
is also diminished when the monitor attributes were reported as unknown, missing entries, or
possibly mischaracterized. In addition, while individual source types, emissions, and distances
to the monitors are presented as quantitative measures, the use of this data can be problematic.
This is because 1) source characteristics can change over time, 2) it is largely unknown what
source(s) influence many of the ambient monitors and by how much, 3) there is uncertainty in
source emission estimates, and 4) even similar source types will not have the same emission
July 2009 90
-------
characteristics. Staff considered several ways to link the statistical model developed from
monitors reporting 5-minute maximum concentrations to the broader SC>2 ambient monitoring
network, including the use of the ambient monitoring site characteristics. Staff decided that the
measured concentrations had the most to offer in efficiently designing such a linkage given the
strong relationships between averaging times, concentration variability, and the frequency of
peak concentrations. Where possible, staff compared the relevant monitor attributes described in
section 7.2.2 with selected variability metrics used in developing and applying the statistical
model.
The purpose of the first analysis that follows is to determine an appropriate variable to
reasonably connect the statistical model derived from 5-minute and 1-hour concentrations to any
1-hour SC>2 concentration data set where there are no 5-minute 862 measurements. Staff first
evaluated variability metrics associated with 5-minute and 1-hour 862 ambient monitoring
concentrations as a basis for linking the statistical model to 1-hour concentrations. Next, staff
generated distributions of PMRs for use in estimating 5-minute concentrations. Then the
statistical model was applied to where 5-minute measurements were reported and evaluated
using cross-validation.
7.2.3.1 Relationship Between 5-minute and 1-hour SO2 Concentrations
Because the statistical model employs 5-minute and 1-hour SC>2 concentrations, staff
evaluated the relationship between the concentrations for the two averaging times. The monitors
reporting all twelve 5-minute concentrations within the hour were used for this analysis (n=16).
First, all of the continuous-5 minute data available for each monitor were averaged to generate a
single 5-minute mean concentration (both in an arithmetic and geometric mean form) and their
respective standard deviations, yielding a total of 16 monitor-specific 5-minute SO2 values.31
Staff performed a second calculation to generate similar statistics using the continuous 5-minute
data, though a 1-hour averaging time was of interest. To obtain the 1-hour statistics, the 5-
minute SC>2 concentrations within an hour were averaged to generate 1-hour mean SC>2
concentrations for each monitor, which were then averaged to generate a single 1-hour mean SC>2
31 Each of the 16 continuous-5 monitors was characterized by four statistics, arithmetic and geometric means and
their respective standard deviations.
July 2009 91
-------
concentration (both in an arithmetic and geometric mean form) and their corresponding standard
deviations, yielding a total of 16 monitor-specific 1-hour 862 values.
Ambient Monitor 5-minute COV (%)
0 50 100 150 200 250 300 350
Q C I
3.5 H
3 Q
O J
2.5
c
O
y = 0.75x + 20
R2 = 0.98
' y 250 R
y = 0.95x + 0.07
R2 = 0.99
300
200 §
o
100 S
c
a)
50
01234
Ambient Monitor 5-minute GSD
Figure 7-7. Comparison of hourly and 5-minute concentration COVs and GSDs at sixteen
monitors reporting all twelve 5-minute SO2 concentrations over multiple years of
monitoring.
Staff selected the coefficient of variation (COV)32 and geometric standard deviation
(GSD) as metrics to compare concentration variability in both 1-hour and 5-minute averaging
times, each of which are illustrated in Figure 7-7. As expected, a strong direct linear relationship
exists between the variability in 5-minute and 1-hour SO2 concentrations at each monitor. Even
with the limited geographic representation (these monitors are from only six U.S. States and
Washington DC), there is a wide range in the observed concentration variability for both the 5-
minute and associated hourly measurements (i.e., COVs range from about 75 - 300%, GSDs
range from about 1.7 - 3.7). In general, this analysis demonstrates that variability in 5-minute
32 The COV used here is calculated by dividing the standard deviation by the arithmetic mean, then multiplying by
100. The statistic gives a relative measure of variation, to better facilitate the comparison of data having different
mean concentrations or units of measure.
July 2009
92
-------
SC>2 concentrations is directly related to the variability in 1-hour SC>2 concentrations, and these
measures of variability may be used to describe the potential variability in concentrations
measured at any ambient SC>2 monitor, similarly for either the 1-hour or 5-minute measured
concentrations. Note that there is a difference in the slope of the two lines, indicating that there
is not a constant relationship between the COV and GSD. This means that in characterizing the
variability at any ambient monitor, an identified COV (e.g., either low or high COV) does not
necessarily correspond to the same GSD characterization.
Next, staff compared the variability in 1-hour SO2 concentrations using data from the
monitors reporting 5-minute maximum SO2 concentrations (n=98) to variability observed for the
broader SO2 monitoring network (n=809). The objective of this evaluation was to determine if
the distribution of the observed hourly concentration variability was similar for the two sets of
data. As done above for the monitors reporting 5-minute maximum SO2 concentrations, four
statistics were generated for each ambient monitor within the broader SO2 monitoring network
using the 1-hour concentrations, with the variability at each monitor represented by its COV and
GSD. Figure 7-8 illustrates the cumulative density functions (CDFs) for the hourly COVs and
GSDs at each of the 98 monitors that reported 5-minute maximum SO2 concentrations (i.e., the
data set used for developing the statistical model) and the 809 monitors from the broader SO2
monitoring network (i.e., the final 1-hour SO2 data set having valid site-years). While the subset
of monitors reporting the 5-minute maximum SO2 concentrations exhibit greater variability in
hourly concentration at most percentiles of the distribution, the overall shape and span of the
distribution is very similar to that of the monitors within the broader SO2 monitoring network
using either variability metric. The similarity in variability distributions could indicate that the
monitor proximity to sources, the magnitude and temporal profile of source emissions, and the
types of sources affecting concentrations at either set of data (i.e., the monitors reporting 5-
minute SO2 concentrations versus the broader SO2 monitoring network) are similar. This,
combined with the meta-data evaluation and the source type, distance, and emissions analysis
that indicated similar source type emission proportions between the two sets of ambient
monitoring data (7.2.2), provides support for using concentration variability as a variable to
extrapolate information from the 5-minute SO2 monitors to the 1-hour SO2 monitors.
July 2009 93
-------
100%
90%
80%
= 70%
2
concentrations where only 1-hour average SC>2 concentrations were measured is the peak-to-
mean ratio (PMR). Peak-to-mean ratios are obtained by dividing the 5-minute maximum SCh
concentration occurring within an hour by the 1-hour SC>2 concentration. The use of a PMR or
distributions of PMRs in estimating 5-minute maximum SC>2 concentrations is not new to the
current NAAQS review. Both individual PMRs and distributions of PMRs were used in the
July 2009
94
-------
previous NAAQS review in characterizing 5-minute SC>2 air quality (Thrall et al, 1982; EPA,
1986a; 1994b; Thompson 2000) and in estimating human exposures to 5-minute SC>2
concentrations (Burton et al. 1987; EPA, 1986a, 1994b; Stoeckenius et al. 1990; Rosenbaum et
al., 1992; Science International, 1995). In this review, staff generated distributions of PMRs to
estimate 5-minute maximum SO2 concentrations at ambient monitors (this chapter) and at air
quality modeled census block centroid receptors (chapter 8). The distributions of PMRs used
here build upon recent PMR analyses conducted by Thompson (2000).33 In the current PMR
analysis, staff developed several distributions of PMRs using more recent 5-minute SO2
monitoring data (through 2007) and used concentration level and variability as categorical
variables in defining the distributions of PMRs.
Concentration variability has been identified as a potential attribute in characterizing
sources affecting concentrations measured at the ambient monitors (section 7.2.3.1). Instead of
designing a continuous function from the variability distribution, staff chose to use categorical
variables to describe the monitors comprising each data set. The approach involved the creation
of variability bins, such that PMR data from several monitors would comprise each bin. Staff
decided this approach would better balance the potential number of PMRs available in
generating the distributions of PMR given the variable number of samples collected and years of
monitoring at monitors that reported the 5-minute maximum SC>2 concentrations (Appendix A-
2). Using the hourly COV or GSD distributions in illustrated Figure 7-8, staff assigned one of
three COV or GSD bins to each of the 98 monitors reporting the 5-minute maximum SC>2
concentrations: for COV, the bins were defined as low (COV < 100%), mid (100% < COV <
200%), and high (COV > 200%). These three COV bins were selected to capture the upper and
lower tails of the variability distribution and a mid-range area.34 Similarly and based on the
same percentile ranges selected for binning the COV, three GSD bins were selected as follows:
low (GSD < 2.17), mid (2.17 < GSD < 2.94), and high (GSD > 2.94).
In addition, the level of the 1-hour mean SO2 concentration has been identified as an
important consideration in defining an appropriate PMR distribution to use in estimating 5-
minute maximum SO2 concentrations (EPA, 1986a). Therefore, staff further stratified the PMRs
33 In the Thompson (2000) analysis, a single distribution of PMRs was employed based on 6 ratio bins and assumed
independence between the ratio and the 1-hour SO2 concentration.
34 For monitors reporting the 5-minute maximum SO2 concentrations, these groupings corresponded to
approximately the 25th and the 84th percentile of the variability distribution.
July 2009 95
-------
by seven 1-hour mean concentration ranges: 1-hour mean < 5 ppb, 5 < 1-hour mean < 10 ppb, 10
< 1-hour mean < 25 ppb, 25 < 1-hour mean < 75 ppb, 75 < 1-hour mean < 150 ppb, 150 < 1-
hour mean < 250 ppb, and 1-hour mean > 250 ppb.35 Staff selected these 1-hour concentration
stratifications to maximize any observed differences in the PMR distributions within a given
variability and concentration bin and to limit the total possible number of PMR distributions for
computational manageability.
Based on the concentration variability and 1-hour concentration bins, staff generated a
total of 19 separate PMR distributions.36 Due to the large number of PMRs available for several
of the variability and concentration bins (the number of samples ranged from 100 to 800,000), all
of the empirical data were summarized into distributions using the cumulative percentiles
ranging from 0 to 100, by increments of 1. Figure 7-9 illustrates two patterns in the PMR
distributions when comparing the different stratification bins. First, the monitors with the
highest COVs or GSDs contain the highest PMRs at each of the percentiles of the distribution
(bottom graph of each variability bin in Figure 7-9) when compared with monitors from the other
two variability bins (top and middle graphs), while the mid-range variability bins (middle graph)
had a greater proportion of high PMRs than the low variability bin (top graph). These
distinctions in the PMR distributions are consistent with the results illustrated in Figure 7-7, that
is, the variability in the hourly average concentrations is directly related to the variability in the
5-minute concentrations as summarized across monitors.
Second, differences were observed in the PMR distributions within each variability bin
when stratified by 1-hour SC>2 concentration. This is most evident in the highest variability bin
(bottom graph of Figure 7-9); the highest 1-hour concentration category (> 250 ppb) had lower
PMRs at each of the distribution percentiles compared with the PMR distributions derived for the
lower concentration categories, most prevalent at the upper percentiles of the distribution. In
fact, the maximum PMRs for the > 250 ppb concentration bin were only 5.4 and 3.6 for the COV
and GSD high variability bin, respectively, compared with maximum PMRs of about 11.5 at
35 While PMR distributions were generated for 1-hour SO2 concentrations < 5 ppb, it should be noted that any
estimated 5-minute maximum SO2 concentration would be below that of the lowest potential health effect
benchmark level of 100 ppb.
36 Although there were a total of 21 PMR distributions possible (i.e., 3 x 7), the COV < 100% and GSD <2.17
categories had only three 1-hour concentrations above 150 ppb. Therefore, the two highest concentration bins do
not have a distribution, and concentrations > 75 ppb constituted the highest concentration bin in the low COV or low
GSD bins. All PMR distributions are provided in Appendix A-3.
July 2009 96
-------
many of the other concentration bins. Again, this inverse relationship between the PMR and
concentration level has been shown by other researchers (EPA, 1986a). The stratification of
PMRs by the 1-hour concentration was done to avoid applying high PMRs calculated from low
hourly concentrations to high hourly concentrations. The observed patterns in the PMR
distributions support the staff selection of variability bins and 1-hour concentration stratifications
in controlling for the aberrant assignment of PMRs to particular 1-hour concentrations.
COV Bins
GSD Bins
COV < 100%
— 1-hour < 5 ppb
— 5 < 1-hour < 10 ppb
— 10 < 1-hour < 25 ppb
— 25 < 1-hour < 75 ppb
— 1-hour > 75 ppb
— 1-hour < 5 ppb
— 5 < 1-hour < 10 ppb
— 10 < 1-hour < 25 ppb
— 25 < 1-hour < 75 ppb
— 1-hour > 75 ppb
100 250 ppb
— 1-hour < 5 ppb
— 5 < 1-hour < 10 ppb
— 10 < 1-hour < 25 ppb
— 25 < 1-hour < 75 ppb
— 75 < 1-hour < 150 ppb
— 150 < 1-hour < 250 ppb
— 1-hour > 250 ppb
COV > 200%
GSD > 2.940/
1-hour < 5 ppb
— 5 < 1-hour < 10 ppb
10 < 1-hour < 25 ppb
25 < 1-hour < 75 ppb
75 < 1-hour < 150 ppb
— 150 < 1-hour < 250 ppb
1-hour > 250 ppb
1-hour < 5 ppb
— 5 < 1-hour < 10 ppb
10 < 1-hour < 25 ppb
25 < 1-hour < 75 ppb
75 < 1-hour < 150 ppb
150 < 1-hour < 250 ppb
1-hour > 250 ppb
567
Peak to Mean Ratio (PMR)
4567
Peak to Mean Ratio (PMR)
Figure 7-9. Peak-to-mean ratio (PMR) distributions for three COV and GSD variability bins and
seven 1-hourSO2 concentration stratifications.
July 2009
97
-------
Staff then evaluated the assigned concentration variability bin using two ambient
monitoring site characteristics described in section 7.2.2 and using the observed number of
benchmark exceedances at each monitor. The purpose of this analysis was to determine to what
extent the selected variability bins were representing variability local source characteristics and
the likelihood of benchmark exceedances. First, staff compared the total emissions within 20 km
of each monitor with the assigned concentration variability bin using the monitors reporting 5-
minute maximum SC>2 concentrations and the broader SC>2 monitoring network (Figure 7-10).
The purpose of this comparison was to determine whether increased emissions were associated
with greater variability in monitoring concentrations. In general, a pattern of increased
emissions was associated with an increase in the concentration variability bin, though the pattern
was more prominent when considering the COV bins. This indicates the variability bins may be
useful as a surrogate for local source emission characteristics.
= 12
£ 2 250000
•>S 200000
Q. CM
iiO
<2 ^ 150000
o .E
y, E 100000
E?
m ° 50000
«|
£ o o~
w\_/v DIMS
n = 21 n = 54 n = 16
n
n
T ^
i — i — i
i — ± — i ^_
-
\joi_r E^ina
n = 21 n = 55 n = 15
n
n — i —
i 1 _^ -^—
is 250000^
.*; o
3 •-
^ = 200000
^"Q" 150000
0W
•3; ^ 100000
E "°
£ £ 50000
— ^
P (M 0J
n = 196 n = 462 n = 80
n
g
D
D
B ~r
1
T
T +
1 — ± — 1 _i_ _i_
n = 228 n = 435 n = 75
n n
Low
Mid
High
Low
Mid
High
Figure 7-10. Distribution of total SO2 emissions (tpy) within 20 km of monitors by COV (left) and
GSD (right) concentration variability bins: monitors reporting 5-minute maximum SO2
concentrations (top) and the broader SO2 monitoring network (bottom).
July 2009
98
-------
The second ambient monitoring site characteristic evaluated using the selected
concentration variability bins was the monitoring objective, principally when it was noted as
source-oriented. The purpose of this analysis was to determine whether high variability in 862
concentration was related to source-oriented monitor siting. Staff calculated the percent of
source-oriented monitors in each variability bin for the two sets of data; the set comprised of
monitors that reported 5-minute maximum SC>2 concentrations and those within the broader SC>2
monitoring network. In general, there is an increasing percent of source-oriented monitors in the
higher concentration variability bins when using either the COV or GSD metrics (Figure 7-11),
though the pattern is more consistent with the COV metric than with the GSD metric. This
comparison also indicates that the concentration variability metric may be useful as a surrogate
for local source emission characteristics.
T3
3
"c
0
'C
O
0
3
O
il
12 $
50
c
O
30
25
20
15 :=:=:=:
• COV (5-Minute SO2 Monitors)
HGSD (5-Minute SO2 Monitors
D COV (All SO2 Monitors)
Q GSD (All SO2 Monitors)
Low
Mid
Variability Bin
High
Figure 7-11. Percent of monitors within each concentration variability bin where the monitoring
objective was source-oriented: monitors reporting 5-minute maximum SO2
concentrations (solid) and the broader SO2 monitoring network (slotted).
July 2009
99
-------
Staff evaluated the number of measured benchmark exceedances in a site-year given the
variability bins used to characterize the ambient monitors. The purpose of this analysis was to
determine whether monitors exhibiting greater variability in SC>2 concentration also have a
greater number of benchmark exceedances. Figure 7-12 summarizes the distribution of
exceedances of the 200 and 400 ppb benchmark level by each of the COV and GSD variability
bins (patterns for the 100 ppb and 300 ppb benchmarks were similar). Clearly, monitors having
the greatest variability in 1-hour SC>2 concentration are the monitors most likely to have 5-minute
SC>2 benchmark exceedances and a greater number of exceedances per year. This analysis
provides further support to the binning of monitors by concentration variability to appropriately
extrapolate the relationships derived from monitors reporting 5-minute maximum concentrations
to monitors reporting only 1-hour 862 concentrations (and at the dispersion model receptors).
>200 ppb >400 ppb
cT 10°-
ETO go "
0)
3 >- f*
E fc S 60
'x Q- >
TO (fl Q 4Q -
2 0) *•
11 --
c -a
'£ o> °
i (_)
«> X 100
'TO if so
Q TO
"Si I 60
J- O Q
r> S f"i 40 '
E m ~
1
0
i
. i
_|_ C±3 ^
1
_
i
c
n
= i
X i b
3
BO
50
40
30
20
10
0
60
50
40
30
20
10
0
§
B
°
|
a
B
B
^ i D
B
D
°
J
fi
I
_s_ J- n
low mid high low mid
Concentration Variability Bin
high
Figure 7-12. Distribution of the measured number of daily 5-minute maximum SO2 concentrations
above 200 ppb (left) and 400 ppb (right) in a year by hourly concentration COV (top) and
GSD (bottom) variability bins. Data were from the 98 ambient monitors reporting 5-
minute maximum concentrations (471 site-years).
July 2009
100
-------
7.2.3.3 Application of Peak to Mean Ratios (PMRs)
As described above in section 7.2.3.2 regarding the monitors reporting 5-minute
maximum 862 concentrations, staff characterized the monitors within the broader 862
monitoring network (n=809) by their respective hourly concentration variability and assigned to
one of the three COV bins (COV < 100%, 100% < COV < 200%, and COV > 200%) and GSD
bins (GSD < 2.17, 2.17 < GSD < 2.94, and GSD > 2.94). Based on the monitor's assigned
concentration variability bin (either from the COV or GSD, not mixed) and the 1-hour SC>2
concentration, PMRs can be randomly sampled37 from the appropriate PMR distribution to
estimate a 5-minute maximum SC>2 concentration using the following equation:
Cmax_5 =PMRV xC,^ equation (7-1)
where,
Cmax-s = estimated 5-minute maximum SO2 concentration (ppb) for each hour
PMRy = peak-to-mean ratio (PMR) randomly sampled from the /' concentration
variability andy 1-hour mean SC>2 concentration distribution
Ci,i-hour= measured 1-hour average SC>2 concentration at an /' concentration
variability monitor
As a result of this calculation, every 1-hour ambient 862 concentration has an estimated
_ -minute maximum SC>2 concentration.38 These statistically modeled 5-minute maximum SC>2
concentrations were then summarized using the output metrics described in section 7.2.5.
5
7.2.3.4 Evaluation of Statistical Model Performance
Staff evaluated the performance of the statistical model using cross-validation (Stone,
1974). Details of the evaluation are provided by Langstaff (2009). Briefly, PMR distributions
were estimated using 97 of the 98 monitors that reported both the 1-hour and 5-minute maximum
SC>2 concentrations. All ambient monitors were characterized using the same variability bins
described in section 7.2.3.2. The 1-hour concentrations were also characterized using the same
37 The random sampling was based selection of a value from a uniform distribution {0,100}, whereas that value was
used to select the PMR from the corresponding distribution percentile value.
38 When the 1-hour SO2 concentration was > 0, otherwise the 5-minute maximum SO2 concentration was estimated
as zero).
July 2009 101
-------
stratifications discussed earlier. Then staff used the newly constructed PMR distributions from
the 97 monitors and equation 7-1 to predict the 5-minute maximum 862 concentrations at the
single monitor not included in developing the PMR distributions. This modeling was performed
98 times, i.e., removing every single monitor (one monitor at a time), generating new PMR
distributions, and predicting 5-minute maximum SC>2 concentrations at the removed monitor.
Staff then compared the predicted and measured daily 5-minute maximum SO2 concentrations to
generate a distribution of model prediction errors (e.g., median errors, median absolute errors)
and general model statistics (i.e., the root mean square error or RMSEs, and R2, a measure of the
amount of variance explained by the model).
Four statistical models were evaluated: two models constructed from the variability bins
(either COV or GSD) using all percentiles of the PMR distributions, and two similar models
constructed without the minimum and maximum percentiles of the PMR distributions. The
models were evaluated at the benchmark concentration levels as well as at selected percentiles in
the 5-minute SC>2 concentration distribution. In comparing the model predictions, the model
using variability bins defined by the COV and excluding the minimum and maximum percentiles
had the lowest prediction errors (e.g., see Table 7-5).39 Based on these results, staff used this
COV model (excluding the 0th and 100th percentiles of the PMR distribution) to estimate 5-
minute maximum SO2 concentrations from 1-hour SO2 concentrations.
39 Table 7-5 presents a few of the prediction error statistics used to compare each of the models, though several other
prediction errors were evaluated (e.g., the 75th and 99th). Results for the other percentiles were consistent with
median results discussed in the text, that is the alt. COV model had the lowest error when compared with the other
models evaluated. See Langstaff (2009) for the additional percentile comparisons for each of the models.
July 2009 102
-------
Table 7-5. Comparison of prediction errors and model variance parameters for the four models
evaluated.
Benchmark Level
(ppb)
100
200
300
400
Model1
cov
alt. COV
GSD
alt. GSD
COV
alt. COV
GSD
alt. GSD
COV
alt. COV
GSD
alt. GSD
COV
alt. COV
GSD
alt. GSD
Median
Prediction Error2
2.6
0.4
2.5
0.3
1
0.1
1.3
0.4
0.6
0
0.6
0.1
0.3
0
0.3
0
RMSE
18.9
14.1
24.8
19.8
10.7
8.6
12.8
10.2
6.5
5.6
8.2
7.1
4.5
3.9
6
5.5
R2
0.72
0.81
0.48
0.63
0.66
0.74
0.49
0.64
0.73
0.78
0.55
0.64
0.76
0.8
0.55
0.61
Notes:
1 The "alt." abbreviation denotes the alternative model was used: the minimum and
maximum percentiles of the PMR distributions were not used.
2 The absolute value of the prediction differences is calculated (predicted minus the
observed number of exceedances in a year), generating a distribution of prediction
errors. The value reported here is the (50th percentile) of that distribution.
Staff performed supplementary evaluations using the prediction errors associated with the
selected statistical model. Additional percentiles of the prediction error distribution were
calculated to estimate the magnitude and direction of the statistical model bias. Table 7-6
summarizes the prediction errors for each benchmark level. When considering paired percentiles
(e.g., the 25th and the 75th or prediction intervals) and the 50th percentile as a pivot pointthere
appears to be an over-estimation bias at each of the benchmark levels. For example, there is a
greater overestimation of the 400 ppb benchmark level at the 95th percentile (i.e., 5 exceedances),
than compared with the under estimation at the 5th percentile (i.e., one exceedance). However,
there is good agreement in the predicted versus observed number of exceedances, whereas 90%
of the predicted exceedances of 400 ppb were within -1 to 5 exceedances per year. There is a
wider range in the prediction intervals at the lower benchmark levels, partly a function of the
July 2009
103
-------
greater number of exceedances at the lower benchmark levels rather than the degree of
agreement (Table 7-6). At the extreme ends of the distribution for each of the benchmarks, the
agreement between the predicted and observed exceedances widens, indicating that for some
site-years (approximately 2%), the number of days with a benchmark exceedance can be over- or
under-estimated by 20 to 50 in a year.
Table 7-6. Prediction errors of the statistical model used in estimating 5-minute maximum SO2
concentrations above benchmark levels.
Percentile
1
5
25
50
75
95
99
Prediction Error at Benchmark Level1
100
-31
-15
-1
0
7
32
48
200
-17
-7
0
0
1
20
43
300
-18
-3
0
0
1
10
26
400
-19
-1
0
0
0
5
14
Benchmark
Observed
Predicted
Mean Number of Benchmark Exceedances2
100
148
150
200
81
100
300
69
67
400
56
45
Notes:
1 The percentiles are based on the distribution of predicted minus the
observed values for each benchmark. Units are the number of
exceedances per year.
2 This is the average of all site-years. Units are the number of
exceedances per year.
7.2.4 Adjustment of Ambient Concentrations to Evaluate the Current and Potential
Alternative Air Quality Scenarios
Staff evaluated multiple hypothetical air quality scenarios in this assessment, each
defined by the form and level of a selected standard. Collectively, the purpose of these air
quality scenarios was to estimate the relative level of public health protection associated with just
meeting the current and potential alternative standards. The measured ambient SO2
concentrations needed adjustment to reflect concentrations that might be observed given the
hypothetical air quality scenarios. To maintain a computationally manageable data set given the
number of air quality scenarios (i.e., eight) and potential health effect benchmark levels
investigated (i.e., four), staff used the recent ambient monitoring data from 40 counties,
July 2009
104
-------
specifically years 2001 through 2006.40 The following two sections discuss the concentration
adjustment approach and the selection criteria used for selecting counties for analysis.
7.2.4.1 Approach
There are two important considerations in developing an approach to adjust air quality
concentrations. One is the relative contribution of policy-relevant background (PRB) to ambient
concentrations and the other is in understanding how the distribution of ambient concentrations
measured at a particular monitor has changed over time.
In developing a simulation approach to adjust air quality to meet a particular standard
level, PRB levels in the U.S. were first considered. As described in section 2.3, PRB is well
below concentrations that might cause potential health effects and constitutes a small percent
(<1%) of the total ambient 862 concentrations at most locations. Based on the small
contribution, PRB will not be considered separately in any characterization of health risk
associated with as is air quality or air quality just meeting the current or potential alternative
standards. In monitoring locations where PRB is expected to be of particular importance
however (e.g., Hawaii County, HI), data were noted by staff as influenced by significant natural
sources rather than anthropogenic sources and were not used in any of the air quality analyses.
While annual average concentrations have declined significantly over time, the
variability in the SC>2 concentrations (both the 5-minute and 1-hour concentrations) has remained
relatively constant. This trend is present when considering ambient concentration data
collectively (section 7.4.2.3) and when considering monitors individually (Rizzo, 2009). For
example, Figure 7-13 compares the distribution of daily maximum 862 1-hour concentration
percentiles at the two ambient monitors in Beaver County, Pa. that were in operation as far back
as 1978 and are currently part of the broader SO2 monitoring network. Staff selected a recent
year of data (2007) to constitute a low concentration year along with an historical year of data
(1992) constituting a high concentration year, with each year of ambient monitoring common to
both monitors. As shown in Figure 7-13, the relationships between the low and high
concentration years at each of the daily maximum concentration percentiles are mostly linear,
with regression coefficients of determination (R2 values) greater than 0.98. Where deviation
40 As described in the section 7.2.1, at the time the 1-hour concentrations were downloaded, none of the monitors
had a complete year of data for 2007. All data from 2007 were excluded from the 1-hour monitor simulations.
July 2009 105
-------
from linearity did occur (as was observed in many of the other low-to-high concentration
comparisons performed), it occurred primarily at the extreme upper or lower portions of the
distribution, often times at the maximum daily maximum or the minimum daily maximum 1-
hour SC>2 concentration (Rizzo, 2009). In addition, the absolute values for the simple linear
regression intercepts were typically 1-3 ppb (Rizzo, 2009). This indicates that the rate of
decrease in ambient air quality concentrations at the mean value for the monitors evaluated is
consistent with the rate of change at the lower and upper daily maximum 1-hour concentration
percentiles. This evaluation provides support for the use of a proportional approach to adjust
current ambient concentrations to represent air quality under both the current and alternative
standard scenarios.
High Year: 1992 Low Year: 2007
000 0.05 010 015 0.20 025 030
0.20 -
CL
Q.
c
o
O
a>
o> 0.10
03
0)
q 0.05 -
0.00 -
420070002
RA2: 0.99
420070014
RA2: 0.98
000 005 010 015 0.20 025 0.30
High Year Percentile Cone, (ppm)
Figure 7-13. Comparison of measured daily maximum SO2 concentration percentiles in Beaver
County, PA for a high concentration year (1992) versus a low concentration year (2007)
at two ambient monitors (from Rizzo, 2009).
The current deterministic form of each standard was used to approximate concentration
adjustment factors to simulate just meeting the current 24-hour and annual 862 NAAQS. The
July 2009
106
-------
24-hour standard of 140 ppb is not to be exceeded more than once per year, therefore, the 2nd
highest 24-hour average observed at each monitor was used as the target for adjustment. The
rounding convention, which is part of the form of the standard, defines values up to 144 ppb as
just meeting the 24-hour standard. The form of the current annual standard requires that a level
of 30 ppb is not to be exceeded; therefore, with a rounding convention to the fourth decimal,
annual average concentrations of up to 30.4 ppb would just meet the current annual standard.
Staff limited the analysis of alternative air quality scenarios to particular locations using
designated geographic boundaries (not just the monitors individually). Counties were used to
define the locations of interest in the alternative air quality standard scenarios. Use of a county is
consistent with current policies on the designation of appropriate boundaries of non-attainment
areas (Meyers, 1983).
For each location (/') and year (/), 24-hour and annual 862 concentration adjustment
factors (F) were derived by the following equation:
Fy=S/Cm^ equation (7-2)
where,
Fy = Adjustment factor derived from either the 24-hour or the annual
average concentrations at monitors in location /' for year 7 (unitless)
S = concentration values allowed that would just meet the current NAAQS
(either 144 ppb for 24-hour or 30.4 ppb for annual average)
Cmax,ij = the maximum 2nd highest 24-hour average 862 concentration at a
monitor in location /' and year7 or the maximum annual average 862
concentration at a monitor in location / and year7 (ppb)
In adjusting concentrations to just meet the current standard, the highest monitor (in
terms of concentration) within a county was adjusted so that it just meets either a 30.4 ppb
annual average or a 144 ppb 24-hour average (2nd highest), whichever was the controlling
standard.41 For monitors in each county and calendar year, all hourly 862 concentrations were
41 The controlling standard by definition would be the standard that allows air quality to just meet either the annual
concentration level of 30.4 ppb (i.e., the annual standard is the controlling standard) or the 2nd highest 24-hour
concentration level of 144 ppb (i.e., the 24-hour standard is the controlling standard). The factor selected is derived
July 2009 107
-------
multiplied by the same constant value F, though only one monitor would have an annual mean
equal to 30.4 ppb or the 2nd highest 24-hour average equal to 144 ppb for that county and year.
For example, of five monitors measuring hourly 862 in Cuyahoga County for year 2001
(Figure 7-14, top), the maximum annual average concentration was 7.5 ppb (ID 390350060),
giving an adjustment factor of F = 30.4/7.5 = 4.06 for that year. The 2nd highest 24-hour SO2
concentration at a monitor in a year was 35.5 (ID 390350038), giving an adjustment factor of F =
144/35.5 = 4.05 for year 2001. Because the adjustment factor derived from the 24-hour average
concentration was lower, the 24-hour average concentration was the controlling standard. All 1-
hour concentrations measured at all five monitoring sites in Cuyahoga County were multiplied
by 4.05, resulting in an upward scaling of hourly 862 concentrations to simulate air quality just
meeting the current standard for that year. Therefore, one monitoring site in Cuyahoga County
for year 2001 would have a 2nd highest 24-hour average concentration of 144 ppb, while all other
monitoring sites would have a 2nd highest 24-hour average concentration below that value,
although still proportionally scaled up by 4.05 (Figure 7-14, bottom).
Proportional adjustment factors were also derived considering the form, averaging time,
and levels of the potential alternative standards under consideration. Discussion regarding the
staff selection of each of these components of the potential alternative standards is provided in
Chapter 5 of this document. The 98th and 99th percentile 1-hour daily maximum SC>2
concentrations averaged across three years of monitoring were used in calculating the adjustment
factors at each of five standard levels as follows:
f 3 A
Vr-
Fm =
equation (7-3)
where,
= SC>2 concentration adjustment factor in location /' given alternative standard
percentile form k and standard level / across a 3-year period (unitless)
from a single monitor within each county (even if there is more than one monitor in the county) for a given year. A
different (or the same) monitor in each county could be used to derive the factor for other years; the only
requirement for selection is that it be the lowest factor, whether derived from the annual or 24-hour standard level.
July 2009
108
-------
Si = Standard level / (i.e., 50, 100, 150, 200, and 250 ppb 1-hour SO2 concentration)
= Selected percentile k (i.e., 98th or 99th) 1-hour daily maximum 862
concentration at a monitor in location /' for each yeary (ppb)
CUYAHOGA COUNTY AIR QUALITY (2001)
(As Is)
Cumulative Percentile
-iNjOJ-e-cnro-Nioococ
3OOOOOOOOOC
100 -
90 -
80 -
1 70-
CD
| 60-
Q.
o> 50 -
« 40 -
3
| 30 -
O
20 -
10 -
0.
1
C
1
i
i
>
i
K--- t
-
'_____
1
1
J
k H
A
t
—
r
i
> \
\-
>
— -
V
t
. +•
tf
-t-
f
><
•
tj
f
k'
I
1
K<
'
-
_•
—
.•
— -
MM
— •
••
•P
-
m
-
m
-
r*
--
M 1* ' ' ' •-— •--•—••-•-
----
« CUY390350038
•• CUY390350045
• CUY390350060
x CUY390350065
+ CUY390356001
CUYAHOGA COUNTY AIR QUALITY (2001)
Adjusted to Just Meet the Current Standard (CS)
3
J
I
f
«
t
'
L
1-
»
:
a
X
*
*
„«*
t
*:
h*
"t&
f
I
41 CUY390350038_cs
n CUY390350045_cs
* CUY390350060_cs
x CUY390350065_cs
-i- CUY390356001_cs
10 100 1000
1-hour SO2 (ppb)
Figure 7-14. Distributions of hourly SO2 concentrations at five ambient monitors in Cuyahoga
County, as is (top) and air quality adjusted to just meet the current 24-hour SO2
standard (bottom), Year 2001.
July 2009
109
-------
As described above for adjustments made in simulating just meeting the current
standards, the highest monitor (in terms of the 3-year average at the 98th or 99th percentile) was
adjusted so that it just meets the level of the particular 1-hour alternative standard. All other
monitor concentrations in that location were adjusted using the same factor, only resulting in
concentrations at those monitors below the level of the selected 1-hour alternative standard.
Since the alternative standard levels range from 50 ppb through 250 ppb, both proportional
upward and downward adjustments were made to the 1-hour ambient SC>2 concentrations. Due
to the form of the alternative standards, the expected utility of such an analysis, and the limited
time available to conduct the analysis, only the more recent air quality data were used (i.e., years
2001-2006). The 1-hour ambient SC>2 concentrations were adjusted in a similar manner
described above for just meeting the current standard, however, due to the form of these
standards, only one factor was derived for two 3-year periods (i.e., 2001-2003, 2004-2006),
rather than one factor for each calendar year.
7.2.4.2 Selection of Locations
The first criterion used to select locations for the alternative air quality analyses was
whether monitors had a high number of daily 5-minute maximum SCh concentrations at or above
the potential health effect benchmark levels. Ambient monitors located in two counties in
Missouri (Iron and Jefferson) had the most frequently measured daily 5-minute maximum SC>2
concentrations above the potential health effect benchmarks (see Appendix A-5). While there
were limited data available from these ambient monitors (4 and 2 years out of 8 total site-years
did not met the completeness criteria for each of Jefferson and Iron counties, respectively), it was
decided by staff that lack of a complete year should not preclude their use in this focused
analysis given the high number of measured daily 5-minute maximum SO2 concentrations at
these monitors. All other monitoring data used in this focused analysis were selected from where
1-hour ambient monitoring met the completeness criteria described in section 7.2.1.
Staff selected an additional 38 counties based on the relationship of the ambient SC>2
concentrations within the county to the current annual and 24-hour NAAQS to expand the
number of counties investigated to a total of 40.42 An additional criterion to be met for county
selection included having at least two monitors operating in the county for at least five of the six
42 In the 1st draft SO2 REA, a total of 20 counties were selected to evaluate the current standard scenario only.
July 2009 110
-------
possible years of monitoring.43 First, the 24-hour and annual concentration adjustment factors
were derived by equation 7-2 for each county and year. Then the mean 24-hour and mean annual
factor for each county was calculated by averaging the site-years available at each monitor, with
the selection of the lowest mean factor retained to characterize the county. Each county was then
ranked in ascending order based on this selected mean factor. The 38 counties were selected
from the top 38 values, that is, those counties having the lowest mean adjustment factors and
having at least two monitors.
The complete list of the 40 counties selected and the mean factors used to select each
location given the above selection criteria are provided in Table 7-7. In addition, Table 7-7 gives
the number of monitors in each COV bin that were used to characterize the air quality in the 40
counties. The locations of ambient monitors comprising the 40 county dataset (i.e., the third data
analysis group) are illustrated in Figure 7-15. Compared with the two other data analysis groups,
the 40 county data set has a greater number of mid and high COV bin monitors and notably
fewer low COV bin monitors (Figure 7-16). This is not unexpected given the concentration-
based selection criteria used in identifying the 40 counties.
Following the selection of the 40 counties, staff retained the adjustment factors
calculated for each monitoring site-year (not simply the mean factor that was used for the county
selection) to simulate air quality just meeting the current standard (either the daily or annual
factor, whichever was lower). These adjustment factors are given in Appendix A, Table A.4-1.
Then using equation 7-3, staff calculated the adjustment factors needed for evaluating the
potential alternative standards. Each of these alternative air quality scenarios were used as an
input to the statistical model to estimate 5-minute maximum SO2 concentrations (equation 7-1).
Then, air quality characterization metrics of interest were estimated for each site and year as
described in section 7.2.5.
43 In the 1st draft SO2 REA, having at least three monitors for all six years of the monitoring period was required.
These earlier criteria were relaxed in the 2nd draft and in this final REA to allow for additional locations that may
have ambient concentrations close to the current annual and daily standard levels.
July 2009 111
-------
Table 7-7. Counties selected for evaluation of air quality adjusted to just meeting the current and
potential alternative SO2 standards and the number of monitors in each COV bin.
State
Arizona
Delaware
Florida
Iowa
Illinois
Indiana
Michigan
Missouri
New Hampshire
New Jersey
New York
Ohio
Oklahoma
Pennsylvania
Tennessee
Texas
Virginia
US Virgin Islands
West Virginia
County1
Gila
New Castle
Hillsbo rough
Linn
Muscatine
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Wayne
Greene
Iron"
Jefferson3
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
St Croix
Brooke
Hancock
Monongalia
Wayne
Mean Factor
3.44
2.80
3.81
3.58
3.46
3.78
3.39
4.38
2.60
4.41
4.80
3.13
4.47
5.49
3.53
2.98
3.90
3.81
3.09
4.19
3.17
4.51
2.99
3.13
4.61
2.65
2.39
3.26
1.74
3.19
1.86
4.08
3.45
4.38
4.80
4.60
2.32
2.32
2.93
3.07
Closest
Standard2
A
D
D
D
D
D
D
D
D
D
D
D
D
A
D
D
A
A
A
D
D
A
D
D
A
D
D
A
D
A
D
D
D
D
A
D
A
A
D
D
# of Monitors in COV bin4
Low
2
2
2
2
2
1
2
1
3
1
Mid
1
5
4
3
1
4
2
2
2
3
1
3
1
1
5
2
2
3
5
3
1
2
1
2
2
2
3
2
2
9
2
3
High
1
2
2
2
2
1
2
2
2
4
1
1
1
3
Notes:
1 Listed counties were selected based on lowest mean concentration adjustment factor, derived from
at least 2 monitors per year for years 2001-2006 and >5 years of data.
2 Ambient concentrations were closest to either the annual (A) or daily (D) NAAQS level.
3 County selected based on frequent 5-minute benchmark level exceedances.
4 COV bins were low (COV<100%); mid (100%200%).
July 2009
112
-------
Figure 7-15. Locations of the 128 ambient monitors comprising the 40 County data set (i.e., the
third data analysis group).
5-Minute Max SO2 Monitors (n=98) Broader SO2 Network (n=809)
40 County SO2 Monitors (n=128)
High COV
16%
Low COV
24%
Mid COV
60%
High COV
11%
High COV
Low COV 20%
27%
Low COV
14%
Mid COV
62%
Mid COV
66%
Figure 7-16. Percent of monitors in each COV bin for the three data analysis groups: monitors
reporting 5-minute maximum SO2 concentrations, the broader SO2 monitoring network,
and SO2 monitors selected for detailed analysis in 40 counties.
July 2009
113
-------
7.2.5 Air Quality Concentration Metrics
For each of the data analysis groups and air quality scenarios considered, several
concentration metrics were calculated; these included the annual average, 24-hour, and 1-hour
daily maximum 862 concentrations for each site-year of data and the number of exceedances of
the potential health effect benchmark levels. The numbers of daily maximum 5-minute
concentration exceedances in a year were counted (i.e., either 1 or none per day) rather than total
number of exceedances (i.e., which confounds numbers of exceedances and days with
exceedances). To characterize the relationship between the number of days with a 5-minute
benchmark exceedance and the ambient concentration levels, staff generated two additional
outputs given the different concentration averaging times.
The first output was a comparison of the annual average 862 concentration and the
number of daily 5-minute maximum SC>2 concentrations above the benchmark levels in a year.
The output of this is the number of days per year a monitor had a measured or modeled
exceedance, given an annual average SO2 concentration. In general, these results are graphically
depicted in this REA, though most of the individual results displayed in the figures are provided
in Appendix A-5. When considering the 40 counties used for detailed analysis, the results are
presented at the county-level, some of which had multiple ambient monitors. Therefore, the
results for the monitors within counties were aggregated to generate mean values representing
the central tendency of the county's annual average concentrations and the numbers of days in a
year with benchmark exceedances.
The second output was the probability of potential health effect benchmark exceedances
given concentrations of short-term averaging times. It was proposed in Chapter 5 that the 1-hour
daily maximum SO2 concentration would be of an appropriate averaging time in controlling the
number of daily 5-minute maximum SC>2 concentrations. Staff evaluated such a relationship
using the measured 5-minute and 1-hour ambient SC>2 concentrations to determine if this indeed
was the case. A tally was made every time a daily 5-minute maximum SC>2 concentration
occurred during the same hour of the day as the 1-hour daily maximum SC>2 concentration. The
results of this analysis, separated by benchmark exceedance level, are given in Table 7-8. The
co-occurrence of the daily 5-minute maximum and the 1-hour daily maximum 862
July 2009 114
-------
concentrations is greater than 70% at each of the benchmark levels indicating a strong
relationship between the two concentration averaging times.
Table 7-8. The co-occurrence of daily 5-minute maximum and 1-hour daily maximum SO2
concentrations using measured ambient monitoring data.
Concentration/Level
All concentrations
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Co-occurring 5-
minute and 1-hour
daily maxim urns1
(n)
106,115
6,192
2,030
1,067
700
Total Paired
Samples2
(n)
130,296
8,817
2,793
1,476
961
Percent
Co-occurring
(%)
81.4
70.2
72.7
72.3
72.8
Notes:
1 the number of events the 5-minute maximum occurred in the same hour as the 1 -hour
daily maximum.
2 total events with both a 5-minute maximum and 1-hour SO2 concentration
measurement.
Given the form of the current 24-hour standard, the form of the potential alternative
standards (1-hour daily maximum), and the frequency of 5-minute SC>2 benchmark exceedances
(i.e., either one or none per concentration), staff generated probability functions to estimate the
likelihood of a 5-minute benchmark exceedance. These functions are useful in estimating the
probability of a 5-minute benchmark exceedance given a range of SC>2 concentrations at
alternative averaging times (i.e., either a 24-hour average or 1-hour daily maximum
concentration). Two approaches were used to generate the probability functions: the first was
empirically-based while the second employed a logistic regression model.
To generate the empirically-based probability functions, concentration data were first
stratified into bins using concentration midpoints, with each bin separated by 10 ppb. For
example a concentration of 53 ppb would be included in the 50 ppb bin, while a concentration of
55 ppb would fall within the 60 ppb bin. Then, the presence or absence of a daily 5-minute
benchmark exceedance given the number of values in each concentration bin (that originate from
all monitored concentrations within the bin range) was used to estimate the probability of an
exceedance. For example, if there were 105 exceedances of the 200 ppb benchmark level out of
July 2009
115
-------
44
239 instances of a 1-hour daily maximum binned concentration of 110 ppb , the probability of a
200 ppb benchmark exceedance would be 105/239 = 0.44 or 44 % given a 1-hour daily
maximum concentration of around 110 ppb. An example of an output from this empirically-
based probability function is illustrated in Figure 7-17 for each of the four benchmark levels.
100 H
_ 90 -
3
.1 80 +
x
re
= — 60 -F
EO)
, o
in =
•>. J5 50
40
30
20 -
10 -'-
o
n- X
O 111
£1
re
1
Q.
> 100 ppb
> 200 ppb
> 300 ppb
- - - > 400 ppb
50 100 150 200 250 300 350
Daily Maximum 1-hour SO2 Concentration (ppb)
400
Figure 7-17. Example of empirically-based probability curves. The probability of a 5-minute SO2
benchmark exceedance (P) was estimated by dividing the number of days with an
exceedance by the total number of days within each 1-hour daily maximum SO2
concentration bin.
In constructing the empirical probability curves, staff noted there were fewer samples
with increasing concentrations (either 1-hour daily maximum or 24-hour average). Having too
few samples generated instability in the empirically-based probability curves at the highest 1-
hour daily maximum or 24-hour average concentrations. For example, there were very few
measured 1-hour daily maximum 862 concentrations above the 130 ppb bin considering the high
44 Therefore, there were 134 instances whereby the 1-hour daily maximum of 110 ppb did not correspond to a 5-
minute maximum concentration above 200 ppb.
July 2009
116
-------
population density group (Table 7-9). A total of 116 1-hour daily maximum SC>2 concentrations
out of 26,983 were scattered across the bins of 140 through 620 ppb, concentrations associated
with the presence or absence of a 300 ppb 5-minute benchmark exceedance. There were
increasing probabilities of 5-minute benchmark exceedances with increasing 1-hour daily
maximum SO2 concentration starting at 100 ppb; however, at 170, 210, and 230 ppb there were
lower estimated probabilities of exceedances than the preceding lower 1-hour daily maximum
SC>2 concentration. If using the probability data alone in Table 7-9, this would imply that at 1-
hour daily maximum concentrations of about 210-230 ppb, the likelihood of an exceedance is
less than that when considering 1-hour daily maximum concentrations between 190-200 ppb.
This is likely not the case, and in this instance, the wide range in estimated probabilities are more
a function of the small sample sizes (no more than 3 samples per bin in this case) rather than the
1-hour daily maximum SC>2 concentrations. Therefore, in viewing the occurrence of this issue at
small sample sizes, staff selected concentration bins having at least thirty 1-hour daily maximum
(or 24-hour average) concentrations (whether it was all, none, or a mixture of exceedances) for
inclusion in the empirically-based probability curves. As a result, the sample size limits
compressed the range of predictability offered by the empirically-based probability curves. As
an example, Figure 7-17 indicates that there were fewer than 30 samples available for
concentration bins above a 1-hour daily maximum SC>2 concentration of 200 ppb (note the 200
ppb bin contained 37 samples).
July 2009 117
-------
Table 7-9. Example of how the probability of exceeding a 400 ppb 5-minute benchmark would be
calculated given 1-hour daily maximum SO2 concentration bins.
Daily Maximum
1-hour bin
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
250
Number of times:
With no
exceedances
71
45
43
34
17
15
11
10
8
1
1
1
1
2
0
0
With one
exceedance
0
2
1
1
1
2
4
2
3
4
3
0
2
0
2
2
Probability of
Exceedance
(%)
0
4
2
3
6
12
27
•*17<-
27
80
75
-»0<-
67
-»0<-
100
100
Notes:
-> % <- notes sharp decrease in probability from prior concentration
bin.
Data used in this table is from the high population density monitors
reporting 5-minute concentrations.
In the second approach, we generated probability curves for each of the four benchmark
levels and the time-averaged SC>2 concentrations (i.e., 1-hour daily maximum or 24-hour average
concentration) usingproc logistic and a probit link function (SAS, 2004). The probit link
function used can be described with the following:
~' I2dt equation (7-4)
where x denotes the time averaged SC>2 concentration (either 1-hour daily maximum or
the 24-hour average in ppb), y denotes the corresponding probability of a 5-minute exceedance,
and ft and y are two model estimated parameters used to generate predicted values. The logistic-
modeled predictions were then used to generate probability curves using all available
measurements, thereby extending the range of predictability beyond that of the empirically-based
curves. Figure 7-18 illustrates an example of logistic-modeled probability curves using the same
July 2009
118
-------
data used in generating the probability curves shown in Figure 7-17. Note that predictions for
the modeled curves extend beyond the 1-hour daily maximum limits of 200 ppb when using the
empirical curves.
Prior to estimating either the empirically-based or logistic-modeled probability curves,
staff separated the monitors within each data analysis group by the population density groups;
either low (< 10,000 persons within 5 km), mid (10,001 to < 50,000 persons within 5 km), or
high (> 50,000 persons within 5 km). Staff hypothesized that there may be different exceedance
probabilities in dense population areas compared with locations having fewer residents given the
siting characteristics of the monitors with regard to the presence of emission sources. This
separation of the monitoring results by the surrounding population should be useful in
appropriately characterizing the air quality because the monitoring data are used as indicators of
potential human exposure; the results from monitors sited within greater population densities
should be more representative of potential population exposure.
> 1OO ppb
> 200 ppb
> 300 ppb
> 4OO ppb
_3 O 50 100 150 2OO 25O 3OO 350 400
15
°- 1 —Hour Daily Maximum SO2 Concentration (ppb)
Figure 7-18. Example of logistic-modeled probability curves. The data used to generate these
modeled curves were the same used in generating the empirically-based curves in
Figure 7-17.
July 2009
119
-------
7.3 RESULTS
7.3.1 Measured 5-minute Maximum and Measured 1-Hour SOi Concentrations at
Ambient Monitors -As Is Air Quality
In this first data analysis group, staff analyzed the as is air quality data solely based on
the SC>2 ambient monitor measurements. Ambient monitoring data were evaluated at the 98
locations where both the 1-hour and 5-minute maximum SC>2 concentrations were reported for
years 1997 through 2007. Due to the large size of the data set (i.e., 471 site-years), staff
summarized the number of potential health effect benchmark exceedances in a series of figures.
This analysis centered on the relationship between various concentration averaging times and the
daily 5-minute maximum SC>2 concentration exceedances. Descriptive statistics for the measured
daily 5-minute maximum and the 1-hour 862 concentrations are provided in Appendix A-5 and
in the SOX ISA (ISA, section 2.5.2), the latter of which includes additional discussion of the
spatial and temporal variability of the 5-minute maximum and continuous 5-minute 862
concentrations. Staff performed two broad analyses using this data analysis group; first staff
evaluated the relationship between annual average concentrations and number of days per year
with at least one 5-minute concentration above benchmark levels and then estimated the
probability of having at least one 5-minute concentration above benchmark levels given short-
term averaging times (i.e., 1-hour daily maximum and 24-hour average).
First, staff evaluated the occurrence of the daily 5-minute maximum 862 concentration
exceedances in a year. Figure 7-19 compares the number of days with 5-minute maximum SC>2
concentrations above the potential health effect benchmark levels along with the corresponding
annual average SO2 concentration from each max-5 monitor. Overall, there are few days in a
year with 5-minute maximum SC>2 concentrations above each of the potential health effect
benchmark levels. Given the data in Table 7-8, no more than 7% of the total days with
measurements had 5-minute maximum SC>2 concentrations above the 100 ppb benchmark, while
approximately 2%, 1%, and 0.7% of days had daily 5-minute maximum SC>2 concentrations
above the 200, 300, and 400 ppb levels, respectively. None of the monitors in this data set had
annual average 862 concentrations above the current annual NAAQS of 30 ppb. However,
several of the monitors in several years frequently had daily 5-minute maximum SC>2
concentrations above the potential health effect benchmark levels. Many of those monitors
where frequent 5-minute benchmark exceedances occurred had annual average SO2
July 2009 120
-------
concentrations between 5 and 15 ppb, with little to no correlation between the annual average
SC>2 concentration and the number of daily 5-minute maximum 862 concentrations above the
potential health effect benchmark levels. These data are useful in determining the number of
days in a year a particular monitor had a daily maximum exceedance of a selected benchmark
level, however from a practical perspective, the annual average concentration would be
ineffective at controlling daily 5-minute maximum SC>2 concentrations given the observed weak
relationships.
b 250
n
£ 225
1 **
O 175
0 150
5 125
1 100
® 75
»
« 50
>
0 25
*« o-
0 250
| 225
.1 200
| 175
| 150
•| 125
«0 100
§ 75
•5 so
fc 25
i o-
> 100 ppb
0
°£'°° o 0 00
0 J.J.J 00?
« ft 0 ° Q, 0
°°° °°0= " 0°
a Sg*0 g*1,, „ o
^&^f?.
> 300 ppb
WO °
B° ° °
° O ° O
-
> 200 ppb
0 ° 0
o
a co Hf ^
CQ O O OQ
° ° 0 °
> 400 ppb
°°f °
0° 0
Od?
-------
Second, the probability of potential health effect benchmark exceedances was estimated
given the 24-hour average and 1-hour daily maximum 862 concentrations. Figure 7-20 presents
the empirically-based and logistic-modeled probability curves given the 24-hour average 862
concentrations and separated by the three population densities. There is an increasing probability
of a daily 5-minute maximum SO2 concentration exceedance with increasing 24-hour average
concentrations at each of the potential health effect benchmark levels and for each of the
population density groups. Some deviation from increasing probability occurs near the end of
the empirically-based curves derived from the mid-population density monitors. As discussed
earlier, this observed behavior is likely a function of the small sample size rather than variability
in 24-hour SC>2 concentrations. The logistic-modeled curves are consistent with the empirically-
based curves; however, the modeled curves illustrate an extended concentration range and a
consistent pattern of increasing probability of 5-minute benchmark exceedances with increasing
24-hour concentration.
Probability curves generated from monitors sited in low-population density areas exhibit
a steeper slope when compared with the other population density groups, indicating a greater
probability of a 5-minute SC>2 benchmark exceedance given the same 24-hour SC>2 concentration.
For example, the probability of exceeding a daily 5-minute maximum concentration of 200 ppb
using the empirically-based curves is 30% at the low-population density monitors given a 24-
hour average concentration of about 20 ppb. In comparison, empirically-based curves generated
from the mid- and high-population density monitors indicate that the probability of a 5-minute
benchmark exceedance at the same 24-hour concentration of 20 ppb is only about 14% and 3%,
respectively. There is a small probability (about 10%) of exceeding the 300 and 400 ppb in the
high-population density areas given a 24-hour average concentration of about 40 ppb (using
either the empirical or modeled curves), though at monitors sited in the low-population areas this
probability is greater than 50%.
The empirically-based curves are limited to estimating exceedance probabilities at or
below 24-hour concentrations of 60 ppb, with mostly unknown probabilities associated with
many of the benchmark levels and at concentrations approaching the current 24-hour standard.
For example, while the estimated probability of a daily 5-minute maximum 862 concentration
above 100 ppb is at or near 100% considering any of the population density groups, little can be
July 2009 122
-------
construed from the other empirically-based curves at 24-hour concentrations above 60 ppb,
particularly at monitors sited in mid- to high-population density areas. The logistic-modeled
curves however provide the probability of benchmark exceedances at higher 24-hour
concentrations. For example, according to Figure 7-20 there would be a 100% probability of
exceeding all benchmark levels at about a 24-hour concentration of 100-120 ppb, when
considering monitors in either the mid- or high-population density areas.
July 2009 123
-------
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Mid
Population
Density
— > 100 ppb
— > 200 ppb
— > 300 ppb
— > 400 ppb
High
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Mid
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
4 Lii High
/;./' ; Population
/ / / \ Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
24-hour Average SO2 Concentration (ppb)
Figure 7-20. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 24-hour average SO2 concentration, using empirical data
(left) and a fitted log-probit model (right), 1997-2007 air quality as is. Both the 5-minute
maximum and 24-hour SO2 concentrations were from measurements collected at 98
ambient monitors and separated by population density.
July 2009
124
-------
Figure 7-21 presents similar probability curves generated from the 5-minute and 1-hour
ambient measurement data, but the probabilities of benchmark exceedances are associated with
the 1-hour daily maximum 862 concentrations instead of 24-hour average concentrations. At
each of the benchmark levels and population densities, Figure 7-21 shows increasing
probabilities of exceedances with increasing 1-hour daily maximum SO2 concentrations.
Further, the probability curves have steeper slopes associated with the low-population density
group compared to the slopes of the higher population density groups. Note that while there is
uncertainty regarding the extrapolation beyond the limits imposed on the empirically-based
curves (i.e., 30 or greater samples per bin), one can be assured that the probability of an
exceedance of a daily 5-minute maximum SC>2 concentration of 400 ppb is 100% given a 1-hour
daily maximum SC>2 concentration of 400 ppb (and so on for the other 5-minute benchmark/1-
hour daily maximum SC>2 concentration combinations).45 As observed using the 24-hour average
concentrations, the shape of the curves beyond the imposed limits of the empirical data can be
informed by the logistic regression modeling (right column, Figure 7-21). In using the logistic-
modeled benchmark curves, a 100% probability of an exceedance is estimated to occur at about a
1-hour daily maximum concentration 50-100 ppb less that of the respective 5-minute benchmark
level.
It also should be noted that when comparing any of the 24-hour average probability
curves with corresponding 1-hour daily maximum probability curves (e.g., Figure 7-20 and
Figure 7-21) the relative slopes of the 24-hour curves are steeper. Therefore, changes in 24-hour
average 862 concentration (either higher or lower) will effectively result in greater changes in
the probability of exceedances when compared to a similar 1-hour daily maximum concentration
shift. For example, to reduce the likelihood of a 200 ppb benchmark exceedance from about
90% to 10%, 24-hour average concentrations would need to go from a level of about 50 to 20
ppb using the logistic-modeled mid-population curves. This same reduction in probability would
correspond to a 1-hour daily maximum concentration reduction of about 150 ppb to 70 ppb.
45 Technically, if all 5-minute concentrations were exactly 400 ppb, the 1-hour average concentration would be 400
ppb and the 5-minute maximum would not actually exceed 400 ppb. However, note that probability of exceeding
the 100 or 200 ppb benchmarks approaches 100% at less than a 1-hour daily maximum of 100 an 200 ppb,
respectively (Figure 7-18).
July 2009 125
-------
(D
O
c
03
•o
(D
(D
O
03
03
O
C
(D
DO
CM
O
(0
x
03
10
_>*
'<5
Q
03
M—
O
.0
s
2
Q_
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Mid
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
High
Population
Density
— > 100 ppb
— > 200 ppb
— > 300 ppb
> 400 ppb
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Population
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
501001502002503003504000 50100150200250300350400
1-Hour Daily Maximum SO2 Concentration (ppb)
Figure 7-21. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SO2 concentration, using
empirical data (left) and a fitted log-probit model (right), 1997-2007 air quality as is.
Both the 5-minute maximum and 1-hour SO2 concentrations were from measurements
collected at 98 ambient monitors and separated by population density.
July 2009
126
-------
7.3.2 Measured 1-Hour and Modeled 5-minute Maximum SOi Concentrations at All
Ambient Monitors -As Is Air Quality
In the second data analysis group, staff analyzed the as is air quality using a combination
of measurement and modeled data. As described in section 7.2.3, a statistical model was applied
to 1-hour ambient 862 measurements to estimate 5-minute maximum 862 concentrations. This
was done because there are a greater number of monitors in the broader 862 monitoring network
compared to subset of monitors reporting the 5-minute maximum SO2 concentrations (section
7.3.1). This larger monitoring data set included 809 ambient monitors in operation at some time
during the years 1997 through 2006 that met the completeness criteria described in section 7.2.1.
This data set included 4,692 site-years of data, and combined with the estimated 5-minute SC>2
concentrations using the measured 1-hour values, allowed for a comprehensive characterization
of the hourly and 5-minute 862 air quality at ambient monitors located across the U.S.
Descriptive statistics for the measured 1-hour 862 concentrations are provided in the SOX ISA
(ISA, section 2.5.1) including additional discussion of the spatial and temporal variability in 1-
hour SO2 concentrations.
Staff performed twenty separate model simulations to estimate the 5-minute maximum
SC>2 concentration associated with each 1-hour measurement. The individual simulation results
at each monitor were averaged to generate a mean number of days per year with a 5-minute
benchmark exceedance. The modeled (5-minute maximum) and measurement (1-hour) data
were analyzed in a similar manner as performed on the measured 5-minute maximum and 1-hour
862 concentrations described in section 7.3.1. The results provided in this section were
generated using the modeled daily 5-minute maximums and the measured hourly 862
concentrations considering 1-hour, 24-hour, and annual averaging times. Staff performed two
broad analyses; first staff evaluated the relationship between annual average concentrations and
number of days per year with at least one 5-minute concentration above benchmark levels and
then estimated the probability of having at least one 5-minute concentration above benchmark
levels given short-term averaging times (1-hour daily maximum and 24-hour average).
First, Figure 7-22 shows the number of days per year with a 5-minute SC>2 concentration
above benchmark levels versus the annual average 862 concentration. Fewer than 5% of total
days per year had a 5-minute 862 concentration above the 100 ppb benchmark, while
approximately 1%, 0.5%, and 0.2% of days had at least one 5-minute concentration above the
July 2009 127
-------
200, 300, and 400 ppb benchmark levels, respectively. None of the site-years of data had annual
average 862 concentrations at or above the level of the current annual NAAQS (30 ppb).
However as described above, several site-years had predicted 5-minute SO2 concentrations above
the potential health effect benchmark levels. Many of the monitors with frequent 5-minute
benchmark exceedances had annual average SC>2 concentrations between 10 and 20 ppb, with a
pattern of increasing number of days per year with at least one 5-minute concentration above the
benchmark levels with increasing annual average concentrations. This pattern was most
prominent at the 100 ppb benchmark level, with progressively weaker relationships between the
number of 5-minute benchmark exceedances and annual average concentrations at each of the
higher benchmark levels.
July 2009 128
-------
5 10 15 20 25 30
Annual Arithmetic Mean SO2 (ppb)
35 0
5 10 15 20 25 30
Annual Arithmetic Mean SO2 (ppb)
Figure 7-22. The number of days per year with modeled daily 5-minute maximum SO2
concentrations above potential health effect benchmark levels at 809 ambient monitors
given the annual average SO2 concentration, 1997-2006 air quality as is. The level of
the annual average SO2 NAAQS of 30 ppb is indicated by the dashed line.
Next, empirical and logistic-modeled probability curves were generated for this second
data analysis group. Figure 7-23 illustrates the probability of benchmark exceedances using the
modeled daily 5-minute maximum SC>2 concentrations and 24-hour average concentrations.
These probability curves exhibit patterns similar to that described using the pure measurement
data (Figure 7-20). For example, the probability curves generated from low-population density
area monitors are steeper than those generated using the higher population density monitors at
July 2009
129
-------
each of the benchmark levels considered. In addition, the slopes of the probability curves are
generally consistent between the measured and modeled 5-minute maximum data, where
comparable 24-hour average concentrations exist.
The broader SC>2 monitoring network to estimate daily 5-minute maximum SC>2
concentrations provides insight as to the potential shape of each empirically-based probability
curve at greater 24-hour average concentrations. The upper range of 24-hour concentrations
extends to around 70-100 ppb (Figure 7-23), while at the monitors reporting 5-minute maximum
SC>2 concentrations the maximum 24-hour average concentrations extends to at most between 50
and 60 ppb (Figure 7-20). The extended range of 24-hour concentrations in the empirically-
based curves provides additional support to what was stated earlier using the pure measurement
data, that is, there is a strong likelihood of 5-minute peak concentrations above the benchmark
levels at 24-hour average concentrations well below the level of the current standard. This is
further confirmed by the logistic-modeled probability curves that estimate all benchmark levels
would be exceeded at about a 24-hour concentration of 60-100 ppb, the level of which dependent
on where the monitor is sited.
The probability curves generated using the modeled 5-minute maximum and 1-hour daily
maximum 862 concentrations (Figure 7-24) also exhibit patterns consistent with those patterns
observed using the pure measurement data (Figure 7-21). Again, a wider range of 1-hour daily
maximum concentrations is observed in using the broader monitoring network when compared
with the results using the monitors reporting the 5-minute maximum SC>2 concentrations, giving
greater ability to discern the probability of benchmark exceedances at higher 1-hour daily
maximum SC>2 concentrations. When using either the empirically-based or logistic modeled
curves, a 100% probability of exceeding the 100, 200, 300, and 400 ppb benchmarks is estimated
to occur at 1-hour daily maximum concentrations of about 80, 150, 225, and 300 ppb,
respectively.
July 2009 130
-------
Low
Population
Density
> 100 ppb
— > 200 ppb
— > 300 ppb
> 400 ppb
Mid
Population
Density
> 100 ppb
— > 200 ppb
— > 300 ppb
> 400 ppb
Low
Population
Density
— > 100 ppb
— > 200 ppb
— > 300 ppb
—- > 400 ppb
High
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
../..
/; 7 'l i ; ;
/ ;/ ,';; Mid
/ ;/ i { ; Population
/ ?..../...;I. Density
/ 1-: i ; ;- > 100 ppb
— > 200 ppb
— > 300 ppb
— > 400 ppb
/ ' ;'
/..../...:/..
High
Population
Density
— > 100 ppb
— > 200 ppb
— > 300 ppb
> 400 ppb
0 20 40 60 80 100 120 1400 20 40 60 80 100 120 140
24-hour Average SO2 Concentration (ppb)
Figure 7-23. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 24-hour average SO2 concentrations, using empirical
data (left) and a fitted log-probit model (right), 1997-2006 air quality as is. The 5-minute
maximum SO2 concentrations were modeled from 1-hour measurements collected at
809 ambient monitors and then separated by population density.
July 2009
131
-------
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Mid
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
High
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
i ' i / =/
....;../M»'-
!./ I/ !..
Low
Population
Density
> 100 ppb
> 200 ppb
> 300 ppb
> 400 ppb
Mid
Population
Density
> 100 ppb
— > 200 ppb
— > 300 ppb
> 400 ppb
High
Population
Density
— > 100 ppb
— > 200 ppb
— > 300 ppb
---- > 400 ppb
Q_ 0 50100150200250300350 4000 50 100 150 200 250 300 350 400
1-Hour Daily Maximum SO2 Concentration (ppb)
Figure 7-24. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SO2 concentrations, using
empirical data (left) and a fitted log-probit model (right), 1997-2006 air quality as is. The
5-minute maximum SO2 concentrations were modeled from 1-hour measurements
collected at 809 ambient monitors and then separated by population density.
July 2009
132
-------
7.3.3 Modeled 1-Hour and Modeled 5-minute Maximum SOi Concentrations at Ambient
Monitors in 40 Counties - Air Quality Adjusted to Just Meet the Current and Potential
Alternative Standards
Staff selected forty counties to analyze 5-minute benchmark exceedances under several
air quality scenarios: as is air quality and air quality adjusted to just meeting the current and
alternative standards. The forty counties were selected using criteria discussed in section 7.2.4.
Specifically, we chose the 38 counties with 1-hour ambient monitor 862 concentrations nearest
the current NAAQS levels and two counties with a high frequency of measured daily 5-minute
maximum SC>2 concentrations above the potential health effect benchmark levels. The 1-hour
SO2 measurement data were from 128 ambient monitors and totaled 610 site-years of
monitoring, a subset of data from the broader SO2 monitoring network (see section 7.3.2). Staff
evaluated multiple alternative air quality scenarios by first adjusting the 1-hour ambient
monitoring concentrations to just meet a particular standard level (section 7.4). Then, as was
done in section 7.3.2, staff performed twenty simulations to estimate the 5-minute maximum SC>2
concentration associated with each 1-hour adjusted concentration using the statistical model
described in section 7.2.3. These simulation results were combined to generate a mean estimate
for each of the metrics of interest (e.g., the number of days in a year with 5-minute maximum
SO2 concentrations > 200 ppb) selected here as the best estimate from the twenty simulations.
Staff 1) evaluated the relationship between annual average concentrations and number of days
per year with at least one 5-minute concentration above benchmark levels, 2) summarized the
number of days per year with at least one 5-minute concentration above benchmark levels for
each air quality scenario, 3) compared number of days per year with at least one 5-minute
concentration above benchmark levels using two percentile forms of the potential alternative 1-
hour daily maximum standards (i.e., 98th and 99th percentile), and 4) estimated the probability of
having at least one 5-minute concentration above benchmark levels given short-term averaging
times (1-hour daily maximum and 24-hour average).
First, staff evaluated the relationship between the short-term peak concentrations and the
level of the current annual SC>2 NAAQS in the selected counties. Figure 7-25 illustrates the
number of days per year with 5-minute daily maximum SO2 concentrations above the potential
health effect benchmark levels along with the corresponding annual average concentrations.
Each data point represents a monitor site-year generated from the modeled 5-minute peaks and
July 2009 133
-------
air quality adjusted to just meeting the current 862 standards. None of the site-years in the
selected counties had annual average concentrations above the level of the current NAAQS (30
ppb) by design46, however there are many more site-years with a greater number of modeled
daily 5-minute maximum SC>2 concentrations above the potential health effect benchmark levels
than compared with that of the as is air quality. There are a decreasing number of exceedances
with increasing benchmark concentrations, though there is a greater proportion of monitors with
exceedances when considering concentrations adjusted to just meeting the current standard than
when using the as is air quality (e.g., see Figure 7-19). When considering concentrations
adjusted to just meeting the current standard, there is a stronger relationship between the annual
average concentrations and the number of benchmark exceedances than observed previously
with the as is air quality however, the strength of that relationship weakens with increasing
benchmark levels.
46 The current annual SO2 NAAQS is 30 ppb. Concentrations of up to 30.4 ppb are possible due to a rounding
convention. This is why there are several data points just to the right of the dashed line in Figure 7-22.
July 2009 134
-------
(0
a)
•u
(75
>_
a)
a.
"55
a)
•a
a)
ts
_a>
a)
(/)
SI
o
CM
O
E
1
re
§
Q)
"3
c
I
IO
Q
M-
O
1_
o>
^
E
350
325
300
275
250
225
200
175
150
125
100
75
50
25
0
350
325
300
275
250
225
200
175
150
125
100
75
50
25
0
> 100 ppb
> 300 ppb
5 10 15 20 25 30
Annual Arithmetic Mean SO2 (ppb)
> 400 ppb
35 0 5 10 15 20 25 30
Annual Arithmetic Mean SO2 (ppb)
Figure 7-25. The number of days per year with modeled 5-minute maximum SO2 concentrations
above potential health effect benchmark levels per year at 128 ambient monitors in 40
selected counties given the annual average SO2 concentration, 2001-2006 air quality
adjusted to just meet the current NAAQS. The level of the annual average SO2 NAAQS
of 30 ppb is indicated by the dashed line.
July 2009
135
-------
Similar relationships are present between the annual average 862 concentrations and the
number of benchmark exceedances when considering the potential alternative standards. As a
reminder, to just meet the current and potential alternative standards staff estimated a unique
adjustment factor to simulate the alternative air quality. The direction of the adjustment factor
(either upwards or >1; downwards or <1) and magnitude of the adjustment factor used has a
direct impact on the estimated number of 5-minute benchmark exceedances. In general, the air
quality distributions that just meet the potential alternative standards were enveloped by the as is
air quality (i.e., a distribution with low concentrations) and the air quality adjusted to just
meeting the current standard (i.e., a distribution with generally high concentrations). Therefore,
the estimated number of days with exceedances also fell within the range of exceedances
generated using the as is air quality or the air quality adjusted to just meet the current standard.
For example, a comparison of the annual average SC>2 concentrations and number of daily 5-
minute maximum exceedances of 200 ppb is presented in Figure 7-26 for six air quality
scenarios: four of the 99th percentile 1-hour daily maximum potential alternative standards (i.e.,
the 100, 150, 200, and 250 ppb); the air quality adjusted to just meet the current standards; and
as is air quality.
Clearly, in using the air quality adjustment procedure combined with the statistical model
to estimate 5-minute maximum 862 concentrations, the current standard air quality scenario
allows for the greatest estimated number of days per year with potential health effect benchmark
exceedances (Figure 7-26). However, at a minimum the annual standard does provide protection
against annual average concentrations above the level of the current standard. While there were
fewer 5-minute benchmark exceedances using the 1-hour daily maximum forms of a potential
alternative standard, two of the levels (1-hour daily maximums of 200 and 250 ppb) did not
prevent annual average concentrations from exceeding the current annual standard (Figure 7-26).
High annual average concentrations become less of an issue when considering the lower levels of
the 1-hour daily maximum potential alternative standards. Even though the 99th percentile 1-
hour daily maximum standards of 100 or 150 ppb allow for greater annual average
concentrations than when considering as is air quality, all but one site-year are below the level of
the current annual standard and there are fewer estimated days per year with benchmark
exceedances. These results further demonstrate the stronger relationship 5-minute peak
concentrations have with 1-hour SC>2 concentrations than with annual average concentrations.
July 2009 136
-------
»_ 350
(0
0
0)
0
Q.
Q.
Q.
O
O
01
O
J2
E
3
E
'S
ra
o
Q
"o
0)
E
300
250
100
501
0
350
300
250
200
100
50
th
99 percentile
Daily 1-hour
maximum of 100 ppb
th
99 percentile
Daily 1-hour
maximum of 200 ppb
99th percentile
Daily 1-hour
maximum of 150 ppb
99th percentile
Daily 1-hour
maximum of 250 ppb
As is air quality
Current annual or
daily standard
5 101520253035404550550
AnnuaJ Arithmetic Mean SO2 (ppb)
5 10152025303540455055
Annual Arithmetic Mean SO2 (ppb)
0 5101520253035404550550
Annual Arithmetic Mean SO2 (ppb)
Figure 7-26. The number of modeled daily 5-minute maximum SO2 concentrations above 200 ppb per year at 128 ambient monitors in 40
selected counties given the annual average SO2 concentration, 2001-2006 air quality as is and that adjusted to just the
current and four potential alternative standards (text in graph indicate standard evaluated). The level of the annual average
SO2 NAAQS of 30 ppb is indicated by the dashed line.
July 2009
137
-------
Table 7-10. Percent of days having a modeled daily 5-minute maximum SO2 concentration above
the potential health effect benchmark levels given air quality as is and air quality adjusted to just
meeting the current and each of the potential alternative standards.
Air Quality
Scenario1
as is
CS
99-50
99-100
98-100
99-150
99-200
99-250
98-200
Percent of Days With Daily 5-minute Maximum SO2
Concentrations Above Benchmark Levels
> 100 ppb
9.1
41.0
0.7
4.5
6.9
10.6
17.2
23.6
22.5
> 200 ppb
2.4
17.2
0.0
0.7
1.2
2.2
4.5
7.4
6.9
> 300 ppb
0.9
9.1
0.0
0.2
0.3
0.7
1.6
2.9
2.6
> 400 ppb
0.5
5.3
0.0
0.0
0.1
0.3
0.7
1.3
1.2
Notes:
1 as ;'s air quality is unadjusted; CS is air quality adjusted to just meet the
current standard; x-y are the xth percentile form of a 1 -hour daily maximum
level of y.
Second, staff summarized the number of days per year with 5-minute maximum SC>2
concentrations above benchmark levels within the 40-county data set for additional comparisons
of the air quality scenarios. Table 7-10 provides the percent of all days above each of the
benchmark levels considering each of the air quality scenarios. Again, the scenario where air
quality just meets the current standard has the greatest percent of days with benchmark
exceedances. With each progressive decrease in the 1-hour daily maximum 862 concentration
levels of the potential alternative standards, there are fewer days with benchmark exceedances.
The percent of all days with benchmark exceedances using as is air quality was between a
potential 1-hour daily maximum alternative standard level of 100 and 150 ppb (99th percentile
form), or similar to that of the 98th percentile form at a level of 100 ppb.
Third, staff evaluated two forms of the potential alternative standards: the 99th and 98th
percentile forms, each having a 1-hour daily maximum level of either 100 or 200 ppb. For
example, Figure 7-27 indicates that nearly all site-years have a greater estimated number of days
per year with benchmark exceedances given the 98th percentile form when compared with a 99th
percentile form at the same level. This is expected given the number of allowable 1-hour 862
concentrations above the 200 ppb level for each of the percentile forms. The two air quality
scenarios were compared on a monitor-to-monitor basis and on average, the 98th percentile form
allowed for approximately 46, 68, 84, and 86% more benchmark exceedances considering the
July 2009
138
-------
100, 200, 300, and 400 ppb benchmark levels, respectively when compared with the 99th
percentile form.
If
I 5
15 o
<1> -T
>- T-
50 100 150 200 250 300 350
25 50 75 100 125 150 175 200
<1>
Q.
(0
Q
a>
.Q
a)
S
a)
a.
00
0)
10 20 30 40 50 60 70 80 90 100 0 5 10 15 20 25 30 35 40 45 50 55 60
Number of Days per Year with a 5-Minute Benchmark Exceedance
- 99th Percentile 1-hour Daily Maximum 200 ppb Standard
Figure 7-27. The number of days per year with modeled 5-minute maximum SO2 concentrations
above benchmark levels given the 99th and 98th percentile forms, using the 40-county
air quality data set adjusted to just meet a 1-hour daily maximum level of 200 ppb.
-,th
When a 1-hour daily maximum level of 100 ppb was considered, on average the 98
percentile form of the potential alternative standard allowed for approximately 68, 90, 84, and
July 2009
139
-------
74% more benchmark exceedances at each monitor considering the 100, 200, 300, and 400 ppb
benchmark levels, respectively when compared with the 99th percentile form. While generally
there were greater differences in the percent of exceedances for the two forms when considering
the 100 ppb level compared with the 200 ppb level, there were far fewer site-years with
benchmark exceedances (Figure 7-28).
180
0 5 10 15 20 25 30 35 0 5 10 15 20
Number of Days per Year with a 5-Minute Benchmark Exceedance
- 99th Percentile 1-hour Daily Maximum 100 ppb Standard
Figure 7-28. The number of days per year with modeled 5-minute maximum SO2 concentrations
above benchmark levels given the 99th and 98th percentile forms, using the 40-county
air quality data set adjusted to just meet a 1-hour daily maximum level of 100 ppb.
July 2009
140
-------
Fourth, staff estimated the probability of potential health effect benchmark exceedances
given the adjusted air quality scenarios and short-term averaging times. Again, patterns in the
curves were consistent with what was observed and described previously; monitors within low-
population density areas had steeper probability curves compared with those in higher population
density areas. Further, there were similarities in the shape and the steepness of the curves when
comparing the adjusted air quality probability curves with the curves developed from the
corresponding as is air quality. Therefore, for the sake of brevity, all of the probability curves
for each of the alternative standards are not presented. However, there were some differences in
the probability curves worthy of presentation and discussion, using the empirically-based curves
for the demonstration.
Figure 7-29 presents the probability of a 5-minute benchmark exceedance using as is air
quality and air quality adjusted to just meet the current standard, given 1-hour daily maximum
SC>2 concentrations. In general, all of the corresponding probability curves for all of the air
quality scenarios overlap when considering the 100 and 200 ppb benchmark levels. However,
the probability curves associated with exceeding the 300 and 400 ppb benchmark levels were of
similar slope, but shifted to the left when considering the as is air quality compared with the
current standard scenario. This is likely a function of the non-linear form of the statistical model
used to estimate the 5-minute maximum 862 concentrations, the proportional adjustment
procedure to simulate alternative standards, and the form of the air quality characterization
metric used.
When adjusting the 1-hour SC>2 concentrations upwards using a proportional factor, a
corresponding proportional increase in the number of days per year with benchmark exceedances
does not necessarily follow. The statistical model uses multiple distributions of PMRs, not
linearly related to 1-hour SC>2 concentrations. Certainly, the total number of days in a year with
benchmark exceedances will increase with an upward adjustment of air quality, and does so as
observed in Figure 7-26. However, the greatest proportion of monitoring days within any of the
air quality scenarios is comprised of days without an exceedance (see Table 7-10). The
frequency of exceedances of the higher benchmarks is very low using the as is air quality; the
few added days with estimated exceedances of 300 or 400 ppb using the simulated air quality is
not proportional to the universal increase in hourly concentrations applied to all 1-hour
concentrations. Therefore the probability curves tend to be less steep with the upward 1-hour
July 2009 141
-------
concentration adjustments when considering the higher benchmark levels. Furthermore, days
already having an exceedance are only counted once, that is, if there were an exceedance on a
given day using the as is air quality, it is likely that the same day would also have an exceedance
using the adjusted air quality, only it is associated with a greater 1-hour (or 24-hour average)
concentration. Again, the 1-hour concentrations are increased without corresponding
proportional increase in the number of exceedances when comparing the two air quality
scenarios. Conversely, it could also be argued that there may be an increased probability of daily
5-minute exceedances of 300 and 400 ppb when using air quality with a relatively low
concentration distribution (such as with the as is air quality) compared with a distribution of
higher concentrations (such as with the current standard scenario). However, it should be noted
that the total number of benchmark level exceedances in a year (and the absence of exceedances
at the same high 1-hour daily maximum concentration) under either of these scenarios would be
very few, with far fewer numbers of exceedances associated with the relatively low
concentration air quality.
This discussion of probability curves can be extended to each of the potential alternative
standards. For example, Figure 7-30 illustrates a range in each of the probability curves given
each of the alternative air quality scenarios and using monitors sited within high-population
density areas. The 100 and 200 ppb benchmark level probability curves exhibit a narrow range
across each of the adjusted air quality scenarios. While the estimated 300 and 400 ppb
probability curves are wider than the 100 and 200 ppb curves, there is still agreement in the
estimated probabilities at many of the 1-hour daily maximum SC>2 values. The range in
probability curves tended to be widest at the lowest probabilities/1-hour daily maximum SC>2
concentrations within a given benchmark, likely indicating a greater uncertainty in the
relationship between exceedance of the daily 5-minute maximum SC>2 concentrations of 300 and
400 ppb and 1-hour daily maximum 862 concentrations less than 130 ppb and 180 ppb,
respectively.
July 2009 142
-------
100 -|
90
80
70
2
o.
g 60
I 50
40
30
20
10
X
n
0
100 i
90
80
70
a> en
o bU
n S
2 w
>
!5
1
Q.
50 -:
40
30
20
10
Si
o
2
a.
0
100
90
80
70
60
50
40
30
20
10
0
Low
Population
Density
100ppb CS
200 ppb CS
300 ppb CS
400 ppb CS
100 ppb AS IS
200 ppb AS IS
300 ppb AS IS
400 ppb AS IS
Mid
Population
Density
100 ppb CS
200 ppb CS
> 300 ppb CS
- - - > 400 ppb CS
—•— > 100 ppb AS IS
—e— > 200 ppb AS IS
300 ppb AS IS
—a~ > 400 ppb AS IS
High
Population
Density
100 ppb CS
200 ppb CS
> 300 ppb CS
- - - > 400 ppb CS
—•— > 100 ppb AS IS
—e—> 200 ppb AS IS
--e-->300 ppb AS IS
---o-- >400 ppb AS IS
100 150 200 250 300 350
Daily Maximum 1-hour SO2 Concentration (ppb)
400
Figure 7-29. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SO2 concentrations, 2001-2006 air
quality as is and that adjusted to just meet the current NAAQS. The 5-minute maximum
concentrations were modeled from 1-hour measurements collected at 128 ambient
monitors from 40 selected counties and then separated by population density within 5
km of monitors.
July 2009
143
-------
E
3
E
'x
ra
100
90
80
70 --
.8
Igeo
tl 50
| | 40
"5
>, 30
20
A
3
£ 1°
^&W"
•> 100 ppb
- > 200 ppb
— > 300 ppb
• - > 400 ppb
0 50 100 150 200 250 300 350 400
Daily Maximum 1-hour SO2 Concentration (ppb)
Figure 7-30. Probability of daily 5-minute maximum SO2 concentrations above potential health
effect benchmark levels given 1-hour daily maximum SO2 concentrations, 2001-2006 air
quality adjusted to just meet the current and each of the potential alternative standards
(99th percentile form). The 5-minute maximum concentrations were modeled from 1-
hour measurements collected at 128 ambient monitors from 40 selected counties, high-
population density monitors.
While there are similarities in the probability of daily 5-minute maximum benchmark
exceedances for each of the potential alternative standard scenarios given either the 1-hour daily
maximum or 24-hour average SO2 concentrations, there are large differences in the total number
of exceedances given a particular county and air quality scenario. Table 7-11 presents the mean
number of days in a year where the daily 5-minute maximum SO2 concentration was above 100
ppb in each of the 40 selected counties and for all air quality scenarios. In considering air quality
adjusted to just meeting the current standard and the level of the highest potential alternative
standards (200 and 250 ppb 1-hour daily maximum), counties such as Hudson NJ, Tulsa OK, and
Wayne WV were estimated to have the greatest number of benchmark exceedances. On average
there would be between 100 and 200 days of the year with 5-minute maximum SO2
concentrations above 100 ppb in these counties. Most of the other locations though had fewer
than 100 benchmark exceedances in a year, particularly when considering the two potential
alternative 1-hour daily maximum standards. Air quality simulating just meeting the current
standard was associated with the greatest number of estimated exceedances at most locations.
This consistent pattern was observed with each of the benchmark levels (see below) indicating
the limited influence the current standard has on the estimated number of 5-minute benchmark
July 2009
144
-------
exceedances. Decreases in the potential alternative standard level corresponded with decreases
in the number of days per year with benchmark exceedances. Most counties have fewer mean
estimated 5-minute benchmark exceedances of 100 ppb using air quality adjusted to just meeting
the 99th percentile daily 1-hour maximum concentration of 100 ppb, than that estimated using the
as is air quality. The were 11 counties that only achieve reduction in the number of benchmark
level exceedances from as is air quality when considering the 99th percentile daily 1-hour
maximum concentration of 50 ppb. This means that to improve current air quality in most
locations, a level below 100 ppb would need to be selected when using a 99th percentile 1-hour
daily maximum standard form.
In addition, the two percentile forms of the alternative standards (98th and 99th) were
evaluated each at two 1-hour daily maximum standard levels (100 and 200 ppb) (Table 7-11).
The estimated number of exceedances using a 98th percentile 1-hour daily maximum alternative
standard level of 100 ppb fell within those estimated using 99th percentile levels of 100 and 150
ppb. The estimated number of exceedances using a 98th percentile 1-hour daily maximum
alternative standard level of 200 ppb was similar to the 99th percentile using a 250 ppb 1-hour
concentration level. Both of these patterns were consistent when comparing the different
standard forms for each the 5-minute benchmarks (see Tables 7-12 through 7-14).
There were fewer estimated exceedances of 200 ppb given the potential alternative
standards than compared with the current standard scenario (Table 7-12). Most counties had
fewer than forty days per year with 5-minute SO2 concentrations above 200 ppb considering the
1-hour daily maximum standards, while the number of exceedances was approximately double
that when using air quality adjusted to just meet the current standard. With progressive
decreases in the 1-hour daily maximum standard level, the number of days per year with 5-
minute maximum SC>2 concentrations also decreases. In 75% of counties, the estimated number
of benchmark exceedances using as is air quality was above that estimated using 1-hour daily
maximum standard level of 100 ppb. The 99th percentile 1-hour daily maximum concentration
level of 50 ppb was associated with the fewest days with 5-minute maximum 862 concentrations
above 200 ppb. On average most locations had zero exceedances of the 200 ppb benchmark
level.
Similar results are presented for each the 300 ppb (Table 7-13) and the 400 ppb (Table 7-
14) 5-minute benchmark levels, though the difference in the number of exceedances between the
July 2009 145
-------
current standard and the other air quality scenarios is much greater than was observed for the
lower benchmark levels. Most counties had a 5-fold (or greater) number of days with daily 5-
minute maximum SC>2 concentrations above 300 or 400 ppb when considering air quality just
meeting the current standard compared with air quality adjusted to just meet the 99th percentile 1-
hour daily maximum level of 250 ppb. The number of exceedances given as is air quality was
still within the range of values estimated using the potential standard levels of 100 and 200 ppb;
in most counties it was fewer than 10 days per year. Most counties did not have any estimated
days per year with 5-minute maximum SC>2 concentrations above 400 ppb given a 99th percentile
1-hour daily maximum of 100 ppb, while 75% of the counties had 1 or fewer exceedances of 300
ppb considering this same potential alternative standard.
July 2009 146
-------
Table 7-11. Modeled mean number of days per year with 5-minute maximum concentrations
above 100 ppb in 40 selected counties given 2001-2006 air quality as is and air quality adjusted to
just meet the current and alternative standards.
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Iron
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
as is1
119
21
22
24
42
47
58
17
27
29
34
29
20
65
70
46
3
2
8
38
60
16
44
51
26
30
76
14
63
25
62
11
75
24
0
76
78
39
30
8
CS1
234
123
127
166
139
211
122
186
184
103
123
134
92
108
150
118
145
117
124
172
163
203
164
198
202
159
194
130
110
185
116
144
201
132
109
220
207
172
201
67
99th percentile1
50
9
1
3
1
6
8
8
3
2
8
9
2
8
9
6
7
1
1
2
6
13
2
3
3
4
1
2
2
3
2
3
3
2
3
1
3
2
3
4
1
100
36
8
12
11
17
24
23
20
12
25
26
18
24
30
22
31
20
16
28
18
34
23
20
23
43
8
11
25
17
21
19
13
20
19
17
25
21
15
33
4
150
63
19
23
25
30
43
37
41
27
42
41
40
37
40
37
52
62
51
71
33
52
55
41
51
93
22
30
56
33
53
42
26
49
40
54
62
52
26
83
11
200
89
34
37
42
43
62
50
64
44
56
54
62
47
48
50
68
111
98
115
50
68
93
61
81
133
41
55
87
48
88
63
39
74
58
98
101
86
38
138
20
250
111
50
50
60
54
81
63
91
63
68
68
80
59
55
61
81
161
141
155
70
83
122
80
110
162
65
83
114
62
125
83
53
94
75
143
140
118
50
180
30
98th percentile1
100
47
12
18
18
29
34
29
31
21
32
32
25
30
34
31
37
35
25
39
23
39
39
27
30
62
12
18
41
25
29
26
21
40
24
29
40
32
22
47
10
200
107
46
53
61
64
83
61
93
68
66
65
76
57
54
61
76
150
122
137
65
75
129
73
96
154
58
79
127
62
110
75
57
100
68
129
135
110
54
166
37
Notes:
1 These are the air quality scenarios evaluated: as ;'s is unadjusted air quality; CS is air quality adjusted to just
meet the current standard; the levels of the two percentile forms (99th and 98th) of a 1 -hour daily maximum
potential alternative standard are given.
July 2009
147
-------
Table 7-12. Mean number of modeled days per year with 5-minute maximum concentrations
above 200 ppb in 40 selected counties given 2001-2006 air quality as is and that adjusted to just
meet the current and alternative standards.
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Iron
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
as is1
55
4
6
5
17
17
28
2
6
10
14
5
6
44
38
14
0
0
0
15
29
1
11
11
2
5
17
2
25
3
19
2
21
5
0
16
17
15
3
2
CS1
171
38
50
66
75
117
70
80
90
53
57
61
47
77
99
68
31
22
32
88
86
85
71
96
112
52
88
40
52
66
54
35
121
71
21
96
96
63
71
24
99th percentile1
50
0
0
1
0
1
1
1
0
0
2
1
0
1
0
0
1
0
0
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
9
1
3
1
6
7
8
3
2
8
9
2
8
9
6
7
1
1
2
6
13
2
3
3
5
1
2
3
3
2
3
3
2
3
1
3
2
3
4
1
150
22
4
7
5
11
16
16
10
6
17
18
9
16
21
14
18
7
6
11
11
24
10
10
12
19
3
5
10
9
10
10
7
9
10
6
12
9
9
16
3
200
36
8
12
11
17
24
22
20
12
25
26
18
24
29
22
30
20
15
27
18
34
23
20
24
42
8
11
25
17
21
20
13
21
19
17
26
21
15
33
4
250
49
13
17
17
23
33
30
31
19
34
34
29
31
36
29
42
39
31
48
25
43
38
30
37
69
14
20
40
25
36
31
20
35
29
34
43
36
21
58
7
98th percentile1
100
15
2
5
3
11
12
11
6
4
12
12
4
12
13
11
10
3
2
3
8
15
5
4
4
9
1
3
5
6
4
4
5
6
5
2
6
4
6
6
2
200
47
12
18
18
29
34
29
31
21
33
32
25
30
34
31
37
34
24
38
24
38
38
26
31
62
12
18
41
25
29
26
21
39
25
28
40
32
22
48
10
Notes:
1 These are the air quality scenarios evaluated: as ;'s is unadjusted air quality; CS is air quality adjusted to just
meet the current standard; the levels of the two percentile forms (99th and 98th) of a 1 -hour daily maximum
potential alternative standard are given.
July 2009
148
-------
Table 7-13. Mean number of modeled days per year with 5-minute maximum concentrations
above 300 ppb in 40 selected counties given 2001-2006 air quality as is and that adjusted to just
meet the current and alternative standards.
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Iron
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
as is1
31
1
3
1
9
8
16
0
2
5
6
1
2
33
24
5
0
0
0
9
17
0
3
2
0
1
6
1
11
1
7
0
7
1
0
5
4
7
1
0
CS1
130
17
27
35
50
75
47
42
49
35
39
32
32
61
72
46
7
5
9
52
59
39
41
51
60
21
42
16
31
28
28
19
83
43
5
45
48
36
31
11
99th percentile1
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
2
0
1
0
2
3
3
1
1
4
4
0
3
1
1
3
0
0
0
2
5
0
0
1
1
0
0
1
1
0
0
1
0
1
0
1
0
1
1
0
150
9
1
3
1
6
8
7
3
2
8
9
2
7
9
6
7
1
1
2
6
13
2
2
3
4
1
2
3
3
2
3
3
2
3
1
4
2
3
4
1
200
18
3
6
4
9
13
13
8
5
14
15
6
13
17
11
14
4
3
7
10
20
7
7
8
12
2
4
7
7
7
7
6
6
7
4
8
6
6
10
2
250
27
5
8
7
13
18
18
13
8
19
20
12
19
24
17
22
10
8
16
13
27
13
13
15
26
4
7
15
11
13
13
9
12
13
9
16
12
11
21
3
98th percentile1
100
4
1
2
1
6
5
4
2
1
6
6
1
5
4
4
4
0
0
0
3
6
1
1
1
2
0
1
1
1
1
1
2
2
1
0
2
1
2
1
1
200
25
5
9
7
17
19
17
13
9
19
19
10
18
23
18
19
9
6
11
12
24
13
10
12
22
4
6
14
11
10
10
10
15
10
7
15
10
12
16
4
Notes:
1 These are the air quality scenarios evaluated: as ;'s is unadjusted air quality; CS is air quality adjusted to just
meet the current standard; the levels of the two percentile forms (99th and 98th) of a 1 -hour daily maximum
potential alternative standard are given.
July 2009
149
-------
Table 7-14. Mean number of modeled days per year with 5-minute maximum concentrations
above 400 ppb in 40 selected counties given 2001-2006 air quality as is and that adjusted to just
meet the current and alternative standards.
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Iron
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
as is1
18
0
2
0
6
5
10
0
1
2
3
0
1
25
16
2
0
0
0
6
10
0
1
1
0
0
3
0
5
0
3
0
3
1
0
2
1
3
0
0
CS1
102
9
17
21
36
52
34
23
30
24
28
18
23
50
54
31
2
1
2
34
44
19
25
30
30
10
22
7
19
13
15
12
58
27
1
24
25
25
14
6
99th percentile1
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
0
0
1
0
1
1
1
0
0
2
2
0
1
0
0
1
0
0
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
150
3
0
2
0
4
4
4
1
1
5
5
1
4
3
2
3
0
0
0
3
7
1
1
1
1
0
1
1
1
1
1
1
1
1
0
1
0
1
1
0
200
9
1
3
1
6
8
8
3
2
9
9
2
8
9
6
7
1
1
2
6
13
2
3
3
4
1
2
3
3
2
3
3
2
3
1
3
2
3
4
1
250
15
2
5
3
8
11
11
6
3
12
13
5
12
15
10
12
3
2
5
9
18
5
6
6
10
2
3
6
6
5
5
5
5
6
3
7
5
5
8
2
98th percentile1
100
1
0
1
0
3
3
2
1
1
3
2
0
2
1
1
1
0
0
0
2
3
0
0
0
0
0
0
1
0
0
0
1
0
1
0
0
0
1
0
0
200
14
2
5
3
10
12
12
6
4
12
12
4
11
13
11
10
3
2
3
8
15
5
4
5
8
1
3
5
6
4
4
5
6
5
2
7
4
6
6
2
Notes:
1 These are the air quality scenarios evaluated: as ;'s is unadjusted air quality; CS is air quality adjusted to just
meet the current standard; the levels of the two percentile forms (99th and 98th) of a 1 -hour daily maximum
potential alternative standard are given.
July 2009
150
-------
7.4 VARIABILITY ANALYSIS AND UNCERTAINTY
CHARACTERIZATION
As discussed in section 6.6, there can be variability and uncertainty in risk and exposure
assessments. This section presents a summary of and associated discussions regarding the
degree to which variability was incorporated in the air quality analyses and how the uncertainty
was characterized for the estimated air quality benchmark exceedances.
7.4.1 Variability Analysis
To the maximum extent possible given the data, time, and resources available for the
assessment, staff accounted for variability within the two main components of the air quality
characterization: the ambient monitoring concentrations and the statistical model used to
estimate 5-minute maximum 862 concentrations. The variability accounted for in this analysis is
summarized in Table 7-15.
Table 7-15. Summary of how variability was incorporated into the air quality characterization.
Component
Variability
Comment
Temporal: 10 to 11 years of 1 -
hour and 5-minute monitoring
data
Broader SO2 monitoring network and monitors
reporting 5-minute maximum concentrations.
Subset of 40 counties for detailed analyses
comprised two 3-year periods (2001-2003; 2004-
2006)
Ambient SO2
Monitoring Data
Spatial: 48 states plus 3 US
territories totaling 407
counties.
Broader SO2 monitoring network. Other analyses
considered monitor results separated by
population density. Subset of 40 counties for
detailed analyses comprised 18 states and 1 US
territory.
9 air quality scenarios
40 county analysis included air quality as ;'s, just
meeting the current standard and 5 levels (50,
100, 150, 200, 250 ppb) of two percentile forms
(98th and 99th); effectively creating a varying
decision surface.
5-Minute Peak
Statistical Model
19 peak-to-mean (PMR)
distributions
PMR distributions used non-parametric form
derived from measurement data (complete range
of values from 1 to <12). Three monitor
concentration variability bins used as a surrogate
for variability in local source emissions, along with
seven concentration bins. Twenty simulations
using random sampling generated a best estimate
of exceedances per site-year of data.
7.4.2 Uncertainty Characterization
As discussed in section 6.6, the approach for evaluating uncertainty was adapted from
guidelines outlining how to conduct a qualitative uncertainty characterization (WHO, 2008).
Staff selected the mainly qualitative approach given the limited data available to inform a
July 2009
151
-------
probabilistic uncertainty characterization, and time and resource constraints. This qualitative
approach used here varies from that of WHO (2008) in that the primary focus is placed on
evaluating the impact of the uncertainty; that is, staff qualitatively rate how the source of
uncertainty, in the presence of alternative and possibly improved data or information, may affect
the estimated number of days with benchmark exceedances. In addition, and consistent with the
WHO (2008) guidance, staff discuss the uncertainty in the knowledge-base (e.g., the accuracy of
the data used, acknowledgement of data gaps) and decisions made (e.g., selection of particular
model forms), though qualitative ratings were assigned only to uncertainty regarding the
knowledge-base.
After identifying the key sources of the assessment that may contribute to uncertainty, staff
subjectively scaled the magnitude47 of each identified source of uncertainty and the associated
direction of potential influence to the number of benchmark exceedances. We used a three level
scale to rate the magnitude: low indicated that large changes within the source of uncertainty
would have only a small effect on the estimated number of exceedances, medium implied that a
change within the source of uncertainty may have a proportional effect on the results, and high
indicated that a small change in the source would have a large effect on results. The direction of
influence on number of exceedances was subjectively assigned as over-estimated., under-
estimated, both (uncertainty affects assessment endpoint in either direction), or unknown (no
evidence to judge the uncertainty). Staff also subjectively scaled the knowledge-base uncertainty
associated with each identified source using a three level scale: low indicated significant
confidence in the data used and its applicability to the assessment endpoints, medium implied
that there were some limitations regarding consistency and completeness of the data used or
scientific evidence presented, and high indicated the knowledge-base was extremely limited.
Table 7-16 provides a summary of the sources of uncertainty identified in the air quality
characterization, the level of uncertainty, and the overall judged bias of each. Further discussion
regarding each of these sources of uncertainty and how conclusions were drawn is given in the
sections that follow.
47 This is synonymous with the "level of uncertainty" discussed in WHO (2008), section 5.1.2.2.
July 2009 152
-------
Table 7-16. Summary of qualitative uncertainty analysis for the air quality and health risk characterization.
Source
Type
Influence of Uncertainty on
Air Quality Benchmark
Exceedances
Direction
Magnitude
Knowledge-
Base
Uncertainty
Comments1
Air Quality Data
Database
Quality
Over
Low
Low
INF: There may be a limited number of poor quality high
concentration data within the analytical data sets,
potentially influencing the number of benchmark
exceedances.
KB: Data used in the analyses are of high quality. There
is no other source of monitoring data as comprehensive.
Data are being used in a manner consistent with one of
the defined purposes of ambient monitoring.
Ambient
Measurement
Technique
Interference
Both
Low -
Medium
Medium
INF: Potential interferences can be controlled; the
influence may be of greater magnitude when considering
upward concentration adjustment procedure.
KB: Limited knowledge on concentration dependencies at
high concentrations. Limited knowledge of interference
controls applied at individual monitors.
Temporal
Representation
of Monitoring
Data
Scale
Unknown
Low -
Medium
Medium
INF: Temporal scale is appropriate for analysis performed.
Most data used are screened for temporal completeness;
however where 5-minute concentrations were reported,
data were not screened for completeness.
KB: Limited knowledge on direction or magnitude;
however 60% of data used would have passed
completeness criteria.
Missing Data
Under
Low
Low
INF: Staff assumed there was an equal probability of
missing low and high concentration 5-minute
measurements; there could be a few missing high
concentration data that would lead to underestimation in
benchmark exceedances. No interpolation was
performed.
KB: All available data are quality assured; most of the
data used were temporally complete.
July 2009
153
-------
Source
Type
Years
Evaluated
Influence of Uncertainty on
Air Quality Benchmark
Exceedances
Direction
Over
Magnitude
Low
Knowledge-
Base
Uncertainty
Low
Comments
INF & KB: Little variation in COV and PMRs over years of
analysis. Estimates of the probability of exceedances are
likely not affected. Estimated number of exceedances
could be influenced by historically high concentrations.
Spatial
Representation
of Monitoring
Network
Broader SO2
Network and 40
County Data
Set
Under
Medium
High
INF: It is possible that the current network is not
adequately capturing 1-hour SO2 from a few localized
sources. However, given the purpose of the network and
purpose of the assessment, staff judges there may be at
most a medium level of influence on results with improved
spatial representation.
KB: Many site-years available from monitors reporting 1-
hour concentrations; However, there are no data available
to evaluate the spatial representativeness of existing
network.
5-minute
Maximum SO2
Under
Medium
High
INF: Distribution of sources potentially influencing
monitors is similar to that of the broader SO2 network even
with limited geographic span.
KB: Very few site-years available from monitors reporting
5-minute measurements.
Air Quality
Adjustment
Procedure
Proportional
Approach Used
Both
Low -
Medium
Medium
INF: Depends on the degree of proportionality in the air
quality distribution and the magnitude of the ambient
concentration adjustment.
KB: Proportional approach judged adequate in
representing the alternative air quality scenarios.
However, evaluation only conducted in 7 of 40 counties,
was dependent on historic air quality as representative of
alternative scenarios, and there was some evidence of
deviation from proportionality. Also only one adjustment
method was investigated.
July 2009
154
-------
Source
Statistical Model
Used for
Estimating 5-
minute SO2
Concentrations
Potential Health
Risk Endpoints
Used2
Type
Spatial Scale
Data Screening
Temporal
Variation in
PMRs
Distribution
Form of PMRs
Accuracy
Reproducibility
Ambient SO2 as
an Indicator of
SO2 Exposure
Influence of Uncertainty on
Air Quality Benchmark
Exceedances
Direction
Both
Over
None
None
Both
None
Over
Magnitude
Medium
Low
Low
Low
Low-
Medium
Low
Medium
Knowledge-
Base
Uncertainty
High
Low
Low
Low
Medium
Low
High
Comments1
INF: The rate of change in concentrations overtime was
moderately different at monitors within a county.
KB: Analysis is dependent on historic air quality as
representative of alternative air quality scenarios. There
is lack of knowledge regarding how changes in emissions
would affect multiple monitors in a county.
INF & KB: Less than 2% of data were removed.
Physically realistic PMR bounds were set. Screened data
were mostly of low 1-hour concentrations that would never
generate a benchmark exceedance.
INF: Consistency in PMRs across period of analysis.
KB: Consistency in PMRs when compared with late 1980s
and early 1990s ambient monitoring data.
INF & KB: Non-parametric distributions were determined
the most appropriate for the analysis.
INF: Accuracy assessment indicated good agreement,
though at upper and lower tails of prediction distribution,
the number of exceedances were under- and over-
estimated, respectively.
KB: Though cross-validation results were reasonable,
there may be additional influential variables that may be
important in the model construction and possibly not
available in extrapolating to the broader data set.
INF & KB: Limited variation observed in the estimated
mean number of benchmark exceedances following
random sampling error analysis.
INF: Long-term time averaging comparisons indicate a
strong proportional relationship between ambient
concentration and personal exposure.
KB: The relationship between 5-minute personal exposure
and ambient concentration is not known.
July 2009
155
-------
Source
Type
Consideration
of Susceptible
Populations
Averaging Time
Single Counts
of
\J I
Fyrpprlflnrp^
^yvot? t?u ci i iot?o
versus Multiple
Fyppprlflnrp^
^yvoc/c/ucii 10^0
per day
Influence of Uncertainty on
Air Quality Benchmark
Exceedances
Direction
Unknown
None
Under
Magnitude
Low
Low
Low
Knowledge-
Base
Uncertainty
Medium
Low
Medium
Comments1
INF & KB: Severe asthmatics are typically not challenged
in clinical studies due to expectations of a significant
adverse response. Potential health risk could be over- or
under-estimated depending on the level of the lowest
benchmark selected to represent susceptible individuals.
KB: There is no clear quantitative evidence indicating
lowest benchmark would either be health protective or at a
level a susceptible individual would respond.
INF & KB: consistently no difference reported in observed
responses from either 5- or 10-minute clinical studies.
INF: Potential health risk may be under-estimated
because approximately 50% of days with a single
exceedance correspond with another (or more)
exceedance(s) in that same day. However, in this air
quality analysis, time of exposure is not considered, thus
limiting the relevance of multiple exceedances.
KB: Frequency of multiple exceedances per day using
existing measurement data is known for limited number of
monitoring sites.
Notes:
1 INF refers to comments associated with the influence rating; KB refers to comments associated with the knowledge-base rating.
2 In these cases the influence of the uncertainty to the potential health risk is discussed, not the influence to the estimated number of exceedances.
July 2009
156
-------
7.4.2.1 Air Quality Data
The purpose of this section is to discuss staff assumptions and potential uncertainties
associated with the data used to construct the various analytical data sets. While the data are
being used in a manner consistent with one of the defined purposes of ambient monitoring (i.e.,
assessing population exposure), both the source of data and its associated quality are discussed.
The uncertainty regarding temporal and spatial components of the ambient monitoring data sets
is discussed in sections 7.4.2.3 and 7.4.2.4, respectively.
The Air Quality System (AQS) contains ambient SO2 concentrations collected by EPA,
state, local, and tribal air pollution control agencies from hundreds of monitoring stations across
the U.S. There are no alternative ambient monitoring data sets available that are as
comprehensive as those within AQS. There might be ambient monitoring data available that are
not included in the AQS however, staff assumed that given similar collection techniques and
quality assurance methods that they would be complementary to AQS monitoring data.
One basic assumption is that the AQS 862 air quality data used are quality assured
already. Methods exist for ensuring the precision and accuracy of the ambient monitoring data
(e.g., EPA, 1983). Reported concentrations contain only valid measures, since values with
quality limitations are not entered into the system or are removed following determination of
being of lower quality or flagged. There is likely no selection bias in retaining data that are not
of reasonable quality if the data are in error; it was assumed that selection of high concentration
poor quality data would be just as likely as low concentration data of poor quality. However, the
retention of poor quality high concentration data would have greater impact on estimated
numbers of exceedances than poor quality low concentration data. Given the numbers of
measurements used for the analyses though, it is likely that even if a few poor quality high
concentration data are present in the analytical data sets, they would not have a large impact on
the results presented here. In addition, a quantitative analysis of available duplicate measures
(i.e., originating from co-location of ambient monitors or by duplicate reporting of ambient
concentrations, see Appendix A-3) indicated little to no difference in the duplicate values or in
the selection of one particular reported (or measured) value over another.
Based on this evaluation, the source and the quality of the ambient monitoring data used
likely contribute minimally to uncertainty in the estimated number of benchmark exceedances.
July 2009 157
-------
Thus, there is both a low level of uncertainty in the knowledge-base and in the subjectivity of
choices made by staff.
7.4.2.2 Ambient Measurement Technique
One potential source of uncertainty within the SO2 air quality measurements is from
interference with other compounds. The ISA notes several sources of positive and negative
interference that could increase the uncertainty in the measurement of ambient SC>2
concentrations (ISA, sections 2.3.1 and 2.3.2). Many of the identified sources (e.g., polycyclic
aromatic hydrocarbons, stray light, collisional quenching) were described as having limited
impact on 862 measurement due to the presence of instrument controls that prevent the
interference.
The actual impact on any individual monitor though is unknown; the presence of either
negative or positive interference, and the degree of interference contributed by one or the other,
has not been quantified for any ambient monitor. In addition, it is not known whether there is a
concentration dependence on the amount of interference. This may be an important uncertainty
in considering the air quality concentrations adjusted to just meet the current and potential
alternative standards.
Reported ambient monitoring concentrations could be either over- or under-estimated
depending on the type of interference present. Staff judges the magnitude of influence as low to
medium, given the potential range of instrument controls present (low magnitude) and possibility
for concentration dependence (medium magnitude). The uncertainty in the knowledge-base is
judged as medium given the limited quantitative evidence available to assess the potential
direction and magnitude of interference at individual monitors, as well as limited evidence
regarding the presence of concentration dependence.
7.4.2.3 Temporal Representation of Monitoring Data
Three components of uncertainty were evaluated regarding the temporal representation of
the monitoring data. These include uncertainty in the temporal scale (i.e., averaging time of
measurements and completeness criteria), how missing data were treated in the analysis, and
long term trends in ambient monitoring and concentration variability.
The air quality analysis relied on quality assured 5-minute and 1-hour average 862
measurement data (see section 7.4.2.1) and are of the same temporal scale as identified potential
July 2009 158
-------
health effect benchmarks, where 5-minute measurements were reported. There are frequent
missing values within a given valid year that may increase the level of uncertainty in temporal
concentration distributions and model estimations (see below); however, given the level of the
benchmark concentrations and the low frequency of benchmark exceedances and overall
completeness of the monitoring data, it is likely of limited consequence. The magnitude of
impact on estimated benchmark exceedances could be significant if some seasons, day-types
(e.g., weekday/weekend), or times of the day (e.g., nighttime or daytime) were not equally
represented in the data analysis group. For the analyses performed using the broader 862
monitoring network and the 40-county data set, a valid year of ambient monitoring was based on
75 percent complete hours/day and days/quarter, and having all four complete quarters/year. The
process of assuring temporal completeness prevented potentially influential monitoring data from
adversely affecting the air quality characterization using these data sets.
However, there is greater uncertainty in the temporal representation of the combined 5-
minute and 1-hour measurement data set because all of the available data were used without
considering the standard 75% completeness criteria. Staff elected to use all of the 5-minute SC>2
measurement data rather than further reducing the already limited number of samples and
locations represented. The 5-minute measurement data set did however undergo a limited
screening that improved the quality of the data set. This included removal of duplicate
reporting/measurements, exclusion of concentrations < 0.1 ppb, and screening for technically
impossible PMRs (see section 7.4.2.6). These screenings and use of the 5-minute data without
the same completeness criteria as the other data analysis groups though would tend to decrease
the temporal representation, potentially influencing the observed probability and the estimated
number of benchmark exceedances.
Therefore, staff judges the magnitude of influence from this source of uncertainty as low
to medium, with a greater magnitude of influence assigned to observations reported for the 5-
minute data set and its application in the statistical model. While staff has not performed
analyses to determine direction and magnitude of impact in applying the completeness criteria to
July 2009 159
-------
the 5-minute data set, the uncertainty in the knowledge-base is judged as medium given the
overall temporal representation of most of the site-years of data.48
Data were not interpolated in the analysis; missing data were not substituted with
estimated values and concentrations reported as zero were used as is. For the missing data, it is
assumed here that missing values are not systematic, i.e., both high and low concentration data
would be absent in equal proportions. There are methods available that can account for time-of-
day, day-of-week, and seasonal variation in ambient monitoring concentrations. However, if a
method were selected, it would have to not simply interpolate the data but also accurately
estimate the probability of peak 1-hour 862 concentrations that could occur outside the
predictive range of the method. It was judged that if such a method was available or one was
developed to substitute data, it would likely add to a similar level of uncertainty as not choosing
to substitute the missing values. Again, this can be viewed as having a limited impact on the
estimated number of exceedances because using the validity criteria selected for the most
temporally representative and complete ambient monitoring data sets possible. In addition, when
using the concentrations reported as zero, there is likely limited impact on the estimated number
of exceedances and associated probability of exceedances. It is possible that some missing data
could have been at a high enough concentration to either exceed a benchmark or result in an
estimated benchmark exceedance, implying the direction of influence is towards under-
estimating benchmark exceedances. However, given the temporal completeness of much of the
data used characterizing air quality, staff judges both the magnitude of influence of missing data
and the uncertainty associated with the knowledge-base to be low.
There is uncertainty associated with the selection of monitoring years, particularly if
concentrations vary significantly between monitors and across the two averaging times. When
using historical monitoring data, staff assumed that the sources present at that time have similar
emissions and emission profiles as the current sources. It is clear that the number of 862
monitoring sites in the U.S. has changed over time, with a trend of decreasing number of
monitors most evident for those reporting the 5-minute maximum SC>2 concentrations (Figure 7-
31). Five-minute SC>2 concentrations have been reported in fewer monitors than the 1-hour SC>2
concentrations; generally only a few site-years of data exist for 5-minute SC>2 concentrations
48 Screening for completeness using the 75% hours/day and days/year criteria would have resulted in only 85 site-
years of data. However, this screened data set would include 1,431,470 hours or 60% of the data set used in the
current analyses.
July 2009 160
-------
(Appendix A, Table A.l-1). This is the reason why, given the limited number of measurements,
all of the 5-minute maximum 862 data were used in developing the statistical relationships and
for the model evaluation without requiring the 75% completeness criteria to be met.
o>
c
60
C 50
= 40
I S 30:
•5 I 20
SI 10
11122222222
99900000000
99900000000
7890
123456
Year
Figure 7-31. Temporal trends in the number of ambient monitors in operation per year for
monitors reporting both 5-minute and 1-hour SO2 concentrations.
However, the variability in monitoring concentrations (both the 1-hour and 5-minute
maximum 802) does not change significantly across most monitoring years (i.e., years 1997
though 2004) and there is a comparable range between the two averaging times (Figure 7-32).
There is some compression in the range of COVs considering some of the more recent years of
data, most notable for year 2007. This is possibly due to the reduction in the number of ambient
monitors in operation (Figure 7-31) rather than a reduction in the temporal variability in 5-
minute or 1-hour concentrations at particular monitors. There may be an over-estimate in the
number of benchmark exceedances where there is a broad range of years used in the
characterization. However, the estimated probability of exceedances is likely not influenced by
year given that the analysis controls for concentration levels and variability changes that may
have occurred over time. Furthermore, the selection of a subset of the recent air quality data
(2001-2006) used for detailed analyses may reduce the potential impact from changes in
national- or location-specific source influences (if one is present). Therefore, due to the limited
variation in temporal trends in COV for both 5-minute and 1-hour 862 and analysis design (i.e.,
July 2009
161
-------
controlling for concentration level changes, limiting the span of years analyzed) the overall
magnitude of influence is expected to be low.
5-Minute Maximum SO2
1-hourSO2
500
400
gsoo
8 200
100
0
T T T
nnnPfU
TYTTTT
111222
999000
999000
78901 2
I l_
TT
2 2
0 0
0 0
3 4
T
2
0
0
5
T
2
0
0
6
5
1
?
-
-
-
-
-
j
T T
11 11111
TT???TT
2 1112222
0 9990000
0 9990000
7 7890123
Year
Year
111 j
TTT?
2222
0000
0000
4567
Figure 7-32. Temporal trends in the coefficient of variability (COV) for 5-minute maximum and 1 -
hour concentrations at the monitors that reported both 5-minute and 1-hour SO2
concentrations. The number of monitors operating in each year is depicted in Figure 7-
31.
7.4.2.4 Spatial Representation of Monitoring Network
The spatial representativeness of the monitoring network can be a source of uncertainty,
particularly if the monitoring network is not dense enough to resolve the spatial variability in
ambient SC>2 concentrations and if the monitors are not effectively distributed to reflect
population exposure. Relative to the physical area, staff acknowledges there are only a few
monitors, particularly when considering the set of monitors that reported 5-minute maximum
SC>2. The magnitude and direction of influence on the modeled or measured benchmark
exceedances will depend on ambient monitoring objectives, monitoring scale, the distribution of
SO2 emission sources, and whether there is large variability in monitoring surface, i.e., areas of
differing terrain that are not adequately represented by the current distribution of monitors.
These elements will be broadly discussed for each of the data sets used in the air quality
characterization and how they could potentially affect the number and probability of benchmark
exceedances. The three data sets of interest include monitors from the broader 862 network
July 2009
162
-------
(including monitors within the 40 selected counties) and the monitors reporting 5-minute SC>2
concentrations.
The broader 1-hour monitoring network, by definition, is the most comprehensive data
set of the three when considering the number of monitors (n=809) and geographic representation
(48 U.S. States, Washington DC, Puerto Rico, and U.S. Virgin Islands). The air quality
characterization is improved with the inclusion of modeled 5-minute benchmark exceedances in
these areas where 5-minute measurements were not reported. In addition, the use of the broader
SC>2 monitoring network in this assessment could assist in identifying and prioritizing locations
to begin reporting 5-minute 862 measurements. However, the broader geographic span of
ambient monitoring does not necessarily confer spatial representativeness. The spatial
representativeness of the broader SO2 monitoring network would remain dependent on the siting
of the monitors with respect to important emission sources and potentially exposed populations.
Staff assumes that the network design, to a large degree, provides adequate spatial representation
of the ambient SC>2 air quality. This may apply to a greater degree to the 40-County data set that
used a minimum number of monitors (i.e., >2) to represent a set geographical area (i.e., a
county).
Staff acknowledges that in using the broader 862 monitoring network and 40-County
data set as an indicator of exposure, there could be local areas that are spatially under-
represented. Furthermore, portions of the air quality characterization used monitors meeting a 75
percent completeness criterion, without taking into account the monitoring objectives, scale, or
land use. Thus, there may be a reduction in spatial representation due to either the inclusion or
exclusion of monitors sited near local SC>2 source emissions as a result of the completeness
screening. Staff estimates that the magnitude of influence to the number of benchmark
exceedances may be at most a medium level in the presence of supplemental spatial monitoring,
given the purposes of both the current monitoring network and the air quality characterization.
We also judge there would be limited influence on the probability of exceedances with improved
spatial representation, given that the probability estimate is driven by ambient concentration
level and concentration variability, two variables that have been well characterized by the current
ambient monitoring network. In the absence of additional measurements or modeling of the
spatial heterogeneity of 1-hour ambient SC>2 concentrations though, staff assigns a high level of
uncertainty to the knowledge-base.
July 2009 163
-------
The overall SC>2 monitoring network design is also responsible for siting monitors that
reported 5-minute concentrations. As a result, staff expects that monitor siting is appropriate and
spatially representative for the same reasons discussed above. However, because the monitors
reporting 5-minute concentrations are not part of a designed 5-minute SC>2 monitoring network
but are entirely voluntary, the direction and magnitude of influence on observed or estimated
benchmark exceedances is largely unknown. Note that there were far fewer monitors reporting
5-minute concentrations used in certain analyses (n=98), representing a limited geographic scope
in comparison with the broader 862 monitoring network. In addition, a greater percentage of
monitors reporting 5-minute concentrations had a source-oriented objective (Figure 7-3).
However, an analysis of the monitoring attributes indicated similar distributions in the types of
sources and the total emissions potentially impacting both sets of data (Figure 7-5). This
suggests that the spatial representation of the monitors reporting 5-minute concentrations may be
similar to that of the broader SC>2 monitoring network regarding proximity to similar SC>2
sources. In the absence of additional measurements or modeling of the spatial heterogeneity of
5-minute ambient SC>2 concentrations, staff assigns a high level of uncertainty to the knowledge-
base.
7.4.2.5 Air Quality Adjustment Procedure
There is uncertainty in the air quality adjustment procedure due to the uncertainty of the
true relationship between the adjusted concentrations that are simulating a hypothetical scenario
and the as is air quality. The adjustment factors used for the current and the potential alternative
standards each assumed that all hourly concentrations will change proportionately at each
ambient monitoring site. Two elements of this source of uncertainty are discussed, namely
uncertainty regarding the proportional approach used and the universal application of the
approach to all ambient monitors within each location.
Different sources have different temporal emission profiles, so that equally applied
changes to the concentrations at the ambient monitors to simulate hypothetical changes in
emissions may not correspond well within all portions of the concentration distribution. When
adjusting concentrations upward to just meeting the current standard, the proportional adjustment
used an equivalent multiplicative factor derived from the annual mean or daily mean
concentration and equally applied that factor to all portions of the concentration distribution, that
is, the upper tails were treated the same as the area of central tendency. This may not necessarily
July 2009 164
-------
reflect changes in an overall emissions profile that may result from, for example, an increase in
the number of sources in a location. It is possible that while the mean concentration measured at
an ambient monitor may increase with an increase in the source emissions affecting
concentrations measured at the monitor, the tails of the hourly concentration distribution might
not have the same proportional increase. The increase in concentration at the tails of the
distribution could be greater or it could be less than that observed at the mean and is dependent
largely on the type of sources influencing the monitor and the source operating conditions.
Adjusting the ambient concentrations upwards to simulate the potential alternative standards also
carries a similar level of uncertainty although the multiplicative factors were derived from the
upper percentiles of the 1-hour daily maximum 862 concentrations, rather than the mean, and
then applied to the 1-hour SO2 concentrations equally. If there are deviations from
proportionality, the magnitude of influence is likely related to the magnitude of the concentration
adjustment factor used. Therefore, there is likely greater uncertainty in the estimated benchmark
levels when evaluating the current and the 250 ppb 99th percentile alternative standards (which
have the highest adjustment factors), than when considering the 50 ppb and 100 ppb 99th
percentile alternative standards (which have the lowest adjustment factors).
In each of these instances of adjusting the concentrations upwards, one could argue that
there may be an associated over-estimation in the concentrations at the upper tails of the
distributions, possibly leading to over-estimation in the numbers of exceedances of benchmark
levels. An analysis was performed using monitors from seven counties evaluated in the air
quality characterization to investigate how distributions of hourly SC>2 concentrations have
changed over time (Rizzo, 2009). The analysis indicates that a proportional approach is a
reasonable model for simulating higher concentrations at most monitoring sites, since
historically, SC>2 concentrations have decreased linearly across the entire concentration
distribution at each of the monitoring sites and counties evaluated.
At some of monitoring sites analyzed however, there were features not consistent with a
completely proportional relationship. This included deviation from linearity primarily at the
maximum or minimum percentile concentrations, some indication of curvilinear relationships,
and the presence of either a positive or negative regression intercept (Rizzo, 2009). Where
multiple monitors were present in a location, there tended to be a mixture of each of these
conditions including proportionality (e.g., see Figure 7-33). Not all of the counties analyzed as
July 2009 165
-------
part of the air quality characterization were included in the evaluation, thus staff assumed that the
findings of the Rizzo (2009) analysis were applicable of the 40-County data set. Given the
observed range of deviations from proportionality and the level of the concentration adjustment,
we judge the magnitude of influence to the estimated benchmark exceedances as between low to
medium. The estimated number of benchmark exceedances could be either over- or under-
estimated, dependent largely on an individual monitor's air quality distribution and its
relationship with proportionality. While staff judged the proportional approach as appropriate, it
was based on analyses using historical monitoring data. The uncertainty about future source
emission control scenarios is largely unknown. In addition, only one approach was investigated,
suggesting that the level of the knowledge-base uncertainty is medium.
Staff applied the proportional adjustment approach universally to all monitors in each
county for consistency. The purpose was to preserve the inherent variability in the concentration
distribution which has been shown to be relatively consistent with large changes in concentration
level. There is however uncertainty associated with emission changes that would affect the
concentrations at the monitor having the highest concentration (e.g., the highest annual mean,
98th or 99th percentile 1-hour concentration) that may not necessarily be reflected in the same
proportion at other lower concentration sites. This could result in either over- or under-
estimations in the number of exceedances at lower concentration sites within a county where the
current or alternative standard scenarios were evaluated. For example, Figure 7-33 shows the
daily maximum 1-hour SC>2 concentration percentiles for five ambient monitors in Allegheny
County PA, where each of the ambient monitors were in operation for years 1998 and 2007.
While all five of the monitors generally demonstrate features of proportionality, the differences
in regression slope indicate that the rate of change in the concentration distribution was not equal
when comparing these monitors for these two monitoring years. These results suggest that even
if all monitors within a county demonstrate proportionality, there may be either over- or under-
estimations in 862 concentrations following the 1-hour concentration adjustment. Staff had
limited time and resources to investigate the potential impact of this on the number of benchmark
exceedances, though we estimate the magnitude of influence as medium based on the range of
observed slopes in the seven counties investigated. The level of uncertainty in the knowledge-
base is judged high. This rating is based on the uncertainty regarding how the historical and
recent ambient data comparisons relate to the simulated air quality scenario and the lack of
July 2009 166
-------
knowledge regarding how source emission changes would affect multiple monitors within a
county.
High Year: 1998 Low Year: 2007
0,00
I I
0.05
0.10
I
0.15
i
0.20
i
E
Q_
0.
O
o
0>
£ 0.25 -
1 0.20 -
| 0.15-
0.10 -
0.05 -
0.00 -
420030064
RA2: 0,96
420030002
RA2: 0.99
420030067
RA2: 0.97
- 0.25
- 0.20
- 0.15
- 0.10
- 0.05
- 0.00
420030010
RA2: 0.99
420030021
RA2: 0.97
0.00 0.05 0.10 0.15 0.20
0.00 0.05 0.10 0.15 0.20
High Year Percentile Cone, (ppm)
Figure 7-33. Comparison of measured daily maximum SO2 concentration percentiles in Allegheny
County PA for one high concentration year (1998) versus a low concentration years
(2007) at five ambient monitors.
7.4.2.6 Statistical Model Used for Estimating 5-minute SO 2 Concentrations
Five components of uncertainty were identified regarding the statistical model and its
impact on the estimated number of benchmark exceedances. These include 1) the impact from
how the PMR data were screened, 2) the temporal representation of data used in the statistical
model development, 3) the form of the distribution used to represent the PMRs, 4) the accuracy
of the model in predicting daily 5-minute maximum concentrations, and (5) the reproducibility of
the model predictions.
Staff identified data for removal from the final combined 5-minute and 1-hour ambient
measurement data set using the PMR as a screening criterion. The calculation of PMRs less than
1 implies the 5-minute peak is less than the 1-hour average, a physical impossibility, and values
July 2009
167
-------
>12 are a mathematical impossibility. The 5-minute ambient monitoring data were screened for
values outside of these bounds,49 increasing confidence in the relevance of PMRs used for
development of the statistical model. While a total of 40,665 data points were excluded from the
data set using the PMR criterion, this comprised less than 2% of the data available to develop the
PMR relationship. It was assumed that the criterion used for the data removal would not
adversely influence the estimated number of benchmark exceedances in the modeling performed
since it was only directed towards identifying unrealistic 5-minute and 1-hour concentration
combinations.
Analysis of the data screened by staff revealed that nearly all of the data are for where the
calculated PMR was less than one (98% of screened samples) and most of the 1-hour
concentrations (approximately 95%) were less than or equal to 5 ppb (Table 7-17). An
alternative approach to developing the PMR distributions could have been to include the
screened data with an assigned PMR value of one (for where the original PMR was less than
one) or twelve (for where the original PMR was greater than twelve) based the 5-minute and 1-
hour concentration distributions. If included, these data would have virtually no influence on the
estimated number of benchmark exceedances. This is because 1-hour concentrations < 8.3 ppb
combined with the PMR distribution principally affected by inclusion of newly assigned ratios
(i.e., the < 5 ppb concentration bin) would never generate a benchmark exceedance. Given the
limited number of samples removed from further analysis and recognizing there would be less
uncertainty when using a data set comprised of PMRs with realistic bounds rather than one using
all possible PMR values, staff judges the magnitude of the influence associated with the
screening of the 5-minute data as low. In excluding the mostly lower concentration data (as
compared to the final data set used) there may be an over-estimation in the percent and
probability of exceedances.
49 It is possible to have a PMR equal to 12. This value is achieved with one 5-minute concentration above zero and
the other eleven 5-minute values reporting concentrations of zero. Data used in developing the statistical
relationship were screened for values with a PMR equal to 12 however, because it could not be used in the
AERMOD/APEX modeling. It is of little consequence because the distributions chosen in estimating the 5-minute
concentrations included the 1st through the 99th percentiles, not the minimum and maximum values.
July 2009 168
-------
Table 7-17. Summary of descriptive statistics for the data removed using peak-to-mean ratio
criterion and the final 1-hour and 5-minute maximum SO2 data set used to develop PMRs.
Statistic1
mean
p99
p95
p50
p5
P1
Data removed
PMR<1
(n = 39,861)
5-min max
(ppb)
1
6
3
1
1
0.2
1-hour
(ppb)
2
10
5
1.6
1.1
0.45
PMR > 12
(n = 804)
5-min max
(ppb)
29
174
82
15.5
12
4
1-hour
(ppb)
2
10
4
1
0.9
0.1
Final data set
(n = 2,367,686)
5-min max
(ppb)
10
100
37
3
1
1
1-hour
(ppb)
6
50
21
2
1
0.2
Notes:
1 mean is the arithmetic average; p99, p95, p50, p5, p1 are the 99th, 95th, 50th, 5th and 1st
percentiles of the concentration distribution.
The use of all screened 5-minute maximum SC>2 data (1997 to 2007) in developing the
PMR distributions assumes that the source emissions present at that time of measurement are
similar to other year source emissions. It could be possible that there is greater uncertainty in the
estimated number of exceedances in areas where year-to-year source emissions deviate from a
consistent pattern. However, as noted with the concentration variability, the PMRs derived from
the 5-minute maximum measurement data do not have a clear trend with monitoring year. Over
the 11 -year period, the mean of each monitor's annual average PMR is about 1.6 (medians of
1.5; 25th percentiles of 1.4; 75th percentiles of 1.7) (Figure 7-34). This general trend in mean
PMRs is consistent with the population-based value used by Stoeckenius et al. (1990) for
exposure analyses (mean of 1.6; median of 1.5) and ambient monitor concentration analyses
conducted by SAI (1996) (mean 1.7; median 1.5).50 While there is some indication of greater
variability in the PMRs for years 2004-2005 compared with some of the other years used, overall
the consistent pattern over time indicates that the use of the older ambient monitoring data in
developing the statistical model would have a negligible impact on the predicted concentrations
and subsequently the estimated number of benchmark exceedances (i.e., low influence with no
apparent direction). Given the consistency of the PMRs derived using recent air quality with that
of the earlier analyses, the uncertainty regarding the knowledge-base is judged as low.
' Data from Table 2-18 of Stoeckenius (1990) for the Scottish Rites monitoring site and Table 5-2 of SAI (1996).
July 2009
169
-------
11122222222
99900000000
99900000000
78901 234567
Year
Figure 7-34. Distributions of annual average peak-to-mean ratios (PMRs) derived from the 98
monitors reporting both 5-minute maximum and 1-hour SO2 concentrations, Years 1997
through 2007.
The PMRs distributions for each COV and concentration bin were represented by a non-
parametric form condensed to single percentiles, with each value from the distribution having an
equal probability of selection. While there may be other distribution forms that could be
alternatively selected, staff judged that use of a fitted distribution would not improve the
representation of the true population of PMRs compared with a non-parametric form, and that
there would likely be no reduction in the uncertainty of estimated number of exceedances if
using a parameterized distribution. While some of the PMR distributions were similar to a
lognormal distribution (for example see Figure 7-35), 93 of 95 possible statistical tests performed
indicated the distributions were statistically distinct (p<0.01) from any of the tested forms (i.e.,
normal, lognormal, Weibull, gamma, and exponential) (see Figure 7-35 as an example). The
PMRs derived from monitors having the greatest COV (all concentration bins) and those derived
from the lowest concentration bins (all COV bins) were most common in exhibiting atypical
distribution forms. Even when considering practical judgments regarding a potential parametric
form (i.e., beyond simply using statistically significant differences as a criterion), most of the
observed PMR distributions had large deviations from parametric distributions such as that
illustrated by Figure 7-36.
July 2009
170
-------
COVbin = b COVconcbin=4
COVbin=b COVconcbin=4
PMR distributions
N 3807
Mean 1.689
Std Dev 0.572
Skewness 2.875
10 H
P
M
R
5.0
10.0
12.5
PMR
Normal(Mu=1.6692 Sgma=0.5724)
Lognormal(Thela=0 Shape=.28 Scale=.47)
Exponential(Theta=0 Scale=1.67)
Weibull(Theta=0 Shape=2.7 Scale=1.9)
GammaCTheta=0 Shape =11.5 Scale=0.15)
Threshold=0.9328, Scale = -0549
.01 50 90 99 99.9 99.99
Lognormal Percentiles (Sigma=0.70339)
Figure 7-35. Example histogram of peak-to-mean ratios (PMRs) compared with four fitted
distributions derived from monitors reporting the 5-minute maximum and 1-hour SO2
concentrations (left) and the same PMRs compared with expected lognormal
percentiles (right). PMRs were derived from monitors with medium level variability
(COVbin = b) and 1-hour concentrations between 75 and 150 ppb (COVconcbin = 4).
In addition, while there is uncertainty associated with the use of the empirically-derived
data in representing the true population of PMRs, assuming a fitted distribution would not be
without its own uncertainties. For example, using a lognormal distribution may underestimate
the observed frequency of certain values of PMRs while overestimating others. For PMR
distributions that are of similar form with the lognormal distribution, it is likely that the small
variation in PMRs selected from a fitted lognormal distribution would have only limited impact
on the estimated 5-minute maximum SC>2 concentrations. For distributions exhibiting no
similarities to any parametric distribution, experimental justification criteria would need to be
developed in selecting the most appropriate form of the distribution, likely requiring multiple test
iterations, potentially yielding distributions with greater uncertainty than those of a non-
parametric form (e.g., WHO, 2008 page 28). Each of these additional evaluations and iterations
would require time and resources not available to staff. Furthermore, the sample sizes for many
of the PMR distributions used are well above 1,000 (only 5 of the 19 distributions had fewer than
July 2009
171
-------
1,000, with all distributions having greater than 100 samples), providing support that the true
distribution may be well-represented by the non-parametric form. Each of these factors
mentioned (uncertainty in the form of the distribution, limits on time and resources available, and
numbers of samples available) were considered and it was decided by staff that the non-
parametric distribution derived from the measurement data would be most appropriate.
Therefore, it is judged that the magnitude of influence on the estimated benchmark exceedances
is low along with no apparent direction of influence. Since staff employed both statistical and
practical comparisons in selection of the distribution form to the maximum extent allowable, the
uncertainty regarding the knowledge-base is judged as low.
COVbin = c COVconcbin=2
121
10
8
P
M 6
R
- Threshold=O.S7B7, Scale=0.0643
.001 95 99 99.9 99.99 99399
Lognorrnal Percerrtles (Sigma=0.904401}
Figure 7-36. Example of a measured peak-to-mean ratio (PMRs) distribution with the percentiles
of a fitted lognormal distribution. PMRs were derived from monitors with high COV
(COVbin = c) and 1-hour concentrations between 5 and 10 ppb (COVconcbin = 2).
The accuracy in the predicted daily 5-minute maximum 862 concentrations above each of
the benchmark levels was evaluated using measured concentrations. The results indicated that
on average, the statistical model performed well in estimating of these short-term peak
concentrations (section 7.2.3.4). There was reasonable agreement in observed versus predicted
July 2009
172
-------
numbers of benchmark exceedances for most of the monitoring site-years (i.e., about 90% of the
data set) and for all of the benchmark levels. Based on this overall assessment of model
accuracy, the magnitude of influence the selected model has on contributing uncertainty to the
estimated number of exceedances is judged by staff to be low. There was no particular direction
of influence; model predictions were equally over- or under-estimated (Figure 7-37, Table 7-6).
The accuracy assessment indicated the estimated number of days with benchmark
exceedances could be either over- or under-estimated by as many as 20 to 50 days in a year,
primarily at the tails of the prediction distribution. These model prediction errors were limited to
several site-years from a few monitors. Figure 7-37 illustrates the model predicted versus the
observed number of benchmark exceedances at each of the benchmark levels. While there is
generally uniform agreement between the predicted and observed values at the 100 ppb
benchmark, there is deviation in the agreement at the greatest and lowest number of days with
exceedances for the 200, 300, and 400 ppb benchmark levels. For example, there were a few
site-years without any observed benchmark exceedances of 400 ppb, although the statistical
model predicted between 2-15 days in a year. This could indicate that a few of the site-years
may have moderate over-estimations in the number of days with 5-minute maximum SC>2
concentration exceedances, where the estimated number of exceedances is 15 or less. In
addition, site-years with the greatest number of observed exceedances of 400 ppb (about 50 per
year) were consistently under-estimated by the model by about 30%. This could imply that when
the estimated number of days with 5-minute maximum SC>2 concentrations above 400 ppb is 40
per year, the under-estimate may be as large as 15 days per year.
Neither of these model errors appeared systematically related to an individual source
type. Additional monitors sited in the same areas impacted by similar source types had good
agreement between the observed and predicted concentrations. For example, at the monitor with
the greatest number of measured benchmark exceedances (ID 290930030) and largest under-
prediction error, one could argue that variable terrain may be an influential factor. This monitor
is about 1.7 km from a primary smelter and located proximal to a ravine running between the
source and the monitoring site. The nearby monitor (ID 290930030) sited in elevated terrain
(Hogan Mountain) at about 4.6 km from the same source had small prediction errors. These
differences in agreement suggest that when considering any individual monitor, there may be
factors not accounted for by the statistical model that are important in estimating benchmark
July 2009 173
-------
exceedances (e.g., terrain). Based on this model accuracy assessment, the magnitude of
influence the selected model has on contributing uncertainty to the estimated number of
exceedances for individual monitors is likely medium at the lower and upper tails of the
prediction distribution. The direction of the influence is likely over-estimation at the lower
number of exceedances and under-estimation at the greatest number of exceedances.
Though the cross-validation results are encouraging, there may be additional influential
variables not included in the construction of the statistical model that may be important and have
the potential to improve the agreement between the observed and predicted values. There is also
the possibility of influential variables that are not within the data set used for statistical model
development, but exist in the broader 1-hour 862 monitoring data set. Staff judged the
concentration variability and level as appropriate variables for linking the statistical model with
the 1-hour measurement data. In addition, the comparison of ambient monitoring attributes (e.g.,
objectives, local source emissions) also indicated consistency between the monitors reporting 5-
minute maximum concentrations and those reporting only 1-hour average concentrations.
However, in the absence of additional 5-minute measurements in areas where there may be
unique conditions (e.g., terrain or climatologic influences), staff judges there remains a medium
level of uncertainty in the knowledge-base regarding the accuracy in the extrapolation using the
statistical model.
July 2009 174
-------
20 40 60 80 100 120 140 160 180
0 10 20 30 40 50 60 70 80 90 100 110
0
10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70
Observed number of daily benchmark exceedances in a year
Figure 7-37. Comparison of observed and predicted number of daily benchmark exceedances in a
year at the 98 monitors reporting 5-minute maximum SO2 concentrations.
Staff needed to evaluate the reproducibility of the statistical model because random
sampling was employed in generating the PMRs used to estimate 5-minute 862 concentrations.
The purpose of this analysis was to determine the effect of random sampling error on the
estimated number of benchmark exceedances. First, to define terminology used in this analysis:
a model simulation is where each monitor had all of its years of 1-hour data SC>2 used in
estimating 5-minute maximum concentrations and as a result, the number benchmark
exceedances was calculated; a model run is comprised of twenty such independent simulations
(i.e., differing by random number seed) and used to generate a mean number of daily 5-minute
July 2009
175
-------
maximum 862 concentration exceedances for each site-year. This is the same process (i.e., a
model run) that was used in generating the air quality characterization.
The reproducibility of the estimated number of benchmark exceedances was evaluated by
performing ten independent modeling runs (with twenty simulations per model run) using the 40-
county as is air quality data set (i.e., having 610 site-years per model simulation). The output
from each model run was the mean number of days per site-year an exceedance occurred;
therefore, ten mean numbers of exceedances were generated for each of the four benchmarks
using the 610 site-years of data. The maximum difference in those ten means was calculated (the
minimum mean value subtracted from the maximum mean value) giving the range of the ten
means for each benchmark and site-year. For example, in one site-year there were 51, 52, 52,
53, 52, 52, 52, 51, 52, and 52 estimated mean numbers of exceedances of 100 ppb from the 10
model runs. Therefore the range (or maximum difference) is equal to two.
The distributions of the range in mean exceedances by benchmark level are illustrated in
Figure 7-38. The range in the mean number of exceedances based on the ten model runs is less
than five for all benchmark levels and consistently decreases with increasing benchmark level.
On average, maximum difference in the estimated mean numbers of exceedances of 100 ppb was
2 exceedances, while at greater benchmark levels the range was 1 or less. This indicates that the
random sampling error has a low impact to the estimated mean number of exceedances per site-
year.
July 2009 176
-------
n
n
100ppb >200ppb >300ppb >400ppb
benchmark level
Figure 7-38. Distributions of the maximum difference in the estimated mean number of
exceedances per site-year given 10 independent model runs (with 20 simulations per
run). Data used are from 40 county as is air quality (610 site-years). Box represents the
inner quartile range (IQR, or the 25th to 75th percentile), + indicates the mean, whiskers
are 1.5 times the IQR.
July 2009
177
-------
7.4.2.7 Potential Health Risk Endpoints Used
The choice of potential health effect benchmarks levels and the use of those benchmarks
to characterize risks are important uncertainties in the air quality characterization results. Human
exposure is characterized by contact of a pollutant with a person, and as such, the air quality
characterization assumes that the ambient monitoring concentrations can serve as an indicator of
exposure. The ISA reports that personal exposure measurements (PEM) are of limited use since
ambient SO2 concentrations are typically below the detection limit of the personal samplers.
There is no method to quantitatively assess the relationship between 5-minute ambient
monitoring data and 5-minute personal exposures, particularly since personal exposures are time-
averaged over days to weeks, and never by 5-minute averages. Therefore the fraction of actual
5-minute maximum personal exposure concentrations attributed to 5-minute maximum ambient
862 is unknown and thus contributes to uncertainty when using ambient air quality data as an
indicator of human exposure.
An evaluation in the ISA indicates the relationship between longer-term averaged
ambient monitoring concentrations and personal exposures is strong, particularly when ambient
concentrations are above the limit of detection. The strength of the relationship between
personal and ambient SC>2 concentrations is supported further by the limited presence of indoor
sources of 862; much of an individuals' personal exposure is of ambient origin. However, 862
personal exposure concentrations are reportedly a small fraction of ambient concentrations. This
is because local outdoor 862 concentrations are typically half that of the ambient monitoring 862
concentrations, and indoor concentrations about half that of the local outdoor SO2 concentrations
(ISA). Therefore, while the relationship between personal exposures and ambient SC>2
concentrations is strong, the use of monitoring data as an indicator of SC>2 exposure may lead to
an overestimate in the number of peak concentrations those individuals might encounter. While
the magnitude of the uncertainty about the true relationship between actual human exposure and
any given ambient monitor short-term concentration exceedance is unknown, it is judged by staff
to be of a medium magnitude given what is known regarding the relationship between longer-
term PEM and ambient 862 concentrations.
There is uncertainty regarding how susceptible populations were considered in
developing the potential health benchmark levels. The human clinical exposure studies
July 2009 178
-------
evaluated airways responsiveness in mild to moderate asthmatics. Health effect symptoms and
responses were observed in these test subjects exposed to concentrations as low as 200 ppb in the
free-breathing chamber studies. As such, a concentration of 200 ppb could well represent a
lower range of the benchmark level for mild to moderate asthmatics. However, for ethical
reasons, adults with severe asthma and younger asthmatics are not commonly challenged in air
pollutant studies. This is because severe asthmatics and/or asthmatic children may be more
susceptible than mild asthmatic adults to the effects of SC>2 exposure. Therefore, exposure levels
(and hence selected benchmark levels) lower than those used in free-breathing chamber studies
may be important in representing populations with greater susceptibility. Staff selected 100 ppb
as the lowest benchmark level based on effects observed in mild to moderate asthmatics using
facemasks at that level and to consider potential effects in susceptible populations at lower 5-
minute concentrations. In the absence of strong quantitative evidence it is difficult to determine
if 100 ppb would be health protective for asthmatics (mild, moderate, or severe) or if 100 ppb is
a concentration that would elicit an adverse effect. Based on this, staff acknowledges there is
medium uncertainty in the knowledge-base regarding representativeness of the lowest
benchmark level selected, but judge that the magnitude of influence to the estimated health risk
is low given the inclusion of the 100 ppb level.
Staff also acknowledges that there may be uncertainty in the selected potential health
effect benchmark averaging time. For example, the used in this assessment were from studies
where volunteers were exposed to 862 for varying lengths of time. Typically, the SO2 exposure
durations in the controlled human studies were between 5 and 10 minutes. This could be an
important uncertainty because the potential health effect benchmark levels were compared to
concentration exceedances occurring over 5-minutes. That is, if there were a difference in the
response rate for a given concentration level and averaging time, the use of a 5-minute averaging
time could either lead to over- or under-estimation in the health risk characterization. The true
exposure-response relationship may be dependent on both the combined concentration level and
the exposure duration, that is, it is possible that a particular response rate observed at a 10-minute
exposure level of concentration x may be similar to that of a 5-minute exposure level equal to or
greater than concentration x. In this hypothetical scenario, if benchmarks were derived from 10
minute exposures and applied in the evaluation of 5-minute ambient concentrations, the risk
characterization may well be over-estimated. However, the ISA did not distinguish between
July 2009 179
-------
health effects observed following either 5- or 10-minute exposures. Therefore the direction of
influence to the potential health risk is judged as none, and given a general consistency in the
observed responses involving either 5- or 10-minute exposures, staff judges the uncertainty in the
knowledge-base as low.
The health effect endpoint used in the air quality characterization was the observed or
estimated number days the maximum 5-minute SC>2 concentration exceeded a particular
benchmark level. Staff acknowledges that this choice could result in the risk characterization
under-estimating the health risk because there can be multiple exceedances of the benchmark
levels in a day (Table 7-18). Using the monitors reporting 5-minute 862 maximum
concentrations, approximately half of the time there was a single benchmark exceedance in a
day. For most days having an exceedance (about 80-90%), there were no more than three that
occurred in a day. There were several days having many benchmark exceedances within a day
(e.g., > 5), particularly when considering the lowest benchmark levels. However in this air
quality analysis, none of the elements of exposure are considered (e.g., whether or not time of
exposure occurs coincident with elevated activity level), thus limiting the relevance of multiple
exceedances within a day. While the risk characterization could be considered under-estimated,
the magnitude of influence by this source of uncertainty is judged by staff as low given the
defined limits of the air quality characterization. Furthermore, staff acknowledges that multiple
benchmark exceedances of 5-minutes can occur within an hour. This issue and its implications
for characterizing health risk are more relevant to human exposure than the air quality analysis
and are discussed in greater detail in section 8.11.
July 2009 180
-------
Table 7-18. The number and percent of days having multiple benchmark exceedances occurring
in the same day, using monitors reporting the 5-minute maximum SO2 concentrations.
Number of
Exceedances
per Day1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Sum
5-minute SO2 Benchmark Level
> 100 ppb
days2
3806
1923
1093
640
424
286
185
127
100
68
45
38
18
27
7
11
3
6
5
2
0
0
3
0
8817
percent2
43
22
12
7
5
3
2
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
> 200 ppb
days
1390
613
327
152
114
60
52
27
21
14
7
7
4
1
1
2
0
1
0
0
0
0
0
0
2793
percent
50
22
12
5
4
2
2
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
> 300 ppb
days
740
349
183
87
48
25
22
8
4
5
2
0
1
0
1
0
0
1
0
0
0
0
0
0
1476
percent
50
24
12
6
3
2
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
> 400 ppb
days
512
248
111
46
19
15
8
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
961
percent
53
26
12
5
2
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Notes:
1 The number of 5-minute maximum benchmark exceedances within a day could range from 1 to 24
given the number of hours in a day.
The total number of days having the given number of multiple exceedances within the day.
3 The percent of days having an exceedance with the given number of multiple exceedances per
day.
7.5 KEY OBSERVATIONS
Presented below are key observations resulting from the 862 air quality characterization:
• For unadjusted as is air quality at ambient monitors measuring 5-minute maximum
concentrations, nearly 70% of the 471 site-years analyzed had at least one daily 5-minute
maximum concentration above 100 ppb and over 100 site-years (more than 21%) had >
25 days with a daily 5-minute maximum concentration above 100 ppb. Less than half
(44%) of the site-years had at least one daily 5-minute maximum concentration above
200 ppb and only 36 site-years had > 25 days with a daily 5-minute maximum
July 2009
181
-------
concentration above 200 ppb. Approximately 25% and 17% of the 471 site-years
analyzed had at least one daily 5-minute maximum concentration above 300 and 400 ppb,
respectively, with 23 and 12 site-years having > 25 days with a daily 5-minute maximum
concentration above 300 and 400 ppb, respectively (Appendix A, Table A.5-1).
• For any of the air quality scenarios considered, the probability of exceeding the 5-minute
maximum benchmark levels was consistently greater at monitors sited in low-population
density areas compared with high-population density areas. In addition, an increased
probability of any 5-minute benchmark exceedance was consistently related to either
increased 24-hour average or 1-hour daily maximum concentrations.
• For unadjusted air quality in the 40 counties selected for detailed analysis, most counties
are estimated to have, on average, fewer than 50 days per year where the daily 5-minute
maximum ambient SC>2 concentrations are > 100 ppb. Most counties are estimated to
have, on average, 25 days per year with daily 5-minute maximum ambient SC>2
concentrations > 200 ppb. Very few counties are estimated to have more than ten days
with 5-minute maximum SC>2 concentrations > 300 ppb, while nearly half did not have
any days with 5-minute maximum SC>2 concentrations > 400 ppb (Tables 7-11 to 7-14).
• When air quality is adjusted to simulate just meeting the current annual standard in the 40
counties selected for detailed analysis, a hypothetical scenario requiring air quality to be
adjusted upward, all locations evaluated are estimated to have multiple days per year
where 5-minute maximum ambient SC>2 concentrations are > 100 ppb. Most counties are
estimated to have, on average, 100 days or more per year with 5-minute maximum
ambient SC>2 concentrations > 100 ppb, while eight of the forty counties are estimated to
have 200 days or more per year with 5-minute maximum ambient SC>2 concentrations >
100 ppb. Fewer benchmark exceedances are estimated to occur with higher benchmark
levels. For example, only five counties are estimated to have 60 or more days per year
with 5-minute maximum ambient SC>2 concentrations that exceed 300 ppb (Table 7-13)
and only four counties are estimated to 50 or more days per year with 5-minute maximum
ambient 862 concentrations that exceed 400 ppb (Table 7-14).
• In all 40 counties, potential alternative standard levels of 100 and 150 ppb are estimated
to result in fewer days per year with 5-minute maximum SC>2 concentrations > 300 and >
400 ppb than with the current standards and the potential alternative standard levels of
200 and 250 ppb (Tables 7-13 and 7-14).
• When considering the potential 1-hour daily maximum potential alternative standard
levels of 100 and 200 ppb in all 40 counties, corresponding annual average 862
concentrations were typically between 3 and 15 ppb, similar to a range of concentrations
using unadjusted air quality (Appendix A). When considering the potential alternative
standard levels of 200 and 250 ppb, corresponding annual average SC>2 concentrations
were typically between 10 and 30 ppb, similar to the range of concentrations observed
when using adjusted air quality that just meets the current annual standard.
• Of the fifteen uncertainties qualitatively judged to influence the estimated number of days
with air quality benchmark exceedances, three may be associated with over-estimation,
three may be associated with under-estimation, while the remaining uncertainties could
affect results in both directions (four sources), no direction (four sources), or unknown
July 2009 182
-------
direction (one source) (see Table 7-16). The magnitude of influence for four of the six
uncertainties associated with either over- or under-estimation was estimated as low (or
negligible magnitude of influence). Staff judged the two remaining uncertainties as
having a medium magnitude of influence in under-estimating the number of days with
benchmark exceedances, both of which were associated with the spatial representation of
the monitoring network. Based on this overall characterization regarding the direction
and magnitude of influence identified sources of uncertainty, there may be a medium
level under-estimate in the number of days with air quality benchmark exceedances.
For the most part, the knowledge-base uncertainty for sources with unknown or
bidirectional influence ranged from low (four sources) to medium (four sources), though
uncertainty regarding the spatial scale of the air quality adjustment procedure (direction
of influence was both, medium magnitude) was judged as high. The knowledge-base
uncertainty was low for four of the six sources associated with either an under- or over-
estimation direction of influence. A high degree of uncertainty in the knowledge-base
was assigned to the spatial representation of the monitoring network. Based on this
overall characterization regarding the knowledge-base, there is a high level of uncertainty
associated with the most influential source.
Staff identified four other sources of uncertainty in the air quality characterization as
having influence on the characterization of health risk. The most influential and most
uncertain source of the four is associated with the direct use of air quality benchmark
exceedances as an indicator of exposure. The number of days with 5-minute exposures
above benchmark levels would likely be lower than the number of days where there were
ambient 862 concentrations above benchmark levels. Thus, the air quality
characterization may over-estimate the health risk due to this factor
July 2009 183
-------
8. EXPOSURE ANALYSIS
8.1 OVERVIEW
This section documents the methodology and data staff used in the inhalation exposure
assessment and associated health risk characterization for SC>2 conducted in support of the
current review of the SC>2 primary NAAQS. Two important components of the analysis include
the approach for estimating temporally and spatially variable 862 concentrations and simulating
human contact with these pollutant concentrations. The approach was designed to better reflect
exposures that may occur near SO2 emission sources, not necessarily reflected by the existing
ambient monitoring data alone.
Staff used a combined air quality and exposure modeling approach to generate estimates
of 5-minute maximum, 24-hour, and annual average SC>2 exposures within Greene County, MO.
and three Counties within the St. Louis Metropolitan Statistical Area (MSA) for the year 2002.
AERMOD, an EPA recommended dispersion model, was used to estimate 1-hour ambient 862
concentrations using emissions estimates from stationary, non-point, and port sources. The Air
Pollutants Exposure (APEX) model, an EPA human exposure model, was used to estimate 5-
minute population exposures using the census block level hourly SO2 concentrations estimated
by AERMOD and the statistical model described in section 7.2.3. Staff used the person-based
exposure profiles to calculate the number of days per year an individual had at least one 5-minute
exposure above the potential health effect benchmark levels of 100, 200, 300, and 400 ppb.
Exposure and potential health risk were characterized considering recent air quality
conditions (as is), for air quality adjusted to just meet the current 862 primary standards (0.030
ppm, annual average; 0.14 ppm, 24-hour average), and for just meeting potential alternative
standards (see Chapter 5 for selection justification). Specifically, APEX reported the number of
times an individual experienced a day with a 5-minute exposure in excess of 100 ppb through
800 ppb.51 The exposures for each individual were estimated over an entire year therefore,
multiple occurrences of exposures above the benchmark levels are also available.
51 The complete output from APEX includes 5-minute exposure concentrations at 50 ppb increments through 800
ppb which served as an input to the risk assessment performed in Chapter 9. The health effect benchmarks
evaluated in the exposure assessment were defined as 100 to 400 ppb by increments of 100 ppb.
July 2009 184
-------
The approaches used for assessing exposures in Greene County and St. Louis are
described below. Additional model input data and supporting discussion of APEX modeling are
provided in Appendix B. Briefly, the discussion in this Chapter includes the following.
• Description of the inhalation exposure model and associated input data used for Green
County and St. Louis;
• Evaluation of estimated SC>2 air quality concentrations and exposures; and
• Assessment of the quality and limitations of the input data for supporting the goals of the
SO2 NAAQS exposure and risk characterization.
The overall flow of the exposure modeling process performed for this SC>2 NAAQS
review is illustrated in Figure 8-1. Several models were used in addition to APEX and
AERMOD including emission factors and meteorological processing models, as well as a
number of databases and literature sources to populate the model input parameters. Each of
these is described within this Chapter, supplemented with additional details in Appendix B.
8.2 OVERVIEW OF HUMAN EXPOSURE MODELING USING APEX
The EPA has developed the APEX model for estimating human population exposure to
criteria and air toxic pollutants. APEX serves as the human inhalation exposure model within
the Total Risk Integrated Methodology (TRIM) framework (EPA 2009a; 2009b). APEX was
recently used to estimate population exposures in 12 urban areas for the O3 NAAQS review
(EPA, 2007d; 2007e) and in estimating population NC>2 exposures in Atlanta as part of the NC>2
NAAQS review (EPA, 2008d).
APEX is a probabilistic model designed to account for sources of variability that affect
people's exposures. APEX simulates the movement of individuals through time and space and
estimates their exposure to a given pollutant in indoor, outdoor, and in-vehicle
microenvironments. The model stochastically generates a sample of simulated individuals using
census-derived probability distributions for demographic characteristics. The population
demographics are drawn from the year 2000 Census at the tract, block-group, or block-level, and
a national commuting database based on 2000 census data provides home-to-work commuting
flows. Any number of simulated individuals can be modeled, and collectively they approximate
a random sampling of people residing in a particular study area.
July 2009 185
-------
Upper Air
Meteorological
Data
(NOAA
Radiosonde
Database)
Surface
Meteorological
Data
(Integrated
Surface Hourly
Database)
L
J
AERMET/
AERSURFACE
'Major Stationary
Source
Emissions
(NE!)
Non-Point
Industrial and
Commercial/
Institutional
Emissions (NEI)
Onroad, Non-
road, and
Residual
Emissions (NEI)
Port Emissions
(NEI)
(Meteorological /
Land Surface
Data
AERMOD
SOj Outdoor Total
Hourly Concentration
Estimates
Model-to-
Monitor
Comparison
/
Human Aetvity
Patterns
(CHAD)
/
Census Block
Populations
(US Census)
^— __^
i
/
Home-toWork
Commuting
(US Census)
(
r
/ SO2 Outdoor
Total Hourly
Concentration
Estimates
/S-Minute Peak
/ to Houly-Mean
Ratio
Distributions
(Monitoring
Data)
/SO2 Indoor
/ Decay Rates
(Published
Studies and
Monte Carlo
Modeling)
y£ir Exchange
/ Rates /
Penetration
Factors
(Published
Studies)
SO2 Exposure
Concentratio
Estimates
Persons per year wit
5-minute SO2
exposures above
benchmark level
Person-days per yea
with 5-minute SO2
exposures above
benchmark levels
Figure 8-1. General process flow used for SO2 exposure assessment.
July 2009
186
-------
Daily activity patterns for individuals in a study area, an input to APEX, are obtained
from detailed diaries that are compiled in the Consolidated Human Activity Database (CHAD)
(McCurdy et al., 2000; EPA, 2002). The diaries are used to construct a sequence of activity
events for simulated individuals consistent with their demographic characteristics, day type, and
season of the year, as defined by ambient temperature regimes (Graham and McCurdy, 2004).
The time-location-activity diaries input to APEX contain information regarding an individuals'
age, gender, race, employment status, occupation, day-of-week, daily maximum hourly average
temperature, the location, start time, duration, and type of each activity performed. Much of this
information is used to best match the activity diary with the generated personal profile, using
age, gender, employment status, day of week, and temperature as first-order characteristics. The
approach is designed to capture the important attributes contributing to an individuals' behavior,
and of likely importance in this assessment (i.e., time spent outdoors) (Graham and McCurdy,
2004). Furthermore, these diary selection criteria give credence to the use of the variable data
that comprise CHAD (e.g., data collected were from different seasons, different states of origin,
etc.).
APEX has a flexible approach for modeling microenvironmental concentrations, where
the user can define the microenvironments to be modeled and their characteristics. Typical
indoor microenvironments include residences, schools, and offices. Outdoor microenvironments
include for example near roadways, at bus stops, and playgrounds. Inside cars, trucks, and mass
transit vehicles are microenvironments which are classified separately from indoors and
outdoors. APEX probabilistically calculates the concentration in the microenvironment
associated with each event in an individual's activity pattern and sums the event-specific
exposures within each hour to obtain a continuous series of hourly exposures spanning the time
period of interest. The estimated microenvironmental concentrations account for the
contribution of ambient (outdoor) pollutant concentration and influential factors such as the
penetration rate into indoor microenvironments, air exchange rates, decay/deposition rates,
proximity to important outdoor sources, and indoor source emissions. Each of these influential
factors are dependent on the microenvironment modeled, available data to define model inputs,
and estimation method selected by the model user. And, because the modeled individuals
represent a random sample of the population of interest, the distribution of modeled individual
exposures can be extrapolated to the larger population within the modeling domain.
July 2009 187
-------
The exposure modeling simulations can be summarized by five steps, each of which is
detailed in the subsequent sections of this document. Briefly, the five steps are as follows:
1. Characterize the study area. APEX selects the census blocks within a study
area - and thus identifies the potentially exposed population - based on user-
defined criteria and availability of air quality and meteorological data for the area.
2. Generate simulated individuals. APEX stochastically generates a sample of
hypothetical individuals based on the demographic data for the study area and
estimates anthropometric and physiological parameters for the simulated
individuals.
3 Construct a sequence of activity events APEX constructs an exposure event
sequence spanning the period of the simulation for each of the simulated
individuals using time-location-activity pattern data.
4. Calculate 5-minute and hourly concentrations in microenvironments. APEX
users define microenvironments that people in the study area would visit by
assigning location codes in the activity pattern to the user-specified
microenvironments. The model calculates all 5-minute concentrations occurring
within the hour (one maximum along with eleven other 5-minute values
normalized to the hourly mean) in each microenvironment for the period of
simulation, based on the user-provided microenvironment descriptions, the hourly
air quality data, and peak-to-mean ratios (PMRs; see section 7.2.3).
Microenvironmental concentrations are calculated independently for each of the
simulated individuals.
5. Estimate exposures. APEX estimates a concentration for each exposure event52
based on the microenvironment occupied during the event. In this assessment,
APEX estimated 5-minute exposures. These exposures can also be averaged by
clock hour to produce a sequence of hourly average exposures spanning the
specified exposure period. The values may be further aggregated to produce
daily, monthly, and annual average exposure values.
8.3 CHARACTERIZATION OF STUDY AREAS
8.3.1 Study Area Selection
The selection of areas to include in the exposure analysis takes into consideration the
availability of ambient monitoring, the presence of significant and diverse 862 emission sources,
population demographics, and results of the ambient air quality characterization. Although it
could be useful to characterize SC>2 exposures nationwide, because the exposure modeling
approach is both time and labor intensive, a regional and source-oriented approach was selected
52 An exposure event is a continuous period of time during which the factors that affect exposure (microenvironment
inhabited, activity performed, ventilation rate, and pollutant concentration) can be considered constant.
July 2009 188
-------
to make the analysis tractable and with the goal of focusing on areas most likely to have elevated
SC>2 peak concentrations and with sufficient data to conduct the analysis.
A broad study area was first identified based on the results of a preliminary screening of
the 5-minute ambient SC>2 monitoring data that were available. The state of Missouri was one of
only a few states reporting both 5-minute maximum and continuous 5-minute SC>2 ambient
monitoring data (14 total monitors), as well as having over thirty monitors in operation at some
time during the period from 1997 to 2007 that measured 1-hour SC>2 concentrations. In addition,
the air quality characterization described in Chapter 7 estimated frequent exceedances above the
potential health effect benchmark levels at several of the 1-hour ambient monitors within
Missouri. In a ranking of estimated SC>2 emissions reported in the National Emissions Inventory
(NEI), Missouri ranked 7th out of all U.S. states for the number of stacks with annual emissions
greater than 1,000 tons. These stack emissions were associated with a variety source types such
as electrical power generating units, chemical manufacturing, cement processing, smelters, and
emissions associated with port operations.
In the 1st draft SC>2 REA, several modeling domains were characterized within the
selected state of Missouri to assess the feasibility of the modeling methods. These modeling
domains were defined as areas within 20 km of a major point source of SC>2 emissions. While
modeled air quality and exposure results were generated for several of these domains in the 1st
draft REA, changes in the methodology used in this 2nd draft REA precluded additional analysis
for most of the domains originally selected. Staff judged the availability of relevant ambient
monitoring data within the model domain as essential in evaluating the dispersion model
performance, increasing confidence in the predicted air quality and exposure modeling results.
For example, when comparing the modeled air quality to ambient monitoring data in Greene
County in the 1st draft REA, it was judged by staff that non-point source emissions may
contribute to a large proportion of measured ambient concentrations. Addressing non-point
source emissions then added a layer to the already complex modeling performed, further limiting
the potential number of locations analyzed. Second, to assess the impact of potential alternative
standards, baseline conditions (as is air quality) need to be known, again requiring ambient
monitoring data. Because Greene County had a number of ambient monitors and most of the
model input data were already well-defined, it was selected for further modeling in the 2nd draft
REA. Additionally, staff decided that modeling a large urban area would be advantageous in
July 2009 189
-------
combining both large emission sources and large potentially exposed populations. Modeling for
St. Louis, Mo. was already underway at the time the 1st draft REA was completed, therefore it
was decided that exposure modeling in this domain should be continued and expanded for other
sources for the 2nd draft and the final REAs.
8.3.2 Study Area Descriptions
8.3.2.1 Greene County, Mo.
The greater Springfield, Mo., Metropolitan Statistical Area (MSA) consists of five
counties in southwestern Missouri including Christian, Dallas, Greene, Polk, and Webster
counties. The only city in the region with a population greater than 150,000 is Springfield, in
Greene County. Greene County has a total area of approximately 678 mi2 (1,756 km2). Due to
the complexity of the air quality and exposure modeling performed in this exposure assessment
and the focus on receptors within 20 km of stationary sources, the modeling domain was limited
to Greene County (see Figure 8-2). The Springfield-Branson Regional Airport (WBAN 13995)
served as the source of meteorological data used in the Greene County modeling domain.
8.3.2.2 St. Louis, Mo. Area
The greater St. Louis Metropolitan Statistical Area (MSA) is the 18th largest MSA in the
United States and includes the independent City of St. Louis; the Missouri counties of St. Louis,
St. Charles, Jefferson, Franklin, Lincoln, Warren, and Washington; as well as the Illinois
counties of Madison, St. Clair, Macoupin, Clinton, Monroe, Jersey, Bond, and Calhoun. The
total MSA has an area of approximately 8,846 mi2 (22,911 km2). Due to the complexity of the
air quality and exposure modeling performed in this exposure assessment and the focus on
receptors within 20 km of stationary sources, staff limited the modeling domain to three counties
directly surrounding the city of St. Louis: St. Louis City, St. Louis County, and St. Charles
County (see Figure 8-3). These three counties comprise much of the urban center of the St.
Louis MSA, with a combined population of about 1.15 million (2000 Census), which is
approximately 45 percent of the Greater St. Louis MSA population.
July 2009 190
-------
Lawrence County
; ' ":^'^^-''-^ " ^
/:.. :'^ :0 r^^Spf c:::
• •/ -*3W rt&/ •
0 3.5 7
I I I I I
14 Kilometers
_i I
ster County
qp Surface Meteorological Station (WBAN = 13995)
A Point Sources
Monitors (Monitor IDs)
^ 290770026
)( 290770032
^ 290770037
^ 290770040
J( 290770041
Census Block Receptors
Area Sources
Christian County
Figure 8-2. Modeling domain for Greene County Mo., along with identified emissions sources, air
quality receptors, ambient monitors, and meteorological station.
The St. Louis modeling domain defined in this REA was assembled from three separate
modeling domains described in the 1st draft SC>2 REA, aggregated to utilize the most reliable
hourly meteorological data available (St. Louis International-Lambert Field; WBAN 13994). It
was then reduced to just the three counties of the urban core described above. Figure 8-3 shows
the modeling domain for the greater St. Louis, MO area.
July 2009
191
-------
Surface Meteorological Station (WBAN = 13994)
In-State Point Sources
\ Cross-Border Point Sources
Monitors (Monitor IDs)
291890004
291890006
291893001
291895001
291897003
295100007
295100086
Port Area Sources
Census Block Receptors (Subset ID)
• 1
2
3
] Area Sources
10
20
40 Kilometers
Figure 8-3. Three county modeling domain for St. Louis, Mo., along with identified emissions sources, air quality receptors, ambient
monitors, and meteorological station.
July 2009
192
-------
8.3.3 Time Period of Analysis
Calendar year 2002 was simulated for both modeling domains to characterize the most
recent year of emissions data available for the study locations. Year 2002 temperature and
precipitation used in the dispersion modeling was compared with 30-year climate normal period
data from 1978 through 2007. For Greene County, 2002 temperatures were similar to the 30-
year normal (56.2 °F compared to 56.3 °F) though drier than the 30-year normal (37.8 in.
compared to 40.2 in.). For St. Louis, 2002 temperatures were warmer on average than the 30-
year normal (57.9 °F compared to 56.8 °F) and received an annual rainfall total that was similar
with the 30-year normal (40.9 in. compared to 39.1 in.). See Appendix B, Attachment 1 for
further details.
8.3.4 Populations Analyzed
The exposure assessment included the total population residing in each modeled area and
population subgroups that were considered more susceptible as identified in the ISA. These
population subgroups include:
• Asthmatic children (5-18 years in age)
• All Asthmatics (all ages)
In addition, based on the observed responses in the human clinical trials, all asthmatic
exposures were characterized only when the individual was at moderate or greater exertion levels
during the exposure events (see sections 8.5.5 and 8.8.2).
8.4 CHARACTERIZATION OF AMBIENT HOURLY AIR QUALITY DATA
USING AERMOD
8.4.1 Overview
Air quality data used for input to APEX were generated using AERMOD, a steady-state,
Gaussian plume model (EPA, 2004a). For both modeling domains, the following steps were
performed.
1 Collect and analyze general input parameters. Meteorological data, processing
methodologies used to derive input meteorological fields (e.g., temperature, wind
speed, precipitation), and information on surface characteristics and land use are
needed to help determine pollutant dispersion characteristics, atmospheric
stability and mixing heights.
July 2009 193
-------
2. Define sources and estimate emissions. The emission sources modeled included:
a. Major stationary emission sources within the domain,
b. Major stationary emission sources outside the domain (cross-border
stacks)
c. Non-point source area emissions,
d. Emissions from ports, and
e. Background sources not otherwise captured.
However, note that not all source categories were present in both modeling
domains.
3. Define air quality receptor locations. Two sets of receptors were identified for
the dispersion modeling, including ambient monitoring locations (where
available) and census block centroids.
4. Estimate concentrations at receptors. Full annual time series of hourly
concentration were estimated for 2002 by summing concentration contributions
from each of the emission sources at each of the defined air quality receptors.
Estimated hourly concentrations output from AERMOD were then used as input to the
APEX model to estimate population exposure concentrations. Details regarding both modeling
approaches and input data used are provided below. Supplemental information regarding model
inputs and methodology is provided in Appendix B.
8.4.2 General Model Inputs
8.4.2.1 Meteorological Inputs
All meteorological data used for the AERMOD dispersion model simulations were processed
with the AERMET meteorological preprocessor, version 06341. The National Weather Service
(NWS) served as the source of input meteorological data for AERMOD. Tables 8-1 and 8-2 list
the surface and upper air NWS stations chosen for the two areas. A potential concern related to
the use of NWS meteorological data is the often high incidence of calms and variable wind
conditions reported for the Automated Surface Observing Stations (ASOS) in use at most NWS
stations. A variable wind observation may include wind speeds up to 6 knots, but the wind
direction is reported as missing. The AERMOD model currently cannot simulate dispersion
under these conditions. To reduce the number of calms and missing winds in the surface data for
each of the four stations, archived one-minute winds for the ASOS stations were used to
calculate hourly average wind speed and directions, which were used to supplement the standard
archive of winds reported for each station in the Integrated Surface Hourly (ISH) database.
Details regarding this procedure are described in Appendix B, Attachment 1.
July 2009 194
-------
Table 8-1. Surface stations for the SO2 study areas.
Area
Greene
County
St. Louis
Station
Springfield-
Branson
Regional
AP
Lambert-St.
Louis
International
AP
Identifier
SGF
STL
WMO
(WBAN)
724400
(13995)
724340
(13994)
Latitude1
37.23528
38.7525
Longitude1
-93.40028
-90.37361
Elevation
(m)
387
161
Time
Zone2
6
6
Notes:
1 Latitude and longitude are the best approximation coordinates of the meteorological towers.
2 Time zone is the offset from UTC/GMT to LST in hours.
Table 8-2. Upper air stations for the SO2 study areas.
Area
Greene
County
St. Louis
Station
Springfield-
Branson
Regional
AP
Lincoln-
Logan
County AP,
IL
Identifier
SGF
ILX
WMO
(WBAN)
724400
(13995)
724340 (4833)
Latitude
37.23
40.15
Longitude
-93.40
-89.33
Elevation
(m)
394
178
Time
Zone1
6
6
Notes:
1 Time zone is the offset from UTC/GMT to LST in hours.
8.4.2.2 Surface Characteristics and Land Use Analysis
The AERSURFACE tool (US EPA, 2008e) was used to determine surface characteristics
(albedo, Bowen ratio, and surface roughness) for input to AERMET. Surface characteristics
were calculated for the location of the ASOS meteorological towers, approximated by using
aerial photos and the station history from the National Climatic Data Center (NCDC). A draft
version of AERSURFACE (08256) that utilizes 2001 National Land Cover Data (NLCD) was
used to determine the surface characteristics for this application since the 2001 land cover data
will be more representative of the meteorological data period than the 1992 NLCD data
supported by the current version of AERSURFACE available on EPA's SCRAM website. All
stations considered were located at an airport. Monthly seasonal assignments were defined as
shown in Table 8-3 and because the AERSURFACE default seasonal assignments were not used,
the surface characteristics were output by month. Note, the winter options can be winter (no
July 2009
195
-------
snow) or winter (continuous snow on ground).53 The exposure modeling domains experienced
less than 28.5 days per year of at least one inch (25.4 mm) of ground snow depth according to
CLIMAP contours,54 so no month was expected to have continuous snow on ground and hence
the designation of winter (no snow) only.
Table 8-3. Seasonal monthly assignments.
Station
SGF
STL
Winter (no snow)
December, January,
February, March
December, January,
February
Spring
April, May
March, April, May
Summer
June, July, August
June, July, August
Autumn
September,
October,
November
September,
October,
November
Seasonal definitions
Winter (no snow)
Spring
Summer
Autumn
Late autumn after frost and harvest, or winter with no snow
Transitional spring with partial green coverage or short annuals
Midsummer with lush vegetation
Autumn with unharvested cropland
8.4.3 Stationary Sources Emissions Preparation
8.4.3.1 Emission Sources and Locations
Point Sources
Point sources at major facilities were identified and paired to a representative surface
meteorological station. Any stacks listed as in the same location with identical release
parameters within a certain resolution (typically to the nearest integer value) were aggregated
into a single stack to simplify modeling but retain all emissions. For this analysis, major
facilities were defined as those with an SC>2 emission total exceeding 1,000 tpy in 2002. Within
such facilities, every stack emitting more than one tpy was included in the modeling inventory.
This process resulted in the identification of 11 (combined) stacks in Greene County and 38
(combined) stacks in St. Louis. Additionally, 45 (combined) stacks were identified across the
state border that could influence concentrations in St. Louis. These cross-border stacks were
modeled the same as the within-state stacks. The locations of all emitting stacks were corrected
based on GIS analysis. This was necessary because many stacks in the NEI are assigned the
The designation of winter (continuous snow) would tend to increase wintertime albedo and decrease wintertime
Bowen ratio and surface roughness for most land-use types compared to snow-free areas.
54 NCDC Climate Maps of the United States database (CLIMAPS). See http://cdo.ncdc.noaa.gov/cgi-
bin/climaps/climaps.pl.
July 2009
196
-------
same location, which often corresponds to a location in the facility - such as the front office -
rather than the actual stack locations. To correct for this, stack locations were reassigned
manually with the Microsoft® Live Maps® Virtual Earth® tool to visually match stacks from
the NEI database to their locations within the facilities using stack heights as a guide to stack
identification. All release heights and other stack parameters were taken from the values listed in
the NEI. Table B.3-1 (in Appendix B) lists all stacks in both domains.
Port-Related Sources
Only the St. Louis modeling domain has relevant port emissions. The Port of St. Louis is
one of the nation's largest inland river ports. Activity from this port was modeled as fourteen
area sources along the waterfront. All port-related emission sources were considered as non-
point area emissions with boundaries based on GIS analysis of aerial photographic images. A
release height of 5.0 m with a plume initial vertical standard deviation (
-------
though a series of sensitivity runs to characterize model performance at the ambient monitor
locations.
For the St. Louis domain, staff chose a slightly different approach to characterize non-
point emissions sources. During model-to-monitor comparisons, it became clear that the spatial
allocation of county-wide non-point emissions to tracts, based on SAFs, resulted in an inaccurate
spatial pattern of emissions. Therefore, the spatial resolution of non-point sources in this domain
was retained at the county level. However, to improve the numerical representation of these
emissions in the model, the two counties with the highest non-point source emissions - St. Louis
City and St. Louis County - were subdivided into regular grid cells. St. Louis County grid cells
were 5 km by 5 km; St. Louis City grid cells were 1 km by 1 km, more closely approximating the
smaller and denser census tracts in that region. All county-wide non-point source emissions
were spatially allocated uniformly to the grid cells. St. Charles County was modeled as a single
area source, with edges approximating the full county boundaries.
The release parameters for the St. Louis domain varied according to the urban and rural
designation of individual grid cells. Rural grid cells have a release height of 10 m and initial
dispersion length of 4.67 m. Urban grid cells have a release height of 20 m and initial dispersion
length of 9.34 m.
Background Sources
For the Greene County modeling domain, background sources were assembled to account
for any emissions not otherwise included. These were comprised of any point sources in
facilities not meeting the 1,000 tpy selection criteria and any residual non-point sources, as well
as on-road and non-road mobile sources. In addition, all emission sources in neighboring
Christian County were modeled as a rural, county-wide non-point area source with uniform
density. Both background sources were characterized as county-wide polygon rural area sources
with release heights of 10.0 m and initial dispersion length of 4.67 m.
For the St. Louis modeling domain, emissions from residual point sources, on-road
mobile sources, and non-road mobile sources were combined with the county-wide non-point
sources as described above. Thus, no separate background sources were simulated.
July 2009 198
-------
8.4.3.2 Urban vs. Rural Designations
This section describes how urban and rural designations were determined for each
emission source type. AERMOD has somewhat different treatment for urban and rural sources.
For example, when regulatory default settings are employed as they were in this application, no
chemical decay is assumed for rural sources, while a 4-hour half-life is assumed for urban
sources. Another difference in AERMOD's treatment of urban and rural sources is that for urban
sources, additional dispersion is simulated at night to account for increased surface heating
within an urban area under stable atmospheric conditions. The magnitude of this effect is weakly
proportional to the urban area population.
Point Sources
Urban or rural designations for point sources were made according to EPA guidance
based on the land use within 3 km of the source. The 2001 NLCD database was used to make
this determination. Table 8-4 lists the land use categories in the 2001 NLCD.
Table 8-4. NLCD2001 land use characterization.
Category
11
12
21
22
23
24
31
32
41
42
43
51
52
71
72
Land Use Type
Open Water
Perennial Ice/Snow
Developed, Open Space
Developed, Low Intensity
Developed, Medium Intensity
Developed, High Intensity
Barren Land (Rock/Sand/Clay)
Unconsolidated Shore1
Deciduous Forest
Evergreen Forest
Mixed Forest
Dwarf Scrub
Shrub/Scrub
Grassland/Herbaceous
Sedge/Herbaceous
Category
73
74
81
82
90
91
92
93
94
95
96
97
98
99
Land Use Type
Lichens
Moss
Pasture/Hay
Cultivated Crops
Woody Wetlands
Palustrine Forested Wetland1
Palustrine Scrub/Shrub Wetland1
Estuarine Forested Wetland1
Estuarine Scrub/Shrub Wetland1
Emergent Herbaceous Wetlands
Palustrine Emergent Wetland (Persistent) 1
Estuarine Emergent Wetland1
Palustrine Aquatic Bed1
Estuarine Aquatic Bed1
Notes:
1 Coastal NLCD class only.
July 2009
199
-------
Each stack where more than half the land use within 3 km fell into categories 21-24 were
designated as urban. These categories are consistent with those considered developed by
AERSURFACE.56
Non-Point Sources
Non-point area sources were defined as rural or urban using a similar methodology as
that for the point sources. As noted in the 2008 AERMOD Implementation Guide,57 in some
cases, a population density is more appropriate than a land use characterization. Therefore, non-
point area sources were evaluated from both a land use and population density perspective.
In Greene County, area sources were defined as corresponding to the census tract
boundaries. Each tract was then considered urban or rural by considering both the population
density and land use fraction from NLCD2001. If the population density was greater than 750
persons/km2 or the developed land use categories 22-24 throughout the tract was greater than 50
percent, the tract was designated as urban. In addition, if a tract was surrounded by urban tracts
it was designated as urban, since the emissions from such a tract would likely be subject to urban
dispersion conditions.
As explained above, for the St. Louis modeling domain, the counties with the greatest
non-point emissions - St. Louis City and St. Louis County - were subdivided into regular grid
cells, while St. Charles County was represented as a polygon area source with its political
boundaries. The urban or rural designation was then assigned to each based on population
density. St. Charles County and all but eleven of the 5 km grid cells in St. Louis County were
designated rural; the remaining cells in St. Louis County and all of St. Louis City were
designated urban.
Port-Related Sources
Only the St. Louis modeling domain has relevant port emissions. The fourteen port-
related non-point area sources described above were designated urban, given their location in the
urban core along the waterfront and their associated industrial activities.
56 AERSURFACE User's Guide, U.S. EPA, OAQPS, Research Triangle Park, NC, EPA-454/B-08-001, January
2008.
57 AERMOD IMPLEMENTATION GUIDE, AERMOD Implementation Workgroup, US EPA, OAQPS, Air Quality
Assessment Division, Research Triangle Park, NC, Revised January 9, 2008,
July 2009 200
-------
Background Sources
Background area sources for Greene County were classified with the same procedures as
for non-point area sources. Both Greene and Christian counties were designated rural.
8.4.3.3 Source Terrain Characterization
All corrected locations for the final list of major facility stacks in St. Louis and Greene
County domains were processed with a pre-release version of the AERMAP terrain
preprocessing tool. This version is functionally equivalent to the current release version of the
tool (version 08280). In particular, this updated version allows use of 1 arc-second terrain data
from the USGS Seamless Server58 which allows for more highly resolved values of the source
and receptor heights as well as the hill height scales.
Terrain height information for point sources was processed through AERMAP with input
data taken from the USGS server. For all area sources (non-point and background source types),
the outputs from AERMAP were modified. In these cases, rather than using a single point to
represent these large areas, the terrain height for each vertex of the area was estimated with
AERMAP. The terrain height for the entire source polygon was then characterized as the
average terrain height from all vertices.
8.4.3.4 Emissions Data Sources
Point Sources
Data for the parameterization of major facility point sources in the two modeling domains
comes primarily from three sources: the 2002 NEI (EPA, 2007f), Clean Air Markets Division
(CAMD) Unit Level Emissions Database (EPA, 2007g), and temporal emission profile
information contained in the EMS-HAP (version 3.0) emissions model.59 The NEI database
contains stack locations, emissions release parameters (i.e., height, diameter, exit temperature,
exit velocity), and annual SC>2 emissions. The CAMD database has information on hourly SC>2
emission rates for all the electric generating units in the US, where the units are the boilers or
equivalent, each of which can have multiple stacks.60 These two databases generally contain
58 http://seamless.usgs.gov/index.php
59 http ://www. epa. gov/ttn/chief/emch/proj ection/emshapS 0 .html
60 The CAMD database also contains hourly NO2 emission data for both electric generating units and other types of
industrial facilities. In the case of facilities for which CAMD has hourly NO2 data but not SO2 data, SO2 relative
July 2009 201
-------
complimentary information, and were first evaluated for matching facility data. However,
CAMD lacks 862 emissions data for facilities other than electric-generating units. To convert
annual total emissions data from the NEI into hourly temporal profiles required for AERMOD, a
three tiered prioritization was used, as follows.
1. CAMD hourly concentrations to create relative temporal profiles.
2. EMS-HAP seasonal and diurnal temporal profiles for source categorization codes
(SCCs).
3. Flat profiles, that is, a uniform emission rate throughout the day.
Details of these processes were as follows:
Tier 1: CAMD to NEI Emissions Alignment and Scaling
Of the 94 major facility stacks within the model domains identified above (11 in Greene
County and 45 cross-border and 38 within-state in the St. Louis domain), 35 (11 in Greene
County and 7 cross-border and 17 in-state in the St. Louis domain) were able to be matched
directly to sources within the CAMD database. Stack matching was based on the facility name,
Office of Regulatory Information Systems (ORIS) identification code (when provided) and
facility total SO2 emissions. For these stacks the relative hourly profiles were derived from the
hourly values in the CAMD database, and the annual emissions totals were taken from the NEI.
Hourly emissions in the CAMD database were scaled to match the NEI annual total emissions by
proportionally scaling each hour. Although the CAMD emissions may be more accurate than the
corresponding values in the NEI because they are based on direct emissions monitoring, because
CAMD emissions estimates were available for only a subset of sources, the NEI emission totals
were used so that the emission estimates would be consistent across all sources.
Tier 2: EMS-HAP to NEI Emissions Profiling
Of the 94 major facility stacks within the two MO domains, 38 stacks (all of which are
cross-border stacks in the St. Louis domain) could not be matched to a stack in the in the CAMD
database, but had SCC values that corresponded to SCCs that have temporal profiles included in
the EMS-HAP emissions model. In these cases, the SCC-specific seasonal and hourly variation
(SEASHR) values from the EMS-HAP model were used to characterize the temporal profiles of
emissions for each hour of a typical day by season and day type.
temporal profiles could be approximated by NO2 temporal profiles. However, there were no such cases for MO
facilities.
July 2009 202
-------
Tier 3: Other Emissions Profiling
Of the 94 major facility stacks within the two MO model domains, 21 (all from the St.
Louis in-state domain) could not be matched to a stack in CAMD database, or to profiles in the
EMS-HAP model by SCC code. In these cases, a flat profile of emissions was assumed. That is,
emissions were assumed to be constant for all hours of every day, but with an annual total that
equals the values from the NEI. A summary of the point source emissions used for the two
modeling domains is given in Table 8-5. Appendix B, Table B.3-1 contains all 94 stacks within
the modeling domains and the data source used to determine their emissions profiles.
Nearly all of the point sources in both domains were accounted for directly in the
dispersion modeling. Table 8-5 shows the point source contribution captured directly within each
modeling domain.
Port-Related Sources
Ports were the only non-road sector explicitly simulated in either modeling domain. Only
the St. Louis domain had port emissions. All relevant port emissions were directly captured,
comprising 51 percent of the total non-road emissions for the domain. Emission profiles for
port-related activity were taken from the EMS-HAP model for sectors matching the modeled
activity. Table 8-5 shows the port source contribution modeled directly within each modeling
domain and compares it to the total non-road emissions.
Non-Point and Background Sources
Non-point polygon area sources were developed to capture non-point
commercial/institutional and industrial emissions within the domains, as specified in the NEI.
For the St. Louis modeling domain, all non-point emissions were included either in gridded area
sources over St. Louis City and St. Louis County or a polygon area source over St. Charles
County, as described above. For the Greene County modeling domain, commercial/institutional
and industrial non-point area source polygons were created to represent the individual census
tracts within the county that captured approximately 87 percent of the relevant emissions
countywide from the NEI. Other non-point sources, as well as on-road mobile and non-road
mobile sources were included in the background source
Because non-point area source and background area source temporal profiles are
unknown, staff derived profiles that provided a best-fit match between the model predictions and
monitor data. To determine the most representative average non-point area source emission
July 2009 203
-------
profile across each modeling domain, we first selected monitors where ambient concentrations
were expected to be primarily influenced by area sources. Due to their locations relative to
sources, all but one monitor (ID 290770032) in Greene County indicated ambient concentrations
were primarily influenced by point source emissions. In St. Louis, all seven ambient monitors
(IDs 291890004, 291890006, 291893001, 291895001, 291897003, 295100007, and 295100086)
indicated significant influence from area source emissions. Next, simulations were conducted
with all sources modeled in detail - except area sources, which were modeled with uniform
emission profiles. A weighting function was then determined based on the modeled error for
each hour of the day at the one Greene County monitor and as an average of the errors at the
seven individual St. Louis area monitors. In both cases, the error function was defined as the
ratio of the total observed concentration, minus the total concentration due to all non-point
sources, to the concentration predicted by the non-point sources alone. This diurnal error
function was then normalized such that its average value is unity. Finally, a corrected non-point
emission profile was determined by combining this normalized weighting function with the
uniform emission profile.
Figures 8-4 and 8-5 show the diurnal emissions profiles derived for both the St. Louis and
Greene County domains compared to other profiles for industrial and commercial/institutional
area sources derived from commonly used emissions models, such as SMOKE and EMS-HAP.
The shape of the derived temporal profiles imply that the emission sources are active almost
exclusively during the daytime from approximately 8 am to 8pm, in contrast to those derived
from SMOKE and EMS-HAP, which show less extreme daytime-dominated patterns. Given the
large uncertainties about the actual emission sources represented by the industrial and
commercial/institutional non-point category and given that such sources are likely to be small
facilities, it is reasonable to assume that their cumulative emissions occur almost exclusively
during daytime hours. Table 8-5 shows the non-point source contribution modeled directly
within each modeling domain and compares it to the total non-point emissions.61
61 Table 8-5 does not have the relevant background contribution for each domain. This is because the total
background in each domain includes not only the counties in the modeling domain (three in the St. Louis domain
and one in the Greene County domain), but also adjacent counties that could influence concentrations within the
modeling domain. In those cases, the total countywide emissions are included in the background. Thus, directly
expressing those values would be confusing and are thus omitted.
July 2009 204
-------
Table 8-5. Summary of NEI emission estimates and total emissions used for dispersion modeling
in Greene County and St. Louis modeling domains.
Modeling
Domain
Greene Co.
St. Louis
Point Sources
NEI
Emissions
(tpy)
9,255
70,016
Modeled
Emissions
(tpy)
9,047
68,656
(%)
98%
98%
Area Sources
NEI
Emissions
(tpy)
2,055
15,137
Modeled
Emissions
(tpy)
1,781
15,137
(%)
87%
100
%
Non-road Sources
NEI
Emissions
(tpy)
N/A
3,058
Modeled
Emissions
(tpy)
N/A
1,559
(%)
N/A
51%
I/)
I/)
c
£
'tn
y>
HI
1
TO
-»- EMS-HAP: Industrial
-•-SMOKE
- -*- EMS-HAP: Commercial/Institutional
Monitors
12
Hour of the Day
18
24
Figure 8-4. Derived best-fit non-point area source diurnal emission profile for the St. Louis
domain, compared to other possible profiles.
July 2009
205
-------
0.12
0.1
i/)
i/)
o>
--•-- EMS-HAP: Industrial
-•--SMOKE
-A- - EMS-HAP: Commercial/Institutional
^^~ Best-fit to Monitors
I 0.08
12
Hour of the Day
18
24
Figure 8-5. Derived best-fit non-point area source diurnal emission profile for the Greene County
domain, compared to other possible profiles.
8.4.4 Receptor Locations
Two sets of receptors were chosen to represent the locations of interest within each of the
modeling domains. The first set was selected to represent the locations of the residential
population of the modeling domain. These receptors were US Census block centroids in the
Greene County and St. Louis modeling domains, (Figures 8-2 and 8-3, respectively), that lie
within 20 km (12 miles) of any of the major facility stacks.62 Each of these receptors was
modeled at ground level. A total of 17,703 receptors were selected in the St. Louis modeling
domain and a total of 5,359 receptors were selected in the Greene County modeling domain.
The second set of receptors included the locations of the available ambient 862 monitors.
These receptors were used in evaluating the dispersion model performance. In Greene County,
there were five ambient monitors with valid ambient monitoring concentrations (Figure 8-2).
Within the three St. Louis counties, there were seven monitors (Figure 8-3).
\)
The block centroids used for this analysis are actually population-weighted locations reported in the ESRI
database. They were derived from geocoded addresses within the block taken from the Acxiom Corporation
InfoBase household database (Skuta and Wombold, 2008; ESRI, 2008). These centroids differ from the "internal
points" reported by the US Census, which are often referred to as centroids because they are designed to represent
the approximate geographic center of the block.
July 2009
206
-------
8.4.5 Modeled Air Quality Evaluation
The hourly 862 concentrations estimated from each of the sources within a modeling
domain were combined at each receptor. These concentration predictions were then compared
with the measured concentrations at ambient SO2 monitors. Rather than compare concentrations
estimated at a single modeled receptor point to the ambient monitor concentrations, a distribution
of concentrations was developed for the predicted concentrations for all receptors within a 4 km
distance of the monitors. Further, instead of a comparison of central tendency values (mean or
median), the full modeled and measurement concentration distributions were used for
comparison.
As an initial comparison of modeled versus measured air quality, all modeled receptors
within 4 km of each ambient monitor location were used to generate a prediction envelope.63
This envelope was constructed based on selected percentiles from the modeled concentration
distribution at each receptor for comparison to the ambient monitor concentration distribution.
The 2.5th and 97.5th percentiles from all monitor distribution percentiles64 were selected to create
the lower and upper bounds of the envelope. The full 1-hour distributions for the ambient
measurement data, the modeled monitor receptor,65 and the prediction envelope were compared
using their respective cumulative density functions (CDFs). When illustrating these
distributions, the percentiles were plotted on a log-scale as the difference between 100 and the
CDF value to allow for visual expansion of the extreme upper percentiles of the distribution. For
illustrative purposes, the maximum concentration was defined as 100-99.99 (or 0.01) because the
logarithm of zero is undefined.
A second comparison between the modeled and monitored data was performed to
evaluate the diurnal variation in SC>2 concentrations. AERMOD receptor concentrations during
each hour-of-the-day were averaged (i.e., 365 values for hour 1, 365 values for hour 2, and so
on) to generate an annual average SC>2 concentration for each hour at each modeled receptor.
Prediction envelopes were constructed similar to that described above from modeled receptors
63 500 m to 4 km is the area of representation of a neighborhood-scale monitor, according to EPA guidance.
64 As an example, suppose there are 1,000 receptors surrounding a monitor, each receptor containing 8,760 hourly
values used to create a concentration distribution. Then say the 73rd percentile concentration prediction is to be
estimated for each receptor. The lower bound of the 73rd percentile of the modeled receptors would represented by
the 2.5th percentile of all the calculated 73rd percentile concentration predictions, i.e., the 25th highest 73rd percentile
concentration prediction across the 1,000 73rd percentile values generated from all of the receptors. Note that at any
given percentile along either of the envelope bounds as well as at the central tendency distribution (the receptor 50th
percentile), the concentration from a different receptor may be used.
65 The modeled monitor is the modeled air quality at the ambient monitoring location.
July 2009 207
-------
located within 4 km of each ambient monitor. The measured ambient monitoring data was also
averaged to generate the diurnal profile. Then, annual averaged concentrations for the ambient
measurement data, the modeled monitor receptor, and the prediction envelope were plotted by
hour-of-the-day for comparison.
Staff also evaluated potential impact of the differences between the predicted and
measured 1-hour SO2 concentrations by comparing the modeled and measured number of 5-
minute air quality benchmark exceedances that would result from using each 1-hour
concentration distribution. The full year of 1-hour ambient monitored and AERMOD modeled
SO2 concentrations (at the monitor receptor location) were used as input to the 5-minute
statistical model and processed as described in section 7.2.5. Measured 5-minute maximum 862
concentrations were only available for two of the monitors in Greene County (290770026 and
290770040). These monitoring locations were used to generate the number of days per year with
at least one benchmark exceedance. Further, the concentration distributions given by the
AERMOD prediction envelopes (i.e., the 2.5th and 97.5th) were used to approximate lower and
upper prediction bounds for the number of days per year with 5-minute benchmark exceedances.
To do this, first the total numbers of benchmark exceedances in a year66 were estimated for each
monitor using the 1-hour concentration percentiles representing each AERMOD distribution
(i.e., the AERMOD monitor receptor, the AERMOD 2.5th, and the AERMOD 97.5th). Then,
scaling factors were calculated by dividing each the AERMOD 2.5th and AERMOD 97.5th
benchmark exceedance results by that of the exceedances estimated using the AERMOD monitor
receptor. These scaling factors were then applied to the full AERMOD monitor receptor
predictions that estimated the number of days per year with exceedances to estimate the lower
and upper bounds.
8.4.5.1 Greene County Modeled Air Quality Evaluation
For Greene County, there were five monitors used for comparison with the AERMOD 1-
hour concentration estimates. For each monitor, staff plotted the model-predicted versus
ambient measured concentrations using two methods; the first used a CDF, the second used the
66 Because the AERMOD p2.5 and p97.5 prediction envelopes are not representing a particular time but are a
temporal and spatial mixture of low and high concentrations surrounding each monitor, specific counts of days per
year could not be calculated. Staff assumed a proportional relationship existed between the total number of
exceedances in a year and the number of days per year with exceedances. Thus, scaling factors can be calculated
using the AERMOD monitor receptor data, which had both the percentile form and 8,760 concentrations at specific
hours of the day and days of the year.
July 2009 208
-------
diurnal profile. In each plot, four concentration distributions were used; the distribution of the
modeled 1-hour 862 concentrations estimated for the monitor receptor, the upper and lower
bounds of the receptor envelope (i.e., generated from all receptors within 4 km of monitor
receptor), and the hourly concentration distribution measured at each ambient monitor. The
results for Greene County are provided in Figures 8-6 to 8-8. The data used to generate the
figures are provided in Appendix B.
When considering the total hourly distribution or CDFs, monitor concentration
distributions are generally bounded by the modeled distributions. At some of the upper
percentiles of the distributions, the deviations were of varying direction (over- or under-
prediction) and magnitude (a few ppb to tens of ppb). For example, monitor ID 290770026
(Figure 8-6) exhibits higher measured concentrations at the upper percentiles of the distribution
that extend beyond the AERMOD prediction envelope, however the deviation occurred beyond
the 99.5th percentile (maximum observed =114 ppb, AERMOD 97.5th = 101 ppb). At monitor ID
290770032 (Figure 8-6), the measured concentrations fall below the prediction envelope,
beginning just beyond the 95th percentile 1-hour concentration.
Even though ambient monitors 290770040 and 290770041 (Figure 8-2) are located
approximately 150 m from one another, they exhibited very different measured concentrations at
the extreme upper percentiles (Figure 8-7). The greatest difference is in comparing the
maximum observed concentrations; 203 ppb versus 33 ppb. The AERMOD predictions followed
a similar pattern at the upper percentiles, i.e., the modeled concentrations for the monitor
location were greater (50 to 100%) at monitor ID 290770040 when compared with 290770041,
but not nearly as great a difference noted at the maximum measured concentrations. The
AERMOD prediction envelope was similar for both of these monitors, encompassing the
ambient measured concentrations from the 80th through the 99.5th percentiles for both, while
completely enveloping all 1-hour concentrations at monitor ID 290770041.
The pattern in the AERMOD modeled concentrations at the monitor location and the
ambient measurement concentration distribution for monitor ID 290770037 is nearly identical.
The only difference observed is that the measured concentrations are 1-3 ppb greater than the
modeled concentrations within the 99th percentile of the distribution. Much of the measured
distribution falls within the AERMOD prediction envelope, with deviation occurring just beyond
the 99.5th percentile.
July 2009 209
-------
The diurnal pattern observed at each of the ambient monitors is represented well by the
modeled concentrations; in general concentrations are elevated during the midday hours and
lowest during the late-night and early-morning hours. In addition, most of the measured
concentrations fall within the AERMOD prediction envelopes at all hours of the day, with a few
exceptions. For example, all observed concentrations for monitor ID 290770032 are below that
of the upper AERMOD prediction envelope, though at monitor ID 290770026, measured
concentrations are above those modeled during the early-morning and late-night hours (Figure 8-
6). Much of the deviation during these hours-of-the-day is likely a result of the concentrations at
or below the 80th percentile, where measured concentrations were always greater than any of the
predicted concentrations at corresponding percentiles of the distribution. While the prediction
envelopes encompassed the diurnal pattern observed at monitor IDs 290770040 and 290770041
(Figure 8-7), the results for the modeled concentrations at the monitor locations were not equally
representative. The diurnal pattern and magnitude of concentrations was well reproduced at
monitor ID 290770041, while modeled concentrations at the monitor location during the midday
and evening hours were greater than the measured concentrations at monitor ID 290770040.
Staff evaluated the potential impact the predicted 1-hour concentrations would have on 5-
minute air quality benchmark exceedances (Table 8-6). In general, the results for the estimated
numbers of days per year with 5-minute concentrations above benchmark levels followed similar
patterns to those observed above when considering comparisons of the 1-hour 862 concentration
distributions. The numbers of benchmark exceedances at monitor ID 290770026 were under-
predicted by AERMOD just as was the 1-hour SO2 concentrations at that monitoring location.
However, the number of days with 5-minute concentrations above the benchmark levels for both
the measured and modeled ambient concentrations fell within the range of the AERMOD
prediction envelopes. There was good agreement in the number of days per year with air quality
benchmark exceedances at each of the four other monitors, whether there were none, a few, or
several days with expected benchmark exceedances. These results indicate that the magnitude of
observed differences in predicted versus measured 1-hour SO2 concentration does not result in
unexpected differences in the number of days per year having 5-minute SO2 concentrations
above the benchmark levels.
July 2009 210
-------
Table 8-6. Measured and modeled number of days in year 2002 with at least one 5-minute SO2
benchmark exceedance at ambient monitors in Greene County.
Monitor ID
290770026
290770032
290770037
290770040
290770041
5-minute SO2
Benchmark
(ppb)
100
200
300
400
100
200
300
400
100
200
300
400
100
200
300
400
100
200
300
400
Number of Days per Year with a 5-minute SO2
Concentration Above Air Quality Benchmark Level
Ambient Monitor1
Modeled
57
18
6
2
0
0
0
0
33
14
7
4
7
3
1
0
0
0
0
0
Measured
27
0
0
0
-
-
-
-
44
12
1
0
-
-
-
-
-
-
-
-
AERMOD2
p2.5
2
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
Monitor
19
2
0
0
0
0
0
0
40
13
5
2
25
3
0
0
2
0
0
0
p97.5
103
9
2
0
0
0
0
0
81
22
6
3
42
5
0
0
17
0
0
0
Notes:
1 The modeled numbers of 5-minute benchmark exceedances were generated from 1-hour
SO2 ambient monitor measurements input to the 5-minute statistical model. The measured
numbers of 5-minute benchmark exceedances were calculated from ambient monitors
reporting 5-minute SO2 concentrations. Both of these values were normalized to a full year
(n=365 days) for comparison with the AERMOD predictions.
AERMOD monitor 5-minute benchmark exceedances were generated from 1-hour SO2
ambient predictions (at monitor receptor location) input to the 5-minute statistical model.
AERMOD p2.5 and p97.5 benchmark exceedances were generated from the
corresponding hourly prediction envelope distribution and input to the 5-minute statistical
model.
July 2009
211
-------
Monitor ID 290770026
Monitor ID 290770026
100
Ambient Monitor
- - AERMOD P2.5
AERMOD P97.5
AERMOD Monitor
§
0.01
100
0 10
30 40 50 60 70 80 90
1 -hour SO2 Concentration (ppb)
Monitor ID 290770032
100 110 120
Ambient Monitor
AERMOD P2.5
AERMOD P97.5
AERMOD Monitor
0.01
10 20 30 40 50
1-hour SO2 Concentration (ppb)
I6
1 5
o
AERMOD P2.5
•AERMOD P97.5
-Ambient Monitor
-AERMOD Monitor
18
24
Hour of the Day
Monitor ID 290770032
4 --
3 ---
$
c
a
2 --
AERMOD P2.5
•AERMOD P97.5
-Ambient Monitor
-AERMOD Monitor
* . •
12
Hour of the Day
18
24
Figure 8-6. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitors 290770026 and 29077032 in Greene County, Mo. Maximum 1-hour concentration
percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
212
-------
Monitor ID 290770040
Monitor ID 290770040
100
—Ambient Monitor
- AERMODP2.5
• 'AERMODP97.5
—AERMOD Monitor
0.01
100
40
60 80 100 120 140 160
1-hour SO2 Concentration (ppb)
Monitor ID 290770041
180 200
Ambient Monitor
- - AERMOD P2.5
— 'AERMOD P97.5
AERMOD Monitor
0.01
20
30 40 50 60 70 80
1-hour SO2 Concentration (ppb)
90 100 110
6 +
3 44-
- - AERMOD P2.5
— 'AERMODP97.5
Ambient Monitor
AERMOD Monitor
12
Hour of the Day
Monitor ID 290770041
24
I 6
•s
AERMOD P2.5
•AERMOD P97.5
-Ambient Monitor
-AERMOD Monitor
\
18
24
Hour of the Day
Figure 8-7. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitors 290770040 and 29077041 in Greene County, Mo. Maximum 1-hour
concentration percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
213
-------
Monitor ID 290770037
Monitor ID 290770037
100
-Ambient Monitor
AERMOD P2.5
'AERMODP97.5
-AERMOD Monitor
0.01
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
1-hour SO2 Concentration (ppb)
.26-
5
u
AERMOD P2.5
• AERMOD P97.5
•Ambient Monitor
-AERMOD Monitor
V
\
V
12
Hour of the Day
24
Figure 8-8. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitor 290770037 in Greene County, Mo. Maximum 1 -hour concentration percentile is
defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
214
-------
8.4.5.2 St. Louis Modeled Air Quality Evaluation
For St. Louis, there were seven monitors used for comparison with the AERMOD
concentration estimates. The distribution of the modeled 1-hour 862 concentrations estimated
for the monitor receptor, the receptor envelope (i.e., all receptors within 4 km of monitor
receptor), and the hourly concentration distribution measured at each ambient monitor are
provided in Figures 8-9 to 8-12. Data used to generate the figures is provided in Appendix B.
There are distinct differences in the comparison of modeled versus measured
concentration distributions at ambient monitoring locations in St. Louis when compared with
Greene County. Most noticeable is the width of the prediction envelopes; St. Louis prediction
envelopes were not as wide as those generated for Greene County. This indicates that, in
comparison with the Greene County modeling domain, there is less spatial variability in the
concentrations modeled at receptors surrounding the ambient monitoring locations in St. Louis.
This is likely a result of the emission source contributions; four of five ambient monitors in
Greene County were primarily influenced by point sources, while most of the concentration
contribution for St. Louis monitors was from area source emissions.
The modeled concentrations at the monitor locations and ambient measured concentration
distributions showed better overall agreement at the St. Louis monitors, though many of the
measured concentrations are outside of the prediction envelopes. For example, at monitor ID
291890006 all measured concentrations up to the 99th percentile fell below the prediction
envelope (Figure 8-9) (the maximum was within). Note however that the difference in the
measured concentrations was only about 1 ppb when compared with concentrations at any of the
envelope percentiles and at most 2 ppb when compared with the modeled concentrations at the
monitor receptor. In addition, because most of these under-predictions occur at concentrations
well below levels of interest, it is not of great consequence. At the upper percentiles, many of
the ambient concentrations fell within the prediction envelopes; 6 of 7 monitors at the maximum
percentile were within, 3 of 7 monitors at the 99th percentile were within, and 4 of 7 monitors at
the 95th percentile were within the prediction envelopes. Where measured upper percentile
concentrations were outside of the prediction envelopes, it was consistently beneath the 2.5th
prediction, possibly indicating AERMOD over-prediction at these monitors at certain percentiles
of the distribution. When comparing the AERMOD monitor concentrations with the measured
ambient concentrations between the 80th and 99th percentile of the distribution, most of the
July 2009 215
-------
predicted values were greater than the measured concentrations. The magnitude of this over-
prediction ranged from about 1 to 2 ppb, although one monitor had a 7 ppb difference at the 99th
percentile. Predictions at the maximum concentrations were more balanced; 4 of the 7 monitors
had over-predictions, while all predictions (under or over) were approximately within 10 to 35
ppb of the measured concentrations.
The diurnal pattern was reproduced at the St. Louis monitoring locations, with some of
the prediction envelopes encompassing much of the measured ambient concentrations (e.g.,
Figure 8-9, monitor ID 291890004; Figure 8-11 monitor ID 291897003). Again where deviation
did occur at a few of the monitors, the contribution of the lower concentrations (i.e., mostly those
beneath the 90th percentile) likely played a role in the magnitude of the disagreement. This can
be seen at monitor ID 291890006 (Figure 8-10) where most (99%) of the predicted
concentrations are consistently above the measure concentrations by 1 to 2 ppb. It is not
surprising to see that the difference in comparing the measured versus modeled diurnal profile at
every hour-of-the-day is also between 1 to 2 ppb.
July 2009 216
-------
Monitor ID 291890004
Monitor ID 291890004
100
Ambient Monitor
AERMOD P2.5
AERMOD P97.5
AERMOD Monitor
0.01
100
0.01
c
o
- - AERMOD P2.5
— • AERMOD P97.5
Ambient Monitor
AERMOD Monitor
3--V
1 --
20
30 40 50 60 70 80
1- hour SO2 Concentration (ppb)
Monitor ID 291890006
90 100 110
12
Hour of the Day
Monitor ID 291890006
18
24
Ambient Monitor
- - AERMOD P2.5
— 'AERMOD P97.5
AERMOD Monitor
- - AERMOD P2.5
— • AERMOD P97.5
Ambient Monitor
AERMOD Monitor
30 40 50 60 70
1- hour SO2 Concentration (ppb)
12
Hour of the Day
18
24
Figure 8-9. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitors 291890004 and 291890006 in St Louis, Mo. Maximum 1-hour concentration
percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
217
-------
Monitor ID 291893001
Monitor ID 291893001
100
Ambient Monitor
- - AERMODP2.5
— 'AERMODP97.5
AERMOD Monitor
0.01
100
0.01
Q.
Q.
1 --
- - AERMOD P2.5
— • AERMOD P97.5
Ambient Monitor
AERMOD Monitor
X
20
30 40 50 60 70
1- hour SO2 Concentration (ppb)
Monitor ID 291895001
80
90
100
12
Hour of the Day
Monitor ID 291895001
18
24
- Ambient Monitor
- - AERMOD P2. 5
— AERMOD P97. 5
- AERMOD Monitor
5 ---
a>
2 4
3 ---
re
« 2
1 --
- - AERMOD P2.5
— • AERMOD P97.5
Ambient Monitor
AERMOD Monitor
40
60 80 100 120 140
1- hour SO2 Concentration (ppb)
160 180 200
12
Hour of the Day
18
24
Figure 8-10. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitors 291893001 and 291895001 in St Louis, Mo. Maximum 1-hour concentration
percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
218
-------
Monitor ID 291897003
Monitor ID 291897003
100
0.01
100
0.01
Ambient Monitor
AERMOD P2.5
AERMOD P97.5
•AERMOD Monitor
20 30 40 50 60 70 80 90 100 110 120 130 140
1- hour SO2 Concentration (ppb)
Monitor ID 295100007
Ambient Monitor
AERMOD P2.5
AERMOD P97.5
•AERMOD Monitor
20
30 40 50 60 70 80 90 100 110 120 130 140
1- hour SO2 Concentration (ppb)
1 --
- - AERMOD P2.5
— • AERMOD P97.5
Ambient Monitor
AERMOD Monitor
12
Hour of the Day
Monitor ID 295100007
18
24
.0
IS
u
c
o
o
d1
W
c
re
10
9
8
7
6
5
4
3
2
1
0
AERMOD P2.5
•AERMOD P97.5
•Ambient Monitor
-AERMOD Monitor
12
Hour of the Day
18
24
Figure 8-11. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitors 291897003 and 295100007 in St Louis, Mo. Maximum 1-hour concentration
percentile is defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
219
-------
Monitor ID 295100086
Monitor ID 295100086
100
Ambient Monitor
AERMOD P2.5
AERMOD P97.5
AERMOD Monitor
0.01
Q.
Q.
C
O
«
I
u
o
o
d'
CO
20
30 40 50 60 70 80 90 100 110 120 130
1- hour SO2 Concentration (ppb)
AERMOD P2.5
•AERMOD P97.5
-Ambient Monitor
-AERMOD Monitor
12
Hour of the Day
18
24
Figure 8-12. Comparison of measured ambient monitor SO2 concentration distribution and diurnal profile with the modeled monitor
receptor and receptors within 4 km of monitor 295100086 in St Louis, Mo. Maximum 1-hour concentration percentile is
defined as 0.01 (or 100-99.99) because log(O) is undefined.
July 2009
220
-------
8.4.4.3. Using unadjusted AERMOD predicted SO2 concentrations
The SC>2 concentrations estimated using AERMOD do not have a particular directional
influence in over- or under-estimating concentrations, save for small over-estimation primarily
observed at the lowest concentrations and some difficulty in reproducing some of the maximum
measured concentrations. Most ambient monitoring concentrations fell within the modeled
prediction envelopes constructed of modeled receptors surrounding the monitor. In generating
the modeled air quality, staff made judgments in appropriately modifying model inputs including
an adjustment of the area source temporal emission profile to improve the comparison of the
model predictions with the measurement data. Staff went through several iterations of evaluating
the model performance in each modeling domain following model input adjustments to obtain
the current modeled air quality results. Given the time and resources to perform this assessment,
the good agreement in the model-to-monitor comparisons, the degree of confidence in the
dispersion modeling system, the spatial representation of the monitors compared with receptors
modeled, and the number of comparisons available, staff did not perform any further adjustments
to the modeled concentrations to improve the relationship between modeled versus measured
concentration at each monitor. Additional details on the staffs reasoning are provided in section
8.11.
8.5 SIMULATED POPULATION
APEX takes population characteristics into account to develop accurate representations of
study area demographics. Specifically, age- and gender-specific population counts and
employment probability estimates, asthma prevalence rates, and home-to-work commuting
locations and probabilities were used to develop representative profiles of hypothetical
individuals used in the exposure modeling simulation. In addition, body surface area (BSA) and
activity-specific ventilation rates are two important attributes used by APEX to characterize
when simulated individuals were at moderate or greater activity levels. Each of these is
discussed in the following sections.
8.5.1 Population Counts and Employment Probabilities
Block-level population counts were obtained from the 2000 Census of Population and
Housing Summary File 1 (SF-1). Estimates of employment were also developed form census
information (US Census Bureau, 2007) and separated into gender and age groups. Children
July 2009 221
-------
under 16 years of age were assumed to be not employed. Staff also assumed that employment
probabilities for a census tract apply uniformly to the constituent census blocks. Further details
are provided in Appendix B.2.2.2.
8.5.2 Asthma Prevalence
The population subgroups included in this exposure assessment are asthmatics and
asthmatic children. Evaluating exposures of these subgroups with APEX requires the estimation
of children's asthma prevalence rates. The proportion of the population of children characterized
as being asthmatic was estimated by statistics on asthma prevalence rates recently used in the
NAAQS review for O3 (US EPA, 2007d). See Appendix B, Attachment 2 for details on the
derivation. Specifically, an analysis of data provided in the National Health Interview Survey
(NHIS) for 2003 (CDC, 2007) generated age and gender specific asthma prevalence rates for
children ages 0-17. Staff used these data rather than the aggregate data available at the county
level, to retain the variability in asthma prevalence observed with children of different ages.
Adult asthma prevalence rates were estimated by gender and for each particular modeling
domain based on Missouri regional data (MO DOH, 2002). Table 8-7 provides a summary of the
asthma prevalence used in the exposure analysis, stratified by age and gender.
The total population simulated within the two modeling domains was approximately 1.4
million persons, of which there was a total simulated population of about 130,000 asthmatics.
The model simulated over 360,000 children ages 5 through 17, of which there were nearly
50,000 asthmatics. The individual populations for each modeling domain and subpopulation of
interest are provided in Table 8-8. For comparison, staff weighted the asthma prevalence by
population in the three counties reported by the MO Department of Health (2003) for all ages
(i.e., St. Charles-8.8%, St. Louis-5.8%, and St. Louis City-16.4%) to generate an asthma
prevalence of 8.8%. This asthma prevalence is similar to the 9.2% modeled here using APEX.
In Greene County, the reported asthma prevalence was 10.2% (MO Department of Health,
2003), while 9.8% of the simulated population was asthmatic.
July 2009 222
-------
Table 8-7. Asthma prevalence rates by age and gender used in Greene County and St. Louis
modeling domains.
Modeling
Domain
(Region)
Greene Co.
and St. Louis
(Midwest)
Greene Co.
St. Louis
Age1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
>17
>17
Asthma Prevalence (%)
Females
7.0
7.1
7.3
7.5
8.1
9.5
9.2
9.0
8.6
11.0
16.2
19.6
21.2
17.0
14.0
13.3
14.0
16.5
10.7
9.3
Males
3.1
6.3
10.8
15.8
21.6
17.8
12.8
12.1
12.8
14.7
17.7
19.0
19.5
16.9
16.8
18.0
20.1
23.7
6.1
5.3
Notes:
1 Ages 0-17 from the National Health Interview Survey
(NHIS) for 2003 (CDC, 2007); ages >17 from (MO DOH,
2002).
Table 8-8. Population modeled in Greene County and St. Louis modeling domains.
Modeling
Domain
Green Co.
St. Louis
Population
All Ages
224,145
1,151,094
Children (5 -18)
54,373
308,939
Asthmatic Population
All Ages
21,948
105,456
Children (5 -18)
7,285
41,714
8.5.3 Commuting Database
Commuting data were originally derived from the 2000 Census, collected as part of the
Census Transportation Planning Package (CTPP) (US DOT, 2007). The data used here contain
counts of individuals commuting from home-to-work locations at a number of geographic scales.
These data were processed to calculate fractions for tract-to-tract flow on a national level (all 50
U.S. states and Washington, D.C.). A software pre-processor was then developed to generate
July 2009
223
-------
block-level commuting files for APEX using the tract-level commuting data and finely-resolved
land use data, assuming the frequency of commuting to a workplace block within a tract is
proportional to the amount of commercial and industrial land in the block. Further details are
provided in Appendix B.2.2.2.
Note that while travel on roads was accounted for by APEX for other individuals (e.g.,
unemployed, children, persons who work at home) it was assumed that the vehicle travel (e.g.,
car, bus, train) occurred within the block the individual resides.
8.5.4 Body Surface Area
Age- and gender-specific BSA is estimated for each simulated individual. Briefly, the
BSA calculation is based on logarithmic relationships developed by Burmaster (1998) that use
body mass (BM) as an independent variable as follows:
e-2'2m 5M06821 equation (8-1)
where,
BSA = body surface area (m2)
BM = body mass (kg)
Each simulated individual's body mass was randomly sampled from age- and gender-
specific body mass distributions generated from National Health and Nutrition Examination
Survey (NHANES) data for the years 1999-2004,67 Details in their development and the
parameter values are provided in Appendix B, Attachment 3.
8.5.5 Activity-Specific Ventilation Rates
Ventilation is a general term describing the movement of air into and out of the lungs.
The rate of ventilation is determined by the type of activity an individual performs which in turn
is related to the amount of oxygen required to perform the activity. Minute or total ventilation
rate is used to describe the volume of air moved in or out of the lungs per minute.
67 Demographic (Demo) and Body Measurement (BMX) datasets for each of the NHANES studies were obtained
fromhttp://www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm.
July 2009 224
-------
Quantitatively, the volume of air breathed in per minute (V1 ) is slightly greater than the volume
•
expired per minute (V E ). Clinically, however, this difference is not important, and by
•
convention, the ventilation rate is always measured on an expired sample or V E .
•
The rate of oxygen consumption ( V 02 ) is related to the rate of energy usage in
performing activities as follows:
Vo2=EEx ECF equation (8-2)
where,
•
V 02 = Oxygen consumption rate (liters Commute)
EE = Energy expenditure (kcal/minute)
ECF = Energy conversion factor (liters
The ECF shows little variation and typically, a value between 0.20 and 0.21 is used to
represent the conversion from energy units to oxygen consumption units. In this REA, APEX
randomly sampled from a uniform distribution defined by these lower and upper bounds to
estimate an ECF once for each simulated individual. The activity-specific energy expenditure is
highly variable and can be estimated using metabolic equivalents (METs). The METs are ratios
of the rate of energy consumption for non-rest activities to the resting rate of energy
consumption. Thus energy expenditure can be represented by the following:
EE=MET x RMR equation (8-3)
where,
EE = Energy expenditure (kcal/minute)
MET = Metabolic equivalent of work (unitless)
RMR = Resting metabolic rate (kcal/minute)
The CHAD database (EPA, 2002) contains distributions of METs for all activities that
might be performed by simulated individuals. APEX randomly samples from the various METs
distributions to obtain values for every activity performed by each individual. Age- and gender-
July 2009 225
-------
specific RMR are estimated once for each simulated individual using a linear regression model
(see Johnson et al., 2000)68 as follows:
RMR = [b0 + \ (BM) + s]F equation (8-4)
where,
RMR = Resting metabolic rate (kcal/min)
b0 = Regression intercept (MJ/day)
bj = Regression slope (MJ/day/kg)
BM = body mass (kg)
e = randomly sampled error term, N{0, se)69 (MJ/day)
F = Factor for converting MJ/day to kcal/min (0.166)
Finally, Graham and McCurdy (2005) describe an approach to estimate VE using V02.
In that report, a series of age- and gender-specific multiple linear regression equations were
derived from data generated in 32 clinical exercise studies. The algorithm accounts for
variability in ventilation rate due to variation in oxygen consumption, the variability within age
groups, and both inter- and intra-personal and variability. The basic algorithm follows:
\n(VE/BM) = b0+bl \n(V02/BM) + b2 In(1 + age) + b3 gender + eb+ew equation (8-5)
where,
In = natural logarithm of variable
VE! BM = activity-specific ventilation rate, body mass normalized (liter air/kg)
bj = see below
V02! BM = activity-specific oxygen consumption rate, body mass normalized
(liter/O2/kg)
age = the age of the individual (years)
gender = gender value (-1 for males and +1 for females)
eb = randomly sampled error term for between persons N{0, se), (liter air/kg)
ew = randomly sampled error term for within persons N{0, se), (liter air/kg)
68 The regression equations were adapted by Johnson et al. (2000) using data reported by Schofield (1985). The
regression coefficients and error terms used by APEX are provided in Appendix B Attachment 3.
69 The value used for each individual is sampled from a normal distribution (N) having a mean of zero (0) and
variability described by the standard error (se).
July 2009 226
-------
As indicated above, the random error (e) is allocated to two variance components and
used to estimate the between-person (inter-individual variability) residuals distribution (eb) and
within-person (intra-individual variability) residuals distribution (ew). The regression parameters
bo, bi, b2, and b3 are assumed to be constant over time for all simulated persons, eb is sampled
once per person, while whereas ew is sampled from event to event. Point estimates of the
regression coefficients and standard errors of the residuals distributions are given in Table 8-9.
Table 8-9. Ventilation coefficient parameter estimates (£>/) and residuals distributions (e,) from
Graham and McCurdy (2005).
Age
group
<20
20-<34
34-<61
en-
Regression Coefficients1
bo
4.3675
3.7603
3.2440
2.5826
bi
1.0751
1.2491
1.1464
1.0840
b2
-0.2714
0.1416
0.1856
0.2766
b3
0.0479
0.0533
0.0380
-0.0208
Random Error1'2
eb
0.0955
0.1217
0.1260
0.1064
ew
0.1117
0.1296
0.1152
0.0676
Notes:
1 These are the values of the coefficients and residuals distributions described by
equation 8-5.
2 The unique value used for each individual is sampled from a normal distribution (N)
having a mean of zero (0) and variability described by the standard error (se).
8.6 CONSTRUCTION OF LONGITUDINAL ACTIVITY SEQUENCES
Exposure models use human activity pattern data to predict and estimate exposure to
pollutants. Different human activities, such as spending time outdoors, indoors, or driving, will
result in varying pollutant exposure concentrations. To accurately model individuals and their
exposure to pollutants, it is critical to understand their daily activities. EPA's CHAD provides
data for where people spend time and the activities they perform. Typical time-activity pattern
data available for inhalation exposure modeling consist of a sequence of location/activity
combinations spanning 24-hours, with 1 to 3 diary-days for any single study individual.
The exposure assessment performed here requires information on activity patterns over a
full year. Long-term multi-day activity patterns were estimated from single days by combining
the daily records using an algorithm that represents the day-to-day correlation of activities for
individuals. The algorithm first uses cluster analysis to divide the daily activity pattern records
into groups that are similar, and then select a single daily record from each group. This limited
July 2009
227
-------
number of daily patterns is then used to construct a long-term sequence for a simulated
individual, based on empirically-derived transition probabilities. This approach is intermediate
between an assumption of no day-to-day correlation (i.e., re-selection of diaries for each time
period) and perfect correlation (i.e., selection of a single daily record to represent all days).
Details regarding the algorithm and supporting evaluations are provided in Appendix B,
Attachments 4 and 5.
8.7 CALCULATING MICROENVIRONMENTAL CONCENTRATIONS
Probabilistic algorithms are used to estimate the pollutant concentration associated with
each exposure event. The estimated pollutant concentrations account for temporal and spatial
variability in ambient (outdoor) pollutant concentration and factors affecting indoor
microenvironments, such as a penetration, air exchange rate, and pollutant decay or deposition
rate. APEX calculates air concentrations in the various microenvironments visited by the
simulated person by using the ambient air data estimated for the relevant blocks/receptors, the
user-specified algorithm, and input parameters specific to each microenvironment. The method
used by APEX to estimate the microenvironmental concentration depends on the
microenvironment, the data available for input to the algorithm, and the estimation method
selected by the user. The current version of APEX calculates hourly concentrations in all the
microenvironments at each hour of the simulation for each of the simulated individuals using one
of two methods: a mass balance model or a transfer factors method. Details regarding the
algorithms used for estimating specific microenvironments and associated input data derivations
are provided in Appendix B.
Briefly, the mass balance method simulates an enclosed microenvironment as a well-
mixed volume in which the air concentration is spatially uniform at any specific time. The
concentration of an air pollutant in such a microenvironment is estimated using the following
processes:
• Inflow of air into the microenvironment
• Outflow of air from the microenvironment
• Removal of a pollutant from the microenvironment due to deposition, filtration, and
chemical degradation
• Emissions from sources of a pollutant inside the microenvironment.
July 2009 228
-------
A transfer factors approach is simpler than the mass balance model; however, most
parameters are derived from distributions rather than single values to account for observed
variability. The transfer factors approach does not calculate concentration in a
microenvironment from the concentration in the previous hour as is done by the mass balance
method and contains only two parameters. A proximity factor is used to account for proximity
of the microenvironment to sources or sinks of pollution, or other systematic differences between
concentrations just outside the microenvironment and the ambient concentrations (at the
measurements site or modeled receptor). The second parameter, a penetration factor, quantifies
the amount of outdoor pollutant that penetrates into the microenvironment.
8.7.1 Approach for Estimating 5-Minute Maximum SOi Concentrations
Five-minute maximum SC>2 concentrations in each exposure modeling domain were
estimated using the empirically-derived PMRs (developed from recent 5-minute SO2 ambient
monitoring data, see section 7.2) and the AERMOD predicted 1-hour SC>2 concentrations. Thus,
for every 1-hour SC>2 concentration estimated at every receptor, an associated 5-minute
maximum SC>2 concentration was generated (i.e., twenty-four 5-minute maximum SC>2
concentrations per day). These statistically modeled 5-minute maximum SC>2 concentrations
were then used to estimate the eleven other 5-minute concentrations that occur within every hour
(see below). This spatially complete (at the block level) and consecutive time-series of 5-minute
SC>2 concentrations then served as the ambient concentrations input to algorithms within APEX
that estimate the microenvironmental concentrations.
The current version of APEX can use ambient concentrations of almost any time step,
including an averaging time of 5-minutes. However, if all of the individual block-level receptor
files were generated as an input to APEX in this assessment, the size and number of files would
become an issue. In this exposure assessment, each of the thousands of receptor files generated
by AERMOD would increase by a factor of twelve, creating disk space, pre-processing, and
exposure modeling difficulties. In addition, the APEX default exposure output for modeled
individuals is the single greatest exposure within a day, thus requiring model changes to obtain
output of a different form. Staff believed that to reasonably estimate multiple peak
concentrations that might occur within an hour by addressing these issues would further
encumber the limited time and resources already available to staff to conduct the assessment.
July 2009 229
-------
Staff elected to use a simplified approach to generate all other 5-minute 862
concentrations that occur within the hour. The objective of the approach used was not to
estimate each of the other eleven 5-minute concentrations with a high degree of certainty; each
of these concentrations, by definition, would be lower than the maximum for that hour. While
the occurrence of multiple peak concentrations above benchmark levels within an hour is
possible, staff assumed that use of the twenty-four 5-minute maximum SC>2 concentrations could
provide an accurate estimate of the maximum exposure an individual might experience in a
day.70 Further discussion regarding multiple peak exposures within an hour is given in section
8.11.
The technical approach to estimating 862 concentrations real-time within the APEX
model rather than modeled externally is as follows. An algorithm was incorporated into the
flexible time-step APEX model to estimate the 5-minute maximum SC>2 concentrations using the
1-hour SC>2 concentration, an appropriate PMR (section 7.2), and equation 7-1. The additional
eleven 5-minute concentrations within an hour at each receptor were approximated using the
following:
nC-P
X = equation (8-6)
n-l
where,
X = 5-minute SO2 concentration in each of non-peak concentration periods in
the hour at a receptor (ppb)
C = 1-hour SO2 concentration estimated at a receptor (ppb)
P = estimated 5-minute maximum SC>2 concentration at a receptor (ppb) using
equation 7-1.
n = number of time steps within the hour (or 12)
In addition to the level of the 5-minute maximum SC>2 concentration, the actual time of
when the contact occurs with a person is also of importance. There is no reason to expect a
temporal relationship of the peak concentrations within the hour, thus clock times for peak
values were estimated randomly (i.e., any one of the 12 possible time periods within the hour).
The PMR assignment also assumes a standard frequency during any hour of the day.
70 Note that the model still uses all of the statistically-modeled twenty-four 5-minute maximum SO2 concentrations
(one for every hour in the day) in estimating microenvironmental concentrations and personal exposures.
July 2009 230
-------
8.7.2 Microenvironments Modeled
In APEX, microenvironments represent the exposure locations for simulated individuals.
For exposures to be estimated accurately, it is important to have realistic microenvironments that
match closely to the locations where actual people spend time on a daily basis. As discussed
above, the two methods available in APEX for calculating pollutant levels within
microenvironments are mass balance or a transfer factors approach. Table 8-10 lists the
microenvironments used in this study, the calculation method used, and the type of parameters
used to calculate the microenvironment concentrations.
Table 8-10. List of microenvironments modeled and calculation methods used.
Microenvironment
Indoors - Residence
Indoors - Bars and restaurants
Indoors - Schools
Indoors - Day-care centers
Indoors - Office
Indoors- Shopping
Indoors - Other
Outdoors - Near road
Outdoors - Public garage - parking lot
Outdoors - Other
In-vehicle - Cars and Trucks
In-vehicle - Mass Transit (bus, subway,
train)
Calculation
Method
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Factors
Factors
Factors
Factors
Factors
Parameter Types
used1
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
AER and DE
PR
PR
None
PE and PR
PE and PR
1 AER=air exchange rate, DE=decay-deposition rate, PR=proximity factor,
PE=penetration factor
8.7.3 Microenvironment Descriptions
8.7.3.1 Microenvironment 1: Indoor-Residence
The Indoor-Residence microenvironment uses several variables that affect 862 exposure:
whether or not air conditioning is present, the average outdoor temperature, the 862 removal
rate, and an indoor concentration source.
Air conditioning prevalence rates
Since the selection of an air exchange rate distribution is conditioned on the presence or
absence of an air-conditioner, for each modeled area the air conditioning status of the residential
microenvironments is simulated randomly using the probability that a residence has an air
July 2009
231
-------
conditioner. A value of 96% was used to represent the air conditioning prevalence rate in both
Greene County and St. Louis, using the data obtained from the St. Louis American Housing
Survey of 2004 (AHS, 2005). Air conditioning prevalence is noted as distinct from usage rate,
the latter being represented by the air exchange rate distribution and dependent on temperature
(see next section).
Air exchange rates
Air exchange rate data for the indoor residential microenvironment were the same used in
APEX for the most recent Os NAAQS review (EPA, 2007d; see Appendix B, Attachment 6).
Briefly, data were reviewed, compiled and evaluated from the extant literature to generate
location-specific AER distributions categorized by influential factors, namely temperature and
presence of air conditioning. In general, lognormal distributions provided the best fit, and are
defined by a geometric mean (GM) and standard deviation (GSD). To avoid unusually extreme
simulated AER values, bounds of 0.1 and 10 were selected for minimum and maximum AER,
respectively. Table 8-11 summarizes the AER distributions used in modeling indoor residential
exposures, separated by A/C prevalence and temperature categories. See Appendix B,
Attachment 6 for additional details.
Table 8-11. Geometric means (GM) and standard deviations (GSD) for air exchange rates by A/C
type and temperature range.
A/C Type1
Central or
Room A/C
No A/C
Temp
(°C)
<=10
10-20
20-25
25-30
>30
<=10
10-20
>20
N
179
338
253
219
24
61
87
44
GM
0.9185
0.5636
0.4676
0.4235
0.5667
0.9258
0.7333
1.3782
GSD
1.8589
1 .9396
2.2011
2.0373
1 .9447
2.0836
2.3299
2.2757
Notes:
1 All distributions derived from data reported in non-California cities. See
Appendix B, Attachment 6 for details in the data used and distribution
derivation.
The AER data obtained was limited in the number of samples, particularly when
considering these influential factors. When categorizing by temperature, a range of temperatures
was used to maintain a reasonable number of samples within each category to allow for some
variability within the category, while still allowing for differences across categories. Several
July 2009
232
-------
distribution forms were investigated (i.e., exponential, log-normal, normal, and Weibull) and in
general, lognormal distributions provided the best fit. Fitted lognormal distributions were
defined by a geometric mean (GM) and standard deviation (GSD). Because no fitted distribution
was available specifically for St. Louis or Greene County, distributions were selected from other
locations thought to have similar characteristics, qualitatively considering factors that might
influence AERs including the age composition of housing stock, construction methods, and other
meteorological variables not explicitly treated in the analysis, such as humidity and wind speed
patterns.
SO2 Removal Rate
Staff estimated distributions of indoor 862 deposition rates by applying a Monte Carlo
sampling approach to configurations of indoor microenvironments of interest. The relative
composition of particular surface materials (e.g., painted wall board, wall paper, wool carpet,
synthetic carpet, synthetic floor covering, cloth) within various sized buildings were
probabilistically modeled to estimate 1,000 SC>2 deposition rates that in turn were used to
parameterize lognormal distributions (Table 8-12). The modeling was fundamentally based on a
review of 862 deposition conducted by Grontoft and Raychaudhuri (2004) for a variety of
building material surfaces under differing conditions of relative humidity. Details on the data
used and derivation of removal rates are provided in Appendix B, section 4.
Table 8-12. Final parameter estimates of SO2 deposition distributions in several indoor
microenvironments modeled in APEX.
Microenv-
ironment
Residence
Office
School/
Day Care
Center
Restaurant
Other
Indoors
Heating or Air Conditioning in Use
or Low Ambient Humidity1
Geom.
Mean
(hr'1)
3.14
3.99
4.02
2.36
2.82
Geom.
Stand.
Dev.
(hr'1)
1.11
1.04
1.02
1.28
1.21
Lower
Limit
(hr1)
2.20
3.63
3.90
1.64
1.71
Upper
Limit
(hr1)
5.34
4.37
4.21
4.17
4.12
Air Conditioning Not in Use
(Summertime Ambient Morning
Relative Humidity of 90%)
Geom.
Mean
(hr1)
13.41
N/A
N/A
N/A
N/A
Geom.
Stand.
Dev.
(hr1)
1.11
N/A
N/A
N/A
N/A
Lower
Limit
(hr-1)
10.31
N/A
N/A
N/A
N/A
Upper
Limit
(hr-1)
26.96
N/A
N/A
N/A
N/A
Notes:
1 Summertime ambient afternoon relative humidity of 50%.
N/A not applicable, assumed by staff to always have A/C in operation.
July 2009
233
-------
8.7.3.2 Microenvironments 2-7: All Other Indoor Microenvironments
The remaining six indoor microenvironments, which represent Bars and Restaurants,
Schools, Day Care Centers, Office, Shopping, and the broadly defined Other Indoor
microenvironments, were all modeled using the same data and functions. An air exchange rate
distribution (GM = 1.109, GSD = 3.015, Min = 0.07, Max = 13.8) was based on an indoor air
quality study (Persily et al., 2005). This is the same distribution in APEX used for the most
recent O3 NAAQS review (EPA, 2007d) and NO2 REA (EPA, 2008d). See Appendix B,
Attachment 6 for details in the data used and derivation. The SO2 removal rates in these six
indoor microenvironments were estimated as explained in section 8.7.3.1, and described in more
detail in Appendix B, section 4. The resulting lognormal distributions for removal rates are
presented in Table 8-12. These microenvironments are all assumed to have air-conditioning.
8.7.3.3 Microenvironments 8-10: Outdoor Microenvironments
All outdoor microenvironmental concentrations are well represented by the modeled
concentrations. Therefore, both the penetration factor and proximity factor for this
microenvironment were set to 1.
8.7.3.4 Microenvironments 11 and 12: In Vehicle- Cars and Trucks, and Mass Transit
There were no available measurement data for SO2 penetration factors, therefore the
penetration factors used were developed from NO2 data provided in Chan and Chung (2003) and
used in the recent NO2 NAAQS review (EPA, 2008d). NO2 and SO2 are expected to have
similar penetration rates inside vehicles since both are gases. Although the in-vehicle NO2
measurements used in the in-vehicle-to-outdoor-ratios might include a small amount of in-
vehicle emissions, resulting in some discrepancy between effective penetration factors for NO2
and SO2, the additional uncertainty is expected to be small compared to the overall uncertainty
implied by the broad uniform distributions.
Inside-vehicle and outdoor NO2 concentrations were measured for three ventilation
conditions: air-recirculation, fresh air intake, and with windows open. Mean in-vehicle-to-
outdoor ratio values ranged from about 0.6 to just over 1.0, with higher values associated with
increased ventilation (i.e., window open). A uniform distribution U{0.6, 1.0} was selected for
the penetration factor for Inside-Cars/Trucks due to the limited data available to describe a more
July 2009 234
-------
formal distribution and the lack of data available to reasonably assign potentially influential
characteristics such as use of vehicle ventilation systems for each location. Mass transit systems,
due to the frequent opening and closing of doors, was assigned a uniform distribution U{0.8,
1.0} based on the reported mean values for fresh-air intake (0.796) and open windows (1.032) on
urban streets.
8.8 EXPOSURE MEASURES AND HEALTH RISK CHARACTERIZATION
8.8.1 Estimation of Exposure
APEX calculates exposure as a time-series of exposure concentrations that a simulated
individual experiences during the simulation period. APEX calculates exposure by identifying
concentrations in the microenvironments visited by the person according to the composite diary.
In this manner, a time-series of event exposures are found. Then, the time-step exposure
concentration at any clock hour during the simulation period is calculated using the following
equation:
N
V c t
Z_^ ^ time -step (j) (j)
equation (8-7)
where,
d = Time-step exposure concentration at clock hour / of the simulation
period (ppm)
N = Number of events (i.e., microenvironments visited) in time-step /'
of the simulation period.
C ' time_step(n = Time-step concentration in microenvironmenty (ppm)
t(j) = Time spent in microenvironmenty (minutes)
T = Length of time-step (or 5 minutes in this analysis)
From the time-step exposures, APEX calculates time-series of 5-minute, 1-hour, 24-hour,
and annual average exposure concentrations that a simulated individual would experience during
the simulation period. APEX then statistically summarizes and tabulates the 5-minute time-step
(or daily, or annual average) exposures. From this, APEX can calculate two general types of
exposure estimates: counts of the estimated number of people whose exposure exceeded a
specified SC>2 concentration level 1 or more times in a year and the number of times per year that
they are so exposed; the latter metric is in terms of person-occurrences or person-days. The
July 2009 235
-------
former highlights the number of individuals whose exposure exceeded at least one or more times
per modeling period the health effect benchmark level of interest. APEX can also report counts
of individuals with multiple exposures. This person-occurrences measure estimates the number
of times per season that individuals are exposed to the exposure indicator of interest and then
accumulates these estimates for the entire population residing in an area.
In this exposure assessment, APEX tabulates and displays the two measures for
exposures above levels ranging from 0 to 800 ppb by 50 ppb increments for all exposures. These
results are tabulated for the total population and subpopulations (i.e., asthmatics, asthmatic
children) of interest.
8.8.2 Estimation of Target Ventilation Rates
Human activities are variable over time, a wide range of activities are possible even
within a single hour of the day. The type of activity an individual performs, such as sleeping or
jogging, will influence their breathing rate. As discussed above in section 8.5.5, APEX estimates
minute-by-minute ventilation rates that account for the expected variability in the activities
performed by simulated individuals. The ISA indicates that the adverse lung function responses
associated with short-term peak exposures at levels below 1,000 ppb coincide with moderate to
heavy exertion levels. Therefore, staff needed to identify a target ventilation rate in the
simulated individuals to further characterize the estimated exposures of interest.
The target ventilation for adults (both a mix of males and females) experiencing effects
from 5-10 minute 862 exposures in many of the controlled human exposure studies was
approximately between 40-50 L/min (Table 3-1, ISA).71 Since there were limited controlled
human exposure study data available for asthmatic children, the ventilation targets needed to be
normalized. Normalized ventilation rates allow for extrapolation of the adult target ventilation
rate and, hence the health effect response associated with that ventilation rate to asthmatic
children. One method used to normalize ventilation rate is to generate an equivalent ventilation
rate (EVR) based on normalizing the simulated individuals activity-specific ventilation rate (VE)
to their body surface area (BSA). Staff has used EVR in previous 63 NAAQS reviews to also
71 Note that study subjects were free-breathing; thus it is expected that there was a mixture of nasal, oral, and
oronasal breathing that occurred across the study subjects. Without information regarding the breathing method
used by any subject and their corresponding health response, staff assumed that the mixture in breathing method is
representative for the simulated population.
July 2009 236
-------
identify comparable activity-specific ventilation rates for children and adults (EPA, 2007d;
Whitfield et al., 1996). In these reviews, an EVR ranging from 16-30 L/min-m2 was associated
with moderate exertion over a 1-hour exposure event, while an EVR ranging from 13-27 L/min-
m2 was associated with moderate exertion over an 8-hour exposure event.
As was done in the Os NAAQS reviews, target ventilation rates were identified in this
exposure assessment by normalizing ventilation rates reported in the clinical studies on adults
(i.e., 40-50 L/min, also see Table 9-3) to body surface area (BSA) to allow for such an
extrapolation from adults to children. Body surface area was not measured in the controlled
human exposure studies and the relevant ventilation data were not separated by gender. Staff
obtained median estimates of BSA for males (1.94 m2) and females (1.69 m2) (EPA, 1997) and
calculated a mean value of 1.81 m2. Based on this data, an EVR = 40/1.81 = 22 L/min-m2 was
used to characterize the minimum target ventilation rate of interest. Individuals at or above an
EVR of 22 L/min-m2 (children or adult) for a 5-minute exposure event were characterized as
performing activities at or above a moderate ventilation rate.
8.8.3 Adjustment for Just Meeting the Current and Alternative Standards
We used a different approach to simulate just meeting the current and alternative
standards than was used in the Air Quality Characterization (see section 7.2.4). In this case,
instead of proportionally adjusting the ambient concentrations, we proportionally adjusted the
health effect benchmark levels used in each exposure modeling domain. The benchmark levels
were adjusted rather than the air quality to reduce the processing time associated with the
modeling of several thousands of receptors in each of the large exposure modeling domains. A
proportional adjustment of the selected benchmark level (i.e., division by the adjustment factor)
is mathematically equivalent to a proportional adjustment of the air quality concentrations (i.e.,
multiplication by the adjustment factor).72 Therefore, the end effect of adjusting exposure model
input concentrations upward versus adjusting exposure model benchmark levels downward is
identical.
For example, an adjustment factor of 5.10 was determined for year 2002 in Cuyahoga
County to simulate ambient concentrations just meeting the current standard. This value was
72 To evaluate the current and most of the proposed alternative standards, 1-hour ambient concentrations were
typically adjusted upwards to just meet the standards. This would correspond to downward adjustments to the
benchmark levels.
July 2009 237
-------
based on an annual average 862 concentration of 5.96 ppb observed at an ambient monitor (ID
390350060) for that year (see Appendix A, section A.3). Therefore in the exposure analysis, the
5-minute potential health effect benchmark levels of 100, 200, 300, and 400 ppb were
proportionally adjusted downward to 19.6, 39.2, 58.8, and 78.4 ppb, respectively for year 2002.
APEX reported the number of days an individual was exposed above each of the adjusted
benchmark levels using the as is air quality as the ambient concentration input. To illustrate the
relationship between the two procedures (air quality adjustment versus benchmark adjustment), a
comparison of the distributions and benchmark exceedances is presented in Figure 8-13. This
example used the distribution of hourly 862 concentrations measured at one ambient monitor (ID
390350045) within the Cuyahoga County modeling domain for year 2002. Staff used the
statistical model (section 7.2.3) to estimate 5-minute maximum SO2 concentrations from both the
adjusted and unadjusted 1-hour SC>2 concentrations. If one were interested in the number of days
per year with 5-minute SC>2 benchmark exceedances of 400 ppb under the current standard
scenario for example, this would be equivalent to counting the number of days with 5-minute
maximum SC>2 concentrations above 78.4 ppb using the as is air quality.
For additional clarity, the same ambient air quality data are presented in Figure 8-14, only
with expansion of the highest percentiles on the graph to allow for improved visualization of the
number of exceedances. When using the air quality adjusted to just meet the current standard,
there were 14 days where the maximum 5-minute concentration was greater than 400 ppb.73
When considering the as is air quality without adjustment but with a downward adjustment of
the benchmark by the same factor of 5.10, there are the same number of days with exceedances
(i.e., 14 exceedances). Due to the relationship between the two procedures, the estimated
number of exceedances at each of the other benchmark levels is identical (Table 8-13).
The values for each adjusted benchmark level considering each of the air quality standard
scenarios are given in Table 8-14. Staff applied the benchmark adjustment in each of the
exposure modeling domains to simulate exposures associated with just meeting the current and
alternative standards.
73 Only 12 points are observed in Figure 8-13 however, three peak concentrations were identical within each of the
simulations.
July 2009 238
-------
MODELED CUYAHOGA COUNTY Daily 5-Minute Max (2002)
As Is and Adjusted to Just Meet the Current Standard (CS)
o>
100
90
80
70 ---
0)
Jj3 60 +
Q.
Q) 50
re 40
3
| 30
O
20
10
0
T^ *• -
/\
'
/ t
-/— |— -
1
i /
' \/
t . /
t ]/
'!""*""
U-t-
ts \
** *
^
s
-**-
/*
Adjusted 400 ppb
Benchmark Level
to Simulate Just
Meeting the
Current Standard
(78.4 ppb)
-*"** *
„..*•'
r.v. » >-*---"-•
400 ppb
j Benchmark
Level
« 390350045 - As Is Air Quality
» 390350045 - Adjusted Air Quality
0 50 100 150 200 250 300 350 400 450 500 550 600 650 700
5-minute Daily Maximum SO2 (ppb)
Figure 8-13. Comparison of adjusted ambient monitoring concentrations or adjusted benchmark
level (dashed line) to simulate just meeting the current annual average standard at one
ambient monitor in Cuyahoga County for year 2002.
100 -
flj nn
Cumulative Percentill
) CD CD C!
) --J 00 C!
ao -
QE; _
I
1
\ 0
\
\ °
<>
lo
1
II
A \
<\
o.
1
Adju:
Benc
toS
M
Cum
r
i
i
400 ppb
Benchmark
Level
sted 400 ppb
hmark Level
mulate Just
seting the
snt Standard
?8.4 ppb)
i i
i i
i i
i i
---
4
»
>
4
^*
>
;
*
0 390350045 - As Is Air Quality
• 390350045 - Adjusted Air Quality
1 1 1 1 1
50 100 150 200 250 300 350 400 450 500 550 600 650 700
Daily 5-minute Maximum SO2 (ppb)
Figure 8-14. Comparison of the upper percentile modeled daily 5-minute maximum SO2
concentrations using either adjusted 1-hour ambient SO2 concentrations or an adjusted
benchmark level (with as is air quality) to simulate just meeting the current annual
standard at monitor 390350045 in Cuyahoga County for year 2002. Complete
distributions are provided in Figure 8-13.
July 2009
239
-------
Table 8-13. Comparison of benchmark levels, adjusted benchmark levels to just meet the current
standard, the benchmark level distribution percentiles, and the number of 5-minute SO2
benchmark exceedances at monitor 390350045 in Cuyahoga County for year 2002.
Benchmark
Level
(ppb)
100
200
300
400
Adjusted
Benchmark Level1
(ppb)
19.6
39.2
58.8
78.4
Concentration
Distribution
Percentile2
37.3
76.7
92.0
97.0
Number of Days
with a Benchmark
Exceedance3
230
86
30
14
Notes:
1 The adjustment factor to simulate just meeting the current standard was 5.10.
2 The percentile of the distribution for each benchmark and adjusted benchmark
level was the same.
3 The number of days with a benchmark exceedance when using either air quality
adjusted to just meet the current standard or applying adjusted benchmarks to as
;'s air quality was the same.
July 2009
240
-------
Table 8-14. Exposure concentrations and adjusted potential health effect benchmark levels used by APEX to simulate just meeting the
current and potential alternative standards in the Greene County and St Louis modeling domains.
Modeling
Domain
Greene
County
St. Louis
Form1
98
99
99
99
99
99
CS
as is
98
99
99
99
99
99
CS
as is
Level2
200
50
100
150
200
250
200
50
100
150
200
250
Exposure Concentrations and Adjusted Potential Health Effect Benchmark Levels (ppb)3
50
20.3
94.3
47.2
31.4
23.6
18.9
14.4
50
13.3
63.3
31.7
21.1
15.8
12.7
8
50
100
40.5
188.7
94.3
62.9
47.2
37.7
28.8
100
26.5
126.7
63.3
42.2
31.7
25.3
16
100
150
60.8
283
141.5
94.3
70.8
56.6
43.2
150
39.8
190
95
63.3
47.5
38
24
150
200
81
377.3
188.7
125.8
94.3
75.5
57.6
200
53
253.3
126.7
84.4
63.3
50.7
32
200
250
101.3
471.7
235.8
157.2
117.9
94.3
72
250
66.3
316.7
158.3
105.6
79.2
63.3
40
250
300
121.5
566
283
188.7
141.5
113.2
86.4
300
79.5
380
190
126.7
95
76
48
300
350
141.8
660.3
330.2
220.1
165.1
132.1
100.8
350
92.8
443.3
221.7
147.8
110.8
88.7
56
350
400
162
754.7
377.3
251.6
188.7
150.9
115.2
400
106
506.7
253.3
168.9
126.7
101.3
64
400
450
182.3
849
424.5
283
212.3
169.8
129.6
450
119.3
570
285
190
142.5
114
72
450
500
202.5
943.3
471.7
314.4
235.8
188.7
144
500
132.5
633.3
316.7
211.1
158.3
126.7
80
500
550
222.8
1037.7
518.8
345.9
259.4
207.5
158.3
550
145.8
696.7
348.3
232.2
174.2
139.3
88
550
600
243
1132
566
377.3
283
226.4
172.7
600
159
760
380
253.3
190
152
96
600
650
263.3
1226.3
613.2
408.8
306.6
245.3
187.1
650
172.3
823.3
411.7
274.4
205.8
164.7
104
650
700
283.5
1320.7
660.3
440.2
330.2
264.1
201.5
700
185.5
886.7
443.3
295.6
221.7
177.3
112
700
750
303.8
1415
707.5
471.7
353.8
283
215.9
750
198.8
950
475
316.7
237.5
190
120
750
800
324
1509.3
754.7
503.1
377.3
301.9
230.3
800
212
1013.3
506.7
337.8
253.3
202.7
128
800
Notes:
1 The form of the standard used to adjust the air quality. 98 is the 98th percentile 1 -hour daily maximum alternative standard, 99 is the 99th percentile 1-hour
daily maximum alternative standard, CS is either the current annual average or 24-hour SO2 NAAQS (whichever had the lowest factor), as ;'s is unadjusted
air quality.
2 The level of the potential alternative standards, i.e., the 1-hour daily maximum at the noted percentile of the distribution.
3 Exposure levels were defined in 50 ppb increments from 0 through 800 ppb even though the selected potential health effect benchmark levels were 100
to 400 ppb in 100 ppb increments.
July 2009
241
-------
8.9 EXPOSURE MODELING AND HEALTH RISK CHARACTERIZATION
RESULTS
Exposure results are presented for simulated asthmatic populations residing in the two
modeling domains in Missouri. For each individual, APEX estimates the number of days with a
5-minute 862 exposure above the potential health effect benchmark levels year 2002. These
short-term exposures were evaluated for all asthmatics and asthmatic children when the exposure
corresponded with moderate or greater activity levels (i.e., the simulated individuals EVR during
a 5-minute exposure event was >22 L/minute-m2). The number of persons and days with at least
one 5-minute SC>2 exposure at or above any level from 0 through 800 ppb in 50 ppb increments
was reported by APEX. Therefore, for each concentration level, an individual at a moderate (or
higher) exertion level while exposed would have at most one exceedance of a particular level per
day, or 365 per year.
Multiple air quality scenarios were evaluated, including unadjusted air quality (termed as
is), air quality adjusted to just meet the current NAAQS, and air quality adjusted to just meet
several potential alternative 1-hr daily maximum standards. Exposure results are presented in a
series of figures that allow for simultaneous comparison of exposures associated with each air
quality scenario. Four types of results are provided for each exposure modeling domain: (1) the
number of persons in the simulated subpopulation exposed at or above selected levels 1 or more
times in a year, (2) the percent of the simulated subpopulation exposed at or above selected
levels 1 or more times in a year, (3) the total number of days in a year the simulated
subpopulation is exposed (or person days) at or above selected levels, and (4) the percent of time
associated with the exposures at or above the selected levels. Tables summarizing all of the
exposure results for each modeling domain, air quality scenario, exposure level, and
subpopulation are provided in Appendix B.4.
8.9.1 Asthmatic Exposures to 5-minute SOi Concentrations in Greene County
When considering the lowest 5-minute benchmark level of 100 ppb, approximately one
thousand asthmatics are estimated to be exposed at least once in the year 2002 while at moderate
or greater exertion and when considering the current standard air quality scenario (top of Figure
8-15). Each of the potential alternative 1-hr standard air quality scenarios as well as the as is air
quality scenario result in fewer asthmatics exposed when compared with the current standard
scenario, and progressively fewer persons were exposed with decreases in the 1-hour daily
July 2009 242
-------
maximum concentration levels of the potential alternative standards. The 99th percentile 1-hour
daily maximum standard levels of 50 and 100 ppb produced the same number of persons with at
least one 5-minute exposure at or above 100 ppb as the as is air quality (i.e., 13). With
progressive increases in benchmark level, there were corresponding decreases in the number of
individuals exposed. None of the asthmatics had a day where 5-minute exposures were above
100 ppb when considering the as is air quality scenario. Asthmatic children exhibited similar
patterns in the estimated number of exposures at each of the exposure levels, thus comprising a
large proportion of the total asthmatics exposed (bottom of Figure 8-15).
The difference between all asthmatics and asthmatic children is best demonstrated by
comparing the percent of the subpopulation exposed. Asthmatic children have nearly double the
percentage of the subpopulation exposed at any of the benchmark levels considered when
compared with that of all asthmatics (Figure 8-16). For example, approximately 1% of asthmatic
children experience at least one day with a 5-minute SC>2 exposure at or above 200 ppb in a year
in considering the current standard scenario, while approximately 0.6% of all asthmatics
experienced a similar exposure. As observed with the numbers of persons exposed, a lower
estimated percent of persons was exposed at the higher benchmark levels, though again, the
current standard scenario contains the greatest percent of asthmatics exposed when compared
with all of the other 1-hour air quality standard scenarios analyzed.
The number of person days or occurrences of exposures is greater than the number of
persons exposed, indicating that some of the simulated asthmatics had more than one day with 5-
minute exposures above selected benchmark levels (Figure 8-17). For example, when
considering all asthmatics and the current standard scenario, there were approximately 22 person
days with exposures at or above 300 ppb. This corresponds with the 18 asthmatics estimated to
experience at least one day with a 5-minute SC>2 concentration above this level, indicating that a
number of persons may have experienced at least 2 benchmark exceedances in the year. For
both subpopulations considered, there were no estimated exposures above 300 ppb when
considering the 99th percentile 1-hour daily maximum alternative standard level of 200 ppb.
Staff evaluated the microenvironments where the peak exposures frequently occurred.
There were very few persons exposed to benchmark levels of 100 ppb or higher considering the
as is air quality, though 99% or greater experienced their 5-minute maximum SC>2 exposure in an
outdoor microenvironment (i.e., outdoors or outdoors near-roads) when considering any of the
July 2009 243
-------
benchmark levels. For the current standard air quality scenario, approximately 7% of persons
were exposed to the 100 ppb benchmark level indoors (i.e., primarily in the persons residence),
though with increasing benchmark level (e.g., 300 ppb) the percent of persons with any
benchmark exceedances indoors approached zero (i.e., > 99% occurred outdoors). The inside
vehicle microenvironment also comprised a small percent of the cases where the exposures
above selected levels occurred; at most 2% of benchmark exceedances occurred inside vehicles
when considering the lowest benchmark levels.
Two forms of the potential alternative standard were evaluated in Greene County, i.e., the
99th and 98th forms of a 1-hour daily maximum level of 200 ppb. The difference in the exposure
results generated for each of these air quality scenarios is provided in Table 8-15. The 99th
percentile form of the potential alternative standard results in fewer persons, person-days, and
percent of asthmatic persons exposed when compared with estimated exposures using air quality
adjusted to just meet a 200 ppb 1-hour daily maximum 98th form. The values listed in the table
are small, but from a relative perspective, the percent difference can be large. For example, there
is approximately a 40% reduction in the percent of persons exposed when considering the 99th
percentile form and the 100 ppb benchmark level. Where there were other higher benchmark
levels that were exceeded, the reduction was greater (66% to 100%). For additional relative
comparisons for these two standard forms, see the corresponding Figures 8-15 to 8-17.
Table 8-15. Absolute difference in APEX exposure estimates for Greene County using either a 98
th
..th
or 99 percentile form potential alternative standard at a 1-hour daily maximum level of 200 ppb.
Population
All Asthmatics
(21 ,948)
Asthmatic
Children
(7,285)
Benchmark
Level (ppb)
100
200
300
400
100
200
300
400
Absolute Difference in Estimated Exposures
using 98th and 99th form1
Number of
Person-days
274
27
13
0
161
18
4
0
Number of
Persons
157
27
13
0
81
18
4
0
Percentage
Points2
0.7
0.1
0.1
0
0.4
0
0
0
Notes:
1 Both the 98th and 99th 1 -hour daily maximum air quality scenarios were simulated by
APEX, using a level of 200 ppb. The value reported is the difference between the 98th
and the 99th.
2 Difference between the percent of persons exposed (98th-200 minus the 99th-200) at
each benchmark level.
July 2009
244
-------
1000
o>
.a
E
D
re
o>
100
Q) i_
> 0)
O **
li
.- O
O !_
«s
LLI
0)
O
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
1,000
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-15. Number of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute SO2 exposure above selected benchmark levels in Greene
County, year 2002 air quality as is and adjusted to just meeting the current and
potential alternative standards.
July 2009
245
-------
re
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum Exposure SO2 Level (ppb)
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-16. Percent of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute SO2 exposure above selected benchmark levels in Greene
County, year 2002 air quality as is and adjusted to just meeting the current and
potential alternative standards.
July 2009
246
-------
10,000
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
10,000
o
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-17. Number person days all asthmatics (top) and asthmatic children (bottom) experience
a 5-minute SO2 exposure above selected benchmark levels in Greene County, year 2002
air quality as is and adjusted to just meeting the current and potential alternative
standards.
July 2009
247
-------
8.9.2 Asthmatic Exposures to 5-minute SOi in St. Louis
The patterns in the number of persons (either asthmatics or asthmatic children) exposed
in St. Louis were different from those observed in Greene County; a greater number of persons
were estimated to be exposed in St. Louis at each of the corresponding benchmark levels and air
quality scenarios (Figure 8-18). For example, nearly 80,000 asthmatics were estimated to
experience at least one day with a 5-minute SC>2 exposure at or above 100 ppb when considering
the current standard scenario compared to the one thousand asthmatics estimated in Greene
County (section 8.9.1). In addition, there were more persons exposed to the higher benchmark
levels in St. Louis compared with Greene County. For example, none of the asthmatics
experienced a 5-minute 862 concentration exposure above 450 ppb in Greene County
considering any of the air quality scenarios. In St. Louis many of the air quality scenarios had
persons with exceedances of 450 ppb; the estimated number of persons experiencing at least one
day with a 5-minute SC>2 exposure above 450 ppb ranged from a low of 16 (the 99th percentile 1-
hour daily maximum standard level of 100 ppb) to over 10,000 (the current standard air quality
scenario). We note though, in considering the as is air quality scenario, none of the asthmatics in
St. Louis had 5-minute SC>2 exposures above a 450 ppb exposure level.
There were also differences in the estimated percent of asthmatics and asthmatic children
exposed to concentrations above the benchmark levels in St. Louis when compared with Greene
County. For example, over 40% of asthmatic children were estimated to experience at least one
day with a 5-minute exposure above 300 ppb in St. Louis considering the current standard air
quality scenario, while less than 1% of asthmatic children in Greene County experienced a
similar exposure (Figure 8-19). Just as observed with the Greene County estimates though, there
were decreases in the percent of persons exposed with decreases in the 1-hour daily maximum
level of the potential alternative standards. For example, less than 3% of asthmatic children
were estimated to have at least one day with a 5-minute SC>2 exposure above 300 ppb when
considering a 99th percentile 1-hour daily maximum standard level of 150 ppb.
The discussion regarding the patterns observed in the number of persons exposed in St.
Louis can be extended to the number of person days (i.e., both a greater number and at higher
benchmark levels when compared with Greene County). In addition, St. Louis had a greater
number of persons with multiple exceedances when compared with Greene County (Figure 8-
20). For example, given the 22 person days at or above 300 ppb in Greene County experienced
July 2009 248
-------
by the 18 asthmatics considering air quality just meeting the current standard, on average this
amounts to approximately 1.2 exposures per person per year. In contrast, approximately 26,000
asthmatics had nearly 50,000 person days at the same benchmark level and air quality scenario in
St. Louis; on average each person is estimated to experience 1.9 exposures exceeding this
benchmark level in a year.
Staff also evaluated the microenvironments where the peak exposures occurred in St.
Louis, and again, there were differences when compared with the exposures in Greene County.
In St. Louis, there were a greater percentage of benchmark exceedances within indoor and inside
vehicle microenvironments, although overall still comprising a small percentage of where the
exceedances were occurring. At the 100 ppb benchmark level, approximately 10% of the
exposures occur within indoor microenvironments (i.e., principally inside residences) and about
5% occur inside vehicles considering as is air quality (Figure 8-21). The percentage increases
when considering air quality adjusted to just meeting the current standard, with approximately
30% of benchmark exceedances of 100 ppb occurring indoors and 20% occurring inside
vehicles. Just beyond the benchmark level of 400 ppb, nearly all of the exceedances occur
outdoors when considering the as is air quality, while indoor microenvironments still contribute
to around 10% of exceedances, up to a 5-minute exposure level of 800 ppb. For comparison, air
quality adjusted to just meet a 99th percentile 1-hour daily maximum standard level of 150 ppb is
also shown, and falls within the range of values provided by the as is and current standard
scenarios.
Two forms of potential alternative standards were also evaluated in St. Louis, using the
99th and 98th percentile forms of a 1-hour daily maximum level of 200 ppb. The difference in the
exposure results generated for each of these air quality scenarios is provided in Table 8-16. The
99th percentile form of the potential alternative standard results in fewer persons, person-days,
and percent of asthmatic persons exposed when compared with estimated exposures using air
quality adjusted to just meet a 200 ppb 1-hour daily maximum 98th percentile form. The impact
of the different scenario is greater than that observed in Greene County from a pure numbers
perspective given so few persons exposed to concentrations above the benchmark levels in
Greene County. From a relative perspective, the percent difference between the two scenarios
can also be large. The reduction in the percent of persons exposed when considering the 99th
July 2009 249
-------
percentile form ranges from approximately 10% to 50%. For additional relative comparisons
between these two standard forms, see the corresponding Figures 8-18 to 8-20.
July 2009 250
-------
*- re
re «
Q) '=
sfl
^H ^ ™
*i fl)
5 _i uj
2 ^ •-
.s ^ fl)
111
£ >- O
w ° t
IS
M
100,000
10,000
1,000
9950
^^99150
^^99250
-e-CS
100
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
re
a) c
> o
100,000
10,000
1,000
100
> o> x
il s
1 o!
O ° o
o re ^
're 2 °
C = 0)
| S 2
| &5
^ UJ O
O 0) S
1_ C *-
o> O re
-Q jj fl)
E 18 ^
3 S £
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-18. Number of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute SO2 exposure above selected benchmark levels in St. Louis,
year 2002 air quality as /sand adjusted to just meeting the current and potential
alternative standards.
July 2009
251
-------
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
100% -F
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-19. Percent of all asthmatics (top) and asthmatic children (bottom) experiencing at least
one day with a 5-minute SO2 exposure above selected benchmark levels in St. Louis,
year 2002 air quality as is and adjusted to just meeting the current and potential
alternative standards.
July 2009
252
-------
*«
CO t.
Q) O
W Q_ *-
(A
re
Q
c
o
0)
Q.
« O
O 0)
x-2
LLI 0)
= 1
O ^
re o>
™ 0) .*,
0) — LLI
_i ^
** ^
re >
o fll —
o>
^2
E
s §
C 0)
0) _l
0)
Z °-
X
LLI
150 -©-99200
^^99250 ^^98200
-e-CS
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
1,000,000
Q)
-S c
1
> -c
~~ (D
LJ o > ra
O 0) -Q m
w 'C ^ ^
n & O ^.
100,000
10,000
1,000
E = O
iS x
^ LLI
O
100
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5- Minute Daily Maximum SO2 Exposure Level (ppb)
Figure 8-20. Number person days all asthmatics (top) and asthmatic children (bottom) experience
a 5-minute SO2 exposure above selected benchmark levels in St. Louis, year 2002 air
quality as is and adjusted to just meeting the current and potential alternative
standards.
July 2009
253
-------
Air Quality As Is
S! s?
3 ' — '
x E
UJ =
c a
I s
§•£
£S
uu%
90%
i
50%
30%
i
no/, .
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 — • — • •"
-T ;--;..,
*~ — •
k- - -i
' •
>--•
/
k
X
k-.J
h.-,
P — •
L=-M
i i i
i i i
i i i
i i i
—•—Indoors
• Outdoors
• *• Vehicles
i i i
i i i
i i i
i i i
i i i
i i
i i
L. -L -L A
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5-Minute Daily Maximum SO2 Exposure (ppb)
Air Quality Adjusted to Just Meet the Current Standard
100% -F
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5-Minute Daily Maximum SO2 Exposure (ppb)
Air Quality Adjusted to Just Meet the 99th %ile 150 ppb Standard
10% ---
0%
*»-—¥•
iiiir r r i"
—•—Indoors
• Outdoors
- A- Vehicles
I 1 J J J I L L L L I I
1 1_ —* '
t •» -t^ - T^ - -
II
100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
5-Minute Daily Maximum SO2 Exposure (ppb)
Figure 8-21. The frequency of estimated exposure level exceedances in indoor, outdoor, and
vehicle microenvironments given as is air quality (top), air quality adjusted to just
meeting the current standard (middle) and that adjusted to just meeting a 99th
percentile 1-hour daily maximum standard level of 150 ppb (bottom) in St. Louis.
July 2009
254
-------
th
Table 8-16. Absolute difference in APEX exposure estimates for St. Louis using either a 98 or
nth
99 percentile form potential alternative standard at a 1-hour daily maximum level of 200 ppb.
Population
All Asthmatics
(105,456)
Asthmatic
Children
(41,714)
Benchmark
Level (ppb)
100
200
300
400
100
200
300
400
Absolute Difference in Estimated
Exposures using 98th and 99th form1
Number of
Person-
days
91490
64531
31441
16705
69420
11682
3496
1449
Number
of
Persons
9142
22194
16922
11330
3826
4856
2425
1150
Percentage
Points2
8.7
6.7
3.2
1.5
9.2
11.6
5.8
2.8
Notes:
1 Both the 98th and 99th 1 -hour daily maximum air quality scenarios were
simulated by APEX, using a level of 200 ppb. The value reported is the
difference between the 98th and the 99th.
2 Difference between the percent of persons exposed (98th-200 minus the
99th-200) at each benchmark level.
8.10 REPRESENTATIVENESS OF EXPOSURE RESULTS
8.10.1 Introduction
Due to time and resource constraints the exposure assessment evaluating the current and
alternative standards was only applied to the two locations in Missouri. A natural question is
how might the estimates from this assessment of exposures in Greene County and St. Louis
compare with other areas in the United States that may have elevated short-term 862
concentrations. To address this question, additional data were compiled and analyzed to provide
context to the exposure modeling results. Because most estimated exceedances were associated
with the outdoor microenvironments, this analysis and discussion is centered on time spent
outdoors to allow for comparison of the two modeling domains with several other broad regions.
In addition, further context is given regarding the SC>2 emissions and air quality in these locations
with respect the 39 other counties evaluated in the air quality characterization. The distribution
of air conditioning and asthma prevalence rates in the U.S. U.S. and how that distribution
compares with those estimated for the two modeling domains is also discussed.
July 2009
255
-------
8.10.2 Time spent outdoors
The time spent outdoors by children age 5-17 was calculated from CHAD-Master74 for
five regions of the country. The U.S. states used in the air quality characterization (Chapter 7)
were of interest, which already includes Missouri (representing the two exposure modeling
domains). Staff analyzed the outdoor time by broad geographic regions because it was thought
that the regional climate would have influence on each population. In addition, most of the
location descriptors are already broadly defined to protect the identity of persons in CHAD; finer
spatial scale such as at a city-level is uncommon. Table 8-17 has the States used to identify
CHAD diaries available to populate a data set for each of the five regions. Staff further
separated the diaries by time-of-year (school year versus summer)75 and the day-of-week
(weekdays versus weekends), both important factors influencing time spent outdoors (Graham
and McCurdy, 2004). Summer days were not separated by day of week; staff assumed that the
variation in outdoor time during the summer would not be greatly influenced by this factor for
children. The results for time spent outdoors in each region are given in Table 8-18.
Table 8-17. States used to define five regions of the U.S. and characterize CHAD data diaries.
Region
Mid-Atlantic (MA)
Midwest (MW)
Northeast (NE)
Southeast (SE)
Southwest (SW)
States
New York, New Jersey, Delaware, Maryland, District of Columbia, Virginia,
Virginia, Pennsylvania
West
Ohio, Iowa, Missouri, Illinois, Indiana, Kentucky
Maine, New Hampshire, Vermont, Massachusetts, Connecticut, Rhode Island
North Carolina, South Carolina, Georgia, Florida, Tennessee, Alabama,
Mississippi, Arkansas, Louisiana
Nevada, Utah, Colorado, Arizona, New Mexico, Texas, Oklahoma
Participation rates for the selected time of year and day of week groupings were similar
for each of the regions. In general, a smaller percent of children spend time outdoors during the
school year (about 45-50%) compared to the summer (about 70-77%). There was no apparent
pattern in the day-of week participation rates considering the school year days. However,
children did spend more time outdoors on weekend days compared to weekdays at all percentiles
of the distribution and within all regions. In addition, children consistently spent more time
74Currently available through EPA at mccurdy.tom@epa.gov.
75A traditional school year was considered (months of September-May); summer months included June-August.
July 2009 256
-------
outdoors during summer days within all regions. There were few differences in outdoor time
when comparing each of the regions. Children in Northeastern States had the widest range in the
distributions for time spent outdoors. In this region of the U.S., children spent the least amount
of time outdoors during the school-year days-of-the-week and the greatest amount of time
outdoors on average during the summer. Based on this analysis, it is not expected that the results
generated for the two Missouri modeling domains would be largely different from results
generated in most areas of the U.S. when considering time spent outdoors, though there may be
differences in exposures estimated in Northeastern states.76 Depending on when the peak
exposure events occur in the year, the exposures estimated in these states may be lower or
higher.
Table 8-18. Time spent outdoors by geographic region for children ages 5-17 based on CHAD
time-location-activity diaries.
Region
MA
MW
NE
SE
SW
Time of
Year
school
summer
school
summer
school
summer
school
summer
school
summer
Day of
Week
weekdays
weekends
all
weekdays
weekends
all
weekdays
weekends
all
weekdays
weekends
all
weekdays
weekends
all
Doers1
(n)
400
317
474
336
258
154
70
54
23
641
593
244
253
232
273
(%)
45
43
71
42
41
71
48
43
77
49
52
70
46
50
76
TimeS
Mean
113
158
193
109
152
193
106
148
217
120
157
185
119
162
187
SD
97
159
140
92
131
180
89
128
148
98
126
147
106
142
137
Min
1
2
5
2
1
5
2
15
30
2
1
5
1
7
2
pent Outdoors (minutes)
Med
90
120
165
88
116
143
75
115
175
95
123
150
90
120
150
P95
301
365
462
300
422
565
290
480
465
325
404
480
315
405
450
Max
700
1440
1210
550
870
1250
335
574
635
555
810
935
650
1390
840
GM
73
105
146
73
102
131
66
105
172
84
112
135
80
116
136
GSD
3.0
2.7
2.3
2.7
2.7
2.6
3.1
2.4
2.1
2.6
2.5
2.4
2.8
2.4
2.5
Notes:
1 Doers are those engaged in the particular activity, in this case those children that had at least 1 minute of
outdoor time recorded in their CHAD time-location-activity diary. The participation rate (%) was estimated by
the total number of persons in each subgroup (not included). The n indicates the person-days of diaries used
to calculate the outdoor time statistics.
76 Note however that all of the Northeastern data have the fewest number of person days available, in particular the
summer days (n=23).
July 2009
257
-------
8.10.3 SOi Emissions and Ambient Concentrations
St. Louis was not one of the 40 selected counties for the Air Quality Characterization due
to its not meeting the selection criteria (see section 7.2.4.2). To provide additional perspective
on the exposure results for both the Greene County and St. Louis modeling domains, staff
compared the air quality in each of these locations with the other 39 counties, beginning with the
estimated number of benchmark exceedances using the available ambient monitoring data.77
Five-minute maximum SC>2 concentrations were estimated in St. Louis as was done with the
other 40 Counties (including Greene County) using the hourly ambient monitoring data (2001-
2006). Staff simulated all air quality scenarios (as is, current standard, potential alternative
standards) and estimated 5-minute maximum 862 concentrations using the statistical model.
Then, the mean number of days with a 5-minute maximum concentration above a benchmark
level in a year for St. Louis were combined with the exceedance results for the 40-counties and
ranked in descending order. In addition, two other rank statistics were generated; the average
total SC>2 emissions within 20 km of ambient monitors and the average population within 5km of
the ambient monitors, both statistics considering the 40 counties and St. Louis area. Each of the
two additional variables was also ranked in descending order.
Greene county estimated air quality exceedances rank within the upper quartile (i.e.,
having some of the highest estimated number of days with 5-minute benchmark exceedances) for
many alternative standard scenarios (Table 8-19). Most scenarios have exceedances ranked
within upper 50th percentile (including the as is scenario), while having the 37th highest ranked
emissions. The population ranking was moderate (19th of 41 locations). St. Louis air quality
exceedances rank within the 50th-75th percentile for most of the alternative standard scenarios,
with a few of the scenarios (e.g., the current standard, and the higher alternative standard) ranked
in the upper quartile, while having moderately ranked emissions (26th highest). The number of
days with benchmark exceedances for the as is scenario in St. Louis was ranked low in
comparison with the other 39 counties (approximately the 90th-95th percentile). The mean
estimated population surrounding the monitors is ranked in the upper quartile (9th of 41).
77 The exposure modeling domain was comprised of three counties (St. Charles, St. Louis, and St. Louis City), while
the available ambient monitoring data was only available for the latter two counties for years 2001-2006.
July 2009 258
-------
Table 8-19. Ranking of selected exposure locations using the modeled number of days with 5-
minute benchmark exceedances and the total emissions within 20 km of ambient monitors.
Exposure Modeling
Domain
Greene County, MO
Population -19th
Emissions - 37th
St. Louis, MO
Population - 9th
Emissions - 26th
Air Quality
Scenario
as is
Current Standard
99-50
99-100
99-150
99-200
99-250
98-200
as is
Current Standard
99-50
99-100
99-150
99-200
99-250
98-200
Benchmark Exceedance Rank (out of 41 )1
100 ppb
31
40
8
13
27
32
34
36
38
2
30
20
13
9
8
8
200 ppb
23
33
4
6
9
14
22
21
37
3
22.5
30
27
21
15
16
300 ppb
22
27
4
5
7
8
9
9
39
8
27
25
30
29
27
24
400 ppb
21
23
22.5
4
5
8
7
8
38.5
14
22.5
24
28.5
30
28
26
Notes:
1 Benchmark exceedances for the exposure modeling domains were compared with the 40
counties selected for the air quality characterization.
Given these ranked statistics and the results of the exposure assessment (i.e., St. Louis
had a much higher percent of asthmatics exposed above benchmark levels than Greene County),
the number and percent of persons exposed above benchmark levels are likely more a function of
the population density and where the persons reside, rather than just total SC>2 emission levels or
the number of air quality benchmark exceedances. In addition, total SC>2 emissions are not
necessarily a good indicator of estimated air quality exceedances. Greene County has a high
ranking for most of the air quality scenarios but only a moderate ranking for total emissions.
Ambient monitors with a high COV (>200%) account for the greatest number of days/year with
air quality benchmark exceedances. For example, in Gila County AZ, one of the two monitors in
the county had a high COV and was located within 2 km of primary smelter emissions. This
county ranked 1st in days/year with exceedances using as is air quality, though ranked only 36th
for SC>2 emissions (18,000 tpy). Figure 7-10 provided support for the variability bins selected
and their relationship with the number of measured air quality benchmark exceedances. Clearly,
ambient monitors with the greatest variability in 1-hour SC>2 concentration are the monitors most
likely to have 5-minute SC>2 benchmark exceedances.
July 2009
259
-------
Greene County was retained in the final exposure assessment based analyses in the 1st
draft SC>2 REA. At the time of the analysis, it was noted by staff that the county had a number of
ambient monitors available for use in calibrating the dispersion model (two of which were rated
as having high COVs), there were some measured benchmark exceedances using as is air
quality, and there was a moderate population density surrounding the monitors/source emissions.
However, based on the air quality characterization and exposure modeling performed here that
includes St. Louis, it appears that a less dense population surrounding the potentially important
SC>2 emission sources in Greene County primarily contributed to the resultant small percent of
asthmatics exposed. This is a common attribute noted at the high COV monitors; most of these
monitors are located in areas having low population density. Eighty-nine of the 809 monitors in
the broader SC>2 monitoring network were rated as having a high COV; 52 of these monitors
(58%) were associated with low population density (<10,000 persons within 5km), 28 moderate
population density (31%, 10,000-50,000 persons within 5km), and 9 high population density
(10%, >50,000 persons within 5km). It is possible that, in areas having several days/year with
air quality benchmark exceedances and a low to moderate population density, the exposure
results would be similar to that estimated for Greene County. For example, if an exposure
assessment was performed in Gila County AZ (ranked 1st in as is air quality benchmark
exceedances), it is possible that the percent of persons exposed would be low (ranked 38th in
population).
Staff also calculated the total SO2 emissions from marine vessels, generally referred to as
port emissions in this document. Using the data in the 2002 NEI, the total port emissions were
calculated for each of the 40 counties used in the air quality characterization and ranked (Table
8-20). The St. Louis modeling domain had the 5th highest total port SC>2 emissions when
considering the 40 counties, though these emissions only comprise 2% of the total SC>2 emissions
in St. Louis. Thirteen of the 40 counties did not have port emissions, one of which was Greene
County. The amount of port emissions in St. Louis was also compared with the top 40 counties
in the U.S that had the highest port emissions (Table 8-21). The total 862 emissions from ports
in St. Louis were ranked 28th, while seven counties had greater port emissions than Jefferson
County TX (one of the 40 counties included in the air quality characterization). Note that most
of the counties with the greatest port emissions were not evaluated in the air quality
characterization because they did not meet the high SC>2 concentration-based selection criterion.
July 2009 260
-------
Table 8-20. Total SO2 emissions and total port SO2 emissions in the St. Louis and the 40 Counties
used in the air quality characterization.
State
TX
PA
FL
NJ
MO
DE
NJ
TN
OH
NY
OH
IN
WV
Ml
WV
MO
PA
WV
NY
OK
IL
IA
TN
PA
WV
IN
NY
VA
IN
IN
PA
MO
NH
TN
AZ
IA
OH
MO
PA
VI
IL
County
Jefferson County
Allegheny County
Hillsborough County
Hudson County
St. Louis (3-County Area)
New Castle County
Union County
Shelby County
Cuyahoga County
Bronx County
Lake County
Lake County
Hancock County
Wayne County
Wayne County
Jefferson County
Beaver County
Brooke County
Erie County
Tulsa County
Madison County
Muscatine County
Blount County
Washington County
Monongalia County
Floyd County
Chautauqua County
Fairfax County
Gibson County
Vigo County
Northampton County
Iron County
Merrimack County
Sullivan County
Gila County
Linn County
Summit County
Greene County
Warren County
St Croix
Wabash County
SO2 Emissions1
Total
(tpy)
33,608
56,411
70,231
22,300
90,135
53,626
3,840
31,023
12,681
3,747
73,316
40,063
2,055
74,832
1,071
40,481
42,685
1,355
50,858
8,181
27,396
24,890
5,164
8,189
92,677
48,653
57,835
3,741
127,934
66,170
58,598
47,562
31,812
30,999
18,594
17,324
12,868
11,819
5,222
122
55
Ports
(tpy)
4,489
2,666
2,168
2,044
1,860
1,693
1,657
1,243
631
295
294
209
177
177
150
132
130
119
108
90
81
71
43
41
20
20
9
1
-
-
-
-
-
-
-
-
-
-
-
-
-
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
29
29
29
29
29
29
29
29
29
29
29
29
%of
Total
13.4%
4.7%
3.1%
9.2%
2.1%
3.2%
43.2%
4.0%
5.0%
7.9%
0.4%
0.5%
8.6%
0.2%
14.0%
0.3%
0.3%
8.7%
0.2%
1.1%
0.3%
0.3%
0.8%
0.5%
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%
Rank of
%
3
9
12
4
13
11
1
10
8
7
18
16
6
23
2
19
20
5
24
14
21
22
15
17
27
25
28
26
29
29
29
29
29
29
29
29
29
29
29
29
29
Notes:
1 SO2 emissions were calculated from the 2002 NEI. Emissions originating from ports were
calculated using SCC for marine vessels: 2280002100, 2280002200, 2280003100,
2280003200, 2282020005.
July 2009
261
-------
Table 8-21. The top 40 counties with the greatest total port SO2 emissions, including SO2
emissions from ports in the St. Louis modeling domain.
State
CA
LA
CA
TX
CA
MD
LA
TX
TX
LA
LA
TX
OR
LA
PA
AL
WV
CA
AK
NH
FL
NY
PA
NH
NH
MN
VA
MO
NY
CA
Ml
CA
DE
NH
CA
CA
TX
WA
NY
Ml
TN
County Name
Los Angeles
St. John the Baptist Parish
Santa Barbara County
Harris County
San Diego County
Baltimore City
Orleans Parish
Jefferson, Co
Nueces County
East Baton Rouge Parish
Iberville Parish
Galveston County
Multnomah County
Calcasieu Parish
Allegheny County
Mobile County
Cabell County
Ventura County
Valdez-Cordova
Cheshire County
Hillsborough County
Kings County
Philadelphia County
Stratford County
Hillsborough County
St. Louis County
Norfolk City
St. Louis 3-County Area
Richmond County
Orange County
Presque Isle County
Contra Costa County
New Castle County
Union County
Orange County
San Francisco County
Brazoria County
Clallam County
Queens County
Alger County
Shelby County
Port
Emissions
(tpy)
13,817
10,605
8,831
8,142
5,408
4,582
4,579
4,489
3,545
3,435
3,179
3,123
3,004
2,728
2,666
2,582
2,575
2,406
2,243
2,231
2,168
2,112
2,069
2,044
1,998
1,987
1,980
1,860
1,818
1,770
1,748
1,716
1,693
1,657
1,615
1,530
1,367
1,356
1,341
1,284
1,243
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Notes:
1 SO2 emissions were calculated from the 2002 NEI. Emissions originating from ports were calculated
using SCC for marine vessels: 2280002100, 2280002200, 2280003100, 2280003200, 2282020005.
July 2009
262
-------
Table 8-22. SO2 emission density the two exposure modeling domains and several counties
within selected U.S. Cities.
State
NY
OH
Ml
PA
IN
MO
PA
FL
NY
IL
TX
TX
GA
MA
MO
CA
City
New York
Cleveland
Detroit
Philadelphia
Gary
St. Louis
Pittsburgh
Tampa
Buffalo
Chicago
Beaumont-Port Arthur
Houston
Atlanta
Boston
Springfield
Los Angeles
FIPS1
36005
36047
36061
36081
36085
39035
39085
26163
42101
18089
29183,
29189,
29510
42003
12057
36029
17031
48245
48201
13067
13089
13121
13135
25017
25019
25021
29077
06037
County
Bronx
Kings
New York
Queens
Richmond
Cuyahoga
Lake
Wayne
Philadelphia
Lake County
St. Charles
St. Louis
St. Louis (city)
Allegheny
Hillsborough
Erie County
Cook
Jefferson
Harris
Cobb
DeKalb
Fulton
Gwinett
Middlesex
Norfolk
Suffolk
Greene County
Los Angeles
Total SO2
Emissions2
(tpy)
38,036
85,997
74,832
11,614
40,063
90,135
56,411
70,231
50,858
35,191
33,608
60,924
48,606
23,712
11,819
17,175
Land
Area3
(miles2)
303
686
614
135
497
1,130
730
1,051
1,044
946
904
1,729
1,570
1,282
675
4,061
Emission
Density
(tons/miles2)
125
125
122
86
81
80
77
67
49
37
37
35
31
19
18
4
Notes:
1 Federal Information Processing Standard Code
2 The emissions totals come from tier 1 data in the 2002 NEI (02ne_v3tier_summary_oct_15_2007.zip).
3 The county land area statistics come from the Census 2000 STF1 . Available at :
http://factfinder.census.gov/servlet/
Staff evaluated the emission density within the two exposure modeling domains and for
counties within several highly populated U.S. Cities. The emission density was calculated by
dividing the total emissions (tpy) by the physical area (mile2) of the location. These data are
presented in Table 8-22. Greene County (or Springfield, Mo.) has one of the lowest emission
densities, another attribute of the county that could have led to the few estimated number of
persons exposed above benchmark levels. On the other hand, St. Louis has a medium-to-high
July 2009
263
-------
emission density, likely one of the factors contributing to the much greater estimated numbers of
persons exposed above benchmark levels. The emission density in St. Louis is similar in
magnitude with counties in Philadelphia PA, Gary IN, and Pittsburgh PA, though much higher
than several counties within large U.S cities such as Atlanta, Boston, Chicago, Houston, and Los
Angeles. Three cities had a distinctly higher emission density than St. Louis: New York,
Cleveland, and Detroit. We note that four counties within these cities with the greatest emission
density were evaluated in the air quality characterization: the Bronx, Cuyahoga, Lake, and
Wayne.
In considering the air quality benchmark exceedance rankings of other counties combined
with their emissions and population density rankings, one could possibly argue for other
locations to conduct an exposure analysis that may provide different results for the as is air
quality scenario.78 Staff began assessing two additional locations for detailed exposure
modeling, i.e., Allegheny and Cuyahoga counties.79 Unresolved technical issues remained
regarding the agreement between dispersion-modeled and ambient measured concentrations,
preventing their inclusion in this final REA. The numbers of estimated air quality benchmark
exceedances in these two counties were ranked similarly to St. Louis (both counties were within
the 50th-75th percentiles). In addition, all of the monitors in Allegheny and Cuyahoga County
had at most moderately rated COVs (between 100-200%), suggesting that exposure results
estimated in those locations would be similar to that estimated in St. Louis. However, the high
emission density for Cuyahoga and Lake Counties (Cleveland) could indicate that a greater
number of persons might be exposed above benchmark levels when using the as is air quality.
While locations such as Los Angeles have greater estimated emissions originating from ports, the
SC>2 concentration levels measured at ambient monitoring data in these locations did not
approach the levels used for selection in the air quality characterization. In addition, the
emission density in Los Angeles County was the lowest of all of the cities selected for that
evaluation. Given each of the above rankings and available monitoring data, staff judges St.
Louis and Greene County as reasonable choices for the detailed exposure assessment,
particularly considering the range of air quality scenarios investigated.
78 For example, Hillsborough Fl. has a few bin C monitors, ranks 7th in population, 21st in emissions within 20 km of
monitors, 21st in countywide port emissions, and medium emission density.
79 Allegheny county ranked 10th in population, 31st in SO2 emissions within 20 km of monitors, and 23rd in
countywide port emissions. Cuyahoga county ranked 5th in population, 25th in SO2 emissions within 20 km of
monitors, though not ranked within the top 40 counties using port emissions.
July 2009 264
-------
8.10.4 American Housing Survey (AHS) Data
The American Housing Survey (AHS), conducted by the Bureau of the Census for the
Department of Housing and Urban Development (HUD), collects data on the nation's housing.
Relevant housing characteristic data, including residential prevalence of air conditioning are
summarized for 13 locations using the available metropolitan areas surveyed by the AHS (Table
8-23). Because survey years differ for each location and some locations contained more than one
survey, the most recent data or data closest to 2002 were selected (the year for the exposure
modeling). The A/C prevalence can vary greatly across urban areas, based largely on climate
differences. The air conditioning prevalence can influence the air exchange rate in a residence,
potentially affecting the infiltration rate of outdoor air concentrations into the indoors residential
microenvironment. St. Louis was estimated to have one of the highest air conditioning
prevalence rates, though similar rates could be found in Miami, Phoenix, Atlanta, and
Washington D.C. A few of the urban areas listed have much lower A/C prevalence rates,
including Los Angeles with 57.4% and Boston with 63.1%. For locations having a low A/C
prevalence, it is expected that the number of indoor residential exposures to daily maximum NC>2
concentrations above selected benchmarks would be greater compared to those estimated in St.
Louis. However, given the limited contribution of the indoor microenvironment to the number
of exceedances even considering much lower A/C prevalence rates (section 8.11.2.2.9; also EPA
2008d), modeled increases in the numbers of persons exposed in these other locations would
likely be small.
July 2009 265
-------
Table 8-23. Residential A/C prevalence for housing units in several metropolitan locations in the
U.S. (AHS, 2008).
Location
Atlanta
Boston
Chicago
Cleveland
Denver
Detroit
Los Angeles
Miami
New York
Philadelphia
Phoenix
St. Louis^
Washington DC
AHS
Survey
Year
2004
1998
2003
2004
2004
2003
2003
2002
2003
2003
2002
2004
1998
A/C
Prevalence1
(%)
97.2
63.1
89.6
75.8
66.9
82.4
57.4
98.1
83.3
91.4
94.4
96.7
96.0
Notes:
1 Represents the percent of total year-round housing
units having central or room unit air conditioners
(AHS, 2008).
2 Note, a truncated value of 96% was used as input
to APEX. The effect of this to estimated exposures is
negligible. See section 8.1 1 .2.2.9.
8.10.5 Asthma Prevalence
Staff compared regional asthma prevalence statistics for children <18 years in age and all
persons. For children, the estimated age-adjusted percents of ever having asthma are presented
in Table 8-24 using data from Dey et al. (2004). There are similar prevalence rates for asthmatic
children in three of the four regions of the U.S. (Midwest, South, and West), suggesting that
exposure analyses conducted in these broader regions may result in similar distributions in the
percent of asthmatics exposed to the two Missouri modeling domains used in this assessment.
The Northeastern U.S. has a higher percentage of asthmatic children. This suggests that there
may be a greater percentage of peak exposures to asthmatic children in the Northeast than
compared with the percent modeled in St. Louis or Greene County, holding all other influential
variables are constant (e.g., time spent outdoors, a similar air quality distribution).
Staff weighted the BRFSS 2002 state-level adult asthma prevalence rates (self-reported)
to generate prevalence rates for five U.S regions (Table 8-25).80 Similar rates (between 7.6-
1 http://www.cdc.gov/asthma/brfss/02/current/tableCl.htm. Regions were mapped using Table 8-12.
July 2009
266
-------
7.9%) were estimated for three of the five regions (Mid-Atlantic, Midwest, and the Southwest),
suggesting that exposure analyses conducted in these broader regions may result in similar
distributions in the percent of asthmatics exposed to the two Missouri modeling domains used in
this assessment. Consistent with that observed for asthmatic children, the Northeastern U.S. has
the greatest percent of asthmatic adults. The Southeastern states on average were estimated to
have the lowest adult asthma prevalence. This suggests that there may be a greater percentage of
peak exposures to asthmatic adults in the Northeast and a lower percentage of peak exposures in
the Southeast when compared with the percent modeled in St. Louis or Greene County, holding
all other influential variables are constant (e.g., time spent outdoors, a similar air quality
distribution).
Table 8-24. Asthma prevalence rates for children in four regions of the U.S.
Region
Northeast
Midwest
South
West
Asthma Prevalence1
(%)
15.2
11.6
11.9
11.1
Notes:
1 prevalence is based on the question, "Has
a doctor or other health professional ever
told you that [child's name] had asthma?'"
(Deyetal.,2004)
Table 8-25. Asthma prevalence rates for adults in five regions of the U.S.
Region1
Mid-Atlantic
Midwest
Northeast
Southeast
Southwest
Asthma Prevalence2
(%)
7.9
7.7
8.9
6.9
7.6
Notes:
1 Table 8-17 was used in mapping the states to regions.
2 state level data obtained from
http://www.cdc.qov/asthma/brfss/02/current/tableC1.htm.
July 2009
267
-------
8.11 VARIABILITY ANALYSIS AND UNCERTAINTY
CHARACTERIZATION
As discussed in section 6.6, there can be variability and uncertainty in risk and exposure
assessments. This section presents a summary and discussion of the degree to which variability
was incorporated in the exposure analyses and how the uncertainty was characterized for the
estimated number of persons and person days with exposure benchmark exceedances.
8.11.1 Variability Analysis
To the maximum extent possible given the data, time, and resources available for the
assessment, staff accounted for variability within the exposure modeling. APEX has been
designed to account for variability in nearly all of the input data, including the physiological
variables that are important inputs to determining exertion level. As a result, APEX addresses
much of the variability in exposure estimates given variability in factors that affect human
exposure. The variability accounted for in this analysis is summarized in Table 8-26.
Table 8-26. Summary of how variability was incorporated into the exposure assessment.
Component
Simulated
Individuals
Ambient Input
Physiological
Factors Relevant to
Ventilation Rate
Variability Source
Population data
Activity patterns
Block-level commuting
Employment
Modeled ambient SO2
concentrations
Meteorological data
Resting metabolic rate
Metabolic equivalents
by activity (METS)
Oxygen uptake per unit
of energy expended
Body mass
Comment
Individuals are randomly sampled from U.S. census
blocks used in model domains, by age and gender.
Data diaries are stratified from CHAD based on 30 day-
type (summer weekday, non-summer weekday,
weekend) and demographic group (males/females, ages
0-4,5-11, 12-17, 18-64,65+).
An individuals' commuting location is randomly sampled,
using adjusted U.S. census tract data that account for
fine-scale land use at the block level.
Work status is randomly generated from U.S. census
data at the tract-level by age and gender.
Spatial: modeled ambient SO2 to block-level receptors.
Temporal: 1-hour SO2for an entire year predicted using
AERMOD; 5-minute SO2 within each hour estimated
using APEX.
Spatial: Local surface and upper air NWS stations used.
Temporal: 1-hour NWS wind data for 2002
(supplemented by1 -minute ASOS data).
Six age-group and two gender-specific regression
equations using body mass as an independent variable
(Johnson etal., 2000).
Values randomly sampled from distributions developed
for specific activities (some age-specific) (EPA, 2002).
Values randomly sampled from a uniform distribution
(Johnson etal., 2000).
Values randomly sampled from lognormal distributions
by gender and age (Isaacs and Smith, 2005).
July 2009
268
-------
Component
Physical Factors
Relevant to
Microenvironmental
Concentrations
Variability Source
Body surface area
Height
Air exchange rates
Air conditioning
prevalence rates
Removal rates
Penetration factors
Comment
Gender specific exponential equations using body mass
as independent variable (Burmaster, 1998).
Separate regression equation for children and adults,
both gender and age-specific (4-groups); children use
age as an independent variable; adults use body weight
(Johnson etal., 2000).
Residential values randomly selected from lognormal
distributions, stratified by 4 temperature groups and
presence/absence of air conditioning. Other indoor
values randomly sampled from a separate lognormal
distribution.
Values randomly sampled AHS survey data for St.
Louis.
Values randomly selected for 5 microenvironment-
specific distributions, stratified by air conditioning usage.
Indoor/outdoor ratios randomly sampled from two
uniform distributions for inside-vehicle
microenvironments.
8.11.2 Uncertainty Characterization
The methods and the models used in this exposure assessment conform to the most
contemporary modeling methodologies available. A similar combined dispersion and exposure
modeling approach has been used recently in estimating human exposures for the NC>2 NAAQS
REA (EPA, 2008d). This increased level of complexity in the type and number of models used,
the overall exposure modeling approaches, and its application in exposure assessments does not
necessarily confer decreased levels of uncertainty. Staff believes however, that these types of
complex assessments serve as an important step towards raising the degree of confidence in
estimating exposures, particularly when the sources of uncertainty are systematically evaluated.
Following the same general approach described in sections 6.6 and 7.8 and adapted from
WHO (2008), staff performed a qualitative characterization of the components contributing to
uncertainty in the exposure results. First, staff identified the important uncertainties. Then, we
qualitatively characterized the magnitude (low, medium, and high) and direction of influence
(over, under, both, and unknown) the source of uncertainty may have on the estimated number of
persons and person days above benchmark levels. Finally, staff also qualitatively rated the
uncertainty in the knowledge-base regarding each source using low, medium, and high
categories. Even though uncertainties in AERMOD concentrations predictions are an APEX
input uncertainty, the uncertainties associated with each of the models are addressed separately
here for clarity. Table 8-27 summarizes the results of the qualitative uncertainty analysis
conducted by staff for the SO2 exposure assessment.
July 2009
269
-------
Table 8-27. Summary of qualitative uncertainty analysis for the exposure assessment.
Source
Type
Influence of Uncertainty
on Exposure Benchmark
Exceedances
Direction
Magnitude
Knowledge-
Base
Uncertainty
Comments
Algorithms
Unknown
Low
Low
INF & KB: Multiple historical model evaluations consistently demonstrate
unbiased ambient concentrations under variety of conditions. Some
potential dispersion scenarios may not be adequately represented and are
unknown as to how they apply in this application. However, model-to-
monitor comparisons in this application indicate very good agreement.
Meteorological
Data
Unknown
Low-
Medium
Low
AERMOD
Inputs and
Algorithms2
INF: A limited number of missing hours of wind data remain, potentially
leading to under-estimation. Model predictions have low to medium
sensitivity to surface roughness characteristics, as long as they are
appropriate for the site of the meteorological data inputs.
KB: Data are from a well-known and quality-assured source. One minute
ASOS wind data used to supplement 1-hour data for improved
completeness, reducing the number of calms and missing data.
Point Source
Emissions and
Profiles
Both
Low
Low
INF: Temporal emission characteristics are well represented for most
modeled point sources.
KB: Most temporal data are from a well-known quality-assured source of
direct measurements.
Area Source
Emissions and
Profiles
Both
Low -
Medium
High
INF: Temporal concentration characteristics were well represented when
using a generalized area source emission profile, i.e., an aggregate profile
covering a variety of emission source types. However, the temporal profile
selected can be very influential to 1-hour concentrations where area sources
are a significant contributor to emissions.
KB: While there were two alternative profiles available, one of which was
evaluated, a local generalized temporal emission profile was selected based
on yielding the best model-to-monitor agreement. It is largely unknown
whether the generalized profile is an appropriate representation of the true
temporal profiles that exist for modeled area sources.
APEX
Inputs and
Algorithms
AERMOD
Modeled 1-hour
Concentrations
Both
Low-
Medium
Medium
INF: Model-to-monitor comparisons indicated very good agreement. Most of
the overestimations in concentration occurred at the lowest 1-hour
concentrations (Figures 8-8 and 8-9), limiting the magnitude of influence on
estimated 5-minute concentrations. The spatial representation of ambient
concentrations using modeling is likely an improvement over using
concentrations from the limited number of ambient monitors.
KB: While model-to-monitor agreement was very good, it is unknown how
well all other modeled receptors are represented.
July 2009
270
-------
Accuracy of 5-
minute
Exposure
Estimation
Both
Low -
Medium
High
INF: The accuracy of the statistical model used in calculating 5-minute SO2
ambient concentrations was rated as having at most a medium level of
influence (see section 7.4.2.6 and Table 7-16).
KB: APEX annual average SO2 exposures are comparable reported
personal exposures of daily to multi-day averaging time. However, there are
no 5-minute SO2 personal exposure data that can be used to evaluate APEX
output.
Population
Database
Both
Low
Low
INF & KB: Data are from a reliable, quality assured source. Staff assumed
the limited uncertainty in the database would have negligible influence on
exposure results.
Commuting
Database and
Algorithm
Both
Low
Medium
INF: Most exposures above benchmark levels occur outdoors, not inside
vehicles. Also note there is limited modeled spatial heterogeneity in SO2
concentrations in St. Louis.
KB: Data are from a reliable, quality assured source. However land-use data
was used as a surrogate for distributing the tract-level commuting data to the
block-level.
Activity Pattern
Database
Over
Low-
Medium
Medium
INF: Most of the potentially influential factors are within the expected (or
assumed) bounds or are controlled for by the exposure modeling approach.
Though most components are rated as potentially having a low magnitude of
influence in either direction, not accounting for averting behavior by
asthmatics could result in a medium level of over-estimation.
KB: Data are from a reliable, quality assured source. Available published
literature was used for many of the comparisons, though some were limited
in direct correspondence and applicability.
Longitudinal
Profile
Algorithm
Both
Low-
Medium
Medium
INF: The magnitude of potential influence would be mostly directed toward
estimates of multiday exposures.
KB: Method compared reasonably well with available measurement data
and two other methods, however long-term (i.e., monthly, annual) diary
profiles do not exist for a population.
Meteorological
Data
Both
Low
Low
INF: Daily maximum temperatures are only used when selecting appropriate
diaries to simulate individuals and in selecting air exchange rate
distributions.
KB: Data are from a well-known and quality-assured source. One minute
ASOS wind data used to supplement 1-hour data for improved
completeness, reducing the number of calms and missing data.
Air Exchange
Rates
Under
Low -
Medium
Medium
INF: Most peak exposures occur outdoors, though indoor exposures may be
underestimated when not using all 5-minute concentrations within the hour
(section 8.11.2.2.11).
KB: Data used are not specific to St. Louis or Greene County Mo.
July 2009
271
-------
A/C Prevalence
Under
Low
Low
INF: Most peak exposures occur outdoors, though indoor exposures may be
underestimated when not using all 5-minute concentrations within the hour
(section 8.11.2.2.11). However a previous sensitivity analysis (EPA, 2008d)
indicates extremely low A/C prevalence has little influence on number and
percent of persons exposed.
KB: Data used are specific for St. Louis, there is limited variability in the
estimate, and compares reasonably with data from a different source.
Indoor
Removal Rate
Unknown
Low
Medium
INF: Most peak exposures occur outdoors, though indoor exposures may be
underestimated when not using all 5-minute concentrations within the hour
(section 8.11.2.2.11).
KB: Data used were obtained from comprehensive review of SO2 removal
rates, however many assumptions were needed in developing the removal
rate distributions.
Occurrence of
Multiple
Exceedances
Within an Hour
Under
Low-
Medium
Medium
INF: Analyses indicate that ignoring multiple peaks within the hour under-
estimates exposure and hence the number of persons exposed upwards to
35%.
KB: While the frequency of multiple exceedances within an hour can be
estimated, there are limited continuous 5-minute data available. The
representativeness of the available data to modeled receptors is unknown.
Asthma
Prevalence
Rate
Both
Low-
Medium
Low
INF: The percent of asthmatics for Greene county's simulated population
was similar to that of another independent estimate. County specific asthma
distributions were not used in St. Louis, there may be an over or under
estimate in the number of persons exposed.
KB: Data for asthma prevalence are from reliable and quality assured
sources.
Notes:
1 INF refers to comments associated with the influence rating; KB refers to comments associated with the knowledge-base rating.
2 The magnitude/direction of influence and the uncertainty associated with the knowledge-base for each source identified for AERMOD is characterized for
the predicted 1-hour concentrations, not the 5-minute benchmark exceedances.
July 2009
272
-------
8.11.2.1 Dispersion Modeling Uncertainties
Air quality data used in the exposure modeling was determined through use of EPA's
recommended regulatory air dispersion model, AERMOD (version 07026 (EPA, 2004a)), with
meteorological data and emissions data discussed above. Parameterization of meteorology and
emissions in the model were made in as accurate a manner as possible to ensure best
representation of air quality for exposure modeling. Thus, the resulting air quality values are
likely free of systematic errors to the best approximation available through application of
modeled data.
The characterization of uncertainty associated with this application of AERMOD is
separated into two main sources: 1) model algorithms, and 2) model inputs. While it is
convenient to discuss uncertainties in this context, it is also important to recognize that there is
some interdependence between the two in the sense that an increase in the complexity of model
algorithms may entail an increase in the potential uncertainty associated with model inputs. In
the characterization that follows, AERMOD uncertainties are discussed regarding the impact to
predicted 1-hour SO2 concentrations.
8.11.2.1.1 Algorithms
The AERMOD model was promulgated by EPA in 2006 as a "refined" dispersion model
for near-field applications (with plume transport distances nominally up to 50 kilometers), based
on a demonstration that the model produces largely unbiased estimates of ambient concentrations
across a range of source characteristics, as well as a wide range of meteorological conditions and
topographic settings (Perry, etal., 2005; EPA, 2003). While a majority of the 17 field study
databases used in evaluating the performance of AERMOD are associated with elevated plumes
from stationary sources (i.e., typically electrical generating units), a number of evaluations
included low-level releases. Moreover, the range of dispersion conditions represented by these
evaluation studies provides some confidence that the fundamental dispersion formulations within
the model will provide robust performance in other settings.
AERMOD is a steady-state, straight-line plume model, which implies limitations on the
model's ability to simulate certain aspects of plume dispersion. For example, AERMOD treats
each hour of simulation as independent, with no memory of plume impacts from one hour to the
next. As a result, AERMOD may not adequately treat dispersion under conditions of
July 2009 273
-------
atmospheric stagnation or recirculation when emissions may build up within a region over
several hours. This could lead to ambient concentration under-predictions by AERMOD during
such periods. On the other hand, AERMOD assumes that each plume may impact the entire
domain for each hour, regardless of whether the actual transport time for a particular source-
receptor combination exceeds an hour. This could lead to ambient concentration over-
predictions by AERMOD. While these assumptions imply some degree of physically unrealistic
behavior when considering the impacts of an individual plume simulation, their importance in
terms of overall uncertainty will vary depending upon the application. The degree of uncertainty
attributable to these basic model assumptions is likely to be more significant for individual
plume simulations than for a cumulative analysis based on a large inventory. This question
deserves further investigation to better define the limits and capabilities of a modeling system
such as AERMOD for large scale exposure assessments such as this. The evidence provided by
the model-to-monitor comparisons presented in section 8.4.5 is encouraging as to the viability of
the approach in this application when adequate meteorological and other inputs are available.
However, each modeling domain and inventory will present its own challenges and will require a
separate assessment based on the specifics of the application.
One of the improvements in the AERMOD model formulations relative to the Industrial
Source Complex - Short Term (ISCST) model which it replaced is a more refined treatment of
enhanced turbulence and other boundary layer processes associated with the nighttime heat
island influence in urban areas. The magnitude of the urban influence in AERMOD is scaled
based on the urban population specified by the user. Since the sensitivity of AERMOD model
concentrations to the user-specified population is roughly proportional to population to the l/4th
power, this is not a significant source of uncertainty. The population areas of interest for this
application are also well-defined, thus reducing any uncertainty associated with specification of
the population or with defining the extent of the modeling domain treated as urban.
Therefore, based on the evidence in historical and recent model evaluations and the
improved AERMOD model formulations, staff judges that algorithm uncertainty has a low
magnitude of influence on the estimated 1-hour concentrations. The direction of influence is
largely unknown, given the limitations in determining how the basic model assumptions apply to
a large-scale analysis. While the AERMOD model algorithms are not considered to be a
significant source of uncertainty for this assessment, the representativeness of modeled
July 2009 274
-------
concentrations for any application are strongly dependent on the quality and representativeness
of the model inputs. The main categories of model inputs that may contribute to uncertainty are
the meteorological input data and emissions estimates. These issues are addressed in the
following sections.
8.11.2.1.2 Meteorological Data
Details regarding the representativeness of the meteorological data inputs for AERMOD
are addressed separately in section 8.4.2 and in Attachment 1 in Appendix B. The data are from
a well-known, reliable source (NWS) and assumed vetted for extraordinary values by the
database architects and data users. Calm and missing 1-hour wind data have been supplemented
with 1-minute ASOS data averaged to the hour, decreasing the number of each within the input
data sets used. A limited number of missing values remained (1.1- 1.5%), however staff
expects these to have a negligible effect on the overall 1-hour concentration profile.
An important issue associated with representativeness is the sensitivity of the AERMOD
model to surface roughness, because the roughness at the location of the meteorological tower
site used to process the meteorological data for use in AERMOD may be very different from the
surface roughness across the full domain of sources. This issue has been shown to be more
significant for low-level sources due to the importance of mechanical shear-stress induced
turbulence on dispersion for such sources. A previous application of the AERMOD model to
support the REA for the NO2 NAAQS review (EPA, 2008d) provided an opportunity for a direct
assessment of this issue by comparing AERMOD modeled concentrations based on processed
meteorological data from the Atlanta Hartsfield airport (ATL) with concentrations based on
processed meteorological data from a Southeast Aerosol Research and Characterization study
(SEARCH) monitoring station located on Jefferson Street (1ST) near Georgia Tech. The ATL
data were representative of an open exposure, low roughness, site typical for an airport
meteorological station. The 1ST data were representative of a higher roughness exposure more
typical of many locations within an urban area. Surface roughness lengths were generally about
an order of magnitude higher at the 1ST site relative to the ATL site. A comparison of
AERMOD modeled concentrations for the mobile source NOX inventory, representing near
ground-level emissions, showed relatively good agreement in modeled concentrations based on
the two sets of meteorological inputs, at least for the peak of the concentration distribution at
four monitor locations across the modeling domain. This suggests that the sensitivity of
July 2009 275
-------
AERMOD model results to variations in surface roughness may be less significant than
commonly believed, provided that meteorological data inputs are processed with surface
characteristics appropriate for the meteorological site.
Therefore, based on the improved completeness of the wind data used and the low
sensitivity of peak model predictions to surface roughness characteristics, as long as they are
appropriate for the site of the meteorological data inputs, staff judges the potential magnitude of
influence from the meteorological data as low to medium. While it is possible that 1-hour
concentrations may be under-estimated based on missing wind data, it is largely unknown what
the overall direction of influence might be when considering the potential influence of other
meteorological parameters such as surface roughness.
8.11.2.1.3 Point Source Emissions and Profiles
As explained in section 8.4.3, point source emission levels were derived from the NEI
with source locations independently verified with GIS analysis of aerial photography. Temporal
profiles were derived from a variety of databases. Temporal profiles for all the modeled point
sources in Greene County and almost half of those in the St. Louis modeling domain were
derived from the CAMD database, which provides hourly emission profiles. For the remaining
modeled stacks inside the St. Louis domain, a uniform temporal profile was used. For most of
point sources located outside of the St. Louis domain but close enough to influence its air
quality, the temporal profiles were from the EMS-HAP emission model.
Therefore, given that the emissions data are from well-known quality-assured sources,
the emission source locations were independently verified, and that the temporal profiles for
most of the emission sources were known, staff judges the magnitude of influence from this
potential source of uncertainty as low and assumes there is an equal tendency to over- or under-
estimate 1-hour SC>2 concentrations. Further, staff also characterizes the knowledge-base for this
source as having a low level of uncertainty.
8.11.2.1.4 Area Source Emissions and Profiles
Details regarding the modeling of non-point and background area sources in AERMOD
were addressed in Section 8.4.3. In the case of SC>2, the area source emissions category for
AERMOD represents a cumulative approximation of several lesser point sources, such as small
commercial/industrial boilers, which are not represented as individual sources within the existing
emissions inventories due to their limited emissions. There is a lack of detailed information
July 2009 276
-------
regarding the location and release characteristics of these small emission sources, thus estimated
emissions are typically aggregated at a county level within the emission inventories. Given these
limitations in terms the emission inventory, two of the main uncertainties associated with
modeling these sources are the temporal and spatial profiles used in simulating their releases.
Lacking detailed location information, the emissions are assumed to be uniformly distributed
across a specified area, typically at a county or census tract level since the emissions are
aggregated at the county level, and spatially allocated using population as one of the surrogates.
An additional uncertainty associated with the area source category for 862 emissions is the
likelihood that the actual emissions may be associated with some plume buoyancy that cannot be
explicitly treated using the area source algorithm within the dispersion model. At best, the
anticipated aggregate effect of plume buoyancy can be reflected through the release height
assigned to the area source.
As discussed in Section 8.4.3, all emissions in the regions of interest were simulated,
either through their representative group (point sources, port-related sources, or other non-point
area sources) or through cumulative background sources. Staff obtained emission estimates from
the 2002 National Emissions Inventory (NET) however, only annual total emissions at the county
level are provided. To better parameterize these emissions for the hourly, census block-level
dispersion modeling conducted here, we relied on additional data and an algorithm to optimize
model performance based on available model-to-monitor comparisons.
Additional data related to the spatial distribution of non-point emissions was used to
spatially allocate county-wide emissions to census tracts in the Greene County domain. Staff
used the spatial allocation factors (SAFs), based on land use patterns, from EPA's EMS-HAP
database to allocate 87% of the non-point emissions to the subset of specific tracts expected to
contain the most emissions. Emissions within each modeled tract were simulated as uniform
over the tract, while emissions outside the modeled tracts and other residual emissions were
characterized as uniform over an entire county. The performance obtained by using tract-level
emission sources in Greene County was verified by model-to-monitor comparisons. In the St.
Louis area, model performance evaluations using factors from the EMS-HAP database made it
apparent that the spatial allocations were mischaracterized for this area. Thus, in the St. Louis
area, spatial bias was avoided by modeling non-point emissions with a uniform density
throughout each of the counties of interest instead of allocating emissions to specific census
July 2009 277
-------
tracts. In both cases, using spatially uniform emissions resolved to the tract or county level
improves spatial representation and reduces the overall level of uncertainty.
Unlike point sources, where the temporal profile was based largely on direct observations
via the CAMD database, these non-point emission profiles are based on generalized emissions
surrogates and may not well represent a specific source or local group of sources. Model
performance evaluations of diurnal profiles suggested that temporal factors derived from the
EMS-HAP emission model inadequately represented the true, aggregate, temporal release
profile.81 Unlike the spatial allocations, however, uniformly distributing the emissions in time
resulted in significantly worse model-to-monitor agreement than using these sample profiles. In
order to account for these uncertainties in the temporal profiles of area source emissions, an
algorithm was developed to determine the optimal temporal emission release profile in each area.
Examination of the diurnal profiles of modeled and monitored concentrations with uniform and
with EMS-HAP emission profiles for monitors in locations dominated by area sources showed
that, while monitored concentrations increased during the daytime, modeled concentrations
actually decreased. An examination of the dispersion characteristics showed that increased
dilution during the daytime overcame the small increase in emission strength predicted using the
EMS-HAP profile, which lacks local emission information. Thus, it is reasonable to conclude
that industrial and commercial/institutional area source emissions in the St. Louis and Greene
County areas would have a more pronounced diurnal cycle than is reflected in the EMS-HAP
temporal profile.
This method of determining an appropriate, local, non-point source emission profile has
the advantage of preserving total emissions reflected in the emission inventory while deducing
what the actual temporal emission profile from these local sources should be, based on the
observed trends in each region. Essentially, it derives an emission profile that best agrees with
observations when coupled with local meteorology and pollutant dispersion. This is justified
given the lack of detail regarding emission characteristics of local area sources. This derived
profile implies that the emission sources are active almost exclusively during the daytime from
approximately 8 am to 8pm. Given that the emission sources represented by the industrial and
commercial/institutional non-point category are small, the possibility that their cumulative
81 Figures 8-4 and 8-5 also show the corresponding temporal profile from the SMOKE emission model, which is
very similar to the temporal profile obtained from the EMS-HAP model.
July 2009 278
-------
emissions occur almost exclusively during daytime hours is plausible. However, in knowing that
there are large variations in the assumed local emission characteristics versus limited and broadly
defined emission characteristics for potential area sources, there is high level of uncertainty in
the knowledge-base. The selected approach though effectively mitigates the magnitude of
influence the uncertainty has on the modeling results by the application of a systematic approach
to minimize discrepancies between predicted and observed values. Based on the discussion
regarding the use of spatial allocation factors and the adjustments made to the area source
temporal profile, staff judges the magnitude of influence to range from low to medium.
8.11.2.2 Exposure Modeling Uncertainties
APEX is a powerful and flexible model that allows for the reasonable estimation of air
pollutant exposure to individuals. Since it is based on actual human time-location-activity
diaries and accounts for the most important variables known to affect exposure (i.e., where
people are located and what they are doing), it has the ability to effectively approximate actual
human exposure conditions. In addition, staff selected to the best available input data to
temporally and spatially represent the ambient concentrations and exposures given the time and
resources allocated for the assessment. However, there are constraints and uncertainties
associated with the input data and modeling approaches that may correspond to uncertainties in
the modeling results.
In the characterization that follows, exposure modeling uncertainties are discussed
regarding their influence to the estimated number of persons and person-days above benchmark
exceedances. Staff primarily focused on the uncertainties and assumptions associated with SO2
specific exposure model inputs, their utilization, and application in this exposure assessment.
Note also that some sensitivity analyses for certain components of APEX (see EPA, 2007d;
Langstaff, 2007) or input variables (EPA, 2008d) have been performed previously in other
NAAQS reviews. Those previous analyses that are relevant to the current SC>2 NAAQS review
are also included, though only summarized below.
8.11.2.2.1 AERMOD Modeled 1-hour Concentrations
The AERMOD model-to-monitor comparisons (section 8.4.5) indicated very good
agreement. Most over-estimations in 1-hour SC>2 concentrations occurred at the lowest 1-hour
concentrations, effectively limiting the potential magnitude of influence on estimated 5-minute
July 2009 279
-------
air quality and exposure concentrations. At the upper tails of the distribution (> 80th percentile),
there was a mixture of over- and under-estimation in 1-hour SC>2 concentrations, most of which
were on the order of 1-2 ppb. Staff performed an additional evaluation in Greene County to
compare estimated benchmark exceedances resultant from the variable concentration
distributions given by the ambient monitoring data and AERMOD predictions (rather than
simply comparing the 1-hour concentrations). The results indicated there was not a significant
influence to the estimated air quality benchmark exceedances from the limited differences
observed in the upper percentiles of the 1-hour concentration distributions.
Further, AERMOD was used in this exposure assessment to improve the spatial
representation of ambient concentrations given the limited number of ambient monitors in each
modeling domain. The dispersion modeling of SO2 concentrations to census block receptors is
judged by staff as improvement over using monitored concentrations alone as an input to APEX.
This may be of greater importance in Greene County where there was greater variability in the
modeled concentrations at the receptors surrounding each ambient monitor (see section 8.4.5).
In addition, the use of concentrations estimated at the census block centroids is judged by staff a
reasonable. This is because the centroids are not expected to be at systematically farther
distances from emission source than specific percentages of the population residing within the
census block.
Therefore, based on the above discussion, staff judges the potential magnitude of
influence from this source of uncertainty as low to medium, recognizing there could be some
conditions that would lead to over- or under-estimation of 5-minute SC>2 concentrations. While
there are limited differences in the modeled versus measured data, it is unknown how the model-
to-monitor agreement represents all other modeled receptors in the absence of additional ambient
monitoring data. Based on the discussion above regarding the current and historical AERMOD
performance evaluations (section 8.11.2.1.1), staff judges the knowledge-base as having a
medium level of uncertainty.
8.11.2.2.2 Accuracy of 5-minute Exposure Estimation
Uncertainties in the accuracy of the statistical model used in calculating 5-minute SO2
ambient concentrations was rated as having between low and medium levels of influence
(section 7.4.2.6 and Table 7-16). Staff assumes, because of the strong relationship between
ambient concentrations and personal exposures (in the absence of indoor sources), the same
July 2009 280
-------
influence rating would apply here with mainly limited opportunities for both over- and under-
estimation of 5-minute benchmark exceedances. This strong relationship between ambient
concentration and personal exposure though is noted as based solely on longer term averaging
times (single day to weeks in duration) and was discussed earlier in section 7.4.2.7.
Staff performed an additional qualitative analysis using the personal exposure
measurements reported in the ISA. As a default output from the APEX model, annual average
exposures were generated for each simulated individual (i.e., the full population rather than just
the identified subpopulation). Exposure results for the entire population (e.g., annual average
exposure concentrations) are assumed by staff as representative of exposures the asthmatic
population would receive because the asthmatic population should not have its
microenvironmental concentrations estimated any differently from those of the total
population.82
Selected percentiles of the distribution of annual average exposures for the APEX
simulated individuals is given in Table 8-28. Annual average AERMOD predicted ambient SC>2
concentrations were calculated for every receptor in the two modeling domains. The selected
percentiles of the distribution of annual average concentrations for the AERMOD predicted
ambient 862 is also given in Table 8-28. As expected, the APEX exposure concentrations are
consistently lower than the AERMOD predicted ambient concentrations. The relationship
between exposure and ambient, as determined by the ratio of the medians, are approximately
0.18 and 0.23 for St. Louis and Greene Counties, respectively. For general comparison, the
range of values developed from personal/ambient concentration linear regression slopes reported
by the ISA (ISA, section 2.3.6.2) is generally from 0.07 to 0.13. These measurement values
describing the relationship between personal exposure and ambient concentrations may be lower
than expected due to presence of personal exposure measurements below the limit of detection.
Note, the upper range (i.e., 0.13) was reported from a study containing the greatest percent of
samples above the limit of detection (ISA, section 2.3.6.2). We also lack information regarding
the value of the regression intercepts in these studies (i.e., if any were non-zero) to approximate
ratios that would be more comparable to the modeled values presented here.
For additional comparison, personal exposure measurements conducted in Baltimore,
Boston, and Steubenville are presented in Table 8-29 (see ISA Tables 2-14 and 2-15). While
82 Assumptions regarding activity patterns of asthmatics and non-asthmatics is discussed further in section 8.11.2.5
July 2009 281
-------
there are large differences in averaging time, sample size, study year, and city selected, the
personal exposure measurement concentrations compare well with selected percentiles of the
APEX exposure concentration distribution for the total simulated population in Greene County
and St. Louis.
Table 8-28. Distribution of APEX estimated annual average SO2 exposures for simulated
individuals in the Greene County and St. Louis modeling domains.
Annual Average
S02 (ppb)1
mean
std
pO
P1
p5
p10
p25
p50
p75
p90
p95
p99
p100
Greene County
(n=50,000)2
APEX - Exposure
0.4
0.2
0.1
0.1
0.2
0.2
0.3
0.4
0.5
0.6
0.6
0.8
1.1
AERMOD -
Ambient
2.0
1.5
0.1
0.2
0.2
0.3
0.6
1.6
3.1
4.2
4.7
5.5
6.0
St. Louis
(n=150,000)2
APEX-
Exposure
1.4
0.3
0.4
0.8
1.0
1.1
1.2
1.4
1.6
1.8
2.0
2.4
8.6
AERMOD -
Ambient
8.2
2.4
1.2
2.3
4.5
5.7
6.8
7.9
10.0
11.2
11.6
13.2
45.2
Notes:
1 mean is the arithmetic mean; std is the arithmetic standard deviation; percentile of the
distribution is given by number following "p" (e.g., p25 is the 25th percentile).
2 number of simulated individuals.
July 2009
282
-------
Table 8-29. Personal SO2 exposure measurement data from the extant literature.
Study1
City
Season
Averaging
Time
n3
Sarnat
(2000)
Baltimore
Winter
12 days
14
Sarnat
(2001 )2
Baltimore
Winter
1 day
45
Sarnat
(2005)
Boston
Summer
1 day
28
Sarnat
(2005)
Boston
Winter
1 day
29
Brauer
(1989)
Boston
Summer
1 day
48
Sarnat
(2006)
Steubenville
Summer
1 1 Weeks
10
Sarnat
(2006)
Steubenville
Winter
12 Weeks
10
SO2 Personal Exposures (ppb)4
mean
std
pO
P5
p10
p90
p95
p100
-
-
ND
-
-
-
-
1.2
-
-
-
ND
-
-
3
-
0.3-0.5
-
-
-
-
-
-
-
ND-1.9
-
-
-
-
-
-
-
-
-
-
-
0.4
1.8
-
-
1.5
3.3
-
-
-
-
-
-
0.7
1.9
-
-
-
-
-
-
Notes:
1 See ISA Tables 2-14 and 2-15 for further details regarding study conditions. Reference is provided here using
primary author and year of publication.
The cohort for Sarnat (2001) consisted of 15 seniors, 15 children, and 15 COPD patients. Seniors and COPD
patients had similar exposures, with children having somewhat higher exposure.
number persons in study.
4 mean is the arithmetic mean; std is the arithmetic standard deviation; percentile of the distribution is given by
number following "p" (e.g., p10 is the 10th percentile); ND is not detected.
APEX modeled exposures have previously been compared with personal exposure
measurements for O?, (EPA, 2007d). Briefly, APEX Os simulation results were compared with
weekly personal O?, concentration measurements for children ages 7-12 (Xue et al., 2005; Geyh
et al., 2000). Two separate areas of San Bernardino County were surveyed: urban Upland CA,
and the combined small mountain towns of Lake Arrowhead, Crestline, and Running Springs,
CA. Available ambient monitoring data for these locations were used as the air quality input to
APEX. APEX predicted personal exposures for both locations reasonably well for much of the
concentration distribution, but tended to underestimate exposures at the upper percentiles of the
distribution. The average difference between the weekly means was less than 1 ppb, with a
range of-11 ppb to 8 ppb, though predicted upper bounds for a few weeks with higher exposure
concentrations were under-predicted by up to 24 ppb (e.g., Figure 8-22). In addition, modeled
exposure concentration variability was less than that observed in the personal exposure
measurements. These differences appear to be driven by under-estimation of the spatial
variability of the outdoor concentrations (EPA, 2007d).
July 2009
283
-------
Q. ou
o
•jjj /n
jB 4U
5r on
s 20
C ,-,,=,- - p p; T
O n f. ! Li « 1 '. fa < ' Lj !.
3 ° T
l range
measured
.-7pgPEi*3
I
•L APEX
]
T ^- Ilirl'
» Measured « APEX
f T TT TI
Figure 8-22. Means of weekly average personal O3 exposures, measured and modeled (APEX),
Upland Ca. Figure obtained from EPA (2007d).
In addition, APEX modeled exposures have previously been compared with personal
NC>2 exposure measurements in Atlanta (EPA, 2008d). Daily personal NC>2 exposure
measurements were obtained from Suh (2008) for 30 participants of a 1999-2000 Atlanta
epidemiological study conducted by Wheeler et al. (2006) across two seasons.83 An exposure
distribution was constructed for each individual, simply using the individual's minimum,
median, and maximum daily mean exposures (e.g., Figure 8-23, top). Daily mean NC>2
exposures estimated using APEX were also evaluated in a similar manner, by stratifying the
results based on the same two seasons. The specific period from 1999-2000 was not modeled by
APEX; simulation results for year 2002 were used in the comparison. A distribution of each
person's estimated daily exposure was also constructed, using the median daily exposure to
represent the central tendency and a 95 % prediction interval to represent the lower and upper
bounds of exposure (e.g., Figure 8-23, bottom). The distributions of median daily exposures
compared better for the spring season, along with the range of estimated daily mean exposures
given by the 95% prediction interval. However, APEX estimated exposures were greater during
the fall. Median estimated daily exposures were consistently about 2 ppb higher than the
personal exposure measurements across most of the percentiles of the distribution, and the APEX
83 The minimum number of exposure measurements per subject was three days, the maximum was seven days. Fall
was designated for sample collection dates reported in the months of September, October, and November 1999;
Spring was designated where sample collection dates were reported in the months of April and May 2000. Only
personal NO2 from ambient sources are discussed here.
July 2009 284
-------
upper prediction intervals ranged consistently higher (between 10 and 40 ppb) compared with the
maximum personal exposure measurement day (between 10 and 20 ppb).84
Personal NO2 Exposure Measurements-Spring 2000
100 -
90 -
80 -
g 70-
0) f
t> (
0)
1501
f
2 40 -
| 30-
20 -
10 -
APEX Personal NO2 Exposures - Spring 2002
100 -H I I - 4
10
Daily Mean NO2 Exposure (ppb)
Figure 8-23. Daily average personal NO2 exposures, measured and modeled (APEX), Atlanta Ga.
Figure obtained from EPA (2008d).
It is encouraging that the APEX longer-term exposure estimates are comparable to
personal exposure measurements. When also noting that there is a strong relationship between
ambient 862 concentration and exposure, staff believes that the estimated numbers of days with
5-minute exposures above benchmark levels are also likely reasonable. However, without the
While a direct comparison of APEX estimated maximum daily exposure concentrations with the maximum
observed daily personal exposure concentrations is considered qualitative given the large discrepancy in sample
sizes and the difference in years compared, it should be noted that considering both seasons, approximately 99.1%
of APEX simulated persons had their estimated maximum daily exposure concentrations within the maximum
observed daily personal exposure measurement of 78.2 ppb.
July 2009
285
-------
availability of 5-minute personal exposure measurements that more closely represent the
modeled population, the level of uncertainty in the knowledge-base is judged as high.
8.11.2.2.3 Population Database
The population data are drawn from U.S. Census data from the year 2000. This is a high
quality data source for nationwide population data in the U.S., there is none considered as
complete and as appropriate for its application in our exposure assessment. As such, uncertainty
regarding the knowledge-base is considered low. The data do have some limitations. The
Census used random sampling techniques instead of attempting to reach all households in the
U.S., as it has in the past. While the sampling techniques are well established and trusted, they
may serve as a limited source of uncertainty in exposure results. The Census has a quality
section (http://www.census.gov/quality/) that discusses these and other issues with Census data.
It is likely the uncertainty in population representation within this data would not affect the
APEX exposure results in any particular direction, and given the use of randomly sampled
demographics to represent the simulated population, it is expected that the magnitude of
influence this source of uncertainty has on the exposure results is low.
8.11.2.2.4 Commuting Database and Algorithm
Commuting pattern data were also derived from the 2000 U.S. Census, again a well-
documented, quality-assured source. The data are used in addressing home-to-work travel,
certainly within the bounds of the objectives associated with the original data collection. Staff
had to make a few simplifying assumptions to allow for practical use of this database to reflect a
simulated individual's commute. First, there were a few commuter identifications that
necessitated a restriction of their movement from a home-block to a work-block. This is not to
suggest that they never travelled on roads, only that their home and work blocks were the same
and served as the only source of ambient concentration data for those individuals. Persons
restricted to a single block for ambient concentrations include the population not employed
outside the home, individuals indicated as commuting within their home-block, and individuals
that commute over 120 km a day. This could lead to either over- or under-estimations in
exposures if they were in fact to visit a block with either higher or lower SCh concentrations.
Given that the number of individuals who meet these conditions is likely a small fraction of the
total population, staff considers the magnitude of influence as low and associated with either
small over- or under-estimation of exposure benchmark exceedances.
July 2009 286
-------
Second, although several of the APEX microenvironments account for time spent in
travel, the travel is assumed to always occur in basically a composite of the home- and work-
blocks. No other provision is made for the possibility of passing through other census blocks
during travel. This could also contribute to either over- or under-estimating exposure
concentrations, dependent on the number of blocks the simulated individual would actually
traverse and the spatial variability of the concentration across different blocks. This could
potentially affect a large portion of the population, since we expect that at the block-level, many
persons would have a commute transect that included more than two blocks, although the actual
number of persons and the number of blocks per commute and the spatial variability across
blocks has not been directly quantified. In addition, the commuting route (i.e., which roads
individuals are traveling on during the commute) is not accounted for. From a practical
perspective though, if staff was to consider multi-block commuting in an exposure modeling
exercise, further complexity would need to be added to the modeling while also requiring
additional input data that is not readily available (e.g., commuting route data for simulated
individuals). These model adjustments would come with a number of additional uncertainties
and require additional time and resources not available for the assessment. Therefore, staff
elected to not account for multi-block commuting. Note however that the modeled spatial
variability within 4 km of ambient monitors in St. Louis was much less than that of the modeled
spatial variability Greene County, suggesting that ignoring multi-block commuting transects may
be of lesser importance in St. Louis.
Furthermore, the estimation of block-to-block commuter flows relied on the assumption
that the frequency of commuting to a workplace block within a tract is proportional to the
amount of commercial and industrial land in the block. This assumption could result in over-
estimating exposures if 1) the blocks with greater commercial/industrial land density also have
greater concentrations when compared with lower density commercial/industrial density blocks,
and 2) most persons commute to lower commercial/industrial density blocks. It should also be
noted that recent surveys, notably the National Household Transportation Survey (NHTS), have
found that most trips taken and most VMT accrued by households are non-work trips,
particularly social/recreational and shopping-related travel (Hu and Reuscher, 2004). In
addition, geographic differences in infrastructure could lead to differences in commuting method
that is not weighted by either the CHAD diaries or the Census commuter dataset. These
July 2009 287
-------
constitute non-quantified sources of uncertainty that are not addressed by the Census commuter
dataset.
Overall, in assessing the influence the commuting database and algorithm have on
estimated exposures above benchmark levels, staff judges the magnitude to be low even in
Greene County particularly since most benchmark exceedances occur outdoors and not inside
vehicles or indoor microenvironments. Even though staff judged the use of land-use is a
reasonable surrogate for identifying where people might work, staff believes that, in the absence
of block-to-block commuting information to further support this relationship, the uncertainty
regarding the knowledge-base is medium.
8.11.2.2.5 Activity Pattern Database
The CHAD time-location activity diaries used are the most comprehensive source of such
data and realistically represent where individuals are located and what they are doing. The
diaries are sequential records of each persons activities performed and microenvironments
visited. There are, however, uncertainties in the exposure results as a result of the CHAD diaries
used for simulating individuals. Specific elements of uncertainty include an evaluation of 1) the
representativeness of CHAD in reflecting recent human activity patterns, 2) the approach used to
allow for geographical representation of influential characteristics, 3) the similarities of
asthmatic and non-asthmatic activity patterns, and 4) response of asthmatics to air quality
notifications. Discussion regarding the use of individual CHAD diary days in developing
longitudinal profiles is presented in section 8.11.2.2.6.
First, a large percentage of the data used to generate the daily diaries were gathered from
survey studies conducted between 20 to 30 years ago. While the trends in people's daily
activities may not have changed much over the years, it is certainly possible that some
differences do exist such as the amount of time spent outdoors, time spent performing activities
at a particular level of exertion, and the microenvironments where moderate or greater exertion is
likely to occur. It would be extremely difficult to determine real differences in the distribution of
these factors that may influence SO2 exposure. For example, much of the data that is available to
test such differences is survey-based. The survey methods used to collect data are not entirely
consistent with one another and most of the studies collecting time-location-activity data did not
have exposure modeling objectives in their design (Graham and McCurdy, 2004). If one were to
test the hypothesis of no observed differences in time spent outdoors using historical and recent
July 2009 288
-------
data, it is likely significant effects would result from differences in survey methods or overall
study design rather than measurable changes in population activities. Staff assumed that if there
were a difference between the time spent outdoors (the most important microenvironment for
SC>2 exposures) for the simulated population and historical data diaries used to represent them,
the difference would be negligible. Therefore, staff judges the magnitude of influence on the
number of days with exposures above benchmark levels as low.
Second, CHAD is a collection of data from numerous activity pattern surveys, many
having differing data collection objectives. Some of the studies were single city surveys,
although a large portion of the data is from National surveys designed to be representative of the
U.S. population. In addition, study collection periods occur at different times of the year,
possibly resulting in seasonal variation not representative of the modeled locations.
Furthermore, the CHAD diaries selected by APEX to represent the Greene County and St. Louis
population are not necessarily from individuals residing in these cities, the State of Missouri, or
from the Midwest, albeit some of the diaries may be. Each of these factors could contribute to
uncertainty in the exposure results if there are location-specific characteristics of the CHAD
surveyed population that are distinct from those of the simulated population. However, a few of
the limitations associated with the use of diaries from different locations or seasons are corrected
by the sampling approaches used in the exposure modeling. For example, diaries used are
weighted by population demographics (i.e., U.S. census based age and gender distributions at the
modeled census block) and temperature is used as a classification variable to account for
expected differences in a location's climate and its effect on human activities.
A sensitivity analysis was recently performed to evaluate the effect that using different
CHAD studies has on APEX results for the recent 63 NAAQS review (see Langstaff (2007) and
EPA (2007d)). Briefly, O3 exposure results were generated using APEX with all of the CHAD
diaries and compared with results generated from running APEX using only the CHAD diaries
from the National Human Activity Pattern Study (NHAPS), a nationally representative study in
CHAD. There was good agreement between the APEX exposure results for the 12 metropolitan
areas evaluated (one of which was St. Louis), whether all of CHAD or only the NHAPS
component of CHAD is used. The absolute difference in percent of persons above a particular
concentration level ranged from -1% to about 4%, indicating that the exposure model results are
not being overly influenced by any single study in CHAD. It is likely that similar results would
July 2009 289
-------
be obtained here for 862 exposures. Therefore, staff judges the magnitude of influence from
using appropriately sampled CHAD diaries in representing the simulated population as low.
Third, due to limited number of CHAD diaries with health-specific information, all
diaries are assumed as appropriate for any simulated individual, provided they concur with age,
gender, temperature, and microenvironmental time selection criteria. In addition, data
summaries85 output from the current version of APEX could only be output for the entire
simulated population rather than the particular subpopulation. This is a reasonable modeling
assumption when considering the calculation of the microenvironmental concentrations, because
it is not expected that the asthmatic population would have microenvironmental concentrations
different from those of the total population. However, there is uncertainty in the use of all
CHAD diaries in simulating any individual without considering the health status of both the
surveyed population and the simulated population if in fact health status affects the activity
pattern of the simulated individual. In this exposure assessment it was shown that the most
important location for contacting the 5-minute peak concentration were outdoor
microenvironments. Therefore, if there is a difference in the time spent outdoors (e.g., total time,
time-of-day) and activities performed outdoors between asthmatics and healthy individuals, there
may be a greater impact to the estimated number of asthmatics exposed (and number of person
days) than if there were no difference.
Briefly, the assumption of modeling asthmatics similarly to healthy individuals (i.e.,
using the same time-location-activity profiles) is supported by the findings of van Gent et al.
(2007), at least when considering children 7-10 years in age. These researchers used three
different activity-level measurement techniques; an accelerometer recording 1-minute time
intervals, a written diary considering 15-minute time blocks, and a categorical scale of activity
level. Based on analysis of 5-days of monitoring, van Gent et al. (2007) showed no difference in
the activity data collection methods used as well as no difference between asthmatic children and
healthy children when comparing their respective activity levels. Contrary to this, an analysis of
2000 BRFSS data by Ford et al. (2003) indicated a statistically significant difference between the
percent of current asthmatics (30.9%) and non asthmatics (27.8%) characterized as inactive. In
addition, these researchers found significant differences in the percent of asthmatic (26.6%) and
85 For example, the time spent in microenvironments at or above a potential health effect benchmark level.
July 2009 290
-------
non-asthmatic (28.1%) adults achieving recommended levels of physical activity (i.e., either
moderate or greater activity levels).
Note though, the issue is not just outdoor time and activity levels, but the intersection of
the two that are of importance as well as recognizing the performance capabilities of persons
with asthma. A person's overall physical activity level is strongly linked with their time spent
outdoors and is considered an important correlate in encouraging increased physical activity
among children and adults alike (e.g., Sallis et al., 1998). In addition, introducing regular
exercise has been shown to improve physical fitness in asthmatic children, with statistically
significant increases in ventilation measures such as maximum minute ventilation rate (VEmax)
maximum oxygen uptake (VO2max) (e.g., van Vledhoven et al., 2001). Further, in other related
research, Santuz et al. (1997) indicated no statistically significant difference between asthmatic
and non asthmatic children when comparing maximum exercise performance levels, provided the
individuals were conditioned through habitual exercise. Thus it appears that asthmatics are
likely to perform activities at elevated levels and do so in outdoor microenvironments.
To support the assumption that there is no difference in CHAD activity patterns used to
represent the asthmatic population, staff compared the amount of time spent outdoors at elevated
activity levels obtained from three individual asthma studies with estimates of the same metric
using the CHAD database. In addition, some of the studies incorporated in CHAD reported
whether the individual was asthmatic, non-asthmatic, or not classified. Therefore, staff
categorized the data and results as such in this analysis. Table 8-30 summarizes data reported
from the three studies and results generated using CHAD data and the known health status.
When considering the three asthma studies, the amount of time spent outdoors at
moderate activity level ranges from a low of approximately 2% to a high of about 11% of
waking hours. The estimates of outdoor time associated with moderate activity level using
CHAD diaries fall within that range (i.e., between 6.5 and 7.5%) with small differences observed
between the CHAD asthmatic and CHAD non-asthmatic population. This limited comparison
indicates that the CHAD diaries may reasonably approximate the amount of time spent outdoors
at moderate activity levels. In addition, comparison of the CHAD asthmatic and non asthmatic
population supports the assumption that all CHAD diaries are appropriate in representing
asthmatic individuals, regardless of health status. However, the percent of outdoor time
associated with strenuous activities using the CHAD database was lower when compared with
July 2009 291
-------
the three asthma studies. It is difficult to judge whether the time spent outdoors at strenuous
activity levels is under-represented by CHAD or it is over-represented by the three asthma
studies.
Staff recognizes that there are a number of differences that exist among the three
asthmatic studies used along with the CHAD diary data that could contribute to variation in the
time spent outdoors at elevated activity levels. This would include: the diary/survey collection
methods used, the classification of activities performed and associated activity levels, the number
of study subjects, and sample selection methods. The particulars regarding how each of these
were addressed across the various studies is wide ranging and could potentially influence the
results. However, based on the comparable results observed in time spent outdoors at moderate
activity levels, staff judges the magnitude of influence as low with no apparent direction in over-
or under-estimation.
Table 8-30. Percent of waking hours spent outdoors at an elevated activity level.
Location
Time of Year
Population
n
Mean age
(min-max)
Activity Level
Moderate
Strenuous
EPRI (1988)'
Los Angeles
April
Asthmatic
52
-
EPRI (1992)^
Cincinnati
August
Asthmatic
136
26
(1-78)
Shamoo(1994)J
Los Angeles
Summer
Asthmatic
48
Winter
Asthmatic
45
33
(18-50)
CHAD"
All
Any
Asthmatic
1,475
23
(<0 - 99)
Not
Asthmatic
15,848
27
(<1 - 93)
Unknown
4,821
31
(<1 - 94)
Percent of Asthmatic Waking Hours Spent Outdoors at Given Activity Level
7
2.4
11
3.3
1.9
0.2
1.7
0.2
7.5
0.04
6.5
0.01
6.7
0.2
Notes:
1 Hour diary questionnaire form used for up to three activities per hour. Non-random sample of 26
mild/moderate, 26 moderate/severe asthmatics selected from voluntary clinical studies.
2 Hour diary questionnaire form used for up to three activities per hour. Random digit dialing and
multiplicity sampling used.
3 Number of minutes performing three self-rated activity levels for three locations per hour. Non-random
sample selected from voluntary clinical studies.
4 Combination of random and non random selection studies, national and city-specific, as well as varying
diary protocol (see Graham and McCurdy, 2004). Original CHAD database (n=22,968; EPA, 2002) was
screened for persons with no age (n=223) and no sleep (n=601) reported. Median METS values from
each activity-specific distribution were assigned to each person's activities. Moderate and vigorous
activity levels were selected based on activities having a METS value of 3 to <6 and >6, respectively.
Finally, there is also a possibility that information regarding bad air quality may affect
the activities performed by the asthmatic population. There has been research regarding averting
behavior, that is, there is a reduction in time spent outdoors when the individual is informed of
the potential for bad air quality days (e.g., Bresnahan, et al. 1997; Mansfield, 2005; KDEH,
July 2009
292
-------
2006; Wen et al., 2009). One study reviewed by staff reported no effect on outdoor time (e.g.,
Yen et. al. 2004). Of the limited studies reviewed by staff, most were focused on the population
response to ozone (or smog) air pollution alerts, EPAs Air Quality Index (AQI), or simply self-
perceived bad air quality.
In the most recent U.S. study conducted in six states,86 it was reported that approximately
25-30% of asthmatic adults altered their outdoor activity due to either perceived bad air quality
or media alerts, compared with about half as many (12%-16%) non asthmatics altering their
outdoor activities (Wen et al., 2009). The media alert response rate was requisite on awareness
of the bad air quality media alert for both children (Mansfield et al., 2005) and adults (KDEH,
2006; Wen, 2009). Parents of asthmatic children checked air quality alerts more frequently than
parents of non-asthmatic children and, though reported as statistically significant, only about
25% of parents of asthmatic children checked the air quality on a daily basis (Mansfield et. al.,
2005). Approximately half of asthmatic and non asthmatic adults were aware of the media alerts
(Wen et al., 2009), though among all adults living in the Kansas City MSA,87 the percent aware
is much greater (70%; KDEH, 2006). Of the persons that reported altering their outdoor
activities, approximately 60% did so three or fewer times per year.
If there is averting behavior by asthmatics in response to air pollution events, the degree
to which an asthmatic's SO2 exposure would be altered is highly uncertain. Staff acknowledges
that there may be fewer asthmatics exposed using APEX if accounting for averting behavior.
However, information missing from the published studies that are of importance include 1) the
amount outdoor time was reduced, 2) the time-of-day the outdoor time reduction occurred, 3) the
distinction between all outdoor activities or moderate or greater activities, 4) influence of asthma
severity on aversion rate, 4) the relationship between ozone air quality and the occurrence of
short-term SO2 pollution events modeled here. Given the above averting behavior statistics,
there could be at most a 30% over-estimation in the number of persons exposed (i.e., a medium
level), though the over-estimation is likely to be less given how the unknown conditions noted
above affect averting behavior.
86 The six states were Colorado, Florida, Indiana, Kansas, Massachusetts, Wisconsin.
87 Note that Kansas City is in close geographic proximity to both of the Mo. exposure modeling domains.
July 2009 293
-------
8.11.2.2.6 Longitudinal Profile Algorithm
Some of the surveys comprising CHAD collected only a single diary-day while others
collected several diary days per individual. In this exposure assessment, individuals are
simulated for an entire year. APEX creates the annual sequences of daily activities for a
simulated individual by sampling human activity data from more than one subject. Therefore,
each simulated person essentially becomes a composite of several actual people from within the
underlying activity data. Certain aspects of the personal profiles are held constant, though in
reality they may change as an individual ages (e.g., body mass). This is likely more important
for simulations with long timeframes (e.g., over a year or more), particularly when simulating
young children. The method used to link the individual activity diaries together could influence
the estimated number of persons exposed, although there would be greater uncertainty in
estimating multiple exposures per individual per year rather than single exposures per year. Note
however, estimating multiple exposures per individual was not a focus of the exposure
assessment.
In a prior analysis, staff evaluated the cluster algorithm used in constructing longitudinal
profiles against a sequence of available multiday diaries sets collected as part of the Harvard
Southern California Chronic Ozone Exposure Study (Xue et al., 2005; Geyh et al., 2000). Diary
data were collected from children between the ages 7 and 12 for six consecutive days/month for
an entire year. See Appendix B, Attachment 4 and 5 for details of the comparison. Briefly, the
activity pattern records were characterized according to time spent in each of five aggregate
microenvironments: indoors-home, indoors-school, indoors-other, outdoors, and in-transit. The
predicted value for each stratum was compared to the value for the corresponding stratum in the
actual diary data using a mean normalized bias statistic. The evaluation indicated the cluster
algorithm can replicate the observed sequential diary data, with some exceptions. The predicted
time-in-microenvironment averages matched well with the observed values. For combinations of
microenvironment/age/gender/season, the normalized bias ranges from -35% to +41%. Sixty
percent of the predicted averages have bias between -9% and +9%, and the mean bias across any
microenvironment ranges from -9% to +4%. Although on occasion there were large differences
in replicating variance across persons and within-person variance subsets, about two-thirds of the
predictions for each case were within 30% of the observed time spent in each microenvironment.
July 2009 294
-------
The longitudinal approach used in the exposure assessment was an intermediate between
random selection of diaries (a new diary used for every day for each person in the year) and
perfect correlation (same diary used for every day for each person in the year). The cluster
algorithm used here was also compared with two other algorithms; one that used random
sampling and the other employing diversity (D) and autocorrelation (^4) statistics (see EPA,
2007g for details on this latter algorithm). The number of persons with at least one or more
exposure to a given Os concentration was about 30% less when using the cluster algorithm than
when using random sampling, while the number of multiple exposures for those persons exposed
was greater using the cluster algorithm (by about 50%). The algorithm employing the D and A
statistics exhibited similar patterns, although were lower in magnitude when compared with
random sampling (about 5% fewer persons with one or more exposures, about 15% greater
multiple exposures). These exposure results using the cluster algorithm in APEX appeared to be
the result of a greater correlation of diaries selected in comparison with the other two algorithms.
This outcome conforms to an expectation of correlation between the daily activities of
individuals. While the evaluation was performed using 8-hour O?, as the exposure output, it is
expected that similar results would be obtained for 5-minute 862 exposures. That is, the
characteristics of the diaries that contribute greatly to any pollutant exposure above a given
threshold (e.g., time spent outdoors, vehicle driving time, time spent indoors) are likely a strong
component in developing each longitudinal profile. Given these results and that the REA is not
necessarily focused on health effects resulting from multiday exposures, staff judges the
longitudinal approach may have a low to medium magnitude of influence on estimated number
of persons exposed. When comparing the modeled profiles with the measurement data, there
was a balanced mix of over- and under-estimation of microenvironmental time. Therefore, the
direction of influence on the estimated number of persons exposed could be in either direction.
Uncertainty in the knowledge-base is rated as medium given the limited longitudinal
measurement data available for comparison.
8.11.2.2.7 Meteorological Data
Details regarding the representativeness of the meteorological data inputs for APEX are
addressed separately in section 8.4.2 and in Attachment 1 in Appendix B. In addition,
uncertainties associated with the data are discussed in section 8.11.2.1.2. Briefly, meteorological
data are taken directly from monitoring stations in the assessment areas. Staff assumed that
July 2009 295
-------
most of the data used are error free and have undergone required quality assurance review. One
strength of these data is that it is relatively easy to see significant errors if they appear in the data.
Because general climatic conditions are known for the simulated area, it would have been
apparent upon review if there were outliers in the dataset, and at this time none were identified.
If there were errors remaining in the data, it would be expected to be limited in extent and occur
randomly. In addition, to reduce the number of calms and missing winds in the 1-hour MET
data, archived one-minute winds for the ASOS stations in each model domain were used to
calculate hourly average wind speed and directions. This approach reduces the number of
estimated zero concentrations that would be output by AERMOD if not supplemented by the
additional wind data, thus preventing a downward bias in the predicted 1-hour 862
concentrations. Therefore, staff judges the MET data as having a low level of influence and
equally applied to either under- or over-estimation in the number of persons exposed.
There are some limitations in the use of the meteorological data in APEX. APEX only
uses the 1-hour daily maximum temperature in selecting an appropriate CHAD diary and indoor
microenvironment air exchange rate. Because the model does not represent hour-to-hour
variations in meteorological conditions throughout the day, there could be uncertainty in some of
the exposure estimates associated with indoor microenvironments (see the next section).
8.11.2.2.8 Air Exchange Rates (AER)
The residential air exchange rate (AER) distributions used to estimate indoor exposures
may contribute to uncertainty in the exposure results. Three components of the AER analyzed
previously by EPA (2007d) include 1) the extrapolation of air exchange rate distributions
between-CMSAs, 2) analysis of within-CMSA uncertainty due to sampling variation, and 3) the
uncertainty associated with estimating daily AER distributions from AER measurements with
different averaging times. The results of those previous investigations are briefly summarized
here. See Appendix B, Attachments 7 and 8 for details in the data used to generate the AER and
the sensitivity analyses performed. It should be recognized that in this assessment, the indoor
microenvironments have been shown to be largely unimportant in estimating exposure
exceedances. Note however, that in ignoring all twelve 5-minute concentrations, the influence of
the indoor-residential microenvironment may be under-estimated (section 8.11.2.2.11).
Extrapolation of AER among locations
July 2009 296
-------
Air exchange rate (AER) distributions were assigned in the APEX model, as described in
the indoors-residential microenvironment. Because location-specific AER data for St. Louis and
Greene County were not available and that there were no AER data from cities thought to have
similar influential characteristics affecting AER,88 staff constructed an aggregate distribution of
the available AER data from cities outside California to represent the distribution of AERs in St.
Louis and Greene County (see Appendix B, Attachment 7).
In the absence of location-specific data for the microenvironments modeled by APEX
within each model domain, only limited evaluations were performed. To assess the uncertainty
associated with deriving AERs from one city and applying those to another city, between-
location uncertainty was evaluated by examining the variation of the geometric means and
standard deviations across several cities and originating from several different studies. The
evaluation showed a relatively wide variation across different cities in their AER geometric
means and standard deviations, stratified by air-conditioning status, and temperature range. For
example, Figure 8-24 illustrates the GM and GSD of AERs estimated for several cities in the
U.S. where A/C was present and within the temperature range of 20-25 °C. The wide range in
GM and GSD pairs implies that the modeling results may be very different if the matching of
modeled location to a particular study location was changed. For example, the 862 exposure
estimates may be sensitive to use of an alternative distribution, say those in New York City,
compared with results generated using the aggregate non-California AER distributions. It is
possible though that the true distribution could be more similar to the selected distribution from
all non-California cities than that of the specific locations given the population of available AER
data. It is unclear as to the direction of influence given the limited number of data available for
comparison. It is likely that the impact to the number of exceedances is low, given that most of
the exceedances occurred outdoors for most of the air quality scenarios evaluated.
88 Such potential influential factors would include age, composition of housing stock, construction methods used,
and other meteorological variables not explicitly treated in the analysis, such as humidity and wind speed patterns.
July 2009 297
-------
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
.2 2.5-
o 7 0-
-------
(GSD). The analysis of the non-California city data used to represent Greene County and St.
Louis indicated that the GSD uncertainty for a given AER temperature group tended to have a
range within ±0.3 fitted GSD (hr"1), with smaller intervals surrounding the GM (i.e., about ±0.10
fitted GM (hr"1) (Figure 8-25). Broader ranges were generated from the bootstrap simulation for
AER distributions used for Greene County and St. Louis homes without A/C (Figure 8-26),
although both still within ±0.5 of the fitted GM and GSD values. Given the limited range in
GMs and GSDs, staff judges the magnitude of influence as low and mainly associated with both
under- and over estimation of indoor exposure concentrations. See Appendix B, Attachment 7
for further details.
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Outside California
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5 —
I 3.0-1
2.5 —
2.0—
0.5
1.0 1.5 2.0
Geometric Mean
2.5
•Bootstrapped Data ++-(Jriginal Data
Figure 8-25. Example of boot strap simulation results used in evaluating random sampling
variation of AER (h~1) distributions (data from cities outside California). Parameters of
the original distribution are given by the intersection of the two inner grid lines
July 2009
299
-------
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Outside California
Air Conditioner Type: No A/C
Temperature Range: >2() Degrees Celsius
4.0
3.5
3.0
S 2.0-
O
1.5-
1.0-
0.5
I
1.0
I
1.5
Geometric Mean
I
2.5
•••Bootstrapped Data ++-Oiiginal Data
Figure 8-26. Example of boot strap simulation results used in evaluating random
sampling variation of AER (h~1) distributions (data from cities outside California).
Parameters of the original distribution are given by the intersection of the two inner
grid lines
Variation in AER measurement averaging times
Although the averaging periods for the air exchange rates in the study data varied from
one day to seven days, the analyses did not take the measurement duration into account and
treated the data as if they were a set of statistically independent daily averages. To investigate
the uncertainty of this assumption, correlations between consecutive 24-hour air exchange rates
measured at the same house were investigated using data from the Research Triangle Park Panel
Study (Appendix B, Attachment 8). The results showed extremely strong correlations, providing
support for the simplified approach of treating multi-day averaging periods as if they were 24-
hour averages. Therefore, staff judges the magnitude of influence as low with unknown
direction on the number of persons exposed.
8.11.2.2.9 Air Conditioning Prevalence
Because the selection of an air exchange rate distribution is conditioned on the presence
or absence of an air-conditioner, the air conditioning status of the residential microenvironment
was simulated randomly using the probability that a residence has an air conditioner, i.e., the
residential air conditioner prevalence rate. For this study we used location-specific data for St.
Louis (AHS, 2005) and applied that data to Greene County as well. EPA (2007d) details the
July 2009
300
-------
specification of uncertainty estimates in the form of confidence intervals for the air conditioner
prevalence rate, and compares these with prevalence rates and confidence intervals developed
from the Residential Energy Consumption Survey (RECS) of 2001 for several aggregate
geographic subdivision (e.g., states, multi-state Census divisions and regions) (EIA, 2001).
Briefly, the A/C prevalence rates used for St. Louis were 96%, with reported standard
errors of 1.7% (AHS, 2003). Estimated 95% confidence intervals were also small and span
approximately 6.5 percentage points (AHS, 2003). The RECS prevalence estimate for Census
Divisions was 92% (ranging between 86.4% and 98.4%), while the Census Region prevalence
estimate was 83.6% (ranging between 80.0% and 87.2%). This suggests that the A/C prevalence
used, while likely being representative of a city in Missouri, may be over-estimated for non-
urban locations (such as Greene County).
Furthermore, a sensitivity analysis was performed using a low (55%) and high (97%) A/C
prevalence rates as input to APEX in an Atlanta, Ga. exposure assessment used for the recent
NC>2 NAAQS review (EPA, 2008d). Upper percentile benchmark exceedances were also of
interest in that exposure assessment, only the averaging time was 1-hour instead of 5-minutes
used here. Indoor microenvironments were also found in the NC>2 exposure assessment to be
unimportant in estimating exposure exceedances. Results from the sensitivity analysis indicated
that there was no difference in the percent of the asthmatic population with NC>2 exposure
benchmark exceedances with a decreased A/C prevalence. Only a few additional persons (about
100 out of a simulated population of 200,000) experienced exposures above exceedances when
using the lower A/C prevalence. Based on the above discussion, staff judges the magnitude of
influence to estimated exposures as low, particularly given that indoor exposures to
concentrations above the benchmark levels rarely occurs.
8.11.2.2.10 Indoor Removal Rate
There may be uncertainty in the exposure results when considering the estimated
parameters, the form (i.e., lognormal) and limits (limited by the bounds of the measurement data)
of the distribution used to represent indoor decay. The data used to develop the distribution were
obtained from a review of several studies that analyzed SC>2 removal for a variety of building
material surfaces (Grontoft and Raychaudhuri, 2004). Potential influential factors such as
humidity and air exchange rate were accounted for in developing and applying the removal
distributions within the indoor microenvironments. In addition, the distributions were based on a
July 2009 301
-------
large empirical database and likely well represent expected 862 removal within indoor
microenvironments.
However, several assumptions were made to characterize the materials used within a
simulated indoor microenvironment, some of which were data-based, others in the absence of
supporting data, were based solely on professional judgment (see Appendix B.4.1). Staff
performed a Monte Carlo simulation using the removal data and 1,000 simulated interior rooms
of buildings to generate a distribution of SC>2 removal rates, weighted by the approximated room
configurations and proportion of materials present. There are many assumptions staff made that
could be modified with newly available data, particularly where inputs were based on
professional judgment. It is largely unknown what the direction of influence is in the absence of
new or refined input data. While some of the assumptions used may be largely uncertain, the
magnitude of the influence is judged by staff as low given the relative contribution of the indoor
microenvironments to exposure concentrations above the potential health effect benchmark
levels.
8.11.2.2.11 Occurrence of Multiple Exceedances within an Hour
The statistical model described in section 7.2 was used within APEX to estimate a single
5-minute maximum SC>2 concentration for every hour. However, multiple short-term peak
concentrations above selected levels are possible within any hour. Analysis of the 5-minute
continuous monitoring data indicates that multiple occurrences of 5-minute concentrations above
the 100, 200, 300, and 400 ppb within the same hour can be common. Using the continuous
monitoring data obtained from years 1997-2007, multiple peak concentrations (i.e., 2 or more) at
or above 400 ppb within the same hour occurred with a 61% frequency (Table 8-31). The
frequency of multiple exceedances was similar for the lower 5-minute SC>2 concentration levels,
where 63, 56, and 53% of the time there were two or more exceedances within the same hour at
the 100, 200, and 300 ppb benchmark levels, respectively. These results may suggest that a
single peak approach (i.e., 24 peak concentrations per day) for estimating the number of persons
and days with 5-minute SO2 exposures as a surrogate for all possible peak exposure events may
lead to an under-estimate in the number of potential exposures.
July 2009 302
-------
Table 8-31. Number of multiple exceedances of potential health effect benchmark levels within an
hour.
Number of Exceedances of
5 -minute SO2 in 1-hour1
1
2
3
4
5
6
7
8
9
10
11
12
Total
Number of Hours with Multiple 5-minute SO2
> 100 ppb
1248
658
411
257
242
153
125
89
64
49
50
73
3419
> 200 ppb
267
122
78
35
28
25
14
11
6
6
3
5
600
> 300 ppb
76
31
21
10
6
4
5
2
3
1
0
1
160
> 400 ppb
26
20
7
5
4
1
1
1
1
1
0
0
67
Notes:
1 The analysis is based on the 16 monitors reporting all 5-minute SO2 concentrations in an
hour(n=3,328,725).
In using the data in Table 8-31 alone, the magnitude of the under-estimation may be
somewhat overstated however, particularly when considering the benchmark levels of 200, 300,
and 400 ppb. A detailed analysis of the multiple exceedances by each monitor indicated that one
of the monitors (ID 420070005) was highly influential in generating the values in Table 8-31,
contributing greatly to the multiple peak occurrences at the higher benchmark levels. This
Beaver Pa. urban-scale monitor is identified as population-based, within a rural setting, and
having agricultural land use (Appendix A). Five out of eight of the sources located within 20 km
of this monitor had SC>2 emissions <250 tpy, one smelter emitting about 7,000 tpy was within 2.5
km, and two power generating facilities located approximately 3.4 and 7.5 km from the monitor
had SC>2 emissions of 3,000 and 30,000 tpy, respectively. Of the number of hours having
multiple exceedances, monitor 420070005 contributed to 61, 73, and 80% of the hours with
multiple peaks >200, >300, and >400 ppb, respectively. Following removal of this monitor from
the full data set, the occurrence of multiple exceedances of each the 200, 300, and 400 ppb
benchmark lowered to approximately 40% of all hours having co-occurring peaks.
This suggests there would be increased uncertainty in the exposure results if the
continuous monitoring data were used to design an approach for estimating multiple exceedances
within an hour. These continuous monitoring data were available only from 16 ambient
July 2009
303
-------
monitors, each having a limited number of monitoring years. The analyses above indicated that
one of the monitors contributed to most of the hours with multiple peak concentrations. How
this one monitor (as well as any other monitor having multiple exceedances) reflects what may
occur at the APEX modeled receptors in St. Louis and Greene County (or other different
locations) is unknown. There is no simple extrapolation possible using the continuous
monitoring data because the time of the peak (and hence multiple peak) concentrations modeled
are not known with respect to the simulated individuals' time spent outdoors.
The PMR statistical model is based on both concentration and variability measures,
implemented by APEX in estimating a single maximum 5-minute SC>2 concentration for every
hour at every receptor. This is based on known concentration and variability relationships
described in section 7.2. While APEX can model all twelve 5-minute concentrations, staff chose
to normalize the eleven remaining 5-minute SC>2 concentrations within an hour to the 1-hour
mean concentration. This decision was based on the already large size of the air quality files
used (thousands of receptors across a year) that also required a time consuming post-processing
step prior to input in APEX and ultimately, the run time associated with the exposure model
simulations. Estimating the 5-minute maximum SC>2 concentrations and the other 11
concentrations within APEX was more efficient than pre-processing all twelve 5-minute SC>2
concentrations.
Having all eleven other 5-minute 862 concentrations normalized to the mean could result
in under-estimating the number of persons exposed. The exposure simulation could miss a
persons' exposure that might have occurred if in fact there are multiple peak concentrations
within the same hour (a likely event given the continuous monitoring data, roughly between 40-
60%). The CHAD time-location-activity diaries used in APEX are fixed, that is, the modeled
time spent outdoors is based on the actual time of day and amount of time recorded by the
surveyed individual. APEX models exposure on a minute-by-minute basis; if most persons
spend time outdoors for a short time (e.g., 5-minutes), then it is possible that persons are not
realistically encountering peak concentrations given the normalization of the eleven 5-minute
SO2 concentrations. Therefore, staff analyzed outdoor activities in the CHAD diaries used by
APEX to determine the duration of time spent outdoors for each outdoor event.
Figure 8-27 illustrates the distribution of time spent outdoors, given activity outdoor
events defined by clock-hour increments (already part of the CHAD design). Thirty-five percent
July 2009 304
-------
of all outdoor events are for the entire hour; if the event corresponds with the same hour as a
simulated peak concentration, there would be no under-estimation in exposure occurring during
these events. Therefore, occurrence of multiple peaks within an hour is potentially not an issue
for 35% of all exposure events that occur outdoors. However, at each of the other outdoor
events, there is a probability of under-estimating the exposure, given by the duration of the event
divided by 60 minutes. For example, approximately 15% of outdoor events were 30 minutes. If
these outdoor events occurred at the time where there was a second estimated peak concentration
in the same hour, there is a 50% chance that the exposure is missed. The probability of missing a
potential exposure increases with decreasing duration of the outdoor event and, given the data in
Figure 8-27, this could be a frequent occurrence (i.e., about 65% of outdoor events may have
some probability of missing an exposure). This analysis does not account for multiple outdoor
events that may increase an individual's chance of an exposure above a benchmark level,
regardless of the event duration. It also assumes the each of the outdoor events evaluated have
an equal probability of occurring at the time of the peak concentration, which may or may not be
the case. In addition, the outdoor time distribution is based on all of the CHAD diary days,
potentially not the same distribution of diaries that were used in the APEX exposure simulations.
A better method to determine the potential number of missing exposures is to model the
exposures using two input data sets: air quality with all continuous 5-minute measurements, and
air quality having the measured 5-minute maximum and the eleven other 5-minute
concentrations within the hour normalized to the 1-hour mean. Staff constructed a data set using
measurements from the continuous-5 ambient monitoring. While there were two monitors
reporting continuous 5-minute measurements in Greene County (monitor IDs 290770037 and
290770026), there were only two years with exceedances of the 200 ppb benchmark level, and
no exceedances of the 300 or 400 ppb benchmarks. To explore the maximum effect of multiple
peak concentrations within an hour, staff used two years of data from monitor ID 420070005,
noted above as having the greatest number of air quality benchmark exceedances in a year (years
2002 and 2005 were selected).
July 2009 305
-------
t 15-
53O
out doort i
Figure 8-27. Duration of time spent outdoors (in minutes) using all CHAD events
First, staff replaced missing concentrations (approximately 5% of each year) using the
time-of-day monthly averaged SC>2 concentration. This data set served as the multiple peak air
quality data set to be tested; all measured 5-minute concentrations were used as is. Next, staff
constructed a similar data set, only this second data set had the maximum measured 5-minute
concentration retained and all other eleven 5-minute concentrations within the hour were
normalized using the 1-hour mean. This single peak data set reflects what was being modeled by
APEX. Each of the data sets were used as the air quality input to an APEX simulation,
controlling for all model sampling, the algorithms used, microenvironments modeled, and
persons simulated. The only difference in the two runs was the air quality input. Fifty thousand
persons were simulated using APEX, 13% of which were asthmatic children. Figure 8-28
illustrates the percent of asthmatic children exposed to selected 5-minute maximum
concentrations for each of the two scenarios; a multiple peak scenario and a single maximum
peak concentration, using two site-years of continuous monitoring data with the greatest number
of benchmark exceedances. As expected, there are more asthmatic children exposed when
considering the occurrence of multiple peaks in an hour. The difference in the percent of
asthmatic children exposed at each of the benchmark levels is small, about 2-5 percentage points
July 2009
306
-------
differ between the two simulations. However, considering the percent difference in the numbers
of persons exposed at most of the benchmarks levels, the simulations using the single peak air
quality method had between 20-35% fewer persons exposed than the multiple peak simulation.
Similar results were generated in simulations using the site-year with the 2nd highest number of
exceedances only the under-estimation using the single peak method was about 15-30% (Figure
8-29). Based on these analyses, at most the estimated number of persons exposed in St. Louis
and Greene County may be under-estimated by 35% when using a single peak method. The
actual amount of under-estimation is likely smaller given that these results were generated using
site-years of monitoring data having the greatest numbers of exceedances and contributing
significantly to the high frequency of multiple peak exceedances.
The location where exposures occur may also be influenced by the presence or absence of
multiple peak concentrations. In particular, the modeled indoor 5-minute maximum
concentrations may be markedly diluted if the indoor air exchange rate is low and all eleven
other 5-minute values within the same hour are normalized to the 1-hour mean concentration.
APEX estimates all microenvironmental concentrations using a mass balance method for 5-
minute time-steps (equation 8-7) that accounts for estimated microenvironmental concentrations
from the previous time-step (EPA, 2009b). While dilution of the indoor air is not an unusual
circumstance considering the physical process modeled, it is possible that the number of
exposure events from indoor sources is under-estimated when the prior time-step concentration is
artificially reduced.
Staff evaluated the microenvironments where peak exposures occurred, by aggregating
the time 5-minute exposures occurred into three broad microenvironmental groups: indoors,
outdoors, and in-vehicles. A comparison of the APEX simulations using the two air quality
input simulations (i.e., multiple peak versus single peak, monitor 420070005 - year 2002) and
considering how often peak exposures occur indoors is presented in Figure 8-30. The
differences in the percent of indoor exposure exceedances are consistent with the design of the
model and the particular input data used. For exposures less than the 400 ppb level, a greater
percent of the overall exposures occur indoors using the single peak method than compared with
the multiple peak data set. For exposures at or above the 400 ppb level, a smaller percent of the
overall exposures occur indoors using the single peak method than compared with the multiple
peak data set. In fact, the multiple peak simulation had indoor peak exposures at levels not
July 2009 307
-------
observed using the single peak method. This is likely a function of the normalized
concentrations, that when used in the mass balance equation as the prior time-step
microenvironmental concentration, the microenvironmental concentration at time t is less than
what would be expected.
While this analysis and its findings are encouraging, context is needed to assign relevance
to the current exposure analyses in St. Louis and Greene County. As stated earlier, the data set
used had the greatest number of benchmark exceedances, designed by staff to observe the effect
that multiple peaks within the hour has on estimated exposures. The observed differences in the
contribution from the indoor microenvironment may be more appropriately applied in
discussions regarding air quality scenarios with high concentrations distributions (e.g., air quality
adjusted to just meeting the current standard, Figure 8-21). While the differences in the highest
benchmark exceedances are likely of greatest interest when investigating the possibility of
missing exposure events, it should be noted that the greatest proportion of all exposure events
still occur outdoors (in this simulation, >70% of exposures above 400 ppb occurred outdoors).
In addition, the differences observed at the lower benchmarks indicated the role of indoor
exposures was fairly similar. At most the difference was four percentage points, with the
multiple peak simulation having a consistently lower contribution of exceedances from indoor
exposures. Therefore, based on the above discussion, staff judges the magnitude of the potential
under-estimation as low to medium.
July 2009 308
-------
80%
5
i 15°%
114°%
| .3 30%
I I
^ jj[ 20%
"5 x
^ 10% +
0%
^•5-minute Max with Normalized 11
CZI Multiple Peak Data Set
Percent Difference in Exposures
Jl—-- V
I I I d~l dl d~l d"l -n -rn
0%
-5%
-10%
I
x
UJ
-15% So
w .E
0) CO
-20% Q- w
•E c
9
-------
ou -
(/)
8
•a
c oc;
i/>
8
1 9n
X
LLI
«<1 C
D
I/)
a
.^ -in
•5
S1
£ =;
1
LL.
n
-i
-,
• 5-minute Max with Normalized 11
D Multiple Peak Data Set
f
\ llrfl_n „
o
o
o
LO
o
o
CM
O
LO
CM
O
O
CO
O
LO
CO
O
O
o
o
LO
o
o
CD
o
LT>
CD
O
O
O
LT>
o
o
oo
SO2 Exposure Level (ppb)
Figure 8-30. Frequency of exposure exceedances indoors for two APEX simulations: one using
multiple peak concentrations in an hour, the other assuming a single peak
concentration. Continuous 5-minute monitoring data (ID 42007005, year 2002) were
used as the air quality input.
8.11.2.2.12 Asthma Prevalence Rate
The best estimate of asthma prevalence used in this analysis was generated using a
comprehensive and widely used data set (CDC, 2007). Staff judged that variability in the asthma
prevalence based on age was an important attribute to represent in simulating SC>2 exposures, one
of the principal reasons for selection of the particular data set. There are however limitations in
using the data that may add to uncertainty in the generated exposure results. The percent of
asthmatics simulated by APEX using a combined regional (children by age) and local (adults all
ages) prevalence was comparable with an independent estimate of the percent of asthmatics
within the four counties modeled (9.3% versus 8.8% of the population, respectively). Therefore,
the uncertainty in the overall total percent of asthmatics exposed is likely low, particularly in
Greene County. In Greene County, 9.8% of the simulated population was asthmatic and
compares well with the 10.2% asthma prevalence reported by MO DOH (2003). However, the
asthma prevalence across the three-county domain in St. Louis was variable, with St. Louis City
County having a high estimated prevalence rate (16.4%) and St Louis County having a much
lower prevalence rate (5.8%). This variable distribution was not represented in the exposure
July 2009
310
-------
modeling simulation; all children and adults in each of the counties used the data summarized in
Table 8-7. Therefore in St. Louis City County, the asthma prevalence may have been under-
estimated, while in St. Louis County the asthma prevalence may have been over-estimated. This
may add to medium level of influence to the total number of asthmatics exposed in St. Louis (not
the percent of asthmatics exposed), though the direction of influence is largely unknown because
individual county level exposures are not output by the model.
8.12 KEY OBSERVATIONS
Presented below are key observations resulting from the exposure assessment:
• 5-minute exposures to 862 were estimated for two areas in Missouri (i.e., Greene County
and St. Louis), with both locations having significant SC>2 emission sources. Air quality
scenarios investigated by staff included as is air quality, air quality adjusted to simulate
just meeting the current annual and 24-hour 862 standards, and just meeting several
alternative 1-hour daily maximum 862 standards.
• A number of factors would be expected to contribute to differences in 862 exposures
across different locations. These include differences such as population density, SO2
emission density, location and types of SO2 sources, prevalence of air conditioning, time
spent outdoors, and asthma prevalence (section 8.10). As discussed in section 8.10, St.
Louis County has a medium-to-high 862 emissions density and a medium-to-high
population density relative to other urban areas. Relative to the St. Louis study area,
Greene County is a more rural county having much lower population density and much
lower SC>2 emissions density. Taken together, the estimated exposures for these two
locations provide useful insights about urban and rural counties with 862 emission
sources.
• St. Louis had both a greater number and percent of asthmatic children and adults exposed
above the benchmark levels than did Greene County for all air quality scenarios. This is
not unexpected given the greater population density and the much greater 862 emissions
density in St. Louis. Staff believes that the St. Louis exposure estimates provide a useful
perspective on the likely overall magnitude and pattern of exposures associated with
various SC>2 air quality scenarios in urban areas within the U.S. that have similar
population densities, 862 emissions densities, and asthma prevalence. Similarly, staff
believes that the results for Greene County provide perspective on exposures in more
rural areas within the U.S. that have similar emission and population attributes to Greene
County.
• Modeled concentrations are reasonable given comparisons to available measurement data
- AERMOD 1-hour SC>2 concentrations at ambient monitoring receptors and their
associated prediction envelops generally replicate and encompass those measured at the
ambient monitor. Model-to-monitor agreement was better in St. Louis than in Greene
County.
July 2009 311
-------
The degree of under- or over-estimation of 1-hour SC>2 concentrations by AERMOD at
ambient monitoring locations in Greene County did not appreciably affect the estimated
number of days per year with 5-minute concentrations above benchmark levels.
APEX-modeled annual mean SC>2 exposures in St. Louis and Green County (arithmetic
means, 0.5-1.4 ppb) are comparable to daily and weekly personal exposure
measurements in other locations (arithmetic means, 0.3-1.9 ppb).
Estimated exposures above 5-minute potential health effect benchmark levels at moderate
or greater exertion using APEX occurred most frequently outdoors (around 50 to >90%,
depending on the air quality scenario and modeling domain).
Simulating air quality that just meets the current annual standard resulted in the greatest
number and percent of asthmatic persons exposed at all benchmark levels. The value
depended on both the benchmark level and modeling domain. For example, the percent
of asthmatic children exposed at least one day above a benchmark concentration ranged
from 0% (400 ppb benchmark) to 8% (100 ppb benchmark) in Greene County, while in
St. Louis the corresponding range was 24% to 97 %.
The exposure results using as is air quality were similar to that estimated using air quality
adjusted to a 99th percentile 1-hour daily maximum of 50 or 100 ppb in Greene County,
though in each of these scenarios, there were only a few persons exposed. In St. Louis,
the estimated exposure associated with as is air quality was also between that estimated
by simulating the 50 and 100 ppb 99th percentile 1-hour daily maximum air quality
scenario.
Staff compared exposure results using the 50 ppb 99th percentile air quality scenario
relative to as is air quality in St. Louis to estimate the reduction in the number and
percent of asthmatic children exposed above each 5-minute health effect benchmark
level. No asthmatic children were exposed above the 400 ppb 5-minute benchmark for
either the as is or 50 ppb 99th percentile alternative standard scenario. There were 121
fewer asthmatic children exposed above the 200 ppb 5-minute benchmark, corresponding
to a 76% reduction in exposures, when considering the 50 ppb standard level. Similarly,
reductions also were observed at the 100 ppb 5-minute benchmark when considering the
50 ppb standard compared with as is air quality: 1,641 (59%) fewer asthmatic children
were exposed. (Appendix B.4).
In both St. Louis and Greene County, there were no reductions in the numbers or percent
of persons exposed at any of the 5-minute benchmark levels when comparing exposure
results using the 100 ppb 99th percentile air quality standard scenario relative to as is air
quality.
Using a 99th versus a 98th percentile form at the same standard level (i.e., 200 ppb)
resulted in fewer persons being exposed above benchmark levels when using the 99th
percentile. Approximately 1,000 to 5,000 fewer asthmatic children, 1,000 to 90,000
fewer person days, and 2 to 12 fewer percent of persons were exposed above benchmark
levels in St. Louis.
Of the fifteen uncertainties qualitatively judged to influence the estimated number of
persons with at least one exposure above the 5-minute SC>2 benchmark levels, one may be
associated with over-estimation, three could result in under-estimations, while the
July 2009 312
-------
remaining uncertainties could affect exposure results in both (nine sources) or unknown
direction (two sources) (see Table 8-27). Nine of these eleven sources with bidirectional
influence were rated by staff as being low-medium magnitude of influence. The
magnitude of influence for three of the four uncertainties associated with either over- or
under-estimation was estimated as being low to medium influence, while the remaining
source (i.e., A/C prevalence) was ranked as being low or a negligible magnitude of
influence. Two of these four sources of uncertainty (i.e., A/C prevalence and indoor
AERs) were parameters used to estimate indoor exposures, which staff believes do not
contribute significantly to exposures above benchmark levels. The remaining two
sources (i.e., uncertainty in the activity pattern database used and the occurrence of
multiple exceedances within an hour) could have an offsetting influence in estimating the
number of persons exposed. This is because both of these sources were rated by staff as
being low to medium in magnitude, though in opposing direction. Based on this overall
characterization related to the direction and magnitude of influence for identified sources
of uncertainty, we are unable to characterize the likelihood of the estimates being either
over- or under-estimated with respect to the number of persons exposed above
benchmark levels.
• The knowledge-base uncertainty for sources with unknown or bidirectional influence
ranged from low (five sources) to medium (four sources). Note that most of these
sources were rated above as being of low-medium magnitude of influence. A high
degree of uncertainty in the knowledge-base was assigned to two sources: the area source
emission profile (direction of influence characterized as both, with low-medium rated
magnitude) and the accuracy of 5-minute exposures estimated by APEX (direction of
influence characterized as both, with low-medium rated magnitude). The knowledge-
base uncertainty was medium for three of the four sources identified above that were
associated with either under- or over-estimating 5-minute exposures (the remaining
source was rated as low). Based on this overall characterization, there is a low-medium
level of uncertainty in the knowledge-base for most sources. While two sources were
rated as having high knowledge-base uncertainty, they were noted as having similar
magnitude of influence on the estimated 1-hour or 5-minute concentrations.
July 2009 313
-------
9. HEALTH RISK ASSESSMENT FOR LUNG FUNCTION
RESPONSES IN ASTHMATICS ASSOCIATED WITH 5-MINUTE
PEAK EXPOSURES
9.1 INTRODUCTION
In the previous review, it was clearly established that subjects with asthma are more
sensitive to the respiratory effects of SC>2 exposure than healthy individuals (ISA, section
3.1.3.2). As discussed above in section 4.2, asthmatics exposed to SC>2 concentrations as low as
200-300 ppb for 5-10 minutes during exercise have been shown to experience moderate or
greater bronchoconstriction, measured as an increase in sRaw (>100%) or decrease in FEVi
(>15%) after correction for exercise-induced responses in clean air. These studies exposed
asthmatic volunteers to 862 in the absence of other pollutants that often confound associations in
the epidemiological literature. Therefore, these controlled human exposure studies provide
direct evidence of a causal relationship between exposure to SO2 and respiratory health effects.
Staff judges the controlled human exposure evidence presented in the ISA with respect to lung
function effects in exercising asthmatic subjects as providing an appropriate basis for conducting
a quantitative risk assessment for this health endpoint and exposure scenario.
As described in Chapters 5 and 6, staff is utilizing both the epidemiological evidence in
the ISA, and an air quality analysis based on U.S. and Canadian ED visit and hospitalization
studies for all respiratory causes and asthma to qualitatively inform: (1) the selection of potential
1-hour daily maximum alternative standards to be analyzed in the air quality, exposure, and risk
chapters of this document (see Chapter 5), and (2) the adequacy of the current, and potential
alternative standards (Chapter 10). However, for the reasons discussed in more detail in section
6.1, staff did not find the overall breadth of the epidemiological evidence to be robust enough to
support a quantitative assessment of risk.
A brief description of the approach used to conduct this health risk assessment is
presented below. More detailed discussion of the approach can be found in the risk assessment
technical support document, prepared by Abt Associates, which is included as Appendix C to
this document. The goals of this 862 risk assessment are: (1) to develop health risk estimates of
the number and percent of the asthmatic population that would experience moderate or greater
lung function decrements in response to 5-minute daily maximum peak exposures while engaged
in moderate or greater exertion for several air quality scenarios (described below); (2) to develop
July 2009 314
-------
a better understanding of the influence of various inputs and assumptions on these risk estimates;
and (3) to gain insights into the risk levels and patterns of risk reductions associated with
meeting several alternative 1-hour daily maximum SO2 standards. Health risks for lung function
effects in exercising asthmatics have been estimated for the following three scenarios: (1) "as is"
ambient levels of SC>2, (2) air quality adjusted to simulate just meeting the current 24-hour
standard, and (3) air quality adjusted to simulate just meeting several alternative 1-hour
standards.
As discussed in Chapter 8, the geographic scope of the assessment includes selected
locations encompassing a variety of 862 emission source types in two areas within the state of
Missouri (i.e., Greene County and St. Louis). These areas were identified based on the results of
a preliminary screening of the 5-minute ambient SO2 monitoring data that were available. The
state of Missouri was one of only a few states having both 5-minute maximum and continuous 5-
minute SC>2 ambient monitoring, as well as having over 30 1-hour SC>2 monitors in operation at
some time during the period from 1997 to 2007. In addition, the air quality characterization,
described in Chapter 7, estimated frequent exceedances above the potential health effect
benchmark levels at several of the 1-hour ambient monitors in Missouri. In a ranking of
estimated 862 emissions reported in the National Emissions Inventory (NEI), Missouri ranked
7th for the number of stacks with > 1000 tpy SOX emissions out of all U.S. states. These stack
emissions were associated with a variety of source types such as electrical power generating
units, chemical manufacturing, cement processing, and smelters. For all these reasons, the
current SC>2 lung function risk assessment focuses on Missouri and, within Missouri, on those
areas within 20 km of a major point source of SC>2 emissions in Greene County and the St. Louis
area.
9.2 DEVELOPMENT OF APPROACH FOR 5-MINUTE LUNG FUNCTION
RISK ASSESSMENT
The lung function risk assessment is based on the health effects information evaluated in
the ISA and discussed above in Chapter 4. The basic structure of the risk assessment reflects the
fact that we have available controlled human exposure study data from several studies involving
volunteer asthmatic subjects who were exposed to SC>2 concentrations at specified exposure
levels while engaged in moderate or greater exertion for 5- or 10-minute exposures. As
discussed in the ISA (section 3.1.3.5), among asthmatics, both the magnitude of SO2-induced
July 2009 315
-------
lung function decrements observed in responding individuals and the percent of individuals
affected in the group exposed have been shown to increase with increasing 5- to 10-minute 862
exposure levels in the range of 200 to 1,000 ppb. Therefore, for the SC>2 lung function risk
assessment we have developed probabilistic exposure-response relationships based on these data.
The analysis was based on the combined data set consisting of all available individual data that
describe the relationship between a measure of personal exposure to SC>2 and measures of lung
function recorded in these studies. For the purposes of this risk assessment, all of the individual
data, including both 5- and 10-minute exposure duration, were combined and treated as
representing 5-minute responses. These probabilistic exposure-response relationships were then
combined with 5-minute daily maximum peak exposure estimates for mild and moderate
asthmatics engaged in moderate or greater exertion associated with the various air quality
scenarios mentioned above. A more detailed description of the exposure assessment that was the
source of the estimated daily maximum 5-minute peak exposures under moderate or greater
exertion is provided above in Chapter 8.
9.2.1 General Approach
The major components of the lung function health risk assessment are illustrated in
Figure 9-1. As shown in Figure 9-1, under the lung function risk assessment, exposure estimates
for mild and moderate asthmatics for a number of different air quality scenarios (i.e., "as is" air
quality (representing 2002), just meeting the current 24-hour standard, just meeting alternative
standards) are combined with probabilistic exposure-response relationships derived using a
combined data base consisting of data from several controlled human exposure studies to
develop risk estimates. The air quality and exposure analysis components that are integral to this
risk assessment are discussed in greater detail in Chapters 7 and 8 of this document and in the
Exposure Assessment TSD (included as Appendix B to this document). Only the air quality and
exposure aspects affecting the scope of the lung function risk assessment are briefly discussed in
section 9.2.2. A description of the overall approach to estimating the exposure-response
relationship is included in section 9.2.3 below.
Two types of risk measures were generated for the lung function risk assessment. The
first type included estimates of the number and percentage of all asthmatics (or asthmatic
children) experiencing one or more occurrences of a defined lung function response associated
with 5-minute exposures to SC>2 while engaged in moderate or greater exertion under a given air
July 2009 316
-------
quality scenario. The second type of risk measure generated for each defined lung function
response is the number of occurrences of the lung function response in asthmatics (or asthmatic
children) in a year associated with 5-minute exposures at moderate or greater exertion under a
given air quality scenario. Since asthmatic school age children are a subset of all asthmatics, the
risk estimates presented for these two groups should not be combined.
To obtain risk estimates associated with SC>2 concentrations under different scenarios, we
estimated expected risk given the personal exposures associated with SC>2 concentrations under
each scenario - i.e., associated with
• "as is" ambient SC>2 concentrations representing 2002 air quality,
• SC>2 air quality levels simulating just meeting the current 24-hour and annual standards,
and
• SC>2 air quality levels simulating just meeting specified alternative 1-hour standards.
Note that, in contrast to the headcount risk estimates calculated for the Os health risk
assessment, the headcount risk estimates calculated for the SC>2 health risk assessment reflect
risks associated with all ambient SC>2 concentrations, not just risks in excess of estimated policy-
relevant background ambient SC>2 concentrations. This is because policy-relevant background
SC>2 concentrations are estimated to be at most 30 parts per trillion and they contribute less than
1% to present day 862 ambient concentrations (ISA, section 2.4.6) and thus would have little
impact on the risk estimates.
The first measure of risk (i.e., the number or percent of individuals in the designated
population to experience at least one lung function response in a year) is calculated as follows:
1) From the exposure modeling described in Chapter 8, we obtain the number of
individuals exposed at least once to x ppb 862 or higher, for x = 0, 50, 100, ... to 800;
2) We then calculate the number of individuals exposed at least once to SC>2
concentrations within each SC>2 exposure bin defined above (item 2 in the illustrative
example in Table 9-1 below);
3) We then multiply the number of individuals in each exposure bin (item 2 in Table 9-1
below) by the response probability (item 3 in Table 9-1 below) corresponding to the
midpoint of the exposure bin (item 1 in Table 9-1 below); and
4) We sum the results across all of the bins.
July 2009 317
-------
Air Quality
Exposure
Ambient Modeling
for Selected Areas
Air Quality
Adjustment
Procedures
Current and
Alternative
Proposed
Standards
Exposure Estimates
Associated with:
•Recent Air Quality
•Current Standard
•Alternative
Standards
Exposure-Response
Controlled Human
Exposure Studies
Probabilistic
Exposure -
Response
Relationships
Health
Risk
Model
Risk Estimates:
• Recent Air
Quality
• Current
Standard
• Alternative
Standards
Figure 9-1. Major components of 5-minute peak lung function health risk assessment based on controlled human exposure studies.
July 2009
318
-------
Because response probabilities are calculated for each of several percentiles of a
probabilistic exposure-response distribution, estimated numbers of individuals with at least one
SO2-related lung function response are similarly percentile-specific. For example, the kth
percentile number of individuals, Yk associated with SC>2 concentrations under a given air quality
scenario is:
Yk=^NI}.x(Rk e}.) (equation 9-1)
;=i
where:
Cj = (the midpoint of) the jth category of personal exposure to 862, given "as is" ambient
SC>2 concentrations;
NIj = the number of individuals whose highest exposure is to 6j ppb SO2, given ambient
SC>2 concentrations under the specified air quality scenario;
RRk | e j = the kth percentile response rate at SC>2 concentration e/, and
n = the number of intervals (categories) of SC>2 personal exposure concentration.
The kth percentile estimate of the total number responding is then calculated by multiplying the
kth percentile risk by the number of people in the relevant population. An example is given in
Table 9-1, for the median (i.e., 50th percentile) risk estimate using personal exposures associated
with a 99th percentile 100 ppb 1-hour daily maximum 862 standard for asthmatics in the St.
Louis modeling domain. We note that this calculation assumes that individuals who do not
respond at the highest SO2 concentration to which they are exposed will not respond to any lower
SO2 concentrations to which they are exposed.
The second type of risk measure, the number of occurrences of a defined lung function
response in the designated population (i.e., asthmatics or asthmatic children) in a year associated
with SC>2 concentrations under a given air quality scenario is calculated as follows:
1) From the exposure modeling described in Chapter 8, we obtain the number of
exposure occurrences among the population at and above each benchmark level (i.e.,
0 ppb, 50 ppb, 100 ppb, ... 800 ppb);
2) We then calculate the number of exposure occurrences within each 50 ppb exposure
"bin" (e.g., < 50 ppb, 50-100 ppb, etc.) 89(item 2 in the illustrative example in Table 9-2
below);
89 The final exposure bin was from 750 to 800 ppb SO2. In at least one of the alternative standard scenarios, there
were a few individuals whose exposure was greater than 800 ppb. For anyone whose exposure exceeded 800 ppb,
July 2009 319
-------
Table 9-1. Example calculation of the number of asthmatics in st. louis engaged in moderate or
greater exertion estimated to experience at least one lung function response (defined as an
increase in sRaw > 100%) associated with exposure to SO2 concentrations just meeting a 99"
percentile, 1-hour 100 ppb standard.
nth
SO2 Exposure Bin (ppb)
Lower
Bound
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
Upper
Bound
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Midpoint
(1)
25
75
125
175
225
275
325
375
425
475
525
575
625
675
725
775
Number of
Asthmatics with
At Least One
Exposure in Bin
(2)
53711
34236
9835
3059
929
368
145
84
31
22
8
0
0
8
0
0
Probability of
Response at Midpoint
SO2 Level
(3)
0.00406
0.02334
0.05162
0.08563
0.12300
0.16220
0.20210
0.24190
0.28060
0.31830
0.35430
0.38850
0.42090
0.45150
0.46600
0.49380
Estimated Number of
Asthmatics Experiencing
at Least One Lung
Function Response
=(2) x (3)
218
799
508
262
114
60
29
20
9
7
3
0
0
4
0
0
Total: 102436 Total: 2032
3) We then multiply the number of occurrences in each exposure bin (item 2 in Table 9-2
below) by the response probability (item 3 in Table 9-2 below) corresponding to the
midpoint (item 1 in Table 9-2 below) of the exposure bin; and
4) We sum the results across all of the bins.
Similar to the first type of risk measure discussed above, because response probabilities
are calculated for each of several percentiles of a probabilistic exposure-response distribution,
estimated numbers of occurrences are similarly percentile-specific. The kth percentile number of
occurrences, Ok, associated with 862 concentrations under a given air quality scenario is:
e,)
(equation 9-2)
where:
•th
j = (the midpoint of) the j category of personal exposure to 862;
we assumed a final bin from 800 to 850 ppb, and assigned them the midpoint value of that bin,825 ppb. This will
result in a slight downward bias in the estimate of risk.
July 2009
320
-------
Nj = the number of exposures to e}- ppb SC>2, given ambient SC>2 concentrations under the
specified air quality scenario;
Rk ej = the kth percentile response probability at 862 concentration e^ and
n = the number of intervals (categories) of SC>2 personal exposure concentration.
An example calculation is given in Table 9-2.
Table 9-2. Example calculation of number of occurrences of lung function response (defined as
an increase in sRaw > 100%), among asthmatics in St. Louis engaged in moderate or greater
exertion associated with exposure to SO, concentrations that just meet a 99th percentile 1-hour,
100 ppb standard.
SO2 Exposure Bin (ppb)
Lower
Bound
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
Upper
Bound
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Midpoint
(D
25
75
125
175
225
275
325
375
425
475
525
575
625
675
725
775
Number of
Exposures
(2)
16519000
136621
15760
3826
1051
413
175
83
31
24
8
0
0
8
0
0
Probability of
Response at Midpoint
SO2 Level
(3)
0.00406
0.02334
0.05162
0.08563
0.12300
0.16220
0.20210
0.24190
0.28060
0.31830
0.35430
0.38850
0.42090
0.45150
0.46600
0.49380
Expected Number of
Occurrences of Lung
Function Response
=(2) x (3)
67067
3189
814
328
129
67
35
20
9
8
3
0
0
4
0
0
Expected Number
Total Number of Exposures: 16677000 of Occurrences: 71672
9.2.2 Exposure Estimates
As noted above, exposure estimates used in the lung function risk assessment were
obtained from running the APEX exposure model for the population of individuals with asthma
for selected locations encompassing a variety of SC>2 emission source types within two areas in
the state of Missouri (i.e., St. Louis and Greene County). Chapter 8 provides additional details
about the inputs and methodology used to estimate 5-minute daily maximum peak 862 exposures
while engaged in moderate or greater exertion for the asthmatic population in these two areas.
July 2009
321
-------
These 5-minute exposure estimates for asthmatic children and adult asthmatics have been
combined separately with probabilistic exposure-response relationships for lung function
response associated with 5-minute 862 exposures. Only the highest 5-minute peak exposure
(with moderate or greater exertion) on each day has been considered in the lung function risk
assessment, since the controlled human exposure studies have shown an acute-phase response
that was followed by a short period where the individual was relatively insensitive to additional
SC>2 challenges. Staff recognizes that consideration of only the highest 5-minute exposure (with
moderate or greater exertion) on each day likely leads to some underestimation of health risks
since we are not including the health impact of other 5-minute exposures (with moderate or
greater exertion) occurring on the same day.
As described in section 8.8.1, instead of adjusting upward90 the air quality concentrations
to simulate just meeting the current SC>2 standards and potential alternative 1-hr daily maximum
standards, to reduce computer processing time, the exposure assessment simulated exposures
associated with just meeting various standards by adjusting the health effect benchmark levels by
the same factors described for each specific modeling domain and simulated year (see Table 8-
11). Since it is a proportional adjustment, the end effect of adjusting concentrations upwards
versus adjusting benchmark levels downward within the model is the same. The same follows
for where as is concentrations were in excess of an alternative standard level (e.g., 50 ppb for the
99th percentile averaged over three years), only the associated benchmarks are adjusted upwards
(i.e., a higher threshold concentration that would simulate lower exposures).
9.2.3 Exposure-Response Functions
Similar to the approach used in the ozone lung function risk assessment (Abt Associates,
2007), we have used a Bayesian Markov Chain Monte Carlo approach to estimate probabilistic
exposure-response relationships for lung function decrements associated with 5-minute daily
maximum peak exposures while engaged in moderate or greater exertion using the WinBUGS
software (Spiegelhalter et al., 1996).91 The combined data set includes all available individual
data from controlled human exposure studies of mild-to-moderate asthmatic individuals exposed
for 5- or 10-minutes while engaged in moderate or greater exertion that was summarized in the
90 To evaluate the current and most of the alternative 1-hr standards analyzed, "as is" ambient concentrations were
lower than air quality that would just meet the standards.
91 See Gleman et al. (1995) or Gilks et al. (1996) for an explanation of these methods.
July 2009 322
-------
final ISA. As noted above, for the purposes of this risk assessment, all of the individual response
data, including both 5- and 10-minute exposure durations, have been combined and treated as
representing 5-minute responses. Table 9-3 summarizes the available controlled human
exposure data that have been used to develop the probabilistic exposure-response relationships
for the lung function risk assessment.
July 2009 323
-------
Table 9-3. Percentage of asthmatic individuals in controlled human exposure studies experiencing SO2-induced decrements in lung function.
SO2
Level
(ppb)
200
250
300
400
500
600
Exposure
Duration
10 min
10 min
5 min
5 min
10 min
10 min
10 min
10 min
10 min
10 min
10 min
5 min
10 min
10 min
10 min
No. of
Subjects
40
40
19
9
28
20
21
20
21
40
40
10
28
45
40
Ventilation
(L/min)
-40
-40
-50-60
-80-90
-40
-50
-50
-50
-50
-40
-40
-50-60
-40
-30
-40
Lung
Funct.
sRaw
FEV-,
sRaw
sRaw
sRaw
sRaw
sRaw
FEV1
FEV-,
sRaw
FEV-,
sRaw
sRaw
sRaw
sRaw
Cumulative Percentage of
Responders
(Number of Subjects)1
> 100% t
> 15% 4-
5% (2)
13% (5)
32% (6)
22% (2)
4%(1)
10% (2)
33% (7)
15% (3)
24% (5)
23% (9)
30% (12)
60% (6)
18% (5)
36% (16)
35% (14)
sRaw
> 200% t
FEVi
> 20% 4-
0
5% (2)
16% (3)
0
0
5%(1)
10% (2)
0
14% (3)
8% (3)
23% (9)
40% (4)
4%(1)
16% (7)
28% (11)
> 300% t
> 30% 4-
0
3%(1)
0
0
0
5%(1)
0
0
10% (2)
3%(1)
13% (5)
20% (2)
4%(1)
13% (6)
18% (7)
Reference
Linnetal. (1987)^
Linn etal. (1987)
Bethel et al. (1985)
Roger etal. (1985)
Linnetal. (1988)J
Linnetal. (1990)J
Linn etal. (1988)
Linn etal. (1990)
Linnetal. (1987)
Linn etal. (1987)
Bethel etal. (1983)
Roger etal. (1985)
Magnussen et al.
(1990)4
Linn etal. (1987)
Respiratory Symptoms:
Supporting Studies
Limited evidence of SO2-
induced increases in respiratory
symptoms in some asthmatics:
Linnetal. (1983; 1984; 1987;
1988; 1990), Schacter et al.
(1984)
Stronger evidence with some
statistically significant increases
in respiratory symptoms:
Balmes et al. (1987)4, Gong
etal. (1995), Linnetal. (1983;
1987), Roger etal. (1985)
Clear and consistent increases
July 2009
324
-------
S02
Level
(ppb)
1,000
Exposure
Duration
10 min
10 min
10 min
10 min
10 min
10 min
10 min
No. of
Subjects
20
21
40
20
21
28
10
Ventilation
(L/min)
-50
-50
-40
-50
-50
-40
-40
Lung
Funct.
sRaw
sRaw
FEV-,
FEV-,
FEV-,
sRaw
sRaw
Cumulative Percentage of
Responders
(Number of Subjects)1
> 100% t
> 15% 4-
60% (12)
62% (13)
53% (21)
55% (11)
43% (9)
50% (14)
60% (6)
sRaw
> 200% t
FEVi
> 20% 4-
35% (7)
29% (6)
48% (19)
55% (11)
33% (7)
25% (7)
20% (2)
> 300% t
> 30% 4-
10% (2)
14% (3)
20% (8)
5%(1)
14% (3)
14% (4)
0
Reference
Linnetal. (1988)
Linn etal. (1990)
Linn etal. (1987)
Linnetal. (1988)
Linnetal. (1990)
Roger etal. (1985)
Kehrletal. (1987)
Respiratory Symptoms:
Supporting Studies
in SO2-induced respiratory
symptoms: Linn etal. (1984;
1987; 1988; 1990), Gong et al.
(1995), Horstman etal. (1988)
Notes:
1Data presented from all references from which individual data were available. Percentage of individuals who experienced greater than or equal to a 100, 200, or
300% increase in specific airway resistance (sRaw), or a 15, 20, or 30% decrease in FEV-i. Lung function decrements are adjusted for effects of exercise in
clean air (calculated as the difference between the percent change relative to baseline with excersise/SC>2 and the percent change relative to baseline with
exercise/clean air). Quality control of data was performed by two EPA staff scientists.
2Responses of mild and moderate asthmatics reported in Linn et al. (1 987) have been combined. Data reported only for the first 1 0 min period of exercise in the
first round of exposures.
3Analysis includes data from only mild (1988) and moderate (1990) asthmatics who were not receiving supplemental medication.
4One subject was not exposed to 1 ,000 ppb due to excessive wheezing and chest tightness experienced at 500 ppb. For this subject, the values used for 500
ppb were also used for 1 ,000 ppb under the assumptions that the response at 1 ,000 ppb would be equal to or greater than the response at 500 ppb.
Indicates studies in which exposures were conducted using a mouthpiece rather than a chamber.
Source: ISA, Table 3-1 (EPA, 2008c, p.3-10).
July 2009
325
-------
The combined data set from Linn et al. (1987, 1988, 1990), Bethel et al. (1983,
1985), Roger et al. (1985), and Kehrl et al. (1987), summarized in Table 9-3, provide data
with which to estimate exposure-response relationships between responses defined in
terms of sRaw and 5-minute exposures to SO2 at levels of 200, 250, 300, 400, 500, 600,
and 1,000 ppb (the exposure levels included in these studies).92 Two definitions of
response have been used: (1) an increase in sRaw > 100% representing moderate or
greater responses and (2) an increase in sRaw > 200% reflecting severe decrements in
lung function.
Likewise, the combined data set from Linn et al. (1987, 1988, 1990), summarized
in Table 9-3, provide data with which to estimate exposure-response relationships
between responses defined in terms of FEVi and 5-minute exposures to SC>2 at levels of
200, 300, 400, and 600 ppb (the exposure levels included in these studies). Again, two
definitions of response have been used in the health risk assessment: (1) a decrease in
FEVi > 15% representing moderate or greater responses and (2) a decrease in FEVi >
20% representing severe decrements in lung function.
Before estimating exposure-response relationships for 5-minute exposures, we
corrected the data from these controlled human exposure studies for the effect of exercise
in clean air to remove any systematic bias that might be present in the data attributable to
an exercise effect. This correction is reflected in the summary of the response data
provided in Table 9-3.93 Generally, this correction for exercise in clean air is small
relative to the total effects measures in the SO2-exposed cases.
Public comments on the 2nd draft REA stated that there were errors in the data
used to create Table 9-3 (UARG, 2009). Johns (2009) describes EPA's evaluation of
these data, building upon an initial EPA analysis conducted in the previous NAAQS
review (Smith, 1994). The vast majority of the alleged errors were described as rounding
errors of the second decimal place introduced by the original study authors. Of the 640
92 Data from Magnussen et al. (1990) were not used in the estimation of sRaw exposure-response functions
because exposures in this study were conducted using a mouthpiece rather than a chamber.
93 Corrections were subject-specific. A correction was made by subtracting the subject's percent change
(in FEVi or sRaw) under the no-SO2 protocol from his or her percent change (in FEVi or sRaw) under the
given SO2 protocol, and rounding the result to the nearest integer. For example, if a subject's percent
change in sRaw under the no-SO2 protocol was 110.12% and his percent change in sRaw under the 0.6 ppm
SO2 protocol was 185.92%, then his percent change in sRaw due to SO2 is 185.92% -110.12% = 75.8%,
which rounds to 76%.
July 2009 326
-------
values of sRaw and FEVi from Linn et al. (1987), commenters identified 11
discrepancies between the original EPA analysis (Smith, 1994) and what was included in
the analysis conducted more recently by EPA (Johns, 2009). EPA has reviewed these
comments, and recognizes that some discrepancies were clearly due to transcription
errors, while others were due to difficulties reading the last decimal place of the raw data.
Commenters also identified 9 cases where the calculated average of individual lung
function measurements did not equal the average values presented in Smith (1994).
While staff placed more confidence in the average values presented rather than the
calculated average of the individual measurements, EPA nonetheless conducted a
preliminary re-analysis using the 20 apparent "corrected" values provided by
commenters. This resulted in relatively minor and variable changes in SO2-induced
changes in lung function, which did not substantively change the percent responders as
presented in Table 9-3. Further, incorporating these 20 changes resulted in an increase in
the percent of responders in three table entries, while no decreases in the percent of
responders were observed. Although the data presented in Table 9-3 were subjected to
quality control procedures (see Johns, 2009), EPA is currently in the process of
conducting a full quality assurance review of the data in response to these public
comments and expects to present the quantitative results of its evaluation as part of the
record for the November proposal. The risk assessment results presented in this
document are based on the Johns (2009) summary.
We considered two different functional forms for the exposure-response
functions: a 2-parameter logistic model and a probit model. In particular, we used the
data in Table 9-3 to estimate the logistic function,
and the probit function,
y(x: B. v) = - . ,. . . (equation 9-3)
J \ i f •> I ! ,. ?+*lnx \ VI /
1 'T ,*/7
y(x;J3,r)= fn ... e t/r (equation 9-4)
July 2009 327
-------
for each of the four lung function responses defined above, where x denotes the 862
concentration (in ppm) to which the individual is exposed, ln(x) is the natural logarithm
of x, y denotes the corresponding probability of response (increase in sRaw > 100% or >
200% or decrease in FEVi > 15% or > 20%), and ft and y are the two parameters whose
values are estimated. 94
We assumed that the number of responses, sf, out of TV, subjects exposed to a
given SC>2 concentration, xt, has a binomial distribution with response probability given
by equation (9-3) when we assume the logistic model and equation (9-4) when we
assume the probit model. The likelihood function is therefore
L(fl,y;datd) = Yl ' X^/W' [1-X*,-; #r)f'~* • (equation 9-5)
Some subjects in the controlled human exposure studies participated in more than
one study and were exposed to a given 862 concentration more than once. However,
because there were insufficient data to estimate subject-specific response probabilities,
we assumed a single response probability (for a given definition of response) for all
individuals and treated the repeated exposures for a single subject as independent
exposures in the binomial distribution.
For each model, we derived a Bayesian posterior distribution using this binomial
likelihood function in combination with uniform prior distributions for each of the
unknown parameters.95 We used 4,000 iterations as the "burn-in" period followed by
10,000 iterations, a number sufficient to ensure convergence of the resulting posterior
distribution. Each iteration corresponds to a set of values for the parameters of the
logistic or probit exposure-response function.
For any 862 concentration, x, we could then derive the nih percentile response
value, for any «, by evaluating the exposure-response function at x using each of the
18,000 sets of parameter values. The resulting median (50th percentile) exposure-
94 For ease of exposition, the same two Greek letters are used to indicate two unknown parameters in the
logistic and probit models; this does not imply, however, that the values of these two parameters are the
same in the two models.
95 We used the following uniform prior distributions for the 2-parameter logistic model: (3 ~ U(-10, 0); and
y ~ U(-10,0); we used the following normal prior distributions for the probit model: (3 ~ N(0, 1000); and y ~
N(0,1000).
July 2009 328
-------
response functions based on the 2-parameter logistic and probit models are shown
together, along with the data used to estimate these functions, for increases in sRaw >
100% and > 200% and decreases in FEVi > 15% and > 20% in Figures 9-2, 9-3, 9-4, and
9-5, respectively. The 2.5th percentile, median, and 97.5th percentile curves, along with
the response data to which they were fit, are shown separately for each of the eight
combinations of (four) response definitions and (two) exposure-response models in
Appendix C.
We note that there were only limited data with which to estimate the logistic and
probit exposure-response functions, and that the logistic and probit models both appear to
fit the data equally well. We also note that since the data being fit has already been
corrected to account for the lung function response due to exercise in clean air, then the
response must by definition be zero associated with 0 ppm SC>2 exposure. While the
CAS AC panel in its comments on the 2nd draft REA suggested a possible a priori reason
to prefer the probit model (based on a hypothesized lognormal distribution of individual
thresholds for response), in staff s judgment there is not sufficient evidence to select one
model over the other. Therefore, we have chosen to include both the 2-parameter logistic
and probit models to develop the risk estimates associated with exposure to 862 under the
different air quality scenarios considered. While the estimated exposure-response
relationships using the two alternative models do not appear to be that different based on
visual inspection of Figures 9-2 through 9-5, the differences do translate into substantial
differences in the estimated aggregate number of sRaw and FEVi responses for St. Louis
as discussed later in this chapter.
July 2009 329
-------
100%
Response Rate
The data
probit
2-parameter logistic
0.4 0.6
SO2 Concentration (ppm)
Figure 9-2. Bayesian-estimated median exposure-response functions: increase in sRaw >
100% for 5-Minute exposures of asthmatics under moderate or greater
exertion.*
100%
90% -
80% -
70% -
60% -
50% -
40% -
30% -
20% -
10% -
0% -
A The data
probit
2-parameter logistic
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure 9-3. Bayesian-estimated median exposure-response functions: increase in sRaw >
200% for 5-minute exposures of asthmatics under moderate or greater
exertion.*
'Derived using method described in text based on all of the individual response data from Linn et al.
(1987), Linn et al. (1988), Linn et al. (1990), Bethel et al. (1983), Bethel et al. (1985), Roger et al. (1985),
and Kehrletal. (1987).
July 2009
330
-------
Response Rate
100% -
90% -
80% -
70% -
60% -
50% -
40% -
30% -
20% -
10% -
0%
The data
probit
2-parameter logistic
0
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure 9-4. Bayesian-estimated median exposure-response functions: decrease in FEV1 >
15% for 5-minute exposures of asthmatics under moderate or greater
exertion*.
100%
A The data
probit
2-parameter logistic
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure 9-5. Bayesian-estimated median exposure-response functions: decrease in FEV1 >
20% for 5-minute exposures of asthmatics under moderate or greater exertion.*
'Derived using method described in text based on all of the individual response data from Linn et al.
(1987), Linn et al. (1988), and Linn et al. (1990).
July 2009
331
-------
9.3 LUNG FUNCTION RISK ESTIMATES
In this section, we present and discuss risk estimates associated with several air
quality scenarios, including "as is" air quality represented by 2002 monitoring data. In
addition, risk estimates are presented for several hypothetical scenarios, equivalent to
adjusting air quality upward to simulate just meeting the current annual 862 24-hour
standard and to adjusting air quality (either up or down) to simulate just meeting potential
alternative 98th and 99th percentile daily maximum 1-h standards. As discussed
previously in Chapter 5, potential alternative 1-h standards with levels set at 50, 100, 150,
200, and 250 ppb have been included in the risk assessment. Only selected risk estimates
are presented in this section and additional risk estimates are presented in Appendix C.
Throughout this section and Appendix C the uncertainty surrounding risk estimates
resulting from the statistical uncertainty in the 862 exposure-response relationships due
to sampling error is characterized by ninety-five percent credible intervals around
estimates of occurrences, number of asthmatics experiencing one or more lung function
response, and percent of total incidence that is SO2-related.
Risk estimates for selected lung function responses for all asthmatics and
asthmatic children associated with 5-minute exposures to ambient SC>2 concentrations
while engaged in moderate or greater exertion are presented in Tables 9-4 through 9-9.
Tables 9-4 through 9-6 are for all asthmatics and Tables 9-7 through 9-9 are for asthmatic
children. Each table includes risk estimates for both Greene County and St. Louis,
Missouri. Each table also includes risk estimates based on use of both the 2-parameter
logistic and probit exposure-response models. As discussed in section 9.2.3, the risk
assessment included two types of lung function responses (i.e., sRaw and FEVi) and two
levels of response for each type of lung function response (> 100 and 200% increase for
sRaw and > 15 and 20% decrease for FEVi). Risk estimates using sRaw as the measure
of lung function response are included in this section because the exposure-response
relationships were developed based on a larger set of data from individual subjects, which
gives us more confidence in the exposure-response relationship. Additional risk
estimates using FEVi as the indicator of lung function response are included in Tables 4-
3, 4-4, 4-7, and 4-8 in Appendix C and show similar patterns across the current and
alternative standards for the two study areas.
July 2009 332
-------
Tables 9-4 and 9-5 summarize the estimated number and percent of asthmatics
that would experience 1 or more lung function responses in a year, where lung function
response was defined as > 100% and > 200% increase in sRaw, in all asthmatics
associated with ambient 5-minute SC>2 exposures estimated to occur under "as is" air
quality (i.e., air quality based on 2002 monitored and modeled SC>2 air quality data) and
under air quality representing just meeting the current SC>2 standards and several
alternative 1-hour daily maximum SC>2 standards. Tables 9-7 and 9-8 present the same
types of estimates for asthmatic children. The median estimates are presented in each
cell of the table with the 95% credible intervals based on statistical uncertainty
surrounding the SC>2 coefficient in the exposure-response relationship shown in
parentheses below the median estimates.
Tables 9-6 and 9-9 summarize the estimated number of occurrences of two
defined levels of lung function response (> 100% and > 200% increase in sRaw) in all
asthmatics and in asthmatic children, respectively, associated with ambient 5-minute SC>2
exposures estimated to occur under "as is" air quality (i.e., air quality based on 2002
monitored and modeled 862 air quality data) and under air quality representing just
meeting the current 862 standards and several alternative 1-hour daily maximum 862
standards.
The current primary 862 standards include a 24-hour standard set at 0.14 parts per
million (ppm), not to be exceeded more than once per year, and an annual standard set at
0.03 ppm, calculated as the arithmetic mean of hourly averages. In St. Louis, SC>2
concentrations that are predicted to occur if the current standards were just met are
substantially higher than "as is" air quality (based on 2002 monitoring and modeling
data) and also substantially higher than they would be under any of the alternative 1-hr
standards considered in this analysis. Consequently, the levels of response that would be
seen if the current standard were just met are well above the levels that would be seen
under the "as is" air quality scenario or under any of the alternative 1-hr standards - for
asthmatics and for asthmatic children, and for all four definitions of lung function
response. We also note that the only standard resulting in decreases in lung function
responses relative to the "as is" scenario is the 50 ppb, 99th percentile 1-hr daily
maximum standard (corresponding to the 99/50 column in Tables 9-6 through 9-9).
July 2009 333
-------
Table 9-4. Number of asthmatics engaged in moderate or greater exertion estimated to experience at least one lung function response
associated with exposure to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
S02
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
90
(20 - 390)
10
(0-180)
210
(80 - 620)
110
(40-410)
80
(20 - 380)
10
(0-170)
90
(20 - 390)
10
(0-180)
100
(20 - 420)
20
(0-210)
120
(30 - 460)
40
(10-250)
160
(50 - 520)
70
(20-310)
140
(40 - 500)
60
(20 - 280)
Sf. Louis, MO
2-Parameter Logistic
Probit
1010
(340-3010)
500
(140-1990)
13460
(9740-18510)
13050
(9430-18100)
730
(220 - 2490)
290
(70-1470)
1990
(860 - 4690)
1340
(520 - 3690)
3650
(1900-7100)
2930
(1450-6200)
5520
(3230 - 9490)
4810
(2760-8710)
7500
(4770-11850)
6860
(4310-11190)
7050
(4410-11320)
6400
(3950-10640)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
30
(0-210)
0
(0-80)
70
(20-310)
30
(10-180)
30
(0-210)
0
(0 - 80)
30
(0-210)
0
(0 - 90)
30
(0 - 220)
10
(0-100)
40
(10-240)
10
(0-110)
50
(10-270)
20
(0-140)
50
(10-260)
10
(0-130)
Sf. Louis, MO
2-Parameter Logistic
Probit
330
(70-1520)
120
(20 - 880)
5520
(3400 - 8960)
5180
(3150-8570)
230
(40-1290)
60
(10-660)
670
(210-2270)
350
(90-1590)
1280
(510-3360)
870
(310-2680)
2010
(940 - 4470)
1560
(690 - 3820)
2830
(1470-5590)
2380
(1200-5000)
2640
(1340-5330)
2190
(1070-4730)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest ten.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
July 2009
334
-------
Table 9-5. Percent of asthmatics engaged in moderate or greater exertion estimated to experience at least one lung function response
associated with exposure to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
1%
(0.4% - 2.9%)
0.5%
(0.2% -1.9%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.5%
(0.1% -2%)
0.1%
(0% - 1 %)
0.6%
(0.2% -2.1%)
0.2%
(0%-1.2%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
0.7%
(0.2% - 2.3%)
0.3%
(0.1% -1.3%)
Sf. Louis, MO
2-Parameter Logistic
Probit
1%
(0.3% - 2.9%)
0.5%
(0.1% -1.9%)
13.1%
(9. 5% -18.1%)
12.7%
(9.2% -17.7%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
1.9%
(0.8% - 4.6%)
1.3%
(0.5% - 3.6%)
3.6%
(1.9% -6. 9%)
2.9%
(1.4% -6.1%)
5.4%
(3.2% - 9.3%)
4.7%
(2.7% - 8.5%)
7.3%
(4.7% - 1 1 .6%)
6.7%
(4.2% -10.9%)
6.9%
(4. 3% -11.1%)
6.2%
(3. 9% -10.4%)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.2%
(0%-1%)
0%
(0% - 0.5%)
0.2%
(0%-1.1%)
0%
(0% - 0.5%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.2%
(0%-1.2%)
0.1%
(0% - 0.6%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.9%)
5.4%
(3.3% - 8.7%)
5.1%
(3.1% -8.4%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.7%
(0.2% - 2.2%)
0.3%
(0.1% -1.6%)
1.3%
(0.5% - 3.3%)
0.8%
(0.3% - 2.6%)
2%
(0.9% - 4.4%)
1.5%
(0.7% - 3.7%)
2.8%
(1.4% -5. 5%)
2.3%
(1.2% -4.9%)
2.6%
(1.3% -5.2%)
2.1%
(1%-4.6%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
July 2009
335
-------
Table 9-6. Number of occurrences
exertion associated with exposure
(in hundreds) of a lung function response among asthmatics engaged in moderate or greater
to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
125
(24 - 572)
16
(0 - 256)
127
(25 - 577)
18
(1 -261)
125
(24 - 572)
16
(0 - 256)
125
(24 - 572)
16
(0 - 256)
125
(24 - 573)
16
(1-257)
126
(24 - 573)
16
(1-257)
126
(24 - 575)
17
(1 -258)
126
(24 - 574)
17
(1 -258)
Sf. Louis, MO
2-Parameter Logistic
Probit
657
(128-2985)
90
(4-1346)
1672
(663 - 4740)
933
(393-3107)
652
(125-2975)
86
(3-1336)
686
(141 -3041)
111
(11 -1402)
762
(176-3184)
170
(33-1543)
880
(234 - 3398)
264
(72-1756)
1036
(315-3673)
392
(128-2031)
997
(295 - 3604)
360
(114-1963)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
38
(4-310)
2
(0-123)
39
(4-312)
3
(0-124)
38
(4-310)
2
(0-122)
38
(4-310)
2
(0-123)
38
(4-310)
2
(0-123)
38
(4-310)
2
(0-123)
39
(4-311)
2
(0-123)
39
(4-311)
2
(0-123)
Sf. Louis, MO
2-Parameter Logistic
Probit
201
(21 -1614)
13
(0 - 643)
560
(165-2407)
258
(86-1388)
199
(20-1609)
12
(0 - 639)
211
(24-1639)
18
(1 - 666)
237
(32-1703)
33
(5 - 725)
278
(47-1799)
59
(12-814)
332
(68-1923)
95
(24 - 930)
319
(63-1892)
86
(21 -901)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03
ppm, calculated as the arithmetic mean of hourly averages.
July 2009
336
-------
Table 9-7. number of
response associated
asthmatic children engaged in moderate or greater exertion estimated to experience at least one lung function
with exposure to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
30
(10-130)
10
(0-60)
110
(40 - 270)
60
(20 - 200)
30
(10-130)
0
(0 - 60)
30
(10-140)
10
(0 - 60)
40
(10-150)
10
(0-80)
50
(20-180)
20
(10-100)
70
(30-210)
40
(10-140)
60
(20 - 200)
30
(10-130)
Sf. Louis, MO
2-Parameter Logistic
Probit
590
(220-1570)
340
(100-1150)
8020
(6080-10370)
7950
(6020-10320)
400
(130-1210)
190
(50 - 790)
1220
(560 - 2620)
890
(360 - 2220)
2240
(1240-4010)
1910
(1000-3690)
3370
(2090 - 5350)
3080
(1860-5110)
4560
(3060 - 6680)
4330
(2870-6510)
4290
(2840 - 6390)
4060
(2640-6210)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
10
(0-70)
0
(0-30)
40
(10-130)
20
(0-90)
10
(0 - 70)
0
(0 - 30)
10
(0 - 70)
0
(0 - 30)
10
(0-80)
0
(0-40)
20
(0 - 90)
10
(0 - 50)
20
(10-110)
10
(0 - 60)
20
(10-100)
10
(0 - 60)
Sf. Louis, MO
2-Parameter Logistic
Probit
190
(50 - 780)
80
(10-500)
3380
(2190-5070)
3290
(2110-5000)
130
(30-610)
40
(10-350)
410
(140-1240)
240
(60 - 950)
800
(340-1870)
580
(220-1590)
1250
(620 - 2500)
1030
(480 - 2250)
1750
(970-3140)
1560
(830 - 2940)
1640
(890 - 3000)
1440
(740 - 2790)
'Numbers are median (50th percentile) numbers of asthmatic children. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest ten.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
July 2009
337
-------
Table 9-8. Percent of asthmatic children engaged in moderate or greater exertion estimated to experience at least one lung function
response associated with exposure to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.9%)
1 .4%
(0.6% - 3.7%)
0.9%
(0.3% - 2.7%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.4%
(0.1% -1.9%)
0.1%
(0% - 0.9%)
0.5%
(0.1% -2.1%)
0.2%
(0%-1.1%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
1%
(0.3% - 2.9%)
0.5%
(0.2% - 1 .9%)
0.9%
(0.3% - 2.7%)
0.4%
(0.1% -1.7%)
Sf. Louis, MO
2-Parameter Logistic
Probit
1 .4%
(0.5% - 3.8%)
0.8%
(0.2% - 2.8%)
19.2%
(14. 6% -24. 9%)
19.1%
(14.4% -24.7%)
0.9%
(0.3% - 2.9%)
0.4%
(0.1% -1.9%)
2.9%
(1.3% -6. 3%)
2.1%
(0.9% - 5.3%)
5.4%
(3% - 9.6%)
4.6%
(2.4% - 8.8%)
8.1%
(5% -12. 8%)
7.4%
(4. 5% -12. 3%)
10.9%
(7. 3% -16%)
10.4%
(6.9% -15.6%)
10.3%
(6. 8% -15. 3%)
9.7%
(6.3% -14.9%)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.5%
(0.1% -1.8%)
0.2%
(0.1% -1.2%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.2%
(0%-1.1%)
0%
(0% - 0.5%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
0.3%
(0.1% -1.4%)
0.1%
(0% - 0.8%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.5%
(0.1% -1.9%)
0.2%
(0% - 1 .2%)
8.1%
(5. 3% -12. 2%)
7.9%
(5% -12%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
1%
(0.3% - 3%)
0.6%
(0.2% - 2.3%)
1.9%
(0.8% - 4.5%)
1 .4%
(0.5% - 3.8%)
3%
(1.5% -6%)
2.5%
(1.2% -5.4%)
4.2%
(2.3% - 7.5%)
3.7%
(2% - 7%)
3.9%
(2.1% -7.2%)
3.4%
(1.8% -6.7%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
July 2009
338
-------
Table 9-9. number of occurrences
exertion associated with exposure
(in hundreds) of a lung function response among asthmatic children engaged in moderate or greater
to SO2 concentrations under alternative air quality scenarios in a year.*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
71
(13-324)
9
(0-145)
72
(14-327)
10
(1 -148)
71
(13-324)
9
(0-145)
71
(14-324)
9
(0-145)
71
(14-324)
9
(0-145)
71
(14-325)
9
(0-146)
71
(14-325)
10
(0-146)
71
(14-325)
10
(0-146)
Sf. Louis, MO
2-Parameter Logistic
Probit
417
(81 -1893)
58
(3 - 855)
1179
(484 - 3209)
692
(296-2176)
413
(80-1885)
55
(2 - 847)
439
(91 -1935)
74
(8 - 896)
497
(118-2043)
118
(25-1004)
586
(162-2206)
189
(53-1166)
704
(222-2413)
286
(96-1373)
674
(207-2361)
262
(85-1321)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
22
(2-175)
1
(0-69)
22
(2-177)
2
(0-71)
22
(2-175)
1
(0 - 69)
22
(2-175)
1
(0 - 69)
22
(2-175)
1
(0-69)
22
(2-176)
1
(0 - 70)
22
(2-176)
1
(0 - 70)
22
(2-176)
1
(0 - 70)
Sf. Louis, MO
2-Parameter Logistic
Probit
128
(13-1023)
8
(0 - 408)
397
(122-1618)
192
(65 - 967)
126
(13-1019)
8
(0-405)
135
(15-1042)
12
(1-425)
155
(22-1091)
24
(4 - 470)
186
(33-1164)
43
(9 - 538)
227
(49-1257)
70
(18-625)
217
(45-1234)
63
(16-603)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
July 2009
339
-------
As an illustration of the changes in the number of occurrences of sRaw increases > 100%
in all asthmatics across the range of standards analyzed in the St. Louis modeling domain, under
the current SO2 standards the median estimate is 117,900. These estimated occurrences decrease
for increasingly more stringent alternative 1-hour standards with the 50 ppb, 99th percentile daily
maximium 1-hour standard, the most stringent alternative standard analyzed, reducing the
median estimated number of occurrences of this lung function response to 41,300. The pattern
of reductions observed for all asthmatics is similar to that observed in asthmatic children.
The estimated occurrences of sRaw responses are much lower in Greene County both due
to a smaller population as well as fewer exposure occurrences of elevated 5-minute 862
concentrations. We also note that the differences in estimated occurrences of lung function
responses associated with all of the air quality scenarios analyzed are much smaller for Greene
County than in St. Louis. The minimal differences observed in Greene County among the air
quality scenarios analyzed is due to the relatively small differences in the distribution of
exposures while engaged in moderate or greater exertion among the air quality scenarios
analyzed.
Figures 9-7 (a) and (b) show the percent of asthmatics based on use of the logistic and
probit exposure-response models, respectively, engaged in moderate or greater exertion in St.
Louis, MO estimated to experience at least one lung function response in a year, defined as an
increase in sRaw > 100%, attributable to exposure to 862 in each exposure "bin" or interval.
Figures 9-8(a) and (b) show these same estimates for the percent of asthmatic children. Figure 9-
6 displays the legend for Figures 9-7 and 9-8 indicating the exposure bins used in these figures
and Table 9-10 provides definitions of the figures' x-axis labels, which represent alternative air
quality scenarios. Similar figures are included in Appendix C for lung function responses
defined in terms of > 15% and > 20% decrements in FEVi for both asthmatics and asthmatic
children. Appendix C also includes similar figures for the Greene County study area. As
apparent in Figures 9-7 (a) and (b) and in Figures 9-8(a) and (b), the pattern of the contribution
of exposures from different concentration intervals on lung function response is very similar for
this risk metric using the two alternative exposure-response models. In comparing the risk
estimates for all asthmatics (Figure 9-7) with the risk estimates for asthmatic children (Figure 9-
8) the total percent responding is higher for asthmatic children. This is due to the greater
percentage of 5-minute exposures while engaged in moderate or greater exertion for asthmatic
July 2009 340
-------
Q Attributable to 500 ppb<=S02
E Attributable to 450 ppb<=S02<500 ppb
D Attributable to 400 ppb<=S02<450 ppb
C Attributable to 350 ppb<=S02<400 ppb
D Attribuable to 300 ppb<=S02<350 ppb
H Attributable to 250 ppb<=S02<300 ppb
E3 Attributable to 200 ppb<=S02<250 ppb
B Attributable to 150 ppb<=S02<200 ppb
D Attributable to 100 ppb<=S02<150 ppb
D Attributable to 50 ppb<=S02<100 ppb
• Attributable to S02<50 ppb
Figure 9-6. Legend for Figures 9-7 and 9-8 showing total and contribution of risk attributable to
SO2 exposure ranges.
Table 9-10. Explanation of labels on the x-axis of Figures 9-7 and 9-8.
Label
"As Is"
Air
Quality
Current
Standard
99/50
99/100
99/150
99/200
99/250
98/200
Explanation
Reflects air quality in 2002
Refers to the current suite of standards, which includes a 24-hr standard of
0. 14 ppm which is not to be exceeded more than once per year and an annual
standard set at 0.03 ppm
Refers to an alternative standard in which the
maximum concentrations must be < 50 ppb.
Refers to an alternative standard in which the
maximum concentrations must be < 100 ppb.
Refers to an alternative standard in which the
maximum concentrations must be < 150 ppb.
Refers to an alternative standard in which the
maximum concentrations must be < 200 ppb.
Refers to an alternative standard in which the
maximum concentrations must be < 250 ppb.
Refers to an alternative standard in which the
maximum concentrations must be < 200 ppb.
99th
99th
99th
99th
99th
98th
percentile
percentile
percentile
percentile
percentile
percentile
of the
of the
of the
of the
of the
of the
1-hr
1-hr
1-hr
1-hr
1-hr
1-hr
daily
daily
daily
daily
daily
daily
July 2009
341
-------
30%
=• 25% -
O
^
in
o>
I 20%
3.
O)
c
•O
g 15% -
Q.
V
o:
1 10% -
U
0)
*=;% -
0% -I
T
- - -
T
ss T
f»-»
i_JA IE ^^
SS T pi SS ilsi
1 i>^i T ^^^^ 7^"^ -^^v ~Zf& s-Zf&
i ^ i m ••' '•'• ^ ^
.— _; o o o o o o
<>.£ in o in o in o
= sw S ^ ^ C! C! C!
V) to +* oj oj oj oj oj co
1 5 g
: 3
O
a) Based on Logistic Exposure-Response Model
30%
=• 25% -
O
in
o>
I 20%
3.
O)
£
1 15% -
Q.
U)
1 10%-
Q.
5% -
0% H
i
_^
V) "5
^
o> o>
o>
p
L-a
C
U
T
c
0
n
!iii
3
)
5
r>
5
&
c
c
a
a
=
&
9
9
i|
?
I>
a
^
o«
c
u
r
(
am
L
L^
i*
9
?
-1
>
E
>
C
C
0
a
3
V
di
9
9
M
a
r>
b) Based on Probit Exposure-Response Model
Figure 9-7. Estimated percent of asthmatics experiencing one or more lung function responses
(defined as > 100% increase in sRaw) per year associated with short-term (5-minute)
exposures to SO2 concentrations associated with alternative air quality scenarios -
total and contribution of 5-minute SO2 exposure ranges (see Figure 9-6 for legend and
Table 9-10 for description of air quality scenarios included on x-axis).
July 2009
342
-------
30% -
=• 25% -
O
in
0>
c 20% -
3.
O)
c
•D
g 15% -
Q.
U)
g 10% -
o
3
. s 07
* 0>
D
3
)
F
C
C
T
C
c
9
9
}
}
]
~
C
1
T
C
C
tv
3
)
1
•>
]
^
r-f.
O^
c
c
r
c
c
™™
III
\
-ii
9
9
4
5
5
^
rs
•^
V
f
>
>,
&
c
Ij
r
c
c
y
p
11
rt
H
>;
v^
9
9
4
9
9
]
'_"
iiT
&
;>
^
c
c
r
c
c
_~.
lit
ft
H
>;
^
9
9
4
i
•>
a) Based on Logistic Exposure-Response Model
30% -
b) Based on Probit Exposure-Response Model
Figure 9-8. Estimated percent of asthmatic children experiencing one or more lung function
responses (defined as > 100% increase in sRaw) per year associated with short-term (5-
minute) exposures to SO2 concentrations associated with alternative air quality
scenarios - total and contribution of 5-minute SO2 exposure ranges (see Figure 9-6 for
legend and Table 9-10 for description of air quality scenarios included on x-axis).
July 2009
343
-------
children compared to all asthmatics due to the higher frequency of exercise in children compared
to adults. Of course the actual number of persons affected is smaller for asthmatic children since
they are a subset of all asthmatics.
The numbers of individuals with at least one lung function response attributable to
exposures in the lowest exposure concentration bin (i.e., 0 to 50 ppb) are typically quite small.
This is because the calculation of numbers of individuals with at least one lung function response
uses individuals' highest exposure only. While individuals may be exposed mostly to low 862
concentrations, many are exposed at least occasionally to higher levels. Thus, the percentage of
individuals in a designated population with at least one lung function response associated with
SC>2 concentrations in the lowest bin is likely to be very small, since most individuals are
exposed at least once to higher SO2 levels. For example, the lowest SO2 exposure bin accounts
for only about 0.2 percent of asthmatics estimated to experience at least 1 SO2-related lung
function response. For this very small percent of the population, the lowest exposure bin
represents their highest SC>2 exposures under moderate exertion in a year. Figure 9-7 (a) shows a
relatively small proportion of asthmatics in St. Louis experiencing at least one response to be
experiencing those responses because of exposures in that lowest exposure bin.
While exposures in the lowest bin are not responsible for the greatest portion of the
estimated risk for the risk metric expressed as incidence or percent incidence of a defined lung
function response 1 or more times per year, exposures in the lowest bin (i.e., 0 to 50 ppb) are
responsible for the bulk of the risks expressed as total occurrences of a defined lung function
response. As noted in public comments on the 2nd draft SC>2 REA, the assignment of response
probability to the midpoint of the exposure bin combined with the lack of more finely divided
intervals in this range can lead to significant overestimation of risks based on total occurrences
of a defined lung function response. This is because the distribution of population exposures for
occurrences is not evenly distributed across the bin, but rather is more heavily weighted toward
the lower range of the bin. Thus, combining all exposures estimated to occur in the lowest bin
with a response probability assigned to the midpoint of the bin results in a significant
overestimate of the risk. Therefore, staff places less weight on the estimated number of
occurrences of lung function responses. This overestimation of total occurrences does not
impact the risk metric expressed as incidence or percent incidence of a defined lung function
July 2009 344
-------
response 1 or more times per year because the bulk of the exposures contributing to these risk
metrics are not skewed toward the lower range of the reported exposure bins.
9.4 CHARACTERIZING UNCERTAINTY AND VARIABILITY
An important issue associated with any population health risk assessment is the
characterization of uncertainty and variability (see section 6.6 for definitions of uncertainty and
variability). This section presents a summary and discussion regarding the degree to which
variability was incorporated in the health risk assessment for lung function responses and how
the uncertainty was characterized for the risk estimates of number and percent of asthmatics and
asthmatic children experiencing defined lung function responses associated with 5-minute SC>2
exposures under moderate or greater exertion associated with alternative air quality scenarios.
With respect to variability, the lung function risk assessment incorporates some of the
variability in key inputs to the analysis by its use of location-specific inputs for the exposure
analysis (e.g., location specific population data, air exchange rates, air quality, and temperature
data). The extent to which there may be variability in exposure-response relationships for the
populations included in the risk assessment residing in different geographic areas is currently
unknown. Temporal variability also is more difficult to address, because the risk assessment
focuses on some unspecified time in the future. To minimize the degree to which values of
inputs to the analysis may be different from the values of those inputs at that unspecified time,
we have used the most current inputs available.
Our approach to characterizing uncertainty includes both qualitative and quantitative
elements. From a quantitative perspective, the statistical uncertainty surrounding the estimated
SC>2 exposure-response relationships due to sampling error is reflected in the credible intervals
that have been provided for the risk estimates in this document. Staff selected a mainly
qualitative approach to address other uncertainties in the assessment given the limited data
available to inform a probabilistic uncertainty characterization, and time and resource
constraints. Following the same general approach described in sections 6.6, 7.4, and 8.11.2 and
adapted from WHO (2008), staff performed a qualitative characterization of the components
contributing to uncertainty in the lung function risks for all asthmatics and asthmatic children
attributable to 5-minute SO2 exposures under moderate or greater exertion. First, staff identified
the important uncertainties. Then, we qualitatively characterized the magnitude (low., medium,
and high) and direction of influence (over, under, both, and unknown) the source of uncertainty
July 2009 345
-------
may have on the estimated number or percent of persons experiencing a defined lung function
response.96 Finally, staff also qualitatively rated the uncertainty in the knowledge-base regarding
each source using low, medium, and high categories. Staffs ratings were based on professional
judgment in the context of the knowledge-base for the criteria air pollutants.
Table 9-11 provides a summary of the sources of uncertainty identified in the health risk
assessment, the level of uncertainty, and the overall judged bias of each. A brief summary
discussion regarding those sources of uncertainty not already examined in Chapters 7 and 8 is
included in the comments section of Table 9-11.
The 5-minute daily maximum exposure estimates for asthmatics and asthmatic children
while engaged in moderate or greater exertion is an important input to the lung function response
risk assessment. A qualitative characterization of uncertainties associated with the exposure
model and the inputs to the exposure model are summarized in Table 8-27 and discussed in
section 8.11.2.
96 Definitions of the rating scales can be found in section 6.6.
July 2009 346
-------
Table 9-11. Characterization of key uncertainties in the lung function response health risk assessment for St. Louis and Greene County,
Missouri.
^rtiii*f*o f\f
ouuruc UT
Uncertainty
Exposure Model
(APEX) Inputs and
Algorithms
Spatial
representation
Air quality
adjustment
Causality
Use of 2-parameter
logistic and probit
probabilistic
exposure-response
relationships
Use of 5- and 10-
minute lung
function response
data to estimate 5-
I influence of Uncertainty on Lung
Function Risk Estimates
Direction
Unknown
Both
Both
Over
Unknown
O\/pr
\J VC? 1
Magnitude
Unknown
Medium
Low-Medium
Low-Medium
Low - for levels at
and above 100
ppb
Medium -for
levels below 100
ppb
I nw
LUVV
Knowledge-
R^ico
Dd9C
Uncertainty
Medium to High
High
Medium
Low - for levels
above 1 00 ppb
Medium -for
levels below
100 ppb
Low - for levels
above 100 ppb
Medium -for
levels below
100 ppb
I nw
HJVV
Comments1
See Table 8-27 and section 8.1 1 .2
See Table 7-16 and discussion in section 7.4.2.4
See Table 7-16 and discussion in section 7.4.2.5
INF: While there is very strong support for SO2 being causally linked to lung function
responses within the range of tested exposure levels (i.e., > 200 ppb) and even
down to thelOO ppb level (where SO2 was administered by mouthpiece (Sheppard
et al. 1981; Koenig et al., 1990)), there is increasing uncertainty about whether SO2
is causally related to lung-function effects at lower exposure levels below 100 ppb.
Since this assessment assumes there is a causal relationship at levels below 100
ppb, the influence of this source of uncertainty would be to over-estimate risk.
KB: The SO2-related lung function responses have been observed in controlled
human exposure studies and, thus there is little uncertainty that SO2 exposures are
responsible for the lung function responses observed for SO2 exposures in the
range of levels tested. Given the lack of chamber data at levels below 1 00 ppb, the
KB uncertainty is rated as medium.
KB: It was necessary to estimate responses at SO2 levels both within the range of
exposure levels tested (i.e., 200 to 1 ,000 ppb) as well as below the lowest exposure
levels used in free-breathing controlled human exposure studies (i.e., below 200
ppb). We have developed probabilistic exposure-response relationships using two
different functional forms (i.e., probit and 2-parameter logistic). Both functional
forms provide reasonable fits to the data in the available range of levels tested. For
the risks attributable to exposure levels below 200 ppb, the lowest level tested in
free-breathing chamber studies, and particularly below 100 ppb, the lowest level
tested in face mask chamber studies, there is greater uncertainty.
INF: It is reasonable to hypothesize that 10-minute exposures might lead to larger
lung function responses, so inclusion of 10-minute response data in the data base
used to estimate 5-minute responses would be more likely to result in over-
estimating risks. However, there is some evidence that responses generally occur
July 2009
347
-------
Source of
Uncertainty
Influence of Uncertainty on Lung
Function Risk Estimates
Direction
Magnitude
Knowledge-
Base
Uncertainty
Comments1
minute lung
function risk
estimates
in the first few minutes of exposure (see ISA, section 3.1.3.2), suggesting that any
overestimation is likely to be very modest in terms of magnitude.
KB: The 5-minute lung function risk estimates are based on a combined data set
from several controlled human exposure studies, most of which evaluated
responses associated with 10-minute exposures. However, since some studies
which evaluated responses after 5-minute exposures found responses occurring as
early as 5-minutes after exposure, we are using all of the 5- and 10-minute
exposure data to represent responses associated with 5-minute exposures. We do
not believe that this factor appreciably impacts the risk estimates.
Use of exposure-
response data from
studies of
mild/moderate
asthmatics to
represent all
asthmatics
Under
Medium
Medium
INF & KB: The data set that was used to estimate exposure-response relationships
included mild and/or moderate asthmatics. There is uncertainty with regard to how
well the population of mild and moderate asthmatics included in the series of SO2
controlled human exposure studies represent the distribution of mild and moderate
asthmatics in the U.S. population. As indicated in the ISA (p. 3-9), the subjects
studied represent the responses "among groups of relatively healthy asthmatics and
cannot necessarily be extrapolated to the most sensitive asthmatics in the
population who are likely more susceptible to the respiratory effects of exposure to
SO2." Thus, the influence of this uncertainty is likely to lead to under-estimating
risks and we judge the magnitude of the influence of this uncertainty on the lung
function risk estimates to be medium.
Reproducibility of
SO2-induced lung
function response
Unknown
Unknown
Low
INF & KB: The risk assessment assumes that the SO2-induced responses for
individuals are reproducible. We note that this assumption has some support in that
one study (Linn et al., 1987) exposed the same subjects on two occasions to 0.6
ppm and the authors reported a high degree of correlation (r > 0.7 for mild
asthmatics and r > 0.8 for moderate asthmatics, p < 0.001), while observing much
lower and nonsignificant correlations (r = 0.0 - 0.4) for the lung function response
observed in the clean air with exercise exposures.
Use of adult
asthmatic lung
function response
data to estimate
exposure-response
relationships for
asthmatic children
Unknown
Unknown
Low to Medium
INF & KB: Because the vast majority of controlled human exposure studies
investigating lung function responses were conducted with adult subjects, the risk
assessment relies on data from adult asthmatic subjects to estimate exposure-
response relationships that have been applied to all asthmatic individuals, including
children. The ISA (section 3.1.3.5) indicates that there is a strong body of evidence
that suggests adolescents may experience many of the same respiratory effects at
similar SO2 levels, but recognizes that these studies administered SO2 via inhalation
through a mouthpiece rather than an exposure chamber. This technique bypasses
nasal absorption of SO2 and can result in an increase in lung SO2 uptake.
July 2009
348
-------
Source of
Uncertainty
Influence of Uncertainty on Lung
Function Risk Estimates
Direction
Magnitude
Knowledge-
Base
Uncertainty
Comments1
Therefore, the uncertainty is greater in the risk estimates for asthmatic children. The
direction and magnitude of this uncertainty on the lung function risk estimates is
unknown.
Exposure history
Both
Low
Medium
INF & KB: The risk assessment assumes that the SO2-induced response on any
given day is independent of previous SO2 exposures. For some pollutants (e.g.,
ozone) prior exposure history can lead to both enhanced and diminished lung
function responses depending on the pattern of exposure. Since the assessment is
only included the highest daily 5-minute exposure under moderate or greater
exertion, and the influence of prior exposures might lead to either enhanced or
diminished response based on what we know about other pollutants (i.e., ozone),
staff rated the magnitude of the influence of this uncertainty to be low. Given the
lack of available information to directly assess this uncertainty for SO2 exposures in
chamber studies staff rated the KB uncertainty to be medium.
Assumed no
interaction effect of
other co-pollutants
on SO2-related
lung function
responses
Under
Medium
Medium
INF: Staff judges that it is more likely that exposure to other pollutants might
increase the magnitude of lung function response and thus increase the risk
estimates. Thus, assuming no interaction is more likely to result in under-estimating
risks.
KB: Because the controlled human exposure studies used in the risk assessment
involved only SO2 exposures, there is little information to judge whether or not
estimates of SO2-induced health responses are affected by the presence of other
pollutants (e.g., PM25, O3, NO2).
Notes:
1INF refers to comments associated with the influence rating; KB refers to comments associated with the knowledge-base rating.
I
July 2009
349
-------
9.5 KEY OBSERVATIONS
Presented below are key observations related to the risk assessment for lung function
responses in asthmatics and asthmatic children associated with 5-minute exposures to 862 while
engaged in moderate or greater exertion:
• Lung function responses estimated to result from 5-minute exposures to SC>2 were
estimated for two areas in Missouri (i.e., Greene County and St. Louis) which have
significant emission sources of 862 for 2002 air quality and for air quality adjusted to
simulate just meeting the current suite of annual and 24-hour SC>2 standards and just
meeting several alternative 1-hour daily maximum SC>2 standards.
• A number of factors would be expected to contribute to differences in estimated SO2-
related lung function responses across different locations. These include exposure-
related differences, such as population density, 862 emission density, location and types
of SC>2 sources, prevalence of air conditioning, and time spent outdoors, which are
discussed in section 8.10, as well as other factors such as differences in population
sensitivity to SO2 and asthma prevalence rates. As discussed in section 8.10, St. Louis
County has a medium to high 862 emission density and a medium to high population
density relative to other medium to high population density urban areas in the U.S.
Relative to the St. Louis study area, Greene County is a more rural county with much
lower SO2 emission density and much lower population density. Taken together, the risk
estimates for these two locations provide useful insights about urban and rural counties
with significant SC>2 emission sources.
• The lung function risk estimates for the St. Louis study area are much higher than for
Greene County, which is not unexpected given the greater population density and the
much greater SO2 emission density. Staff believes that the St. Louis risk estimates
provide a useful perspective on the likely overall magnitude and pattern of lung function
responses associated with various SC>2 air quality scenarios in urban areas within the U.S.
that have similar population densities and SC>2 emission densities.
• Risk estimates for Greene County are considerably lower than for the St. Louis study area
both with respect to estimated number of asthmatics and the percentage of asthmatics
estimated to experience one or more moderate or severe lung function responses. As
discussed above, this is not unexpected given the rural nature of Greene County and the
fact that it has much lower SC>2 emission density and lower population density than the
St. Louis study area.
• Of the alternative regulatory scenarios analyzed, only the 50 ppb/99th percentile daily
maximum 1-hr standard is estimated to reduce risks in one of the two modeling study
areas (i.e., St. Louis) relative to the "as is" air quality scenario. This reduction is
observed for both number and percent of asthmatics and asthmatic children estimated to
experience 1 or more lung function responses per year.
• For the St. Louis study area median risk estimates for 1 or more occurrences of moderate
lung function responses (i.e., based on sRaw > 100%) per year range from about 11%
down to 0.9% of asthmatic children using the 2-parameter logistic exposure-response
July 2009 350
-------
model compared to 10.4% down to 0.4% of asthmatic children using the probit exposure-
response model for alternative 99th percentile daily maximum 1-hour standards ranging
from 250 ppb down to 50 ppb. In general, the risk estimates associated with the use of
the probit exposure-response model are lower than those based on the logistic model.
• For the St. Louis study area median risk estimates for 1 or more occurrences of severe
lung function responses (i.e., based on sRaw > 200%) per year range from 4.2% down to
0.3% of asthmatic children using the 2-parameter logistic exposure-response model
compared to 3.7% down to 0.1% of asthmatic children using the probit exposure-
response model for alternative 99th percentile daily maximum 1-hour standars ranging
from 250 ppb down to 50 ppb.
• In terms of estimated percentage of asthmatics or asthmatic children experiencing 1 or
more lung function responses, risks are greater for asthmatic children, likely because they
spend more time at higher exertion levels than adults.
• A broad range of SC>2 exposure concentration intervals, as high as 500 ppb, contributes to
the estimated risks of experiencing 1 or more lung function responses per year for some
of the standards considered in the assessment. For standards in the range of 100 to 150
ppb SC>2 exposure concentration intervals below 200 ppb contribute most of the estimated
risks of experiencing 1 or more lung function response per year.
• Important uncertainties and limitations associated with the risk assessment which were
discussed above in section 9.3 and which should be kept in mind as one considers the
quantitative risk estimates include:
- uncertainties related to the exposure estimates which are an important input to the
risk assessment which staff rated as medium to high with respect to the knowledge
base and which staff rated the overall influence of these uncertainties on the
magnitude of the lung function risk estimates as unknown;
- uncertainties associated with the air quality adjustment procedure that was used to
simulate just meeting the current annual and several alternative 1-h daily maximum
standards which staff rated as medium with respect to the knowledge base uncertainty
and low-medium in terms of the influence of this uncertainty on the magnitude of the
lung function risk estimates;
- statistical uncertainty due to sampling error which is characterized in the
assessment through presentation of 95% credible intervals;
- uncertainty about the shape of the exposure-response relationship for lung function
responses at levels well below 200 ppb, the lowest level examined in free-breathing
single pollutant controlled human exposure studies which staff rated as low for levels
at and above 100 ppb and medium for levels below 100 ppb with respect to
knowledge base uncertainty and the influence of this uncertainty on the lung function
risk estimates;
- uncertainty with respect to how well the estimated exposure-response relationships
reflect asthmatics with more severe disease than those tested in chamber studies
which staff rated as medium with respect to knowledge base uncertainty and the
influence of this uncertainty on the magnitude of the lung function risk estimates;
July 2009 351
-------
- uncertainty about whether the presence of other pollutants in the ambient air would
enhance the SO2-related responses observed in the controlled human exposure studies
which staff rated as medium with respect to knowledge base uncertainty and the
influence of this uncertainty on the magnitude of the lung function risk estimates;
- uncertainty about the extent to which the risk estimates presented for the two
modeled areas in Missouri are representative of other locations in the U.S. with
significant 862 point and area sources which staff rated as high with respect to
knowledge base uncertainty and medium for the influence of this uncertainty on the
magnitude of the lung function risk estimates;
- other uncertainties such as the assumption about causality, use of both 5- and 10-
minute data to estimate 5-minute effects, the assumption of reproducible responses,
use of adult data to estimate exposure-response for children, and influence of
exposure history were generally rated as low to medium with respect to knowledge
base uncertainty and low or unknown impact on the magnitude of these uncertainties
on the lung function risk estimates.
July 2009 352
-------
10. EVIDENCE-AND EXPOSURE/RISK-BASED
CONSIDERATIONS RELATED TO THE PRIMARY SO2 NAAQS
10.1 INTRODUCTION
This chapter considers the scientific evidence in the ISA (EPA, 2008a) and the air
quality, exposure and risk characterization results presented in this document as they relate to the
adequacy of the current SO2 primary NAAQS and potential alternative primary SO2 standards.
The available scientific evidence includes epidemiologic, controlled human exposure, and animal
toxicological studies. The SO2 air quality, exposure, and risk analyses described in Chapters 7-9
of this document include characterization of air quality, exposure, and health risks associated
with recent SO2 concentrations and with SO2 concentrations adjusted to simulate scenarios just
meeting the current suite of standards and potential alternative 1-hour standards. In considering
the scientific evidence and the exposure- and risk-based information, we have also considered
relevant uncertainties. Section 10.2 of this chapter presents our general approach to considering
the adequacy of the current standards and the need for potential alternative standards. Sections
10.3 and 10.4 focus on evidence- and exposure-/risk-based considerations related to the
adequacy of the current 24-hour and annual standards respectively, while section 10.5 focuses on
such considerations related to the need for potential alternative standards (in terms of the
indicator, averaging time, form, and level).
These considerations are intended to inform the Agency's policy assessment of a range of
options with regard to the SO2 NAAQS. A final decision will draw upon scientific information
and analyses about health effects, population exposure and risks, and policy judgments about the
appropriate response to the range of uncertainties that are inherent in the scientific evidence and
air quality, exposure, and risk analyses. Our approach to informing these judgments, discussed
more fully below, is based on a recognition that the available health effects evidence reflects a
continuum consisting of ambient levels at which scientists generally agree that health effects are
likely to occur through lower levels at which the likelihood and magnitude of the response
become increasingly uncertain. This approach is consistent with the requirements of the
NAAQS provisions of the Act and with how EPA and the courts have historically interpreted the
Act. These provisions require the Administrator to establish primary standards that, in the
Administrator's judgment, are requisite to protect public health with an adequate margin of
safety. In so doing, the Administrator seeks to establish standards that are neither more nor less
July 2009 353
-------
stringent than necessary for this purpose. The Act does not require that primary standards be set
at a zero-risk level but rather at a level of protection that avoids unacceptable risks to public
health, including the health of at risk populations.
10.2 GENERAL APPROACH
This section describes the general approach that staff is taking to inform decisions
regarding the need to retain or revise the current SO2 NAAQS. The current standards, a 24-hour
average of 0.14 ppm (equivalent to 144 ppb), not to be exceeded more than one time per year,
and an annual average of 0.03 ppm (equivalent to 30.4 ppb) were retained by the Administrator
in the most recent review completed in 1996 (61 FR 25566). The decision to retain the 24-hour
standard was largely based on an assessment of epidemiologic studies that supported a likely
association between 24-hour average 862 exposure and daily mortality, aggravation of
bronchitis, and small, reversible declines in children's lung function (EPA 1982, 1994a).
Similarly, the decision to retain the annual standard (see section 10.4) was largely based on an
assessment of epidemiologic studies finding an association between respiratory
symptoms/illnesses and annual average SC>2 concentrations (EPA 1982, 1994a).
The previous review of the SC>2 NAAQS also addressed the question of whether an
additional short-term standard (e.g., 5-minute) was necessary to protect against short-term peak
SC>2 exposures. Based on the scientific evidence, the Administrator judged that repeated
exposures to 5-minute peak levels > 600 ppb could pose a risk of significant health effects for
asthmatic individuals at elevated ventilation rates (61 FR 25566). The Administrator also
concluded that the likely frequency of such effects should be a consideration in assessing the
overall public health risks. Based upon an exposure analysis conducted by EPA (see section
1.1.3), the Administrator concluded that exposure of asthmatics to SC>2 levels that could reliably
elicit adverse health effects was likely to be a rare event when viewed in the context of the entire
population of asthmatics, and therefore did not pose a broad public health problem for which a
NAAQS would be appropriate (61 FR 25566). On May 22, 1996, EPA published its final
decision to retain the existing 24-hour and annual standards and not to promulgate a 5-minute
standard (61 FR 25566). The decision not to set a 5-minute standard was ultimately challenged
by the American Lung Association and remanded back to EPA for further explanation on
January 30,1998 by the D.C. Circuit Court of Appeals (see section 1.1.1). Specifically, the court
gave EPA the opportunity to provide additional rationale to support the Agency judgment that 5-
July 2009 354
-------
minute peaks of 862 do not pose a public health problem when viewed from a national
perspective.
To inform the range of options that the Agency will consider in the current review of the
primary SC>2 NAAQS, the general approach we have adopted builds upon the approaches used in
reviews of other criteria pollutants, including the most recent reviews of the Pb, Os, PM, and
NO2 NAAQS (EPA, 2007i; EPA, 2007e; EPA, 2005, EPA 2008d). As in these other reviews, we
consider the implications of placing more or less weight or emphasis on different aspects of the
scientific evidence and the exposure/risk-based information, recognizing that the weight to be
given to various elements of the evidence and exposure/risk information is part of the public
health policy judgments that the Administrator will make in reaching decisions on the standards.
A series of general questions frames our approach to considering the scientific evidence
and exposure/risk-based information. First, our consideration of the scientific evidence and
exposure/risk-based information with regard to the adequacy of the current standards is framed
by the following questions:
• To what extent does evidence and exposure/risk-based information that has become
available since the last review reinforce or call into question evidence for SCVassociated
effects that were identified in the last review?
• To what extent has evidence for different health effects and/or sensitive populations
become available since the last review?
• To what extent have uncertainties identified in the last review been reduced and/or have
new uncertainties emerged?
• To what extent does evidence and exposure/risk-based information that has become
available since the last review reinforce or call into question any of the basic elements of
the current standards?
To the extent that the available evidence and exposure/risk-based information suggests it
may be appropriate to consider revision of the current standards, we consider that evidence and
information with regard to its support for consideration of standards that are either more or less
protective than the current standards. This evaluation is framed by the following questions:
• Is there evidence that associations, especially causal or likely causal associations,
extend to ambient SO2 concentrations as low as, or lower than, the concentrations that
have previously been associated with health effects? If so, what are the important
uncertainties associated with that evidence?
July 2009 355
-------
• Are exposures above benchmark levels and/or health risks estimated to occur in areas
that meet the current standards? If so, are the estimated exposures and health risks
important from a public health perspective? What are the important uncertainties
associated with the estimated risks?
To the extent that there is support for consideration of a revised standard, we then
consider the specific elements of the standard (indicator for gaseous SOX, averaging time, form,
and level) within the context of the currently available information. In so doing, we address the
following questions:
• Does the evidence provide support for considering a different indicator for gaseous SOX?
• Does the evidence provide support for considering different averaging times?
• What ranges of levels and forms of alternative standards are supported by the evidence,
and what are the associated uncertainties and limitations?
• To what extent do specific averaging times, levels, and forms of alternative standards
reduce the estimated exposures above benchmark levels and estimated risks attributable
to SO2, and what are the uncertainties associated with the estimated exposure and risk
reductions?
The following discussion addresses the questions outlined above and presents staffs
conclusions regarding the scientific evidence and the exposure-/risk-based information
specifically as they relate to the current and potential alternative standards. This discussion is
intended to inform the Agency's consideration of policy options that will be presented during the
rulemaking process, together with the scientific support for such options. Sections 10.3 and 10.4
consider the adequacy of the current standards while section 10.5 considers potential alternative
standards in terms of indicator, averaging time, form, and level. Each of these sections considers
key conclusions as well as the uncertainties associated with the evidence and exposure/risk
analyses.
10.3 ADEQUACY OF THE CURRENT 24-HOUR STANDARD
10.3.1 Introduction
In the last review of the 862 NAAQS, retention of the 24-hour standard was based
largely on epidemiologic studies conducted in London in the 1950's and 1960's. The results of
those studies suggested an association between 24-hour average levels of 862 and increased daily
mortality and aggravation of bronchitis when in the presence of elevated levels of PM (53 FR
14927). Additional epidemiologic evidence suggested that elevated SC>2 levels were associated
July 2009 356
-------
with the possibility of small, reversible declines in children's lung function (53 FR 14927).
However, it was noted that in the locations where these epidemiologic studies were conducted,
high SO2 levels were usually accompanied by high levels of PM, thus making it difficult to
disentangle the individual contribution each pollutant had on these health outcomes. It was also
noted that rather than 24-hour average SC>2 levels, the health effects observed in these studies
may have been related, at least in part, to the occurrence of shorter-term peaks of SC>2 within a
24-hour period (53 FR 14927).
In this review, as described in Chapter 4, the ISA concludes that there is sufficient
evidence to infer "a causal relationship between respiratory morbidity and short-term exposure to
SCV' (ISA, section 5.2). The ISA states that the strongest evidence for this judgment is from
human exposure studies demonstrating decreased lung function and/or increased respiratory
symptoms in exercising asthmatics exposed for 5-10 minutes to > 200 ppb SC>2 (ISA, section
5.2). Supporting this conclusion is a larger body of epidemiologic studies published since the
last review observing positive associations between 1-hour daily maximum or 24-hour average
SC>2 concentrations and respiratory symptoms, ED visits, and hospital admissions (ISA, section
5.2). Thus, the ISA bases its causal determination between short-term 862 exposure and
respiratory morbidity on respiratory effects associated with averaging times from 5-minutes to
24-hours.
Here, we will examine the health information first presented in Chapter 4 as it relates to
the adequacy of the current 24-hour standard (as well as the annual standard, see section 10.4).
Section 10.3.2 will discuss the epidemiologic results. The epidemiologic literature is particularly
relevant for evaluating the adequacy of the current 24-hour standard given that the majority of
these studies examined possible associations between 24-hour average SC>2 concentrations and
respiratory morbidity endpoints (e.g. ED visits or hospitalizations for all respiratory causes).
Section 10.3.3 will then discuss the air quality, exposure, and risk based information as it relates
to the adequacy of the current 24-hour standard. These analyses are first presented in Chapters
7-9 and describe exposures and their associated health risks given air quality just meeting the
current standards. More specifically, these analyses simulate air quality to just meet the current
24-hour or annual standard, whichever is controlling in a given area, and then describe exposure
and health risks associated with 5-minute SO2 benchmark concentrations. As described in
section 6.2, these benchmark concentrations are SC>2 exposure levels found in controlled human
July 2009 357
-------
exposure studies to result in decrements in lung function and/or respiratory symptoms in
exercising asthmatics. Finally, considering the evidence presented in section 10.3.2 and the air
quality, exposure, and risk information presented in section 10.3.3, staff presents conclusions
with regard to the overall adequacy of the current 24-hour standard in section 10.3.4.
10.3.2 Evidence-based considerations
As mentioned above, the ISA found supporting evidence for its conclusion that there is a
causal relationship between short-term SC>2 exposures and respiratory morbidity from the
reported associations observed in epidemiologic studies of respiratory symptoms and ED visits
and hospitalizations. In considering the adequacy of the current 24-hour standard, we note that
many epidemiologic studies demonstrating positive associations between ambient 862 and
respiratory symptoms, ED visits, and hospitalizations were conducted in areas where 862
concentrations were less than the level of the current 24-hour (as well as the annual; see section
10.4) NAAQS. With regard to these epidemiologic studies, we note that the ISA characterizes
the evidence for respiratory effects as consistent and coherent. The evidence is consistent in that
positive associations are reported in studies conducted in numerous locations and with a variety
of methodological approaches (ISA, section 5.2). It is coherent in the sense that respiratory
symptom results from epidemiologic studies predominantly using 1-hour daily maximum or 24-
hour average 862 concentrations are generally in agreement with the respiratory symptom results
from controlled human exposure studies of 5-10 minutes. These results are also coherent in that
the respiratory effects observed in controlled human exposure studies of 5-10 minutes provide a
basis for a progression of respiratory morbidity that could lead to the ED visits and
hospitalizations observed in epidemiologic studies (ISA, section 5.2).
However, it should be noted that interpretation of the epidemiologic literature is
complicated by the fact that SC>2 is but one component of a complex mixture of pollutants
present in the ambient air. The matter is further complicated by the fact that SC>2 is a precursor
to sulfate, which can be a principal component of PM. Ultimately, this uncertainty calls into
question the extent to which effect estimates from epidemiologic studies reflect the independent
contribution of 862 to the adverse respiratory outcomes assessed in these studies. In order to
provide some perspective on this uncertainty, the ISA evaluates epidemiologic studies that
employ multi-pollutant models. The ISA concludes that these analyses indicate that although
copollutant adjustment has varying degrees of influence on SC>2 effect estimates, the effect of
July 2009 358
-------
SO2 on respiratory health outcomes appears to be generally independent of the effects of gaseous
copollutants, including NC>2 and 63 (ISA, section 5.2). With respect to PMio, evidence of an
independent SO2 effect on respiratory health is less consistent, with some of the positive ED visit
and hospitalization results becoming negative (although results were not statistically significantly
negative) after inclusion of PMioin regression models (ISA, section 3.1.4.6). In epidemiologic
studies of respiratory symptoms, the SC>2 effect estimate often remained relatively unchanged
after inclusion of PMi0 in multipolutant models (although the effect estimate may have lost
statistical significance; ISA, section 3.1.4.1). The ISA also finds that SCVeffect estimates
generally remained relatively unchanged in the limited number of studies that included PM2.5
and/or PMio-2.5 in multipolutant models (ISA, section 3.1.4.6). Taken together, the ISA
concludes studies employing multi-pollutant models do suggest that SO2 has an independent
effect on respiratory morbidity outcomes (see Chapter 4; ISA, section 5.2). Thus, the results of
experimental and epidemiologic studies form a plausible and coherent data set that supports a
relationship between SC>2 exposures and respiratory morbidity endpoints, and calls into question
the adequacy of the 24-hour standard to protect public health.
10.3.3 Air Quality, exposure and risk-based considerations
In addition to the evidence-based considerations described above, staff has considered the
extent to which exposure- and risk-based information can inform decisions regarding the
adequacy of the current 24-hour 862 standard, taking into account key uncertainties associated
with the estimated exposures and risks. For this review, we have employed three approaches.
In the first approach, SO2 air quality levels were used as a surrogate for exposure. In the second
approach, modeled estimates of human exposure were developed for all asthmatics and asthmatic
children living in Greene County and St. Louis MO. Notably, this second approach considers
time spent in different microenvironments, as well as time spent at elevated ventilation rates. In
each of the first two approaches, health risks have been characterized by comparing estimates of
air quality or exposure to 5-minute potential health effect benchmarks. These benchmarks are
based on controlled human exposure studies involving known 5-10 minute 862 exposure levels
and corresponding decrements in lung function, and/or increases in respiratory symptoms in
asthmatics at elevated ventilation rates (e.g., while exercising; see section 6.2 for further
discussion of benchmark levels). In addition to these analyses, staff conducted a quantitative risk
assessment for lung function responses associated with 5-minute exposures to characterize SO2-
July 2009 359
-------
related health risks. This assessment combined outputs from the exposure analysis with
estimated exposure-response functions derived from the combined individual data from
controlled human exposure studies to estimate the number and percent of exposed asthmatics
that would experience moderate or greater lung function responses (in terms of FEVi and sRaw)
at least once per year and to estimate the total number of occurrences of these lung function
responses per year (see Chapter 9).
The respiratory effects (i.e., decrements in FEVi, increases in sRaw, and/or respiratory
symptoms) considered in the air quality, exposure, and risk analyses mentioned above are
considered by staff to be adverse to the health of asthmatics. As described in section 4.3, staff
bases this conclusion on: 1) guidelines published by the ATS; 2) conclusions from the ISA and
previous NAAQS reviews; and 3) advice from CAS AC. Being mindful of this conclusion, we
note the following key points from the ISA:
• Approximately 5-30% of exercising asthmatics are expected to experience moderate or
greater lung function decrements (i.e., > 100% increase in sRaw and/or a > 15% decrease
in FEVi) following exposure to 200- 300 ppb SC>2 for 5-10 minutes (ISA, section 3.1).
• Approximately 20-60% of exercising asthmatics are expected to experience moderate or
greater lung function decrements (i.e > 100% increase in sRaw and/or a > 15% decrease
in FEVi) following exposure to 400-1000 ppb SO2 for 5-10 minutes (ISA, Table 5-3).
• At concentrations > 400 ppb, moderate or greater statistically significant decrements in
lung function are frequently associated with respiratory symptoms (ISA, section 3.1).
• There is no evidence to indicate that exposure to 200-300 ppb SC>2 for 5- 10 minutes
represents a threshold below which no respiratory effects occur.
Given the discussion in section 4.3 and the key points presented above, staff concludes
that exposure to 5-10 minute SC>2 concentrations at least as low as 200 ppb can result in adverse
respiratory effects in some asthmatics. We note that this conclusion is in agreement with
CASAC comments offered on the first draft SC>2 REA. The CASAC letter to the Administrator
states: "CASAC believes strongly that the weight of clinical and epidemiology evidence
indicates there are detectable clinically relevant health effects in sensitive subpopulations down
to a level at least as low as 0.2 ppm SO2 (Henderson 2008)." This CASAC letter also states:
"these sensitive subpopulations represent a substantial segment of the at-risk population
(Henderson 2008)." As an additional matter, we note that over 20 million people in the U.S.
have asthma (EPA 2008d), and therefore, exposure to SC>2 likely represents a significant public
health issue.
July 2009 360
-------
Thus, staff finds it is appropriate to consider the air quality, exposure and risk results as
they relate to the adequacy of the current 24-hour standard (as well as the current annual (see
section 10.4) and potential alternative (see section 10.5) standards). This is because these
analyses provide useful information with respect to the current 24-hour standard's ability to
limit: 1) 5-10 minute SC>2 concentrations associated with decrements in lung function and/or
respiratory symptoms in exercising asthmatics; and 2) the estimated number of exercising
asthmatics expected to experience a moderate or greater lung function response.
10.3.3.1 Key Uncertainties
The way in which air quality, exposure, and risk results will inform ultimate decisions
regarding the 862 standard will depend upon the weight placed on each of the analyses when
uncertainties associated with those analyses are taken into consideration. Sources of uncertainty
associated with each of the analyses (air quality, exposure, and quantitative risk) are briefly
presented below and are described in more detail in Chapters 7-9 of this document. Although
we are discussing these uncertainties within the context of the adequacy of the 24-hour standard,
they apply equally to consideration of the annual, as well as alternative 1-hour standards.
Air Quality Analysis
A number of key uncertainties should be considered when interpreting air quality results
with regard to decisions on the standards. A general description of such uncertainties is
highlighted below, and these, as well as other sources of uncertainty are discussed in greater
depth in section 7.4 of this document.
• Staff used the broader SC>2 ambient monitoring network, in addition to subsets of data
from this network, to characterize air quality in the U.S. There was general agreement in
the monitor site attributes and emissions sources potentially influencing ambient
monitoring concentrations for each set of data analyzed. However, staff noted that the
greatest uncertainty, compared to several other sources of uncertainty, was in the spatial
representativeness of both the overall monitoring network and the subsets chosen for
detailed analyses.
• Staff developed a statistical model to estimate 5-minute maximum SO2 concentrations at
monitors that reported only 1-hour 862 concentrations. Cross-validation of the statistical
model for where 5-minute SC>2 measurements existed indicated reasonable model
performance. The greatest difference in the predicted versus observed numbers of
benchmark exceedances occurred at the lower and upper tails of the distribution,
indicating greater uncertainty in the predictions at similarly representative monitors.
July 2009 361
-------
• The air quality characterization assumes that the ambient monitoring data and the
estimated days per year with benchmark exceedances can serve as an indicator of
exposure. Longer-term personal SO2 exposure (i.e., days to weeks) concentrations are
correlated with and are a fraction of ambient 862 concentrations. However, uncertainty
remains in this relationship when considering short-term (i.e., 5-minute) averaging times
because of the lack of comparable measurement data.
St Louis and Greene Counties Exposure Analysis
A number of key uncertainties should be considered when interpreting the St. Louis and
Greene County exposure results with regard to decisions on the standards. Such uncertainties are
highlighted below, and these, as well as other sources of uncertainty, are also discussed in greater
depth in section 8.11 of this document.
• It was necessary for staff to derive an area source emission profile rather than use a
default profile to improve the agreement between ambient measurements and predicted 1-
hour SO2 concentrations. The improved model performance reduces uncertainty in the 1-
hour SC>2 concentrations predictions, but nonetheless remains as an important uncertainty
in the absence of actual local source emission profiles.
• Staff performed the exposure assessment to better reflect both the temporal and spatial
representation of ambient concentrations and to estimate the rate of contact of individuals
with 5-minute SC>2 concentrations while engaged in moderate or greater exertion.
Estimated annual average 862 exposures in the two exposure modeling domains are
consistent with long-term personal exposures (i.e., days to weeks) measured in other U.S.
locations. However, uncertainty remains in the estimated number of persons with 5-
minute SC>2 concentrations above benchmark levels because of the lack of comparable
measurement data, particularly considering both the short-term averaging time and
geographic location.
• While all 5-minute ambient 862 concentrations were estimated by the exposure model,
each hour was comprised of the maximum 5-minute SC>2 concentration and eleven other
5-minute SC>2 concentrations normalized to the 1-hour mean concentration. Staff
assumed that this approach would reasonably estimate the number of individuals exposed
to peak concentrations. Sensitivity analyses revealed that both the number of persons
exposed and where peak exposures occur can vary when considering an actual 5-minute
temporal profile.
July 2009 362
-------
St Louis and Greene Counties Quantitative Risk Analysis
A number of key uncertainties should be considered when interpreting the St. Louis and
Greene County quantitative risk estimated for lung function responses with regard to decisions
on the standards. Such uncertainties are highlighted below, and these, as well as other sources of
uncertainty, are also discussed in greater depth in section 9.3 of this document.
• It was necessary to estimate responses at 862 levels below the lowest exposure levels
used in the free-breathing controlled human exposure studies (i.e., below 200 ppb). We
have developed probabilistic exposure-response relationships using two different
functional forms (i.e., probit and 2-parameter logistic), but nonetheless there remains
greater uncertainty in responses below 200 ppb because of the lack of comparable
experimental data.
• The risk assessment assumes that the SO2-induced responses for individuals are
reproducible. We note that this assumption has some support in that one study (Linn et
al., 1987) exposed the same subjects on two occasions to 600 ppb and the authors
reported a high degree of correlation while observing a much lower correlation for the
lung function response observed in the clean air with exercise exposure.
• Because the vast majority of controlled human exposure studies investigating lung
function responses were conducted with adult subjects, the risk assessment relies on data
from adult asthmatic subjects to estimate exposure-response relationships that have been
applied to all asthmatic individuals, including children. The ISA (section 3.1.3.5)
indicates that there is a strong body of evidence that suggests adolescents may experience
many of the same respiratory effects at similar SC>2 levels, but recognizes that these
studies administered SO2 via inhalation through a mouthpiece (which can result in an
increase in lung 862 uptake) rather than an exposure chamber. Therefore, the uncertainty
is greater in the risk estimates for asthmatic children.
• Because the controlled human exposure studies used in the risk assessment involved only
SC>2 exposures, it is assumed that estimates of SC>2-induced health responses are not
affected by the presence of other pollutants (e.g., PM2.5, O3, NO2).
10.3.3.2 Assessment Results
As previously mentioned, the ISA finds the evidence for an association between
respiratory morbidity and 862 exposure to be "sufficient to infer a causal relationship" (ISA,
section 5.2) and that the "definitive evidence" for this conclusion comes from the results of
controlled human exposure studies demonstrating decrements in lung function and/or respiratory
symptoms in exercising asthmatics (ISA, section 5.2). Accordingly, the exposure and risk
analyses presented in this document focused on exposures and risks associated with 5-minute
peaks of SC>2 in excess of the potential health effect benchmark values of 100, 200, 300, and 400
ppb SC>2 (see section 6.2). In considering the results presented in these analyses, we particularly
July 2009 363
-------
note exceedances or exposures with respect to the 200 and 400 ppb 5-minute benchmark levels.
We highlight these benchmark levels because (1) 400 ppb represents the lowest concentration in
human exposure studies where statistically significant moderate or greater lung function
decrements are frequently accompanied by respiratory symptoms; (2) 200 ppb is the lowest level
at which effects have been observed (and the lowest level tested) for moderate or greater
decrements in lung function in free-breathing human exposure studies. Notably, we also
recognize that there is very limited evidence demonstrating small decrements in lung function at
100 ppb from two mouthpiece exposure studies (see section 6.2). However, as previously noted
(see section 6.2), the results of these studies are not directly comparable to free-breathing
chamber studies, and thus, staff is primarily considering exceedances of the 200 ppb and 400 ppb
benchmark levels in its evaluation of the adequacy of the current standards.
Exposures and risks have been estimated for two study areas in Missouri (i.e., Greene
County and several counties representing the St. Louis urban area) which have significant
emission sources of SC>2. As noted in section 8.10, there were differences in the number of
exposures above benchmark values when the results of the Greene County and St. Louis
exposure assessments were compared. Moreover, given that the results of the exposure
assessment were used as inputs into the quantitative risk assessment, it was not surprising that
there were also far fewer asthmatics at elevated ventilation rates estimated to have a moderate or
greater lung function response in Greene county when compared to St. Louis. The difference in
the St. Louis and Greene County exposure and quantitative risk results are likely indicative of the
different types of locations they represent (see section 8.10). Greene County is a rural county
with much lower population and emission densities, compared to the St. Louis study area which
has population and emissions density similar to other urban areas in the U.S. It therefore follows
that there would be greater exposures, and hence greater numbers and percentages of asthmatics
at elevated ventilation rates experiencing moderate or greater lung function responses in the St.
Louis study area. Thus, when considering the risk and exposure results as they relate to the
adequacy of the current standards (as well as the need for considering potential alternative
standards), the St. Louis results are more informative in that they suggest that the current
standards may not adequately protect public health. Moreover, staff judges that the exposure and
risk estimates for the St. Louis study area provide useful insights into exposures and risks for
other urban areas in the U.S. with similar population and SC>2 emissions densities.
July 2009 364
-------
Air Quality Assessment
The results of our air quality assessment provide additional perspective on the public
health impacts of exposure to ambient levels of SO2. In considering these results, we first note
that the benchmark values derived from the controlled human exposure literature are associated
with a 5-minute averaging time, but very few state and local agencies in the U.S. report
measured 5-minute concentrations since such monitoring is not required. As a result, staff
developed a statistical relationship to estimate the highest 5-minute level in an hour, given a
reported 1-hour average 862 concentration (see section 7.2.3). Thus, many of the outputs of the
air quality analysis are presented with respect to statistically estimated 5-minute concentrations
in excess of potential health effect benchmark values. Results of these analyses, as they relate to
the adequacy of the current standards, are discussed below.
A key output of the air quality analysis is the predicted number of statistically estimated
5-minute daily maximum SC>2 concentrations above benchmark levels given air quality simulated
to just meet the level of the current 24-hour or annual SC>2 standards, whichever is controlling for
a given county. Under this scenario, in 40 counties selected for detailed analysis, we note that
the predicted yearly mean number of statistically estimated 5-minute daily maximum
concentrations > 400 ppb ranges from 1-102 days per year97, with most counties in this analysis
experiencing a mean of at least 20 days per year when statistically estimated 5-minute daily 862
concentrations exceed 400 ppb (Table 7-14). In addition, the predicted yearly mean number of
statistically estimated 5-minute daily maximum concentrations >200 ppb ranges from 21-171
days per year, with about half of the counties in this analysis experiencing > 70 days per year
when 5-minute daily maximum SC>2 concentrations exceed 200 ppb (Table 7-12).
Exposure Assessment
When considering the St. Louis exposure results as they relate to the adequacy of the
current standard, we focus on the number of asthmatics at elevated ventilation rates estimated to
experience at least one benchmark exceedance given air quality that is adjusted upward to
simulate just meeting the current 24-hour standard (i.e., the controlling standard in St. Louis).
We note that in these analyses, if 862 concentrations are such that the St Louis area just meets
the current standard, approximately 13% of asthmatics would be estimated to experience at least
97 Air quality estimates presented in this section represent the mean number of days per year when 5-minute daily
maximum SO2 concentrations exceed a particular benchmark level given 2001-2006 air quality adjusted to just meet
the current standards (see Tables 7-11 to 7-14).
July 2009 365
-------
one SC>2 exposure concentration greater than or equal to a 400 ppb benchmark level while at
elevated ventilation rates (Figure 8-19). Similarly, approximately 46% of asthmatics would be
expected to experience at least one SO2 exposure concentration greater than or equal to a 200
ppb benchmark level while at elevated ventilation rates. When the St. Louis results are restricted
to asthmatic children at elevated ventilation rates, approximately 25% and 73% of these children
would be estimated to experience at least one SC>2 exposure concentration greater than or equal
to the 400 ppb and 200 ppb benchmark levels, respectively (Figure 8-19).
Risk results
When considering the St. Louis risk results as they relate to the adequacy of the current
standard, we note the percent of asthmatics at elevated ventilation rates likely to experience at
least one lung function response given air quality that is adjusted upward to simulate just
meeting the current standards. Under this scenario, 12.7% to 13.1% of exposed asthmatics at
elevated ventilation rates are estimated to experience at least one moderate lung function
response (defined as an increase in sRaw > 100% (Table 9-5))98. Furthermore, 5.1% to 5.4% of
exposed asthmatics at elevated ventilation rates are estimated to experience at least one large
lung function response (defined as an increase in sRaw > 200% (Table 9-5)). We also note that
estimates from this analysis indicate that the percentage of exposed asthmatic children in St.
Louis estimated to experience at least one moderate or large lung function response is somewhat
greater than the percentage for the asthmatic population as a whole (Table 9-8). In addition, we
note that comparable results were observed when moderate or greater lung function responses
were defined in terms of FEVi.
10.3.4 Conclusions regarding the adequacy of the 24-hour standard
As noted above, several lines of scientific evidence are relevant to consider in evaluating
the adequacy of the current 24-hour standard to protect the public health. These include
causality judgments made in the ISA, as well as the human exposure and epidemiologic evidence
supporting those judgments. In particular, we note that numerous epidemiologic studies
reporting positive associations between ambient 862 and respiratory morbidity endpoints were
conducted in locations that met the current 24-hour standard. To the extent that these
98 The risk results presented represent the median estimate of exposed asthmatics expected to experience moderate
or greater lung function decrements. Results are presented for both the probit and 2-parameter logistic functional
forms. The full range of estimates can be found in Chapter 9, and in all instances the smaller estimate is a result of
using the probit function to estimate the exposure-response relationship.
July 2009 366
-------
considerations are emphasized, the adequacy of the current standard to protect the public health
would clearly be called into question. This suggests consideration of a revised 24-hour standard
and/or that an additional shorter-averaging time standard may be needed to provide additional
health protection for sensitive groups, including asthmatics and individuals who spend time
outdoors at elevated ventilation rates. Moreover, this also suggests that an alternative SC>2
standard(s) should protect against health effects ranging from lung function responses and
increased respiratory symptoms following 5-10 minute peak SC>2 exposures, to increased
respiratory symptoms and respiratory-related ED visits and hospital admissions associated with
1-hour daily maximum or 24-hour average SC>2 concentrations.
In examining the exposure- and risk-based information with regard to the adequacy of the
current 24-hour SO2 standard to protect the public health, we note that the results described
above (and in more detail in Chapters 7-9) indicate that 5-minute exposures that can reasonably
be judged important from a public health perspective are associated with air quality adjusted
upward to simulate just meeting the current 24-hour standard. Therefore, exposure- and risk-
based considerations reinforce the scientific evidence in supporting the conclusion that
consideration should be given to revising the current 24-hour standard and/or setting a new
shorter averaging time standard (e.g., 1-hour or less) to provide increased public health
protection, especially for sensitive groups (e.g., asthmatics), from SCVrelated adverse health
effects.
10.4 ADEQUACY OF THE CURRENT ANNUAL STANDARD
10.4.1 Introduction
In the last review of the 862 NAAQS, retention of the annual standard was largely based
on an assessment of qualitative evidence gathered from a limited number of epidemiologic
studies. The strongest evidence for an association between annual 862 concentrations and
adverse health effects in the 1982 AQCD was from a study conducted by Lunn et al (1967). The
authors found that among children a likely association existed between chronic upper and lower
respiratory tract illnesses and annual SC>2 levels of 70 -100 ppb in the presence of 230-301 ug/m3
black smoke. Three additional studies described in the 1986 Second Addendum also suggested
that long-term exposure to SC>2 was associated with adverse respiratory effects. Notably, studies
conducted by Chapman et al. (1985) and Dodge et al. (1985) found associations between long-
July 2009 367
-------
term 862 concentrations (with or without high particle concentrations) and cough in children and
young adults. However, it was noted that there was considerable uncertainty associated with
these studies because they were conducted in locations subject to high, short-term peak SO2
concentrations (i.e., locations near point sources); therefore it was difficult to discern whether
this increase in cough was the result of long-term, low level SC>2 exposure, or repeated short-
term peak SC>2 exposures.
It was concluded in the last review that there was no quantitative rationale to support a
specific range for an annual standard (EPA, 1994b). However, it was also found that while no
single epidemiologic study provided clear quantitative conclusions, there appeared to be some
consistency across studies indicating the possibility of respiratory effects associated with long-
term exposure to SO2 just above the level of the existing annual standard (EPA, 1994b). In
addition, air quality analyses conducted during the last review indicated that the short-term
standards being considered (1-hour and/or 24-hour) could not by themselves prevent long-term
concentrations of SC>2 from exceeding the level of the existing annual standard in several large
urban areas. Ultimately, both the scientific evidence and the air quality analyses were used by
the Administrator to conclude that retaining the existing annual standard was requisite to protect
human health.
10.4.2 Evidence-based considerations
The ISA presents numerous studies published since the last review examining possible
associations between long-term 862 exposure and mortality and morbidity outcomes. This
includes discussion of additional epidemiologic studies examining possible associations between
long-term SC>2 exposure and respiratory effects in children (in part, the basis for retaining the
annual standard in the last review; see section 10.4.1). In addition, the ISA presents results from
epidemiologic and animal toxicological studies published since the last review examining
possible associations between long-term ambient SC>2 concentrations and adverse respiratory,
cardiovascular, and birth outcomes, as well as carcinogenesis. The current ISA also discusses
the possible association between long-term 862 exposure and mortality.
As an initial consideration with regard to the adequacy of the current annual standard,
staff notes that the evidence relating long-term (weeks to years) SO2 exposure to adverse health
effects (respiratory morbidity, carcinogenesis, adverse prenatal and neonatal outcomes, and
mortality) is judged to be "inadequate to infer the presence or absence of a causal relationship"
July 2009 368
-------
(ISA, Table 5-3). That is, the ISA finds this health evidence to be of insufficient quantity,
quality, consistency, or statistical power to make a determination as to whether 862 is truly
associated with these health endpoints (ISA, Table 1-2). With respect specifically to respiratory
morbidity in children, the ISA presents recent epidemiologic evidence of an association with
long-term exposure to SC>2 (ISA, section 3.4.2). However, the ISA finds the strength of these
epidemiologic studies to be limited because of 1) variability in results across studies with respect
to specific respiratory morbidity endpoints, 2) high correlations between long-term average SC>2
and co-pollutant concentrations, particularly PM, and 3) a lack of evaluation of potential
confounding (ISA, section 3.4.2.1).
We also note that many epidemiologic studies demonstrating positive associations
between 1-hour daily maximum or 24-hour average SO2 concentrations and respiratory
symptoms, ED visits, and hospitalizations were conducted in areas where ambient SC>2
concentrations were well below the current annual NAAQS. This evidence suggests that the
current annual standard is not providing adequate protection against health effects associated
with shorter-term SC>2 concentrations.
10.4.3 Risk-based considerations
Results of the risk characterization based on the air quality assessment provide additional
insight into the adequacy of the current annual standard. Analyses in this document describe the
extent to which the current annual standard provides protection against 5-minute peaks of 862 in
excess of potential health effect benchmark levels. Figure 7-16 counts the number of measured
5-minute daily maximum SC>2 concentrations above the 100 -400 ppb benchmark levels for a
given annual average SC>2 concentration. None of the monitors in this data set reported annual
average SC>2 concentrations above the current NAAQS, but several of the monitors in several of
the years frequently reported 5-minute daily maximum concentrations above the potential health
effect benchmark levels. Many of these monitors where frequent exceedances were reported had
annual average 862 concentrations between 5 and 15 ppb, with little to no correlation between
the annual average 862 concentration and the number of 5-minute daily maximum
concentrations above potential health effect benchmark levels. This suggests that the annual
standard adds little in the way of protection against 5-minute peaks of SO2 (see section 7.3.1).
July 2009 369
-------
10.4.4 Conclusions regarding the adequacy of the current annual standard
As noted above, the ISA concludes that the evidence relating long-term (weeks to years)
SC>2 exposure to adverse health effects (respiratory morbidity, carcinogenesis, adverse prenatal
and neonatal outcomes, and mortality) is "inadequate to infer the presence or absence of a causal
relationship" (ISA, Table 5-3). The ISA also reports that many epidemiologic studies
demonstrating positive associations between short-term (i.e. 1-hour daily maximum, 24-hour
average) SC>2 concentrations and respiratory symptoms, as well as ED visits and hospitalizations,
were conducted in areas where annual ambient SC>2 concentrations were well below the level of
the current annual NAAQS. In addition, analyses conducted in this REA suggest that the
current annual standard is not providing protection against 5-10 minute peaks of 862. Thus, the
scientific evidence and the risk and exposure information suggest that the current annual 862
standard: 1) is likely not needed to protect against health risks associated with long term
exposure to 802; and 2) does not provide adequate protection from the health effects associated
with shorter-term (i.e. < 24-hours). This suggests that consideration should be given to either
revoking the annual standard or retaining it without revision, in conjunction with setting an
appropriate short-term standard(s).
10.5 POTENTIAL ALTERNATIVE STANDARDS
10.5.1 Indicator
In the last review, EPA focused on SC>2 as the most appropriate indicator for ambient
SOX. This was in large part because other gaseous sulfur oxides (e.g., SOs) are likely to be found
at concentrations many orders of magnitude lower than SC>2 in the atmosphere, and because most
all of the health effects and exposure information was for SC>2. The current ISA has again found
this to the case, and although the presence of gaseous SOX species other than 862 has been
recognized, no alternative to 862 has been advanced as being a more appropriate surrogate for
ambient gaseous SOX. Importantly, controlled human exposure studies and animal toxicology
studies provide specific evidence for health effects following exposure to SO2. Epidemiologic
studies also typically report levels of SC>2, as opposed to other gaseous SOX. Because emissions
that lead to the formation of SC>2 generally also lead to the formation of other SOX oxidation
products, measures leading to reductions in population exposures to SC>2 can generally be
expected to lead to reductions in population exposures to other gaseous SOX. Therefore, meeting
an SC>2 standard that protects the public health can also be expected to provide some degree of
July 2009 370
-------
protection against potential health effects that may be independently associated with other
gaseous SOX even though such effects are not discernable from currently available studies
indexed by SO2 alone. Given these key points, staff judges that the available evidence supports
the retention of SC>2 as the indicator in the current review. We also note that this would be in
agreement with CAS AC comments offered on the second draft REA. The consensus CAS AC
response to Agency charge questions from the second draft REA states: "For indicator, SC>2 is
clearly the preferred choice (Samet 2009)."
10.5.2 Averaging Time
EPA established the current 24-hour and annual averaging times for the primary 862
NAAQS in 1971. As previously described, (see section 10.3.1) the 24-hour NAAQS was based
on epidemiologic studies that observed associations between 24-hour average 862 levels and
adverse respiratory effects and daily mortality (EPA 1982, 1994b). The annual standard was
supported by a few epidemiologic studies that found an association between adverse respiratory
effects and annual average SC>2 concentrations (EPA 1982, 1994b). Based on currently available
evidence, staff concludes that different averaging time(s) be established for the primary
standard(s) as part of the current review. In reaching this conclusion, staff has considered
causality judgments from the ISA, results from controlled human exposure and epidemiologic
studies, and 862 air quality correlations. These considerations are described in more detail
below.
10.5.2.1 Evidence-based considerations
As an initial consideration regarding the most appropriate averaging time (e.g., short-
term, long-term, or a combination of both) for alternative SO2 standard(s), we note (as in 10.4.1
above) that the ISA finds evidence relating long-term (weeks to years) SC>2 exposures to adverse
health effects to be "inadequate to infer the presence or absence of a causal relationship" (ISA,
Table 5-3). In contrast, the ISA judges evidence relating short-term (5-minutes to 24-hours) SC>2
exposure to respiratory morbidity to be "sufficient to infer a causal relationship" and short-term
exposure to 862 and mortality to be "suggestive of a causal relationship" (ISA, Table 5-3).
Taken together, these judgments most directly support standard averaging time(s) that focus
protection on 862 exposures from 5-minutes to 24-hours.
July 2009 371
-------
In considering the level of support available for specific short-term averaging times, we
first note the strength of evidence from human exposure and epidemiologic studies. Controlled
human exposure studies exposed exercising asthmatics to 5-10 minute peak concentrations of
SC>2 and consistently found decrements in lung function and/or respiratory symptoms.
Importantly, the ISA describes the controlled human exposure studies as being the "definitive
evidence" for its conclusion that there is a causal association between short-term (5-minutes to
24-hours) SC>2 exposure and respiratory morbidity (ISA, section 5.2). Supporting the controlled
human exposure evidence is a relatively small body of epidemiologic studies describing positive
associations between 1-hour daily maximum 862 levels and respiratory symptoms as well as
hospital admissions and ED visits for all respiratory causes and asthma (ISA Tables 5.4 and 5.5).
In addition to the 1-hour daily maximum epidemiologic evidence, there is a considerably larger
body of epidemiologic studies reporting positive associations between 24-hour average SC>2
levels and respiratory symptoms, as well as hospitalizations and ED visits for all respiratory
causes and asthma. However, as in the last review, there remains considerable uncertainty as to
whether these positive associations are due to 24-hour average SC>2 exposures, or exposure (or
multiple exposures) to short-term peaks of 862 within a 24-hour period. More specifically, when
describing epidemiologic studies observing positive associations between ambient 862 and
respiratory symptoms, the ISA states "that it is possible that these associations are determined in
large part by peak exposures within a 24-hour period" (ISA, section 5.2). The ISA also states
that the respiratory effects following 5-10 minute SO2 exposures in controlled human exposure
studies provides a basis for a progression of respiratory morbidity that could result in increased
ED visits and hospital admissions (ISA, section 5.2).
The controlled human exposure evidence described above provides support for an
averaging time that protects against 5-10 minute peak exposures. Results from the
epidemiologic evidence provides support for both 1-hour and 24-hour averaging times.
However, it is worth noting again that the effects observed in epidemiologic studies also may be
due, at least in part and especially in 24-hour epidemiologic studies, to shorter-term peaks of
SC>2. Overall, the evidence mentioned above suggests that a primary concern with regard to
averaging time is the level of protection provided against 5-10 minute peak SC>2 exposures. The
evidence described above also suggests it would be appropriate to consider the degree of
July 2009 372
-------
protection averaging times under consideration provide against both 1-hour daily maximum and
24-hour average 862 concentrations.
10.5.2.2 Air Quality considerations
The shortest averaging time for the current primary SO2 standard is 24-hours. We
therefore evaluate the potential for a standard based on 24-hour average SC>2 concentrations to
limit 5-minute peak SC>2 exposures. Table 10-1 reports the ratio between 99th percentile 5-
minute daily maximum and 99th percentile 24-hour average SC>2 concentrations for 42 monitors
reporting measured 5-minute data for any year between 2004-2006. Across this set of monitors
in 2004, ratios of 99th percentile 5-minute daily maximum to 99th percentile 24-hour average SC>2
concentrations spanned a range of 2.0 to 14.1, with an average ratio of 6.7 (Table 10-1). These
results suggest that a standard based on 24-hour average 862 concentrations would not likely be
an effective or efficient approach for addressing 5-minute peak SO2 concentrations. That is,
using a 24-hour average standard to address 5-minute peaks would likely result in over-
controlling in some areas, while under-controlling in others. This analysis also suggests that a 5-
minute standard would not likely be an effective or efficient means for controlling 24-hour
average SC>2 concentrations.
Table 10-1 also reports the ratios between 99th percentile 5-minute daily maximum and
99th percentile 1-hour daily maximum SC>2 levels from this set of monitors. Compared to the
ratios discussed above (5-minute daily maximum to 24-hour average), there is far less variability
between 5-minute daily maximum and 1-hour daily maximum ratios. More specifically, 39 of
the 42 monitors had 99th percentile 5-minute daily maximum to 99th percentile 1-hour daily
maximum ratios in the range of 1.2 to 2.5 (Table 10-1). The remaining 3 monitors had ratios of
3.6, 4.2 and 4.6 respectively. Overall, this relatively narrow range of ratios suggests that a
standard with a 1-hour averaging time would be more efficient and effective at limiting 5-minute
peaks of SC>2 than a standard with a 24-hour averaging time. These results also suggest that a 5-
minute standard could be a relatively effective means of controlling 1-hour daily maximum 862
concentrations.
July 2009 373
-------
Table 10-1 Ratios of 99th percentile 5-minute daily maximums to 99th percentile 24-hour average
and 1-hour daily maximum SO2 concentrations for monitors reporting measured 5-minute data
from years 2004-2006"
99
Monitor ID
110010041
191770005
290930030
290930031
370670022
120890005
190330018
190450019
191390016
191390017
191390020
191630015
191770006
291630002
380130002
380150003
380590002
380590003
540990003
540990004
540990005
541071002
051190007
051390006
080310002
290770026
290770037
290990004
291370001
301110084
380070002
380130004
380171004
380250003
380530002
380530104
380530111
380570004
380650002
381050103
381050105
420070005
# of years
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5-minute daily max:
24-hour average
3.8
4.1
2.9
3.4
5.5
9.4
8.2
11.2
6.9
9.8
6.2
4.5
3.1
7
8.4
4.8
5.6
8.4
2
5.9
5.3
8.1
4.7
12
5.5
6.6
8.1
14.1
2.4
5.8
6.3
6.1
4.3
5.1
4
7.9
11.6
7.5
7.3
9.7
6.4
10.5
5-minute daily max:1-
hour daily maximum
1.4
1.7
1.2
1.6
1.6
2.2
2
3.6
1.5
2.2
1.8
1.5
1.3
1.8
1.9
1.6
1.9
1.9
1.4
2
2
1.6
2.2
2.3
1.7
1.7
2.2
2.5
1.3
1.6
2.1
1.8
1.6
1.6
1.4
4.2
4.6
2.3
1.9
2.5
2.4
2
99 99th percentile 5-minute daily maximum, 1-hour daily maximum, and 24-hour average values were identified for
each year a given monitor was in operation from 2004-2006. If a monitor was in operation for multiple years over
that span, 99th percentile values were identified for each year, averaged, and then the appropriate ratio was
determined.
July 2009
374
-------
Staff further evaluated the potential of the 1-hour daily maximum standards analyzed in
this REA to provide protection against 24-hour average 862 exposures. The 99th percentile 24-
hour average SO2 concentrations in cities where key U.S. ED visit and hospitalization studies
(for all respiratory causes and asthma) were conducted ranged from 16 ppb to 115 ppb
(Thompson and Stewart, 2009). Moreover, effect estimates that remained statistically significant
in multipollutant models with PM were found in cities with 99th percentile 24-hour average SC>2
concentrations ranging from approximately 36 ppb to 64 ppb. Table 10-2 uses 2004 air quality
data and suggests that a 99th percentile 1-hour daily maximum standard set at a level of 50- 100
ppb would limit 99th percentile 24-hour average 862 concentrations observed in epidemiologic
studies where statistically significant results were observed in multi-pollutant models with PM.
That is, given a 50 ppb 99th percentile 1-hour daily maximum standard, none of the 39 counties
analyzed would be expected to have 24-hour average SC>2 concentrations > 36 ppb; and, given a
100 ppb 99th percentile 1-hour daily maximum standard, only 6 of the 39 counties (Linn, Union,
Bronx, Fairfax, Hudson, and Wayne) included in this analysis would be estimated to have 99th
percentile 24-hour average SC>2 concentrations > 36 ppb. This analysis was also done for the
years 2005 and 2006 and similar results were found (Appendix D).
July 2009 375
-------
nth
Table 10-2. 99 percentile 24-hour average SO2 concentrations for 2004 given just meeting the
alternative 1-hour daily maximum 99th and 98th percentile standards analyzed in the air quality
assessment (note: concentrations in ppb)
100
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
99th percentile
50
6
12
10
12
7
8
9
12
10
21
17
17
12
9
17
19
18
23
13
14
17
10
12
16
12
10
11
11
15
15
17
8
9
23
12
15
10
30
14
100
12
23
20
24
13
15
18
24
19
42
34
33
24
18
33
38
36
47
27
27
34
19
24
32
23
20
23
22
31
31
34
16
17
46
24
29
20
59
27
150
18
35
30
36
20
23
27
36
29
64
51
50
36
27
50
57
54
70
40
41
51
29
36
47
35
30
34
33
46
46
51
24
26
69
37
44
30
89
41
200
25
47
40
48
27
31
36
48
39
85
68
66
48
36
66
76
72
93
54
54
67
39
48
63
47
40
45
44
62
61
68
32
35
92
49
58
40
119
54
250
31
59
50
60
33
39
45
60
48
106
85
83
60
45
83
95
90
117
67
68
84
48
61
79
59
51
56
56
77
77
85
39
44
116
61
73
50
149
68
98th percentile
100
16
28
28
28
19
20
20
31
24
49
38
37
31
25
39
48
44
54
32
30
40
23
27
36
30
25
36
28
36
35
41
23
21
52
31
35
25
67
51
200
32
56
55
56
38
41
41
62
48
98
76
74
62
51
79
96
89
107
65
61
80
47
55
72
60
49
72
56
71
71
81
46
41
103
62
69
51
133
101
100 99th or 98th percentile 1-hour daily maximum concentrations were determined for each monitor in a given county
for the years completed data were available from 2004-2006. These concentrations were averaged, and the monitor
with the highest average in a given county was determined. Based on this highest average, all monitors in a given
county were adjusted to just meet the potential alternative standards defined above, and for each of the years, the
99th percentile 24-hour average SO2 concentration was identified. Iron County did not meet completeness criteria
for any of these years and is therefore not part of this analysis. Results for the years 2005 and 2006 are presented in
Appendix D.
July 2009
376
-------
As an additional matter, we note that a 99th percentile 1-hour daily maximum standard at
a level of 50-150 ppb could have the effect of maintaining SO2 concentrations below the level of
the current 24-hour and annual standards. That is, under these alternative standard scenarios
(using 2004 air quality data), there would be no counties in this analysis with a 2nd highest 24-
hour average greater than 144 ppb (Table 10-3). Similarly, under these alternative standard
scenarios (using 2004 air quality data), there would be no counties in this analysis with an annual
average 862 concentration in excess of the current annual standard (30.4 ppb; Table 10-4).
These analyses were also done with air quality from the years 2005 and 2006 and similar results
were found (Appendix D).
July 2009 377
-------
Table 10-3. 2nd highest 24-hour average SO2 concentrations (i.e., the current 24-hour standard) for
2004 given just meeting the alternative 1-hour daily maximum 99th and 98th percentile standards
analyzed in the air quality assessment (note: concentrations in ppb).101
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
99th percentile
50
7
12
11
14
10
8
11
15
10
28
17
19
17
11
18
21
19
25
21
15
19
13
17
19
13
10
15
13
16
17
19
10
13
26
18
17
12
33
17
100
14
38
23
28
19
17
21
29
20
57
38
38
34
22
37
43
38
51
42
31
38
27
35
38
28
21
30
27
31
34
38
21
25
52
36
35
24
67
34
150
21
57
34
42
29
25
32
44
30
85
57
56
51
33
55
64
57
76
63
46
58
40
52
57
42
31
45
40
50
50
57
31
38
78
54
52
35
100
51
200
27
76
45
55
39
34
43
58
40
113
75
75
67
45
74
86
77
102
83
61
77
54
70
76
56
42
60
54
67
67
76
42
50
104
72
69
47
134
68
250
34
95
57
69
48
42
53
73
50
142
94
94
84
56
92
107
96
127
104
77
96
67
87
95
70
52
75
67
84
84
95
52
63
130
90
86
59
167
85
98th percentile
100
18
45
31
32
28
22
24
38
25
65
43
42
44
31
44
54
47
59
50
35
47
32
39
43
32
25
48
34
36
39
45
30
29
58
46
41
30
75
63
200
36
91
63
65
55
44
48
76
50
130
86
84
87
63
88
109
95
117
100
69
91
65
79
87
71
51
96
68
77
78
90
60
59
117
91
82
60
150
126
101 99th or 98th percentile 1-hour daily maximum concentrations were determined for each monitor in a given county
for the years completed data were available from 2004-2006. These concentrations were averaged, and the monitor
with the highest average in a given county was determined. Based on this highest average, all monitors in a given
county were adjusted to just meet the potential alternative standards defined above, and for each of the years, the 2nd
highest 24-hour maximum concentration was identified. Iron County did not meet completeness criteria for any of
these years and is therefore not part of this analysis. Results for years 2005 and 2006 are presented in Appendix D.
July 2009
378
-------
Table 10-4. Annual average SO2 concentrations for 2004 given just meeting the alternative 99
,th
r»th
and 98 percentile 1-hour daily maximum standards analyzed in the air quality assessment (note:
concentrations in ppb).
102
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
Ml
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
VI
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
St Croix
99th percentile
50
1.7
2.0
1.6
1.8
0.8
1.8
1.5
2.0
1.5
1.8
2.5
2.5
2.0
1.5
2.2
6.4
6.4
7.6
2.6
3.1
3.9
2.3
2.6
3.9
2.9
2.5
4.6
2.3
4.3
3.0
3.5
2.1
1.3
7.7
4.8
4.0
2.2
6.1
1.2
100
3.4
4.0
3.2
3.6
1.6
3.6
2.9
4.1
3.1
3.5
5.0
5.1
4.1
2.9
4.4
12.8
12.7
15.1
5.3
6.1
7.7
4.7
5.1
7.8
5.8
5.1
9.1
4.5
8.7
6.1
7.0
4.2
2.6
15.5
9.6
8.0
4.3
12.2
2.4
150
5.1
6.0
4.7
5.4
2.3
5.3
4.4
6.1
4.6
5.3
7.5
7.6
6.1
4.4
6.5
19.3
19.1
22.7
7.9
9.2
11.6
7.0
7.7
11.7
8.7
7.6
13.7
6.7
13.0
9.1
10.4
6.3
3.9
23.2
14.3
12.0
6.5
18.3
3.7
200
6.8
7.9
6.3
7.2
3.1
7.1
5.9
8.2
6.1
7.0
10.0
10.2
8.2
5.8
8.7
25.7
25.4
30.2
10.5
12.2
15.5
9.3
10.2
15.5
11.6
10.1
18.3
9.0
17.4
12.1
13.9
8.4
5.3
30.9
19.1
16.1
8.7
24.4
4.9
250
8.5
9.9
7.9
9.0
3.9
8.9
7.3
10.2
7.7
8.8
12.5
12.7
10.2
7.3
10.9
32.1
31.8
37.8
13.2
15.3
19.3
11.6
12.8
19.4
14.5
12.7
22.8
11.2
21.7
15.2
17.4
10.4
6.6
38.6
23.9
20.1
10.9
30.6
6.1
98th percentile
100
4.5
4.7
4.4
4.2
2.2
4.7
3.3
5.3
3.8
4.0
5.6
5.7
5.3
4.1
5.2
16.2
15.7
17.4
6.3
6.9
9.2
5.6
5.8
8.9
7.4
6.1
14.6
5.7
10.0
7.0
8.2
6.0
3.1
17.3
12.1
9.5
5.5
13.7
4.5
200
9.0
9.5
8.7
8.5
4.4
9.4
6.7
10.7
7.6
8.1
11.2
11.3
10.6
8.3
10.4
32.5
31.4
34.8
12.7
13.8
18.4
11.2
11.5
17.7
14.8
12.3
29.1
11.3
20.0
14.0
16.5
12.0
6.2
34.6
24.2
19.1
11.1
27.4
9.1
102 99th or 98th percentile 1-hour daily maximum concentrations were determined for each monitor in a given county
for the years completed data were available from 2004-2006. These concentrations were averaged, and the monitor
with the highest average in a given county was determined. Based on this highest average, all monitors in a given
county were adjusted to just meet the potential alternative standards defined above, and for each of the years, the
annual concentration was calculated. Iron County did not meet completeness criteria for any of these years and is
therefore not part of this analysis. Results for the years 2005 and 2006 are presented in Appendix D
July 2009
379
-------
10.5.2.3 Conclusions regarding averaging time
The air quality analyses presented above strongly support that it is likely an alternative
99th percentile (see form discussion in 10.5.3) 1-hour daily maximum standard set at an
appropriate level (see level discussion in 10.5.4) can substantially reduce: (1) 5-10 minute peaks
of SC>2 shown in human exposure studies to result in respiratory symptoms and/or decrements in
lung function in exercising asthmatics, (2) 99th percentile 1-hour daily maximum air quality
concentrations in cities observing positive effect estimates in epidemiologic studies of hospital
admissions and ED visits for all respiratory causes and asthma, and (3) 99th percentile 24-hour
average air quality concentrations found in U.S. cities where ED visit and hospitalization studies
(for all respiratory causes and asthma) observed statistically significant associations in multi-
pollutant models with PM (i.e., 99th percentile 24-hour average 862 concentration > 36 ppb).
Thus, staff concludes that a 1-hour daily maximum standard, with an appropriate form and level,
can provide adequate protection against the range of health outcomes associated with averaging
times from 5-minutes to 24-hours. As an additional matter, we note that this conclusion is in
agreement with CASAC comments offered on the second draft SC>2 REA. The CASAC letter to
the Administrator states: "CASAC is in agreement with having a short-term standard and finds
that the REA supports a one-hour standard as protective of public health (Samet 2009)."
We note that based solely on the controlled human exposure evidence, staff also
considered a 5-minute averaging time. However, staff does not favor such an approach. As in
pastNAAQS reviews, we have considered the stability of the design of pollution control
programs in considering the elements of a NAAQS, since more stable programs are more
effective, and hence result in enhanced public safety. In this review, staff has concerns about the
stability of a 5-minute averaging time standard. Specific concerns relate to the number of
monitors needed and the placement of such monitors given the temporal and spatial
heterogeneity of 5-minute SC>2 concentrations. Moreover, staff is concerned that compared to
longer averaging times (e.g., 1-hour, 24-hour), year-to-year variation in 5-minute 862
concentrations is likely to be substantially more temporally and spatially diverse. Consequently,
staff judges that a 5-minute averaging time would not provide a stable regulatory target and
therefore, is not the preferred approach to provide adequate public health protection. However,
as noted above, staff's view is that a 1-hour averaging time, given an appropriate form (see
July 2009 380
-------
10.5.3) and level (see 10.5.4), can adequately limit 5-minute 862 exposures and provide a more
stable regulatory target than setting a 5-minute standard.
10.5.3 Form
When evaluating alternative forms in conjunction with specific levels, staff considers the
adequacy of the public health protection provided by the combination of level and form to be the
foremost consideration. In addition, we recognize that it is important that the standard have a
form that is reasonably stable. As just explained in the context of a five-minute averaging time,
a standard set with a high degree of instability could have the effect of reducing public health
protection because shifting in and out of attainment could disrupt an area's ongoing
implementation plans and associated control programs.
10.5.3.1 Evidence-based considerations
As previously mentioned, staff recognizes that the adequacy of the public health
protection provided by a 1-hour daily maximum potential alternative standard will be dependent
on the combination of form and level. It is therefore important that the particular form selected
for a 1-hour daily maximum potential alternative standard reflect the nature of the health risks
posed by increasing SC>2 concentrations. That is, the form of the standard should reflect results
from human exposure studies demonstrating that the percentage of asthmatics affected, and the
severity of the respiratory response (i.e. decrements in lung function, respiratory symptoms)
increases as 862 concentrations increase (see section 4.2.2). Taking this into consideration, staff
finds that a concentration-based form is more appropriate than an exceedance-based form. This
is because a concentration-based form averaged over three years (see below) would give
proportionally greater weight to 1-hour daily maximum SO2 concentrations that are well above
the level of the standard, than to 1-hour daily maximum SC>2 concentrations that are just above
the level of the standard. In contrast, an expected exceedance form would give the same weight
to 1-hour daily maximum SC>2 concentrations that are just above the level of the standard, as to
1-hour daily maximum SC>2 concentrations that are well above the level of the standard.
Therefore, a concentration-based form better reflects the continuum of health risks posed by
increasing 862 concentrations (i.e. the percentage of asthmatics affected and the severity of the
response increases with increasing 862 concentrations).
July 2009 381
-------
10.5.3.2 Risk-based considerations
In considering specific concentration-based forms, we recognize the importance of: 1)
minimizing the number of days per year that an area could exceed the level of the standard and
still attain the standard; 2) limiting the prevalence of 5-minute peaks of SO2; and 3) providing a
stable regulatory target to prevent areas from frequently shifting in and out of attainment. Given
this, we have focused on 98th and 99th percentile forms averaged over 3 years. We first note that
in most locations analyzed, the 99th percentile form of a 1-hour daily maximum standard would
correspond to the 4th highest daily maximum concentration in a year, while a 98th percentile form
would correspond approximately to the 7th to 8th highest daily maximum concentration in a year
(Table 10-5; see Thompson, 2009). In addition, results from the air quality analysis suggest that
at a given 862 standard level, a 99th percentile form is appreciably more effective at limiting 5-
minute peak SO2 concentrations than a 98th percentile form (Figures 7-27 and 7-28103).
Compared to the same standard with a 99th percentile form, a 98th percentile 1-hour daily
maximum standard set at 200 ppb allows for on average, an estimated 68 and 86% more days per
year when 5-minute SC>2 concentrations are greater than 200 and 400 ppb respectively (Figure 7-
27). Similarly, compared to the same standard with a 99th percentile form, a 98th percentile 1-
hour daily maximum standard at 100 ppb allows for on average, an estimated 90 and 74% more
days per year when 862 concentrations are greater than 200 and 400 ppb respectively104 (Figure
7-28). We also note that in the 40 counties selected for detailed air quality analysis, the
estimated number of benchmark exceedances using a 98th percentile 1-hour daily maximum
standard level of 200 ppb was similar to the corresponding 99th percentile standard at 250 ppb
(Tables 7-11 through 7-14). Similarly, the estimated number of benchmark exceedances
considering a 98th percentile standard at 100 ppb fell within the range of benchmark exceedances
estimated for 99th percentile standards at 100 and 150 ppb (Tables 7-11 through 7-14).
103 In these figures, the two air quality scenarios were compared on a monitor-to-monitor basis (see section 7.3)
104 Compared to a 200 ppb standard, a standard at 100 ppb results in far fewer site-years experiencing benchmark
exceedances (see Figures 7-27 and 7-28).
July 2009 382
-------
Table 10-5. SO2 concentrations (ppb) corresponding to the 2nd-9th daily maximum and 98th/99th
percentile forms for alternative 1-hour daily maximum standards (2004-2006).105
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
St Croix
State
AZ
DE
FL
IL
IL
IN
IN
IN
IA
IA
Ml
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
VI
SO2 Daily Maximums
2nd
36
17
13
16
21
21
15
15
11
15
14
10
50
16
6
7
8
11
17
9
19
17
11
13
30
19
26
12
21
12
24
15
5
21
18
23
16
3ra
33
15
12
15
19
19
12
13
10
15
13
9
43
16
6
7
7
11
15
9
19
16
9
12
25
17
24
11
20
10
22
14
5
18
17
22
13
4th
28
15
12
14
17
17
11
13
10
14
13
8
41
15
6
6
7
10
13
8
18
15
8
11
23
15
23
10
19
9
21
13
4
16
16
18
9
5tn
26
13
11
14
17
16
11
12
9
13
12
7
34
14
6
5
7
10
12
7
17
14
8
10
22
14
22
10
19
8
19
12
4
14
15
17
8
6th
25
13
11
13
15
14
10
11
10
13
12
7
31
14
5
5
6
9
12
7
16
14
7
10
20
12
19
9
18
8
16
12
4
14
15
16
5
7th
23
12
10
13
13
14
9
11
9
12
12
6
29
13
5
5
6
9
12
7
15
14
7
9
19
10
18
9
17
8
15
11
4
13
14
15
5
8th
22
12
9
12
13
13
9
10
10
12
11
6
28
13
5
5
6
8
11
7
15
13
7
9
19
9
18
9
17
7
14
11
4
12
13
14
5
9th
21
12
8
12
12
12
8
10
9
12
11
6
27
13
5
4
6
8
11
7
14
13
7
9
18
9
18
9
16
7
14
10
4
12
13
14
4
Percentiles
99th
28
15
12
14
15
17
11
13
10
14
13
8
35
15
6
6
6
10
13
8
18
15
8
11
13
15
23
10
19
9
21
13
4
16
16
17
5
98th
22
12
9
12
13
13
9
10
8
12
12
6
25
13
5
5
6
8
11
7
15
13
7
9
11
9
18
9
17
7
14
11
4
12
13
14
4
As an additional matter, staff compared trends in 98th and 99th percentile design values, as
well as design values based on the 4th highest daily maximum from 54 sites located in the 40
counties selected for detailed analysis (see Thompson, 2009). These results suggest that at the
105 Table 10-5 displays the 2nd through 9th highest, and the 99th and 98th percentiles of the daily maximums for
each of the counties. For the alternative daily metrics (the nth maximum and percentiles), the statistics were
computed for each year and then averaged over 2004-2006 (see Thompson 2009).
July 2009
383
-------
vast majority of sites, there would have been similar changes in 98th and 99th percentile design
values over the last ten years (i.e. based evaluating overlapping three year intervals over the last
ten years; see Thompson, 2009). These results also demonstrate that design values based on the
4th highest daily maximum are virtually indistinguishable from design values based on the 99th
percentile. For illustrative purposes, design value trends for four of these sites are presented in
Figure 10-1. As part of this analysis, all of the design values over this ten year period for all 54
sites were aggregated and the standard deviation calculated (see Thompson, 2009). Results
demonstrate similar standard deviations - i.e. similar stability — based on aggregated 98th or
aggregated 99th percentile design values over the ten year period (Figure 10-2; see Thompson
2009).
July 2009 384
-------
SlTE=l20570109
Q
^
ca
=
OX 0 IP
Figure 10-1. Design value trends from 4 of the 54 sites analyzed in Thompson 2009.
106
106 There were 8 possible 3-year design values from 1997 to 2007 (e.g. 1997-1999, 1998- 2000, etc.). Thus, the
design values presented in Figure 10-1 represent the 3-year average of the annual 98th percentile or 99th percentile 1-
hour daily maximum, or the 3-year average of 4th highest of the 1-hour daily maximum. (Thompson 2009).
July 2009
385
-------
o
'-4-»
co
CD
Q
ro
T3
co
0.06-r
0.05-
0.04-
0.03-
0.02-
0.01 -
0.00 -.
I
p98
p99
4th Max
Figure 10-2. Boxplots of the distributions of standard deviations for alternative air quality
standard forms.
10.5.3.3 Conclusions regarding form
Staff concludes that a concentration-based form provides the best protection against the
health risks posed by increasing SO2 concentrations (see 10.5.3.1). We also find that at a given
standard level, a 99th percentile or 4th highest daily maximum form provides appreciably more
public health protection against 5-minute peaks than a 98th percentile or 7th - 8th highest daily
maximum form (see 10.5.3.2). In addition, over the last 10 years and for the vast majority of the
sites examined, there appears to be little difference in 98th and 99th percentile design value
stability (see 10.5.3.2). Thus, staff concludes that consideration be given primarily to a 1-hour
daily maximum standard with a 99th percentile or 4th highest daily maximum form.
July 2009
386
-------
10.5.4 Level
In sections 10.3.3.3 and 10.4.4 staff concluded that the health evidence presented above
in Chapter 4 and the air quality, exposure, and risk information presented in Chapters 7-9 clearly
call into question the adequacy of the current SO2 standards to protect public health with an
adequate margin of safety from the respiratory effects of SCh. In considering potential
alternative standards that would provide increased public health protection against these
respiratory effects, staff concluded in section 10.5.1 that the most appropriate indicator remains
SC>2. In section 10.5.2, staff concluded that an alternative standard with a 1-hour averaging time,
set at an appropriate level, can provide adequate protection against the range of respiratory
effects observed in both controlled human exposure studies of 5-10 minutes, as well as
epidemiologic studies using longer averaging times. In addition, section 10.5.3 concluded that a
99th percentile or 4th highest daily maximum form averaged over three years was most
appropriate for potential standards using a 1-hour averaging time. Here, we consider 99th
percentile 1-hour daily maximum alternative standard levels that would provide greater public
health protection against SO2-related adverse respiratory effects than that afforded by the current
standards. As an initial consideration, we note that Table 10-6 demonstrates that although all
counties in the U.S. meet the current 24-hour and annual standards, all of the potential alternative
1-hour daily maximum standard levels (50-250 ppb) analyzed in the air quality, exposure, and
risk analyses would be estimated to result in counties in the U.S. with air quality above the level
of the given alternative standard. Thus, to varying extents, meeting any of the potential
alternative 1-hour daily maximum standards analyzed in this document would represent
reductions in ambient SC>2 levels based on air quality from 2004-2006, as well as reductions from
SC>2 concentrations that would be allowed under the current standards. All of these potential
standards would consequently result in some increased public health protection.
July 2009 387
-------
Table 10-6. Percent of counties that may be above the level of alternative standards (based on years 2004-2006)
Alternative Standards and
Levels (ppb)
Number of counties with
monitors
Percent of counties, total and by region not likely to meet a given standard
Total Counties
(population in
millions)
211 (96.5)
3 year 99th percentile daily 1-hour max:
250
200
150
100
50
1 (0.4)
3 (0.8)
10(2.4)
22(13.5)
54 (43.5)
3 year 98th percentile daily 1-hour max:
200
1 (0.4)
Northeast
52
0
0
2
8
38
0
Southeast
40
0
3
5
13
55
0
Industrial
Midwest
75
1
4
20
47
81
1
Upper
Midwest
19
0
0
5
5
37
0
Southwest
7
14
14
14
14
14
14
Northwest
9
0
0
0
0
22
0
Southern
CA
6
0
0
0
0
0
0
Outside
Regions
3
33
33
33
33
33
33
July 2009
388
-------
10.5.4.1 Evidence, Air Quality, Exposure and Risk-based considerations
Chapter 4 discussed the controlled human exposure and epidemiologic evidence with
respect to the judgments of causality presented in the ISA. In Chapter 5, our evaluation of the
health evidence informed the selection of potential alternative SO2 standards that would be
analyzed in the air quality, exposure, and risk analyses. In Chapter 6, potential health effect
benchmark values for use in the air quality and exposure analyses were derived from SC>2
concentrations found in controlled human exposure studies to result in decrements in lung
function and/or respiratory symptoms in exercising asthmatics. In this chapter, staff also used
the controlled human exposure and the epidemiologic evidence to inform judgments about the
adequacy of the current 862 standards, and to inform staff conclusions about the indicator,
averaging time, and form for potential alternative 862 standards.
Staff now considers the health evidence as it relates to evaluating 99th percentile 1-hour
daily maximum alternative standard levels.107 In doing so, we have considered the extent to
which a variety of alternative standard levels would limit the magnitude and frequency of 1-hour
SC>2 concentrations to provide sufficient protection for at-risk populations against experiencing
various respiratory health effects including moderate or greater decrements in lung function,
respiratory symptoms, and respiratory-related ED visits and hospitalizations. We note that these
health endpoints are logically linked together in that the controlled human exposure evidence
demonstrating moderate or greater decrements in lung function and/or respiratory symptoms in
exercising asthmatics is recognized by the ISA as supporting the plausibility of associations
between ambient 862 and the respiratory morbidity endpoints (i.e., respiratory symptoms,
emergency department visits, and hospital admissions) reported in epidemiologic studies.
In assessing the extent to which potential alternative standard levels with a 1-hour
averaging time and a 99th percentile form limit the array of health outcomes reported in both
controlled human exposure and epidemiologic studies, we first note the air quality information
provided by authors of key U.S. ED visit and hospitalization epidemiologic studies. This
information was presented earlier in Figures 5-1 to 5-4 and is described in detail in Thompson
and Stewart (2009). This information characterizes 99th percentile 1-hour daily maximum SC>2
air quality levels in cities and time periods corresponding to key U.S. studies of ED visits and
107 We note that these considerations are also relevant for consideration of alternative standard levels in conjunction
with a 4th highest daily maximum form.
July 2009 389
-------
hospitalizations for all respiratory causes and asthma. This information provides the most direct
evidence for effects in cities with particular 99th percentile 1-hour 862 levels, and hence, is of
particular relevance here. This information suggests that the strongest epidemiologic evidence of
an association between ambient SC>2 and ED visits and hospitalizations is in cities where 99th
percentile 1-hour daily maximum SC>2 concentrations ranged from about 75 tolSO ppb. In this
range, there are numerous studies that reported positive associations between ambient SC>2 and
respiratory related ED visits and hospitalizations (although all results were not statistically
significant). In addition, this range of 862 levels importantly contains a cluster of epidemiologic
studies demonstrating statistically significant results in multi-pollutant models with PM. More
specifically, in epidemiologic studies conducted in the Bronx, NY (78 ppb; NYDOH 2006,) and
in NYC, NY (82 ppb; Ito et al., 2007), the SO2 effect estimate remained positive and statistically
significant in multi-pollutant models with PM2.5 (ISA, Table 5-5). Moreover, in an
epidemiologic study conducted in New Haven, CT (150 ppb; Schwartz et al., 1995), the SC>2
effect estimate remained positive and statistically significant in a multi-pollutant model with
PMio. Staff notes that while statistical significance in co-pollutant models is an important
consideration, it is not necessary for appropriate consideration of and reliance on such
epidemiologic evidence.108 However, the existence of these studies particularly supports
consideration of standards levels at and below the range observed in these studies. Given this
body of epidemiologic evidence, staff concludes that alternative standard levels at and below 75
ppb should be considered to provide protection against the effects observed in these studies.
With regard to the epidemiologic studies mentioned above, we also note that most of the
ED visit and hospitalization effect estimates reported in these studies are with respect to 24-hour
average SO2 concentrations. Thus, staff investigated whether a 99th percentile 1-hour daily
maximum standard at approximately 75 ppb would also limit the 99thpercentile 24-hour average
SO2 concentrations observed in the cluster of studies finding statistically significant results in
multipollutant models with PM. Considering these studies, we note that the lowest 99th
percentile 24-hour average SO2 concentration reported in a study location finding statistically
significant associations in a multipollutant model with PM was 36 ppb in Bronx, NY (NYDOH
108 For example, evidence of a pattern of results from a group of studies that find effect estimates similar in direction
and magnitude would warrant consideration of and reliance on such studies even if the studies did not all report
statistically significant associations in single- or multi-pollutant models
July 2009 390
-------
2006). A standard of approximately 75 ppb was not analyzed in the air quality analysis, but
given a 50 ppb 99th percentile 1-hour daily maximum standard, none of the counties analyzed in
our analysis would be expected to have 99thpercentile 24-hour average SC>2 concentrations > 36
ppb (Table 10-2). However, given a 100 ppb 99th percentile 1-hour daily maximum standard,
six of the counties included in the 40-county air quality analysis would be estimated to have 99th
percentile 24-hour average SO2 concentrations > 36 ppb109. Thus, although not directly
analyzed, a 1-hour standard set at 75 ppb would be expected to limit 24-hour average
concentrations from exceeding 36 ppb in most, if not all, these counties. This analysis further
indicates that a 99th percentile 1-hour daily maximum standard level should be considered at or
below 75 ppb to provide protection against the effects observed in this cluster of epidemiologic
studies.
Staff also considered findings from controlled human exposure studies when evaluating
potential alternative standard levels. In doing so, we again note that the ISA finds that the most
consistent evidence of decrements in lung function and/or respiratory symptoms is from
controlled human exposure studies exposing exercising asthmatics to SC>2 concentrations > 400
ppb (ISA, section 3.1.3.5). At 862 concentrations > 400 ppb, moderate or greater
bronchoconstriction occurs in 20-60% of exercising asthmatics, and compared to exposures at
200- 300 ppb, a larger percentage of subjects experience severe bronchoconstriction. Moreover,
at concentrations > 400 ppb, statistically significant moderate or greater bronchoconstriction is
frequently accompanied by respiratory symptoms (ISA, Table 5-1). Controlled human exposure
evidence has also demonstrated decrements in lung function in exercising asthmatics following
5-10 minute SC>2 exposures starting as low as 200-300 ppb in free-breathing chamber studies. At
concentrations ranging from 200 - 300 ppb, the lowest levels tested in free breathing chamber
studies, 5-30% percent of exercising asthmatics are likely to experience moderate or greater
bronchoconstriction. However, at these lower levels, moderate or greater bronchoconstriction
has not been shown to be statistically significant, nor is it frequently accompanied by respiratory
symptoms. On the other hand, for understandable ethical reasons, it must also be noted that the
subjects participating in these controlled human exposure studies do not necessarily represent the
most SC>2 sensitive individuals (e.g. severe asthmatics). Thus, it is reasonable to anticipate that
109 Given a 99th percentile 1-hour daily maximum standard at 100 ppb, 99th percentile 24-hour average SO2
concentrations are estimated to be greater than 36 ppb in Linn, Union, Bronx, Hudson, Fairfax, and Wayne counties
(Table 10-2)
July 2009 391
-------
individuals who are more 862 sensitive would have a greater response at 200-300 ppb 862,
and/or would respond to 862 concentrations even lower than 200 ppb. Similarly, there is no
evidence to suggest that 200 ppb represents a threshold below which no adverse respiratory
effects occur. In fact, very limited evidence from two mouthpiece exposure studies suggests that
exposure to 100 ppb SC>2 can result in small decrements in lung function110. Moreover, while not
directly comparable to free-breathing chamber studies, findings from these mouthpiece studies
may be particularly relevant to those asthmatics who breathe oronasally even at rest (EPA,
1994b). Taken together, staff concludes that the level of a 99th percentile 1-hour daily maximum
SC>2 standard should be set so as to substantially limit the number of estimated 5-minute peaks >
400 ppb, while also appreciably limiting 862 concentrations > 200 ppb.
In evaluating the extent to which alternative standard levels provide substantial protection
against 5-minute SC>2 concentrations > 400 ppb, we first note the results of our 40 county air
quality analysis. As described above, epidemiologic studies support consideration of levels of a
99th percentile 1-hour daily maximum standard at or below 75 ppb. Thus, it would be instructive
to determine if a standard set at approximately 75 ppb would also substantially limit 5-minute
SC>2 concentrations > 400 ppb. Results of the air quality analysis indicate that just meeting a 99th
percentile 1-hour daily maximum standard at 50 ppb would result in 0 days per year when
statistically estimated 5-minute daily maximum 862 concentrations are > 400 ppb, whereas a
standard at 100 ppb would result in at most 2 days per year when statistically estimated 5-minute
daily maximum SO2 concentrations are > 400 ppb (Table 7-14)111. Given the results associated
with 99th percentile 1-hour daily maximum standards at 50 and 100 ppb, it is reasonable to
conclude that a 99th percentile 1-hour daily maximum standard at 75 ppb would also
substantially limit ambient 5-minute SC>2 concentrations > 400 ppb.
In further evaluating the extent to which potential alternative standard levels limit 5-
minute 862 exposures >400 ppb, we consider the results of the St. Louis exposure analysis.112
110 As first noted in Chapter 6, studies utilizing a mouthpiece exposure system cannot be directly compared to
studies involving freely breathing subjects, as nasal absorption of SO2 is bypassed during oral breathing, thus
allowing a greater fraction of inhaled SO2 to reach the tracheobronchial airways. As a result, individuals exposed to
SO2 through a mouthpiece are likely to experience greater respiratory effects from a given SO2 exposure.
Nonetheless, these studies do provide very limited evidence for SO2- induced respiratory effects at 100 ppb.
111 Air quality estimates presented in this section represent the mean number of days per year when 5-minute daily
maximum SO2 concentrations exceed a particular benchmark level given 2001-2006 air quality adjusted to just meet
alternative 99th percentile 1-hour daily maximum standards at 50, 100, or 150 ppb (see Tables 7-11 to 7-14).
112 As described in section 10.3.3.2, staff is primarily considering the St. Louis exposure and risk results when
evaluating the adequacy of the current and potential alternative 99th percentile 1-hour daily maximum SO2 standards.
July 2009 392
-------
Results indicate air quality just meeting a 99th percentile 1-hour daily maximum standard at 50 or
100 ppb would result in an estimated < 1% of asthmatics at elevated ventilation rates
experiencing at least one 5-minute daily maximum SO2 exposure > 400 ppb (Figure 8-19).
Similarly, this analysis also indicates that air quality just meeting a 50 or 100 ppb standard would
result in an estimated < 1% of asthmatic children at elevated ventilation rates experiencing at
least one 5-minute daily maximum SCh exposure > 400 ppb. These results necessarily suggest
that a standard at approximately 75 ppb would also substantially limit exposures of all asthmatics
and asthmatic children to 862 concentrations > 400 ppb.
We next evaluated the extent to which 99th percentile 1-hour daily maximum standard
levels provide appreciable protection against 5-minute 862 concentrations > 200 ppb. Results of
the 40 county air quality analysis indicate that a standard level of 50 ppb would result in at most
2 days per year when statistically estimated 5-minute daily maximum SC>2 concentrations would
be > 200 ppb, whereas a standard level of 100 ppb would result in at most 13 days per year when
statistically estimated 5-minute daily maximum SC>2 concentrations would be > 200 ppb (Table
7-12). Thus, a standard set at 75 ppb would result in somewhere between 2 and 13 days per year
when statistically estimated 5-minute daily maximum 862 concentrations would be > 200 ppb.
Results from the St. Louis exposure analysis estimate that air quality just meeting a 50
ppb, or 100 ppb 1-hour daily maximum standard would result in a corresponding < 1% or 1.5%
of asthmatics at elevated ventilation rates experiencing at least one 5-minute daily maximum 862
exposure > 200 ppb (Figure 8-19). Moreover, just meeting a 50 ppb, or 100 ppb 99th percentile
1-hour daily maximum standard would be estimated to result in a corresponding <1% or 2.7% of
asthmatic children at elevated ventilation rates experiencing at least one 5-minute daily
maximum SC>2 exposure > 200 ppb (Figure 8-19). Thus, a standard set at 75 ppb would be
estimated to result in somewhere between <1 and 1.5% of asthmatics, or <1 and 2.7% of
asthmatic children, at elevated ventilation rates experiencing at least one 5-minute daily
maximum 862 exposure > 200 ppb.
As an additional consideration, we note the results of the St. Louis risk assessment
indicate that a 99th percentile 1-hour daily maximum standard at 75 ppb would likely provide
appreciable protection against moderate or greater lung function responses. More specifically,
given a 99th percentile 1-hour daily maximum standard at 50 ppb, the median percentage of
asthmatics at elevated ventilation rates estimated to experience at least one > 100% increase in
July 2009 393
-------
sRaw ranges from 0.3% to 0.7% (and 0.4% to 0.9% for asthmatic children)113. In addition, given
air quality just meeting a 100 ppb standard, the estimated median percentage of asthmatics at
elevated ventilation rates experiencing at least one > 100% increase in sRaw ranges from 1.3 to
1.9% (and 2.1 to 2.9% for asthmatic children) (Table 9-5). Thus, we can expect that a standard
at 75 ppb would limit risk estimates to somewhere between the risks associated with the 50 and
100 ppb, 99th percentile 1-hour daily maximum standards.
Being mindful that the most severe effects associated with SC>2 exposure are those
observed in epidemiologic studies (i.e. respiratory-related ED visits and hospitalizations), staff
concludes that consideration also should be given to a standard level of 50 ppb. A 99th percentile
1-hour daily maximum standard at 50 ppb would provide an increased margin safety against the
air quality levels observed in the cluster of epidemiologic studies observing statistically
significant positive associations between SC>2 and respiratory-related ED visits and
hospitalizations in studies with multipollutant models with PM (i.e. 99th percentile 1-hour daily
maximum SC>2 concentrations > 78 ppb). Moreover, as demonstrated in TablelO-2, a 99th
percentile 1-hour daily maximum standard set at 50 ppb would also be expected to limit 99th
percentile 24-hour average SO2 concentrations significantly. That is, given a 1-hour daily
maximum standard set at 50 ppb, Table 10-2 demonstrates that most counties included in the 40-
county air quality analysis would have 99thpercentile 24-hour average SC>2 concentrations
below 15 ppb, ranging from 6-30 ppb.
Recognizing that there are important uncertainties associated with the controlled human
exposure evidence, we note that a 99th percentile 1-hour daily maximum standard set at 50 ppb
could also be considered if emphasis is placed on the: 1) uncertainty that the participants in
controlled human exposure studies do not represent the most SC>2 sensitive individuals; and/or 2)
very limited evidence suggesting decrements in lung function down to 100 ppb when SC>2 is
administered via mouthpiece (see section 6.2). Under this scenario, we note that a standard set at
50 ppb would provide increased protection against 5-minute 862 concentrations > 100 ppb.
Results from the 40 county air quality analysis indicate that a 99th percentile 1-hour daily
maximum standard set at 50 ppb would be estimated to result in at most 13 days per year when
statistically estimated 5-minute daily maximum SC>2 concentrations are > 100 ppb (Table 7-11).
113 As first noted in section 10.3.3.2, results are presented for both the probit and 2-parameter logistic functional
forms. The full range of estimates can be found in Chapter 9, and in all instances the smaller estimate is a result of
using the probit function to estimate the exposure-response relationship.
July 2009 394
-------
In addition, the St. Louis exposure analysis estimates that a 50 ppb 99th percentile 1-hour daily
maximum standard would likely result in 1.5% of asthmatics, and 2.7% of asthmatic children at
elevated ventilation rates experiencing at least one SO2 concentration > 100 ppb per year (Figure
8-19).
In considering alternative standard levels > 100 ppb, we first note that as mentioned in
section 10.3.3, staff concluded that exposure to 5-10 minute SO2 concentrations at least as low as
200 ppb can result in adverse respiratory effects in some asthmatics. Thus, in order to limit 5-10
minute 862 concentrations from exceeding 200 ppb, the level of a 99th percentile 1-hour daily
maximum standard would have to be < 200 ppb. We note that this conclusion is in accord with
consensus CAS AC comments following their review of the second draft REA. The CAS AC
letter to the Administrator states: "the draft REA appropriately implies that levels greater than
150 ppb are not adequately supported."
This letter also stated that "an upper limit of 150 ppb posited in Chapter 10 could be
justified under some interpretations of weight of evidence, uncertainties, and policy choices
regarding margin of safety" (Samet 2009). A 99th percentile 1-hour daily maximum SC>2
standard set in this range would have to place considerable weight on the uncertainties in the
epidemiologic health evidence presented in the ISA. That is, the emphasis on the uncertainties
would have to lead to a judgment that effects reported in epidemiologic studies are due in large
part to co-occurring pollutants, rather than to 862. Under this scenario, results of the 40 county
air quality analysis indicate that just meeting a 99th percentile 1-hour daily maximum standard
set at a level of 150 ppb would result in at most 7 days per year when statistically estimated 5-
minute daily maximum SC>2 concentrations would be > 400 ppb (Table 7-14). In addition, the St.
Louis exposure analysis indicates that a 99th percentile 1-hour daily maximum standard at 150
ppb would be estimated to result in < 1% of asthmatics, or asthmatic children at elevated
ventilation rates experiencing at least one 862 exposure > 400 ppb (Figure 8-19). Taken
together, it can reasonable be concluded that a 99th percentile 1-hour daily maximum standard up
to 150 ppb could similarly limit 862 exposures > 400 ppb when compared to standards in the
range of 50-100 ppb114.
114 Given a 50 or 100 ppb standard, the 40 county air quality analysis estimated at most 0 to 2 days per year when
statistically estimated 5-minute daily maximum SO2 concentrations would be > 400 ppb. In addition, the St. Louis
exposure analysis indicated that < 1% of asthmatics, or asthmatic children at elevated ventilation rates would be
expected to experience at least one SO2 exposure > 400 ppb.
July 2009 395
-------
However, it is important to note that a 99th percentile 1-hour daily maximum standard up
to 150 ppb would provide considerably less protection against 5-minute 862 concentrations >
200 ppb than standards in the range of 50 -100 ppb. Results of the 40 county air quality analysis
indicate that a 99th percentile 1-hour daily maximum standard at 150 ppb would result in at most
24 days per year when statistically estimated 5-minute daily maximum SC>2 concentrations would
be > 200 ppb. Moreover, the St. Louis exposure analysis indicates that a 150 ppb standard
would be estimated to result in 6.4% of all asthmatics, and 11.6% of asthmatic children
experiencing an 862 exposure > 200 ppb (Figure 8-19). Finally, we consider the results of the
St. Louis risk assessment. This assessment indicates that given a 150 ppb standard, the estimated
median percentage of exposed asthmatics at elevated ventilation rates estimated to experience at
least one > 100% increase in sRaw per year ranges from 2.9% to 3.6% (and 4.6% to 5.4% for
asthmatic children). Several aspects of these assessment results raise questions as to the
sufficiency of the protection that would be provided by a standard set at this level, when
compared to similar standards at or below 75 ppb.
10.5.4.1 Conclusions regarding level
Staff concludes that the health evidence and the air quality, exposure, and risk
information presented above most strongly support consideration of 99th percentile 1-hour daily
maximum standards in the range of 50- 75 ppb. However, if significant weight is placed on the
uncertainties in the epidemiologic and controlled human exposure evidence, levels up to 150 ppb
could be considered, recognizing the questions that would be raised by levels at the higher end of
this range. Staff recognizes that selecting an appropriate level that will protect public health with
an adequate margin of safety will be based on the relative weight given to different types of
information from the air quality, exposure, and risk assessment, as well as to the evidence, and
the uncertainties associated with the evidence and assessments.
10.5.4.2 Implications for the Current SO2 Standards
Finally, staff recognizes that the particular level selected for a new 1-hour daily
maximum standard will have implications for reaching decisions on whether to retain or revoke
the current 24-hour and annual standards. That is, with respect to SCVinduced respiratory
morbidity, the lower the level selected for a 99th percentile 1-hour daily maximum standard, the
less additional public health protection the current standards would be expected to provide. As
July 2009 396
-------
an initial consideration, we note that all 99th percentile 1-hour daily maximum 862 standard
levels being considered (i.e. 50-150 ppb) are expected to prevent ambient 862 concentrations
in the 40 counties analyzed in the air quality analysis from exceeding the levels of the current 24-
hour and annual standards (Tables 10-3 and 10-4). Moreover, Table 10-6 demonstrates that given
any of the potential alternative 1-hour daily maximum standards in this range, there would be
counties in the U.S. expected to have air quality above the level of that standard. However, this
does not rule out the possibility that the current standards could still offer some degree of
additional protection in some parts of the country not currently monitoring for 862.
Based on these considerations, staff finds it reasonable to conclude that if a new 99th
percentile 1-hour daily maximum standard is selected with a level from the upper end of the
range that staff has identified for consideration, then in addition to setting a 99th percentile 1-
hour daily maximum standard, consideration should also be given to retaining the existing 24-
hour and/or annual standards. However, if the selected level of a 99th percentile 1-hour daily
maximum standard is in the lower end of the range, it could reasonably be concluded that
consideration should be given to revoking the current 24-hour and/or annual NAAQS.
10.6 KEY OBSERVATIONS
The following observations reflect staffs views and conclusions:
• The scientific evidence and the risk and exposure information call into question the
adequacy of the current standards to protect public health with an adequate margin of
safety.
• In considering potential alternative standards, SO2 remains the most appropriate indicator
ambient SOX.
• A 1-hour daily maximum standard, set at an appropriate level, can provide adequate
protection against the range of health outcomes associated with averaging times from 5-
minutes to 24-hours.
• Consideration should be given primarily to establishing a new 1-hour daily maximum
standard with a 99th percentile or 4th highest daily maximum form.
• The health evidence and the air quality, exposure, and risk information presented above
most strongly support consideration of 99th percentile (or 4th highest) 1-hour daily
maximum standards in the range of 50- 75 ppb. Consideration should also be given to
standard levels above this range, up to 150 ppb, to the extent that significant weight is
placed on the uncertainties in the epidemiologic and controlled human exposure
evidence, recognizing the questions that would be raised by levels at the higher end of
this range.
July 2009 397
-------
REFERENCES
Abt Associates. (2007). Ozone Health Risk Assessment for Selected Urban Areas. EPA 452/R-
07-009. Prepared for Office of Air Quality Planning and Standards, U.S. EPA, Research
Triangle Park, NC. July 2007. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_pr_td.html.
AHS. (2003). Source and Accuracy Statement for the 2003 AHS-N Data Chart. Available at:
http://www.census.gov/hhes/www/housing/ahs/03dtchrt/source.html.
AHS. (2005). American Housing Survey for the St. Louis Metropolitan Area: 2004. Available
at: http://www.census.gov/prod/2005pubs/hl70-04-59.pdf
AHS. (2008). American Housing Survey Data: Metropolitan Areas. Available at:
http://www.census.gov/hhes/www/housing/ahs/metropolitandata.html.
American Lung Association. (2001). Urban air pollution and health inequities: a workshop
report. Environ Health Perpect. 109(S3):357-374.
American Thoracic Society. (1985). Guidelines as to what constitutes an adverse respiratory
health effect, with special reference to epidemiologic studies of air pollution. Am Rev Respir
Dis. 131:666-668.
American Thoracic Society. (2000). What constitutes an adverse health effect of air pollution?
Am JRespir Crit CareMed. 161:665-673.
Anderson HR, Ponce de Leon A, Bland JM, Bower JS, Emberlin J, Strachan DP. (1998). Air
pollution, pollens, and daily admissions for asthma in London 1987-92. Thorax. 53:842-
848.
Balmes JR, Fine JM, Sheppard D. (1987). Symptomatic bronchoconstriction after short-term
inhalation of sulfur dioxide. Am Rev Respir Dis. 136:1117-21.
Bethel RA, Epstein J, Sheppard D, Nadel JA, Boushey HA. (1983). Sulfur dioxide-induced
bronchoconstriction in freely breathing, exercising, asthmatic subjects. Am Rev Respir Dis.
128:987-90.
Bethel RA, Sheppard D, Geffroy B, Tarn E, Nadel JA, Boushey HA. (1985). Effect of 0.25 ppm
sulfur dioxide on airway resistance in freely breathing, heavily exercising, asthmatic subjects.
Am Rev Respir Dis. 131:659-61.
Brauer M, Koutrakis P, Spengler JD. (1989). Personal exposures to acidic aerosols and gases.
Environ Sci Tech. 23:1408-1412.
July 2009 398
-------
Bresnahan BW, Dickie M, Gerking S. (1997). Averting behavior and urban air pollution. Land
Economics. 73(3):340-357.
Burnett RT, Cakmak S, Brook JR, Krewski D. (1997). The role of particulate size and
chemistry in the association between summertime ambient air pollution and hospitalization
for cardiorespiratory diseases. Environ Health Perspect. 105:614-620.
Burnett RT, Smith-Doiron M, Stieb D, Cakmak S, Brook JR. (1999). Effects of particulate and
gaseous air pollution on cardiorespiratory hospitalizations. Arch Environ Health. 54:130-
139.
Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, Leech JA, Cakmak
S, Krewski D. (2001). Association between ozone and hospitalization for acute respiratory
diseases in children less than 2 years of age. Am JEpidemiol. 153: 444-452.
Burmaster DE. (1998). LogNormal distributions for skin area as a function of body weight. Risk
Analysis. 18(l):27-32. Available at: http://www.alceon.com/skinarea.pdf.
Burton CS, Stockenius TE, Stocking TS, Carr EL, Austin BS, Roberson RL (1987). Assessment
of Exposures of Exercising Asthmatics to Short-term SC>2 Levels as a Result of Emissions
from U.S. Fossil-fueled Power Plants. Systems Applications Inc., San Rafael, CA.
Publication No. 87/176, September 23, 1987.
Cakmak S, Dales RE, Judek S. (2006). Respiratory health effects of air pollution gases:
modification by education and income. Arch Environ Occup Health. 61: 5-10.
CASAC Transcripts (2008). Sulfur Dioxide Review Docket, July 30-31, 2008, pages 211-213.
Docket ID No. EPA-HQ-OAR-2007-0352. Available at www.regulations.gov.
CDC. (2007). National Center for Health Statistics. National Health Interview Survey (NHIS)
Public Use Data Release (2003). Available at:
http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.htm.
Chan AT and Chung MW. (2003). Indoor-outdoor air quality relationships in vehicle: effect of
driving environment and ventilation modes. Atmos Environ. 37:3795-3808.
Chapmann RS, Calafiore DC, Hasselblad V. (1985). Prevalence of persistent cough and phlegm
in young adults in relation to long-term ambient sulfur dioxide exposure. Am Rev Respir Dis.
132:261-267.
July 2009 399
-------
Chang MC, Sioutas C, Kim S, Gong H, Jr., Linn WS. (2000). Reduction of nitrate losses from
filter and impactor samplers by means of concentration enrichment. Atmos Environ. 34:85-
98.
Dab W, Medina S, Quenel P, Le Moullec Y, Le Tertre A, Thelot B, Monteil C, Lameloise P,
Pirard P, Momas I, Ferry R, Festy B. (1996). Short term respiratory health effects of
ambient air pollution: results of the APHEA project in Paris. JEpidemiol Community
Health. 50(Sl):42-46.
Delfino RJ, Gone H, Linn WS, Pellizzari ED, Hu Y. (2003). Asthma symptoms in hispanic
children and daily ambient exposures to toxic and criteria air pollutants. Environ Health
Perspect. 111:647-656.
Dey AN, Schiller JS, Tai DA. (2004). Summary Health Statistics for U.S. Children: National
Health Interview Survey, 2002. National Center for Health Statistics. Vital Health Statistics
10(221).
Dodge R. et al. (1985). A longitudinal study of children exposed to sulfur oxides. Am J
Epidemiology. 121:720-73 6.
EIA. (2001). Residential Energy Consumption Survey of 2001. Energy Information
Administration. Available at:
http://www.eia.doe.gov/emeu/recs/recs2001_hc/200 ltblhp.html.
EPA. (1982). Air Quality Criteria for Paniculate Matter and Sulfur Oxides. US EPA, Research
Triangle Park, NC: Office of Health and Environmental Assessment.
EPA. (1986a). Review of the National Ambient Air Quality Standards for Sulfur Oxides:
Updated Assessment of Scientific and Technical Information, Addendum to the 1982
OAQPS Staff Paper. Research Triangle Park, NC: Office of Air Quality Planning and
Standards, Strategies and Air Standards Division. EPA/450/5-86-13. Available from: NTIS,
Springfield, VA; PB87-200259/XAB.
EPA. (1986b). Second Addendum to Air Quality Criteria for Particulate Matter and Sulfur
Oxides (1982): Assessment of Newly Available Health Effects Information. US EPA,
Research Triangle Park, NC: Office of Health and Environmental Assessment.
EPA. (1986c). An Analysis of Short-Term Sulfur Dioxide Population Exposures in the Vicinity
of Utility Power Plants. US EPA Office of Air Quality Planning and Standards, Strategies
and Air Standards Division, Research Triangle Park, NC. September 12, 1986.
July 2009 400
-------
EPA. (1994a). Supplement to the Second Addendum (1986) to Air Quality Criteria for
Paniculate Matter and Sulfur Oxides (1982). Research Triangle Park, NC: Office of Health
and Environmental Assessment, Environmental Criteria and Assessment Office. EPA-
600/FP-93/002.
EPA. (1994b). Review of the National Ambient Air Quality Standards for Sulfur Oxides:
Assessment of Scientific and Technical Information, Supplement to the 1986 OAQPS Staff
Paper Addendum. Research Triangle Park, NC: Office of Air Quality Planning and
Standards. EPA-452/R-94/013.
EPA. (2002). Consolidated Human Activities Database (CHAD) Users Guide. Database and
documentation available at: http://www.epa.gov/chadnetl/.
EPA. (2003). AERMOD: Latest Features and Evaluation Results. U.S. Environmental
Protection Agency, Research Triangle Park, NC. EPA-454/R-03-003. Available at:
http ://www. epa.gov/scramOO 1 /7thconf/aermod/aermod_mep. pdf.
EPA. (2004a). AERMOD: Description of Model Formulation. Office of Air Quality Planning
and Standards. EPA-454/R-03-004. Available at:
http://www.epa.gov/scram001/7thconf/aermod/aermod_mfd.pdf.
EPA. (2004b). An Examination of EPA Risk Assessment Principles and Practices. Staff Paper
prepared by the US EPA Risk Assessment Task Force. EPA/100/B-04/001. Available at:
http://www.epa.gov/OSA/pdfs/ratf-fmal.pdf.
EPA. (2005). Review of the National Ambient Air Quality Standards for Particulate Matter:
Policy Assessment of Scientific and Technical Information - OAQPS Staff Paper. Research
Triangle Park, NC: Office of Air Quality Planning and Standards. EPA-452/R-05-005a.
Available at: http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf
EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final).
Office of Research and Development, National Center for Environmental Assessment,
Research Triangle Park, NC. EPA/600/R-05/004aF-cF.
EPA. (2007a). SO2 NAAQS Review Plan - Final. US EPA, Research Triangle Park, NC:
National Center for Environmental Assessment.
EPA. (2007b). Sulfur Dioxide Health Assessment Plan: Scope and Methods for Exposure and
Risk Assessment. US EPA Research Triangle Park, NC: Office of Air Quality Planning and
Standards, Research Triangle Park. Available at:
http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_pd.html.
EPA. (2007c). US EPA Air Quality System (AQS). Download Detailed AQS Data. Available
at: http://www.epa.gov/ttn/airs/airsaqs/detaildata/downloadaqsdata.htm.
July 2009 401
-------
EPA. (2007d). Ozone Population Exposure Analysis for Selected Urban Areas (July 2007).
Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-07-
010. Available at: http://epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_td.html.
EPA. (2007e). Review of the National Ambient Air Quality Standards for Ozone: Assessment
of Scientific and Technical Information, OAQPS Staff paper. Office of Air Quality Planning
and Standards, Research Triangle Park, NC. EPA-452/R-07-007a. Available at:
http://epa.gov/ttn/naaq s/standards/ozone/s_o3_cr_sp.html.
EPA. (2007f). 2002 National Emissions Inventory Data & Documentation. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. Available at:
http://www.epa.gov/ttn/chief/net/2002inventory.html.
EPA. (2007g). Clean Air Markets - Data and Maps. Office of Atmospheric Programs,
Washington, DC. Emissions Prepackaged Data Sets. Available at:
http://camddataandmaps.epa.gov/gdm/index.cfm?fuseaction=emissions. wizard.
EPA. (2007h). Sulfur Dioxide NAAQS Review - Analyses and Data Sets. Available at:
http://staging.epa.gov/ttn/analysis/so2.htm.
EPA. (20071). Review of the National Ambient Air Quality Standards for Lead: Policy
Assessment of Scientific and Technical Information - OAQPS Staff Paper. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-07-013.
Available at: http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf.
EPA. (2008a). Integrated Science Assessment (ISA) for Sulfur Oxides - Health Criteria (Final
Report). EPA/600/R-08/047F. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 98843.
EPA. (2008b). Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standard: First Draft. EPA-452/P-08-003. June 2008.
Available at: http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_rea.html.
EPA. (2008c). Sulfur Dioxide Health Assessment Plan: Scope and Methods for Exposure and
Risk Assessment. US EPA, Research Triangle Park, NC: Office of Air Quality Planning and
Standards, Research Triangle Park. Available at:
http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_pd.html.
EPA. (2008d). Risk and Exposure Assessment to Support the Review of the NO2 Primary
National Ambient Air Quality Standard. EPA-452/R-08-008a. Available at:
http://www.epa.gov/ttn/naaqs/standards/nox/data/20081121_NO2_REA_fmal.pdf
July 2009 402
-------
EPA. (2008e). AERSURFACE User's Guide. EPA-454/B-08-001. US EPA, Research Triangle
Park, NC. Available at:
http://www.epa.gov/scram001/7thconf/aermod/aersurface_userguide.pdf
EPA. (2009a). Total Risk Integrated Methodology (TRIM) - Air Pollutants Exposure Model
Documentation (TRIM.Expo / APEX, Version 4.3) Volume I: User's Guide. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. January 2009. Available at:
http ://www. epa.gov/ttn/fera/human_apex.html.
EPA. (2009b). Total Risk Integrated Methodology (TRIM) - Air Pollutants Exposure Model
Documentation (TRIM.Expo / APEX, Version 4.3) Volume II: Technical Support Document.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. January 2009.
Available at: http://www.epa.gov/ttn/fera/human_apex.html.
EPRI. (1988). A Study of Activity Patterns Among a Group of Los Angeles Asthmatics.
Research Project 940-5. Electric Power Research Institute. Prepared by Roth Associates.
November 1988.
EPRI. (1992). A Survey of Daily Asthmatic Activity Patterns in Cincinnati. TR-101396.
Research Project 940-05. Electric Power Research Institute. Prepared by Roth Associates.
November 1992.
ESRI. (2008). ESRI Demographic Update Methodology: 2008/2013 (page 8). Available at:
http://www.esri.com/library/whitepapers/pdfs/demographic-update-methodology-2008.pdf
Finkelstein MM, Jerrett M, DeLuca P, Finkelstein N, Verma DK, Chapman K, Sears MR.
(2003). Relation between income, air pollution and mortality: a cohort study. CMAJ. 169:
397-402.
Ford ES, Heath GW, Mannino DM, Redd SC. (2003). Leisure-time physical activity patterns
among US adults with asthma. Chest. 124:432-437.
Fung KY, Khan S, Krewski D, Chen Y. (2006). Association between air pollution and multiple
respiratory hospitalizations among the elderly in Vancouver, Canada. Inhalation Toxicol.
18:1005-1011.
Galan I, Tobias A, Banegas JR, Aranguez E. (2003). Short-term effects of air pollution on daily
asthma emergency room admissions. EurRespirJ. 22:802-808.
Gelman et al. (1995). Bayesian Data Analysis. Chapman and Hall, New York.
July 2009 403
-------
Geyh AS, Xue J, Ozkaynak H, Spengler JD. (2000). The Harvard Southern California chronic
ozone exposure study: assessing ozone exposure of grade-school-age children in two
Southern California communities. Environ Health Perspect. 108:265-270.
Gilks WR, Richardson S, Spiegelhalter DJ. (Eds). (1996). Markov Chain Monte Carlo in
Practice. Chapman and Hall, London, UK.
Gong H, Jr., Lachenbruch PA, Harber P, Linn WS. (1995). Comparative short-term health
responses to sulfur dioxide exposure and other common stresses in a panel of asthmatics.
Taxicol Ind Health. 11:467-487.
Gong H, Linn WS, Shamoo DA, Anderson KR, Nugent CA, Clark KW, Lin AE. (1996). Effect
of inhaled salmeterol on sulfur dioxide-induced bronchoconstriction in asthmatic subjects.
Chest. 110:1229-1235.
Gong H, Jr., Linn WS, Terrell SL, Anderson KR, Clark KW. (2001). Anti-inflammatory and
lung function effects of Montelukast in asthmatic volunteers exposed to sulfur dioxide.
Chest. 119:402-408.
Graham SE and McCurdy T. (2004). Developing meaningful cohorts for human exposure
model s. J Expos Anal Environ Epidemiol. 14(l):23-43.
Graham SE and McCurdy T. (2005). Revised Ventilation Rate (VE) Equations for Use in
Inhalation-Oriented Exposure Models. EPA/600/X-05/008. See Appendix A and D of
EPA/600/R-06/129F. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=202543
Grontoft T and Raychaudhuri MR. (2004). Compilation of tables of surface deposition
velocities for Os, NC>2 and SC>2 to a range of indoor surfaces. Atmos Environ. 38:533-544.
Hagen JA, Nafstad P, Skrondal A, Bjorkly S, Magnus P. (2000). Associations between outdoor
air pollutants and hospitalization for respiratory diseases. Epidemiology. 11:136-140.
Hajat S, Haines A, Goubet SA, Atkinson RW, Anderson HR. (1999). Association of air
pollution with daily GP consultations for asthma and other lower respiratory conditions in
London. Thorax. 54:597-605.
Henderson. (2006). Letter to EPA Administrator Stephen Johnson: Clean Air Scientific
Advisory Committee's (CAS AC) Peer Review of the Agency's 2nd Draft Ozone Staff Paper.
EPA-CASAC-07-001. October 24, 2006.
July 2009 404
-------
Henderson. (2008). Letter to EPA Administrator Stephen Johnson: Clean Air Scientific
Advisory Committee's (CAS AC) Peer Review of EPA's Risk and Exposure Assessment to
Support the Review of the SC>2 Primary National Ambient Air Quality Standards (First Draft,
July 2008). EPA-CASAC-08-019. August 22, 2008. Sulfur Dioxide Review Docket.
Docket ID No. EPA-HQ-OAR-2007-0352. Availableatwww.regulations.gov.
Hill AB. (1965). The environment and disease: association or causation? ProcRSocMed.
58:295-300.
Horstman D, Roger LJ, Kehrl H, Hazucha M. (1986). Airway sensitivity of asthmatics to sulfur
dioxide. ToxicolIndHealth. 2:289-98.
Hu PS and Reuscher TR. (2004). Summary of Travel Trends (2001). National Household
Travel Survey. US Department of Transportation, Federal Highway Administration.
Available at: http://nhts.ornl.gov/2001/pub/STT.pdf
Isaacs K and Smith L. (2005). New Values for Physiological Parameters for the Exposure
Model Input File Physiology.txt. Memorandum submitted to the U.S. Environmental
Protection Agency under EPA Contract EP-D-05-065. NERL WA 10. Alion Science and
Technology.
Ito K. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions in
the context of time-series health effects models. JExpos Sci Environ Epidemiol. 17:845-
S60.
Jaffe DH, Singer ME, Rimm AA. (2003). Air pollution and emergency department visits for
asthma among Ohio medicaid recipients, 1991-1996. Environ Res. 91:21-28.
Johns. (2009). Presentation and analysis of controlled human exposure data described in Table
3-1 of the 2008 Integrated Science Assessment (ISA) for Sulfur Oxides; April 29, 2009.
Available at: http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_rea.html
Johnson T, Mihlan G, LaPointe J, Fletcher K, Capel J, Rosenbaum A, Cohen J, Stiefer P.
(2000). Estimation of carbon monoxide exposures and associated carboxyhemoglobin levels
for residents of Denver and Los Angeles using pNEM/CO. Appendices. EPA contract 68-
D6-0064.
KDHE. (2006). Environmental Factors, Outdoor Air Quality, and Activity Level: Results from
2005 Kansas Behavioral Risk Factor Surveillance System. Kansas Department of Health and
Environment. Available at: http://www.kdheks.gov/brfss/PDF/cste_report_fmal.pdf
Kehrl HR, Roger LJ, Hazucha MJ, Horstman DH. (1987). Differing response of asthmatics to
sulfur dioxide exposure with continuous and intermittent exercise. Am Rev Respir Dis. 135:
350-355.
July 2009 405
-------
Kindzierski WB and Ranganathan HKS. (2006). Indoor and outdoor SC>2 in a community near
oil sand extraction and production facilities in Northern Alberta. Journal of Environmental
Engineering and Science. 5:S121-S129.
Koenig JQ, Marshall SG, Horike M, Shapiro GG, Furukawa CT, Bierman CW, Pierson WE.
(1987). The effects of albuterol on sulfur dioxide-induced bronchoconstriction in allergic
adolescents. JAllergy Clin Immunol. 79:54-58.
Koenig JQ, Covert DS, Hanley QS, van Belle G, Pierson WE. (1990). Prior exposure to ozone
potentiates subsequent response to sulfur dioxide in adolescent asthmatic subjects. Am Rev
RespirDis. 141:377-380.
Langstaff JE. (2007). OAQPS Staff Memorandum to Ozone NAAQS Review Docket (OAR-
2005-0172). Subject: Analysis of Uncertainty in Ozone Population Exposure Modeling.
[January 31, 2007]. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_td.html.
Langstaff J. (2009). Analysis of Uncertainty in the Statistical Model for Estimating 5-minute
Maximum Concentrations of SO2. Sulfur Dioxide Review Docket. Docket ID No. EPA-HQ-
OAR-2007-0352. Available atwww.regulations.gov.
Lin M, Chen Y, Burnett RT, Villeneuve PJ, Krewski D. (2003). Effect of short-term exposure
to gaseous pollution on asthma hospitalizations in children: a bi-directional case-crossover
analysis. JEpidemiol Community Health. 57:50-55.
Lin M, Chen Y, Villeneuve PJ, Burnett RT, Lemyre L, Hertzman C, McGrail KM, Krewski D.
(2004a). Gaseous air pollutants and asthma hopitalization of chiildren with low household
income in Vancouver, British Columbia, Canada. Am J Epidemiol. 159:294-303.
Lin S, Hwang S-A, Pantea C, Kielb C, Fitzgerald E. (2004). Childhood asthma hospitalizations
and ambient air sulfur dioxide concentrations in Bronx County, New York. Arch Environ
Health. 59:266-275.
Lin M, Stieb DM, Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory
infections in children younger than 15 years in Toronto: a case-crossover analysis.
Pediatrics. 116:e235-240.
Linn WS, Venet TG, Shamoo DA, Valencia LM, Anzar UT, Spier CE, Hackney JD. (1983).
Respiratory effects of sulfur dioxide in heavily exercising asthmatics. A dose-response study.
Am Rev RespirDis. 127:278-83.
July 2009 406
-------
Linn WS, Avol EL, Shamoo DA, Venet TG, Anderson KR, Whynot JD, Hackney JD. (1984).
Asthmatics' responses to 6-hr sulfur dioxide exposures on two successive days. Arch
Environ Health. 39:313-319.
Linn WS, Avol EL, Peng RC, Shamoo DA, Hackney JD. (1987). Replicated dose-response
study of sulfur dioxide effects in normal, atopic, and asthmatic volunteers. Am Rev Respir
Dis. 136:1127-1134.
Linn WS, Avol EL, Shamoo DA, Peng RC, Spier CE, Smith MN, Hackney JD. (1988). Effect
of metaproterenol sulfate on mild asthmatics' response to sulfur dioxide exposure and
exercise. Arch Environ Health. 43:399-406.
Linn WS, Shamoo DA, Peng RC, Clark KW, Avol EL, Hackney JD. (1990). Responses to
sulfur dioxide and exercise by medication-dependent asthmatics: effect of varying
medication levels. Arch Environ Health. 45:24-30.
Lunn JE, Knowelden J, Handy side AJ. (1967). Patterns of respiratory illness in Sheffield infant
schoolchildren. Br JPrev SocMed. 21:7-16
Luginaah IN, Fung KY, Gorey KM, Webster G, Wills C. (2005). Association of ambient air
pollution with respiratory hospitalization in a government designated "area of concern": The
case of Windsor, Ontario. Environ Health Perspect. 113:290-296.
Magnussen H, Jorres R, Wagner HM, von Nieding G. (1990). Relationship between the airway
response to inhaled sulfur dioxide, isocapnic hyperventilation, and histamine in asthmatic
subjects. Int Arch Occup Environ Health. 62:485-91.
Mansfield C, Van Houtven G, Johnson FR, Crawford-Brown D. (2005). Behavioral Reactions
to Ozone Alerts: What Do They Tell Us About Willingness-to-Pay for Children's Health?
Research Update - 6/17/05.
Meyers S. (1983). Section 107 Designation Policy Summary. Memo to the Directors of Air and
Waste Management Division, Regions II-IV, VI-VIII, X and Director of Air Management
Division Regions I, V, and IX. US EPA Office of Air Quality Planning and Standards. April
21, 1983.
McCurdy T, Glen G, Smith L, Lakkadi Y. (2000). The National Exposure Research
Laboratory's Consolidated Human Activity Database. J Expos Anal Environ Epidemiol.
10:566-578.
MO Department of Health. (2002). Health Risk Behaviors of Adult Missourians, 2002 Annual
Report, Behavioral Risk Factor Surveillance System. Division of Community and Public
Health, Bureau of Health Informatics, Available at:
http://www.dhss.mo.gov/BRFSS/2002AnnualReport.pdf
July 2009 407
-------
MO DOH. (2003). Health and Preventive Practices Status Report from Missouri County-level
Study - Community Data Profiles. Available at:
http://www.dhss.mo.gov/CommunityDataProfiles/.
Mortimer KM, Neas LM, Dockery DW, Redline S, Tager IB. (2002). The effect of air pollution
on inner-city children with asthma. Eur Respir J. 19:699-705.
Neas LM, Dockery DW, Koutrakis P, Tollerud DJ, Speizer FE. (1995). The association of
ambient air pollution with twice daily peak expiratory flow rate measurements in children.
AmJEpidemiol. 141:111-122.
NY DOH. (2006). A Study of Ambient Air Contaminants and Asthma in New York City.
ATSDR Final Report #NTIS PB2006-113523. Albany, NY; New York State Energy
Research and Development Authority; New York State Department of Health, for Atlanta,
GA; Agency for Toxic Substances and Disease Registry; U.S. Department of Health and
Human Services.
O'Byrne PM. (2007). How much is too much? The treatment of mild asthma. Eur Respir J.
30:403-406.
Peel JL, Tolbert PE, Klein M, Metzger KB, Flanders WD, Knox T, Mulholland JA, Ryan PB,
Frumkin H. (2005). Ambient air pollution and respiratory emergency department visits.
Epidemiology. 16:164-174.
Perry SG, Cimorelli AJ, Paine RJ, Erode RW, Weil JC, Venkatram A, Wilson RB, Lee RF,
Peters WD. (2005). AERMOD: a dispersion model for industrial source applications. Part
II: model performance against 17 field study databases. JApplMeteor. 44:694-708.
Persily A, Gorfain J, Brunner G. (2005). Ventilation design and performance in U.S. Office
buildings. ASHRAE Journal. April 2005, 30-35.
Rabe KF, Adachi M, Lai CKW, Soriano JB, Vermeire PA, Weiss KB, Weiss ST. (2004).
Worldwide severity and control of asthma in children and adults: the global asthma insights
and reality surveys. The Journal of Allergy and Clinical Immunology. 114:40-47.
Rizzo M. (2009). Investigation of How Distributions of Hourly Sulfur Dioxide Concentrations
Have Changed Over Time in Six Cities. Sulfur Dioxide Review Docket. Docket ID No.
EPA-HQ-OAR-2007-0352. Available at www.regulations.gov.
Robinson JP, Wiley JA, Piazza T, Garrett K, Cirksena K. (1989). Activity Patterns of California
Residents and their Implications for Potential Exposure to Pollution. California Air
Resources Board, Sacramento, CA. CARB-A6-177-33.
July 2009 408
-------
Roger LJ, Kehrl HR, Hazucha M, Horstman DH. (1985). Bronchoconstriction in asthmatics
exposed to sulfur dioxide during repeated exercise. J ApplPhysiol. 59:784-91.
Rosenbaum AS, Hudischewskyj AB, Roberson RL, Burton CS. (1992). Estimates of Future
Exposures of Exercising Asthmatics to Short-term Elevated SO2 Concentrations Resulting
from Emissions of U.S. Fossil-fueled Power Plants: Effects of the 1990 Amendments to the
Clean Air Act and a 5-Minute Average Ambient SO2 Standard. Publication No. SYSAPP-
92/016. April 23, 1992. Docket No. A-84-25, IV-K-37.
SAL (1996). Summary of 1988-1995 Ambient 5-Minute SO2 Concentration Data. Prepared for
US EPA Office of Air Quality Planning and Standards by Systems Application International.
Contract #68-D3-0101, May 1996.
Sallis JF, Bauman A, Pratt M. (1998). Environmental and policy interventions to promote
physical activity. American Journal of Preventive Medicine. 15(4):379-397.
Samet JM. (2009). Letter to EPA Administrator Lisa P. Jackson: Clean Air Scientific Advisory
Committee's (CASAC) Review of EPA's Risk and Exposure Assessment to Support the
Review of the SO2 Primary National Ambient Air Quality Standards: Second Draft. EPA-
CASAC-09-0007, May 18, 2008. Sulfur Dioxide Review Docket. Docket ID No. EPA-HQ-
OAR-2007-0352. Available atwww.regulations.gov.
Santuz P, Baraldi E, Filippone M, Zacchello F. (1997). Exercise performance in children with
asthma: Is it different from that of healthy controls? Eur Respir J. 10:1254-1260.
Sarnat SE, Coull BA, Schwartz J, Gold DR, Suh HH. (2006). Factors affecting the association
between ambient concentrations and personal exposures to particles and gases. Environ
Health Perspect. 114:649-654.
S AS Institute Inc. (2004). SAS OnlineDoc® 9.1.3. Cary, NC: S AS Institute Inc.
Schachter EN, Witek TJ, Jr, Beck GJ, Hosein HB, Colice G, Leaderer BP, Cain W. (1984).
Airway effects of low concentrations of sulfur dioxide: dose-response characteristics. Arch
Environ Health. 39:34-42.
Schildcrout JS, Sheppard L, Lumley T, Slaughter JC, Koenig JQ, Shapiro GG. (2006). Ambient
air pollution and asthma exacerbations in children: an eight-city analysis. Am JEpidemiol.
164:505-517.
Schofield WN. (1985). Predicting basal metabolic rate, new standards, and review of previous
work. HumNutrClinNutr. 39C(S1):5-41.
July 2009 409
-------
Schwartz J, Dockery DW, Neas LM, Wypij D, Ware JH, Spengler JD, Koutrakis P, Speizer FE,
Ferris BG, Jr. (1994). Acute effects of summer air pollution on respiratory symptom
reporting in children. Am JRespir Crit Care Med. 150:1234-1242.
Schwartz J. (1995). Short term fluctuations in air pollution and hospital admissions of the
elderly for respiratory disease. Thorax. 50:531-538.
Schwartz J. (1996). Air pollution and hospital admissions for respiratory disease.
Epidemiology. 7:20-28.
Sciences International (1995). Estimate of the Nationwide Exercising Asthmatic Exposure
Frequency to Short-term Peak Sulfur Dioxide Concentrations in the Vicinity of Non-Utility
Sources. Prepared for National Mining Association by Sciences International, Inc.,
Alexandria VA. April 1995. Docket No. A-84-25, VIII-D-71.
Sheppard D, Saisho A, Nadel JA, Boushey HA. (1981). Exercise increases sulfur dioxide-
induced bronchoconstriction in asthmatic subjects. Am Rev Respir Dis. 123:486-491.
Sheppard reanalysis (2003). Ambient Air Pollution and Nonelderly Asthma Hospital
Admissions in Seattle, Washington, 1987-1994. In: Revised Analyses of Time-Series
Studies of Air Pollution and Health. Special report. Boston, MA: Health Effects Institute;
pp. 227-230.
Sheppard L, Slaughter JC, Schildcrout J, Liu LJ, Lumley T. (2005). Exposure and measurement
contributions to estimates of acute air pollution effects. J Expos Anal Environ Epidemiol.
15:366-76.
Skuta and Wombold. (2008). The Secret Life of Polygons: Understanding the relative accuracy
of user-defined areas. Available at: http://www.esri.com/news/arcuser/1008/polygons.html.
Smith E. (1994). Subject Data Supplied by the Researchers for the Recent Controlled Human
Studies Analyzed in the Staff Paper Supplement and Accompanying Memorandum.
Memorandum to Docket No. A-84-25, Item IV-B-5
Spier CE, Little DE, Trim SC, Johnson TR, Linn WS, Hackney JD. (1992). Activity patterns in
elementary and high school students exposed to oxidant pollution. J Expos Anal Environ
Epidemiol. 2:277-293.
Spiegelhalter D, Thomas A, Best N, Gilks W. (1996). Bugs.5 Bayesian Inference Using Gibbs
Sampling. Manual, version ii. MRC Biostatistics Unit, Institute of Public Health.
Cambridge, U.K.
July 2009 410
-------
Stieb DM, Beveridge RC, Brook JR, Smith-Doiron M, Burnett RT, Dales RE, Beaulieu S, Judek
S, Mamedov A. (2000). Air pollution, aeroallergens and cardiorespiratory emergency
department visits in Saint John, Canada. J Expos Anal Environ Epidemiol. 10:461-477.
Stoeckenius TE, Garelick B, Austin BS, O'Connor K, Pehling JR. (1990). Estimates of
Nationwide Asthmatic Exposures to Short-Term Sulfur Dioxide Concentrations in the
Vicinity of Non-Utility Sources. Systems Applications Inc., San Rafael, CA. Publication
No. SYSAPP-90/129, December 6, 1990.
Stone M. (1974). Cross-validatory choice and assessment of statistical predictions. Journal of
the Royal Statistical Society. Ser.B, 36:111-147.
Suh H. (2008). Personal communication, August 27, 2008, regarding personal exposure data for
Atlanta.
Sunyer J, Spix C, Quenel P, Ponce-de-Leon A, Ponka A, Barumandzadeh T, Touloumi G,
Bacharova L, Wojtyniak B, Vonk J, Bisanti L, Schwartz J, Katsouyanni K. (1997). Urban
air pollution and emergency admissions for asthma in four European cities: The APHEA
Project.
Thompson R. (2000). Preliminary Analysis of 5-Minute Maximum Ambient SC>2
Concentrations. OAQPS. 12/21/00. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC. Available at:
http ://www. epa.gov/ttn/amtic/so2data.html.
Thompson R. (2009). Sulfur Dioxide Descriptive Statistics Tables. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
NC. Sulfur Dioxide Review Docket. Docket ID No. EPA-HQ-OAR-2007-0352. Available
at www.regulations.gov.
Thompson R and Stewart MJ. (2009). Air Quality Statistics for Cities Referenced in Key U.S.
and Canadian Hospital Admission and Emergency Department Visits for All Respiratory
Causes and Asthma. Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC. Sulfur Dioxide Review Docket. Docket ID
No. EPA-HQ-OAR-2007-0352. Available atwww.regulations.gov.
Thrall AD, Langstaff JE, Liu MK, Burton CS. (1982). On the Variability in Peak Sulfur
Dioxide Concentrations Contained in Longer-Term Averages: An Empirical Study of an
Isolated PowerPlant. Systems Applications Inc.
Tolbert PE, Klein M, Peel JL, Sarnat SE, Sarnat JA. (2007). Multipollutant modeling issues in a
study of ambient air quality and emergency department visits in Atlanta. J Expos Sci
Environ Epidemiol. 17: S29-S3 5
July 2009 411
-------
Trenga CA, Koenig JQ, Williams PV. (1999). Sulphur dioxide sensitivity and plasma
antioxidants in adult subjects with asthma. Occup EnvironMed. 56:544-547.
Tsai SS, Cheng MH, Chiu HF, Wu TN, Yang CY. (2006). Air pollution and hospital
admissions for asthma in a tropical city: Kaohsiung, Taiwan. Inhal Toxicol. 18:549-554.
UARG. (2009). Comments of the Utility Air Regulatory Group on the Risk and Exposure
Assessment to Support the Review of the 862 Primary National Ambient Air Quality
Standards: Second Draft; Hunton & Williams, LLP. Sulfur Dioxide Review Docket. Docket
ID No. EPA-HQ-OAR-2007-0352. Available at www.regulations.gov.
US Census Bureau. (2007). Employment Status: 2000- Supplemental Tables. Available at:
http://www.census.gov/population/www/cen2000/phc-t28.html
US DOT. (2007). Part 3-The Journey to Work files. Bureau of Transportation Statistics (BTS).
Available at: http://transtats.bts.gov/.
Van der Zee SC, Hoek G, Boezen HM, Schouten JP, van Wijnen JH, Brunekreef B. (1999).
Acute effects of urban air pollution on respiratory health of children with and without chronic
respiratory symptoms. Occup Environ Med. 56:802-813.
Van Gent, R, van der Ent K, van Essen-Zandvliet LEM, Rovers MM, Kimpen ILL, de Meer G,
Klijn PHC. (2007). No Difference in physical activity in (un)diagnosed asthma and healthy
controls. PediatricPulmonology. 42:1018-1023.
Van Veldhoven NHMJ, Vermeer A, Bogaard JM, Hessels MGP, Wijnroks L, Colland VT, van
Essen-Zandvliet EEM. (2001). Children with asthma and physical exercise: effects of an
exercise programme. Clinical Rehabilitation. 15:360-370.
Watkins N. (2009). SO2 Network Review and Background; OAQPS; Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
NC. Sulfur Dioxide NAAQS Review Docket. (OAR-2005-0352). Sulfur Dioxide Review
Docket. Docket ID No. EPA-HQ-OAR-2007-0352. Available at www.regulations.gov.
WHO. (2008). Harmonization Project Document No. 6. Part 1: Guidance Document on
Characterizing and Communicating Uncertainty in Exposure Assessment. Available at:
http://www.who.int/ipcs/methods/harmonization/areas/exposure/en/.
Wen X-J, Balluz L, Mokdad A. (2005). Association between media alerts of Air Quality Index
and change of outdoor activity among adult asthma in six states, BRFSS, 2005. J Community
Health. 34:40-46
July 2009 412
-------
Wheeler A, Zanobetti A, Gold DR, Schwartz J, Stone P, Suh HH. (2006). The relationship
between ambient air pollution and heart rate variability differs for individuals with heart and
pulmonary di sease. Environ Health Per sped. 114(4): 5 60-6.
Whitfield R, Biller W, Jusko M, Keisler J. (1996). A Probabilistic Assessment of Health Risks
Associated with Short- and Long-Term Exposure to Tropospheric Ozone. Argonne National
Laboratory, Argonne IL. Available at:
http://www.epa.gov/ttnnaaqs/standards/ozone/data/riskrep.pdf.
Wilson AM, Wake CP, Kelly T, Salloway JC. (2005). Air pollution, weather, and respiratory
emergency room visits in two northern New England cities: An ecological time-series study.
Environ Res. 97:312-321.
Winterton DL, Kaufman J, Keener CV, Quigley S, Farin FM, Williams PV, Koenig JQ. (2001).
Genetic polymorphisms as biomarkers of sensitivity to inhaled sulfur dioxide in subjects with
asthma. Ann Allergy Asthma Immunol. 86:232-238.
Xue J, Liu SV, Ozkaynak H, Spengler J. (2005). Parameter evaluation and model validation of
ozone exposure assessment using Harvard Southern California Chronic Ozone Exposure
Study Data. J Air & Waste Manage Assoc. 55:1508-1515.
Yang Q, Chen Y, Shi Y, Burnett RT, McGrail KM, Krewski D. (2003). Association between
ozone and respiratory admissions among children and the elderly in Vancouver, Canada.
Inhalation Toxicol. 15:1297-1308.
Yang Q, Chen Y, Krewski D, Burnett RT, Shi Y, McGrail KM. (2005). Effect of short-term
exposure to low levels of gaseous pollutants on chronic obstructive pulmonary disease
hospitalizations. Environ Res. 99:99-105.
Yen ST, Shaw WD, Eiswerth ME. (2004). Asthma patients' activities and air pollution: a
semiparametric censored regression analysis. Review of Economics of the Household.
2(1):1569-7152.
Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, Cohen A. (2000).
Exposure measurement error in time-series studies of air pollution? Official statement of the
American Thoracic Society: concepts and consequences. Environ Health Perspect.
108:877-685.
July 2009 413
-------
Appendix A: Supplement to the SO2 Air Quality
Characterization
A-l
-------
Overview
This appendix contains supplementary information on the 862 ambient monitoring data
used in the air quality characterization described in Chapter 7 of the SO2 REA. Included in this
appendix are spatial and temporal attributes important for understanding the relationship between
the ambient monitor and those sources affecting air quality measurements.
In section A.I, important spatial characteristics described include the physical locations
of the ambient monitors (e.g., U.S. states, counties, territories, and cities). Temporal attributes of
interest include, for example, the number of samples collected, sample averaging times, and
years of monitoring data available. Attributes of the monitors that reported both the 5-minute
maximum and the 1-hour SC>2 concentrations are given in Tables A. 1-1 and A. 1-2, while the
supplemental characteristics of the broader ambient monitoring network are given in Table A. 1-3
and A. 1-4. The method for calculating the proximity of the ambient monitors follows, along
with the distance and emission results summarized in Table A. 1-5.
Section A.2 details the analyses performed on simultaneous concentrations, some of
which are the result of co-located monitoring instruments, others the result of duplicate
reporting. Simultaneous measurements were identified by staff using monitor IDs and multiple
concentrations present given the hour-of-day on each available date. Staff estimated a relative
percent difference between the simultaneous measurements at each monitor.
Section A-3 has the tables summarizing the COV and GSD peak-to-mean ratio (PMRs).
Section A-4 has tables summarizing the individual factors used in adjusting ambient air quality
to just meet the current and potential alternative SC>2 air quality standards. Section A-5
summarizes measured 1-hour concentrations and number of days per year with air quality
benchmark exceedances occurring at the 98 monitors reporting 5-minute maximum SC>2
concentrations.
A-2
-------
A.1 Spatial and Temporal Attributes of Ambient SO2 Monitors
A-3
-------
Table A.1-1. Meta-data for 98 ambient monitors reporting 5-minute maximum and corresponding 1-hour SO2
concentrations.
State
AR
AR
AR
CO
DC
DE
FL
IA
IA
IA
IA
IA
IA
IA
IA
IA
LA
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Pulaski
Pulaski
Union
Denver
District of Columbia
New Castle
Nassau
Cerro Gordo
Clinton
Muscatine
Muscatine
Muscatine
Scott
Van Buren
Van Buren
Wood bury
West Baton Rouge
Buchanan
Buchanan
Greene
Greene
Iron
Iron
Jefferson
Jefferson
Jefferson
Jefferson
Monroe
Pike
Saint Charles
Monitor ID
051190007
051191002
051390006
080310002
110010041
100031008
120890005
190330018
190450019
191390016
191390017
191390020
191630015
191770005
191770006
191930018
221210001
290210009
290210011
290770026
290770037
290930030
290930031
290990004
290990014
290990017
290990018
291370001
291630002
291830010
Latitude
34.756111
34.830556
33.215
39.75119
38.897222
39.577778
30.658333
43.16944
41.823283
41.419429
41.387969
41.407796
41.530011
40.689167
40.695078
42.399444
30.501944
39.731389
39.731389
37.128333
37.11
37.466389
37.519444
38.2633
38.267222
38.252778
38.297694
39.473056
39.3726
38.579167
Longitude
-92.275833
-92.259444
-92.668889
-104.98762
-76.952778
-75.611111
-81.463333
-93.202426
-90.211982
-91.070975
-91 .054504
-91 .062646
-90.587611
-91 .994444
-92.006318
-96.355833
-91.209722
-94.8775
-94.868333
-93.261667
-93.251944
-90.69
-90.7125
-90.3785
-90.379444
-90.393333
-90.384333
-91.789167
-90.9144
-90.841111
Objective1
POP
HIC
UNK
HIC
POP
UNK
HIC
UNK
UNK
UNK
UNK
UNK
HIC
UNK
GEN
POP
HIC
GEN
GEN
POP
POP
SRC
UNK
POP
OTH
UNK
HIC
UNK
HIC
UNK
Setting2
URB
RUR
URB
URB
URB
RUR
SUB
SUB
URB
URB
SUB
SUB
URB
RUR
RUR
URB
SUB
URB
URB
SUB
RUR
RUR
RUR
RUR
RUR
SUB
SUB
RUR
RUR
RUR
Land Use3
COM
FOR
COM
COM
RES
AGR
IND
RES
IND
RES
IND
IND
RES
FOR
FOR
RES
COM
IND
IND
RES
RES
RES
AGR
IND
RES
RES
RES
UNK
RES
AGR
Scale4
NEI
NEI
NEI
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
(m)
4
4
4
5
2
4
3
4
4
4
3
3
3
2
3
3
3
4
4
2
3
4
5
5
3
3
Years
n
6
5
11
10
6
2
4
5
5
5
5
5
5
4
2
2
4
4
4
11
11
8
8
4
5
4
3
11
3
2
First
2002
1997
1997
1997
2000
1997
2002
2001
2001
2001
2001
2001
2001
2001
2004
2001
1997
1997
2000
1997
1997
1997
1997
2004
1997
1998
2001
1997
2005
1997
Last
2007
2001
2007
2006
2007
1998
2005
2005
2005
2005
2005
2005
2005
2004
2005
2002
2000
2000
2003
2007
2007
2004
2004
2007
2001
2001
2003
2007
2007
1998
A-4
-------
State
MO
MT
MT
MT
MT
MT
MT
MT
NC
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
PA
PA
PA
PA
County
Saint Charles
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Forsyth
New Hanover
Billings
Billings
Burke
Burke
Burleigh
Cass
Cass
Dunn
McKenzie
McKenzie
McKenzie
Mercer
Mercer
Morton
Morton
Oliver
Steele
Williams
Williams
Allegheny
Allegheny
Allegheny
Allegheny
Monitor ID
291831002
301110066
301110079
301110080
301110082
301110083
301110084
301112008
370670022
371290006
380070002
380070003
380130002
380130004
380150003
380171003
380171004
380250003
380530002
380530104
380530111
380570001
380570004
380590002
380590003
380650002
380910001
381050103
381050105
420030002
420030021
420030031
420030032
Latitude
38.8725
45.788318
45.769439
45.777149
45.783889
45.795278
45.831453
45.786389
36.110556
34.268403
46.8943
46.9619
48.9904
48.64193
46.825425
46.910278
46.933754
47.3132
47.5812
47.575278
47.605556
47.258853
47.298611
46.84175
46.873075
47.185833
47.599703
48.408834
48.392644
40.500556
40.413611
40.443333
40.414444
Longitude
-90.226389
-108.459536
-108.574292
-108.47436
-108.515
-108.455833
-108.449964
-108.523056
-80.226667
-77.956529
-103.37853
-103.356699
-102.7815
-102.4018
-100.76821
-96.795
-96.85535
-102.5273
-103.2995
-103.968889
-104.017222
-101.783035
-101.766944
-100.870059
-100.905039
-101.428056
-97.899009
-102.90765
-102.910233
-80.071944
-79.941389
-79.990556
-79.942222
Objective1
UNK
SRC
POP
UNK
POP
SRC
POP
UNK
POP
GEN
GEN
HIC
SRC
REG
POP
POP
POP
GEN
GEN
SRC
SRC
POP
POP
SRC
SRC
SRC
GEN
SRC
SRC
POP
POP
POP
UNK
Setting2
RUR
RUR
SUB
RUR
URB
SUB
SUB
URB
URB
RUR
RUR
RUR
RUR
RUR
SUB
SUB
SUB
RUR
RUR
RUR
RUR
SUB
RUR
SUB
SUB
RUR
RUR
RUR
RUR
SUB
SUB
URB
SUB
Land Use3
AGR
RES
COM
AGR
COM
AGR
RES
RES
RES
IND
AGR
IND
AGR
AGR
RES
RES
AGR
AGR
AGR
AGR
IND
RES
AGR
IND
IND
AGR
AGR
IND
IND
RES
RES
COM
RES
Scale4
NEI
NEI
NEI
NEI
URB
REG
URB
REG
REG
URB
URB
URB
REG
REG
URB
URB
NEI
URB
NEI
NEI
URB
REG
URB
URB
NEI
NEI
NEI
Height
(m)
2
3.5
4.5
4
3
4
4.5
3
3
3
12.2
4
4
4
4
4
3
4
4
3
3
5
4
4
4
3
3
4
4
6
6
13
5
Years
n
4
7
4
5
3
5
4
1
8
4
10
1
7
5
3
2
10
11
9
10
10
3
9
9
8
11
4
6
6
3
4
3
3
First
1997
1997
1997
1997
2001
1999
2003
1997
1997
1999
1998
1997
1999
2003
2005
1997
1998
1997
1997
1998
1998
1997
1999
1997
1998
1997
1997
2002
2002
1997
1997
1997
1997
Last
2000
2003
2003
2001
2003
2003
2006
1997
2004
2002
2007
1997
2005
2007
2007
1998
2007
2007
2007
2007
2007
1999
2007
2005
2005
2007
2000
2007
2007
1999
2002
1999
1999
A-5
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
UT
WV
WV
WV
County
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Berks
Cambria
Erie
Philadelphia
Philadelphia
Philadelphia
Warren
Warren
Washington
Washington
Washington
Barnwell
Charleston
Charleston
Georgetown
Greenville
Lexington
Oconee
Richland
Richland
Richland
Salt Lake
Wayne
Wayne
Wayne
Monitor ID
420030064
420030067
420030116
420031301
420033003
420033004
420070002
420070005
420110009
420210011
420490003
421010022
421010048
421010136
421230003
421230004
421250005
421250200
421255001
450110001
450190003
450190046
450430006
450450008
450630008
450730001
450790007
450790021
450791003
490352004
540990002
540990003
540990004
Latitude
40.323611
40.381944
40.473611
40.4025
40.318056
40.305
40.56252
40.684722
40.320278
40.309722
42.14175
39.916667
39.991389
39.9275
41.857222
41 .844722
40.146667
40.170556
40.445278
33.320344
32.882289
32.941023
33.362014
34.838814
34.051017
34.805261
34.093959
33.81468
34.024497
40.736389
38.39186
38.390278
38.380278
Longitude
-79.868333
-80.185556
-80.077222
-79.860278
-79.881111
-79.888889
-80.503948
-80.359722
-75.926667
-78.915
-80.038611
-75.188889
-75.080833
-75.222778
-79.1375
-79.169722
-79.902222
-80.261389
-80.420833
-81 .465537
-79.977538
-79.657187
-79.294251
-82.402918
-81.15495
-83.2377
-80.962304
-80.781135
-81.036248
-112.210278
-82.583923
-82.585833
-82.583889
Objective1
POP
GEN
POP
HIC
POP
UNK
REG
POP
HIC
HIC
HIC
HIC
UNK
POP
HIC
HIC
POP
POP
REG
SRC
POP
SRC
SRC
POP
SRC
REG
OTH
GEN
POP
HIC
POP
HIC
HIC
Setting2
SUB
RUR
SUB
SUB
SUB
SUB
RUR
RUR
SUB
URB
SUB
URB
RUR
URB
SUB
RUR
SUB
SUB
RUR
RUR
URB
RUR
URB
URB
SUB
RUR
SUB
RUR
URB
RUR
RUR
RUR
RUR
Land Use3
RES
RES
RES
RES
IND
RES
AGR
AGR
RES
COM
COM
IND
RES
RES
RES
FOR
COM
RES
AGR
FOR
COM
FOR
IND
COM
COM
FOR
COM
FOR
COM
IND
IND
RES
RES
Scale4
NEI
NEI
NEI
NEI
REG
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
REG
URB
NEI
REG
NEI
NEI
NEI
REG
NEI
URB
MID
NEI
NEI
NEI
Height
(m)
8
9
5
9
5
8
3
3
4
12
4
7
5
4
4
4
2
4
4
3.1
4.3
4
2.13
4
3.35
4.3
3
4.42
4
4
3
3
Years
n
4
3
4
3
4
3
2
8
3
3
3
5
3
7
2
2
3
3
2
3
3
3
3
3
2
3
3
3
2
2
1
4
4
First
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2000
2000
2000
2000
2000
2001
2000
2000
2000
2001
1997
2002
2002
2002
Last
2002
1999
2002
1999
2002
1999
1998
2007
1999
1999
1999
2001
1999
2003
1998
1998
1999
1999
1998
2002
2002
2002
2002
2002
2002
2002
2002
2002
2002
1998
2002
2005
2005
A-6
-------
State
WV
wv
County
Wayne
Wood
Monitor ID
540990005
541071002
Latitude
38.372222
39.323533
Longitude
-82.588889
-81.552367
Objective1
HIC
POP
Setting2
RUR
SUB
Land Use3
RES
IND
Scale4
NEI
URB
Height
(m)
3
4
Years
n
4
5
First
2002
2001
Last
2005
2005
Notes:
1 Objectives are POP=Population Exposure; HIC=Highest Concentration; SRC=Source Oriented; GEN=General/Background; REG=Regional
Transport; OTH=Other; UNK=Unknown
2 Settings are R=Rural; U=Urban and Center City; S=Suburban
3 Land Uses are AGR=Agricultural; COM=Commercial; IND=lndustrial; FOR=Forest; RES=Residential; UNK=Unknown
4 Scales are NEI=Neighborhood; MID=Middle; URB=URBAN; REG=Regional
1
2
A-7
-------
Table A.1-2. Population density, concentration variability, and total SO2 emissions associated with 98 ambient
monitors reporting 5-minute maximum and corresponding 1-hour SO2 concentrations.
State
AR
AR
AR
CO
DE
DC
FL
IA
IA
IA
IA
IA
IA
IA
IA
IA
LA
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Pulaski
Pulaski
Union
Denver
New Castle
District of Columbia
Nassau
Cerro Gordo
Clinton
Muscatine
Muscatine
Muscatine
Scott
Van Buren
Van Buren
Wood bury
West Baton Rouge
Buchanan
Buchanan
Greene
Greene
Iron
Iron
Jefferson
Jefferson
Jefferson
Jefferson
Monroe
Pike
Saint Charles
Saint Charles
Monitor ID
051190007
051191002
051390006
080310002
100031008
110010041
120890005
190330018
190450019
191390016
191390017
191390020
191630015
191770005
191770006
191930018
221210001
290210009
290210011
290770026
290770037
290930030
290930031
290990004
290990014
290990017
290990018
291370001
291630002
291830010
291831002
Population Residing Within:
5km
67784
45800
21877
189782
5386
216129
17963
21247
24561
20360
11109
20360
90863
994
994
4449
21249
23253
28224
41036
21784
1121
0
15049
11967
19711
12258
0
645
2637
4587
10km
178348
109372
29073
574752
80025
813665
21386
30341
37638
27101
27101
27101
201277
2252
2252
44815
137455
72613
75073
146752
110681
1121
3799
33379
35082
36471
41709
1439
2077
6349
95765
15km
270266
230200
32652
1158644
192989
1461563
38521
39284
42404
31886
31696
31886
268535
3764
3764
92956
239718
87121
86317
224445
210953
4507
6585
64516
61963
60199
79196
2093
6916
34541
273147
20km
334649
310362
36340
1608099
391157
2029936
48316
45105
45947
40248
36604
40290
293627
6984
6984
112802
366741
93365
93365
256158
254437
8447
8436
124301
125932
116882
170110
5612
11249
90953
431484
Analysis Bins
Population1
hi
mid
mid
hi
low
hi
mid
mid
mid
mid
mid
mid
hi
low
low
low
mid
mid
mid
mid
mid
low
low
mid
mid
mid
mid
low
low
low
low
cov2
a
a
b
b
b
a
c
c
b
b
b
c
b
b
a
b
b
c
b
c
c
c
c
c
c
c
c
a
b
b
b
GSD3
a
a
a
b
b
a
b
c
c
b
b
c
c
b
b
c
b
b
b
b
b
c
b
c
b
b
b
a
b
b
b
Emissions
(tpy)4
20
20
2527
26354
39757
18325
5050
10737
9388
31137
31054
31054
9415
36833
31242
3563
3563
9206
9206
43340
43340
55725
55725
55725
32468
13495
47610
67735
A-8
-------
State
MT
MT
MT
MT
MT
MT
MT
NC
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
PA
PA
PA
PA
PA
County
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Forsyth
New Hanover
Billings
Billings
Burke
Burke
Burleigh
Cass
Cass
Dunn
McKenzie
McKenzie
McKenzie
Mercer
Mercer
Morton
Morton
Oliver
Steele
Williams
Williams
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Monitor ID
301110066
301110079
301110080
301110082
301110083
301110084
301112008
370670022
371290006
380070002
380070003
380130002
380130004
380150003
380171003
380171004
380250003
380530002
380530104
380530111
380570001
380570004
380590002
380590003
380650002
380910001
381050103
381050105
420030002
420030021
420030031
420030032
420030064
Population Residing Within:
5 km
27389
61645
33774
58256
27620
22577
61335
61669
17957
0
0
0
655
49591
48975
2118
0
0
0
0
3280
3280
17925
10305
0
0
0
0
83332
170777
183843
174072
64846
10km
79644
89282
86065
94753
76641
59919
95574
170320
83529
0
888
0
655
67377
134561
91149
0
596
521
0
3280
4428
67959
31348
0
934
1259
1259
277442
560187
580429
558904
201143
15km
98733
102887
104825
103200
98733
97912
103200
258102
145330
1887
1887
0
655
83082
144878
145789
0
596
521
2283
5902
5902
75685
75685
2057
934
1259
1259
651551
921490
877668
922097
520438
20km
107178
114640
108399
1 06046
109475
110980
1 06046
325974
1 70260
1887
1887
625
655
84415
154455
148002
537
596
2283
5771
6465
7455
84415
82584
2670
934
1827
1827
961378
1142754
1145039
1144558
943781
Analysis Bins
Population1
mid
hi
mid
hi
mid
mid
hi
hi
mid
low
low
low
low
mid
mid
low
low
low
low
low
low
low
mid
mid
low
low
low
low
hi
hi
hi
hi
hi
cov2
b
b
b
b
b
b
b
b
c
a
a
b
b
b
b
a
a
a
b
c
b
b
c
b
b
a
c
b
b
b
a
b
b
GSD3
c
a
b
a
b
b
b
b
c
a
a
b
b
a
a
b
a
a
a
a
b
a
c
b
b
a
b
c
b
b
b
b
c
Emissions
(tpy)4
5480
5480
5480
5480
5480
15298
5480
3945
30020
283
283
426
4592
771
756
5
210
823
91617
91617
4592
4592
28565
1605
1605
1964
52447
46957
52447
11490
A-9
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
UT
WV
WV
WV
WV
County
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Berks
Cambria
Erie
Philadelphia
Philadelphia
Philadelphia
Warren
Warren
Washington
Washington
Washington
Barnwell
Charleston
Charleston
Georgetown
Greenville
Lexington
Oconee
Richland
Richland
Richland
Salt Lake
Wayne
Wayne
Wayne
Wayne
Monitor ID
420030067
420030116
420031301
420033003
420033004
420070002
420070005
420110009
420210011
420490003
421010022
421010048
421010136
421230003
421230004
421250005
421250200
421255001
450110001
450190003
450190046
450430006
450450008
450630008
450730001
450790007
450790021
450791003
490352004
540990002
540990003
540990004
540990005
Population Residing Within:
5 km
13277
96820
115432
55221
38588
3434
17292
121330
50440
81199
316944
262592
382995
14142
13965
31276
32125
1359
0
40872
1103
10567
70221
42208
0
35872
1666
87097
0
17320
17320
16553
13314
10km
86792
331624
411867
202092
170065
28961
77240
203799
79710
150626
985213
1102727
985957
19940
18884
68512
52910
15854
4022
132716
1103
18215
173012
131361
2260
121006
4643
213836
4074
62645
59989
54251
48330
15km
324154
704601
766188
509708
461433
68617
143738
250610
102905
190212
1726387
1938877
1718068
25715
28805
111222
83324
43364
13647
273298
9529
22467
284047
257820
11136
255135
13324
300874
35159
124477
123349
122072
114824
20km
610975
996267
1088115
944188
904760
120780
224631
309553
124592
209983
2446142
2607877
2381173
32490
33523
183285
118188
126091
21554
364953
22255
34357
379022
355854
26182
353072
33098
396116
124394
178576
177744
179815
173807
Analysis Bins
Population1
mid
hi
hi
hi
mid
low
mid
hi
hi
hi
hi
hi
hi
mid
mid
mid
mid
low
low
mid
low
mid
hi
mid
low
mid
low
hi
low
mid
mid
mid
mid
cov2
a
b
b
b
b
b
b
a
a
b
a
b
b
b
b
a
b
b
a
b
b
b
a
b
a
a
b
a
a
a
b
b
b
GSD3
b
b
b
c
b
b
c
b
b
b
b
b
b
b
c
b
b
b
a
b
a
b
a
b
a
a
a
a
a
b
b
b
b
Emissions
(tpy)4
1167
1964
52100
11490
11501
187257
41385
14817
16779
4122
18834
6214
21700
4890
4890
8484
7
2566
65
34934
40841
1067
10433
5
613
40492
12935
3735
10172
10172
10172
10172
A-10
-------
State
WV
County
Wood
Monitor ID
541071002
Population Residing Within:
5 km
24917
10km
70324
15km
1 04458
20km
128127
Analysis Bins
Population1
mid
COV2
b
GSD3
b
Emissions
(tpy)4
48124
Notes:
1 Population bins: low (<1 0,000); mid (10,001 to 50,000); hi (>50,000) using population within 5 km of ambient monitor.
2 COV bins: a (<100%); b (>100 to <200); c (>200).
3 GSD bins: a (<2.17); b (>2.17 to <2.94); c (>2.94).
4 Sum of emissions within 20 km radius of ambient monitor based on 2002 NEI.
A-ll
-------
Table A.1-3. Meta-data for 809 ambient monitors in the broader SO2 monitoring network.
State
AL
AL
AL
AL
AL
AL
AL
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AR
AR
AR
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
County
Colbert
Jackson
Jefferson
Lawrence
MOB
MOB
Montgomery
Gila
Gila
Maricopa
Maricopa
Maricopa
Pima
Final
Pulaski
Pulaski
Union
Alameda
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Imperial
Los Angeles
Los Angeles
Los Angeles
Monitor ID
010330044
010710020
010731003
010790003
010970028
010972005
011011002
040070009
040071001
040130019
040133002
040133003
040191011
040212001
051190007
051191002
051390006
060010010
060130002
060130006
060130010
060131001
060131002
060131003
060131004
060132001
060133001
060250005
060371002
060371103
060374002
Latitude
34.690556
34.876944
33.485556
34.589571
30.958333
30.474674
32.40712
33.399135
33.006179
33.48385
33.45793
33.47968
32.208333
32.600479
34.756111
34.830556
33.215
37.7603
37.936
37.9478
38.0313
38.055556
38.010556
37.964167
37.96028
38.013056
38.029167
32.676111
34.17605
34.06659
33.82376
Longitude
-87.821389
-85.720833
-86.915
-87.109445
-88.028333
-88.14114
-86.256367
-110.858896
-110.785797
-112.14257
-112.04601
-111.91721
-110.872222
-110.633598
-92.275833
-92.259444
-92.668889
-122.1925
-122.0262
-122.3651
-122.1318
-122.219722
-121.641389
-122.339167
-122.35667
-122.133611
-121.902222
-115.483333
-118.31712
-118.22688
-118.18921
Objective1
UNK
LINK
HIC
UNK
HIC
POP
HIC
SRC
SRC
UNK
HIC
POP
POP
POP
POP
HIC
UNK
POP
SRC
UNK
POP
SRC
UNK
UNK
POP
UNK
HIC
UNK
UNK
UNK
POP
Setting2
RUR
RUR
SUB
RUR
SUB
RUR
SUB
URB
URB
SUB
URB
SUB
SUB
SUB
URB
RUR
URB
SUB
SUB
URB
URB
SUB
RUR
URB
URB
URB
URB
SUB
URB
URB
SUB
Land Use3
AGR
AGR
RES
AGR
IND
AGR
COM
RES
IND
RES
RES
RES
RES
RES
COM
FOR
COM
RES
RES
IND
COM
IND
AGR
COM
COM
RES
RES
RES
COM
RES
RES
Scale4
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
4
4
4
1
6
4
11.3
5.8
5
4
4
4
4
8.3
8.5
7
7
6
20
9
7
5
11
7
Years
n
9
9
9
2
3
3
1
7
7
1
9
7
9
4
5
5
7
1
9
9
1
8
9
4
3
9
9
6
7
6
9
First
1997
1997
1997
1998
1997
2002
1997
1999
1999
1998
1997
1998
1998
1998
2002
1997
1997
2002
1997
1997
2002
1997
1997
1998
2003
1997
1997
1999
1998
1997
1997
Last
2005
2005
2006
1999
1999
2004
1997
2005
2005
1998
2006
2006
2006
2005
2006
2001
2006
2002
2005
2005
2002
2004
2005
2001
2005
2005
2005
2005
2005
2005
2005
A-12
-------
State
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
County
Los Angeles
Los Angeles
Orange
Riverside
Sacramento
Sacramento
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Diego
San Diego
San Diego
San Francisco
San Luis Obispo
San Luis Obispo
San Luis Obispo
San Luis Obispo
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Monitor ID
060375001
060375005
060591003
060658001
060670002
060670006
060710012
060710014
060710306
060711234
060712002
060714001
060730001
060731007
060732007
060750005
060791005
060792001
060792004
060794002
060830008
060831012
060831013
060831015
060831016
060831019
060831020
060831025
060831026
060831027
060832004
060832011
060834003
Latitude
33.92288
33.9508
33.67464
33.99958
38.712778
38.614167
34.426111
34.5125
34.51
35.763889
34.10002
34.418056
32.631231
32.709172
32.552164
37.766
35.043889
35.125
35.022222
35.028333
34.462222
34.451944
34.725556
34.478056
34.477778
34.475278
34.415278
34.489722
34.479444
34.469167
34.6375
34.445278
34.596111
Longitude
-118.37026
-118.43043
-117.92568
-117.41601
-121.38
-121.366944
-117.563056
-117.33
-117.330556
-117.396111
-117.49201
-117.284722
-117.059075
-117.153975
-116.937772
-122.3991
-120.580278
-120.633333
-120.569444
-120.387222
-120.024444
-120.457778
-120.427778
-120.210833
-120.205556
-120.188889
-119.878611
-120.045833
-120.0325
-120.039444
-120.456389
-119.827778
-120.630278
Objective1
POP
UPW
UNK
POP
UNK
HIC
UNK
UNK
UNK
OTH
POP
UNK
POP
POP
POP
UNK
UNK
UNK
UNK
POP
POP
UNK
UNK
UNK
UNK
UNK
UNK
UNK
UNK
UNK
POP
POP
UNK
Setting2
URB
SUB
SUB
SUB
SUB
SUB
RUR
SUB
SUB
RUR
SUB
SUB
SUB
URB
RUR
URB
RUR
SUB
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
URB
SUB
RUR
Land Use3
COM
RES
RES
RES
RES
RES
COM
RES
RES
DES
IND
RES
RES
COM
MOB
IND
COM
RES
IND
RES
UNK
AGR
AGR
AGR
AGR
AGR
AGR
AGR
AGR
AGR
COM
RES
AGR
Scale4
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
REG
REG
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
2
4
6
7
5
5
4
4
1
5
7
5
5
4
5
4
4
4
Years
n
7
1
9
7
7
9
1
3
7
8
5
1
9
8
8
9
5
5
9
7
9
1
9
1
1
1
6
9
2
2
9
9
8
First
1997
2005
1997
1997
1997
1997
1997
1997
2000
1998
1997
1997
1997
1997
1997
1997
1997
1997
1997
2000
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
Last
2003
2005
2005
2005
2006
2006
1997
1999
2006
2006
2005
1997
2005
2004
2004
2005
2001
2002
2006
2006
2005
1997
2005
1997
1997
1997
2005
2005
1998
1998
2005
2005
2005
A-13
-------
State
CA
CA
CA
CA
CO
CO
CO
CO
CO
CO
CO
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
DE
DE
DE
DE
DE
DE
DC
County
Santa Cruz
Solano
Solano
Ventura
Adams
Adams
Denver
El Paso
El Paso
El Paso
El Paso
Fairfield
Fairfield
Fairfield
Fairfield
Fairfield
Hartford
Hartford
Hartford
New Haven
New Haven
New Haven
New Haven
New London
Tolland
New Castle
New Castle
New Castle
New Castle
New Castle
New Castle
District of
Columbia
Monitor ID
060870003
060950001
060950004
061113001
080010007
080013001
080310002
080416001
080416004
080416011
080416018
090010012
090010017
090011123
090012124
090019003
090031005
090031018
090032006
090090027
090091003
090091123
090092123
090110007
090130003
100031003
100031007
100031008
100031013
100032002
100032004
110010041
Latitude
37.011944
38.052222
38.1027
34.255
39.8
39.83818
39.75119
38.633611
38.921389
38.846667
38.811389
41.195
41.003611
41.399167
41.063056
41.118333
42.015833
41.760833
41.7425
41.301111
41.310556
41.310833
41.550556
41.361111
41.73
39.761111
39.551111
39.577778
39.773889
39.757778
39.739444
38.897222
Longitude
-122.193333
-122.144722
-122.2382
-119.1425
-104.910833
-104.94984
-104.98762
-104.715556
-104.8125
-104.827222
-104.751389
-73.163333
-73.585
-73.443056
-73.528889
-73.336667
-72.518056
-72.670833
-72.634444
-72.902778
-72.915556
-72.916944
-73.043611
-72.08
-72.213611
-75.491944
-75.730833
-75.611111
-75.496389
-75.546389
-75.558056
-76.952778
Objective1
UNK
UNK
UNK
HIC
POP
POP
HIC
UNK
UNK
UNK
UNK
HIC
UNK
UNK
HIC
POP
POP
POP
HIC
POP
UNK
HIC
POP
UNK
UNK
HIC
UNK
UNK
POP
POP
UNK
POP
Setting2
RUR
URB
URB
RUR
URB
RUR
URB
RUR
URB
URB
URB
URB
SUB
SUB
URB
RUR
RUR
URB
SUB
URB
SUB
URB
URB
SUB
SUB
SUB
RUR
RUR
SUB
URB
URB
URB
Land Use3
RES
COM
COM
RES
RES
AGR
COM
IND
RES
RES
COM
RES
RES
RES
RES
FOR
AGR
COM
IND
COM
IND
RES
MOB
RES
COM
RES
AGR
AGR
RES
COM
COM
RES
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
6
8
4
4
4
5
4
4
3
3
3
3
3
3
3
9
3.67
5
5
5
3
3
4
6
Years
n
9
1
9
7
2
9
8
4
3
4
3
9
1
9
8
8
2
1
9
1
1
7
9
2
2
6
3
8
2
2
6
10
First
1997
1997
1997
1997
2002
1997
1997
1997
1997
1997
1998
1997
1997
1997
1997
1998
1997
1997
1997
2005
1997
1997
1997
1997
1997
1997
2002
1997
2004
1997
2000
1997
Last
2006
1997
2005
2003
2003
2005
2006
2000
1999
2000
2000
2005
1997
2005
2004
2005
1998
1997
2005
2005
1997
2003
2005
1998
1998
2002
2006
2006
2006
1998
2006
2006
A-14
-------
State
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
GA
GA
County
B reward
Duval
Duval
Duval
Duval
Escambia
Escambia
Hamilton
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Manatee
Miami-Dade
Nassau
Nassau
Orange
Palm Beach
Pinellas
Pinellas
Pinellas
Pinellas
Polk
Polk
Putnam
Sarasota
Sarasota
Sarasota
Baldwin
Bartow
Monitor ID
120110010
120310032
120310080
120310081
120310097
120330004
120330022
120470015
120570021
120570053
120570081
120570095
120570109
120571035
120574004
120813002
120860019
120890005
120890009
120952002
120993004
121030023
121033002
121035002
121035003
121050010
121052006
121071008
121151002
121151005
121151006
130090001
130150002
Latitude
26.128611
30.356111
30.308889
30.422222
30.367222
30.525
30.544722
30.411111
27.947222
27.886389
27.739722
27.9225
27.856389
27.928056
27.9925
27.632778
25.8975
30.658333
30.686389
28.599444
26.369722
27.863333
27.871389
28.09
28.141667
27.856111
27.896944
29.6875
27.299722
27.306944
27.350278
33.153258
34.103333
Longitude
-80.167222
-81.635556
-81 .6525
-81.621111
-81.594167
-87.204167
-87.216111
-82.783611
-82.453333
-82.481389
-82.465278
-82.401389
-82.383667
-82.454722
-82.125833
-82.546111
-80.38
-81.463333
-81 .4475
-81.363056
-80.074444
-82.623333
-82.691667
-82.700833
-82.739722
-82.017778
-81.960278
-81.656667
-82.524444
-82.570556
-82.48
-83.235807
-84.915278
Objective1
HIC
HIC
HIC
HIC
POP
POP
HIC
UNK
HIC
POP
UNK
HIC
POP
POP
HIC
POP
POP
HIC
HIC
HIC
HIC
POP
HIC
HIC
HIC
HIC
HIC
HIC
HIC
POP
POP
SRC
POP
Setting2
SUB
SUB
SUB
SUB
SUB
SUB
SUB
RUR
RUR
SUB
UNK
SUB
SUB
SUB
SUB
RUR
UNK
SUB
SUB
URB
SUB
RUR
SUB
RUR
SUB
RUR
SUB
RUR
SUB
SUB
SUB
RUR
SUB
Land Use3
RES
COM
COM
RES
COM
IND
COM
IND
RES
RES
UNK
COM
COM
RES
RES
IND
UNK
IND
RES
COM
COM
IND
COM
RES
RES
IND
IND
IND
RES
RES
RES
RES
AGR
Scale4
NEI
NEI
MID
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
REG
Height
4
3
3
4
5
4
6
3
2
4
4
4
3
4
4
4
4
2
4
4
10
4
3
4
4
2
4
5
4
5
5
5
Years
n
8
8
8
8
8
9
8
10
3
9
8
9
9
9
6
5
7
8
1
9
6
9
9
9
7
8
6
10
1
4
4
3
5
First
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2000
1999
1997
1997
1997
1997
1997
1997
1997
1997
1999
1997
1997
1997
1997
1997
2000
1998
1997
Last
2005
2004
2005
2005
2005
2005
2005
2006
1999
2005
2005
2005
2005
2005
2005
2004
2003
2006
1997
2005
2002
2005
2005
2005
2005
2004
2002
2006
1997
2000
2003
2006
2004
A-15
-------
State
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
HI
HI
HI
HI
ID
ID
ID
ID
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
County
Bibb
Chatham
Chatham
Chatham
Dougherty
Fannin
Floyd
Fulton
Fulton
Glynn
Muscogee
Richmond
Honolulu
Honolulu
Honolulu
Honolulu
Bannock
Caribou
Caribou
Power
Adams
Champaign
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Monitor ID
130210012
130510019
130510021
130511002
130950006
131110091
131150003
131210048
131210055
131270006
132150008
132450003
150030010
150030011
150031001
150031006
160050004
160290003
160290031
160770011
170010006
170190004
170310050
170310059
170310063
170310064
170310076
170311018
170311601
170312001
170314002
170314201
170318003
Latitude
32.805244
32.093889
32.06905
32.090278
31.567778
34.985556
34.261113
33.779189
33.720428
31.16953
32.521099
33.393611
21.329167
21.337222
21.310278
21.3475
42.916389
42.661298
42.695278
42.9125
39.93301
40.123796
41.70757
41.6875
41.876969
41.790787
41.7514
41.773889
41.66812
41.662109
41.855243
42.139996
41.631389
Longitude
-83.543628
-81.151111
-81 .048949
-81.130556
-84.102778
-84.375278
-85.323018
-84.395843
-84.357449
-81.496046
-84.944695
-82.006389
-158.093333
-158.119167
-157.858056
-158.113333
-112.515833
-111.591443
-111.593889
-112.535556
-91 .404237
-88.229531
-87.568574
-87.536111
-87.63433
-87.601646
-87.713488
-87.815278
-87.99057
-87.696467
-87.75247
-87.799227
-87.568056
Objective1
POP
HIC
SRC
POP
HIC
POP
POP
HIC
POP
POP
POP
POP
SRC
SRC
POP
UNK
HIC
POP
SRC
SRC
POP
POP
POP
HIC
POP
POP
POP
HIC
POP
HIC
POP
POP
POP
Setting2
RUR
SUB
SUB
URB
SUB
URB
RUR
URB
SUB
SUB
SUB
SUB
RUR
RUR
URB
RUR
RUR
URB
RUR
RUR
URB
SUB
SUB
SUB
URB
SUB
SUB
SUB
SUB
SUB
SUB
SUB
SUB
Land Use3
IND
IND
COM
IND
RES
IND
RES
COM
COM
RES
RES
IND
IND
COM
COM
IND
IND
RES
DES
IND
COM
RES
IND
IND
MOB
RES
RES
IND
RES
IND
RES
RES
RES
Scale4
URB
URB
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
MIC
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
URB
NEI
Height
4
4
10
5
4
3
4
5
5
8
4
4
4
10
3
3
4
9
5
8
10
3
15
4
4
4
9
4
8
4
Years
n
3
1
6
3
1
9
10
8
10
1
2
3
9
6
7
9
9
3
4
1
10
4
10
4
10
1
3
8
10
7
10
2
6
First
1998
2000
1998
2004
1998
1997
1997
1999
1997
1999
1999
1997
1997
2000
1998
1997
1997
1999
2002
2004
1997
1997
1997
1997
1997
1997
2004
1997
1997
1997
1997
2004
1997
Last
2003
2000
2006
2006
1998
2006
2006
2006
2006
1999
2005
2001
2005
2005
2004
2005
2005
2001
2005
2004
2006
2000
2006
2000
2006
1997
2006
2004
2006
2003
2006
2005
2002
A-16
-------
State
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
County
DuPage
La Salle
Macon
Macoupin
Madison
Madison
Madison
Madison
Madison
Peoria
Randolph
Rock Island
Saint Clair
Saint Clair
Saint Clair
Sangamon
Tazewell
Wabash
Wabash
Will
Daviess
Dearborn
Floyd
Floyd
Floyd
Fountain
Gibson
Gibson
Hendricks
Hendricks
Hendricks
Jasper
Jasper
Monitor ID
170436001
170990007
171150013
171170002
171190008
171190017
171191010
171193007
171193009
171430024
171570001
171610003
171630010
171631010
171631011
171670006
171790004
171850001
171851001
171970013
180270002
180290004
180430004
180430007
180431004
180450001
180510001
180510002
180630001
180630002
180630003
180730002
180730003
Latitude
41.813049
41.293015
39.866834
39.396075
38.890186
38.701944
38.828303
38.860669
38.865984
40.68742
38.176278
41.511944
38.612034
38.592192
38.235
39.800614
40.55646
38.397222
38.369444
41.459963
38.572778
39.092778
38.367778
38.273333
38.308056
39.964167
38.361389
38.392778
39.876944
39.863361
39.880833
41.187778
41.135833
Longitude
-88.072827
-89.049425
-88.925594
-89.809739
-90.148031
-90.149167
-90.058433
-90.105851
-90.070571
-89.606943
-89.788459
-90.514167
-90.160477
-90.165081
-89.841944
-89.591225
-89.654028
-87.773611
-87.834444
-88.182019
-87.214722
-84.855
-85.833056
-85.836389
-85.834167
-87.421389
-87.748611
-87.748333
-86.473889
-86.47075
-86.542194
-87.053333
-86.987778
Objective1
POP
SRC
POP
POP
SRC
HIC
SRC
POP
SRC
POP
GEN
HIC
POP
HIC
SRC
SRC
SRC
HIC
HIC
SRC
HIC
HIC
HIC
SRC
POP
HIC
HIC
HIC
HIC
HIC
HIC
HIC
HIC
Setting2
SUB
SUB
SUB
RUR
SUB
URB
SUB
SUB
SUB
SUB
RUR
URB
SUB
SUB
RUR
SUB
SUB
URB
RUR
RUR
RUR
SUB
RUR
RUR
SUB
RUR
RUR
RUR
RUR
SUB
SUB
RUR
RUR
Land Use3
AGR
IND
IND
AGR
IND
MOB
IND
IND
COM
COM
IND
COM
IND
IND
IND
IND
IND
MOB
AGR
IND
AGR
COM
AGR
RES
RES
AGR
AGR
AGR
IND
COM
COM
AGR
AGR
Scale4
NEI
NEI
NEI
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
Height
14
5
5
5
15
3
5
10
7
5
5
8
5
7
5
8
6
2
2
13
2
5
5
4
5
2
5
9
3
3
Years
n
4
1
10
10
6
4
10
10
9
10
10
4
10
6
5
10
10
5
6
10
9
9
5
5
10
6
5
5
2
2
2
10
6
First
1997
2006
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2004
2004
2004
1997
1997
Last
2000
2006
2006
2006
2002
2000
2006
2006
2006
2006
2006
2000
2006
2002
2001
2006
2006
2005
2005
2006
2005
2005
2005
2005
2006
2005
2005
2004
2005
2005
2005
2006
2002
A-17
-------
State
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IA
IA
IA
IA
IA
County
Jefferson
Lake
Lake
LaPorte
LaPorte
Marion
Marion
Marion
Marion
Marion
Morgan
Perry
Perry
Pike
Porter
Porter
Porter
Spencer
Spencer
Sullivan
Vanderburgh
Vanderburgh
Vigo
Vigo
Warrick
Warrick
Wayne
Wayne
Cerro Gordo
Clinton
Clinton
Clinton
Lee
Monitor ID
180770004
180890022
180892008
180910005
180910007
180970042
180970054
180970057
180970072
180970073
181091001
181230006
181230007
181250005
181270011
181270017
181270023
181470002
181470010
181530004
181630012
181631002
181670018
181671014
181730002
181731001
181770006
181770007
190330018
190450018
190450019
190450020
191110006
Latitude
38.776667
41.606667
41.639444
41.716944
41.679722
39.646254
39.730278
39.749019
39.768056
39.789167
39.515
37.99433
37.983773
38.519167
41.633889
41.621944
41.616667
37.9825
37.95536
39.099444
38.021667
37.9025
39.486111
39.514722
37.9375
37.938056
39.812222
39.795833
43.16944
41.824722
41.823283
41.845833
40.392222
Longitude
-85.407222
-87.304722
-87.493611
-86.9075
-86.852778
-86.248784
-86.196111
-86.186314
-86.16
-86.060833
-86.391667
-86.763457
-86.772202
-87.249722
-87.101389
-87.116389
-87.145833
-86.96638
-87.0318
-87.470556
-87.569444
-87.671389
-87.401389
-87.407778
-87.314167
-87.345833
-84.89
-84.880833
-93.202426
-90.212778
-90.211982
-90.216389
-91.4
Objective1
HIC
UNK
HIC
HIC
HIC
POP
HIC
HIC
POP
POP
HIC
UNK
UNK
HIC
HIC
HIC
HIC
HIC
HIC
HIC
POP
UNK
POP
HIC
HIC
HIC
HIC
HIC
UNK
UNK
UNK
HIC
UNK
Setting2
SUB
URB
SUB
URB
RUR
RUR
URB
URB
URB
URB
SUB
RUR
RUR
RUR
RUR
RUR
SUB
RUR
RUR
RUR
URB
RUR
URB
RUR
RUR
RUR
SUB
RUR
SUB
SUB
URB
SUB
URB
Land Use3
COM
IND
COM
IND
RES
AGR
IND
RES
COM
RES
RES
IND
IND
AGR
IND
IND
IND
AGR
AGR
AGR
COM
AGR
RES
COM
IND
IND
IND
IND
RES
RES
IND
COM
IND
Scale4
NEI
NEI
NEI
NEI
URB
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
MID
URB
Height
5
5
4
3
4
9
4
3
5
2
4
4
5
5
2
5
9
5
5
4
4
5
9
4
4
7
5
Years
n
8
8
9
10
6
10
1
10
4
9
2
5
5
8
10
6
6
5
4
7
10
10
10
8
5
4
10
10
9
1
10
1
2
First
1997
1998
1997
1997
1997
1997
1997
1997
1997
1997
1997
1998
1998
1997
1997
1997
1997
1997
2002
1997
1997
1997
1997
1997
1997
1997
1997
1997
1998
1997
1997
1997
1997
Last
2004
2005
2006
2006
2002
2006
1997
2006
2000
2005
2005
2003
2003
2005
2006
2002
2002
2001
2005
2005
2006
2006
2006
2005
2006
2002
2006
2006
2006
1997
2006
1997
1998
A-18
-------
State
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
KS
KS
KS
KS
KS
KS
KS
KS
KS
KY
KY
KY
KY
KY
KY
KY
County
Lee
Linn
Linn
Linn
Linn
Linn
Linn
Linn
Muscatine
Muscatine
Muscatine
Scott
Scott
Van Buren
Van Buren
Van Buren
Woodbury
Linn
Montgomery
Pawnee
Sedgwick
Sumner
Trego
Wyandotte
Wyandotte
Wyandotte
Boyd
Boyd
Boyd
Campbell
Campbell
Daviess
Fayette
Monitor ID
191111007
191130028
191130029
191130031
191130032
191130034
191130038
191130039
191390016
191390017
191390020
191630015
191630017
191770004
191770005
191770006
191930018
201070002
201250006
201450001
201730010
201910002
201950001
202090001
202090020
202090021
210190015
210190017
210191003
210370003
210371001
210590005
210670012
Latitude
40.5825
41.910556
41.974722
41.983333
41.964722
41.971111
41.941111
41.934167
41.419429
41.387969
41.407796
41.530011
41.467236
40.711111
40.689167
40.695078
42.399444
38.135833
37.046944
38.17625
37.701111
37.476944
38.770278
39.113056
39.151389
39.1175
38.465833
38.459167
38.388611
39.065556
39.108611
37.780833
38.065
Longitude
-91.4275
-91.651944
-91 .666667
-91.662778
-91 .664722
-91 .645278
-91.633889
-91.6825
-91.070975
-91.054504
-91.062646
-90.587611
-90.688451
-91.975278
-91.994444
-92.006318
-96.355833
-94.731944
-95.613333
-99.108028
-97.313889
-97.366389
-99.763611
-94.624444
-94.6175
-94.635556
-82.621111
-82.640556
-82.6025
-84.451944
-84.476111
-87.075556
-84.5
Objective1
UNK
HIC
HIC
SRC
UNK
UNK
SRC
SRC
UNK
UNK
UNK
HIC
UNK
HIC
UNK
GEN
POP
REG
POP
POP
POP
REG
GEN
HIC
POP
POP
POP
POP
POP
POP
POP
POP
POP
Setting2
RUR
SUB
URB
URB
URB
URB
SUB
URB
URB
SUB
SUB
URB
RUR
RUR
RUR
RUR
URB
RUR
URB
SUB
URB
RUR
RUR
URB
URB
URB
URB
SUB
SUB
SUB
URB
SUB
SUB
Land Use3
IND
COM
COM
RES
RES
RES
IND
IND
RES
IND
IND
RES
IND
FOR
FOR
FOR
RES
AGR
RES
RES
RES
RES
AGR
COM
IND
RES
RES
RES
IND
RES
RES
COM
RES
Scale4
NEI
NEI
MID
MID
NEI
NEI
REG
NEI
NEI
NEI
REG
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
15
8
16
4
4.5
3
4
4
4
4
3
3
3
3
4
4
3
4
4
4
15
9
4
4
3
5
4
4
4
4
Years
n
2
5
9
10
2
2
8
1
10
10
10
8
1
2
4
2
1
6
7
1
1
3
3
2
1
4
3
4
3
6
3
9
9
First
1998
1997
1997
1997
1998
1998
1999
2001
1997
1997
1997
1997
1997
1997
2000
2005
2002
1999
1998
1997
1997
2001
2002
1997
1997
2000
1997
2002
1997
2000
1997
1997
1997
Last
2000
2001
2006
2006
1999
1999
2006
2001
2006
2006
2006
2005
1997
1998
2003
2006
2002
2004
2005
1997
1997
2005
2005
1998
1997
2005
2000
2005
1999
2005
1999
2005
2005
A-19
-------
State
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
LA
ME
ME
ME
ME
ME
ME
ME
ME
ME
MD
MD
MD
MD
County
Greenup
Hancock
Henderson
Henderson
Jefferson
Jefferson
Jefferson
Livingston
McCracken
McCracken
McCracken
Warren
Bossier
Calcasieu
East Baton
Rouge
Ouachita
St. Bernard
West Baton
Rouge
Androscoggin
Aroostook
Aroostook
Aroostook
Aroostook
Aroostook
Cumberland
Cumberland
Oxford
Allegany
Anne Arundel
Baltimore
Baltimore (City)
Monitor ID
210890007
210910012
211010013
211010014
211110032
211110051
211111041
211390004
211450001
211451024
211451026
212270008
220150008
220190008
220330009
220730004
220870002
221210001
230010011
230030009
230030012
230031003
230031013
230031018
230050014
230050027
230172007
240010006
240032002
240053001
245100018
Latitude
38.548333
37.938889
37.858889
37.871389
38.1825
38.060833
38.23163
37.070833
37.131667
37.058056
37.040833
37.036667
32.53626
30.261667
30.46198
32.509713
29.981944
30.501944
44.089406
47.351667
47.354444
47.351667
46.123889
46.660899
43.659722
43.661944
44.543056
39.649722
39.159722
39.310833
39.314167
Longitude
-82.731667
-86.896944
-87.575278
-87.463333
-85.861667
-85.896111
-85.82672
-88.334167
-88.813333
-88.5725
-88.541111
-86.250556
-93.74891
-93.284167
-91.17922
-92.046093
-89.998611
-91.209722
-70.214219
-68.303611
-68.314167
-68.311389
-67.829722
-67.902066
-70.261389
-70.265833
-70.545833
-78.762778
-76.511667
-76.474444
-76.613333
Objective1
POP
POP
POP
POP
HIC
POP
POP
HIC
HIC
POP
POP
POP
POP
POP
HIC
GEN
SRC
HIC
HIC
UNK
LINK
UNK
UNK
SRC
HIC
HIC
UNK
POP
POP
POP
POP
Setting2
SUB
RUR
SUB
RUR
SUB
SUB
SUB
RUR
RUR
SUB
SUB
RUR
URB
RUR
URB
URB
SUB
SUB
URB
SUB
URB
SUB
URB
RUR
URB
URB
SUB
URB
SUB
SUB
URB
Land Use3
RES
RES
RES
COM
RES
RES
IND
AGR
IND
COM
RES
RES
COM
IND
COM
IND
RES
COM
COM
RES
IND
RES
COM
IND
COM
IND
IND
COM
RES
RES
RES
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
4
4
4
4
4
4
5
4
5
4
6
4
3
5
5
4
2
2
4
1
9
3
4
4
4
4
5
5
5
4
Years
n
9
7
5
2
3
10
8
9
3
6
2
3
10
10
10
10
7
10
4
1
1
1
1
1
1
7
8
1
4
2
2
First
1997
1998
1997
2004
1997
1997
1997
1997
1997
2000
1997
2003
1997
1997
1997
1997
1998
1997
1997
1997
1997
1997
1997
2004
1997
2000
1997
1997
1999
2004
1997
Last
2005
2004
2001
2005
2001
2006
2006
2005
1999
2005
1998
2005
2006
2006
2006
2006
2004
2006
2002
1997
1997
1997
1997
2004
1997
2006
2004
1997
2002
2005
1998
A-20
-------
State
MD
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
County
Baltimore (City)
Bristol
Essex
Essex
Essex
Essex
Hampden
Hampden
Hampshire
Middlesex
Middlesex
Suffolk
Suffolk
Suffolk
Suffolk
Suffolk
Suffolk
Suffolk
Worcester
Worcester
Delta
Genesee
Genesee
Kent
Macomb
Missaukee
St. Clair
Schoolcraft
Wayne
Wayne
Wayne
Wayne
Wayne
Monitor ID
245100036
250051004
250090005
250091004
250091005
250095004
250130016
250131009
250154002
250171701
250174003
250250002
250250019
250250020
250250021
250250040
250250042
250251003
250270020
250270023
260410902
260490021
260492001
260810020
260991003
261130001
261470005
261530001
261630001
261630005
261630015
261630016
261630019
Latitude
39.265
41.683279
42.709444
42.515556
42.525
42.772222
42.108581
42.085556
42.298279
42.474444
42.383611
42.348873
42.316394
42.309417
42.377833
42.340251
42.3294
42.401667
42.267222
42.263877
45.796667
43.047224
43.168336
42.984173
42.51334
44.310555
42.953336
46.288877
42.22862
42.267231
42.302786
42.357808
42.43084
Longitude
-76.536667
-71.169171
-71.146389
-70.931389
-70.934167
-71.061111
-72.590614
-72.579722
-72.333904
-71.111111
-71.213889
-71.097163
-70.967773
-71.055573
-71.027138
-71.03835
-71.0825
-71.031111
-71.798889
-71.794186
-87.089444
-83.670159
-83.461541
-85.671339
-83.005971
-84.891865
-82.456229
-85.950227
-83.2082
-83.132086
-83.10653
-83.096033
-83.000138
Objective1
HIC
HIC
HIC
UNK
UNK
OTH
POP
HIC
OTH
UNK
POP
HIC
OTH
OTH
HIC
POP
POP
POP
HIC
POP
UNK
POP
GEN
POP
POP
GEN
HIC
GEN
POP
HIC
HIC
POP
POP
Setting2
URB
SUB
URB
SUB
SUB
SUB
URB
SUB
RUR
SUB
RUR
URB
RUR
URB
URB
URB
URB
SUB
URB
URB
RUR
URB
RUR
URB
SUB
RUR
SUB
RUR
SUB
SUB
URB
URB
SUB
Land Use3
RES
COM
RES
RES
RES
RES
COM
RES
FOR
RES
AGR
COM
RES
COM
RES
IND
COM
RES
COM
COM
IND
RES
AGR
IND
RES
FOR
RES
FOR
COM
IND
COM
RES
RES
Scale4
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
5
5
4
9
4
5
5
4
5
5
5
4
4
5
4
3
4
4
5
3
4
4
4
4
4
4
Years
n
1
9
5
1
1
3
9
3
9
2
2
8
9
8
9
9
5
3
4
3
7
8
1
8
10
1
10
1
1
4
9
9
7
First
1997
1997
1997
1997
1997
1997
1997
1997
1998
1997
1997
1997
1997
1997
1997
1997
2001
1997
1998
2004
1997
1997
2004
1997
1997
2003
1997
2005
1997
1997
1997
1997
1997
Last
1997
2006
2001
1997
1997
2000
2006
1999
2006
1999
1998
2006
2005
2005
2005
2005
2006
1999
2002
2006
2003
2006
2004
2005
2006
2003
2006
2005
1997
2000
2006
2006
2006
A-21
-------
State
Ml
Ml
Ml
Ml
Ml
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MS
MS
MS
MS
MO
MO
MO
MO
MO
MO
MO
County
Wayne
Wayne
Wayne
Wayne
Wayne
Anoka
Carlton
Dakota
Dakota
Dakota
Dakota
Dakota
Hennepin
Hennepin
Koochiching
Ramsey
Sherburne
Sherburne
Sherburne
Sherburne
Washington
Wright
Harrison
Hinds
Jackson
Lee
Buchanan
Buchanan
Clay
Greene
Greene
Greene
Greene
Monitor ID
261630025
261630027
261630033
261630062
261630092
270031002
270176316
270370020
270370423
270370439
270370441
270370442
270530954
270530957
270711240
271230864
271410003
271410011
271410012
271410013
271630436
271710007
280470007
280490018
280590006
280810004
290210009
290210011
290470025
290770026
290770032
290770037
290770040
Latitude
42.423063
42.292231
42.306674
42.340833
42.296111
45.13768
46.733611
44.76323
44.77553
44.748039
44.7468
44.73857
44.980995
45.021111
48.605278
44.991944
45.420278
45.394444
45.394444
45.369444
44.84737
45.329167
30.446806
32.296806
30.378425
34.263333
39.731389
39.731389
39.183889
37.128333
37.205278
37.11
37.108889
Longitude
-83.426263
-83.106807
-83.148754
-83.0625
-83.116944
-93.20772
-92.418889
-93.03255
-93.06299
-93.043266
-93.02611
-93.00496
-93.273719
-93.281944
-93.402222
-93.183056
-93.871667
-93.8975
-93.885
-93.898056
-92.9954
-93.835833
-89.029139
-90.188306
-88.533985
-88.759722
-94.8775
-94.868333
-94.4975
-93.261667
-93.283333
-93.251944
-93.252778
Objective1
POP
HIC
HIC
POP
HIC
POP
SRC
UNK
UNK
UNK
UNK
UNK
HIC
HIC
UNK
POP
UNK
UNK
UNK
UNK
UNK
UNK
HIC
POP
POP
UNK
GEN
GEN
POP
POP
UNK
POP
SRC
Setting2
SUB
URB
SUB
URB
URB
SUB
RUR
RUR
RUR
RUR
RUR
RUR
URB
URB
URB
SUB
RUR
RUR
URB
RUR
SUB
RUR
SUB
URB
URB
SUB
URB
URB
SUB
SUB
URB
RUR
SUB
Land Use3
COM
IND
IND
RES
RES
RES
AGR
IND
IND
IND
IND
AGR
COM
IND
IND
RES
AGR
IND
MOB
IND
IND
AGR
RES
COM
COM
COM
IND
IND
RES
RES
RES
RES
RES
Scale4
NEI
MID
MID
NEI
MID
URB
NEI
NEI
NEI
NEI
MID
NEI
NEI
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
Height
4
3
5
5
7
4.57
3
3
3.66
4
3
3.5
3
10
10
6
4.88
4
4
4
3
3
4
3
3
4
Years
n
1
3
2
1
1
4
2
8
9
1
7
6
7
6
2
6
1
2
1
2
9
1
8
9
7
1
3
1
5
10
10
10
4
First
1997
1997
1997
1997
1997
2003
2001
1997
1997
1999
2000
2001
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2001
1997
1997
1997
1997
2003
Last
1997
1999
1998
1997
1997
2006
2002
2006
2006
1999
2006
2006
2006
2002
1999
2002
1997
1998
1997
1998
2006
1997
2004
2005
2006
1997
1999
2001
2001
2006
2006
2006
2006
A-22
-------
State
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MT
MT
MT
MT
MT
MT
MT
MT
County
Greene
Iron
Iron
Jackson
Jefferson
Jefferson
Jefferson
Jefferson
Monroe
Pike
Platte
Saint Charles
Saint Charles
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
St. Louis City
St. Louis City
St. Louis City
St. Louis City
Cascade
Cascade
Jefferson
Jefferson
Jefferson
Lewis and Clark
Lewis and Clark
Rosebud
Monitor ID
290770041
290930030
290930031
290950034
290990004
290990014
290990017
290990018
291370001
291630002
291650023
291830010
291831002
291890001
291890004
291890006
291890014
291893001
291895001
291897002
291897003
295100007
295100072
295100080
295100086
300132000
300132001
300430903
300430911
300430913
300490702
300490703
300870700
Latitude
37.108611
37.466389
37.519444
39.104722
38.2633
38.267222
38.252778
38.297694
39.473056
39.3726
39.3
38.579167
38.8725
38.521667
38.5325
38.613611
38.7109
38.641389
38.766111
38.727222
38.720917
38.5425
38.624167
38.682778
38.672222
47.532222
47.53
46.557679
46.548056
46.534722
46.583333
46.593889
45.886944
Longitude
-93.272222
-90.69
-90.7125
-94.570556
-90.3785
-90.379444
-90.393333
-90.384333
-91.789167
-90.9144
-94.7
-90.841111
-90.226389
-90.343611
-90.382778
-90.495833
-90.4759
-90.345833
-90.285833
-90.379444
-90.367028
-90.263611
-90.198611
-90.246667
-90.238889
-111.271111
-111.283611
-111.918098
-111.873333
-111.861389
-1 1 1 .934444
-111.92
-106.628056
Objective1
SRC
SRC
UNK
LINK
POP
OTH
UNK
HIC
UNK
HIC
UNK
UNK
UNK
POP
POP
UNK
HIC
UNK
UNK
POP
POP
HIC
POP
UNK
POP
SRC
SRC
UNK
UNK
UNK
UNK
UNK
UNK
Setting2
SUB
RUR
RUR
URB
RUR
RUR
SUB
SUB
RUR
RUR
SUB
RUR
RUR
SUB
SUB
RUR
RUR
SUB
SUB
SUB
SUB
URB
URB
URB
URB
SUB
SUB
RUR
RUR
RUR
RUR
RUR
SUB
Land Use3
RES
RES
AGR
COM
IND
RES
RES
RES
UNK
RES
MOB
AGR
AGR
RES
RES
RES
RES
COM
COM
RES
RES
RES
COM
RES
RES
AGR
IND
AGR
AGR
AGR
AGR
RES
RES
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
4
2
3
4
5
5
3
3
3
2
3
3
4
3
4
2
4
4
4
14
4
4
3
3.5
4
4
3
3
4
Years
n
4
7
6
9
3
4
2
1
10
1
8
1
3
1
6
8
1
10
8
4
2
10
4
3
7
3
5
4
4
4
4
4
4
First
2003
1997
1997
1997
2004
1997
1999
2002
1997
2006
1997
1997
1997
1997
1999
1997
2006
1997
1997
1997
2002
1997
1997
1997
2000
1997
2001
1997
1997
1997
1997
1997
1998
Last
2006
2003
2003
2006
2006
2000
2000
2002
2006
2006
2004
1997
1999
1997
2004
2004
2006
2006
2004
2000
2003
2006
2000
1999
2006
1999
2005
2000
2000
2000
2000
2000
2001
A-23
-------
State
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
NE
NE
NE
NE
NV
NV
NV
NV
NH
NH
NH
NH
NH
NH
NH
NH
County
Rosebud
Rosebud
Rosebud
Rosebud
Rosebud
Rosebud
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Douglas
Douglas
Douglas
Douglas
Clark
Clark
Clark
Clark
Cheshire
Coos
Coos
Coos
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Monitor ID
300870701
300870702
300870760
300870761
300870762
300870763
301110016
301110066
301110079
301110080
301110082
301110083
301110084
301111065
301112005
301112006
301112007
310550048
310550050
310550053
310550055
320030022
320030078
320030539
320030601
330050007
330070019
330070022
330071007
330110016
330110019
330110020
330111009
Latitude
45.901944
45.863889
45.668056
45.603056
45.648333
45.976667
45.656389
45.788318
45.769439
45.777149
45.783889
45.795278
45.831453
45.801944
45.803889
45.81
45.832778
41.323889
41.332778
41.297778
41.362433
36.390775
35.46505
36.144444
35.978889
42.930556
44.488611
44.458333
44.596667
42.992778
43.000556
43.000556
42.764444
Longitude
-106.637778
-106.557778
-106.518889
-106.464167
-106.556667
-106.660556
-108.765833
-108.459536
-108.574292
-108.47436
-108.515
-108.455833
-108.449964
-108.426111
-108.445556
-108.413056
-108.377778
-95.942778
-95.956389
-95.9375
-95.976112
-114.90681
-114.919615
-115.085556
-114.844167
-72.277778
-71.180278
-71.154167
-71.516667
-71 .459444
-71 .468056
-71 .468056
-71.4675
Objective1
UNK
UNK
SRC
SRC
OTH
UNK
UNK
SRC
POP
UNK
POP
SRC
POP
UNK
UNK
OTH
OTH
HIC
HIC
POP
HIC
REG
REG
POP
POP
UNK
POP
UNK
POP
HIC
UNK
UNK
HIC
Setting2
RUR
RUR
RUR
RUR
RUR
RUR
RUR
RUR
SUB
RUR
URB
SUB
SUB
SUB
SUB
SUB
RUR
URB
URB
URB
SUB
RUR
RUR
SUB
SUB
URB
UNK
RUR
URB
URB
URB
URB
URB
Land Use3
AGR
AGR
FOR
FOR
FOR
IND
AGR
RES
COM
AGR
COM
AGR
RES
RES
IND
AGR
RES
RES
RES
IND
RES
IND
DES
MOB
COM
COM
UNK
IND
IND
COM
COM
COM
COM
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
REG
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
5
5
4
3
4
3.5
4.5
4
3
4
4.5
4
4
3
3
5
6
4
8
3.5
4
3.5
4
4
5
5
5
5
3
Years
n
3
2
5
5
5
1
9
10
2
4
2
3
3
9
9
8
9
1
2
4
2
5
2
8
1
7
5
1
4
1
1
5
3
First
1997
1997
1998
1997
1998
1997
1997
1997
2002
1997
2002
2000
2004
1997
1997
1997
1997
1997
2002
2002
2005
1998
2001
1998
2002
1997
1997
1997
1997
1997
2000
2002
1997
Last
1999
2000
2003
2003
2003
1997
2005
2006
2003
2000
2003
2002
2006
2005
2005
2004
2005
1997
2003
2006
2006
2002
2002
2005
2002
2003
2001
1997
2001
1997
2000
2006
2001
A-24
-------
State
NH
NH
NH
NH
NH
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NM
NM
NM
NM
NM
NM
NM
NM
NM
County
Hillsborough
Merrimack
Merrimack
Merrimack
Merrimack
Rockingham
Rockingham
Rockingham
Sullivan
Atlantic
Bergen
Burlington
Camden
Camden
Cumberland
Essex
Essex
Gloucester
Hudson
Hudson
Middlesex
Morris
Union
Union
Dona Ana
Dona Ana
Eddy
Grant
Grant
Hidalgo
San Juan
San Juan
San Juan
Monitor ID
330111010
330130007
330131003
330131006
330131007
330150009
330150014
330150015
330190003
340010005
340035001
340051001
340070003
340071001
340110007
340130011
340130016
340150002
340170006
340171002
340232003
340273001
340390003
340390004
350130008
350130017
350151004
350170001
350171003
350230005
350450008
350450009
350450017
Latitude
42.701944
43.206944
43.177222
43.132444
43.218491
43.078056
43.075278
43.0825
43.364444
39.53024
40.88237
40.07806
39.92304
39.68425
39.42227
40.726667
40.722222
39.80034
40.67025
40.73169
40.50888
40.78763
40.66245
40.64144
31.930556
31.795833
32.855556
32.759444
32.691944
31.783333
36.735833
36.742222
36.752778
Longitude
-71.445
-71.534167
-71 .4625
-71.45827
-71 .45827
-70.762778
-70.748056
-70.761944
-72.338333
-74.46069
-74.04217
-74.85772
-75.09762
-74.86149
-75.0252
-74.144167
-74.146944
-75.21212
-74.12608
-74.06657
-74.2682
-74.6763
-74.21474
-74.20836
-106.630556
-106.5575
-104.411389
-108.131389
-108.124444
-108.497222
-108.238333
-107.976944
-108.716667
Objective1
UNK
UNK
UNK
OTH
OTH
UNK
POP
POP
UNK
UNK
POP
HIC
POP
GEN
UNK
UNK
POP
UNK
POP
HIC
HIC
UNK
POP
HIC
UNK
SRC
SRC
UNK
SRC
UNK
UNK
SRC
UNK
Setting2
SUB
URB
RUR
SUB
URB
SUB
URB
SUB
URB
RUR
URB
URB
SUB
RUR
RUR
URB
URB
RUR
URB
URB
URB
RUR
URB
SUB
RUR
SUB
URB
SUB
RUR
RUR
RUR
RUR
RUR
Land Use3
IND
COM
RES
RES
COM
COM
RES
COM
RES
RES
COM
COM
RES
COM
IND
IND
IND
AGR
COM
COM
COM
AGR
COM
IND
AGR
COM
COM
IND
IND
UNK
DES
IND
UNK
Scale4
MIC
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
MID
NEI
URB
NEI
NEI
NEI
Height
5
3
3
9
3
2
4
4
4
4
5
4
4
4
5
4
5
4
5
5
5
4
2
4
3
3
Years
n
6
6
7
4
1
3
3
1
5
8
9
9
8
9
9
2
1
9
9
8
9
9
9
9
6
9
9
5
5
5
6
3
1
First
1997
1997
1997
2003
2005
1997
2004
2002
1997
1997
1997
1997
1997
1997
1997
1997
2002
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1999
1997
1997
1997
1997
Last
2002
2003
2003
2006
2005
2000
2006
2002
2001
2005
2005
2005
2005
2005
2005
1998
2002
2005
2005
2005
2005
2005
2005
2005
2002
2005
2005
2001
2005
2001
2002
2005
1997
A-25
-------
State
NM
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
County
San Juan
Albany
Bronx
Bronx
Bronx
Bronx
Chautauqua
Chautauqua
Chautauqua
Chemung
Erie
Erie
Erie
Essex
Franklin
Hamilton
Herkimer
Kings
Kings
Madison
Monroe
Monroe
Monroe
Nassau
New York
New York
Niagara
Onondaga
Putnam
Queens
Queens
Rensselaer
Rensselaer
Monitor ID
350451005
360010012
360050073
360050080
360050083
360050110
360130005
360130006
360130011
360150003
360290005
360294002
360298001
360310003
360330004
360410005
360430005
360470011
360470076
360530006
360551004
360551007
360556001
360590005
360610010
360610056
360632008
360671015
360790005
360810097
360810124
360830004
360831005
Latitude
36.796667
42.68069
40.811389
40.83608
40.86586
40.81616
42.29073
42.49945
42.29073
42.11105
42.87684
42.99549
42.818889
44.39309
44.434309
43.44957
43.68578
40.73277
40.67185
42.73046
43.16545
43.146198
43.161
40.74316
40.739444
40.75917
43.08216
43.05238
41.44151
40.75527
40.7362
42.78187
42.72444
Longitude
-108.4725
-73.75689
-73.91
-73.92021
-73.88075
-73.90207
-79.58958
-79.31888
-79.58658
-76.80249
-78.80988
-78.90157
-78.840833
-73.85892
-74.24601
-74.51625
-74.98538
-73.94722
-73.97824
-75.78443
-77.55479
-77.54813
-77.60357
-73.58549
-73.986111
-73.96651
-79.00099
-76.0592
-73.70762
-73.75861
-73.82317
-73.46361
-73.43166
Objective1
UNK
HIC
UNK
HIC
UNK
OTH
POP
HIC
POP
UNK
POP
HIC
HIC
GEN
GEN
POP
POP
HIC
POP
POP
POP
POP
HIC
UNK
HIC
HIC
POP
POP
UNK
GEN
POP
OTH
GEN
Setting2
UNK
RUR
URB
URB
URB
URB
URB
URB
RUR
URB
URB
SUB
URB
RUR
RUR
RUR
RUR
URB
URB
RUR
SUB
URB
URB
SUB
URB
URB
SUB
SUB
RUR
URB
SUB
RUR
RUR
Land Use3
UNK
AGR
RES
RES
COM
RES
IND
IND
AGR
COM
RES
IND
IND
FOR
COM
COM
FOR
IND
RES
AGR
RES
RES
COM
COM
RES
COM
IND
COM
FOR
RES
RES
FOR
FOR
Scale4
NEI
MID
NEI
REG
NEI
NEI
NEI
NEI
URB
REG
NEI
REG
NEI
NEI
NEI
NEI
MID
NEI
NEI
Height
9
5
13
12
5
4
4
4
4
4
4
4
5
4
13
11
4
12
5
38
10
4
5
12
5
Years
n
7
10
2
3
6
5
4
7
10
9
10
10
2
10
2
10
8
1
2
10
7
2
7
9
2
8
8
10
10
3
5
3
3
First
1997
1997
1997
1997
2001
2000
1997
2000
1997
1998
1997
1997
1997
1997
2005
1997
1997
1998
1997
1997
1997
2005
1997
1997
1997
1997
1999
1997
1997
1999
2002
2002
1998
Last
2005
2006
1998
1999
2006
2006
2000
2006
2006
2006
2006
2006
1998
2006
2006
2006
2006
1998
1999
2006
2003
2006
2003
2006
1999
2006
2006
2006
2006
2001
2006
2004
2000
A-26
-------
State
NY
NY
NY
NY
NY
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
ND
ND
ND
ND
ND
ND
ND
County
Richmond
Schenectady
Suffolk
Suffolk
Ulster
Alexander
Beaufort
Beaufort
Beaufort
Chatham
Cumberland
Davie
Duplin
Edgecombe
Forsyth
Johnston
Lincoln
Martin
Mecklenburg
Mecklenburg
New Hanover
New Hanover
Northampton
Person
Pitt
Swain
Billings
Billings
Burke
Burke
Burleigh
Cass
Cass
Monitor ID
360850067
360930003
361030002
361030009
361111005
370030003
370130003
370130004
370130006
370370004
370511003
370590002
370610002
370650099
370670022
371010002
371090004
371170001
371190034
371190041
371290002
371290006
371310002
371450003
371470099
371730002
380070002
380070111
380130002
380130004
380150003
380171003
380171004
Latitude
40.59733
42.79963
40.74529
40.8275
42.1438
35.903611
35.3575
35.377241
35.377778
35.757222
34.968889
35.809289
34.954823
35.988333
36.110556
35.590833
35.438556
35.81069
35.248611
35.2401
34.364167
34.268403
36.48438
36.306965
35.583333
35.435509
46.8943
47.296667
48.9904
48.64193
46.825425
46.910278
46.933754
Longitude
-74.12619
-73.94019
-73.41919
-73.05694
-74.49414
-81.184167
-76.779722
-76.748997
-76.766944
-79.159722
-78.9625
-80.559115
-77.960781
-77.582778
-80.226667
-78.461944
-81 .27675
-76.89782
-80.766389
-80.785683
-77.838611
-77.956529
-77.61998
-79.09197
-77.598889
-83.443697
-103.37853
-103.095556
-102.7815
-102.4018
-100.76821
-96.795
-96.85535
Objective1
POP
POP
HIC
UNK
POP
GEN
SRC
HIC
SRC
GEN
POP
GEN
GEN
GEN
POP
GEN
GEN
GEN
POP
POP
POP
GEN
SRC
GEN
GEN
GEN
GEN
HIC
SRC
REG
POP
POP
POP
Setting2
SUB
SUB
SUB
SUB
RUR
SUB
RUR
RUR
RUR
RUR
SUB
SUB
URB
RUR
URB
RUR
RUR
RUR
SUB
URB
RUR
RUR
RUR
RUR
RUR
SUB
RUR
RUR
RUR
RUR
SUB
SUB
SUB
Land Use3
RES
RES
IND
RES
COM
COM
IND
FOR
IND
AGR
COM
IND
RES
AGR
RES
AGR
RES
AGR
RES
RES
AGR
IND
COM
AGR
COM
RES
AGR
IND
AGR
AGR
RES
RES
AGR
Scale4
NEI
NEI
NEI
URB
URB
NEI
NEI
NEI
MIC
NEI
NEI
REG
NEI
URB
NEI
URB
NEI
NEI
URB
URB
URB
URB
REG
NEI
REG
NEI
REG
REG
URB
URB
URB
Height
20
5
5
5
3
3
3
4
3
5
5
5
3
3
4
12.2
4
4
4
4
4
3
Years
n
3
10
2
7
9
2
3
2
5
2
2
2
1
2
8
1
2
2
2
6
1
10
2
2
2
2
5
1
6
3
1
1
8
First
1997
1997
1997
2000
1997
1999
1997
1997
2001
1998
1999
1997
1999
1999
1997
1999
1997
1998
1997
2000
2005
1997
1997
1998
1997
1998
2000
1997
2000
2004
2006
1997
1999
Last
1999
2006
1998
2006
2006
2003
1999
1998
2006
2001
2006
2000
1999
2004
2004
1999
2000
2001
1998
2006
2005
2006
2000
2004
2000
2004
2006
1997
2005
2006
2006
1997
2006
A-27
-------
State
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
County
Dunn
McKenzie
McKenzie
McKenzie
McLean
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Morton
Morton
Oliver
Steele
Williams
Williams
Adams
Allen
Ashtabula
Belmont
Butler
Butler
Clark
Clermont
Columbiana
Columbiana
Columbiana
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Monitor ID
380250003
380530002
380530104
380530111
380550113
380570001
380570004
380570102
380570118
380570123
380570124
380590002
380590003
380650002
380910001
381050103
381050105
390010001
390030002
390071001
390133002
390170004
390171004
390230003
390250021
390290016
390290022
390292001
390350038
390350045
390350060
390350065
390356001
Latitude
47.3132
47.5812
47.575278
47.605556
47.606667
47.258853
47.298611
47.325
47.371667
47.385725
47.400619
46.84175
46.873075
47.185833
47.599703
48.408834
48.392644
38.795
40.772222
41.959444
39.968056
39.383333
39.53
39.855556
38.961273
40.634722
40.635
40.620278
41.476944
41.471667
41.493955
41.446389
41.504722
Longitude
-102.5273
-103.2995
-103.968889
-104.017222
-102.036389
-101.783035
-101.766944
-101.765833
-101.780833
-101.862917
-101.92865
-100.870059
-100.905039
-101.428056
-97.899009
-102.90765
-102.910233
-83.535278
-84.051944
-80.5725
-80.7475
-84.544167
-84.3925
-83.9975
-84.09445
-80.546389
-80.546667
-80.580833
-81.681944
-81.657222
-81.678542
-81.661944
-81.623889
Objective1
GEN
GEN
SRC
SRC
POP
POP
POP
SRC
SRC
SRC
SRC
SRC
SRC
SRC
GEN
SRC
SRC
POP
POP
POP
POP
POP
POP
POP
HIC
POP
POP
POP
HIC
POP
POP
HIC
POP
Setting2
RUR
RUR
RUR
RUR
RUR
SUB
RUR
RUR
RUR
RUR
RUR
SUB
SUB
RUR
RUR
RUR
RUR
SUB
UNK
SUB
SUB
SUB
SUB
RUR
URB
SUB
SUB
URB
URB
URB
URB
URB
SUB
Land Use3
AGR
AGR
AGR
IND
AGR
RES
AGR
IND
IND
IND
IND
IND
IND
AGR
AGR
IND
IND
RES
AGR
RES
IND
COM
COM
AGR
RES
RES
COM
COM
IND
IND
COM
RES
COM
Scale4
REG
REG
URB
URB
URB
NEI
URB
URB
URB
URB
URB
NEI
NEI
URB
REG
URB
URB
NEI
URB
URB
NEI
NEI
NEI
NEI
URB
NEI
MIC
NEI
NEI
NEI
NEI
NEI
NEI
Height
4
4
3
3
3
5
4
3
3
4
4
4
4
3
3
4
4
5
6
8
6
7
4
4
5
7
6
20
4
4
4
5
6
Years
n
10
6
9
7
6
1
8
10
10
10
10
8
6
9
2
9
9
10
10
10
7
10
10
10
8
1
5
1
9
10
10
9
6
First
1997
1997
1998
2000
1998
1997
1999
1997
1997
1997
1997
1997
1999
1997
1997
1997
1997
1997
1997
1997
2000
1997
1997
1997
1997
1997
2002
1998
1997
1997
1997
1998
1997
Last
2006
2006
2006
2006
2003
1997
2006
2006
2006
2006
2006
2004
2004
2006
1999
2006
2005
2006
2006
2006
2006
2006
2006
2006
2004
1997
2006
1998
2006
2006
2006
2006
2002
A-28
-------
State
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
County
Franklin
Franklin
Gallia
Hamilton
Hamilton
Jefferson
Jefferson
Jefferson
Lake
Lake
Lawrence
Lorain
Lorain
Lorain
Lucas
Lucas
Mahoning
Mahoning
Meigs
Montgomery
Morgan
Morgan
Scioto
Scioto
Scioto
Stark
Summit
Summit
Tuscarawas
Tuscarawas
Cherokee
Kay
Kay
Monitor ID
390490004
390490034
390530002
390610010
390612003
390810016
390810017
390811001
390850003
390853002
390870006
390930017
390930026
390931003
390950008
390950024
390990009
390990013
391051001
391130025
391150003
391150004
391450013
391450020
391450022
391510016
391530017
391530022
391570003
391570006
400219002
400710602
400719003
Latitude
39.992222
40.0025
38.944167
39.214931
39.228889
40.362778
40.366104
40.321944
41.673056
41.7225
38.520278
41.368056
41.471667
41.365833
41.663333
41.644167
41.098333
41.096111
39.037778
39.758333
39.631667
39.634221
38.754167
38.609048
38.588034
40.827778
41.063333
41.080278
40.516389
40.511416
35.85408
36.705328
36.662778
Longitude
-83.041667
-82.994444
-82.112222
-84.690723
-84.448889
-80.615556
-80.615002
-80.606389
-81 .4225
-81.241944
-82.666667
-82.110556
-82.143611
-82.108333
-83.476667
-83.546667
-80.651944
-80.658611
-82.045556
-84.2
-81 .673056
-81.670038
-82.9175
-82.822911
-82.834973
-81.378611
-81.468611
-81.516389
-81 .476389
-81.639149
-94.985964
-97.087656
-97.074444
Objective1
HIC
POP
POP
POP
HIC
POP
HIC
HIC
UNK
HIC
POP
POP
POP
HIC
HIC
POP
HIC
GEN
POP
HIC
HIC
SRC
HIC
HIC
UPW
HIC
HIC
POP
POP
POP
REG
UNK
POP
Setting2
SUB
URB
SUB
RUR
SUB
URB
URB
URB
SUB
SUB
SUB
URB
SUB
URB
URB
URB
URB
URB
SUB
URB
RUR
RUR
SUB
RUR
RUR
SUB
SUB
URB
URB
RUR
RUR
URB
RUR
Land Use3
COM
COM
RES
IND
IND
COM
COM
IND
RES
COM
RES
COM
IND
COM
IND
IND
COM
RES
RES
COM
AGR
AGR
IND
FOR
IND
RES
IND
COM
IND
RES
RES
RES
RES
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
MID
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
URB
URB
MID
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
Height
5
4
10
5
3
10
3
6
5
16
8
6
5
9
8
8
6
6
4
3
5
4
10
4
4
5
4
3
5
10
3
4
3
Years
n
3
9
5
9
1
4
3
6
10
10
9
3
6
3
7
8
3
7
10
7
9
1
10
2
2
7
10
10
6
3
4
8
2
First
1997
1997
2002
1998
1997
1999
2004
1998
1997
1997
1998
2001
1997
1997
1998
1999
1997
2000
1997
1997
1997
2006
1997
2005
2005
1997
1997
1997
1997
2004
2001
1997
2002
Last
1999
2006
2006
2006
1997
2002
2006
2003
2006
2006
2006
2003
2002
1999
2006
2006
1999
2006
2006
2003
2005
2006
2006
2006
2006
2003
2006
2006
2002
2006
2005
2005
2003
A-29
-------
State
OK
OK
OK
OK
OK
OK
OK
OK
OK
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Kay
Mayes
Muskogee
Oklahoma
Oklahoma
Ottawa
Tulsa
Tulsa
Tulsa
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Beaver
Beaver
Berks
Berks
Blair
Bucks
Cambria
Centre
Dauphin
Delaware
Delaware
Monitor ID
400719010
400979014
401010167
401090025
401091037
401159004
401430175
401430235
401430501
420030002
420030010
420030021
420030031
420030032
420030064
420030067
420030116
420031301
420033003
420033004
420070002
420070004
420070005
420070014
420110009
420110100
420130801
420170012
420210011
420270100
420430401
420450002
420450109
Latitude
36.956222
36.228408
35.793134
35.553056
35.614131
36.922222
36.149877
36.126945
36.16127
40.500556
40.445577
40.413611
40.443333
40.414444
40.323611
40.381944
40.473611
40.4025
40.318056
40.305
40.56252
40.635575
40.684722
40.747796
40.320278
40.335278
40.535278
40.107222
40.309722
40.811389
40.245
39.835556
39.818715
Longitude
-97.03135
-95.249943
-95.302235
-97.623611
-97.475083
-94.838889
-96.011664
-95.998941
-96.015784
-80.071944
-80.016155
-79.941389
-79.990556
-79.942222
-79.868333
-80.185556
-80.077222
-79.860278
-79.881111
-79.888889
-80.503948
-80.230605
-80.359722
-80.316442
-75.926667
-75.922778
-78.370833
-74.882222
-78.915
-77.877028
-76.844722
-75.3725
-75.413973
Objective1
GEN
GEN
SRC
POP
POP
UNK
UNK
SRC
UNK
POP
POP
POP
POP
UNK
POP
GEN
POP
HIC
POP
UNK
REG
HIC
POP
POP
HIC
UNK
POP
POP
HIC
POP
HIC
HIC
UNK
Setting2
RUR
RUR
RUR
SUB
SUB
RUR
SUB
URB
URB
SUB
URB
SUB
URB
SUB
SUB
RUR
SUB
SUB
SUB
SUB
RUR
URB
RUR
URB
SUB
URB
SUB
SUB
URB
RUR
RUR
URB
URB
Land Use3
AGR
AGR
COM
RES
RES
RES
IND
IND
COM
RES
COM
RES
COM
RES
RES
RES
RES
RES
IND
RES
AGR
IND
AGR
RES
RES
COM
IND
RES
COM
AGR
COM
IND
IND
Scale4
NEI
NEI
NEI
URB
URB
NEI
NEI
MID
NEI
URB
NEI
NEI
NEI
NEI
NEI
NEI
REG
NEI
URB
URB
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
3
3
5
4
4
4
4
6
4
6
13
5
8
9
5
9
5
8
3
4
3
4
4
4
6
2
12
3
4
4
Years
n
2
1
9
4
2
3
9
9
6
8
9
7
3
2
10
9
7
4
7
4
10
2
10
10
9
2
10
10
10
3
10
10
3
First
2004
2005
1997
1999
2004
2001
1997
1997
2000
1997
1998
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2004
1997
1997
1997
Last
2005
2005
2005
2002
2005
2004
2005
2005
2005
2006
2006
2006
1999
1998
2006
2006
2005
2000
2005
2000
2006
1998
2006
2006
2005
1998
2006
2006
2006
2006
2006
2006
1999
A-30
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
Rl
County
Erie
Indiana
Lackawanna
Lancaster
Lawrence
Lehigh
Luzerne
Lycoming
Lycoming
Mercer
Montgomery
Northampton
Northampton
Northampton
Perry
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Schuylkill
Warren
Warren
Washington
Washington
Washington
Westmoreland
York
Providence
Monitor ID
420490003
420630004
420692006
420710007
420730015
420770004
420791101
420810100
420810403
420850100
420910013
420950025
420950100
420958000
420990301
421010004
421010022
421010024
421010027
421010029
421010047
421010048
421010055
421010136
421070003
421230003
421230004
421250005
421250200
421255001
421290008
421330008
440070012
Latitude
42.14175
40.56333
41.442778
40.046667
40.995848
40.611944
41.265556
41.2508
41.246111
41.215014
40.112222
40.628056
40.676667
40.692224
40.456944
40.008889
39.916667
40.076389
40.010556
39.957222
39.944722
39.991389
39.922517
39.9275
40.820556
41.857222
41.844722
40.146667
40.170556
40.445278
40.304694
39.965278
41.825556
Longitude
-80.038611
-78.919972
-75.623056
-76.283333
-80.346442
-75.4325
-75.846389
-76.9238
-76.989722
-80.484779
-75.309167
-75.341111
-75.216667
-75.237156
-77.165556
-75.097778
-75.188889
-75.011944
-75.151944
-75.173056
-75.166111
-75.080833
-75.186783
-75.222778
-76.212222
-79.1375
-79.169722
-79.902222
-80.261389
-80.420833
-79.505667
-76.699444
-71 .405278
Objective1
HIC
POP
HIC
HIC
POP
POP
POP
POP
POP
POP
POP
POP
UNK
POP
GEN
POP
HIC
UNK
UNK
POP
POP
UNK
POP
POP
POP
HIC
HIC
POP
POP
REG
POP
HIC
POP
Setting2
SUB
RUR
SUB
SUB
SUB
SUB
SUB
URB
URB
URB
SUB
SUB
SUB
SUB
RUR
URB
URB
SUB
URB
URB
URB
RUR
URB
URB
RUR
SUB
RUR
SUB
SUB
RUR
SUB
SUB
URB
Land Use3
COM
COM
RES
IND
IND
COM
RES
RES
COM
COM
RES
COM
IND
RES
UNK
RES
IND
IND
MOB
COM
RES
RES
RES
RES
RES
RES
FOR
COM
RES
AGR
COM
RES
COM
Scale4
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
NEI
NEI
NEI
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
REG
URB
NEI
NEI
Height
4
3
4
4
4
3
4
3.5
8
3
4
3
3
4
4
7
7
4
5
11
4
5
4
4
4
4
4
2
4
4
4
4
20
Years
n
10
2
10
10
10
10
9
5
4
9
10
9
2
7
10
8
2
2
2
8
2
2
1
7
9
10
10
10
10
10
9
10
10
First
1997
2005
1997
1997
1997
1997
1997
2002
1997
1997
1997
1998
1997
2000
1997
1997
1997
1997
1997
1997
1997
1997
2005
1997
1998
1997
1997
1997
1997
1997
1998
1997
1997
Last
2006
2006
2006
2006
2006
2006
2005
2006
2000
2006
2006
2006
1998
2006
2006
2004
1998
1998
1998
2004
1998
1998
2005
2004
2006
2006
2006
2006
2006
2006
2006
2006
2006
A-31
-------
State
Rl
Rl
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SD
SD
SD
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
County
Providence
Providence
Aiken
Barnwell
Charleston
Charleston
Georgetown
Greenville
Greenville
Lexington
Oconee
Orangeburg
Richland
Richland
Richland
Richland
Custer
Jackson
Minnehaha
Anderson
Blount
Blount
Blount
Bradley
Coffee
Davidson
Hawkins
Humphreys
McMinn
Montgomery
Montgomery
Polk
Polk
Monitor ID
440071005
440071009
450030003
450110001
450190003
450190046
450430006
450450008
450450009
450630008
450730001
450750003
450790007
450790021
450791003
450791006
460330132
460710001
460990007
470010028
470090002
470090006
470090101
470110102
470310004
470370011
470730002
470850020
471070101
471250006
471250106
471390003
471390007
Latitude
41.878333
41.823611
33.342226
33.320344
32.882289
32.941023
33.362014
34.838814
34.899141
34.051017
34.805261
33.29959
34.093959
33.81468
34.024497
33.817902
43.5578
43.74561
43.537626
36.027778
35.775
35.768056
35.63149
35.283164
35.582222
36.205
36.366944
36.051944
35.29733
36.520056
36.504529
35.026111
34.988333
Longitude
-71.378889
-71.411667
-81.788731
-81.465537
-79.977538
-79.657187
-79.294251
-82.402918
-82.31307
-81.15495
-83.2377
-80.442218
-80.962304
-80.781135
-81 .036248
-80.826596
-103.4839
-101.941218
-96.682001
-84.151389
-83.965833
-83.976667
-83.943512
-84.759371
-86.015556
-86.744722
-82.977778
-87.965
-84.75076
-87.394167
-87.396675
-84.384722
-84.371667
Objective1
HIC
HIC
HIC
SRC
POP
SRC
SRC
POP
WEL
SRC
REG
SRC
OTH
GEN
POP
GEN
REG
GEN
POP
UNK
HIC
HIC
GEN
UNK
UNK
POP
UNK
UNK
HIC
UNK
HIC
POP
POP
Setting2
URB
URB
SUB
RUR
URB
RUR
URB
URB
SUB
SUB
RUR
RUR
SUB
RUR
URB
RUR
RUR
RUR
URB
SUB
RUR
SUB
RUR
URB
RUR
URB
RUR
RUR
SUB
RUR
RUR
SUB
URB
Land Use3
RES
COM
RES
FOR
COM
FOR
IND
COM
RES
COM
FOR
FOR
COM
FOR
COM
FOR
FOR
AGR
RES
RES
COM
RES
FOR
RES
AGR
RES
AGR
AGR
AGR
IND
RES
COM
COM
Scale4
NEI
NEI
URB
URB
NEI
REG
NEI
NEI
NEI
NEI
REG
NEI
NEI
URB
MID
MIC
REG
REG
NEI
MID
MID
REG
NEI
NEI
NEI
MID
NEI
NEI
Height
6
3
4.02
3.1
4.3
4
2.13
4
4
3.35
4.3
3.2
3
4.42
4
5
3.35
3
4
3
4
4
10
4
13
1
4
4
3
4
8
1
Years
n
1
10
2
10
10
8
7
9
2
9
9
1
7
4
10
2
2
2
3
8
8
8
1
8
1
10
6
8
8
10
10
9
9
First
1997
1997
1997
1997
1997
1997
1997
1997
2005
1997
1997
2003
1999
2002
1997
1997
2005
2005
2004
1997
1997
1997
1999
1997
1998
1997
1998
1997
1997
1997
1997
1997
1997
Last
1997
2006
1998
2006
2006
2006
2006
2006
2006
2006
2006
2003
2006
2005
2006
1999
2006
2006
2006
2006
2006
2006
1999
2006
1998
2006
2004
2006
2005
2006
2006
2005
2005
A-32
-------
State
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
County
Polk
Polk
Roane
Shelby
Shelby
Shelby
Shelby
Stewart
Sullivan
Sullivan
Sumner
Cameron
Dallas
Ellis
Ellis
Ellis
El Paso
El Paso
El Paso
Galveston
Galveston
Gregg
Harris
Harris
Harris
Harris
Harris
Harris
Harris
Jefferson
Jefferson
Jefferson
Kaufman
Monitor ID
471390008
471390009
471450009
471570034
471570043
471570046
471571034
471610007
471630007
471630009
471651002
480610006
481130069
481390015
481390016
481390017
481410037
481410053
481410058
481670005
481671002
481830001
482010046
482010051
482010059
482010062
482010070
482011035
482011050
482450009
482450011
482450020
482570005
Latitude
34.995833
34.989722
35.947222
35.0434
35.087778
35.272778
35.087222
36.389722
36.534804
36.513971
36.341667
25.892509
32.819952
32.436944
32.482222
32.473611
31.768281
31.758504
31.893928
29.385236
29.398611
32.37871
29.8275
29.623611
29.705833
29.625833
29.735129
29.733713
29.583032
30.036446
29.89403
30.06607
32.564969
Longitude
-84.368333
-84.383889
-84.522222
-90.0136
-90.025278
-89.961389
-90.133611
-87.633333
-82.517078
-82.560968
-86.398333
-97.493824
-96.860082
-97.025
-97.026944
-97.0425
-106.501253
-106.501023
-106.425813
-94.931526
-94.933333
-94.711834
-95.283611
-95.473611
-95.281111
-95.2675
-95.315583
-95.257591
-95.015535
-94.071073
-93.987898
-94.077383
-96.31766
Objective1
UNK
UNK
UNK
HIC
HIC
POP
UNK
OTH
HIC
HIC
OTH
HIC
POP
HIC
GEN
OTH
POP
HIC
POP
HIC
HIC
GEN
POP
SRC
HIC
POP
GEN
POP
HIC
HIC
SRC
SRC
HIC
Setting2
RUR
RUR
SUB
SUB
SUB
SUB
RUR
RUR
SUB
RUR
RUR
URB
URB
SUB
SUB
RUR
URB
URB
URB
URB
SUB
RUR
SUB
SUB
SUB
SUB
SUB
SUB
SUB
SUB
URB
URB
SUB
Land Use3
RES
IND
RES
RES
COM
IND
AGR
AGR
RES
RES
AGR
COM
COM
AGR
AGR
RES
COM
COM
RES
RES
RES
RES
RES
RES
RES
RES
RES
IND
RES
RES
IND
IND
COM
Scale4
NEI
NEI
URB
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
MID
NEI
NEI
NEI
NEI
Height
3
4
4
3
3
3
3
3
3
3
6
4
4
3
5
5
5
4
4
4
6
5
11
6
11
6.31
4
5
5
Years
n
3
3
6
4
2
10
10
7
9
10
7
3
10
9
7
1
9
8
5
2
7
6
8
8
1
9
6
9
5
10
10
8
6
First
1998
1997
1998
2002
1997
1997
1997
1997
1998
1997
1997
1998
1997
1998
1998
2005
1998
1999
2001
2005
1997
2000
1997
1997
1997
1997
2001
1997
2002
1997
1997
1998
2001
Last
2000
2000
2005
2005
1998
2006
2006
2005
2006
2006
2004
2000
2006
2006
2006
2005
2006
2006
2005
2006
2003
2005
2006
2006
1997
2006
2006
2006
2006
2006
2006
2006
2006
A-33
-------
State
TX
TX
TX
UT
UT
UT
UT
UT
UT
VT
VT
VT
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
WA
WA
WA
WA
WA
WA
WA
County
Nueces
Nueces
Nueces
Cache
Davis
Davis
Salt Lake
Salt Lake
Salt Lake
Chittenden
Chittenden
Rutland
Charles
Fairfax
Fairfax
Fairfax
Fairfax
Fairfax
Madison
Roanoke
Rockingham
Rockingham
Alexandria City
Hampton City
Norfolk City
Richmond City
Clallam
Clallam
King
King
Pierce
Pierce
Skagit
Monitor ID
483550025
483550026
483550032
490050004
490110001
490110004
490350012
490351001
490352004
500070003
500070014
500210002
510360002
510590005
510590018
510591004
510591005
510595001
511130003
511611004
511650002
511650003
515100009
516500004
517100023
517600024
530090010
530090012
530330057
530330080
530530021
530530031
530570012
Latitude
27.76534
27.832409
27.804482
41.731111
40.886389
40.902967
40.8075
40.708611
40.736389
44.478889
44.4762
43.608056
37.343294
38.893889
38.7425
38.868056
38.837517
38.931944
38.521944
37.285556
38.389444
38.47732
38.810833
37.003333
36.850278
37.562778
48.113333
48.0975
47.563333
47.568333
47.281111
47.2656
48.493611
Longitude
-97.434272
-97.555381
-97.431553
-111.8375
-1 1 1 .882222
-111.884467
-111.921111
-112.094722
-112.210278
-73.211944
-73.2106
-72.982778
-77.260034
-77.465278
-77.0775
-77.143056
-77.163231
-77.198889
-78.436111
-79.884167
-78.914167
-78.81904
-77.044722
-76.399167
-76.257778
-77.465278
-123.399167
-123.425556
-122.3406
-122.308056
-122.374167
-122.3858
-122.551944
Objective1
POP
HIC
POP
POP
POP
POP
UNK
HIC
HIC
HIC
POP
POP
HIC
POP
UNK
UNK
POP
UNK
UNK
HIC
POP
POP
POP
HIC
POP
HIC
UNK
UNK
HIC
POP
HIC
POP
UNK
Setting2
URB
URB
SUB
URB
SUB
SUB
SUB
SUB
RUR
URB
URB
URB
SUB
RUR
SUB
SUB
SUB
SUB
RUR
SUB
RUR
SUB
URB
SUB
URB
URB
SUB
SUB
SUB
URB
SUB
SUB
SUB
Land Use3
RES
RES
RES
COM
COM
RES
IND
RES
IND
COM
COM
COM
RES
AGR
RES
COM
RES
RES
FOR
RES
AGR
COM
RES
RES
COM
COM
RES
RES
IND
RES
RES
IND
RES
Scale4
NEI
NEI
NEI
NEI
NEI
MID
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
Height
4
6
4
4
3
4
4
6
4
4
5
4
4
11
4
4
7
6
10
4
5
5
4
5
11
5
5
5
5
Years
n
9
8
8
3
6
2
6
9
7
3
1
7
10
9
1
4
4
9
3
10
6
2
10
10
8
8
1
5
2
4
1
1
1
First
1997
1998
1998
2003
1997
2004
1999
1997
1997
1997
2004
1997
1997
1997
1997
1997
2003
1998
2000
1997
1998
2005
1997
1997
1997
1999
1997
1999
1997
2001
1997
1997
1997
Last
2005
2005
2005
2005
2002
2005
2004
2005
2003
1999
2004
2005
2006
2006
1997
2000
2006
2006
2003
2006
2003
2006
2006
2006
2004
2006
1997
2004
1998
2004
1997
1997
1997
A-34
-------
State
WA
WA
WA
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
Wl
Wl
Wl
Wl
County
Skagit
Snohomish
What com
Brooke
Brooke
Cabell
Greenbrier
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Kanawha
Kanawha
Kanawha
Marshall
Monongalia
Monongalia
Monongalia
Ohio
Wayne
Wayne
Wayne
Wayne
Wood
Brown
Dane
FOR
Marathon
Monitor ID
530571003
530610016
530730011
540090005
540090007
540110006
540250001
540290005
540290007
540290008
540290009
540290011
540290014
540290015
540290016
540291004
540390004
540390010
540392002
540511002
540610003
540610004
540610005
540690007
540990002
540990003
540990004
540990005
541071002
550090005
550250041
550410007
550730005
Latitude
48.486111
47.983333
48.750278
40.341023
40.389655
38.424133
37.819444
40.529021
40.460138
40.61572
40.427372
40.394583
40.43552
40.618353
40.411944
40.421539
38.343889
38.3456
38.416944
39.915961
39.649367
39.633056
39.648333
40.12043
38.39186
38.390278
38.380278
38.372222
39.323533
44.516667
43.100833
45.56498
45.028333
Longitude
-122.549444
-122.209722
-122.482778
-80.596635
-80.586235
-82.4259
-80.5125
-80.576067
-80.576567
-80.56
-80.592318
-80.612017
-80.600579
-80.540616
-80.601667
-80.580717
-81.619444
-81.628317
-81 .846389
-80.733858
-79.920867
-79.957222
-79.957778
-80.699265
-82.583923
-82.585833
-82.583889
-82.588889
-81.552367
-87.993889
-89.357222
-88.80859
-89.652222
Objective1
UNK
UNK
UNK
POP
POP
POP
UNK
POP
POP
POP
POP
POP
POP
POP
HIC
HIC
POP
POP
HIC
POP
POP
UNK
UNK
HIC
POP
HIC
HIC
HIC
POP
POP
POP
GEN
HIC
Setting2
RUR
URB
URB
SUB
RUR
SUB
RUR
SUB
RUR
SUB
SUB
SUB
SUB
URB
SUB
SUB
SUB
URB
SUB
SUB
SUB
SUB
SUB
SUB
RUR
RUR
RUR
RUR
SUB
URB
URB
RUR
RUR
Land Use3
IND
COM
IND
IND
RES
COM
AGR
RES
RES
RES
RES
RES
RES
RES
RES
RES
COM
COM
IND
RES
COM
RES
RES
RES
IND
RES
RES
RES
IND
RES
RES
FOR
FOR
Scale4
NEI
NEI
NEI
URB
URB
NEI
NEI
NEI
MID
URB
NEI
NEI
URB
NEI
URB
URB
URB
NEI
NEI
NEI
NEI
NEI
URB
NEI
NEI
REG
MID
Height
3
4
12
4
4
13.6
4
4
5
4
4
3
8
13
4
4
4.6
10.7
8
4
3
3
3
4
11
5
6
5
Years
n
1
1
1
10
10
10
1
10
10
10
10
10
7
10
7
10
2
6
1
10
10
4
9
6
6
8
8
8
10
7
2
2
3
First
1997
1997
1998
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2001
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
2004
1997
Last
1997
1997
1998
2006
2006
2006
1997
2006
2006
2006
2006
2006
2003
2006
2003
2006
1998
2006
1997
2006
2006
2000
2005
2002
2002
2005
2005
2005
2006
2005
1998
2005
1999
A-35
-------
State
Wl
Wl
Wl
Wl
Wl
Wl
Wl
WY
PR
PR
PR
PR
PR
PR
PR
PR
PR
VI
VI
VI
VI
VI
County
Milwaukee
Milwaukee
Milwaukee
Oneida
Sauk
Vilas
Wood
Campbell
Barceloneta
Bayamon
Bayamon
Catano
Catano
Catano
Catano
Guayama
Salinas
St Croix
St Croix
St Croix
St Croix
St Croix
Monitor ID
550790007
550790026
550790041
550850996
551110007
551250001
551410016
560050857
720170003
720210004
720210006
720330004
720330007
720330008
720330009
720570009
721230001
780100006
780100011
780100013
780100014
780100015
Latitude
43.047222
43.061111
43.075278
45.645278
43.435556
46.048056
44.3825
44.277222
18.436111
18.412778
18.416667
18.430556
18.444722
18.440028
18.449964
17.966844
17.963002
17.706944
17.719167
17.7225
17.734444
17.741667
Longitude
-87.920278
-87.9125
-87.884444
-89.4125
-89.680278
-89.653611
-89.819167
-105.375
-66.580556
-66.132778
-66.150833
-66.142222
-66.116111
-66.127076
-66.149043
-66.188014
-66.254749
-64.780556
-64.775
-64.776667
-64.783333
-64.751944
Objective1
POP
POP
HIC
UNK
GEN
GEN
POP
SRC
UNK
HIC
POP
UNK
POP
POP
POP
SRC
SRC
HIC
HIC
POP
POP
SRC
Setting2
URB
URB
URB
URB
RUR
RUR
URB
RUR
RUR
SUB
SUB
SUB
URB
URB
URB
RUR
RUR
RUR
RUR
SUB
RUR
RUR
Land Use3
COM
COM
RES
IND
FOR
FOR
RES
IND
RES
IND
IND
RES
RES
COM
RES
COM
AGR
IND
IND
RES
AGR
AGR
Scale4
NEI
NEI
NEI
REG
REG
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
NEI
Height
7
9
7
6
6
15
7
4
3
3
4
2
4
4
4
4
4
Years
n
4
4
4
9
1
1
2
3
5
6
7
7
1
1
1
4
1
5
5
5
5
4
First
1997
2002
1997
1997
2003
2003
1998
2002
1997
1997
1997
1997
2002
2005
2005
2002
2004
1998
1997
1999
1999
2000
Last
2000
2005
2001
2005
2003
2003
1999
2004
2005
2004
2005
2005
2002
2005
2005
2005
2004
2004
2004
2004
2004
2004
Notes:
1 Objectives are POP=Population Exposure; HIC=Highest Concentration; SRC=Source Oriented; GEN=General/Background; REG=Regional
Transport; OTH=Other; UNK=Unknown; UPW=Upwind Background; WEL=Welfare Related Impacts
2 Settings are R=Rural; U=Urban and Center City; S=Suburban
3 Land Uses are AGR=Agricultural; COM=Commercial; IND=lndustrial; FOR=Forest; RES=Residential; UNK=Unknown; DES=Desert; MOB=Mobile.
4 Scales are NEI=Neighborhood; MID=Middle; URB=URBAN; REG=Regional; MIC=Micro
A-36
-------
Table A. 1-4. Population density, concentration variability, and total SO2 emissions associated with 809 ambient
monitors in the broader SO2 monitoring network.
State
AL
AL
AL
AL
AL
AL
AL
AR
AR
AR
AZ
AZ
AZ
AZ
AZ
AZ
AZ
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
County
Colbert
Jackson
Jefferson
Lawrence
Mobile
Mobile
Montgomery
Pulaski
Pulaski
Union
Gila
Gila
Maricopa
Maricopa
Maricopa
Pima
Pinal
Alameda
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Contra Costa
Imperial
Los Angeles
Los Angeles
Monitor ID
010330044
010710020
010731003
010790003
010970028
010972005
011011002
051190007
051191002
051390006
040070009
040071001
040130019
040133002
040133003
040191011
040212001
060010010
060130002
060130006
060130010
060131001
060131002
060131003
060131004
060132001
060133001
060250005
060371002
060371103
Population Residing within:
5km
2195
1902
76802
3952
5966
3017
45389
67784
45800
21877
7801
1359
197458
144581
91955
111215
4375
236320
136288
119088
29809
53051
4033
146336
125350
34743
64019
27033
167653
378843
10km
7954
8137
1 96682
28674
7758
18106
1 56786
178348
109372
29073
14076
1359
613618
490123
340325
354473
7679
532827
303088
231479
123220
181259
39708
256417
233220
1 55226
152758
31895
827729
1618324
15km
25394
19317
344386
73092
17087
52682
213606
270266
230200
32652
17280
3098
1036233
980730
829051
561487
9577
841443
445297
471471
403137
321500
117118
420619
433669
433934
303597
56234
2001363
3027507
20km
62838
29686
489181
91057
39111
111608
259730
334649
310362
36340
17633
5401
1447648
1612687
1518806
639921
10125
1 342267
598861
968983
685185
610171
173196
856435
876585
807706
478310
84405
3286038
4530714
Analysis Bins
Population1
low
low
hi
low
low
low
mod
hi
mod
mod
low
low
hi
hi
hi
hi
low
hi
hi
hi
mod
hi
low
hi
hi
mod
hi
mod
hi
hi
cov2
c
c
b
b
c
b
a
a
a
b
b
c
a
a
a
a
c
a
b
b
a
b
a
a
a
a
b
b
a
a
GSD3
a
b
b
b
c
a
a
a
a
a
b
c
b
a
a
a
a
a
a
a
a
a
a
a
a
a
a
c
a
a
Emissions
(tpy)4
50041
45357
6478
8937
66130
1187
3650
20
20
2527
18438
186
185
180
3119
369
15056
5032
17834
19592
79
5032
5032
17834
8105
7
51
551
A-37
-------
State
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
County
Los Angeles
Los Angeles
Los Angeles
Orange
Riverside
Sacramento
Sacramento
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Bernardino
San Diego
San Diego
San Diego
San Francisco
San Luis Obispo
San Luis Obispo
San Luis Obispo
San Luis Obispo
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Santa Barbara
Monitor ID
060374002
060375001
060375005
060591003
060658001
060670002
060670006
060710012
060710014
060710306
060711234
060712002
060714001
060730001
060731007
060732007
060750005
060791005
060792001
060792004
060794002
060830008
060831012
060831013
060831015
060831016
060831019
060831020
060831025
060831026
060831027
060832004
060832011
Population Residing within:
5km
240505
276378
94836
200253
78757
92433
132584
6720
58937
59772
0
89732
40799
168237
169117
9376
433367
4725
39236
2135
0
655
0
6617
0
0
0
39222
655
655
655
38688
55496
10km
913176
890302
652173
744882
360234
328190
472019
17620
114149
114149
0
314392
114888
528890
616102
15849
827164
56677
55657
34056
51508
1678
0
41576
0
0
0
71015
1678
1678
1678
49356
105491
15km
1850549
2071144
1628468
1303743
734267
645533
866437
29756
193928
1 93046
1911
650533
174610
866015
1097387
218480
1227784
85064
61709
113260
95245
17486
960
59590
2391
4034
4689
117832
11216
15659
13618
58271
170865
20km
3218392
3561110
2848126
1829713
1141466
916197
1180898
69717
224008
224008
1911
1142460
219525
1177835
1449106
452120
1729715
152491
121393
162669
141786
67965
3201
89777
17826
17826
17826
170206
56132
63963
62298
59279
181894
Analysis Bins
Population1
hi
hi
hi
hi
hi
hi
hi
low
hi
hi
low
hi
mod
hi
hi
low
hi
low
mod
low
low
low
low
low
low
low
low
mod
low
low
low
mod
hi
cov2
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
c
a
b
b
a
a
a
a
a
a
a
a
a
b
a
a
GSD3
b
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
b
a
c
b
a
a
a
a
a
a
a
a
a
a
a
a
Emissions
(tpy)4
5869
6282
2304
68
299
5
58
8
251
251
290
203
32
21
34
21
399
3755
3755
3755
3755
118
1109
1109
18
18
18
118
118
118
118
1109
118
A-38
-------
State
CA
CA
CA
CA
CA
CO
CO
CO
CO
CO
CO
CO
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
CT
DC
DE
DE
DE
DE
DE
DE
County
Santa Barbara
Santa Cruz
Solano
Solano
Ventura
Adams
Adams
Denver
El Paso
El Paso
El Paso
El Paso
Fairfield
Fairfield
Fairfield
Fairfield
Fairfield
Hartford
Hartford
Hartford
New Haven
New Haven
New Haven
New Haven
New London
Tolland
District of Columbia
New Castle
New Castle
New Castle
New Castle
New Castle
New Castle
Monitor ID
060834003
060870003
060950001
060950004
061113001
080010007
080013001
080310002
080416001
080416004
080416011
080416018
090010012
090010017
090011123
090012124
090019003
090031005
090031018
090032006
090090027
090091003
090091123
090092123
090110007
090130003
110010041
100031003
100031007
100031008
100031013
100032002
100032004
Population Residing within:
5km
0
0
27872
102003
47248
45071
81896
189782
0
84979
97849
93065
164887
30184
72689
121109
28181
33414
152497
91965
140329
156879
154781
104191
58457
23441
216129
68790
14297
5386
79498
111236
111609
10km
0
6016
130319
166693
227525
360261
334611
574752
24520
242841
288563
266008
291072
188214
126452
209567
151905
147625
329646
333744
290735
293853
292598
189838
97870
47285
813665
223079
67478
80025
221315
245832
245173
15km
8430
51831
359105
247861
401656
903964
784343
1158644
54194
368203
407401
388801
393358
330125
191805
343909
313449
319902
523045
510929
389117
414381
417546
276310
141173
78649
1461563
369450
178295
192989
386624
400217
411000
20km
51692
124792
620107
374613
427503
1344766
1205604
1608099
111518
430076
448545
438812
528453
672435
277225
476656
546288
484462
693079
671515
529118
552021
557442
447334
182476
115317
2029936
603736
274942
391157
618604
624587
600168
Analysis Bins
Population1
low
low
mod
hi
mod
mod
hi
hi
low
hi
hi
hi
hi
mod
hi
hi
mod
mod
hi
hi
hi
hi
hi
hi
hi
mod
hi
hi
mod
low
hi
hi
hi
cov2
a
a
a
a
a
b
b
b
b
a
b
a
b
b
a
b
b
a
a
a
b
b
b
a
a
a
a
b
b
b
b
b
b
GSD3
a
a
a
a
b
b
b
b
b
a
b
a
b
b
b
b
b
b
b
b
b
b
b
b
a
a
a
b
b
b
b
b
b
Emissions
(tpy)4
1109
722
17821
17763
19
24028
23817
26354
5010
8547
8547
8537
4671
757
766
5039
1268
113
83
4761
5085
5085
430
3898
18325
33133
34382
39757
33133
28868
59518
A-39
-------
State
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
GA
GA
County
Broward
Duval
Duval
Duval
Duval
Escambia
Escambia
Hamilton
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Manatee
Miami-Dade
Nassau
Nassau
Orange
Palm Beach
Pinellas
Pinellas
Pinellas
Pinellas
Polk
Polk
Putnam
Sarasota
Sarasota
Sarasota
Baldwin
Bartow
Monitor ID
120110010
120310032
120310080
120310081
120310097
120330004
120330022
120470015
120570021
120570053
120570081
120570095
120570109
120571035
120574004
120813002
120860019
120890005
120890009
120952002
120993004
121030023
121033002
121035002
121035003
121050010
121052006
121071008
121151002
121151005
121151006
130090001
130150002
Population Residing within:
5km
173204
81831
70468
23862
59980
43464
32534
582
90125
54303
5101
28554
11493
63839
32134
2043
54755
17963
8627
85060
54596
40222
74280
58164
48341
1499
8128
10853
78620
28895
65360
7410
1628
10km
475485
270954
288474
152805
225163
133022
122295
1733
287073
140247
24672
1 92630
81649
244436
66598
18810
283528
21386
18803
389159
222249
180398
310490
1 84586
1 74960
21899
49090
21601
180672
140026
188269
22059
15879
15km
953527
439516
506828
305323
418997
233520
223566
6459
539627
307460
48751
493886
287436
463185
149341
82190
685044
38521
27645
808816
446441
488170
633807
401002
304905
60024
125120
35511
237782
244918
295631
44230
50084
20km
1459284
620929
704506
463770
600591
303319
291695
12479
762352
668911
228142
719140
509661
764479
346648
281383
1386189
48316
59574
1031221
718156
901428
907997
655181
492683
142707
198136
44711
332704
356779
386824
50761
91503
Analysis Bins
Population1
hi
hi
hi
mod
hi
mod
mod
low
hi
hi
low
mod
mod
hi
mod
low
hi
mod
low
hi
hi
mod
hi
hi
mod
low
low
mod
hi
mod
hi
low
low
cov2
c
b
b
c
b
b
c
b
c
b
b
c
c
b
b
b
a
c
b
b
b
b
b
b
b
b
b
b
b
b
b
c
c
GSD3
a
b
b
b
b
b
b
a
b
b
b
b
b
b
a
b
a
c
b
a
a
c
c
b
b
a
b
b
b
a
a
b
a
Emissions
(tpy)4
19178
38010
38015
38001
38010
43573
43573
2264
89751
89830
122051
65362
65352
89751
8617
365
235
5050
5050
46
235
24819
24813
30797
30797
21475
21989
29894
143
73950
162418
A-40
-------
State
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
HI
HI
HI
HI
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
County
Bibb
Chatham
Chatham
Chatham
Dougherty
Fannin
Floyd
Fulton
Fulton
Glynn
Muscogee
Richmond
Honolulu
Honolulu
Honolulu
Honolulu
Cerro Gordo
Clinton
Clinton
Clinton
Lee
Lee
Linn
Linn
Linn
Linn
Linn
Linn
Linn
Muscatine
Muscatine
Muscatine
Scott
Monitor ID
130210012
130510019
130510021
130511002
130950006
131110091
131150003
131210048
131210055
131270006
132150008
132450003
150030010
150030011
150031001
150031006
190330018
190450018
190450019
190450020
191110006
191111007
191130028
191130029
191130031
191130032
191130034
191130038
191130039
191390016
191390017
191390020
191630015
Population Residing within:
5km
5430
24119
47852
40337
28572
3943
2671
139962
103612
22992
63822
30694
24951
16119
197479
16676
21247
24561
24561
25544
11675
1202
9112
72325
76896
66674
63548
30007
30134
20360
11109
20360
90863
10km
38736
107149
121273
113925
73138
9432
22348
429736
409533
38643
167389
124609
89592
58440
344436
66976
30341
37638
37638
36227
18308
11474
77687
146914
148919
131315
146044
1 08042
106631
27101
27101
27101
201277
15km
102539
188444
183343
186077
101552
19045
46960
806001
779857
61789
234866
206847
181585
160177
483321
180191
39284
42404
42404
41370
24246
20995
143283
168250
170320
169310
170320
163636
160903
31886
31696
31886
268535
20km
153254
220328
220814
222588
117779
24026
74655
1253530
1209013
67649
254253
298992
344307
277456
672198
300444
45105
45947
45947
48214
25010
34036
189856
179312
179312
183904
185547
180807
180968
40248
36604
40290
293627
Analysis Bins
Population1
low
mod
mod
mod
mod
low
low
hi
hi
mod
hi
mod
mod
mod
hi
mod
mod
mod
mod
mod
mod
low
low
hi
hi
hi
hi
mod
mod
mod
mod
mod
hi
cov2
b
b
b
c
b
c
c
b
b
b
b
b
a
a
b
b
c
b
b
b
b
b
b
b
c
b
c
c
b
c
b
c
b
GSD3
a
b
b
b
a
b
b
b
b
a
a
a
a
a
a
a
c
b
c
b
c
c
a
b
b
a
b
c
a
c
b
c
c
Emissions
(tpy)4
2694
19069
19069
19069
6773
1900
32455
30375
30375
2464
6960
20025
15617
15617
3130
15617
10737
9388
9388
9388
29
208
15400
15400
15400
15400
15400
15400
15400
31137
31054
31054
9415
A-41
-------
State
IA
IA
IA
IA
IA
ID
ID
ID
ID
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
County
Scott
Van Buren
Van Buren
Van Buren
Wood bury
Bannock
Caribou
Caribou
Power
Adams
Champaign
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
Cook
DuPage
La Salle
Macon
Macoupin
Madison
Madison
Madison
Madison
Madison
Peoria
Randolph
Monitor ID
191630017
191770004
191770005
191770006
191930018
160050004
160290003
160290031
160770011
170010006
170190004
170310050
170310059
170310063
170310064
170310076
170311018
170311601
170312001
170314002
170314201
170318003
170436001
170990007
171150013
171170002
171190008
171190017
171191010
171193007
171193009
171430024
171570001
Population Residing within:
5km
3486
0
994
994
4449
16523
0
0
7702
40173
91239
162765
67237
307232
299183
289574
113572
23495
138992
406933
63731
111959
83416
4862
54806
0
36580
37113
9382
32393
27788
76341
5095
10km
43003
2252
2252
2252
44815
57823
1351
604
50773
49711
126127
649556
496359
1205813
965573
1034471
617444
167647
604707
1482581
232428
456791
401929
26956
92426
5005
84254
201161
70816
71861
69631
167513
10038
15km
159186
3764
3764
3764
92956
64147
3211
3211
64147
54168
1 34689
1310508
1055079
2476802
1758392
2000564
1657665
466741
1380464
2777797
627873
1004517
787802
37974
1 03292
16518
152472
536687
176153
172196
136629
232727
16360
20km
245960
7809
6984
6984
112802
69313
4218
3211
69313
64300
152309
1997666
1759830
3318024
2786664
2971446
3102521
1000711
2117578
3752141
1254146
1682955
1266818
63052
112667
19043
330907
950679
323143
353090
273179
269180
29336
Analysis Bins
Population1
low
low
low
low
low
mod
low
low
low
mod
hi
hi
hi
hi
hi
hi
hi
mod
hi
hi
hi
hi
hi
low
hi
low
mod
mod
low
mod
mod
hi
low
cov2
c
a
b
a
b
b
c
c
b
b
b
b
b
b
b
a
b
b
b
b
b
b
b
c
b
b
b
b
b
b
b
b
c
GSD3
a
b
b
b
b
c
b
c
a
b
b
b
b
b
b
b
b
b
b
b
b
b
b
c
b
a
b
b
b
b
b
c
b
Emissions
(tpy)4
14841
36833
1609
12572
12572
1609
3859
362
42308
36403
23944
50763
33488
24023
45681
39578
24553
659
30075
35837
3561
13757
67657
35077
26719
72660
72512
73334
26296
A-42
-------
State
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
County
Rock Island
Saint Clair
Saint Clair
Saint Clair
Sangamon
Tazewell
Wabash
Wabash
Will
Daviess
Dearborn
Floyd
Floyd
Floyd
Fountain
Gibson
Gibson
Hendricks
Hendricks
Hendricks
Jasper
Jasper
Jefferson
Lake
Lake
LaPorte
LaPorte
Marion
Marion
Marion
Marion
Marion
Morgan
Monitor ID
171610003
171630010
171631010
171631011
171670006
171790004
171850001
171851001
171970013
180270002
180290004
180430004
180430007
180431004
180450001
180510001
180510002
180630001
180630002
180630003
180730002
180730003
180770004
180890022
180892008
180910005
180910007
180970042
180970054
180970057
180970072
180970073
181091001
Population Residing within:
5km
87160
48405
49630
1148
41165
32800
8738
1069
12237
905
11932
17205
65510
45432
788
792
6276
4657
7481
1776
991
1688
11228
40318
97669
28928
29106
19283
53595
79478
115856
100599
4178
10km
228445
274406
269778
9915
123641
50160
9493
10899
66320
9377
21347
86512
228353
169258
2536
10900
9493
29661
31567
11450
8080
4551
22061
152401
293157
42982
54698
109791
301941
349455
380088
357454
26279
15km
275180
621019
593969
18231
1 54447
99136
13312
11617
171777
21937
69595
201325
408246
351938
9505
18174
16779
66108
79685
41400
16959
12127
32050
292371
745205
60818
82651
306701
612446
640054
684608
585925
53331
20km
296786
999843
973751
27769
171401
194767
27993
21643
249868
32380
151228
363262
607160
532952
19361
30700
29981
183728
205437
79693
28865
20725
36387
500754
1339901
97304
112181
564512
863127
909257
922620
880596
105208
Analysis Bins
Population1
hi
mod
mod
low
mod
mod
low
low
mod
low
mod
mod
hi
mod
low
low
low
low
low
low
low
low
mod
mod
hi
mod
mod
mod
hi
hi
hi
hi
low
cov2
b
b
b
c
c
c
c
c
b
b
b
b
b
c
c
c
c
c
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
GSD3
a
b
b
b
b
b
b
b
b
c
b
c
b
b
b
b
c
b
b
b
a
b
b
b
b
b
b
b
b
b
b
b
c
Emissions
(tpy)4
9449
13346
13346
26296
10849
73270
127357
127357
46347
65217
151052
52000
67211
66977
55655
127357
127357
147
27494
27494
38198
50716
36590
12499
9198
51880
51077
51077
51096
50949
18019
A-43
-------
State
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
KS
KS
KS
KS
KS
KS
KS
KS
KS
KY
KY
KY
KY
KY
KY
KY
County
Perry
Perry
Pike
Porter
Porter
Porter
Spencer
Spencer
Sullivan
Vanderburgh
Vanderburgh
Vigo
Vigo
Warrick
Warrick
Wayne
Wayne
Linn
Montgomery
Pawnee
Sedgwick
Sumner
Trego
Wyandotte
Wyandotte
Wyandotte
Boyd
Boyd
Boyd
Campbell
Campbell
Daviess
Fayette
Monitor ID
181230006
181230007
181250005
181270011
181270017
181270023
181470002
181470010
181530004
181630012
181631002
181670018
181671014
181730002
181731001
181770006
181770007
201070002
201250006
201450001
201730010
201910002
201950001
202090001
202090020
202090021
210190015
210190017
210191003
210370003
210371001
210590005
210670012
Population Residing within:
5km
6348
6153
3991
12202
14162
13645
1935
2483
1735
45373
1289
50963
25046
2200
11943
34483
31811
1728
9331
5329
102842
1476
0
63756
41751
61336
31077
34804
14960
67933
153388
25889
92980
10km
13158
15700
7372
44110
59080
79678
4701
5934
8313
141869
30177
82314
72089
27584
28798
51601
48948
3741
14142
6038
276624
13125
0
288005
237368
271585
78140
79205
58723
285451
421973
70609
195446
15km
20298
19228
12598
101993
118122
136098
13255
14936
15494
184521
123286
98561
1 00022
60538
80348
59606
59606
4705
17807
6038
380868
56924
578
588511
491118
571758
124766
119732
117154
616440
754366
81162
267016
20km
30372
29270
29314
210946
223900
256849
32146
32405
25746
225094
201383
115726
118986
123354
155370
71062
72278
6412
21677
6038
426333
120034
578
868652
742170
840225
179511
161810
181371
910551
1016145
92902
309266
Analysis Bins
Population1
low
low
low
mod
mod
mod
low
low
low
mod
low
hi
mod
low
mod
mod
mod
low
low
low
hi
low
low
hi
mod
hi
mod
mod
mod
hi
hi
mod
hi
cov2
c
b
b
b
b
b
b
b
a
b
b
b
b
b
b
b
b
b
b
a
a
b
a
b
b
b
b
b
b
b
b
b
a
GSD3
c
c
a
b
b
b
b
b
a
b
b
b
c
b
b
c
c
a
b
a
a
a
a
a
a
a
b
b
b
b
b
b
b
Emissions
(tpy)4
56262
56262
65217
39173
29995
29975
109391
60394
27810
9032
9032
65055
65055
109088
109088
12892
12892
1873
806
806
19433
19433
19427
11909
11933
10172
74986
5111
60963
626
A-44
-------
State
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
LA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
County
Greenup
Hancock
Henderson
Henderson
Jefferson
Jefferson
Jefferson
Livingston
McCracken
McCracken
McCracken
Warren
Bossier
Calcasieu
East Baton Rouge
Ouachita
St. Bernard
West Baton Rouge
Bristol
Essex
Essex
Essex
Essex
Hampden
Hampden
Hampshire
Middlesex
Middlesex
Suffolk
Suffolk
Suffolk
Suffolk
Suffolk
Monitor ID
210890007
210910012
211010013
211010014
211110032
211110051
211111041
211390004
211450001
211451024
211451026
212270008
220150008
220190008
220330009
220730004
220870002
221210001
250051004
250090005
250091004
250091005
250095004
250130016
250131009
250154002
250171701
250174003
250250002
250250019
250250020
250250021
250250040
Population Residing within:
5km
19411
3345
21591
2594
49825
13446
81560
1695
1336
17904
9706
1865
43077
12932
76518
24260
97021
21249
89767
125952
123377
109921
57974
136483
127283
5182
109401
164954
486825
6913
320320
243006
261273
10km
45899
4280
35051
30452
208276
52332
281755
8508
15733
48907
42285
8137
149478
68406
193981
87999
407863
137455
169077
225058
309194
314258
128881
296109
278577
23547
512228
629764
1141656
437626
899106
887256
962956
15km
85066
20931
126144
135741
375535
121743
485759
15337
28279
63098
62036
23083
247738
137949
321486
116037
672107
239718
221707
376322
545716
523212
316108
450050
447646
50329
1210094
1334022
1582622
1118549
1461574
1488386
1475999
20km
109294
39607
202537
194289
586335
257453
676730
31298
64951
83436
82624
68407
295731
154942
408305
131643
856519
366741
372963
598605
906225
870238
422519
532663
541476
123102
1773702
1860034
1955479
1681211
1895175
1966520
1921168
Analysis Bins
Population1
mod
low
mod
low
mod
mod
hi
low
low
mod
low
low
mod
mod
hi
mod
hi
mod
hi
hi
hi
hi
hi
hi
hi
low
hi
hi
hi
low
hi
hi
hi
cov2
b
b
b
b
b
b
b
b
b
b
b
a
a
b
c
a
b
b
b
b
a
b
a
a
b
a
b
b
a
a
a
a
b
GSD3
b
b
b
b
b
b
b
b
b
a
b
a
a
b
b
a
b
b
b
b
b
b
a
b
b
b
b
b
b
a
a
a
a
Emissions
(tpy)4
4806
109458
9026
109476
86910
39110
68947
1775
61380
1760
1760
52
153
53630
39378
2166
7543
31242
44817
1626
20202
20170
1235
7360
2065
859
7670
7254
7999
7791
8024
7921
7952
A-45
-------
State
MA
MA
MA
MA
MD
MD
MD
MD
MD
ME
ME
ME
ME
ME
ME
ME
ME
ME
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
Ml
County
Suffolk
Suffolk
Worcester
Worcester
Allegany
Anne Arundel
Baltimore
Baltimore (City)
Baltimore (City)
Androscoggin
Aroostook
Aroostook
Aroostook
Aroostook
Aroostook
Cumberland
Cumberland
Oxford
Delta
Genesee
Genesee
Kent
Macomb
Missaukee
St. Clair
Schoolcraft
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Monitor ID
250250042
250251003
250270020
250270023
240010006
240032002
240053001
245100018
245100036
230010011
230030009
230030012
230031003
230031013
230031018
230050014
230050027
230172007
260410902
260490021
260492001
260810020
260991003
261130001
261470005
261530001
261630001
261630005
261630015
261630016
261630019
261630025
261630027
Population Residing within:
5km
441455
260061
155688
151851
28416
40618
99648
360916
105632
46561
3403
3403
3403
6476
2387
65123
67865
5903
7503
94710
4058
122533
116002
0
32599
0
151437
86804
98193
203577
210099
81534
79205
10km
1048879
829040
248143
252264
49750
134276
383980
823207
490543
61938
4534
4534
4534
6476
8245
122951
124508
10118
26225
227235
17555
294283
549258
2308
64545
0
338726
350207
423093
654802
695836
280589
384693
15km
1536036
1436251
316330
318317
66814
372761
785111
1195508
1004531
83767
6561
6561
6561
10298
15656
151066
153138
12717
28725
323367
47495
453477
1171414
7840
82832
0
682793
804947
975303
1283000
1189529
668415
915619
20km
1941989
1951612
404489
403312
79171
829885
1155009
1472306
1351499
101615
9030
9030
9030
11213
21187
187005
190157
17231
31746
388490
126929
553989
1 769656
14456
98014
1389
1135095
1386398
1647773
1934280
1756001
1150319
1 574294
Analysis Bins
Population1
hi
hi
hi
hi
mod
mod
hi
hi
hi
mod
low
low
low
low
low
hi
hi
low
low
hi
low
hi
hi
low
mod
low
hi
hi
hi
hi
hi
hi
hi
cov2
a
b
a
a
a
b
b
a
a
b
b
b
c
b
b
b
b
a
a
b
b
a
b
a
b
b
b
b
b
b
b
b
b
GSD3
a
b
a
a
a
b
b
a
b
b
b
b
b
b
b
b
b
a
a
b
b
a
b
a
c
b
b
b
c
b
b
b
c
Emissions
(tpy)4
7987
22045
690
690
1363
64947
97428
65129
97338
283
90
90
90
48
772
3201
3201
499
4222
166
127
541
718
58
1572
64065
64412
34236
34225
31238
81
64407
A-46
-------
State
Ml
Ml
Ml
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MN
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Wayne
Wayne
Wayne
Anoka
Carlton
Dakota
Dakota
Dakota
Dakota
Dakota
Hennepin
Hennepin
Koochiching
Ramsey
Sherburne
Sherburne
Sherburne
Sherburne
Washington
Wright
Buchanan
Buchanan
Clay
Greene
Greene
Greene
Greene
Greene
Iron
Iron
Jackson
Jefferson
Jefferson
Monitor ID
261630033
261630062
261630092
270031002
270176316
270370020
270370423
270370439
270370441
270370442
270530954
270530957
270711240
271230864
271410003
271410011
271410012
271410013
271630436
271710007
290210009
290210011
290470025
290770026
290770032
290770037
290770040
290770041
290930030
290930031
290950034
290990004
290990014
Population Residing within:
5km
150194
123532
96517
57502
9236
3854
8572
1487
1487
2705
224357
157024
6444
112909
5629
5629
5629
0
46665
5377
23253
28224
40627
41036
96594
21784
18988
24455
1121
0
84236
15049
11967
10km
544634
491879
429048
226660
17582
64533
101147
55052
36683
24656
608888
542309
8075
599029
7667
9806
9806
10957
149177
28368
72613
75073
163217
146752
180831
110681
109888
120781
1121
3799
310816
33379
35082
15km
1115397
1104610
973432
496686
28511
221239
265053
218201
191183
153752
1082178
1022041
8923
1052764
35016
29985
29774
33889
354337
39511
87121
86317
366686
224445
208384
210953
210953
213312
4507
6585
605775
64516
61963
20km
1730610
1 743263
1618949
903982
56009
432974
574966
411081
384938
332905
1517123
1489863
10210
1510602
50427
51661
50884
58410
679510
77671
93365
93365
617013
256158
244406
254437
254437
256766
8447
8436
921037
124301
125932
Analysis Bins
Population1
hi
hi
hi
hi
low
low
low
low
low
low
hi
hi
low
hi
low
low
low
low
mod
low
mod
mod
mod
mod
hi
mod
mod
mod
low
low
hi
mod
mod
cov2
b
b
b
b
b
b
b
b
b
b
b
b
c
b
a
b
a
a
b
a
c
b
b
c
a
c
b
a
c
c
c
c
c
GSD3
b
b
b
a
a
b
a
a
a
a
b
a
a
a
a
a
a
a
b
a
b
b
a
b
a
b
a
a
c
b
b
c
b
Emissions
(tpy)4
34236
34225
64407
13324
362
9155
13685
8949
8639
5567
21921
18443
67
20773
26742
26742
26742
26742
11441
26794
3563
3563
25233
9206
9206
9206
9206
9206
43340
43340
19433
55725
55725
A-47
-------
State
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MS
MS
MS
MS
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
County
Jefferson
Jefferson
Monroe
Pike
Platte
Saint Charles
Saint Charles
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
Saint Louis
St. Louis City
St. Louis City
St. Louis City
St. Louis City
Harrison
Hinds
Jackson
Lee
Cascade
Cascade
Jefferson
Jefferson
Jefferson
Lewis and Clark
Lewis and Clark
Rosebud
Rosebud
Rosebud
Monitor ID
290990017
290990018
291370001
291630002
291650023
291830010
291831002
291890001
291890004
291890006
291890014
291893001
291895001
291897002
291897003
295100007
295100072
295100080
295100086
280470007
280490018
280590006
280810004
300132000
300132001
300430903
300430911
300430913
300490702
300490703
300870700
300870701
300870702
Population Residing within:
5km
19711
12258
0
645
2159
2637
4587
95190
61422
68741
48016
117492
108578
82790
88786
107568
101305
154740
145966
18607
54986
39463
24421
40281
42971
1767
0
0
10126
7706
2353
2353
0
10km
36471
41709
1439
2077
36438
6349
95765
327257
315539
235858
223506
487564
358731
336688
383007
375790
393971
463092
473923
88520
171385
49647
44442
64778
64778
25076
11616
6845
38881
31421
2353
2353
2353
15km
60199
79196
2093
6916
113990
34541
273147
630767
647834
488837
550275
929037
617042
729925
764342
678820
726063
861774
857733
139495
273630
65034
61390
68296
70181
47509
36425
27041
49340
48723
2353
3131
2353
20km
116882
170110
5612
11249
238276
90953
431484
966432
1 020228
927852
1005593
1305061
941386
1170973
1192267
979578
1097105
1168442
1177204
181694
332464
75787
74867
70181
70181
49340
49340
47509
49340
49340
3131
3131
3131
Analysis Bins
Population1
mod
mod
low
low
low
low
low
hi
hi
hi
mod
hi
hi
hi
hi
hi
hi
hi
hi
mod
hi
mod
mod
mod
mod
low
low
low
mod
low
low
low
low
cov2
c
c
a
b
a
b
b
b
b
b
b
b
b
b
b
b
b
b
b
c
b
b
a
b
c
b
c
c
c
b
b
b
b
GSD3
b
b
a
b
a
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
a
b
a
b
b
b
b
b
c
c
a
a
a
Emissions
(tpy)4
55725
32468
13495
11030
47610
67735
24466
22816
190
265
10737
66892
697
7262
24933
13346
13502
13486
25071
256
34318
702
702
234
234
234
234
234
16735
16735
16735
A-48
-------
State
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
County
Rosebud
Rosebud
Rosebud
Rosebud
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Alexander
Beaufort
Beaufort
Beaufort
Chatham
Cumberland
Davie
Duplin
Edgecombe
Forsyth
Johnston
Lincoln
Martin
Mecklenburg
Mecklenburg
New Hanover
New Hanover
Northampton
Monitor ID
300870760
300870761
300870762
300870763
301110016
301110066
301110079
301110080
301110082
301110083
301110084
301111065
301112005
301112006
301112007
370030003
370130003
370130004
370130006
370370004
370511003
370590002
370610002
370650099
370670022
371010002
371090004
371170001
371190034
371190041
371290002
371290006
371310002
Population Residing within:
5km
0
0
0
0
8526
27389
61645
33774
58256
27620
22577
13350
24420
11205
5391
7574
1085
0
0
4146
32970
4799
850
0
61669
9854
10568
573
90874
105796
2584
17957
12284
10km
0
643
0
1536
9747
79644
89282
86065
94753
76641
59919
59574
68288
46767
26316
16738
1762
1762
1762
12138
108671
16224
6058
11321
170320
32163
32515
5282
276915
295729
20636
83529
29917
15km
643
3524
2928
3131
14953
98733
102887
104825
103200
98733
97912
97912
97912
86788
69446
40689
5519
6616
6616
23134
203822
44277
12866
25673
258102
67759
62768
14427
474624
494494
67021
145330
38134
20km
2928
3524
2928
3131
39121
107178
114640
108399
106046
109475
110980
110980
109475
110980
104067
80547
8488
8488
8488
72477
280713
93569
29813
51492
325974
129979
125735
26518
629520
647110
127088
170260
46966
Analysis Bins
Population1
low
low
low
low
low
mod
hi
mod
hi
mod
mod
mod
mod
mod
low
low
low
low
low
low
mod
low
low
low
hi
low
mod
low
hi
hi
low
mod
mod
cov2
c
b
a
a
b
b
b
b
b
b
b
b
b
b
b
a
a
b
b
a
a
a
a
a
b
a
a
a
b
b
b
b
a
GSD3
a
a
a
a
c
c
a
b
a
b
b
b
b
c
c
a
a
a
b
a
a
a
a
a
b
a
a
a
a
b
a
b
a
Emissions
(tpy)4
16735
5480
5480
5480
5480
5480
15298
5480
5480
15298
15298
4730
4730
4730
474
1477
7795
414
325
3945
29
10
3426
1030
821
29923
30020
2416
A-49
-------
State
NC
NC
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NE
NE
NE
NE
NH
NH
County
Person
Pitt
Swain
Billings
Billings
Burke
Burke
Burleigh
Cass
Cass
Dunn
McKenzie
McKenzie
McKenzie
McLean
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Morton
Morton
Oliver
Steele
Williams
Williams
Douglas
Douglas
Douglas
Douglas
Cheshire
Coos
Monitor ID
371450003
371470099
371730002
380070002
380070111
380130002
380130004
380150003
380171003
380171004
380250003
380530002
380530104
380530111
380550113
380570001
380570004
380570102
380570118
380570123
380570124
380590002
380590003
380650002
380910001
381050103
381050105
310550048
310550050
310550053
310550055
330050007
330070019
Population Residing within:
5km
2620
5860
3268
0
0
0
655
49591
48975
2118
0
0
0
0
0
3280
3280
1574
0
0
557
17925
10305
0
0
0
0
50168
45166
82663
13902
16719
9280
10km
8081
10688
8992
0
0
0
655
67377
134561
91149
0
596
521
0
632
3280
4428
4428
1574
557
557
67959
31348
0
934
1259
1259
209209
187855
264396
109385
30003
13603
15km
24203
23742
15036
1887
0
0
655
83082
144878
145789
0
596
521
2283
698
5902
5902
5902
6898
3837
557
75685
75685
2057
934
1259
1259
371395
367828
424100
299381
39998
14203
20km
41995
72588
18230
1887
0
625
655
84415
154455
148002
537
596
2283
5771
698
6465
7455
7455
7455
5981
3903
84415
82584
2670
934
1827
1827
532173
525602
578351
473231
53389
14928
Analysis Bins
Population1
low
low
low
low
low
low
low
mod
mod
low
low
low
low
low
low
low
low
low
low
low
low
mod
mod
low
low
low
low
hi
mod
hi
mod
mod
low
cov2
b
a
a
a
b
b
b
b
b
a
a
a
c
c
b
b
b
b
b
b
b
c
b
b
a
b
b
c
b
c
b
a
b
GSD3
b
a
a
a
a
b
b
a
a
b
a
a
a
a
a
b
a
b
b
b
b
c
b
b
a
a
c
b
a
b
a
b
c
Emissions
(tpy)4
96752
28
283
426
4592
771
756
5
210
823
91617
91617
91617
91617
91617
91617
4592
4592
28565
1605
1605
31850
31850
31850
11535
81
638
A-50
-------
State
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NM
NM
NM
County
Coos
Coos
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Hillsborough
Merrimack
Merrimack
Merrimack
Merrimack
Rockingham
Rockingham
Rockingham
Sullivan
Atlantic
Bergen
Burlington
Camden
Camden
Cumberland
Essex
Essex
Gloucester
Hudson
Hudson
Middlesex
Morris
Union
Union
Dona Ana
Dona Ana
Eddy
Monitor ID
330070022
330071007
330110016
330110019
330110020
330111009
330111010
330130007
330131003
330131006
330131007
330150009
330150014
330150015
330190003
340010005
340035001
340051001
340070003
340071001
340110007
340130011
340130016
340150002
340170006
340171002
340232003
340273001
340390003
340390004
350130008
350130017
350151004
Population Residing within:
5km
8360
2438
107911
104650
104650
72131
37423
27595
8787
8066
9351
25227
25984
25227
11339
6123
209619
71953
1 93686
8015
26454
209592
200779
32432
158136
343775
95281
13515
221266
1 94256
10195
40832
12050
10km
12552
2438
145660
140235
140235
130360
145620
54309
45710
35862
43118
48762
48762
48762
17306
33910
973093
261206
806251
46392
77939
1133321
1136145
107924
930071
1754575
371119
60394
868022
750485
49347
158545
12050
15km
14928
6025
1 96209
189502
189502
219169
333540
75576
74945
1 04656
73240
88743
78775
92738
34644
71617
3404473
561157
1761045
121996
109030
2811759
2763272
537340
3370494
5021807
839280
181888
1790660
1727936
114220
258940
14785
20km
14928
8364
270491
266391
266391
438168
467236
101847
138179
218207
98696
157669
148875
152363
48414
160179
5751193
1133142
2534030
262931
160091
5933785
5837087
1392192
5894707
8159098
1615249
361716
3314852
3277263
181522
387481
16465
Analysis Bins
Population1
low
low
hi
hi
hi
hi
mod
mod
low
low
low
mod
mod
mod
mod
low
hi
hi
hi
low
mod
hi
hi
mod
hi
hi
hi
mod
hi
hi
mod
mod
mod
cov2
b
c
b
b
b
a
b
b
c
b
b
b
b
b
a
a
a
b
b
a
a
a
b
b
a
a
a
b
a
a
a
c
b
GSD3
b
b
b
b
b
a
b
b
c
c
b
b
b
b
a
a
b
b
b
b
b
b
b
b
b
b
b
b
b
a
a
b
b
Emissions
(tpy)4
638
18
30806
30806
30806
454
772
30833
30833
30833
30833
13706
13706
13706
220
27848
15099
10733
17
646
27424
27638
26452
27538
29856
1675
38
23181
23146
37
574
4233
A-51
-------
State
NM
NM
NM
NM
NM
NM
NM
NV
NV
NV
NV
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
County
Grant
Grant
Hidalgo
San Juan
San Juan
San Juan
San Juan
Clark
Clark
Clark
Clark
Albany
Bronx
Bronx
Bronx
Bronx
Chautauqua
Chautauqua
Chautauqua
Chemung
Erie
Erie
Erie
Essex
Franklin
Hamilton
Herkimer
Kings
Kings
Madison
Monroe
Monroe
Monroe
Monitor ID
350170001
350171003
350230005
350450008
350450009
350450017
350451005
320030022
320030078
320030539
320030601
360010012
360050073
360050080
360050083
360050110
360130005
360130006
360130011
360150003
360290005
360294002
360298001
360310003
360330004
360410005
360430005
360470011
360470076
360530006
360551004
360551007
360556001
Population Residing within:
5km
4292
1429
0
22921
2930
0
491
0
0
226197
13570
108841
1215989
1278526
1162835
1205886
3605
14144
3605
41915
150194
80118
66153
492
0
0
2043
1301071
1173879
806
149439
129608
222716
10km
6951
5721
0
41258
18431
6492
2247
0
0
557934
22316
255221
3522226
3040232
2294809
344471 1
6928
29535
6928
68619
458758
328976
237799
2054
2880
0
2043
3958499
3316779
4985
384621
381741
407438
15km
21790
9904
0
51483
32213
10898
11772
0
2836
933583
71616
371301
5762144
5159927
4245952
5621679
15645
39906
15645
82014
680793
575596
503575
7005
5697
454
2043
6872002
5595972
7448
579436
570995
582031
20km
23982
24316
0
68906
58595
10936
16909
10778
2836
1236711
97845
484970
8036800
7489995
6315293
7878863
22519
47684
22519
101244
839570
768392
729503
10934
11358
2054
2043
8807020
7596057
17313
665760
669909
678777
Analysis Bins
Population1
low
low
low
mod
low
low
low
low
low
hi
mod
hi
hi
hi
hi
hi
low
mod
low
mod
hi
hi
hi
low
low
low
low
hi
hi
low
hi
hi
hi
cov2
c
c
c
b
b
b
b
a
a
a
a
b
a
a
a
a
b
c
b
a
b
c
b
b
b
b
b
a
a
b
b
b
b
GSD3
b
b
c
a
a
b
c
a
a
a
a
b
b
b
b
a
c
b
c
a
b
c
b
b
b
c
c
b
b
b
b
a
b
Emissions
(tpy)4
263
263
17344
585
50191
178
362
27101
26825
6659
26965
52177
404
40734
41722
40659
29050
28686
50379
50379
50379
A-52
-------
State
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
County
Nassau
New York
New York
Niagara
Onondaga
Putnam
Queens
Queens
Rensselaer
Rensselaer
Richmond
Schenectady
Suffolk
Suffolk
Ulster
Adams
Allen
Ashtabula
Belmont
Butler
Butler
Clark
Clermont
Columbiana
Columbiana
Columbiana
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Franklin
Franklin
Monitor ID
360590005
360610010
360610056
360632008
360671015
360790005
360810097
360810124
360830004
360831005
360850067
360930003
361030002
361030009
361111005
390010001
390030002
390071001
390133002
390170004
390171004
390230003
390250021
390290016
390290022
390292001
390350038
390350045
390350060
390350065
390356001
390490004
390490034
Population Residing within:
5km
172837
1 062324
1289280
60505
56156
15437
378415
823992
1987
1222
282277
100404
80740
101641
755
4630
15401
11409
17529
68823
47209
19786
7297
21336
21336
25779
136697
151001
116933
132176
191842
133697
157233
10km
677944
3421130
3673609
96530
207136
57790
1589364
2441512
5975
6357
653196
157970
526254
341308
1541
6792
67353
17288
41346
163124
96458
66337
20144
46769
46769
43920
547523
564795
512974
562942
529243
467572
482749
15km
1424915
6487922
6607580
1 76040
329787
111398
3438261
5839274
22806
19071
2026407
233426
950326
551178
7851
15822
90874
23848
95392
276076
1 52032
175311
53435
67377
67377
64319
932680
962245
907112
968826
883601
806703
868013
20km
2365352
8988411
8980807
348603
395331
223357
7138176
8419326
118285
69278
4371801
383092
1417428
802861
10684
22444
114512
42433
120821
487924
287701
410155
96496
101068
101068
92597
1214114
1221356
1201852
1244026
1165619
1042146
1090438
Analysis Bins
Population1
hi
hi
hi
hi
hi
mod
hi
hi
low
low
hi
hi
hi
hi
low
low
mod
mod
mod
hi
mod
mod
low
mod
mod
mod
hi
hi
hi
hi
hi
hi
hi
cov2
b
a
a
b
a
b
a
b
b
b
a
a
a
a
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
GSD3
b
b
a
a
a
c
b
b
c
c
b
a
b
b
c
b
b
b
c
b
b
b
b
b
c
b
b
b
b
b
b
b
b
Emissions
(tpy)4
1806
28873
29021
40748
3280
8183
8043
379
188
24733
96
1404
7344
19670
3977
8655
138904
9979
13912
2034
91822
186262
186262
179205
7403
7403
7403
7403
74869
450
450
A-53
-------
State
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
County
Gallia
Hamilton
Hamilton
Jefferson
Jefferson
Jefferson
Lake
Lake
Lawrence
Lorain
Lorain
Lorain
Lucas
Lucas
Mahoning
Mahoning
Meigs
Montgomery
Morgan
Morgan
Scioto
Scioto
Scioto
Stark
Summit
Summit
Tuscarawas
Tuscarawas
Cherokee
Kay
Kay
Kay
Mayes
Monitor ID
390530002
390610010
390612003
390810016
390810017
390811001
390850003
390853002
390870006
390930017
390930026
390931003
390950008
390950024
390990009
390990013
391051001
391130025
391150003
391150004
391450013
391450020
391450022
391510016
391530017
391530022
391570003
391570006
400219002
400710602
400719003
400719010
400979014
Population Residing within:
5km
1087
15310
71390
28019
30069
21833
48791
40430
26563
58361
54867
58580
62606
134960
79207
78376
5440
123978
1122
1122
15699
4530
3469
99075
104817
140332
26914
2710
993
25029
6614
1123
1947
10km
13134
124569
325799
70995
71838
53684
145694
92415
71376
129195
114602
124277
205665
319708
210961
214611
15029
304826
3168
3168
47369
11216
12081
208779
292059
329963
40238
15439
22584
31461
29697
3516
14224
15km
30170
345879
683493
96550
96408
91514
238216
141902
94538
249878
202571
251182
356815
466184
293714
294367
21812
511565
9162
9871
61292
45697
40548
291216
470747
454363
61526
38518
28182
31461
32746
16273
26265
20km
49474
632705
1079723
122094
122094
1 1 9322
407417
209471
131453
362235
298148
365323
487567
528531
378289
375287
31834
645130
22426
24252
77940
87756
82103
350367
574282
570258
85938
72765
36130
36740
35459
20121
29243
Analysis Bins
Population1
low
mod
hi
mod
mod
mod
mod
mod
mod
hi
hi
hi
hi
hi
hi
hi
low
hi
low
low
mod
low
low
hi
hi
hi
mod
low
low
mod
low
low
low
cov2
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
c
c
b
b
b
a
b
b
b
b
c
b
b
b
a
GSD3
b
c
b
c
c
c
b
c
b
b
b
b
b
b
b
b
b
b
c
b
b
c
c
b
c
b
b
b
a
b
b
a
b
Emissions
(tpy)4
190311
92654
7257
223185
223185
78071
72266
4799
11400
495
53
495
37337
37450
21074
21074
190311
9652
115526
115526
4351
4351
1269
11053
11053
2579
2556
7003
7003
19079
A-54
-------
State
OK
OK
OK
OK
OK
OK
OK
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Muskogee
Oklahoma
Oklahoma
Ottawa
Tulsa
Tulsa
Tulsa
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Beaver
Beaver
Berks
Berks
Blair
Bucks
Cambria
Centre
Dauphin
Delaware
Delaware
Erie
Indiana
Monitor ID
401010167
401090025
401091037
401159004
401430175
401430235
401430501
420030002
420030010
420030021
420030031
420030032
420030064
420030067
420030116
420031301
420033003
420033004
420070002
420070004
420070005
420070014
420110009
420110100
420130801
420170012
420210011
420270100
420430401
420450002
420450109
420490003
420630004
Population Residing within:
5km
5633
78654
46197
6272
53094
65020
46840
83332
168140
170777
183843
174072
64846
13277
96820
1 1 5432
55221
38588
3434
35152
17292
36335
121330
118553
44392
85719
50440
60659
86638
74840
59762
81199
1110
10km
39252
254952
141934
22614
207546
235972
187023
277442
536314
560187
580429
558904
201143
86792
331624
411867
202092
170065
28961
1 04660
77240
82468
203799
202746
72996
324327
79710
76595
219394
237232
209503
1 50626
8662
15km
56271
384825
258441
29716
357175
405434
333482
651551
842237
921490
877668
922097
520438
324154
704601
766188
509708
461433
68617
203430
143738
134467
250610
254794
94779
638218
102905
96267
324647
510590
446058
190212
23057
20km
64455
552894
459371
37508
485641
515780
468989
961378
1114184
1142754
1145039
1144558
943781
610975
996267
1088115
944188
904760
120780
317823
224631
220614
309553
310286
124536
1212911
124592
107078
384070
1091830
812243
209983
57759
Analysis Bins
Population1
low
hi
mod
low
hi
hi
mod
hi
hi
hi
hi
hi
hi
mod
hi
hi
hi
mod
low
mod
mod
mod
hi
hi
mod
hi
hi
hi
hi
hi
hi
hi
low
cov2
b
a
a
b
b
b
b
b
a
b
a
b
b
a
b
b
b
b
b
a
b
b
b
a
b
a
b
a
a
a
a
a
b
GSD3
b
a
a
a
c
c
b
b
b
b
b
b
b
b
b
b
c
b
b
a
c
b
b
b
b
b
b
b
b
b
b
a
b
Emissions
(tpy)4
30011
182
182
62
9377
9377
9377
1964
4688
52447
46957
52447
11490
1167
1964
52100
11490
11501
187257
41170
41385
44003
14817
14774
441
15117
16779
4359
857
38833
38470
4122
14389
A-55
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PR
PR
PR
County
Lackawanna
Lancaster
Lawrence
Lehigh
Luzerne
Lycoming
Lycoming
Mercer
Montgomery
Northampton
Northampton
Northampton
Perry
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Schuylkill
Warren
Warren
Washington
Washington
Washington
Westmoreland
York
Barceloneta
Bayamon
Bayamon
Monitor ID
420692006
420710007
420730015
420770004
420791101
420810100
420810403
420850100
420910013
420950025
420950100
420958000
420990301
421010004
421010022
421010024
421010027
421010029
421010047
421010048
421010055
421010136
421070003
421230003
421230004
421250005
421250200
421255001
421290008
421330008
720170003
720210004
720210006
Population Residing within:
5km
68522
97205
40803
133092
68639
15088
41897
40443
91275
79756
71422
71626
6450
400078
316944
197076
472813
484661
410380
262592
341893
382995
19152
14142
13965
31276
32125
1359
35656
85574
29823
192976
208167
10km
144913
174296
57962
298181
157363
60400
69102
69465
239337
173911
118395
133639
13169
1147634
985213
588104
1348135
1229942
1153434
1102727
1020004
985957
30388
19940
18884
68512
52910
15854
82661
156166
83433
679576
587003
15km
189515
254789
81815
395772
215050
83910
80935
96468
706445
398867
209567
228524
26326
1971579
1726387
1351349
2026206
1999611
1989848
1938877
1774411
1718068
59370
25715
28805
111222
83324
43364
148990
216656
134176
1002864
956783
20km
246604
344292
118770
501878
265123
108961
103969
184589
1623890
513651
317220
330629
49400
2631448
2446142
2063868
2632847
2574304
2573573
2607877
2476647
2381173
100508
32490
33523
183285
118188
126091
213978
284208
243828
1292141
1256603
Analysis Bins
Population1
hi
hi
mod
hi
hi
mod
mod
mod
hi
hi
hi
hi
low
hi
hi
hi
hi
hi
hi
hi
hi
hi
mod
mod
mod
mod
mod
low
mod
hi
mod
hi
hi
cov2
a
b
b
a
a
b
a
b
a
a
a
b
b
b
a
b
a
b
a
b
a
b
a
b
b
a
a
b
a
b
b
b
b
GSD3
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
b
c
a
b
b
b
b
a
b
b
Emissions
(tpy)4
66
375
28854
9143
467
83
83
28
4794
12167
32680
32714
6228
18834
1663
6246
17550
17536
6214
18848
21700
4987
4890
4890
8484
7
2566
72
80487
A-56
-------
State
PR
PR
PR
PR
PR
PR
Rl
Rl
Rl
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SD
SD
SD
TN
TN
TN
TN
TN
TN
TN
County
Catano
Catano
Catano
Catano
Guayama
Salinas
Providence
Providence
Providence
Aiken
Barnwell
Charleston
Charleston
Georgetown
Greenville
Greenville
Lexington
Oconee
Orangeburg
Richland
Richland
Richland
Richland
Custer
Jackson
Minnehaha
Anderson
Blount
Blount
Blount
Bradley
Coffee
Davidson
Monitor ID
720330004
720330007
720330008
720330009
720570009
721230001
440070012
440071005
440071009
450030003
450110001
450190003
450190046
450430006
450450008
450450009
450630008
450730001
450750003
450790007
450790021
450791003
450791006
460330132
460710001
460990007
470010028
470090002
470090006
470090101
470110102
470310004
470370011
Population Residing within:
5km
154575
95500
99778
110439
12086
20645
223521
148802
226940
752
0
40872
1103
10567
70221
56686
42208
0
2904
35872
1666
87097
1666
0
0
65647
11872
28887
36020
0
2540
1286
77459
10km
583552
576841
594607
457427
49373
31312
487990
390751
493584
6505
4022
132716
1103
18215
173012
151862
131361
2260
7856
121006
4643
213836
5435
0
0
1 1 9287
59225
70731
72290
12650
11940
9718
228349
15km
958456
983702
972270
883511
90444
68199
638092
615465
646894
18533
13647
273298
9529
22467
284047
279293
257820
11136
14446
255135
13324
300874
15920
3940
0
138918
153931
105939
104178
44702
46188
23113
410925
20km
1233122
1219701
1238188
1164315
174005
174332
816597
809993
821476
55485
21554
364953
22255
34357
379022
356410
355854
26182
24656
353072
33098
396116
47548
4686
0
147218
292415
198408
189214
81010
84762
35158
583532
Analysis Bins
Population1
hi
hi
hi
hi
mod
mod
hi
hi
hi
low
low
mod
low
mod
hi
hi
mod
low
low
mod
low
hi
low
low
low
hi
mod
mod
mod
low
low
low
hi
cov2
b
b
b
b
a
b
a
b
a
a
a
b
b
b
a
b
b
a
b
a
b
b
b
b
a
b
c
b
b
b
b
b
a
GSD3
b
b
b
b
a
b
b
b
b
a
a
b
a
b
b
a
b
a
a
a
b
a
a
a
a
a
b
c
b
a
a
a
b
Emissions
(tpy)4
2228
2265
2253
21498
65
34934
40841
1067
1082
10433
5
7166
613
40492
12935
42894
496
44761
4263
4263
4263
5437
8019
A-57
-------
State
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
County
Hawkins
Humphreys
McMinn
Montgomery
Montgomery
Polk
Polk
Polk
Polk
Roane
Shelby
Shelby
Shelby
Shelby
Stewart
Sullivan
Sullivan
Sumner
Cameron
Dallas
Ellis
Ellis
Ellis
El Paso
El Paso
El Paso
Galveston
Galveston
Gregg
Harris
Harris
Harris
Harris
Monitor ID
470730002
470850020
471070101
471250006
471250106
471390003
471390007
471390008
471390009
471450009
471570034
471570043
471570046
471571034
471610007
471630007
471630009
471651002
480610006
481130069
481390015
481390016
481390017
481410037
481410053
481410058
481670005
481671002
481830001
482010046
482010051
482010059
482010062
Population Residing within:
5km
6748
2474
2540
21032
16569
1613
2491
2491
2491
8848
74216
94449
18782
886
787
28689
28254
5070
70071
93552
6089
7883
5723
56009
49083
78658
37427
38619
1349
65125
123431
151412
73770
10km
14441
6672
11940
79399
74449
9042
9432
9432
10239
21677
277713
325228
113964
97506
4566
78826
77403
38555
151247
455917
13876
18193
17592
1 82473
163206
126481
62491
65658
17138
350122
372470
475338
352818
15km
22457
13621
37322
112883
109087
14124
19726
17401
17235
37175
497847
534950
273306
277857
8854
112565
117095
53602
160048
991123
35210
68740
50332
337222
325118
299419
98724
98768
52116
756166
896497
902121
695432
20km
39857
23460
84929
139621
138438
24537
24026
25902
24026
57683
695164
751299
473443
484234
20362
153445
151856
119241
167993
1609774
113413
191352
152699
522824
519008
524259
182464
196215
105781
1283440
1380154
1392348
1108749
Analysis Bins
Population1
low
low
low
mod
mod
low
low
low
low
low
hi
hi
mod
low
low
mod
mod
low
hi
hi
low
low
low
hi
mod
hi
mod
mod
low
hi
hi
hi
hi
cov2
c
c
b
a
a
b
c
a
b
c
a
a
b
b
b
b
b
c
a
a
b
b
c
b
b
b
b
b
c
b
b
b
b
GSD3
c
b
a
a
a
a
b
a
a
a
a
b
a
b
a
c
c
b
a
a
b
c
b
b
b
a
b
b
b
a
a
c
b
Emissions
(tpy)4
35493
111597
5501
1330
1330
1900
1900
1900
1900
77881
21675
21675
3945
21847
16682
30097
30156
34373
307
7972
7972
7972
574
574
614
7976
7976
66443
17583
26
25608
25677
A-58
-------
State
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
UT
UT
UT
UT
UT
UT
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VI
VI
VI
County
Harris
Harris
Harris
Jefferson
Jefferson
Jefferson
Kaufman
Nueces
Nueces
Nueces
Cache
Davis
Davis
Salt Lake
Salt Lake
Salt Lake
Charles
Fairfax
Fairfax
Fairfax
Fairfax
Fairfax
Madison
Roanoke
Rockingham
Rockingham
Alexandria City
Hampton City
Norfolk City
Richmond City
St Croix
St Croix
St Croix
Monitor ID
482010070
482011035
482011050
482450009
482450011
482450020
482570005
483550025
483550026
483550032
490050004
490110001
490110004
490350012
490351001
490352004
510360002
510590005
510590018
510591004
510591005
510595001
511130003
511611004
511650002
511650003
515100009
516500004
517100023
517600024
780100006
780100011
780100013
Population Residing within:
5km
153479
99581
23794
33143
13164
35739
6583
99888
16215
48320
49600
56718
52464
57910
31709
0
3370
34561
87725
215952
203219
80603
1316
33161
17897
13821
137533
73011
124263
109306
0
0
0
10km
511407
451485
83705
87386
93985
101563
9190
186846
28033
128230
64094
82741
83909
1 83684
1 07346
4074
32169
183637
293189
660586
670880
358173
4823
123148
58020
47577
622283
182507
379455
309672
0
0
0
15km
991134
891195
224120
182005
121116
177284
28396
231717
92841
228351
80592
178141
1 54925
370433
260423
35159
76679
408647
730360
1410007
1238334
1098236
13930
197615
76316
71219
1320784
356676
703082
524083
0
0
0
20km
1610993
1287766
405297
237033
140687
223336
43226
280479
177008
272861
86020
311810
295333
630857
522228
124394
176978
687195
1388941
2092422
1 844099
2041931
28417
235072
85276
92912
1894197
601943
871632
656099
0
0
0
Analysis Bins
Population1
hi
hi
mod
mod
mod
mod
low
hi
mod
mod
mod
hi
hi
hi
mod
low
low
mod
hi
hi
hi
hi
low
mod
mod
mod
hi
hi
hi
hi
low
low
low
COV2
b
b
a
b
b
b
a
b
b
b
a
b
b
b
b
b
b
a
a
a
a
a
b
a
a
a
b
b
b
b
b
b
c
GSD3
b
b
a
c
c
c
a
b
a
b
a
b
a
a
a
a
b
b
b
a
a
a
c
a
a
a
b
b
b
b
c
c
c
Emissions
(tpy)4
24501
25635
11195
13807
26962
1362
7954
8056
7954
5
2807
2807
2807
5832
3735
86717
156
18204
18303
18405
17221
7
677
277
235
18293
4274
36499
2675
A-59
-------
State
VI
VI
VT
VT
VT
WA
WA
WA
WA
WA
WA
WA
WA
WA
WA
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
Wl
WV
WV
WV
WV
WV
WV
WV
County
St Croix
St Croix
Chittenden
Chittenden
Rutland
Clallam
Clallam
King
King
Pierce
Pierce
Skagit
Skagit
Snohomish
What com
Brown
Dane
Forest
Marathon
Milwaukee
Milwaukee
Milwaukee
Oneida
Sauk
Vilas
Wood
Brooke
Brooke
Cabell
Greenbrier
Hancock
Hancock
Hancock
Monitor ID
780100014
780100015
500070003
500070014
500210002
530090010
530090012
530330057
530330080
530530021
530530031
530570012
530571003
530610016
530730011
550090005
550250041
550410007
550730005
550790007
550790026
550790041
550850996
551110007
551250001
551410016
540090005
540090007
540110006
540250001
540290005
540290007
540290008
Population Residing within:
5km
0
0
50990
54166
21330
17871
20830
131605
116769
68072
55628
3580
1733
46071
60525
79060
79610
1330
5095
248317
214859
137816
8351
2743
934
19525
25010
30794
50835
2158
6006
14924
24095
10km
0
0
89229
87471
30052
26073
27014
394412
423064
250876
275358
22573
21622
1 52230
83632
158940
189421
3913
42173
606921
572784
455868
17018
15039
934
33790
64711
70187
88879
9273
23418
44311
49351
15km
0
0
110853
110853
35316
30255
30036
730218
811856
548806
555755
32120
32120
303720
111425
201226
306132
5514
61417
865925
834939
765734
17018
24240
8639
43315
92813
95823
125923
18280
77160
83167
63727
20km
0
0
133530
133749
46525
37672
36843
1093083
1157199
839357
820805
70660
75069
432356
126291
215144
353861
6669
93151
1037293
1014161
964876
23821
43368
10755
50360
118070
120385
164495
23902
125873
128126
91485
Analysis Bins
Population1
low
low
hi
hi
mod
mod
mod
hi
hi
hi
hi
low
low
mod
hi
hi
hi
low
low
hi
hi
hi
low
low
low
mod
mod
mod
hi
low
low
mod
mod
cov2
c
c
b
a
b
b
b
a
b
b
b
b
b
a
a
b
b
b
c
b
b
b
c
b
a
c
b
b
b
a
b
b
b
GSD3
c
c
a
a
c
b
a
a
a
b
a
b
b
a
a
b
b
a
a
b
a
b
c
a
a
b
b
b
b
a
b
b
b
Emissions
(tpy)4
6
6
756
756
1203
1203
538
538
8951
8951
381
4391
23888
9049
5
12120
15753
15753
15753
2304
63
14245
78071
223185
7504
176554
148520
186262
A-60
-------
State
WV
wv
WV
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
WY
County
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Kanawha
Kanawha
Kanawha
Marshall
Monongalia
Monongalia
Monongalia
Ohio
Wayne
Wayne
Wayne
Wayne
Wood
Campbell
Monitor ID
540290009
540290011
540290014
540290015
540290016
540291004
540390004
540390010
540392002
540511002
540610003
540610004
540610005
540690007
540990002
540990003
540990004
540990005
541071002
560050857
Population Residing within:
5km
20946
31890
22857
20793
19278
24761
46977
48231
21694
13403
43902
44079
46591
20818
17320
17320
16553
13314
24917
3288
10km
61117
76198
58620
45848
70483
63677
80511
83340
61059
32048
65672
63708
61019
60048
62645
59989
54251
48330
70324
11413
15km
90717
95992
89998
65851
96151
91977
120631
123101
111812
55054
80405
80385
77800
91967
124477
123349
122072
114824
104458
23902
20km
115283
115162
120724
102031
114992
115615
164476
172217
164912
95735
98315
98966
99544
126981
178576
177744
179815
173807
128127
25752
Analysis Bins
Population1
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
mod
low
COV2
b
b
b
b
b
b
b
b
b
b
b
b
b
b
a
b
b
b
b
b
GSD3
b
b
b
b
b
b
b
b
b
c
b
b
b
b
b
b
b
b
b
b
Emissions
(tpy)4
148404
223185
148520
186262
169771
169771
6115
6115
113491
138904
91984
97887
96396
74781
10172
10172
10172
10172
48124
10106
Notes:
1 Population bins: low (<1 0,000); mid (10,001 to 50,000); hi (>50,000) using population within 5 km of ambient monitor.
2 COV bins: a (<100%); b (>100 to <200); c (>200).
3 GSD bins: a (<2.17); b (>2.17 to <2.94); c (>2.94).
4 Sum of emissions within 20 km radius of ambient monitor based on 2002 NEI.
A-61
-------
A.1.2 Analysis of SOi Emission Sources Surrounding Ambient Monitors
Distances of the 5-minute and 1-hour ambient monitoring sites to stationary sources
emitting SC>2 were estimated using data from the 2002 National Emissions Inventory1 (NEI).
The NEI database reports emissions of SO2 in tons per year (tpy) for 98,667 unique emission
sources at various points of release. The release locations were all taken from the latitude
longitude values within the NEI. First, all SC>2 emissions were summed for identical latitude and
longitude entries while retaining source codes for the emissions (e.g., Standard Industrial Code
(SIC), or North American Industrial Classification System (NAICS)). Therefore, any facility
containing similar emission processes were summed at the stack location, resulting in 32,521
observations. These data were then screened for sources with emissions greater than 5 tpy,
yielding 6,104 unique 862 emission sources. Locations of these stationary source emissions
were compared with ambient monitoring locations using the following formula:
d = arccos(sin(toj)x sin(to2) + cos(latl)xcos(to2)xcos(/o«2 -lon^))xr
where
d = distance (kilometers)
latj = latitude of a monitor (radians)
Iat2 = latitude of source emission (radians)
lon} = longitude of monitor (radians)
Ion2 = longitude of source emission (radians)
r = approximate radius of the earth (or 6,371 km)
Location data for monitors and sources provided in the AQS and NEI data bases were
given in units of degrees therefore, these were first converted to radians by dividing by 180/Ti.
For each monitor, source emissions within 20 km of the monitor were retained.
Table A. 1-5 contains the summary of the distance of stationary source emissions to each
of the monitors in the broader SC>2 monitoring network. There were varying numbers of sources
emitting >5 tpy of SC>2 and located within a 20 km radius for many of the monitors. Some of the
monitors are point-source oriented, that is, sited to measure ambient concentrations potentially
1 2002 National Emissions Inventory Data & Documentation. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. Available at: http://www.epa.gov/ttn/chief/net/2002inventory.html.
A-62
-------
influenced by a specific single sources (e.g., Missouri monitor IDs 290210009, 290210011,
290930030), or by several sources (e.g., Pennsylvania monitor IDs 420030021, 420030031) of
varying emission strength. A few of the monitors contained no source emissions >5 tpy (e.g.,
Iowa monitor IDs 191770005, 191770006).
A-63
-------
Table A.1-5. Distance of ambient SO2 monitors (all used in analysis) to stationary sources emitting > 5 tons of SO2
per year, within a 20 kilometer distance of monitoring site, and SO2 emissions associated with those stationary
sources.
Monitor ID
010330044
010710020
010731003
010790003
010970028
010972005
011011002
040070009
040071001
040130019
040133002
040133003
040191011
040212001
051190007
051191002
051390006
060010010
060130002
060130006
060130010
060131001
060131002
060131003
060131004
060132001
060133001
060250005
060371002
060371103
060374002
060375001
n
3
3
43
5
10
9
4
0
2
8
9
9
1
0
1
1
6
7
15
9
15
13
3
9
9
15
16
1
3
15
32
31
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
16680
15119
151
1787
6613
132
913
9219
23
21
20
3119
20
20
421
53
1004
559
1189
1507
26
559
559
1189
507
7
17
37
183
203
std
28821
25004
227
3416
13057
154
1183
10723
19
19
19
689
66
2007
789
1977
2036
21
789
789
1977
1104
7
36
313
342
min
30
98
5
6
14
5
180
1637
10
10
6
3119
20
20
8
5
6
5
6
6
6
5
5
6
6
7
10
7
5
5
p2.5
30
98
5
6
14
5
180
1637
10
10
6
3119
20
20
8
5
6
5
6
6
6
5
5
6
6
7
10
7
5
5
p50
51
1276
38
58
214
72
403
9219
19
14
14
3119
20
20
22
14
58
38
419
793
25
38
38
419
48
7
17
29
46
61
p97.5
49960
43983
786
7852
38917
440
2663
16801
69
69
69
3119
20
20
1689
187
7009
1829
7009
7009
48
1829
1829
7009
4337
7
24
119
1503
1503
p100
49960
43983
982
7852
38917
440
2663
16801
69
69
69
3119
20
20
1689
187
7009
1829
7009
7009
48
1829
1829
7009
4337
7
24
119
1503
1503
Distance of monitor to SO2 emission source (km)1
mean
6.0
5.7
11.4
8.4
7.5
7.7
12.7
0.0
1.3
11.0
10.8
12.5
6.1
0.0
6.3
13.7
7.7
8.9
13.5
13.0
8.3
10.1
11.7
12.6
12.8
8.8
9.8
18.0
6.8
13.8
10.4
13.4
std
0.7
2.4
5.5
1.7
5.8
2.1
7.2
0.0
0.5
3.6
6.6
4.7
0.0
0.0
0.0
0.0
4.2
5.4
2.8
6.4
5.6
5.9
1.4
4.8
5.7
5.4
6.1
0.0
2.1
5.1
5.2
5.9
min
5.5
3.1
1.1
5.5
1.4
4.3
4.5
0.0
0.9
5.6
1.9
5.5
6.1
0.0
6.3
13.7
1.9
1.2
9.6
2.5
1.6
0.2
10.1
5.4
4.1
2.3
0.7
18.0
4.7
6.3
4.1
3.7
p2.5
5.5
3.1
1.2
5.5
1.4
4.3
4.5
0.0
0.9
5.6
1.9
5.5
6.1
0.0
6.3
13.7
1.9
1.2
9.6
2.5
1.6
0.2
10.1
5.4
4.1
2.3
0.7
18.0
4.7
6.3
4.1
3.7
p50
5.9
6.2
13.1
8.6
6.1
7.2
13.2
0.0
1.3
10.2
11.2
12.4
6.1
0.0
6.3
13.7
8.8
9.0
13.3
15.0
6.4
9.9
12.4
12.2
13.5
6.7
11.4
18.0
6.9
12.5
9.3
16.4
p97.5
6.8
7.8
16.8
9.8
19.1
10.1
19.9
0.0
1.6
16.9
19.2
18.5
6.1
0.0
6.3
13.7
11.7
16.8
17.8
19.3
19.7
19.8
12.7
19.0
19.1
19.9
18.6
18.0
8.8
19.8
19.5
19.6
p100
6.8
7.8
19.8
9.8
19.1
10.1
19.9
0.0
1.6
16.9
19.2
18.5
6.1
0.0
6.3
13.7
11.7
16.8
17.8
19.3
19.7
19.8
12.7
19.0
19.1
19.9
18.6
18.0
8.8
19.8
19.5
19.6
A-64
-------
Monitor ID
060375005
060591003
060658001
060670002
060670006
060710012
060710014
060710306
060711234
060712002
060714001
060730001
060731007
060732007
060750005
060791005
060792001
060792004
060794002
060830008
060831012
060831013
060831015
060831016
060831019
060831020
060831025
060831026
060831027
060832004
060832011
060834003
060870003
060950001
060950004
061113001
n
12
7
4
1
1
1
2
2
3
2
1
1
3
1
6
7
7
7
7
3
2
2
1
1
1
3
3
3
3
2
3
2
1
13
12
2
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
192
10
75
5
58
8
126
126
97
102
32
21
11
21
66
536
536
536
536
39
554
554
18
18
18
39
39
39
39
554
39
554
722
1371
1480
9
std
332
5
76
132
132
85
112
9
83
1369
1369
1369
1369
43
357
357
43
43
43
43
357
43
357
2071
2124
3
min
6
5
17
5
58
8
32
32
6
22
32
21
5
21
5
6
6
6
6
10
302
302
18
18
18
10
10
10
10
302
10
302
722
6
6
7
p2.5
6
5
17
5
58
8
32
32
6
22
32
21
5
21
5
6
6
6
6
10
302
302
18
18
18
10
10
10
10
302
10
302
722
6
6
7
p50
33
7
50
5
58
8
126
126
110
102
32
21
7
21
39
24
24
24
24
18
554
554
18
18
18
18
18
18
18
554
18
554
722
790
791
9
P97.5
1119
18
181
5
58
8
219
219
175
181
32
21
21
21
224
3642
3642
3642
3642
89
807
807
18
18
18
89
89
89
89
807
89
807
722
7009
7009
11
p100
1119
18
181
5
58
8
219
219
175
181
32
21
21
21
224
3642
3642
3642
3642
89
807
807
18
18
18
89
89
89
89
807
89
807
722
7009
7009
11
Distance of monitor to SO2 emission source (km)1
mean
9.1
13.9
16.8
14.8
9.7
11.9
8.0
8.1
4.9
13.2
6.5
4.0
12.9
16.4
13.3
1.0
10.5
2.5
18.4
9.7
16.5
14.1
15.6
15.2
13.7
7.3
10.9
9.9
10.3
4.2
10.4
16.4
0.8
7.1
13.0
10.4
std
5.9
5.7
4.7
0.0
0.0
0.0
3.0
3.3
5.8
2.9
0.0
0.0
1.3
0.0
6.2
0.2
0.2
0.1
0.1
6.2
0.1
0.1
0.0
0.0
0.0
8.2
8.9
8.0
7.7
0.1
8.4
0.1
0.0
3.3
4.9
5.1
min
2.3
5.3
9.8
14.8
9.7
11.9
5.9
5.7
1.3
11.2
6.5
4.0
11.8
16.4
1.8
0.8
10.2
2.3
18.3
2.8
16.4
14.1
15.6
15.2
13.7
2.0
0.8
0.9
1.7
4.2
3.9
16.4
0.8
2.1
5.5
6.8
p2.5
2.3
5.3
9.8
14.8
9.7
11.9
5.9
5.7
1.3
11.2
6.5
4.0
11.8
16.4
1.8
0.8
10.2
2.3
18.3
2.8
16.4
14.1
15.6
15.2
13.7
2.0
0.8
0.9
1.7
4.2
3.9
16.4
0.8
2.1
5.5
6.8
p50
6.0
15.6
18.8
14.8
9.7
11.9
8.0
8.1
1.9
13.2
6.5
4.0
12.5
16.4
15.2
1.0
10.4
2.6
18.5
11.3
16.5
14.1
15.6
15.2
13.7
3.2
14.2
12.6
12.8
4.2
7.5
16.4
0.8
7.5
13.6
10.4
P97.5
19.8
19.7
19.6
14.8
9.7
11.9
10.1
10.4
11.7
15.3
6.5
4.0
14.4
16.4
18.3
1.5
10.9
2.7
18.5
14.9
16.5
14.2
15.6
15.2
13.7
16.7
17.7
16.2
16.4
4.2
20.0
16.5
0.8
13.8
19.6
14.0
p100
19.8
19.7
19.6
14.8
9.7
11.9
10.1
10.4
11.7
15.3
6.5
4.0
14.4
16.4
18.3
1.5
10.9
2.7
18.5
14.9
16.5
14.2
15.6
15.2
13.7
16.7
17.7
16.2
16.4
4.2
20.0
16.5
0.8
13.8
19.6
14.0
A-65
-------
Monitor ID
080010007
080013001
080310002
080416001
080416004
080416011
080416018
090010012
090010017
090011123
090012124
090019003
090031005
090031018
090032006
090090027
090091003
090091123
090092123
090110007
090130003
100031003
100031007
100031008
100031013
100032002
100032004
110010041
120110010
120310032
120310080
120310081
120310097
120330004
120330022
120470015
n
24
20
24
3
3
3
2
11
3
0
4
10
28
7
6
8
9
9
5
6
0
34
11
24
34
36
39
13
8
14
15
13
14
6
6
3
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
1001
1191
1098
1670
2849
2849
4268
425
252
192
504
45
16
14
595
565
565
86
650
975
3126
1657
975
802
1526
1410
2397
2715
2534
2923
2715
7262
7262
755
std
3352
3657
3356
2857
4920
4920
6026
1198
423
366
1257
106
9
7
1388
1302
1302
96
1088
1619
6528
4554
1619
1272
3681
4437
6653
5784
5617
5965
5784
14101
14101
1268
min
8
8
6
7
7
7
7
5
5
5
5
5
5
5
5
5
5
9
7
5
15
5
5
5
5
7
17
5
5
5
5
6
6
18
p2.5
8
8
6
7
7
7
7
5
5
5
5
5
5
5
5
5
5
9
7
5
15
5
5
5
5
7
17
5
5
5
5
6
6
18
p50
25
28
28
34
10
10
4268
21
11
10
10
12
15
15
32
43
43
28
110
112
103
60
112
97
116
24
41
287
257
317
287
330
330
27
P97.5
15958
15958
15958
4969
8530
8530
8530
4024
741
741
4024
522
30
25
4012
4012
4012
198
2755
6720
19923
19923
6720
5051
19923
16141
18861
20908
20908
20908
20908
35417
35417
2218
p100
15958
15958
15958
4969
8530
8530
8530
4024
741
741
4024
522
30
25
4012
4012
4012
198
2755
6720
19923
19923
6720
5051
19923
16141
18861
20908
20908
20908
20908
35417
35417
2218
Distance of monitor to SO2 emission source (km)1
mean
9.8
8.3
9.2
6.2
13.0
9.9
6.8
6.1
9.6
0.0
7.4
13.2
14.2
7.7
4.5
6.3
7.4
7.3
9.0
8.4
0.0
9.0
10.5
10.9
9.1
10.2
10.9
11.7
11.0
9.0
12.0
7.9
9.2
9.3
7.6
3.0
std
6.1
5.8
4.5
8.6
3.9
7.6
0.5
4.9
6.4
0.0
6.0
5.1
3.6
6.3
5.1
7.3
8.2
8.2
5.7
3.0
0.0
5.2
2.2
6.9
4.8
6.2
6.2
6.5
6.1
4.2
4.7
4.6
4.7
4.2
4.4
0.5
min
2.4
1.6
3.9
0.8
10.7
2.5
6.5
2.1
5.7
0.0
2.3
4.0
3.0
1.9
0.5
1.0
0.7
0.8
0.8
3.3
0.0
1.5
9.1
2.2
2.8
1.1
1.3
0.6
5.1
1.3
1.1
1.3
3.1
4.9
2.4
2.6
p2.5
2.4
1.6
3.9
0.8
10.7
2.5
6.5
2.1
5.7
0.0
2.3
4.0
3.0
1.9
0.5
1.0
0.7
0.8
0.8
3.3
0.0
1.5
9.1
2.2
2.8
1.1
1.3
0.6
5.1
1.3
1.1
1.3
3.1
4.9
2.4
2.6
p50
8.2
5.9
7.0
1.8
10.9
9.6
6.8
4.8
6.2
0.0
6.5
14.0
14.2
3.7
1.8
3.1
2.6
2.7
8.9
8.7
0.0
8.1
9.8
13.9
8.3
9.4
11.1
11.5
7.5
9.1
13.3
6.5
7.7
8.9
8.4
2.7
P97.5
19.7
19.8
19.5
16.1
17.6
17.7
7.2
19.7
17.0
0.0
14.4
19.5
19.9
18.4
11.4
18.6
19.7
19.7
15.2
12.7
0.0
19.8
16.2
19.7
18.9
19.7
19.8
19.8
19.2
18.5
19.7
15.5
19.5
14.6
12.3
3.6
p100
19.7
19.8
19.5
16.1
17.6
17.7
7.2
19.7
17.0
0.0
14.4
19.5
19.9
18.4
11.4
18.6
19.7
19.7
15.2
12.7
0.0
19.8
16.2
19.7
18.9
19.7
19.8
19.8
19.2
18.5
19.7
15.5
19.5
14.6
12.3
3.6
A-66
-------
Monitor ID
120570021
120570053
120570081
120570095
120570109
120571035
120574004
120813002
120860019
120890005
120890009
120952002
120993004
121030023
121033002
121035002
121035003
121050010
121052006
121071008
121151002
121151005
121151006
130090001
130150002
130210012
130510019
130510021
130511002
130950006
131110091
131150003
131210048
131210055
131270006
132150008
n
18
19
18
17
16
18
3
5
7
4
4
5
6
7
6
2
2
9
13
3
0
0
2
2
4
11
14
14
14
4
1
8
7
7
3
4
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
4986
4728
6781
3845
4084
4986
2872
73
34
1262
1262
9
39
3546
4136
15398
15398
2386
1691
9965
71
36975
40604
245
1362
1362
1362
1693
1900
4057
4339
4339
821
1740
std
11445
11180
13097
11285
11610
11445
4949
93
45
1594
1594
4
38
7041
7521
21767
21767
2929
2627
12565
90
52282
80047
468
2664
2664
2664
2220
9625
10445
10445
948
3214
min
6
6
6
6
6
6
11
6
5
11
11
5
5
6
23
7
7
6
6
12
7
6
21
6
8
8
8
5
1900
5
68
68
14
8
p2.5
6
6
6
6
6
6
11
6
5
11
11
5
5
6
23
7
7
6
6
12
7
6
21
6
8
8
8
5
1900
5
68
68
14
8
p50
341
104
1116
61
83
341
19
9
12
765
765
10
32
104
156
15398
15398
1210
230
5799
71
36975
862
17
235
235
235
932
1900
101
169
169
586
197
P97.5
47103
47103
47103
47103
47103
47103
8587
208
130
3509
3509
14
103
18822
18822
30790
30790
8587
8587
24083
135
73943
160673
1576
7969
7969
7969
4905
1900
27594
27993
27993
1865
6559
p100
47103
47103
47103
47103
47103
47103
8587
208
130
3509
3509
14
103
18822
18822
30790
30790
8587
8587
24083
135
73943
160673
1576
7969
7969
7969
4905
1900
27594
27993
27993
1865
6559
Distance of monitor to SO2 emission source (km)1
mean
11.6
10.8
14.4
10.1
9.9
10.6
14.2
7.2
9.6
4.5
4.2
13.8
12.0
7.4
10.3
13.6
9.8
10.4
11.9
5.8
0.0
0.0
15.8
11.3
10.4
10.1
7.1
6.8
6.2
6.3
1.6
1.4
10.3
15.1
3.6
12.5
std
6.5
3.5
4.5
6.2
4.0
5.8
8.5
8.3
5.8
5.0
4.6
2.8
3.1
6.4
4.2
0.1
4.0
3.1
6.2
3.4
0.0
0.0
0.8
5.4
5.2
5.2
4.1
4.4
3.4
5.3
0.0
0.4
2.1
3.3
2.8
2.6
min
1.4
5.9
8.1
2.5
6.7
1.6
4.4
0.7
2.6
1.1
1.1
10.1
7.0
2.3
3.5
13.5
7.0
3.7
2.7
2.6
0.0
0.0
15.2
7.5
2.5
1.5
0.4
1.4
1.6
2.2
1.6
1.1
8.4
8.0
1.8
10.1
p2.5
1.4
5.9
8.1
2.5
6.7
1.6
4.4
0.7
2.6
1.1
1.1
10.1
7.0
2.3
3.5
13.5
7.0
3.7
2.7
2.6
0.0
0.0
15.2
7.5
2.5
1.5
0.4
1.4
1.6
2.2
1.6
1.1
8.4
8.0
1.8
10.1
p50
14.3
12.1
15.1
14.2
7.4
12.9
18.9
2.2
6.9
2.5
2.4
13.6
12.0
3.7
10.0
13.6
9.8
10.8
13.7
5.6
0.0
0.0
15.8
11.3
13.0
8.8
6.6
7.2
6.6
4.4
1.6
1.2
9.2
15.6
2.2
12.4
P97.5
18.2
17.3
19.6
19.3
19.9
16.3
19.3
16.5
19.1
12.0
11.0
17.6
16.7
19.6
15.4
13.7
12.6
14.4
19.9
9.3
0.0
0.0
16.4
15.1
13.0
19.9
12.0
14.0
10.0
14.1
1.6
2.3
14.0
18.1
6.8
15.1
p100
18.2
17.3
19.6
19.3
19.9
16.3
19.3
16.5
19.1
12.0
11.0
17.6
16.7
19.6
15.4
13.7
12.6
14.4
19.9
9.3
0.0
0.0
16.4
15.1
13.0
19.9
12.0
14.0
10.0
14.1
1.6
2.3
14.0
18.1
6.8
15.1
A-67
-------
Monitor ID
132450003
150030010
150030011
150031001
150031006
160050004
160290003
160290031
160770011
170010006
170190004
170310050
170310059
170310063
170310064
170310076
170311018
170311601
170312001
170314002
170314201
170318003
170436001
170990007
171150013
171170002
171190008
171190017
171191010
171193007
171193009
171430024
171570001
171610003
171630010
171631010
n
15
7
7
3
7
2
13
13
2
4
3
47
40
23
50
36
26
12
43
25
4
36
12
4
11
0
15
40
28
28
26
10
2
10
30
30
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
1335
2231
2231
1043
2231
804
967
967
804
965
121
900
910
1041
1015
930
924
3807
920
982
165
835
2986
890
1251
4510
877
954
2595
2789
7333
13148
945
445
445
std
2379
2339
2339
1509
2339
606
2904
2904
606
614
182
1775
1928
1800
1902
1976
1721
5540
1807
1738
230
1797
5690
1527
2596
11972
2339
2564
8875
9193
11752
18554
1612
1152
1152
min
8
79
79
6
79
376
7
7
376
392
10
5
5
5
5
5
5
7
5
5
7
5
6
6
22
6
6
6
6
6
5
28
7
6
6
p2.5
8
79
79
6
79
376
7
7
376
392
10
5
5
5
5
5
5
7
5
5
7
5
6
6
22
6
6
6
6
6
5
28
7
6
6
p50
545
1566
1566
350
1566
804
33
33
804
817
21
65
65
17
51
26
16
1090
64
17
77
70
17
189
164
111
117
183
214
247
67
13148
169
68
68
P97.5
8275
6978
6978
2774
6978
1233
10544
10544
1233
1834
331
5951
7381
6229
6229
8443
6229
15934
6229
6229
498
8443
15934
3178
8032
45960
9663
12063
45960
45960
35748
26268
4963
6250
6250
p100
8275
6978
6978
2774
6978
1233
10544
10544
1233
1834
331
8443
8443
6229
8443
8443
6229
15934
8443
6229
498
8443
15934
3178
8032
45960
12063
12063
45960
45960
35748
26268
4963
6250
6250
Distance of monitor to SO2 emission source (km)1
mean
8.0
5.0
5.7
10.1
6.1
1.3
2.9
1.4
0.8
4.8
1.8
11.0
7.5
11.0
14.6
13.2
10.7
14.1
16.5
9.0
18.0
8.8
16.5
7.2
3.3
0.0
10.1
9.5
10.5
12.2
12.4
13.2
6.4
11.4
9.3
10.4
std
1.4
4.6
5.3
8.2
4.8
0.1
0.4
1.1
0.1
4.5
0.9
5.0
5.2
6.8
4.1
4.0
6.8
6.4
3.2
3.1
3.0
2.4
5.1
6.1
2.3
0.0
3.7
6.6
6.8
6.9
7.5
5.8
0.4
5.6
4.1
4.3
min
4.7
2.5
2.2
0.7
1.8
1.2
2.7
0.8
0.7
1.9
0.8
2.1
1.5
0.9
3.9
4.9
0.5
4.0
3.4
3.9
13.4
2.9
1.5
0.5
1.8
0.0
3.2
0.5
0.7
2.4
2.9
1.3
6.0
2.3
1.3
1.1
p2.5
4.7
2.5
2.2
0.7
1.8
1.2
2.7
0.8
0.7
1.9
0.8
3.4
1.5
0.9
6.4
4.9
0.5
4.0
8.4
3.9
13.4
2.9
1.5
0.5
1.8
0.0
3.2
0.7
0.7
2.4
2.9
1.3
6.0
2.3
1.3
1.1
p50
8.2
3.3
4.1
13.8
5.0
1.3
2.8
1.2
0.8
2.9
2.3
10.2
5.8
9.3
16.4
13.3
11.6
18.5
17.7
9.5
19.4
8.4
18.1
6.7
3.2
0.0
9.5
11.2
15.6
16.1
14.7
15.5
6.4
12.3
9.6
11.7
P97.5
10.0
15.3
17.5
15.7
16.5
1.4
4.3
4.9
0.8
11.5
2.4
19.7
19.3
19.7
19.9
19.7
19.8
19.3
19.3
18.5
19.7
14.7
19.8
14.8
9.9
0.0
19.7
19.0
18.4
18.9
19.8
18.8
6.7
17.2
18.5
19.4
p100
10.0
15.3
17.5
15.7
16.5
1.4
4.3
4.9
0.8
11.5
2.4
19.8
19.5
19.7
19.9
19.7
19.8
19.3
19.9
18.5
19.7
14.7
19.8
14.8
9.9
0.0
19.7
19.6
18.4
18.9
19.8
18.8
6.7
17.2
18.5
19.4
A-68
-------
Monitor ID
171631011
171670006
171790004
171850001
171851001
171970013
180270002
180290004
180430004
180430007
180431004
180450001
180510001
180510002
180630001
180630002
180630003
180730002
180730003
180770004
180890022
180892008
180910005
180910007
180970042
180970054
180970057
180970072
180970073
181091001
181230006
181230007
181250005
181270011
181270017
181270023
n
2
5
6
3
3
19
6
7
8
10
9
3
3
3
0
1
0
4
4
2
50
39
3
2
22
20
20
21
20
3
8
8
6
23
22
21
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
13148
2170
12212
42452
42452
2439
10869
21579
6500
6721
7442
18552
42452
42452
147
6874
6874
19099
1014
938
4166
4599
2358
2554
2554
2433
2547
6006
7033
7033
10869
1703
1363
1427
std
18554
3169
13311
25439
25439
6269
16456
32930
10778
10131
10470
32099
25439
25439
1422
1422
1297
1502
1945
4640
6476
6820
7138
7138
6980
7141
9709
17145
17145
16456
2266
1612
1623
min
28
9
22
27097
27097
6
9
174
12
12
12
10
27097
27097
147
6085
6085
18182
5
5
20
20
5
5
5
5
5
242
7
7
9
20
20
23
p2.5
28
9
22
27097
27097
6
9
174
12
12
12
10
27097
27097
147
6085
6085
18182
6
5
20
20
5
5
5
5
5
242
7
7
9
20
20
23
p50
13148
202
10290
28443
28443
37
2241
1574
484
516
798
28
28443
28443
147
6204
6204
19099
188
72
3301
4599
36
23
23
19
18
561
38
38
2241
1062
1029
1062
P97.5
26268
7210
35748
71817
71817
25224
41536
85699
23995
23995
23995
55617
71817
71817
147
9002
9002
20016
5951
8443
9178
9178
30896
30896
30896
30896
30896
17216
49028
49028
41536
9178
6318
6318
p100
26268
7210
35748
71817
71817
25224
41536
85699
23995
23995
23995
55617
71817
71817
147
9002
9002
20016
6318
8443
9178
9178
30896
30896
30896
30896
30896
17216
49028
49028
41536
9178
6318
6318
Distance of monitor to SO2 emission source (km)1
mean
4.0
7.3
5.4
2.9
5.9
6.6
6.3
4.2
13.4
9.2
10.0
9.8
2.0
2.9
0.0
19.2
0.0
4.3
10.2
4.3
14.1
6.4
9.1
6.0
11.2
3.3
4.2
6.9
13.7
4.3
7.7
6.8
3.0
6.7
5.4
4.1
std
0.6
3.6
5.2
0.1
0.1
4.8
0.6
4.1
3.1
6.4
3.5
8.7
0.1
0.0
0.0
0.0
0.0
1.0
1.2
0.1
4.0
4.1
9.7
0.8
2.9
2.3
2.0
3.5
2.3
2.4
4.3
4.2
4.7
6.2
5.7
4.4
min
3.6
4.9
0.8
2.8
5.8
1.1
5.8
1.2
8.8
1.1
5.0
4.5
1.8
2.9
0.0
19.2
0.0
3.5
9.5
4.3
0.8
1.6
0.4
5.4
7.8
0.9
0.9
0.8
6.2
2.1
2.8
2.1
0.9
2.2
2.0
1.1
p2.5
3.6
4.9
0.8
2.8
5.8
1.1
5.8
1.2
8.8
1.1
5.0
4.5
1.8
2.9
0.0
19.2
0.0
3.5
9.5
4.3
1.8
1.6
0.4
5.4
7.8
0.9
0.9
0.8
6.2
2.1
2.8
2.1
0.9
2.2
2.0
1.1
p50
4.0
5.6
3.6
2.9
5.8
5.2
6.0
3.4
12.3
7.3
9.8
5.1
2.0
2.9
0.0
19.2
0.0
4.0
9.7
4.3
14.6
5.6
7.3
6.0
11.0
2.4
4.3
6.6
14.5
4.0
7.0
5.7
1.1
3.6
2.6
2.4
P97.5
4.4
13.5
13.8
3.1
6.0
18.6
7.3
12.8
17.7
19.9
14.7
19.8
2.1
3.0
0.0
19.2
0.0
5.8
12.1
4.4
19.8
17.6
19.6
6.5
17.0
9.2
9.8
18.7
15.3
6.9
14.3
13.1
12.7
18.7
17.8
14.6
p100
4.4
13.5
13.8
3.1
6.0
18.6
7.3
12.8
17.7
19.9
14.7
19.8
2.1
3.0
0.0
19.2
0.0
5.8
12.1
4.4
19.9
17.6
19.6
6.5
17.0
9.2
9.8
18.7
15.3
6.9
14.3
13.1
12.7
18.7
17.8
14.6
A-69
-------
Monitor ID
181470002
181470010
181530004
181630012
181631002
181670018
181671014
181730002
181731001
181770006
181770007
190330018
190450018
190450019
190450020
191110006
191111007
191130028
191130029
191130031
191130032
191130034
191130038
191130039
191390016
191390017
191390020
191630015
191630017
191770004
191770005
191770006
191930018
201070002
201250006
201450001
n
7
4
3
5
5
6
6
8
8
2
2
4
2
2
2
1
2
7
7
7
7
7
7
7
5
4
4
7
7
0
0
0
4
0
4
0
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
15627
15099
9270
1806
1806
10842
10842
13636
13636
6446
6446
2684
4694
4694
4694
29
104
2200
2200
2200
2200
2200
2200
2200
6227
7763
7763
1345
2120
9208
468
std
24405
25616
8089
2589
2589
25028
25028
16457
16457
9089
9089
3305
839
839
839
105
2428
2428
2428
2428
2428
2428
2428
6934
6956
6956
1810
3515
10818
464
min
7
20
10
5
5
12
12
50
50
19
19
20
4101
4101
4101
29
29
12
12
12
12
12
12
12
83
463
463
17
17
15
11
p2.5
7
20
10
5
5
12
12
50
50
19
19
20
4101
4101
4101
29
29
12
12
12
12
12
12
12
83
463
463
17
17
15
11
p50
66
3589
12846
382
382
417
417
3559
3559
6446
6446
1934
4694
4694
4694
29
104
1954
1954
1954
1954
1954
1954
1954
3790
7345
7345
336
303
7845
428
P97.5
53196
53196
14955
6004
6004
61901
61901
41049
41049
12873
12873
6850
5287
5287
5287
29
179
5480
5480
5480
5480
5480
5480
5480
15901
15901
15901
4963
8983
21127
1006
p100
53196
53196
14955
6004
6004
61901
61901
41049
41049
12873
12873
6850
5287
5287
5287
29
179
5480
5480
5480
5480
5480
5480
5480
15901
15901
15901
4963
8983
21127
1006
Distance of monitor to SO2 emission source (km)1
mean
13.0
12.3
12.1
13.1
8.5
6.8
6.8
2.9
3.0
2.1
3.2
3.9
1.4
1.3
3.4
3.7
13.3
5.8
3.8
4.3
3.5
3.6
3.9
4.6
8.7
3.8
4.9
9.5
9.6
0.0
0.0
0.0
6.4
0.0
5.8
0.0
std
3.6
6.6
6.4
7.7
5.3
3.6
5.9
0.4
0.5
1.4
2.5
3.7
0.9
1.0
0.6
0.0
6.3
2.4
3.1
3.2
2.7
2.5
1.9
3.0
6.9
3.6
4.4
5.0
4.2
0.0
0.0
0.0
4.3
0.0
9.3
0.0
min
8.0
3.3
4.8
3.1
3.4
5.0
1.9
2.5
2.5
1.1
1.4
0.4
0.7
0.6
3.0
3.7
8.8
2.8
0.5
0.5
0.6
0.2
0.6
1.1
2.4
0.6
0.9
1.1
1.1
0.0
0.0
0.0
0.7
0.0
0.5
0.0
p2.5
8.0
3.3
4.8
3.1
3.4
5.0
1.9
2.5
2.5
1.1
1.4
0.4
0.7
0.6
3.0
3.7
8.8
2.8
0.5
0.5
0.6
0.2
0.6
1.1
2.4
0.6
0.9
1.1
1.1
0.0
0.0
0.0
0.7
0.0
0.5
0.0
p50
15.0
14.0
14.7
18.0
9.5
5.5
5.5
3.0
2.9
2.1
3.2
3.2
1.4
1.3
3.4
3.7
13.3
6.7
4.0
4.7
3.1
2.9
4.2
4.2
7.4
3.1
4.0
11.7
11.2
0.0
0.0
0.0
7.1
0.0
1.6
0.0
P97.5
16.6
17.9
16.8
19.6
16.5
14.1
17.3
3.3
3.7
3.1
5.0
8.8
2.0
2.0
3.8
3.7
17.7
8.8
9.2
9.3
8.8
7.4
6.2
10.3
19.2
8.5
10.4
15.1
13.6
0.0
0.0
0.0
10.7
0.0
19.7
0.0
p100
16.6
17.9
16.8
19.6
16.5
14.1
17.3
3.3
3.7
3.1
5.0
8.8
2.0
2.0
3.8
3.7
17.7
8.8
9.2
9.3
8.8
7.4
6.2
10.3
19.2
8.5
10.4
15.1
13.6
0.0
0.0
0.0
10.7
0.0
19.7
0.0
A-70
-------
Monitor ID
201730010
201910002
201950001
202090001
202090020
202090021
210190015
210190017
210191003
210370003
210371001
210590005
210670012
210890007
210910012
211010013
211010014
211110032
211110051
211111041
211390004
211450001
211451024
211451026
212270008
220150008
220190008
220330009
220730004
220870002
221210001
230010011
230030009
230030012
230031003
230031013
n
3
3
0
14
14
13
9
10
8
11
11
4
3
5
9
4
10
14
12
11
4
7
3
3
1
2
16
28
1
18
28
9
1
1
1
3
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
269
269
1388
1388
1494
1323
1193
1271
6817
465
15241
209
961
12162
2256
10948
6208
3259
6268
444
8769
587
587
52
77
3352
1406
2166
419
1116
31
90
90
90
16
std
448
448
2341
2341
2402
2058
1983
2194
20950
664
25506
316
1147
22226
2755
15581
8948
5326
9779
869
13010
1005
1005
21
5531
3913
846
3650
41
17
min
6
6
6
6
6
25
25
25
12
12
26
12
25
7
5
5
38
38
12
6
174
6
6
52
62
6
6
2166
8
6
5
90
90
90
5
p2.5
6
6
6
6
6
25
25
25
12
12
26
12
25
7
5
5
38
38
12
6
174
6
6
52
62
6
6
2166
8
6
5
90
90
90
5
p50
15
15
34
34
40
401
343
343
268
213
3871
42
401
38
1508
2980
516
168
234
11
7435
7
7
52
77
184
45
2166
52
33
23
90
90
90
7
P97.5
785
785
7625
7625
7625
6285
6285
6285
69953
1848
53196
573
2589
53196
6004
41049
23995
14977
23995
1747
37077
1747
1747
52
91
18851
18680
2166
3009
18680
140
90
90
90
36
p100
785
785
7625
7625
7625
6285
6285
6285
69953
1848
53196
573
2589
53196
6004
41049
23995
14977
23995
1747
37077
1747
1747
52
91
18851
18680
2166
3009
18680
140
90
90
90
36
Distance of monitor to SO2 emission source (km)1
mean
11.4
16.3
0.0
9.2
9.0
8.6
12.3
12.8
9.3
12.0
8.5
7.4
3.2
10.9
10.4
10.2
12.9
11.4
10.6
9.1
8.0
7.5
18.2
15.3
19.1
8.7
7.6
5.8
10.1
8.8
5.4
6.6
1.9
1.0
1.3
4.7
std
3.1
2.4
0.0
5.9
6.1
5.5
5.5
5.4
5.4
3.0
3.4
6.8
2.2
6.2
5.1
5.5
1.4
5.6
7.5
7.3
6.9
3.4
2.1
2.2
0.0
0.1
6.1
5.6
0.0
4.2
4.7
4.3
0.0
0.0
0.0
4.5
min
9.0
13.6
0.0
3.5
0.6
3.4
1.6
2.9
1.3
8.1
4.2
2.2
1.2
5.1
1.2
2.0
11.5
2.4
1.6
1.3
3.1
2.0
15.8
12.7
19.1
8.6
1.2
1.5
10.1
0.5
2.4
1.3
1.9
1.0
1.3
1.4
p2.5
9.0
13.6
0.0
3.5
0.6
3.4
1.6
2.9
1.3
8.1
4.2
2.2
1.2
5.1
1.2
2.0
11.5
2.4
1.6
1.3
3.1
2.0
15.8
12.7
19.1
8.6
1.2
1.5
10.1
0.5
2.4
1.3
1.9
1.0
1.3
1.4
p50
10.2
17.3
0.0
7.1
7.7
6.6
14.6
13.8
9.9
10.8
7.5
5.5
2.7
7.6
10.6
12.7
12.8
13.7
14.6
7.7
5.4
9.4
19.3
16.5
19.1
8.7
5.8
3.2
10.1
7.8
3.4
6.5
1.9
1.0
1.3
2.8
P97.5
14.9
18.0
0.0
19.8
18.9
19.1
17.7
19.5
15.4
17.8
15.5
16.5
5.6
19.8
18.9
13.3
16.6
18.3
18.7
19.3
17.9
11.2
19.5
16.7
19.1
8.8
16.7
20.0
10.1
19.0
18.1
13.3
1.9
1.0
1.3
9.9
p100
14.9
18.0
0.0
19.8
18.9
19.1
17.7
19.5
15.4
17.8
15.5
16.5
5.6
19.8
18.9
13.3
16.6
18.3
18.7
19.3
17.9
11.2
19.5
16.7
19.1
8.8
16.7
20.0
10.1
19.0
18.1
13.3
1.9
1.0
1.3
9.9
A-71
-------
Monitor ID
230031018
230050014
230050027
230172007
240010006
240032002
240053001
245100018
245100036
250051004
250090005
250091004
250091005
250095004
250130016
250131009
250154002
250171701
250174003
250250002
250250019
250250020
250250021
250250040
250250042
250251003
250270020
250270023
260410902
260490021
260492001
260810020
260991003
261130001
261470005
261530001
n
4
12
12
2
2
20
22
21
21
24
25
23
22
14
34
32
12
55
57
62
50
58
58
59
60
58
28
28
3
4
2
9
3
1
3
0
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
193
267
267
249
681
3247
4429
3101
4635
1867
65
878
917
88
216
65
72
139
127
129
156
138
137
135
133
380
25
25
1407
42
64
60
239
58
524
std
233
628
628
344
685
9622
11101
9402
11331
8085
148
3071
3137
197
907
148
113
678
663
639
710
660
660
654
649
1952
35
35
1264
24
79
96
287
431
min
7
5
5
6
197
5
5
5
5
6
6
5
5
8
5
5
6
5
5
5
5
5
5
5
5
5
6
6
671
7
7
9
10
58
31
p2.5
7
5
5
6
197
5
5
5
5
6
6
5
5
8
5
5
6
5
5
5
5
5
5
5
5
5
6
6
671
7
7
9
10
58
31
p50
133
16
16
249
681
21
27
22
22
31
26
16
16
25
14
13
29
15
13
14
14
15
14
14
14
15
12
12
685
48
64
12
148
58
715
P97.5
499
2091
2091
492
1166
39974
39974
39974
39974
39593
762
14132
14132
762
5282
671
363
640
460
640
640
640
640
640
640
5007
178
178
2867
63
120
280
560
58
826
p100
499
2091
2091
492
1166
39974
39974
39974
39974
39593
762
14132
14132
762
5282
671
363
5007
5007
5007
5007
5007
5007
5007
5007
14132
178
178
2867
63
120
280
560
58
826
Distance of monitor to SO2 emission source (km)1
mean
8.5
6.1
6.0
1.0
8.9
11.9
11.9
9.1
6.6
7.5
9.6
11.3
11.2
8.6
7.6
8.4
15.8
13.3
12.2
9.6
12.0
10.0
10.6
10.2
9.4
11.0
5.0
5.1
2.5
10.9
19.0
10.5
14.0
10.3
8.7
0.0
std
5.6
4.8
4.7
0.1
4.0
4.5
3.3
4.6
3.3
6.7
6.6
6.3
6.3
4.2
5.2
4.7
3.4
4.6
4.0
6.1
3.8
5.0
4.7
5.3
5.8
4.6
5.9
5.8
1.5
4.5
0.4
5.6
3.4
0.0
5.9
0.0
min
0.3
1.2
0.8
1.0
6.0
2.7
4.6
1.4
1.6
0.1
0.3
0.8
0.7
0.7
0.5
1.7
9.1
0.4
0.6
0.7
0.7
1.1
1.8
1.0
0.5
1.0
0.1
0.6
0.8
4.2
18.8
4.3
10.2
10.3
3.8
0.0
p2.5
0.3
1.2
0.8
1.0
6.0
2.7
4.6
1.4
1.6
0.1
0.3
0.8
0.7
0.7
0.5
1.7
9.1
2.9
5.6
1.1
4.2
3.0
3.4
1.4
0.7
2.1
0.1
0.6
0.8
4.2
18.8
4.3
10.2
10.3
3.8
0.0
p50
10.3
5.0
4.8
1.0
8.9
13.3
12.1
7.6
6.8
3.8
9.2
12.8
11.9
10.1
7.4
7.4
16.8
15.0
12.4
8.6
12.0
9.1
9.3
9.5
9.1
10.4
2.8
2.9
3.3
13.1
19.0
10.6
15.2
10.3
6.9
0.0
P97.5
13.0
16.8
16.6
1.1
11.7
19.9
19.2
16.7
16.0
18.9
19.9
20.0
18.6
14.7
19.2
18.9
19.7
19.4
19.5
19.5
18.1
19.2
19.5
19.5
19.1
19.3
19.5
19.1
3.4
13.1
19.3
19.4
16.7
10.3
15.2
0.0
p100
13.0
16.8
16.6
1.1
11.7
19.9
19.2
16.7
16.0
18.9
19.9
20.0
18.6
14.7
19.2
18.9
19.7
20.0
19.7
19.7
18.4
19.2
20.0
19.8
19.3
19.4
19.5
19.1
3.4
13.1
19.3
19.4
16.7
10.3
15.2
0.0
A-72
-------
Monitor ID
261630001
261630005
261630015
261630016
261630019
261630025
261630027
261630033
261630062
261630092
270031002
270176316
270370020
270370423
270370439
270370441
270370442
270530954
270530957
270711240
271230864
271410003
271410011
271410012
271410013
271630436
271710007
280470007
280490018
280590006
280810004
290210009
290210011
290470025
290770026
290770032
n
36
34
32
31
23
6
33
32
31
33
10
5
15
17
14
12
11
24
21
1
27
1
1
1
1
21
2
2
5
7
0
1
1
15
4
4
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
1780
1894
1070
1104
1358
13
1952
1070
1104
1952
1332
72
610
805
639
720
506
913
878
67
769
26742
26742
26742
26742
545
13397
12535
51
4903
3563
3563
1682
2302
2302
std
5390
5529
2436
2469
2828
14
5605
2436
2469
5605
4067
84
1015
1227
1047
1114
873
2729
2877
2540
997
18873
17718
45
10049
2364
2728
2728
min
5
5
5
5
10
5
5
5
5
5
5
5
9
9
9
9
9
5
5
67
5
26742
26742
26742
26742
7
52
6
15
12
3563
3563
6
5
5
p2.5
5
5
5
5
10
5
5
5
5
5
5
5
9
9
9
9
9
5
5
67
5
26742
26742
26742
26742
7
52
6
15
12
3563
3563
6
5
5
p50
109
117
117
121
121
9
121
117
121
121
11
26
104
205
79
79
54
48
12
67
46
26742
26742
26742
26742
104
13397
12535
30
96
3563
3563
105
1772
1772
P97.5
30171
30171
8913
8913
8913
42
30171
8913
8913
30171
12904
190
3071
3821
3071
3071
2869
12904
12904
67
12904
26742
26742
26742
26742
3821
26742
25064
128
27207
3563
3563
7625
5657
5657
p100
30171
30171
8913
8913
8913
42
30171
8913
8913
30171
12904
190
3071
3821
3071
3071
2869
12904
12904
67
12904
26742
26742
26742
26742
3821
26742
25064
128
27207
3563
3563
7625
5657
5657
Distance of monitor to SO2 emission source (km)1
mean
10.9
6.1
5.5
9.0
17.3
14.8
5.5
5.0
9.0
5.4
14.3
13.7
11.9
11.6
12.5
11.6
12.2
10.9
10.7
0.3
12.0
4.9
1.7
1.8
1.1
11.1
11.8
6.5
7.3
7.0
0.0
0.7
0.9
11.9
8.2
7.8
std
4.0
5.4
4.2
2.7
4.5
2.4
5.2
4.5
2.9
5.1
4.4
6.8
6.1
5.5
5.8
5.7
5.5
5.8
5.3
0.0
4.8
0.0
0.0
0.0
0.0
5.6
6.5
7.9
5.4
4.9
0.0
0.0
0.0
4.8
4.5
3.9
min
5.4
1.2
1.5
3.6
3.7
11.2
0.4
0.4
3.1
0.9
4.7
2.2
0.9
0.4
2.6
1.6
2.3
0.6
0.9
0.3
3.9
4.9
1.7
1.8
1.1
0.9
7.2
0.9
3.2
3.3
0.0
0.7
0.9
2.8
2.3
3.0
p2.5
5.4
1.2
1.5
3.6
3.7
11.2
0.4
0.4
3.1
0.9
4.7
2.2
0.9
0.4
2.6
1.6
2.3
0.6
0.9
0.3
3.9
4.9
1.7
1.8
1.1
0.9
7.2
0.9
3.2
3.3
0.0
0.7
0.9
2.8
2.3
3.0
p50
9.6
4.4
3.8
8.6
18.9
15.2
3.9
4.2
8.5
3.0
15.5
16.4
12.4
12.4
13.1
12.6
13.8
12.2
10.9
0.3
12.6
4.9
1.7
1.8
1.1
11.4
11.8
6.5
6.0
5.4
0.0
0.7
0.9
10.8
9.3
8.5
P97.5
20.0
19.0
17.9
17.0
19.8
17.8
19.7
15.8
17.2
19.9
18.9
19.7
19.6
18.8
20.0
19.0
18.8
19.0
18.3
0.3
19.7
4.9
1.7
1.8
1.1
18.4
16.3
12.1
16.6
17.3
0.0
0.7
0.9
18.2
11.8
11.0
p100
20.0
19.0
17.9
17.0
19.8
17.8
19.7
15.8
17.2
19.9
18.9
19.7
19.6
18.8
20.0
19.0
18.8
19.0
18.3
0.3
19.7
4.9
1.7
1.8
1.1
18.4
16.3
12.1
16.6
17.3
0.0
0.7
0.9
18.2
11.8
11.0
A-73
-------
Monitor ID
290770037
290770040
290770041
290930030
290930031
290950034
290990004
290990014
290990017
290990018
291370001
291630002
291650023
291830010
291831002
291890001
291890004
291890006
291890014
291893001
291895001
291897002
291897003
295100007
295100072
295100080
295100086
300132000
300132001
300430903
300430911
300430913
300490702
300490703
300870700
300870701
n
4
4
4
1
1
14
5
5
5
4
0
2
4
1
15
14
9
7
8
29
35
14
18
19
30
34
32
2
2
1
1
1
1
1
1
1
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
2302
2302
2302
43340
43340
1388
11145
11145
11145
8117
6747
2757
47610
4516
1748
2535
27
33
370
1911
50
403
1312
445
397
421
351
351
234
234
234
234
234
16735
16735
std
2728
2728
2728
2341
10277
10277
10277
8927
934
3602
11970
4547
5610
48
47
1164
7823
75
1461
3936
1152
1088
1118
481
481
min
5
5
5
43340
43340
6
243
243
243
243
6087
19
47610
6
8
8
6
6
6
6
6
6
8
6
6
6
11
11
234
234
234
234
234
16735
16735
p2.5
5
5
5
43340
43340
6
243
243
243
243
6087
19
47610
6
8
8
6
6
6
6
6
6
8
6
6
6
11
11
234
234
234
234
234
16735
16735
p50
1772
1772
1772
43340
43340
34
15223
15223
15223
7889
6747
1693
47610
136
35
13
8
10
60
111
16
37
50
68
61
68
351
351
234
234
234
234
234
16735
16735
P97.5
5657
5657
5657
43340
43340
7625
23258
23258
23258
16447
7408
7625
47610
45960
16447
16447
136
136
6250
45960
277
6250
16447
6250
6250
6250
691
691
234
234
234
234
234
16735
16735
p100
5657
5657
5657
43340
43340
7625
23258
23258
23258
16447
7408
7625
47610
45960
16447
16447
136
136
6250
45960
277
6250
16447
6250
6250
6250
691
691
234
234
234
234
234
16735
16735
Distance of monitor to SO2 emission source (km)1
mean
9.2
9.2
8.6
1.7
4.6
8.7
9.7
9.8
10.2
8.3
0.0
7.3
17.8
1.7
12.6
14.8
14.0
14.7
11.9
15.2
15.1
13.2
14.1
12.5
8.8
10.7
9.8
4.1
4.1
3.3
4.5
4.9
6.2
7.3
19.8
19.0
std
6.1
6.2
5.5
0.0
0.0
4.9
7.4
7.4
7.1
6.6
0.0
6.6
1.3
0.0
3.4
4.4
3.1
4.6
6.2
4.2
3.1
5.7
5.4
6.0
3.8
4.3
3.9
3.6
4.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
min
0.6
0.5
1.2
1.7
4.6
1.4
0.2
0.7
1.6
1.4
0.0
2.7
16.0
1.7
4.3
6.4
9.8
8.4
3.2
5.1
6.7
3.9
3.5
0.5
2.0
0.4
1.7
1.5
0.7
3.3
4.5
4.9
6.2
7.3
19.8
19.0
p2.5
0.6
0.5
1.2
1.7
4.6
1.4
0.2
0.7
1.6
1.4
0.0
2.7
16.0
1.7
4.3
6.4
9.8
8.4
3.2
5.1
6.7
3.9
3.5
0.5
2.0
0.4
1.7
1.5
0.7
3.3
4.5
4.9
6.2
7.3
19.8
19.0
p50
11.0
11.0
9.7
1.7
4.6
8.1
11.4
11.9
10.6
8.2
0.0
7.3
18.1
1.7
13.5
15.9
15.2
15.7
11.3
16.0
15.9
14.6
16.2
14.0
9.7
10.5
10.0
4.1
4.1
3.3
4.5
4.9
6.2
7.3
19.8
19.0
P97.5
14.0
14.1
13.8
1.7
4.6
15.4
17.1
17.5
17.3
15.3
0.0
12.0
19.1
1.7
17.3
19.7
18.2
19.9
19.7
20.0
20.0
20.0
19.4
19.6
19.2
19.7
18.6
6.7
7.5
3.3
4.5
4.9
6.2
7.3
19.8
19.0
p100
14.0
14.1
13.8
1.7
4.6
15.4
17.1
17.5
17.3
15.3
0.0
12.0
19.1
1.7
17.3
19.7
18.2
19.9
19.7
20.0
20.0
20.0
19.4
19.6
19.2
19.7
18.6
6.7
7.5
3.3
4.5
4.9
6.2
7.3
19.8
19.0
A-74
-------
Monitor ID
300870702
300870760
300870761
300870762
300870763
301110016
301110066
301110079
301110080
301110082
301110083
301110084
301111065
301112005
301112006
301112007
310550048
310550050
310550053
310550055
320030022
320030078
320030539
320030601
330050007
330070019
330070022
330071007
330110016
330110019
330110020
330111009
330111010
330130007
330131003
330131006
n
1
0
0
0
1
0
4
4
4
4
4
6
4
4
6
6
5
5
5
3
4
0
0
0
1
1
1
2
3
3
3
11
16
4
4
4
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
16735
16735
1370
1370
1370
1370
1370
2550
1370
1370
2550
2550
6370
6370
6370
3845
45
81
638
638
9
10269
10269
10269
41
48
7708
7708
7708
std
1322
1322
1322
1322
1322
2627
1322
1322
2627
2627
9218
9218
9218
6637
27
4
10386
10386
10386
42
42
9906
9906
9906
min
16735
16735
75
75
75
75
75
75
75
75
75
75
6
6
6
6
16
81
638
638
6
149
149
149
6
6
41
41
41
p2.5
16735
16735
75
75
75
75
75
75
75
75
75
75
6
6
6
6
16
81
638
638
6
149
149
149
6
6
41
41
41
p50
16735
16735
1135
1135
1135
1135
1135
1976
1135
1135
1976
1976
58
58
58
20
44
81
638
638
9
9754
9754
9754
20
38
4945
4945
4945
P97.5
16735
16735
3135
3135
3135
3135
3135
7415
3135
3135
7415
7415
20257
20257
20257
11509
75
81
638
638
12
20902
20902
20902
149
149
20902
20902
20902
p100
16735
16735
3135
3135
3135
3135
3135
7415
3135
3135
7415
7415
20257
20257
20257
11509
75
81
638
638
12
20902
20902
20902
149
149
20902
20902
20902
Distance of monitor to SO2 emission source (km)1
mean
19.8
0.0
0.0
0.0
15.2
0.0
3.1
7.8
2.4
3.4
3.4
10.3
4.7
4.1
10.1
11.4
12.7
13.4
11.3
13.0
3.9
0.0
0.0
0.0
0.3
1.7
2.3
0.6
17.3
17.0
17.0
12.7
13.0
7.3
7.7
8.2
std
0.0
0.0
0.0
0.0
0.0
0.0
0.5
3.0
1.8
2.7
0.7
6.6
2.7
1.8
7.2
3.9
7.5
7.5
5.7
7.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
1.3
2.3
2.3
6.0
3.0
3.9
5.4
8.8
min
19.8
0.0
0.0
0.0
15.2
0.0
2.6
5.8
0.9
1.7
2.7
3.1
0.7
1.5
1.1
4.7
0.5
1.0
3.3
4.7
3.8
0.0
0.0
0.0
0.3
1.7
2.3
0.6
16.5
15.7
15.7
4.4
7.2
1.4
4.0
1.3
p2.5
19.8
0.0
0.0
0.0
15.2
0.0
2.6
5.8
0.9
1.7
2.7
3.1
0.7
1.5
1.1
4.7
0.5
1.0
3.3
4.7
3.8
0.0
0.0
0.0
0.3
1.7
2.3
0.6
16.5
15.7
15.7
4.4
7.2
1.4
4.0
1.3
p50
19.8
0.0
0.0
0.0
15.2
0.0
3.1
6.7
1.9
2.3
3.4
7.4
5.7
4.6
7.6
11.2
13.6
14.7
10.6
16.1
3.9
0.0
0.0
0.0
0.3
1.7
2.3
0.6
16.6
15.7
15.7
14.7
12.0
9.0
5.6
5.8
P97.5
19.8
0.0
0.0
0.0
15.2
0.0
3.7
12.2
5.0
7.3
4.4
18.6
6.7
5.7
18.8
15.3
19.3
19.6
18.0
18.2
3.9
0.0
0.0
0.0
0.3
1.7
2.3
0.7
18.8
19.6
19.6
19.0
19.0
9.6
15.4
19.8
p100
19.8
0.0
0.0
0.0
15.2
0.0
3.7
12.2
5.0
7.3
4.4
18.6
6.7
5.7
18.8
15.3
19.3
19.6
18.0
18.2
3.9
0.0
0.0
0.0
0.3
1.7
2.3
0.7
18.8
19.6
19.6
19.0
19.0
9.6
15.4
19.8
A-75
-------
Monitor ID
330131007
330150009
330150014
330150015
330190003
340010005
340035001
340051001
340070003
340071001
340110007
340130011
340130016
340150002
340170006
340171002
340232003
340273001
340390003
340390004
350130008
350130017
350151004
350170001
350171003
350230005
350450008
350450009
350450017
350451005
360010012
360050073
360050080
360050083
360050110
360130005
n
4
9
9
9
2
0
61
21
60
2
4
59
61
50
59
71
21
2
38
38
1
13
4
1
1
0
7
2
0
8
9
68
66
56
67
0
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
7708
1523
1523
1523
110
457
719
179
8
161
465
453
529
467
421
80
19
610
609
37
44
1058
263
263
2478
293
6274
40
399
406
119
402
std
9906
2990
2990
2990
81
2442
3104
644
1
198
2471
2431
1281
2471
2267
206
8
3074
3075
92
973
2496
378
10983
46
2309
2344
355
2326
min
41
6
6
6
53
6
5
5
8
28
5
5
5
5
5
6
13
5
5
37
5
168
263
263
11
25
11
7
5
5
6
5
p2.5
41
6
6
6
53
6
5
5
8
28
6
6
6
5
5
6
13
5
5
37
5
168
263
263
11
25
11
7
6
6
6
6
p50
4945
52
52
52
110
22
35
25
8
81
25
25
44
25
18
16
19
19
19
37
11
983
263
263
2554
293
2630
20
22
18
19
21
P97.5
20902
8057
8057
8057
168
2302
14266
2378
9
456
1845
1845
4450
1845
2302
958
25
18958
18958
37
345
2099
263
263
5919
560
32847
153
2302
2302
1129
2302
p100
20902
8057
8057
8057
168
18958
14266
4450
9
456
18958
18958
6720
18958
18958
958
25
18958
18958
37
345
2099
263
263
5919
560
32847
153
18958
18958
2302
18958
Distance of monitor to SO2 emission source (km)1
mean
9.3
9.0
9.6
8.9
2.5
0.0
14.8
10.7
9.7
10.2
7.5
13.1
13.4
13.2
13.0
11.9
8.6
17.7
11.5
11.2
17.9
14.8
8.6
6.1
1.5
0.0
17.2
3.3
0.0
6.1
10.8
10.0
10.6
11.2
10.1
0.0
std
2.1
6.9
7.0
7.1
1.7
0.0
3.7
6.7
3.4
0.5
6.6
4.9
5.0
3.7
4.6
5.0
4.6
3.1
5.3
5.6
0.0
4.0
8.4
0.0
0.0
0.0
3.5
2.0
0.0
3.8
5.2
4.9
5.0
5.6
4.9
0.0
min
7.5
2.0
1.0
1.9
1.3
0.0
2.2
1.5
2.0
9.9
1.8
1.6
1.8
2.1
2.0
0.8
1.8
15.5
2.3
0.7
17.9
1.7
0.9
6.1
1.5
0.0
11.9
2.0
0.0
3.2
3.5
3.4
1.8
1.6
2.7
0.0
p2.5
7.5
2.0
1.0
1.9
1.3
0.0
5.2
1.5
2.8
9.9
1.8
2.2
2.7
4.6
3.2
0.8
1.8
15.5
2.3
0.7
17.9
1.7
0.9
6.1
1.5
0.0
11.9
2.0
0.0
3.2
3.5
3.4
3.0
1.8
2.8
0.0
p50
8.6
4.4
5.5
4.1
2.5
0.0
15.7
12.3
9.6
10.2
5.7
14.2
14.3
12.9
13.5
11.6
9.2
17.7
12.4
12.1
17.9
15.7
8.7
6.1
1.5
0.0
19.2
3.3
0.0
3.5
9.0
9.1
9.6
11.3
9.0
0.0
P97.5
12.3
19.2
19.9
19.5
3.7
0.0
19.7
19.9
17.2
10.5
16.8
19.2
19.8
19.2
19.9
19.7
15.8
19.8
20.0
19.9
17.9
17.7
16.1
6.1
1.5
0.0
19.3
4.7
0.0
11.9
18.0
19.2
19.5
19.6
19.2
0.0
p100
12.3
19.2
19.9
19.5
3.7
0.0
19.9
19.9
19.9
10.5
16.8
19.4
19.9
19.7
19.9
19.8
15.8
19.8
20.0
19.9
17.9
17.7
16.1
6.1
1.5
0.0
19.3
4.7
0.0
11.9
18.0
19.7
19.9
19.6
19.7
0.0
A-76
-------
Monitor ID
360130006
360130011
360150003
360290005
360294002
360298001
360310003
360330004
360410005
360430005
360470011
360470076
360530006
360551004
360551007
360556001
360590005
360610010
360610056
360632008
360671015
360790005
360810097
360810124
360830004
360831005
360850067
360930003
361030002
361030009
361111005
370030003
370130003
370130004
370130006
370370004
n
1
0
2
10
16
9
0
0
0
0
77
67
0
4
4
4
12
77
76
13
4
0
60
66
3
2
48
4
9
10
0
0
1
1
1
4
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
52177
202
4073
2608
4518
377
428
12595
12595
12595
151
375
382
3134
820
136
122
126
94
515
24
156
734
4730
4730
4730
119
std
270
12273
9706
12932
2178
2333
14519
14519
14519
301
2178
2192
10777
1602
358
342
106
124
2737
26
344
2013
71
min
52177
11
8
8
8
5
5
8
8
8
6
5
5
8
8
5
5
10
6
5
6
6
11
4730
4730
4730
12
p2.5
52177
11
8
8
8
5
5
8
8
8
6
5
5
8
8
6
6
10
6
6
6
6
11
4730
4730
4730
12
p50
52177
202
182
166
247
18
17
11988
11988
11988
26
17
18
118
24
22
21
153
94
17
14
19
42
4730
4730
4730
148
P97.5
52177
393
38999
38999
38999
2302
2302
26395
26395
26395
1057
2302
2302
38999
3223
1129
1129
217
182
1845
62
1057
6453
4730
4730
4730
165
p100
52177
393
38999
38999
38999
18958
18958
26395
26395
26395
1057
18958
18958
38999
3223
2302
2302
217
182
18958
62
1057
6453
4730
4730
4730
165
Distance of monitor to SO2 emission source (km)1
mean
2.0
0.0
10.2
10.2
10.4
13.5
0.0
0.0
0.0
0.0
10.3
11.6
0.0
11.0
11.3
10.5
11.8
10.4
9.9
9.3
5.9
0.0
14.8
12.5
18.4
17.6
14.0
9.5
9.3
11.3
0.0
0.0
2.2
2.7
1.1
17.2
std
0.0
0.0
13.6
4.7
6.2
5.6
0.0
0.0
0.0
0.0
5.5
4.8
0.0
4.2
4.1
6.8
4.8
5.4
5.4
7.3
4.3
0.0
4.0
4.0
1.8
1.6
4.0
6.6
5.8
5.7
0.0
0.0
0.0
0.0
0.0
3.7
min
2.0
0.0
0.6
2.5
1.6
4.6
0.0
0.0
0.0
0.0
0.7
2.3
0.0
7.6
6.4
5.2
1.9
0.3
0.3
0.3
1.9
0.0
2.9
2.1
16.3
16.5
5.5
2.0
1.9
2.0
0.0
0.0
2.2
2.7
1.1
11.8
p2.5
2.0
0.0
0.6
2.5
1.6
4.6
0.0
0.0
0.0
0.0
1.9
3.1
0.0
7.6
6.4
5.2
1.9
1.4
1.4
0.3
1.9
0.0
5.0
2.3
16.3
16.5
6.2
2.0
1.9
2.0
0.0
0.0
2.2
2.7
1.1
11.8
p50
2.0
0.0
10.2
11.1
12.3
14.7
0.0
0.0
0.0
0.0
10.8
11.5
0.0
10.0
11.9
8.5
11.8
11.1
10.6
12.2
5.2
0.0
15.5
12.4
19.3
17.6
14.2
9.7
7.3
11.9
0.0
0.0
2.2
2.7
1.1
18.6
P97.5
2.0
0.0
19.9
15.4
18.3
19.0
0.0
0.0
0.0
0.0
19.2
19.4
0.0
16.5
15.0
19.8
19.1
19.4
19.9
19.8
11.5
0.0
19.9
19.5
19.6
18.8
19.6
16.5
18.2
19.3
0.0
0.0
2.2
2.7
1.1
19.9
p100
2.0
0.0
19.9
15.4
18.3
19.0
0.0
0.0
0.0
0.0
19.7
19.9
0.0
16.5
15.0
19.8
19.1
19.6
19.9
19.8
11.5
0.0
20.0
20.0
19.6
18.8
19.9
16.5
18.2
19.3
0.0
0.0
2.2
2.7
1.1
19.9
A-77
-------
Monitor ID
370511003
370590002
370610002
370650099
370670022
371010002
371090004
371170001
371190034
371190041
371290002
371290006
371310002
371450003
371470099
371730002
380070002
380070111
380130002
380130004
380150003
380171003
380171004
380250003
380530002
380530104
380530111
380550113
380570001
380570004
380570102
380570118
380570123
380570124
380590002
380590003
n
5
4
5
1
9
2
1
2
12
12
9
12
3
3
2
0
1
0
0
1
1
3
2
1
1
0
2
0
2
2
2
2
2
2
1
1
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
295
1949
83
325
438
15
10
1713
86
68
3325
2502
805
32251
14
283
426
4592
257
378
5
210
411
45808
45808
45808
45808
45808
45808
4592
4592
std
264
3658
132
848
4
2329
121
103
6800
5987
759
54874
3
226
119
522
55924
55924
55924
55924
55924
55924
min
17
13
6
325
5
12
10
66
5
5
6
6
16
5
12
283
426
4592
15
294
5
210
42
6264
6264
6264
6264
6264
6264
4592
4592
p2.5
17
13
6
325
5
12
10
66
5
5
6
6
16
5
12
283
426
4592
15
294
5
210
42
6264
6264
6264
6264
6264
6264
4592
4592
p50
173
175
36
325
46
15
10
1713
11
11
313
50
871
1136
14
283
426
4592
294
378
5
210
411
45808
45808
45808
45808
45808
45808
4592
4592
P97.5
675
7432
317
325
2591
17
10
3360
320
320
20865
20865
1529
95610
16
283
426
4592
462
462
5
210
781
85352
85352
85352
85352
85352
85352
4592
4592
p100
675
7432
317
325
2591
17
10
3360
320
320
20865
20865
1529
95610
16
283
426
4592
462
462
5
210
781
85352
85352
85352
85352
85352
85352
4592
4592
Distance of monitor to SO2 emission source (km)1
mean
15.8
15.3
12.3
16.1
6.3
10.3
10.7
6.6
13.3
12.7
14.5
6.9
4.2
18.8
1.3
0.0
11.4
0.0
0.0
18.6
9.8
7.7
9.0
13.9
17.3
0.0
16.1
0.0
2.5
2.7
5.4
10.7
14.3
18.6
2.6
5.1
std
2.5
4.3
4.9
0.0
5.7
7.5
0.0
7.8
4.7
5.0
4.9
4.8
1.8
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
6.9
1.1
0.0
0.0
0.0
0.1
0.0
2.6
2.0
2.3
2.2
1.4
1.0
0.0
0.0
min
11.5
10.4
4.1
16.1
1.2
5.0
10.7
1.1
6.3
6.3
2.3
0.6
2.1
18.4
1.3
0.0
11.4
0.0
0.0
18.6
9.8
3.0
8.2
13.9
17.3
0.0
16.1
0.0
0.7
1.3
3.8
9.1
13.3
17.9
2.6
5.1
p2.5
11.5
10.4
4.1
16.1
1.2
5.0
10.7
1.1
6.3
6.3
2.3
0.6
2.1
18.4
1.3
0.0
11.4
0.0
0.0
18.6
9.8
3.0
8.2
13.9
17.3
0.0
16.1
0.0
0.7
1.3
3.8
9.1
13.3
17.9
2.6
5.1
p50
16.5
15.6
13.1
16.1
3.9
10.3
10.7
6.6
12.8
12.2
15.4
7.1
5.1
18.7
1.3
0.0
11.4
0.0
0.0
18.6
9.8
4.6
9.0
13.9
17.3
0.0
16.1
0.0
2.5
2.7
5.4
10.7
14.3
18.6
2.6
5.1
P97.5
17.9
19.6
17.0
16.1
17.8
15.6
10.7
12.2
19.8
19.8
19.0
14.5
5.3
19.3
1.3
0.0
11.4
0.0
0.0
18.6
9.8
15.7
9.7
13.9
17.3
0.0
16.2
0.0
4.3
4.1
7.0
12.2
15.3
19.3
2.6
5.1
p100
17.9
19.6
17.0
16.1
17.8
15.6
10.7
12.2
19.8
19.8
19.0
14.5
5.3
19.3
1.3
0.0
11.4
0.0
0.0
18.6
9.8
15.7
9.7
13.9
17.3
0.0
16.2
0.0
4.3
4.1
7.0
12.2
15.3
19.3
2.6
5.1
A-78
-------
Monitor ID
380650002
380910001
381050103
381050105
390010001
390030002
390071001
390133002
390170004
390171004
390230003
390250021
390290016
390290022
390292001
390350038
390350045
390350060
390350065
390356001
390490004
390490034
390530002
390610010
390612003
390810016
390810017
390811001
390850003
390853002
390870006
390930017
390930026
390931003
390950008
390950024
n
1
0
1
1
1
9
5
5
11
9
4
6
9
9
8
10
10
10
10
13
6
6
6
10
11
17
17
13
6
3
8
3
2
3
9
10
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
28565
1605
1605
19670
442
1731
27781
907
1546
509
15304
20696
20696
22401
740
740
740
740
5759
75
75
31718
9265
660
13129
13129
6005
12044
1600
1425
165
27
165
4149
3745
std
535
3761
23029
1265
2186
349
28111
19955
19955
20621
916
916
916
916
16867
74
74
26583
26865
817
20063
20063
15392
24426
2615
2178
241
29
241
4513
4443
min
28565
1605
1605
19670
16
12
795
56
56
105
26
18
18
18
15
15
15
15
8
5
5
9
12
12
10
10
10
8
18
25
6
6
6
204
113
p2.5
28565
1605
1605
19670
16
12
795
56
56
105
26
18
18
18
15
15
15
15
8
5
5
9
12
12
10
10
10
8
18
25
6
6
6
204
113
p50
28565
1605
1605
19670
45
34
35454
233
309
492
145
24766
24766
25596
382
382
382
382
382
64
64
29551
537
268
361
361
234
2390
163
343
47
27
47
3712
2406
P97.5
28565
1605
1605
19670
1469
8458
56009
3998
6275
946
69953
59928
59928
59928
2453
2453
2453
2453
61629
192
192
74452
85699
2164
59928
59928
53414
61629
4618
6285
442
47
442
13581
13581
p100
28565
1605
1605
19670
1469
8458
56009
3998
6275
946
69953
59928
59928
59928
2453
2453
2453
2453
61629
192
192
74452
85699
2164
59928
59928
53414
61629
4618
6285
442
47
442
13581
13581
Distance of monitor to SO2 emission source (km)1
mean
8.5
0.0
2.8
1.8
11.4
8.5
17.3
14.5
14.7
6.5
12.2
15.0
12.7
12.7
11.4
9.8
10.1
10.4
9.8
13.8
8.7
9.5
7.0
16.1
8.7
9.5
9.6
4.9
9.1
5.3
13.7
11.4
3.3
11.6
8.1
11.4
std
0.0
0.0
0.0
0.0
0.0
0.4
0.6
5.1
6.9
6.5
6.1
2.7
3.6
3.6
4.1
4.9
5.5
5.7
4.3
7.1
3.4
3.0
7.4
3.0
5.5
7.1
6.9
5.6
4.2
6.0
6.0
2.2
0.5
2.1
5.5
6.4
min
8.5
0.0
2.8
1.8
11.4
7.9
16.6
6.0
0.9
1.7
5.8
12.7
7.2
7.2
4.6
1.9
1.2
1.0
2.0
1.7
2.9
3.4
1.0
8.6
0.4
1.7
2.0
0.3
5.6
1.1
2.2
8.9
3.0
9.2
2.5
3.9
p2.5
8.5
0.0
2.8
1.8
11.4
7.9
16.6
6.0
0.9
1.7
5.8
12.7
7.2
7.2
4.6
1.9
1.2
1.0
2.0
1.7
2.9
3.4
1.0
8.6
0.4
1.7
2.0
0.3
5.6
1.1
2.2
8.9
3.0
9.2
2.5
3.9
p50
8.5
0.0
2.8
1.8
11.4
8.3
17.2
15.8
18.5
3.3
12.0
14.1
13.5
13.6
10.8
11.7
10.4
13.3
9.8
16.8
9.2
10.4
3.6
16.8
8.0
5.6
5.9
2.9
7.4
2.6
15.5
12.5
3.3
12.5
4.5
9.5
P97.5
8.5
0.0
2.8
1.8
11.4
9.3
18.2
19.8
19.3
19.8
19.2
18.7
18.1
18.2
19.3
14.3
15.8
15.5
14.5
20.0
12.9
11.5
16.5
19.7
19.4
19.0
18.6
18.0
15.2
12.3
19.3
12.8
3.6
13.1
14.6
18.6
p100
8.5
0.0
2.8
1.8
11.4
9.3
18.2
19.8
19.3
19.8
19.2
18.7
18.1
18.2
19.3
14.3
15.8
15.5
14.5
20.0
12.9
11.5
16.5
19.7
19.4
19.0
18.6
18.0
15.2
12.3
19.3
12.8
3.6
13.1
14.6
18.6
A-79
-------
Monitor ID
390990009
390990013
391051001
391130025
391150003
391150004
391450013
391450020
391450022
391510016
391530017
391530022
391570003
391570006
400219002
400710602
400719003
400719010
400979014
401010167
401090025
401091037
401159004
401430175
401430235
401430501
420030002
420030010
420030021
420030031
420030032
420030064
420030067
420030116
420031301
420033003
n
10
10
6
6
2
2
0
3
3
7
4
4
7
6
0
2
2
0
6
8
2
2
1
10
10
10
19
55
64
62
64
54
16
19
57
54
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
2107
2107
31718
1609
57763
57763
1450
1450
181
2763
2763
368
426
3502
3502
3180
3751
91
91
62
938
938
938
103
85
819
757
819
213
73
103
914
213
std
5350
5350
26583
2326
38696
38696
1306
1306
213
2244
2244
741
795
457
457
5200
4529
110
110
1088
1088
1088
137
101
5274
5327
5274
741
105
137
5587
741
min
6
6
9
105
30401
30401
25
25
10
863
863
15
15
3178
3178
173
23
13
13
62
9
9
9
7
5
5
5
5
5
7
7
5
5
p2.5
6
6
9
105
30401
30401
25
25
10
863
863
15
15
3178
3178
173
23
13
13
62
9
9
9
7
7
7
7
7
6
7
7
7
6
p50
353
353
29551
753
57763
57763
1737
1737
43
2091
2091
38
38
3502
3502
713
1130
91
91
62
263
263
263
30
49
47
46
47
52
29
30
47
52
P97.5
17244
17244
74452
6275
85125
85125
2589
2589
510
6009
6009
2017
2017
3825
3825
13428
9866
169
169
62
2729
2729
2729
468
407
5395
468
5395
1164
407
468
5395
1164
p100
17244
17244
74452
6275
85125
85125
2589
2589
510
6009
6009
2017
2017
3825
3825
13428
9866
169
169
62
2729
2729
2729
468
468
42018
42018
42018
5395
407
468
42018
5395
Distance of monitor to SO2 emission source (km)1
mean
12.4
12.4
13.6
13.4
4.8
5.1
0.0
9.6
8.4
6.6
5.0
3.9
12.0
6.4
0.0
3.4
1.8
0.0
4.7
5.9
8.7
8.8
5.2
11.8
10.7
12.6
7.4
14.2
11.7
13.9
11.7
6.0
15.1
7.4
9.9
5.6
std
7.3
7.5
2.2
5.4
0.2
0.3
0.0
6.9
7.5
1.5
2.4
0.7
6.4
6.1
0.0
2.3
2.0
0.0
1.3
4.2
4.5
7.9
0.0
6.9
6.9
6.8
5.9
5.6
3.3
5.1
3.3
5.2
3.5
5.1
4.6
5.4
min
2.0
1.7
11.6
7.3
4.6
4.9
0.0
4.6
2.8
4.5
1.4
3.0
0.6
0.4
0.0
1.8
0.4
0.0
2.7
3.7
5.6
3.2
5.2
1.4
1.5
2.7
0.6
2.5
3.2
1.3
3.1
2.0
6.1
2.1
1.1
1.0
p2.5
2.0
1.7
11.6
7.3
4.6
4.9
0.0
4.6
2.8
4.5
1.4
3.0
0.6
0.4
0.0
1.8
0.4
0.0
2.7
3.7
5.6
3.2
5.2
1.4
1.5
2.7
0.6
2.5
4.8
1.4
4.7
2.0
6.1
2.1
1.1
1.0
p50
15.6
15.8
13.0
13.4
4.8
5.1
0.0
6.7
5.4
5.9
6.0
4.1
13.3
5.3
0.0
3.4
1.8
0.0
5.5
3.7
8.7
8.8
5.2
13.9
13.4
14.2
8.6
15.5
13.1
14.4
13.2
3.1
15.7
7.7
11.0
2.3
P97.5
19.6
19.6
17.8
19.4
4.9
5.3
0.0
17.5
16.9
8.7
6.6
4.6
18.6
14.2
0.0
5.0
3.2
0.0
5.7
15.8
11.9
14.4
5.2
18.3
18.1
19.2
18.1
20.0
18.0
18.7
18.1
17.9
19.7
17.0
17.5
17.8
p100
19.6
19.6
17.8
19.4
4.9
5.3
0.0
17.5
16.9
8.7
6.6
4.6
18.6
14.2
0.0
5.0
3.2
0.0
5.7
15.8
11.9
14.4
5.2
18.3
18.1
19.2
18.1
20.0
18.7
19.8
18.7
18.2
19.7
17.0
17.8
17.8
A-80
-------
Monitor ID
420033004
420070002
420070004
420070005
420070014
420110009
420110100
420130801
420170012
420210011
420270100
420430401
420450002
420450109
420490003
420630004
420692006
420710007
420730015
420770004
420791101
420810100
420810403
420850100
420910013
420950025
420950100
420958000
420990301
421010004
421010022
421010024
421010027
421010029
421010047
421010048
n
55
10
7
8
10
13
12
1
22
4
4
8
57
45
5
3
5
5
9
13
4
3
3
2
28
18
15
16
0
61
66
36
63
67
65
60
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
209
18726
5881
5173
4400
1140
1231
441
687
4195
1090
107
681
855
824
4796
13
75
3206
703
117
28
28
14
171
676
2179
2045
102
285
46
99
262
270
104
std
735
19819
11104
10474
9400
3818
3973
3033
5171
1267
99
1415
1553
1068
5156
5
109
8423
1041
160
28
28
4
704
1020
5602
5439
316
1022
77
311
1007
1022
318
min
5
18
9
9
8
14
14
441
5
34
53
10
5
5
10
1497
6
6
6
7
9
6
6
11
5
7
7
7
5
5
5
5
5
5
5
p2.5
6
18
9
9
8
14
14
441
5
34
53
10
5
5
10
1497
6
6
6
7
9
6
6
11
5
7
7
7
6
5
5
6
5
5
6
p50
49
15912
118
157
157
37
34
441
27
3004
834
78
47
91
228
2154
15
23
28
120
53
18
18
14
15
86
120
86
20
26
13
20
24
26
22
P97.5
1164
59928
30312
30312
30312
13841
13841
441
14266
10738
2638
313
5051
5051
2398
10738
18
264
25551
2888
351
59
59
17
3753
2888
22057
22057
560
4450
407
560
4450
4450
560
p100
5395
59928
30312
30312
30312
13841
13841
441
14266
10738
2638
313
6720
6720
2398
10738
18
264
25551
2888
351
59
59
17
3753
2888
22057
22057
2378
6720
407
2378
6720
6720
2378
Distance of monitor to SO2 emission source (km)1
mean
5.9
13.0
14.5
9.6
12.0
9.8
8.7
1.3
11.1
8.5
10.4
5.4
13.6
12.4
3.1
18.4
10.9
3.7
12.5
12.5
12.3
11.3
15.8
10.8
15.3
13.1
10.4
10.1
0.0
10.5
8.0
13.0
9.8
8.3
7.9
10.4
std
6.0
3.2
5.1
5.6
3.1
7.1
6.3
0.0
6.5
7.4
6.2
4.0
5.5
6.4
1.9
1.4
7.4
3.7
5.6
5.8
3.4
0.7
1.1
11.8
4.5
4.3
5.5
5.9
0.0
5.2
5.6
3.8
4.6
4.7
4.5
4.9
min
0.6
9.2
7.4
2.5
7.1
1.3
1.5
1.3
1.2
1.5
2.3
0.8
1.3
0.5
1.2
17.0
2.1
0.6
0.6
0.3
7.8
10.6
14.9
2.4
1.4
4.0
2.5
0.6
0.0
1.0
0.9
6.3
0.8
1.1
0.6
0.9
p2.5
0.7
9.2
7.4
2.5
7.1
1.3
1.5
1.3
1.2
1.5
2.3
0.8
1.9
1.6
1.2
17.0
2.1
0.6
0.6
0.3
7.8
10.6
14.9
2.4
1.4
4.0
2.5
0.6
0.0
1.3
1.0
6.3
1.7
1.8
0.8
1.7
p50
3.3
11.4
16.0
8.8
12.0
10.3
7.5
1.3
12.4
8.9
11.4
3.7
15.8
13.3
2.6
18.4
8.2
2.7
13.2
12.0
12.9
11.2
15.4
10.8
16.2
14.1
10.7
9.1
0.0
10.9
7.0
12.6
11.0
6.8
6.4
10.7
P97.5
18.8
18.6
19.8
17.1
17.2
19.8
17.2
1.3
19.6
14.9
16.6
12.1
19.8
19.9
5.4
19.8
19.6
10.1
18.0
19.3
15.8
12.0
16.9
19.1
20.0
19.7
19.3
18.8
0.0
19.2
19.4
19.9
19.7
18.9
17.6
18.6
p100
18.8
18.6
19.8
17.1
17.2
19.8
17.2
1.3
19.6
14.9
16.6
12.1
19.8
20.0
5.4
19.8
19.6
10.1
18.0
19.3
15.8
12.0
16.9
19.1
20.0
19.7
19.3
18.8
0.0
19.7
20.0
19.9
19.7
19.6
17.9
19.2
A-81
-------
Monitor ID
421010055
421010136
421070003
421230003
421230004
421250005
421250200
421255001
421290008
421330008
440070012
440071005
440071009
450030003
450110001
450190003
450190046
450430006
450450008
450450009
450630008
450730001
450750003
450790007
450790021
450791003
450791006
460330132
460710001
460990007
470010028
470090002
470090006
470090101
470110102
470310004
n
66
68
6
2
2
33
1
8
3
9
54
55
55
13
1
16
0
7
12
13
11
1
5
10
8
13
10
0
0
1
8
3
3
3
2
0
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
286
319
831
2445
2445
257
7
321
24
8943
41
41
41
1654
65
2183
5834
89
83
948
5
1433
61
5061
995
4289
496
5595
1421
1421
1421
2719
std
1022
1042
687
659
659
945
439
9
22698
90
89
89
2599
6339
14038
136
132
2944
1913
103
12720
2730
11350
14808
2325
2325
2325
3687
min
5
5
8
1979
1979
5
7
7
16
14
5
5
5
8
65
6
6
6
6
5
5
5
5
7
5
7
496
7
6
6
6
112
p2.5
5
5
8
1979
1979
5
7
7
16
14
5
5
5
8
65
6
6
6
6
5
5
5
5
7
5
7
496
7
6
6
6
112
p50
26
27
674
2445
2445
47
7
82
22
171
13
13
13
549
65
28
24
20
19
9
5
211
18
89
52
89
496
34
153
153
153
2719
P97.5
4450
4450
1743
2911
2911
5395
7
1017
34
68932
392
392
392
8275
65
25544
37622
411
411
9820
5
4088
343
36378
9820
36378
496
42188
4104
4104
4104
5326
p100
6720
6720
1743
2911
2911
5395
7
1017
34
68932
521
521
521
8275
65
25544
37622
411
411
9820
5
4088
343
36378
9820
36378
496
42188
4104
4104
4104
5326
Distance of monitor to SO2 emission source (km)1
mean
7.9
8.8
10.4
4.0
3.0
15.7
1.1
15.9
9.8
9.3
8.4
9.1
8.6
15.3
13.2
7.2
0.0
4.6
11.7
10.1
11.5
14.9
8.5
14.0
14.7
10.9
17.5
0.0
0.0
17.5
12.2
5.7
5.4
12.1
2.5
0.0
std
5.4
5.4
7.4
1.2
1.6
4.7
0.0
4.1
1.4
5.8
5.8
5.5
6.0
1.5
0.0
5.0
0.0
4.3
4.5
5.7
5.4
0.0
5.1
4.1
1.2
5.9
3.3
0.0
0.0
0.0
6.5
5.7
5.3
6.9
1.2
0.0
min
1.3
1.1
3.3
3.2
1.9
1.1
1.1
9.3
8.7
0.8
0.3
0.9
0.1
11.4
13.2
1.1
0.0
0.2
2.1
4.0
0.5
14.9
3.4
6.4
12.3
1.4
8.2
0.0
0.0
17.5
0.9
0.7
1.4
4.2
1.6
0.0
p2.5
1.4
1.4
3.3
3.2
1.9
1.1
1.1
9.3
8.7
0.8
0.4
1.0
0.4
11.4
13.2
1.1
0.0
0.2
2.1
4.0
0.5
14.9
3.4
6.4
12.3
1.4
8.2
0.0
0.0
17.5
0.9
0.7
1.4
4.2
1.6
0.0
p50
6.8
9.3
8.8
4.0
3.0
17.5
1.1
17.2
9.3
10.1
5.9
8.4
6.3
15.3
13.2
6.2
0.0
3.4
10.7
5.4
13.0
14.9
9.6
15.9
15.3
10.9
18.9
0.0
0.0
17.5
12.8
4.5
3.3
15.4
2.5
0.0
P97.5
18.8
18.7
19.2
4.9
4.1
18.7
1.1
19.7
11.5
17.7
18.9
18.5
19.5
17.5
13.2
16.3
0.0
13.2
17.4
17.3
19.2
14.9
15.8
18.7
15.6
18.5
19.1
0.0
0.0
17.5
18.8
11.9
11.3
16.7
3.4
0.0
p100
20.0
19.8
19.2
4.9
4.1
18.7
1.1
19.7
11.5
17.7
19.0
19.0
19.9
17.5
13.2
16.3
0.0
13.2
17.4
17.3
19.2
14.9
15.8
18.7
15.6
18.5
19.1
0.0
0.0
17.5
18.8
11.9
11.3
16.7
3.4
0.0
A-82
-------
Monitor ID
470370011
470730002
470850020
471070101
471250006
471250106
471390003
471390007
471390008
471390009
471450009
471570034
471570043
471570046
471571034
471610007
471630007
471630009
471651002
480610006
481130069
481390015
481390016
481390017
481410037
481410053
481410058
481670005
481671002
481830001
482010046
482010051
482010059
482010062
482010070
482011035
n
9
3
6
3
6
6
1
1
1
1
4
18
18
2
19
3
10
12
4
0
9
12
12
12
13
13
16
43
43
5
29
2
38
37
31
39
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
891
11831
18599
1834
222
222
1900
1900
1900
1900
19470
1204
1204
1973
1150
5561
3010
2513
8593
34
664
664
664
44
44
38
185
185
13289
606
13
674
694
790
657
std
2248
10420
44191
3024
401
401
22311
2391
2391
2640
2336
5107
5303
4935
10129
25
993
993
993
92
92
83
611
611
12287
1182
8
1486
1503
1622
1470
min
9
6
12
64
8
8
1900
1900
1900
1900
9
5
5
106
5
21
22
13
88
9
13
13
13
5
5
5
5
5
6
6
7
6
6
6
6
p2.5
9
6
12
64
8
8
1900
1900
1900
1900
9
5
5
106
5
21
22
13
88
9
13
13
13
5
5
5
6
6
6
6
7
6
6
6
6
p50
60
15822
281
112
35
35
1900
1900
1900
1900
19188
32
32
1973
35
6580
495
286
7029
18
57
57
57
11
11
12
22
22
19024
161
13
48
49
161
46
P97.5
6842
19666
108788
5326
1025
1025
1900
1900
1900
1900
39495
6540
6540
3839
6540
10081
16855
16855
20226
69
3003
3003
3003
345
345
345
1937
1937
24837
5097
18
6968
6968
6968
6968
p100
6842
19666
108788
5326
1025
1025
1900
1900
1900
1900
39495
6540
6540
3839
6540
10081
16855
16855
20226
69
3003
3003
3003
345
345
345
3599
3599
24837
5097
18
6968
6968
6968
6968
Distance of monitor to SO2 emission source (km)1
mean
10.4
2.9
3.2
7.6
6.2
7.1
3.1
1.6
1.4
1.0
10.9
11.4
9.6
6.0
3.5
1.8
3.7
5.7
4.2
0.0
12.1
9.5
9.0
9.6
9.7
9.7
13.9
2.3
3.6
18.9
12.8
19.1
10.3
14.8
10.7
8.6
std
3.6
2.1
1.8
10.2
6.9
7.3
0.0
0.0
0.0
0.0
6.7
2.2
1.7
6.7
5.6
0.2
2.6
6.0
1.8
0.0
5.7
5.8
6.3
6.9
1.8
1.6
2.3
1.3
1.1
0.5
3.1
0.6
5.9
3.8
5.3
5.4
min
5.6
1.7
1.6
0.5
1.0
1.5
3.1
1.6
1.4
1.0
5.3
4.8
5.3
1.3
0.5
1.7
1.7
2.0
2.9
0.0
2.0
2.3
2.9
1.9
4.5
5.1
9.5
1.2
2.5
18.6
6.2
18.7
1.8
7.8
2.2
1.6
p2.5
5.6
1.7
1.6
0.5
1.0
1.5
3.1
1.6
1.4
1.0
5.3
4.8
5.3
1.3
0.5
1.7
1.7
2.0
2.9
0.0
2.0
2.3
2.9
1.9
4.5
5.1
9.5
1.3
2.5
18.6
6.2
18.7
1.8
7.8
2.2
1.6
p50
10.7
1.7
2.6
3.0
2.5
3.5
3.1
1.6
1.4
1.0
9.5
11.8
10.0
6.0
0.7
1.7
2.6
2.7
3.5
0.0
12.9
9.4
6.1
7.6
10.0
9.9
14.7
2.0
3.3
18.7
13.1
19.1
8.5
15.7
8.7
7.7
P97.5
17.6
5.2
6.3
19.3
15.0
16.3
3.1
1.6
1.4
1.0
19.1
15.3
11.4
10.8
18.0
1.9
10.7
18.7
6.9
0.0
20.0
16.6
17.4
18.6
12.0
11.9
16.0
3.3
4.6
19.9
19.6
19.5
19.5
20.0
19.5
17.6
p100
17.6
5.2
6.3
19.3
15.0
16.3
3.1
1.6
1.4
1.0
19.1
15.3
11.4
10.8
18.0
1.9
10.7
18.7
6.9
0.0
20.0
16.6
17.4
18.6
12.0
11.9
16.0
9.5
9.5
19.9
19.6
19.5
19.5
20.0
19.5
17.6
A-83
-------
Monitor ID
482011050
482450009
482450011
482450020
482570005
483550025
483550026
483550032
490050004
490110001
490110004
490350012
490351001
490352004
500070003
500070014
500210002
510360002
510590005
510590018
510591004
510591005
510595001
511130003
511611004
511650002
511650003
515100009
516500004
517100023
517600024
530090010
530090012
530330057
530330080
530530021
n
46
16
27
8
0
17
19
17
1
6
6
6
7
3
1
1
0
18
5
10
11
13
11
1
8
7
6
11
15
21
14
1
1
5
5
3
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
243
863
999
170
468
424
468
5
468
468
468
833
1245
6
6
4818
31
1820
1664
1416
1566
7
85
40
39
1663
285
1738
191
756
756
241
241
179
std
1028
2732
2362
306
1086
1032
1086
500
500
500
1006
1415
17274
46
5043
4813
4435
4837
117
36
40
4813
505
7026
363
301
301
213
min
6
6
6
6
6
6
6
5
8
8
8
8
8
6
6
7
8
8
7
7
6
7
5
8
5
7
6
5
6
756
756
63
63
11
p2.5
7
6
6
6
6
6
6
5
8
8
8
8
8
6
6
7
8
8
7
7
6
7
5
8
5
7
6
5
6
756
756
63
63
11
p50
36
80
45
64
43
43
43
5
366
366
366
712
939
6
6
35
11
74
59
59
24
7
34
32
25
59
92
85
16
756
756
117
117
109
P97.5
829
11064
11064
908
3955
3955
3955
5
1332
1332
1332
2788
2788
6
6
73839
114
16141
16141
16141
16141
7
341
108
108
16141
1983
32344
1148
756
756
771
771
419
p100
6968
11064
11064
908
3955
3955
3955
5
1332
1332
1332
2788
2788
6
6
73839
114
16141
16141
16141
16141
7
341
108
108
16141
1983
32344
1148
756
756
771
771
419
Distance of monitor to SO2 emission source (km)1
mean
16.5
14.8
9.0
10.8
0.0
6.7
10.0
3.9
1.8
8.2
9.7
4.9
13.0
9.8
1.6
1.9
0.0
12.1
17.2
13.5
10.9
13.6
14.8
10.8
9.3
12.3
11.4
9.6
11.1
8.3
9.4
5.6
5.3
4.0
5.0
3.2
std
3.9
6.8
5.3
8.1
0.0
3.0
3.3
4.1
0.0
5.8
6.0
3.7
6.5
8.0
0.0
0.0
0.0
7.2
1.6
3.9
3.5
4.3
4.4
0.0
5.5
5.1
5.4
5.1
4.9
3.4
5.8
0.0
0.0
6.0
4.2
1.1
min
5.0
0.4
2.8
1.8
0.0
4.2
4.6
0.4
1.8
1.5
2.3
0.6
2.1
2.4
1.6
1.9
0.0
2.0
15.0
8.4
3.7
4.6
5.1
10.8
2.9
5.1
6.3
1.1
4.0
3.6
1.2
5.6
5.3
0.6
2.5
2.1
p2.5
5.3
0.4
2.8
1.8
0.0
4.2
4.6
0.4
1.8
1.5
2.3
0.6
2.1
2.4
1.6
1.9
0.0
2.0
15.0
8.4
3.7
4.6
5.1
10.8
2.9
5.1
6.3
1.1
4.0
3.6
1.2
5.6
5.3
0.6
2.5
2.1
p50
17.9
18.7
7.0
11.3
0.0
5.2
11.0
1.7
1.8
8.1
9.8
4.5
13.0
8.9
1.6
1.9
0.0
13.6
17.3
15.7
11.2
13.8
16.0
10.8
9.7
13.9
10.3
8.6
11.3
8.3
10.3
5.6
5.3
1.3
3.1
3.2
P97.5
19.1
19.7
18.1
19.9
0.0
16.4
13.6
16.0
1.8
17.7
19.2
8.9
19.6
18.3
1.6
1.9
0.0
19.9
19.4
17.5
16.3
19.0
19.8
10.8
19.1
17.8
17.9
17.9
17.9
18.8
20.0
5.6
5.3
14.7
12.5
4.3
p100
19.9
19.7
18.1
19.9
0.0
16.4
13.6
16.0
1.8
17.7
19.2
8.9
19.6
18.3
1.6
1.9
0.0
19.9
19.4
17.5
16.3
19.0
19.8
10.8
19.1
17.8
17.9
17.9
17.9
18.8
20.0
5.6
5.3
14.7
12.5
4.3
A-84
-------
Monitor ID
530530031
530570012
530571003
530610016
530730011
540090005
540090007
540110006
540250001
540290005
540290007
540290008
540290009
540290011
540290014
540290015
540290016
540291004
540390004
540390010
540392002
540511002
540610003
540610004
540610005
540690007
540990002
540990003
540990004
540990005
541071002
550090005
550250041
550410007
550730005
550790007
n
3
4
4
2
9
13
17
5
0
8
16
9
15
17
16
9
16
16
4
4
5
5
2
4
3
2
8
8
8
8
11
7
7
1
3
9
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
179
2238
2238
191
488
6005
13129
1501
22069
9282
20696
9894
13129
9282
20696
10611
10611
1529
1529
22698
27781
45992
24472
32132
37391
1271
1271
1271
1271
4375
3413
1293
5
4040
1750
std
213
2630
2630
194
695
15392
20063
2677
20983
17668
19955
18112
20063
17668
19955
17732
17732
1146
1146
47491
23029
63840
44468
51128
22660
2194
2194
2194
2194
9095
5045
2743
6715
4858
min
11
21
21
53
8
10
10
124
18
10
18
10
10
10
18
10
10
854
854
750
795
850
850
850
21367
25
25
25
25
7
9
7
5
24
5
p2.5
11
21
21
53
8
10
10
124
18
10
18
10
10
10
18
10
10
854
854
750
795
850
850
850
21367
25
25
25
25
7
9
7
5
24
5
p50
109
1793
1793
191
349
234
361
401
25596
238
24766
243
361
238
24766
302
302
1008
1008
1009
35454
45992
2952
4412
37391
343
343
343
343
1517
850
71
5
303
28
P97.5
419
5345
5345
328
2286
53414
59928
6285
59928
59928
59928
59928
59928
59928
59928
59928
59928
3245
3245
107633
56009
91134
91134
91134
53414
6285
6285
6285
6285
31006
13470
7417
5
11792
14686
p100
419
5345
5345
328
2286
53414
59928
6285
59928
59928
59928
59928
59928
59928
59928
59928
59928
3245
3245
107633
56009
91134
91134
91134
53414
6285
6285
6285
6285
31006
13470
7417
5
11792
14686
Distance of monitor to SO2 emission source (km)1
mean
1.8
2.2
1.7
0.5
16.9
5.3
10.7
13.2
0.0
9.3
13.1
12.1
11.0
10.7
11.8
12.1
10.8
11.5
10.2
9.7
9.1
10.1
4.6
11.8
9.2
13.9
9.7
9.6
9.6
9.5
8.5
4.2
7.4
8.3
10.7
6.5
std
0.9
0.8
0.6
0.1
6.2
5.3
5.3
7.1
0.0
5.3
3.8
4.2
3.5
5.2
4.0
3.5
4.3
3.9
4.3
4.6
5.6
4.7
1.4
8.9
9.7
1.8
5.5
5.5
6.0
6.4
5.4
3.4
4.7
0.0
9.2
3.4
min
1.2
1.3
1.1
0.4
0.5
0.9
3.9
0.5
0.0
4.7
4.8
6.3
1.0
3.2
1.5
7.1
1.1
1.8
6.0
5.2
2.3
2.2
3.6
0.8
1.0
12.7
1.7
1.5
1.0
0.9
2.7
1.1
2.8
8.3
0.1
1.8
p2.5
1.2
1.3
1.1
0.4
0.5
0.9
3.9
0.5
0.0
4.7
4.8
6.3
1.0
3.2
1.5
7.1
1.1
1.8
6.0
5.2
2.3
2.2
3.6
0.8
1.0
12.7
1.7
1.5
1.0
0.9
2.7
1.1
2.8
8.3
0.1
1.8
p50
1.3
2.3
1.7
0.5
19.3
2.7
8.3
16.2
0.0
7.5
13.1
11.2
12.0
8.8
11.1
12.4
10.6
11.8
10.0
9.8
6.7
11.4
4.6
13.5
6.7
13.9
10.6
10.7
11.3
11.4
8.8
3.1
5.2
8.3
15.8
5.9
P97.5
2.8
3.1
2.4
0.6
19.7
16.8
18.8
17.2
0.0
17.6
18.3
19.8
17.7
18.8
19.4
18.2
18.3
19.8
14.8
14.0
15.5
15.0
5.6
19.4
19.9
15.2
16.0
15.8
15.8
16.2
17.0
9.7
14.7
8.3
16.2
12.9
p100
2.8
3.1
2.4
0.6
19.7
16.8
18.8
17.2
0.0
17.6
18.3
19.8
17.7
18.8
19.4
18.2
18.3
19.8
14.8
14.0
15.5
15.0
5.6
19.4
19.9
15.2
16.0
15.8
15.8
16.2
17.0
9.7
14.7
8.3
16.2
12.9
A-85
-------
Monitor ID
550790026
550790041
550850996
551110007
551250001
551410016
560050857
n
9
9
2
2
0
6
4
SO2 emissions (tpy) from sources within 20 km of monitor1
mean
1750
1750
1152
31
2374
2527
std
4858
4858
1617
35
2368
3868
min
5
5
9
7
6
23
p2.5
5
5
9
7
6
23
p50
28
28
1152
31
2032
896
P97.5
14686
14686
2295
56
5782
8291
p100
14686
14686
2295
56
5782
8291
Distance of monitor to SO2 emission source (km)1
mean
7.6
10.1
0.9
14.7
0.0
5.3
4.6
std
3.0
3.0
0.1
7.4
0.0
2.6
6.5
min
3.5
5.9
0.9
9.5
0.0
2.3
1.1
p2.5
3.5
5.9
0.9
9.5
0.0
2.3
1.1
p50
7.5
10.2
0.9
14.7
0.0
4.9
1.6
P97.5
12.8
14.5
1.0
19.9
0.0
9.8
14.4
p100
12.8
14.5
1.0
19.9
0.0
9.8
14.4
Notes:
1 Mean, std , min, p2.5, p50, p97.5, max are the arithmetic average, standard deviation, minimum, 2.5th, 50th, 97.5th percentiles, and maximum distances
and emissions.
2 There were no emissions above 5 tpy for sources located within 20 km of the monitors sited in Puerto Rico and the Virgin Islands.
A-86
-------
A.2 Analysis of Duplicate SO2 Values at Ambient Monitor Locations
During the screening of each of the ambient monitoring data sets, it became evident that
simultaneous measurements were present. Staff analyzed the duplicate SO2 measurements to
discern if there were any differences in the reported/measured values because ultimately only
one value would be selected for use in each of the final screened data sets. Staff was not
interested in whether multiple monitors were present at a particular monitoring site or if there
were duplicate reporting of SC>2 concentrations, only to determine that the selection of a
particular value used in the final data sets were not biased.
In selecting which of the duplicate concentrations to use for final REA data sets, staff
made the following judgements. First, the ambient monitor POC containing the greatest number
of samples was used to populate the max-5 data set. Second, where continous-5 measurements
were available and coincided with max-5 measurements, staff selected the 5-minute maximum
SC>2 concentration from the continuous-5 data set. And finally, where continuous-5 data were
available and used to estimate a 1-hour average SC>2 concentration that coincided with a reported
1-hour ambient monitor concentration, the continuous-5 1-hour average concentration was used.
Staff designed the following analyses to explore the effect the selection of one concentration
over another may have on the final data set used.
Staff calculated the relative percent difference (RPD) for each duplicate concentration,
considering measurements within the 5-max data set (n=300,438), duplicate reporting between
the continuous-5 and the max-5 data sets (n=29,058), and duplicate values between the 1-hour
and the continuous-5 data sets (n=258,457), separately. We anticipated that small fluctuations in
concentration between the duplicate data would have a greater influence on the RPD at lower
concentrations than at higher concentrations. Therefore, staff separated the duplicate values into
concentration groups for this analysis. Two groups were constructed; one with concentrations <
10 ppb and the other conrtaining concentrations > 10 ppb. The following formula calculates the
RPD for each duplicate value:
RPD = ^ ^ x 200 equation A.2-1
(Q+C2)
where,
RPD = Relative percent difference (%)
A-87
-------
C2
First 862 concentration value
Second 862 concentration value
Depending on the difference in concentration, the value for the calculated RPD could be
as low as -200 or as high +200, indicating the maximum difference between any two values,
while an RPD of zero indicates no difference. The sign of the value can also indicate the
direction of bias when comparing the first concentration to the second.
In the first comparison (i.e., the within max-5 duplicates), Ci was selected as the ambient
monitor containing the overall greater sample size/duration. Table A.2-1 summarizes the
distribution of RPDs for where duplicate values of 862 concentrations were less than 10 ppb
within the max-5 monitoring data set. On average, there were relatively small differences in the
duplicate values reported at each of the monitoring locations. Most duplicate values were within
+1-61% of one another, although some were noted at or above 100% (absolute difference). In
considering that these maximum 5-minute SC>2 concentrations are well below that of potential
interest in the exposure and risk analysis, this degree of agreement between the two values at
these concentration levels is acceptable.
Table A.2-1. Distribution of the relative percent difference (RPD) between
duplicate 5-minute maximum SO2 values at max-5 monitors, where
concentrations were < 10 ppb.
Monitor ID
290210009
290210011
290930030
290930031
290990004
290990014
290990017
290990018
291630002
n
25868
22247
54904
48417
22788
33245
21460
17025
11528
Relative Percent Difference (%)1
mean
0
-7
8
-14
-8
-12
2
2
-3
std
34
22
34
29
27
29
30
25
34
min
-196
-143
-181
-122
-120
-133
-120
-156
-164
p5
-50
-40
-40
-67
-50
-67
-50
-40
-40
p50
0
0
0
0
0
0
0
0
0
p95
67
18
67
67
67
29
67
67
67
max
100
67
100
67
100
67
120
100
67
Notes:
1 the mean, std, min, p5, p50, p95, max are the arithmetic average, standard deviation, minimum,
5th, median, 95th, and maximum, respectively.
When considering duplicate values > 10 ppb, the RPD was much lower at each of the
monitors (Table A.2-2). Most of the RPDs are within +/-10%, indicating excellent agreement
among the duplicate values. A small negative bias may exist with selection of the monitor with
A-88
-------
the greatest number of samples as the base monitor, but on average the difference was typically
less than 3%.
Table A.2-2. Distribution of the relative percent difference (RPD) between
duplicate 5-minute maximum SO2 values at max-5 monitors, where
concentrations were > 10 ppb.
Monitor ID
290210009
290210011
290930030
290930031
290990004
290990014
290990017
290990018
291630002
n
2333
2344
8068
7652
8627
4973
5138
2626
1195
Relative Percent Difference (%)1
mean
-2
0
-1
-3
-1
2
-1
0
-6
std
6
3
6
6
4
16
7
6
32
min
-133
-66
-120
-134
-100
-17
-137
-81
-137
P5
-10
-6
-9
-13
-7
-8
-11
-7
-133
p50
0
0
0
-2
0
0
0
0
0
p95
6
5
4
0
5
9
10
10
11
max
18
18
24
10
20
184
32
32
29
Notes:
1 the mean, std, min, p5, p50, p95, max are the arithmetic average, standard deviation, minimum,
5th, median, 95th, and maximum, respectively.
Staff also analyzed data where the max-5 sampling times corresponded with the
continuous-5 monitoring at the same location. Of the 29,058 duplicate measurement values, only
312 contained different values among the two sample types (i.e, a non-zero RPD). This indicates
that the majority of the data are duplicate values reported in each of the two data sets. Since
there were very few samples with RPDs deviating from zero (i.e., 1.1%), the following analysis
included only the samples that had a non-zero difference and at any concentration levels. The
distribution for the RPD given these monitors and duplicate monitoring events is provided in
Table A.2-3. On average there may be a small positive bias in selecting the continuous-5
monitoring concentrations where differences existed, however given that there were only 1% of
samples that differed among the two data sets, the overall impact to the below estimation
procedure is determined as negligible. In addition, selection of the continuous-5 measurement
preserves the relationship between the actual 5-minute maximum and the calculated 1-hour
concentration derived from the multiple 5-minute measurements that occurred within the hour,
not adding to uncertainty regarding the true relationship between the 1-hour and 5-minute
maximum concentrations.
A-89
-------
Table A.2-3. Distribution of the relative percent difference (RPD) between
simultaneous 5-minute SO2 maximum values in the max-5 and continuous-5 data
sets, where concentrations > 0 ppb.
Monitor ID
301110066
301110079
301110082
301110083
n1
76
149
47
40
Relative Percent Difference (%) 2
mean
26
27
25
78
std
57
48
52
64
min
-143
-178
-67
-120
P5
-117
-67
-67
-53
p50
16
29
29
67
p95
133
67
67
160
max
160
164
186
160
Notes:
1 This distribution is for the number of samples where the RPD was non-zero. The majority of the
duplicate measures (n=28,746) were identical.
2 the mean, std, min, p5, p50, p95, max are the arithmetic average, standard deviation, minimum, 5th,
median, 95th, and maximum, respectively.
In the last comparison (i.e., the 1-hour concentration duplicates), the 1-hour
concentration from the continous-5 ambient monitors was selected as Ci in equation A.2-1.
Table A.2-4 summarizes the distribution of RPDs for where duplicate measurements of SO2
concentrations were less than 10 ppb within the max-5 monitoring data set. While nearly 20%
had no difference between the duplicate values, on average, there were greater differences in the
duplicate 1-hour values at most of the monitors than was observed for the 5-minute duplicates.
Nearly 20% of the concentrations were noted at or above 100% one another (absolute
difference), however all of these were due to where reported values were zero at the 1-hour
monitor and concentrations of 1 ppb were reported for the continuous-5 monitor. This factor
contributes to the observed positive bias at most of the monitors, however in considering that
these 1-hour SO2 concentrations are below that of potential interest in the exposure and risk
analysis, this degree of limited agreement between the two data sets at these concentration levels
should be acceptable.
Table A.2-4. Distribution of the relative percent difference (RPD) between
duplicate 1-hour SO2 values in the continuous-5 and 1-hour data sets, where
concentrations were < 10 ppb.
Monitor ID
110010041
120890005
290770026
290770037
301110066
301110079
301110082
301110083
371290006
n
2049
25163
24286
24822
6640
7906
7930
4757
27954
Relative Percent Difference (%)1
mean
0
88
91
41
24
119
69
82
-45
std
7
99
99
80
62
95
92
96
83
min
-34
-175
-105
-46
-100
-133
-165
-105
-193
P5
-12
-11
-9
-13
-13
-9
-13
-9
-133
p50
0
15
15
0
0
200
12
15
-59
p95
12
200
200
200
200
200
200
200
200
max
45
200
200
200
200
200
200
200
200
A-90
-------
Monitor ID
420030021
420030064
420030116
420033003
420070005
540990002
541071002
n
4594
5174
4231
4640
30386
6592
23864
Relative Percent Difference (%)1
mean
34
20
3
23
63
1
1
std
81
71
25
69
91
10
11
min
-175
-172
-61
-67
-133
-40
-156
P5
-18
-29
-18
-23
-10
-13
-13
p50
2
-4
0
-1
6
0
0
p95
200
200
19
200
200
19
17
max
200
200
200
200
200
90
200
Notes:
1 the mean, std, min, p5, p50, p95, max are the arithmetic average, standard deviation, minimum,
5th, median, 95th, and maximum, respectively.
When considering duplicate 1-hour concentrations > 10 ppb, the RPD was much lower at
each of these same monitors (Table A.2-5). Most RPD distributions were within +/-5%,
indicating excellent agreement among the duplicate 1-hour values at concentrations above 10
ppb. A very small positive bias may exist with selection of the continuous-5 monitor data for use
in the air quality characterization when compared with the reported 1-hour concentrations, but on
average, the difference was typically less than 1% when considering concentrations above 10
ppb.
Table A.2-5. Distribution of the relative percent difference (RPD) between
duplicate 1-hour SO2 values in the continuous-5 and 1-hour data sets, where
concentrations were > 10 ppb.
Monitor ID
110010041
120890005
290770026
290770037
301110066
301110079
301110082
301110083
371290006
420030021
420030064
420030116
420033003
420070005
540990002
541071002
n
202
2400
1906
1373
1616
71
176
85
3747
1852
2892
1145
2625
15034
2062
10283
Relative Percent Difference (%)1
mean
0
0
0
0
0
0
0
1
1
1
-2
0
-1
0
0
0
std
2
4
2
2
5
3
2
3
25
14
2
9
5
2
2
2
min
-5
-90
-10
-5
-50
-6
-5
-4
-108
-59
-10
-34
-36
-23
-5
-87
P5
-4
-3
-3
-3
-3
-4
-3
-3
-15
-4
-6
-4
-5
-3
-3
-3
p50
0
0
0
0
0
-1
0
1
-2
0
-2
0
-1
0
0
0
p95
4
3
3
3
4
4
4
5
12
4
0
4
2
3
4
3
max
5
34
7
7
173
6
6
20
186
200
11
200
187
73
10
65
Notes:
1 the mean, std, min, p5, p50, p95, max are the arithmetic average, standard deviation, minimum,
5th, median, 95th, and maximum, respectively.
A-91
-------
A.3 Peak-To-Mean Ratio Distributions
Peak-to-mean ratios (PMR) were calculated using the measured values for each the 5-
minute maximum and 1-hour SC>2 concentrations. PMRs were seperated into 19 groups2 based
on the observed variability (3 bins) and concentrations ranges (7 bins) in measured 1-hour
ambient monitor concentrations (n=2,367,686). Table A.3-1 summarizes the PMR distributions
used for estimating 5-minute maximum concentrations from 1-hour measurements where
ambient monitors were characterized by the 1-hour coefficient of variation (COV). These are the
PMR distributions used in the statistical modeling of 5-minute maximum SC>2 concentrations in
the air quality characterization and in the exposure modeling.3 Table A.3-2 summarizes the
PMR distributions used for estimating 5-minute maximum SC>2 concentrations from 1-hour
measurements where ambient monitors were characterized by the 1-hour coefficient of variation
(GSD). Peak-to-mean ratios estimated by categorizing the ambient monitors by GSD were used
only in evaluating an alternative method of estimating 5-minute SC>2 concentrations.
Although there are 21 PMR distributions possible (i.e., 3 x 7), the COV < 100% and GSD <2.17 categories had
only three 1-hour concentrations above 150 ppb. Therefore, the two highest concentration bins do not have a
distribution, and concentrations > 75 ppb constituted the highest concentration bin in the low COV or low GSD bins
3 Note that the minimum and maximum values of each distribution were not used in the final statistical model to
estimate 5-minute maximum concentrations. This was determined in the model evaluations described in section
7.2.5 of the SO, REA.
A-92
-------
Table A.3-1. Distribution of 5-minute maximum peak to 1-hour mean SO2 concentration ratios (PMRs) using ambient
monitors categorized by 1-hour coefficient of variation (COV) and 1-hour mean concentration.
Concbin1
Pet2 -0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
-12
-13
-14
-15
-16
-17
-18
-19
-20
-21
-22
-23
-24
-25
-26
-27
-28
-29
COV < 100%
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.04
1.11
1.11
1.11
1.11
1.13
1.13
1.13
1.13
1.13
1.13
1.14
2
1.00
1.00
1.00
1.00
1.00
1.05
1.06
1.06
1.06
1.07
1.07
1.08
1.08
1.08
1.08
1.08
1.08
1.09
1.09
1.09
1.09
1.10
1.10
1.10
1.10
1.10
1.11
1.12
1.13
1.13
3
1.00
1.03
1.03
1.04
1.04
1.05
1.06
1.06
1.07
1.07
1.07
1.07
1.08
1.08
1.09
1.09
1.10
1.10
1.11
1.11
1.11
1.12
1.12
1.12
1.13
1.13
1.13
1.14
1.14
1.15
4
1.01
1.02
1.02
1.04
1.04
1.05
1.06
1.07
1.08
1.09
1.09
1.10
1.10
1.11
1.11
1.11
1.12
1.13
1.14
1.15
1.15
1.16
1.17
1.17
1.18
1.18
1.19
1.19
1.20
1.20
100% 200 %
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.06
1.11
1.11
1.13
1.13
1.14
1.14
1.14
1.14
1.17
1.17
1.17
1.17
1.17
1.18
1.20
1.20
2
1.00
1.00
1.06
1.08
1.08
1.09
1.10
1.11
1.14
1.15
1.17
1.18
1.20
1.20
1.22
1.24
1.26
1.27
1.30
1.30
1.33
1.34
1.36
1.38
1.40
1.42
1.44
1.46
1.50
1.50
3
1.00
1.12
1.17
1.21
1.24
1.26
1.29
1.31
1.33
1.36
1.38
1.40
1.42
1.44
1.46
1.48
1.50
1.52
1.53
1.55
1.57
1.59
1.61
1.62
1.64
1.66
1.68
1.70
1.71
1.73
4
1.00
1.14
1.18
1.21
1.24
1.26
1.28
1.30
1.32
1.33
1.35
1.37
1.38
1.40
1.42
1.43
1.45
1.46
1.48
1.50
1.51
1.53
1.54
1.56
1.57
1.59
1.60
1.62
1.64
1.65
5
1.08
1.18
1.21
1.22
1.25
1.28
1.31
1.34
1.37
1.40
1.42
1.44
1.47
1.49
1.51
1.54
1.57
1.59
1.60
1.64
1.65
1.68
1.72
1.75
1.76
1.78
1.80
1.81
1.83
1.87
6
1.13
1.25
1.28
1.29
1.30
1.33
1.34
1.37
1.38
1.43
1.47
1.48
1.50
1.51
1.53
1.54
1.57
1.58
1.59
1.59
1.61
1.61
1.63
1.64
1.64
1.67
1.69
1.71
1.73
1.73
A-93
-------
Concbin1
-30
-31
-32
-33
-34
-35
-36
-37
-38
-39
-40
-41
-42
-43
-44
-45
-46
-47
-48
-49
-50
-51
-52
-53
-54
-55
-56
-57
-58
-59
-60
-61
-62
COV<100%
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.05
1.11
1.20
1.25
1.25
1
1.14
1.14
1.14
1.14
1.14
1.15
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.22
1.22
1.24
1.25
1.25
1.25
2
1.13
1.14
1.14
1.15
1.15
1.16
1.17
1.17
1.17
1.18
1.18
1.18
1.18
1.19
1.20
1.20
1.20
1.20
1.20
1.21
1.21
1.22
1.23
1.24
1.25
1.25
1.25
1.27
1.27
1.27
1.29
1.29
1.30
3
1.15
1.16
1.16
1.16
1.17
1.17
1.18
1.18
1.19
1.19
1.19
1.20
1.21
1.21
1.22
1.22
1.23
1.23
1.24
1.24
1.25
1.25
1.26
1.27
1.27
1.28
1.28
1.29
1.30
1.31
1.31
1.32
1.32
4
1.23
1.23
1.23
1.24
1.24
1.24
1.24
1.25
1.26
1.29
1.29
1.29
1.30
1.30
1.31
1.31
1.32
1.32
1.34
1.35
1.35
1.36
1.37
1.39
1.40
1.41
1.42
1.42
1.45
1.45
1.46
1.46
1.47
100% 200 %
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.04
1.11
1.17
1.25
1.25
1
1.20
1.20
1.20
1.20
1.20
1.20
1.22
1.24
1.25
1.27
1.29
1.29
1.33
1.33
1.33
1.34
1.38
1.40
1.40
1.40
1.40
1.43
1.44
1.50
1.50
1.50
1.56
1.57
1.60
1.60
1.63
1.67
1.67
2
1.53
1.55
1.57
1.60
1.62
1.64
1.67
1.69
1.71
1.74
1.76
1.80
1.82
1.84
1.87
1.90
1.92
1.94
2.00
2.00
2.03
2.07
2.09
2.11
2.15
2.18
2.20
2.24
2.27
2.30
2.34
2.38
2.41
3
1.75
1.76
1.78
1.80
1.81
1.83
1.85
1.87
1.88
1.90
1.92
1.94
1.96
1.98
2.00
2.02
2.04
2.06
2.08
2.10
2.12
2.14
2.17
2.19
2.21
2.24
2.26
2.29
2.31
2.34
2.37
2.39
2.42
4
1.67
1.69
1.70
1.73
1.74
1.77
1.78
1.80
1.82
1.84
1.86
1.88
1.90
1.93
1.95
1.97
1.99
2.01
2.04
2.06
2.09
2.12
2.15
2.18
2.21
2.24
2.27
2.30
2.34
2.36
2.40
2.44
2.48
5
1.90
1.91
1.93
1.96
1.97
1.99
2.02
2.05
2.08
2.10
2.14
2.16
2.18
2.20
2.21
2.23
2.24
2.26
2.28
2.30
2.31
2.34
2.36
2.38
2.41
2.43
2.44
2.47
2.50
2.53
2.57
2.60
2.62
6
1.76
1.77
1.78
1.79
1.79
1.80
1.81
1.82
1.82
1.83
1.84
1.84
1.87
1.89
1.91
1.91
1.93
1.94
1.96
1.96
1.97
1.97
1.98
2.01
2.02
2.04
2.06
2.08
2.09
2.13
2.14
2.15
2.17
A-94
-------
Concbin1
-63
-64
-65
-66
-67
-68
-69
-70
-71
-72
-73
-74
-75
-76
-77
-78
-79
-80
-81
-82
-83
-84
-85
-86
-87
-88
-89
-90
-91
-92
-93
-94
-95
COV<100%
0
1.25
1.25
1.30
1.33
1.33
1.33
1.33
1.33
1.43
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.58
1.67
1.75
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
1
1.29
1.29
1.29
1.33
1.33
1.33
1.33
1.33
1.33
1.38
1.38
1.40
1.40
1.40
1.40
1.40
1.43
1.44
1.50
1.50
1.50
1.55
1.57
1.60
1.60
1.63
1.67
1.71
1.78
1.80
1.86
2.00
2.00
2
1.30
1.31
1.31
1.33
1.33
1.35
1.36
1.36
1.38
1.39
1.40
1.40
1.42
1.43
1.45
1.46
1.48
1.50
1.50
1.53
1.55
1.58
1.60
1.63
1.65
1.69
1.71
1.75
1.80
1.83
1.90
1.95
2.05
3
1.33
1.34
1.35
1.36
1.37
1.38
1.39
1.40
1.42
1.42
1.44
1.44
1.46
1.47
1.48
1.50
1.52
1.52
1.54
1.57
1.59
1.61
1.64
1.67
1.70
1.72
1.76
1.80
1.85
1.89
1.96
2.02
2.10
4
1.52
1.54
1.55
1.57
1.58
1.58
1.60
1.64
1.64
1.65
1.65
1.65
1.66
1.67
1.68
1.69
1.70
1.71
1.74
1.75
1.77
1.77
1.82
1.85
1.86
1.88
1.91
1.98
2.10
2.25
2.26
2.30
2.50
100% 200 %
0
1.25
1.33
1.33
1.33
1.33
1.42
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.60
1.67
1.71
1.85
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.44
2.67
3.00
3.00
3.75
1
1.74
1.78
1.80
1.83
1.86
1.89
2.00
2.00
2.00
2.10
2.14
2.18
2.22
2.29
2.34
2.40
2.46
2.56
2.60
2.67
2.78
2.83
3.00
3.00
3.17
3.29
3.40
3.56
3.68
3.86
4.00
4.29
4.57
2
2.45
2.50
2.53
2.56
2.60
2.64
2.69
2.73
2.77
2.82
2.87
2.92
2.96
3.00
3.07
3.13
3.18
3.25
3.31
3.38
3.46
3.54
3.62
3.70
3.80
3.90
4.00
4.12
4.25
4.39
4.56
4.76
5.00
3
2.45
2.48
2.51
2.54
2.57
2.61
2.64
2.68
2.72
2.76
2.80
2.84
2.89
2.93
2.97
3.03
3.09
3.14
3.20
3.26
3.33
3.41
3.48
3.57
3.67
3.77
3.90
4.04
4.18
4.35
4.55
4.77
5.03
4
2.52
2.56
2.60
2.66
2.71
2.76
2.80
2.85
2.89
2.95
3.01
3.06
3.11
3.16
3.22
3.30
3.35
3.41
3.47
3.57
3.65
3.72
3.80
3.90
4.00
4.10
4.21
4.35
4.44
4.62
4.82
5.03
5.24
5
2.64
2.67
2.70
2.73
2.77
2.80
2.84
2.88
2.90
2.93
2.97
2.99
3.02
3.06
3.10
3.16
3.19
3.24
3.26
3.32
3.38
3.42
3.49
3.55
3.62
3.69
3.80
3.88
3.94
4.07
4.18
4.28
4.40
6
2.17
2.19
2.21
2.24
2.27
2.28
2.30
2.31
2.33
2.33
2.35
2.37
2.41
2.44
2.49
2.52
2.53
2.55
2.57
2.60
2.64
2.65
2.67
2.70
2.71
2.74
2.82
2.84
2.94
2.98
3.03
3.09
3.13
A-95
-------
Concbin1
-96
-97
-98
-99
-100
n3
COV<100%
0
2.33
2.67
3.00
4.00
11.67
352735
1
2.14
2.29
2.50
2.89
10.60
74053
2
2.14
2.27
2.50
2.91
10.08
42876
3
2.22
2.37
2.63
3.02
6.81
6895
4
2.53
2.56
3.12
3.61
6.10
147
100% 200 %
0
4.75
6.00
10.00
10.00
11.75
475572
1
4.88
5.29
5.86
6.86
11.50
55341
2
5.27
5.69
6.30
7.27
11.93
35502
3
5.37
5.80
6.51
7.50
11.45
20077
4
5.48
5.94
6.48
7.29
11.39
4019
5
4.48
4.63
5.06
5.36
6.48
989
6
3.33
3.38
3.48
3.70
5.39
341
Notes:
1 1-hour SO2 concentration bins are: 0 = 1-hour mean < 5 ppb; 2 = 5 < 1-hour mean < 10 ppb; 2 = 10 < 1-hour mean < 25 ppb; 3 = 25 < 1-hour mean < 75
ppb; 4 = 75 < 1-hour mean < 150 ppb ; 5 = 150 < 1-hour mean 250 ppb; 6 = 1-hour mean > 250 ppb.
pet -x indicates the percentile of the distribution.
3 n is the number of 5-minute maximum and 1-hour SO2 measurements used to develop distribution.
A-96
-------
Table A.3-2. Distribution of 5-minute maximum peak to 1-hour mean SO2 concentration ratios (PMRs) using ambient
monitors categorized by 1-hour geometric standard deviation (GSD) and 1-hour mean concentration.
Concbin1
Pet2 -0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
-12
-13
-14
-15
-16
-17
-18
-19
-20
-21
-22
-23
-24
-25
-26
-27
-28
-29
GSD < 2.1 7
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.07
1.11
1.11
1.11
1.13
1.13
1.13
1.14
2
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.04
1.06
1.06
1.07
1.07
1.07
1.08
1.08
1.08
1.08
1.09
1.09
1.09
1.09
1.10
1.10
1.10
1.10
1.10
1.11
1.12
1.13
1.14
3
1.00
1.03
1.04
1.04
1.06
1.07
1.08
1.10
1.11
1.12
1.13
1.14
1.14
1.15
1.16
1.17
1.19
1.19
1.20
1.21
1.22
1.23
1.24
1.25
1.26
1.27
1.28
1.29
1.30
1.30
4
1.01
1.07
1.17
1.21
1.22
1.24
1.25
1.27
1.29
1.30
1.30
1.31
1.32
1.32
1.33
1.34
1.35
1.36
1.36
1.37
1.40
1.43
1.44
1.44
1.46
1.47
1.48
1.49
1.50
1.50
2.17 < GSD < 2.94
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.05
1.10
1.11
1.11
1.11
1.13
1.13
1.13
1.13
1.14
1.14
1.14
1.14
1.14
1.17
1.17
1.17
1.17
2
1.00
1.00
1.00
1.00
1.05
1.06
1.06
1.07
1.07
1.08
1.08
1.08
1.09
1.09
1.10
1.10
1.10
1.10
1.11
1.12
1.13
1.13
1.14
1.15
1.15
1.16
1.17
1.17
1.18
1.18
3
1.00
1.00
1.03
1.04
1.05
1.06
1.07
1.08
1.08
1.09
1.10
1.11
1.12
1.12
1.13
1.14
1.15
1.15
1.16
1.17
1.17
1.18
1.19
1.20
1.21
1.21
1.22
1.23
1.24
1.25
4
1.00
1.05
1.08
1.10
1.12
1.14
1.16
1.17
1.18
1.20
1.21
1.23
1.24
1.25
1.26
1.28
1.29
1.30
1.31
1.32
1.34
1.35
1.36
1.38
1.39
1.40
1.41
1.42
1.43
1.44
5
1.05
1.13
1.21
1.23
1.26
1.28
1.29
1.31
1.32
1.34
1.35
1.38
1.39
1.40
1.42
1.43
1.44
1.46
1.47
1.49
1.50
1.52
1.53
1.55
1.57
1.57
1.59
1.60
1.61
1.64
6
1.13
1.19
1.26
1.28
1.29
1.30
1.33
1.34
1.35
1.37
1.38
1.43
1.45
1.46
1.47
1.48
1.50
1.51
1.52
1.53
1.54
1.54
1.56
1.58
1.58
1.58
1.59
1.59
1.61
1.63
GSD > 2.94
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.03
1.05
1.07
1.09
1.10
1.11
1.11
1.13
1.13
1.13
1.14
1.14
1.14
1.15
1.16
1.17
1.17
1.17
1.18
1.19
1.20
1.20
1.20
2
1.00
1.00
1.04
1.06
1.07
1.08
1.08
1.09
1.10
1.10
1.11
1.12
1.13
1.14
1.15
1.16
1.17
1.18
1.18
1.19
1.20
1.20
1.21
1.23
1.24
1.25
1.25
1.27
1.27
1.29
3
1.00
1.05
1.07
1.08
1.09
1.10
1.12
1.13
1.14
1.15
1.16
1.17
1.18
1.19
1.20
1.21
1.22
1.23
1.24
1.25
1.26
1.27
1.29
1.30
1.31
1.32
1.33
1.34
1.35
1.37
4
1.00
1.08
1.10
1.12
1.14
1.15
1.17
1.18
1.19
1.20
1.21
1.22
1.24
1.25
1.26
1.27
1.29
1.30
1.31
1.32
1.33
1.34
1.35
1.36
1.37
1.39
1.40
1.41
1.42
1.43
5
1.07
1.14
1.17
1.20
1.21
1.21
1.22
1.23
1.25
1.27
1.28
1.30
1.32
1.35
1.36
1.38
1.39
1.42
1.43
1.44
1.46
1.49
1.50
1.53
1.56
1.58
1.60
1.62
1.64
1.65
6
1.02
1.14
1.16
1.18
1.24
1.25
1.29
1.30
1.31
1.33
1.36
1.37
1.38
1.45
1.46
1.46
1.47
1.49
1.51
1.54
1.55
1.57
1.58
1.60
1.61
1.64
1.64
1.64
1.68
1.68
A-97
-------
Concbin1
-30
-31
-32
-33
-34
-35
-36
-37
-38
-39
-40
-41
-42
-43
-44
-45
-46
-47
-48
-49
-50
-51
-52
-53
-54
-55
-56
-57
-58
-59
-60
-61
-62
GSD<2.17
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.07
1.14
1.22
1.25
1
1.14
1.14
1.14
1.14
1.15
1.17
1.17
1.17
1.17
1.17
1.17
1.17
1.18
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.20
1.22
1.25
1.25
1.27
1.29
1.29
1.32
1.33
2
1.14
1.15
1.16
1.17
1.17
1.18
1.18
1.18
1.20
1.20
1.20
1.20
1.21
1.22
1.23
1.25
1.25
1.27
1.27
1.27
1.29
1.30
1.30
1.31
1.32
1.33
1.35
1.36
1.37
1.38
1.40
1.40
1.42
3
1.31
1.32
1.33
1.34
1.35
1.36
1.37
1.39
1.40
1.40
1.42
1.43
1.44
1.44
1.46
1.47
1.48
1.50
1.51
1.52
1.54
1.56
1.57
1.59
1.61
1.63
1.65
1.67
1.69
1.71
1.72
1.74
1.75
4
1.52
1.53
1.55
1.57
1.58
1.59
1.60
1.60
1.62
1.63
1.63
1.64
1.65
1.66
1.67
1.70
1.70
1.71
1.72
1.74
1.74
1.75
1.76
1.76
1.77
1.78
1.80
1.81
1.81
1.83
1.83
1.84
1.91
2.17 2.94
0
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.03
1.05
1.11
1.11
1.11
1.15
1.18
1.21
1.25
1.25
1.25
1.25
1.29
1.33
1.33
1.33
1.33
1.38
1.43
1.43
1.48
1.50
1.50
1.50
1.50
1
1.20
1.20
1.22
1.23
1.25
1.25
1.26
1.28
1.29
1.29
1.32
1.33
1.33
1.33
1.35
1.37
1.38
1.40
1.40
1.40
1.41
1.43
1.44
1.47
1.50
1.50
1.50
1.54
1.57
1.58
1.60
1.60
1.63
2
1.30
1.30
1.32
1.33
1.34
1.36
1.36
1.38
1.39
1.40
1.42
1.43
1.45
1.46
1.47
1.50
1.50
1.53
1.54
1.56
1.58
1.59
1.61
1.63
1.65
1.67
1.69
1.71
1.73
1.76
1.79
1.81
1.83
3
1.38
1.39
1.40
1.42
1.43
1.44
1.46
1.47
1.49
1.50
1.52
1.53
1.55
1.56
1.58
1.59
1.61
1.63
1.64
1.66
1.68
1.69
1.71
1.73
1.75
1.77
1.79
1.80
1.82
1.84
1.86
1.88
1.90
4
1.45
1.46
1.47
1.48
1.50
1.51
1.52
1.53
1.54
1.56
1.57
1.58
1.60
1.61
1.62
1.64
1.65
1.67
1.68
1.69
1.71
1.72
1.74
1.77
1.78
1.80
1.82
1.84
1.86
1.88
1.91
1.93
1.96
5
1.67
1.71
1.73
1.75
1.76
1.77
1.80
1.82
1.85
1.90
1.92
1.94
1.96
1.99
2.02
2.03
2.08
2.12
2.16
2.17
2.21
2.22
2.24
2.26
2.28
2.31
2.37
2.38
2.41
2.43
2.47
2.50
2.52
6
1.74
1.76
1.77
1.79
1.81
1.83
1.83
1.84
1.90
1.91
1.91
1.93
1.96
1.96
1.96
1.97
1.98
1.98
2.00
2.01
2.03
2.04
2.06
2.08
2.09
2.11
2.14
2.15
2.16
2.18
2.19
2.20
2.26
A-98
-------
Concbin1
-63
-64
-65
-66
-67
-68
-69
-70
-71
-72
-73
-74
-75
-76
-77
-78
-79
-80
-81
-82
-83
-84
-85
-86
-87
-88
-89
-90
-91
-92
-93
-94
-95
GSD<2.17
0
1.25
1.29
1.33
1.33
1.33
1.33
1.36
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.67
1.75
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.25
2.50
1
1.33
1.33
1.33
1.38
1.40
1.40
1.40
1.40
1.40
1.43
1.43
1.49
1.50
1.50
1.56
1.57
1.60
1.60
1.63
1.67
1.71
1.77
1.80
1.83
1.86
1.97
2.00
2.00
2.14
2.20
2.29
2.40
2.56
2
1.44
1.45
1.47
1.50
1.50
1.53
1.55
1.58
1.60
1.62
1.64
1.67
1.69
1.71
1.73
1.77
1.80
1.83
1.87
1.91
1.94
2.00
2.00
2.09
2.13
2.20
2.27
2.33
2.42
2.54
2.64
2.79
2.93
3
1.78
1.80
1.81
1.84
1.87
1.89
1.91
1.94
1.97
2.00
2.02
2.04
2.07
2.10
2.12
2.15
2.19
2.22
2.27
2.30
2.36
2.43
2.48
2.56
2.63
2.69
2.79
2.88
2.97
3.08
3.24
3.50
3.61
4
1.93
1.93
1.99
2.00
2.05
2.08
2.09
2.11
2.15
2.18
2.19
2.23
2.26
2.27
2.28
2.31
2.40
2.46
2.47
2.47
2.49
2.53
2.68
2.74
2.78
2.81
2.85
2.96
3.06
3.24
3.39
3.55
3.68
2.17 2.94
0
1.50
1.54
1.62
1.67
1.67
1.71
1.80
1.86
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.00
2.29
2.50
2.50
2.75
3.00
3.00
3.33
3.33
4.00
4.67
5.00
5.50
10.00
10.00
1
1.67
1.67
1.70
1.72
1.76
1.80
1.80
1.83
1.86
1.89
1.98
2.00
2.00
2.03
2.13
2.17
2.20
2.24
2.31
2.37
2.40
2.50
2.57
2.63
2.73
2.83
2.94
3.00
3.18
3.33
3.50
3.71
4.00
2
1.86
1.89
1.92
1.94
1.99
2.00
2.03
2.07
2.10
2.14
2.18
2.21
2.25
2.29
2.33
2.38
2.43
2.49
2.54
2.60
2.65
2.71
2.79
2.87
2.94
3.00
3.10
3.20
3.31
3.46
3.61
3.77
4.00
3
1.92
1.95
1.97
2.00
2.02
2.04
2.07
2.10
2.13
2.15
2.18
2.21
2.25
2.28
2.32
2.36
2.40
2.44
2.48
2.53
2.58
2.63
2.69
2.75
2.82
2.89
2.96
3.06
3.16
3.28
3.41
3.57
3.78
4
1.98
2.01
2.04
2.07
2.10
2.14
2.18
2.21
2.24
2.29
2.33
2.37
2.42
2.48
2.54
2.60
2.68
2.77
2.85
2.95
3.05
3.13
3.24
3.35
3.47
3.62
3.73
3.86
4.03
4.22
4.41
4.65
4.94
5
2.56
2.59
2.62
2.63
2.67
2.70
2.72
2.77
2.81
2.85
2.90
2.93
2.98
3.01
3.03
3.09
3.12
3.17
3.21
3.25
3.30
3.35
3.39
3.46
3.54
3.59
3.68
3.78
3.88
3.98
4.10
4.18
4.35
6
2.28
2.29
2.31
2.31
2.33
2.35
2.36
2.37
2.39
2.41
2.44
2.50
2.51
2.53
2.53
2.54
2.56
2.58
2.60
2.64
2.65
2.69
2.71
2.73
2.81
2.84
2.84
2.94
2.97
3.02
3.06
3.12
3.16
A-99
-------
Concbin1
-96
-97
-98
-99
-100
n3
GSD<2.17
0
3.00
3.00
3.50
4.00
11.75
456580
1
2.71
3.00
3.29
4.00
11.57
54454
2
3.13
3.40
3.85
4.68
11.94
16117
3
3.91
4.23
4.71
5.58
10.14
1925
4
3.93
4.13
4.49
5.09
6.10
150
2.17 2.94
0
10.00
10.00
10.00
10.00
11.67
297365
1
4.25
4.67
5.22
6.20
11.67
63582
2
4.23
4.57
5.04
5.91
11.93
55615
3
4.07
4.48
5.06
6.19
11.45
28545
4
5.23
5.59
6.11
6.89
11.39
3952
5
4.44
4.58
4.89
5.45
6.63
759
6
3.27
3.33
3.35
3.51
3.62
224
Notes:
1 1-hour SO2 concentration bins are: 0 = 1-hour mean < 5 ppb; 2 = 5 < 1-hour mean < 10 ppb; 2 = 10 < 1-hour mean < 25 ppb; 3 = 25 < 1-hour mean < 75
ppb; 4 = 75 < 1-hour mean < 150 ppb ; 5 = 150 < 1-hour mean 250 ppb; 6 = 1-hour mean > 250 ppb.
pet -x indicates the percentile of the distribution.
3 n is the number of 5-minute maximum and 1-hour SO2 measurements used to develop distribution.
A-100
-------
A.4 Factors Used in Adjusting Air Quality to Just Meet the Current and Potential
Alternative SO2 Air Quality Standards
The adjustment factors used for simulating just meeting particular forms and levels of
SO2 standards are described here in two sections. This was done given the difference in how the
adjustment factors were derived and applied to each of the air quality scenarios and given the
number of factors generated for the potential alternative standards. The first section includes the
factors used for adjusting air quality to just meet the current standards (either the 24-hour or
annual average), while the second section note the concentrations used in deriving the factors
applied to simulate just meeting potential alternative standards.
A.4.1 Adjustment factors for just meeting the current standard
Both annual and daily adjustment factors were calculated for all selected counties in
evaluating the current annual and daily standards however, the lowest value of the two was
selected for use in adjusting concentrations (see REA section 7.2.4). The adjustment factors for
each county, year, and the standard from which the factors were derived is given in Table A.4-1.
In addion, the coefficient of variation (i.e., COV) was used as a measure to indicate the
variability associated with each of the calculated factors when considering all of the monitors in
a county. Within a given year, the COV generally indicates the extent of spatial variability in
ambient concentrations, considering the number of monitors in operation. Variation in the COV
across different years can indicate the temporal variability in a county however, year-to-year
differences in the number and location of ambient monitors may confound this comparison.
Lower COVs indicate similarity in that concentration metric in the county, while higher values
indicate less homogeneity in concentrations (whether spatially or temporally).
Table A.4-1. Adjustment factors used in simulating air quality just meeting the
current SO2 NAAQS in selected counties by year.
State
Abbreviation
AZ
AZ
AZ
AZ
AZ
AZ
DE
DE
County
Gila
Gila
Gila
Gila
Gila
Gila
New Castle
New Castle
Year
2001
2002
2003
2004
2005
2006
2001
2002
Monitors
(n)
2
2
2
2
2
2
4
4
Adjustment
Factor
3.12
3.53
3.82
3.04
3.33
4.40
3.38
2.67
COV
4
5
12
21
5
1
16
9
Closest
Standard1
D
A
A
A
D
D
D
D
A-101
-------
State
Abbreviation
DE
DE
DE
DE
FL
FL
FL
FL
FL
FL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IA
IA
County
New Castle
New Castle
New Castle
New Castle
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Hillsbo rough
Madison
Madison
Madison
Madison
Madison
Madison
Wabash
Wabash
Wabash
Wabash
Wabash
Wabash
Floyd
Floyd
Floyd
Floyd
Floyd
Floyd
Gibson
Gibson
Gibson
Gibson
Gibson
Gibson
Lake
Lake
Lake
Lake
Lake
Lake
Vigo
Vigo
Vigo
Vigo
Vigo
Vigo
Linn
Linn
Year
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
Monitors
(n)
5
4
4
4
7
7
6
6
6
6
4
4
3
3
3
3
2
2
2
2
2
2
3
3
3
2
3
3
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
5
3
Adjustment
Factor
2.75
2.58
2.73
2.68
3.14
3.09
3.09
4.95
4.40
4.19
3.51
2.88
3.60
3.61
4.19
4.90
3.25
3.33
2.95
3.98
3.80
3.01
3.98
4.85
4.14
5.04
3.98
3.64
2.34
2.68
1.17
2.99
4.78
1.67
4.87
4.43
4.94
4.39
3.39
8.12
2.47
4.65
4.06
5.28
4.57
6.97
3.53
4.70
cov
9
13
11
14
13
16
19
32
25
29
7
12
6
18
11
16
1
3
5
1
7
5
2
6
5
6
11
5
6
19
13
10
3
16
0
17
7
14
16
0
16
18
13
1
5
5
18
5
Closest
Standard1
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
A
A
D
D
D
D
D
D
D
D
D
D
D
D
A
D
A
A
D
D
D
D
D
A-102
-------
State
Abbreviation
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
Ml
Ml
Ml
Ml
Ml
Ml
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
NH
NH
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
County
Linn
Linn
Linn
Linn
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Greene
Greene
Greene
Greene
Greene
Greene
Iron
Iron
Iron
Iron
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Merrimack
Merrimack
Merrimack
Merrimack
Merrimack
Merrimack
Hudson
Hudson
Hudson
Hudson
Hudson
Hudson
Union
Union
Union
Union
Year
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
Monitors
(n)
3
3
3
3
3
3
3
3
3
3
6
3
3
3
3
3
3
5
5
5
5
5
2
2
2
2
3
1
1
1
1
1
2
3
3
2
2
2
2
1
2
2
2
2
2
2
2
2
Adjustment
Factor
3.45
2.29
3.41
4.10
4.20
3.87
4.09
2.78
2.90
2.94
3.21
2.97
3.30
2.99
3.35
2.95
3.57
3.47
5.12
5.29
4.87
4.46
2.26
2.11
2.44
7.96
5.74
3.89
5.65
1.87
2.13
1.93
3.07
3.71
3.31
2.59
2.70
2.51
3.39
5.26
3.52
3.67
3.67
6.25
3.71
3.52
3.70
3.99
cov
5
10
9
35
12
11
11
16
17
10
9
15
5
12
7
13
17
32
26
29
34
19
0
2
2
22
10
0
0
0
0
0
21
18
10
17
18
28
6
0
6
4
1
5
7
11
8
8
Closest
Standard1
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
A
D
D
D
D
D
D
D
D
D
D
D
D
A
A
A
A
A
D
A
A
A
A
A-103
-------
State
Abbreviation
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
OK
PA
PA
PA
PA
County
Union
Union
Bronx
Bronx
Bronx
Bronx
Bronx
Bronx
Chautauqua
Chautauqua
Chautauqua
Chautauqua
Chautauqua
Chautauqua
Erie
Erie
Erie
Erie
Erie
Erie
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Cuyahoga
Lake
Lake
Lake
Lake
Lake
Lake
Summit
Summit
Summit
Summit
Summit
Summit
Tulsa
Tulsa
Tulsa
Tulsa
Tulsa
Tulsa
Allegheny
Allegheny
Allegheny
Allegheny
Year
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
Monitors
(n)
2
2
1
2
2
2
1
2
3
2
2
2
2
2
2
2
2
2
2
2
5
5
5
4
4
4
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
4
7
5
7
7
Adjustment
Factor
4.12
7.98
2.95
3.04
2.82
2.96
3.26
3.44
1.85
2.34
2.30
3.42
5.78
9.47
2.66
2.01
1.85
3.65
4.14
4.72
4.05
5.10
3.98
4.54
3.43
4.25
3.78
3.34
2.79
3.05
1.87
2.51
3.25
2.39
2.65
2.75
3.76
3.79
4.16
4.51
3.65
4.07
4.57
5.69
2.72
2.80
2.23
2.81
cov
7
4
0
3
1
3
0
6
12
18
13
16
11
2
13
16
16
20
14
17
6
11
5
11
6
8
8
15
10
13
13
16
3
8
2
11
14
9
10
2
6
3
4
59
5
4
5
6
Closest
Standard1
A
D
A
A
D
A
A
A
D
D
D
D
D
D
D
D
D
D
D
D
D
A
D
D
D
D
A
A
D
D
D
D
D
D
D
D
A
D
A
D
D
D
A
D
D
A
D
D
A-104
-------
State
Abbreviation
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
County
Allegheny
Allegheny
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Northampton
Northampton
Northampton
Northampton
Northampton
Northampton
Warren
Warren
Warren
Warren
Warren
Warren
Washington
Washington
Washington
Washington
Washington
Washington
Blount
Blount
Blount
Blount
Blount
Blount
Shelby
Shelby
Shelby
Shelby
Shelby
Shelby
Sullivan
Sullivan
Sullivan
Sullivan
Sullivan
Sullivan
Jefferson
Jefferson
Jefferson
Jefferson
Year
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
Monitors
(n)
7
6
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
2
2
2
2
2
2
3
3
3
3
4
3
2
2
2
2
2
2
3
3
3
3
Adjustment
Factor
2.17
2.97
2.01
1.91
1.73
2.64
2.42
2.67
2.15
5.01
3.73
2.28
3.55
0.98
1.66
1.45
1.40
2.37
1.91
1.68
2.95
3.11
2.99
3.42
3.07
3.48
1.62
2.05
1.88
2.22
1.61
1.79
3.47
4.79
3.75
4.46
3.90
4.12
2.95
3.26
3.28
3.33
3.72
3.33
2.68
4.82
4.30
4.47
cov
7
8
5
6
6
6
7
8
28
0
18
21
3
19
11
15
11
15
17
19
6
6
8
2
5
6
18
10
12
1
7
10
19
20
21
20
46
44
8
10
4
3
4
3
8
4
4
13
Closest
Standard1
D
D
D
D
D
A
A
D
A
A
A
A
A
D
D
D
D
D
D
D
A
A
A
A
D
A
D
D
D
D
D
D
D
D
D
D
D
D
A
D
D
D
D
D
D
D
D
D
A-105
-------
State
Abbreviation
TX
TX
VA
VA
VA
VA
VA
VA
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
WV
VI
VI
VI
VI
VI
VI
County
Jefferson
Jefferson
Fairfax
Fairfax
Fairfax
Fairfax
Fairfax
Fairfax
Brooke
Brooke
Brooke
Brooke
Brooke
Brooke
Hancock
Hancock
Hancock
Hancock
Hancock
Hancock
Monongalia
Monongalia
Monongalia
Monongalia
Monongalia
Monongalia
Wayne
Wayne
Wayne
Wayne
Wayne
St Croix
St Croix
St Croix
St Croix
St Croix
St Croix
Year
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2006
2001
2002
2003
2004
2005
2001
2002
2003
2004
2005
2006
Monitors
(n)
3
3
2
2
3
3
3
3
2
2
2
2
2
2
9
9
9
7
7
7
3
2
2
2
2
2
4
4
4
3
3
5
5
5
5
5
5
Adjustment
Factor
5.67
4.31
4.50
4.49
4.89
4.80
4.79
5.35
2.13
2.49
2.63
2.02
2.16
2.50
2.20
2.38
2.30
2.38
2.22
2.34
2.37
2.22
3.26
3.25
3.13
3.20
2.85
3.31
3.41
2.87
2.02
3.41
3.46
3.66
3.26
9.25
4.59
cov
7
4
18
14
15
19
19
18
5
4
3
6
5
8
3
3
3
4
5
4
3
2
1
1
3
1
4
3
7
9
11
83
64
66
56
15
25
Closest
Standard1
D
D
A
A
A
A
A
A
A
A
A
A
A
A
A
D
D
A
A
A
D
D
D
D
A
D
D
A
D
D
D
D
D
D
D
D
D
Notes:
1 Ambient SO2 concentrations were closest to either the annual (A) or daily (D) NAAQS level.
A-106
-------
A.4.2 Adjustment factors for just meeting the potential alternative standards
Five potential alternative standards (i.e., 50, 100, 150, 200, and 250 ppb daily maximum
1-hour) given a 99th percentile form and one alternative standard (200 ppb daily maximum 1-
hour) given a 98th percentile form were selected for evaluation (for details, see REA chapter 5).
Adjustment factors were derived for each of two 3-year groups of recent air quality (i.e., 2001-
2003 and 2004-2006). For the sake of brevity, only the maximum 3-year averaged
concentrations for each of the percentile forms are provided in Table A.4-2, rather than all of the
adjustment factors. The actual adjustment factors used in simulating air quality can be derived
for each of the concentration levels by dividing by the county concentration for each year goup.
For example, the adjustement factor applied to the 2002 hourly mean concentrations in New
Castle DE to simulate just meeting a 99th percentile daily maximum 1-hour of 100 ppb is
100/164 = 0.61. That is to say, to meet this particular standard, the hourly concentrations need to
be adjusted downward by a factor of 0.61. The COV is also used to represent the temporal
variability over the three years of monitoring (where such data exist).
Table A.4-2. Concentrations used in developing adjustment factors when
simulating air quality just meeting potential alternative SO2 NAAQS in selected
counties by year.
State
Abbreviation
AZ
AZ
DE
DE
FL
FL
IA
IA
IA
IA
IL
IL
IL
IL
IN
IN
IN
County
Gila
Gila
New Castle
New Castle
Hillsbo rough
Hillsbo rough
Linn
Linn
Muscatine
Muscatine
Madison
Madison
Wabash
Wabash
Floyd
Floyd
Gibson
Year Group
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
98th Percentile
Years
(n)
3
2
2
3
3
2
3
3
3
3
3
3
1
1
3
3
2
Cone
(ppb)
226
222
138
123
117
93
82
96
92
120
110
123
139
131
124
129
185
COV
(%)
10
6
5
20
12
8
21
17
13
10
22
5
17
14
12
99th Percentile
Years
(n)
3
2
2
3
3
2
3
3
3
3
3
3
1
1
1
3
2
Cone
(ppb)
260
294
164
147
146
128
105
111
113
135
144
144
216
187
151
170
235
COV
(%)
10
1
0
31
2
8
12
27
9
8
24
7
6
19
A-107
-------
State
Abbreviation
IN
IN
IN
IN
IN
Ml
Ml
MO
MO
MO
MO
MO
MO
NH
NH
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OH
OK
OK
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
TN
TN
TN
County
Gibson
Lake
Lake
Vigo
Vigo
Wayne
Wayne
Greene
Greene
Iron
Iron
Jefferson
Jefferson
Merrimack
Merrimack
Hudson
Hudson
Union
Union
Bronx
Bronx
Chautauqua
Chautauqua
Erie
Erie
Cuyahoga
Cuyahoga1
Cuyahoga1
Lake
Lake
Summit
Summit
Tulsa
Tulsa
Allegheny
Allegheny
Beaver
Beaver
Northampton
Northampton
Warren
Warren
Washington
Washington
Blount
Blount
Shelby
Year Group
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
98th Percentile
Years
(n)
1
3
2
3
2
2
3
3
3
3
1
1
3
3
3
2
2
3
2
2
2
3
3
3
3
2
3
3
3
3
3
3
3
2
1
2
3
3
3
3
3
3
3
3
1
3
3
Cone
(ppb)
199
68
87
114
110
102
115
81
63
289
20
230
244
110
127
54
51
52
49
64
59
238
84
206
114
76
67
67
129
146
131
133
63
82
149
144
200
188
55
92
218
180
99
89
189
168
70
cov
(%)
5
1
7
8
3
2
13
29
20
10
30
2
9
3
13
10
1
7
2
47
10
33
1
8
18
10
5
12
9
22
32
16
28
6
9
41
6
22
10
10
5
29
99th Percentile
Years
(n)
1
2
2
3
2
2
3
3
3
3
1
1
3
3
3
2
2
3
2
2
2
3
3
3
3
2
3
3
3
3
3
3
2
1
2
3
3
3
3
3
3
3
3
1
3
3
Cone
(ppb)
226
84
113
159
136
126
128
94
81
341
22
234
346
125
151
61
65
57
60
71
68
285
101
225
129
101
80
145
175
148
150
76
93
164
183
245
228
65
146
270
226
111
102
204
194
101
COV
(%)
52
3
25
2
4
2
13
25
9
16
34
9
1
1
7
9
7
2
12
54
8
24
1
9
4
9
12
13
7
33
36
31
8
3
65
12
15
11
11
6
35
A-108
-------
State
Abbreviation
TN
TN
TN
TN
TX
TX
VA
VA
VI
VI
WV
WV
WV
WV
WV
WV
WV
WV
County
Shelby1
Shelby1
Sullivan
Sullivan
Jefferson
Jefferson
Fairfax
Fairfax
St Croix
St Croix
Brooke
Brooke
Hancock
Hancock
Monongalia
Monongalia
Wayne
Wayne
Year Group
2004-2006
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
2001-2003
2004-2006
98th Percentile
Years
(n)
3
3
3
3
3
3
3
3
2
1
3
3
3
3
3
2
3
2
Cone
(ppb)
72
72
157
145
92
109
38
37
103
70
154
125
182
134
163
148
93
67
cov
(%)
35
2
13
7
20
49
15
8
6
20
8
17
24
22
3
7
11
99th Percentile
Years
(n)
3
3
3
3
3
3
3
2
1
3
3
3
3
3
2
3
2
Cone
(ppb)
85
195
208
103
129
48
41
126
130
180
158
217
159
218
188
109
75
COV
(%)
33
19
17
16
46
24
11
18
17
19
23
19
26
16
14
0
Notes:
1 Two monitors in the county had the same average 98th percentile daily 1-hour maximum
concentrations. Concentrations, monitoring years, and COVs for both monitors are indicated.
A-109
-------
A.5 Supplementary Results Tables for 5-minute Measurement Data
A-110
-------
Table A.5-1. Annual average SO2 concentrations and number of measured 5-minute daily maximum SO2
concentrations above potential health effect benchmark levels. Data used were from 98 monitors that reported both
the 5-minute maximum and 1-hour SO2 concentrations for years 1997 through 2007.
State
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
AR
CO
CO
CO
CO
CO
CO
CO
CO
County
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Pulaski
Union
Union
Union
Union
Union
Union
Union
Union
Union
Union
Union
Denver
Denver
Denver
Denver
Denver
Denver
Denver
Denver
Monitor ID
051190007
051190007
051190007
051190007
051190007
051190007
051191002
051191002
051191002
051191002
051191002
051390006
051390006
051390006
051390006
051390006
051390006
051390006
051390006
051390006
051390006
051390006
080310002
080310002
080310002
080310002
080310002
080310002
080310002
080310002
Year
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2001
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2001
2002
2003
2004
Days
(n)
339
365
359
350
365
90
365
329
275
352
364
365
313
275
357
364
275
364
334
249
365
90
365
360
156
137
360
365
362
337
Hours
(n)
7138
7799
7687
6702
8356
2062
6607
5997
3833
5596
6529
7624
6766
5101
5792
7474
6296
7239
4267
4922
8364
2061
7014
4311
1626
2434
5575
6830
6250
4412
Annual Hourly (ppb)
Mean
2.76
2.47
2.08
1.91
3.2
2.88
2.33
1.62
2.31
2.38
2.28
5.27
6.4
5.39
6.21
3.09
2.92
2.14
2.15
2.36
2.89
2.99
6.77
7.37
6.77
6.53
6.63
5.36
3.83
3.68
std
1.43
1.3
1.61
1.17
1.13
1.12
1.5
1.3
1.51
1.63
1.18
11.3
7.45
6.94
10.95
3.86
2.27
5.13
2.74
2.58
2.19
1.3
9.36
9.45
8.21
8.62
8.85
7.27
4.62
4.09
GM
2.44
2.18
1.69
1.65
3.03
2.71
1.99
1.35
1.85
1.77
2.02
3.28
5.14
3.66
3.76
2.28
2.5
1.59
1.63
1.94
2.61
2.81
3.75
4.29
4.01
3.84
3.84
3.11
2.54
2.48
GSD
1.65
1.64
1.84
1.69
1.39
1.39
1.74
1.74
2.04
2.44
1.63
2.15
1.73
2.44
2.29
2.06
1.65
1.88
1.89
1.76
1.49
1.39
2.86
2.79
2.76
2.69
2.75
2.67
2.34
2.31
Number of 5-minute Daily Maximum
> 100 ppb
1
0
0
0
0
0
0
0
0
0
0
30
17
12
44
5
1
2
3
2
1
0
23
18
3
4
8
6
1
0
> 200 ppb
0
0
0
0
0
0
0
0
0
0
0
11
3
1
7
1
0
2
2
1
1
0
0
2
0
0
0
0
0
0
> 300 ppb
0
0
0
0
0
0
0
0
0
0
0
5
1
0
2
1
0
2
0
0
1
0
0
0
0
0
0
0
0
0
> 400 ppb
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
A-lll
-------
State
CO
CO
DE
DE
DC
DC
DC
DC
DC
DC
FL
FL
FL
FL
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
County
Denver
Denver
New Castle
New Castle
District of Columbia
District of Columbia
District of Columbia
District of Columbia
District of Columbia
District of Columbia
Nassau
Nassau
Nassau
Nassau
Cerro Gordo
Cerro Gordo
Cerro Gordo
Cerro Gordo
Cerro Gordo
Clinton
Clinton
Clinton
Clinton
Clinton
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Monitor ID
080310002
080310002
100031008
100031008
110010041
110010041
110010041
110010041
110010041
110010041
120890005
120890005
120890005
120890005
190330018
190330018
190330018
190330018
190330018
190450019
190450019
190450019
190450019
190450019
191390016
191390016
191390016
191390016
191390016
191390017
191390017
191390017
191390017
191390017
Year
2005
2006
1997
1998
2000
2001
2002
2003
2004
2007
2002
2003
2004
2005
2001
2002
2003
2004
2005
2001
2002
2003
2004
2005
2001
2002
2003
2004
2005
2001
2002
2003
2004
2005
Days
(n)
337
349
330
257
160
358
365
181
119
268
357
365
275
175
38
254
296
366
173
70
345
333
353
177
91
365
353
365
181
83
364
365
363
181
Hours
(n)
3599
6199
7490
4898
3731
7774
8365
4267
2765
6394
8415
8662
6507
4120
513
3325
5032
8141
3528
1276
6516
5939
7093
3323
1733
7391
6570
6664
3629
1373
7242
7586
7322
3441
Annual Hourly (ppb)
Mean
3.92
3.38
10.29
8.86
8.64
7
6.89
8.63
7.88
5.05
6.39
3.44
3.2
4.06
1.22
1.16
1.88
0.8
0.69
2.14
3.29
2.89
2.83
3.99
3.27
4.07
3.87
3.92
4.22
2.14
3.12
3.93
3.56
3.16
std
4.2
3.62
17.99
14.99
6.17
6.51
5.62
5.92
5.51
3.74
15.33
8.95
7.18
10.16
3.38
3.83
7.57
2.84
1.49
1.69
3.37
3.2
3.06
4.31
4.61
5.36
7.01
5.67
7.55
1.86
3.82
4.26
3.92
4.14
GM
2.57
2.33
5.23
4.35
7.25
4.83
5.29
7.28
6.3
4.24
2.65
1.6
1.68
1.65
0.44
0.33
0.27
0.23
0.31
1.52
1.96
1.68
1.67
2.35
1.89
2.78
2.21
2.43
2.34
1.42
2.05
2.69
2.24
2.03
GSD
2.42
2.26
2.86
3.03
1.77
2.45
2.11
1.75
2.06
1.76
2.95
2.5
2.37
2.71
3.16
4.07
4.83
3.4
3.16
2.54
3.02
3.14
3.12
3.11
2.93
2.39
2.86
2.7
2.79
2.66
2.62
2.51
2.79
2.59
Number of 5-minute Daily Maximum
> 100 ppb
0
1
103
64
1
3
1
5
1
1
69
26
11
26
0
0
4
0
0
0
3
4
3
2
0
4
4
5
9
0
3
4
2
4
> 200 ppb
0
0
33
16
0
1
0
1
0
1
23
5
5
4
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
> 300 ppb
0
0
1
2
0
1
0
1
0
1
6
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
> 400 ppb
0
0
0
0
0
0
0
1
0
0
2
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A-112
-------
State
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
IA
LA
LA
LA
LA
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Muscatine
Muscatine
Muscatine
Muscatine
Muscatine
Scott
Scott
Scott
Scott
Scott
Van Buren
Van Buren
Van Buren
Van Buren
Van Buren
Van Buren
Wood bury
Wood bury
West Baton Rouge
West Baton Rouge
West Baton Rouge
West Baton Rouge
Buchanan
Buchanan
Buchanan
Buchanan
Buchanan
Buchanan
Buchanan
Buchanan
Greene
Greene
Greene
Greene
Monitor ID
191390020
191390020
191390020
191390020
191390020
191630015
191630015
191630015
191630015
191630015
191770005
191770005
191770005
191770005
191770006
191770006
191930018
191930018
221210001
221210001
221210001
221210001
290210009
290210009
290210009
290210009
290210011
290210011
290210011
290210011
290770026
290770026
290770026
290770026
Year
2001
2002
2003
2004
2005
2001
2002
2003
2004
2005
2001
2002
2003
2004
2004
2005
2001
2002
1997
1998
1999
2000
1997
1998
1999
2000
2000
2001
2002
2003
1997
1998
1999
2000
Days
(n)
92
363
365
366
181
85
364
364
336
177
65
353
358
305
53
181
85
280
277
353
354
361
361
364
362
264
72
329
331
253
339
350
362
366
Hours
(n)
1909
7682
7695
7757
3931
1345
7505
7451
6696
3436
597
6350
7118
5011
877
3349
1578
3875
4966
7566
7272
7360
8484
8161
7415
5297
1672
6412
6457
5141
4763
5810
7242
8721
Annual Hourly (ppb)
Mean
5.36
5.27
5.31
7.36
5.55
1.15
2.28
2.09
2.11
2.56
0.9
1.03
1.1
0.88
0.85
0.9
1.32
1.5
7.04
7.52
6.4
7.3
8.3
7.06
2.77
2.37
5.27
3.7
4.01
4.06
4.32
5.73
4.09
4.97
std
9.76
10.27
11.2
15.39
13.61
2.13
3.17
2.68
2.65
3.05
0.92
0.92
0.91
1.45
0.94
0.79
2.28
2.94
12.51
10.67
9.59
11.13
31.64
24.17
3.07
3.04
8.53
5.3
7.33
7.04
9.65
11.66
7.53
10.21
GM
2.04
2.22
2.11
3.02
2.02
0.45
0.87
1.02
1.06
1.17
0.64
0.72
0.78
0.5
0.55
0.69
0.77
0.7
3.94
5.03
4.01
4.51
2.77
2.8
2.08
1.81
3.45
2.52
2.52
2.59
2.02
2.35
2.22
2.41
GSD
3.95
3.61
3.71
3.2
3.64
4
4.7
3.79
3.68
4.17
2.33
2.48
2.48
2.87
2.53
2.09
2.45
3.14
2.65
2.29
2.44
2.46
2.8
2.64
2
1.88
2.15
2.15
2.23
2.25
2.69
3.07
2.5
2.67
Number of 5-minute Daily Maximum
> 100 ppb
1
31
42
60
27
0
0
0
0
0
0
0
0
0
0
0
0
0
42
50
55
76
94
92
3
7
8
6
21
13
20
39
13
52
> 200 ppb
0
1
5
14
12
0
0
0
0
0
0
0
0
0
0
0
0
0
13
18
12
26
79
67
0
0
0
0
0
0
2
1
1
1
> 300 ppb
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
4
2
1
7
57
44
0
0
0
0
0
0
0
0
0
0
> 400 ppb
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
39
19
0
0
0
0
0
0
0
0
0
0
A-113
-------
State
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Greene
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Iron
Monitor ID
290770026
290770026
290770026
290770026
290770026
290770026
290770026
290770037
290770037
290770037
290770037
290770037
290770037
290770037
290770037
290770037
290770037
290770037
290930030
290930030
290930030
290930030
290930030
290930030
290930030
290930030
290930031
290930031
290930031
290930031
290930031
290930031
290930031
290930031
Year
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2001
2002
2003
2004
1997
1998
1999
2000
2001
2002
2003
2004
Days
(n)
365
360
362
274
365
365
272
356
361
363
341
355
335
363
274
365
365
272
365
365
356
324
356
354
363
90
352
363
341
332
365
365
350
91
Hours
(n)
8304
7054
7935
6574
8756
8753
6520
6559
8134
8554
5318
6707
6373
8179
6575
8760
8745
6496
8575
8475
6546
4071
5388
7960
6963
1846
6177
7991
7918
5170
8426
8665
8230
2172
Annual Hourly (ppb)
Mean
4.52
4.28
3.5
3.21
2.95
3.15
3.2
4.98
4.27
3.13
6.36
4.04
4
3.32
2.71
3.05
3.26
2.42
8.24
7.9
9.33
14.3
9.32
6.95
7.58
2.47
8.09
7.56
8.41
8.27
6.62
6.32
6.6
3.82
std
9.62
9.08
6.16
6.41
5.94
6.77
7.07
14.73
7.37
7.72
17.9
10.65
9.68
6.96
4.79
6.06
8.44
6.03
26.43
25.09
28.07
46.11
32.18
23.55
23.2
2.56
24.57
22.94
25.99
24.93
23.42
18.53
21.05
2.74
GM
2.17
1.94
2.02
1.64
1.58
1.58
1.59
1.89
2.76
1.72
2.13
1.91
2.15
1.93
1.79
1.93
1.57
1.37
3.12
2.73
3.29
3.2
2.37
2.2
2.69
1.76
2.92
3.03
3.93
2.81
2.47
3.19
2.89
3.2
GSD
2.63
2.72
2.36
2.45
2.35
2.42
2.43
2.78
2.18
2.23
3.04
2.49
2.27
2.21
2.05
2.11
2.38
2.08
2.89
2.99
3.09
3.95
3.41
2.98
2.94
2.11
3.17
2.94
2.63
3.21
2.79
2.35
2.64
1.74
Number of 5-minute Daily Maximum
> 100 ppb
36
27
5
3
5
8
9
52
30
31
46
37
40
19
13
20
37
16
93
85
83
95
88
99
99
0
77
88
92
86
95
88
88
0
> 200 ppb
0
0
0
0
0
0
0
21
2
3
23
9
11
1
0
1
4
0
78
70
74
77
74
73
81
0
55
57
54
53
60
54
54
0
> 300 ppb
0
0
0
0
0
0
0
8
0
0
3
2
1
0
0
0
0
0
63
62
63
69
64
58
64
0
37
37
37
35
40
28
39
0
> 400 ppb
0
0
0
0
0
0
0
5
0
0
0
0
0
0
0
0
0
0
54
52
49
55
56
52
48
0
27
22
23
23
22
19
23
0
A-114
-------
State
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
County
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Jefferson
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Monroe
Pike
Pike
Pike
Saint Charles
Saint Charles
Saint Charles
Saint Charles
Monitor ID
290990004
290990004
290990004
290990004
290990014
290990014
290990014
290990014
290990014
290990017
290990017
290990017
290990017
290990018
290990018
290990018
291370001
291370001
291370001
291370001
291370001
291370001
291370001
291370001
291370001
291370001
291370001
291630002
291630002
291630002
291830010
291830010
291831002
291831002
Year
2004
2005
2006
2007
1997
1998
1999
2000
2001
1998
1999
2000
2001
2001
2002
2003
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2005
2006
2007
1997
1998
1997
1998
Days
(n)
346
351
343
90
359
365
363
361
132
289
360
355
74
219
352
272
364
364
365
366
309
321
336
316
348
338
51
311
348
68
365
230
365
362
Hours
(n)
8033
7144
6524
2125
7174
7770
7591
6588
2433
5721
7289
7153
1044
3492
6305
6009
8280
8411
8714
8617
4346
5358
5948
5123
6518
6169
526
4879
6469
1019
8152
4810
8514
8122
Annual Hourly (ppb)
Mean
10.32
11.41
13.12
6.31
8.38
4.57
4.6
3.87
3.15
7.37
8.65
6.06
7.72
5.33
5.51
4.41
2.92
2.35
3.58
2.93
1.78
1.81
1.82
2.29
2.03
1.73
1.86
4.37
3.94
3.08
4.35
4.32
5.72
6.31
std
22.63
24.87
27.2
11.92
19
9.67
9.49
7.06
5.64
18.87
22.19
16.54
16.53
11.74
14.84
10.38
2.86
2.25
2.36
2.06
1.44
1.48
1.48
2.31
1.81
1.26
2
5.43
4.67
3.69
7.95
5.69
6.95
7.9
GM
4.78
4.62
4.3
3.08
4.14
2.62
2.48
2.36
1.95
3.47
3.8
2.87
3.69
2.53
2.59
2.4
2.38
1.86
3.13
2.54
1.47
1.48
1.51
1.77
1.63
1.47
1.48
2.89
2.78
2.09
2.6
2.77
3.65
4.02
GSD
2.96
3.34
4.02
2.88
2.79
2.48
2.57
2.35
2.25
2.87
3.01
2.77
3.02
2.84
2.75
2.54
1.79
1.87
1.63
1.65
1.74
1.75
1.73
1.91
1.81
1.68
1.8
2.33
2.2
2.24
2.45
2.38
2.5
2.5
Number of 5-minute Daily Maximum
> 100 ppb
106
118
134
21
87
37
32
28
7
59
90
59
13
34
56
27
0
0
0
0
0
0
0
0
0
0
0
5
3
0
5
1
23
25
> 200 ppb
41
68
78
8
54
23
19
7
1
33
57
40
9
18
36
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
0
> 300 ppb
26
47
53
4
31
13
11
4
0
22
42
26
5
9
24
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
> 400 ppb
13
28
41
1
23
6
5
2
0
16
29
17
3
6
18
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
A-115
-------
State
MO
MO
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
MT
NC
NC
NC
County
Saint Charles
Saint Charles
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Yellowstone
Forsyth
Forsyth
Forsyth
Monitor ID
291831002
291831002
301110066
301110066
301110066
301110066
301110066
301110066
301110066
301110079
301110079
301110079
301110079
301110080
301110080
301110080
301110080
301110080
301110082
301110082
301110082
301110083
301110083
301110083
301110083
301110083
301110084
301110084
301110084
301110084
301112008
370670022
370670022
370670022
Year
1999
2000
1997
1998
1999
2000
2001
2002
2003
1997
2001
2002
2003
1997
1998
1999
2000
2001
2001
2002
2003
1999
2000
2001
2002
2003
2003
2004
2005
2006
1997
1997
1998
1999
Days
(n)
363
331
362
357
352
355
365
364
347
180
55
353
350
363
358
350
360
150
169
365
361
112
341
357
360
166
99
294
291
273
177
362
364
352
Hours
(n)
7969
6421
6873
7198
5767
6099
6872
8347
5691
3166
837
8034
5107
5433
5371
5588
5999
2015
2605
8212
5173
2087
3845
5604
6847
1641
759
2465
2577
1983
2579
7822
7122
6428
Annual Hourly (ppb)
Mean
5.61
4.6
8.06
7.14
7.75
7.72
7.77
6.81
7.37
3.84
4.64
1.9
3.02
7.54
6.85
6.36
6.22
5.55
4.19
2.32
2.93
8.07
4.68
4.36
2.31
2.29
2.99
3.48
2.96
2.75
3.96
7.06
6.98
5.85
std
7.24
5.45
10.76
9.49
9.65
10.26
10.46
11.61
9.92
4.06
3.71
1.91
2.55
10.11
9.12
7.81
7.65
6.3
4.62
2.77
3.25
8.01
5.36
5.59
3.21
3.08
4.51
5.45
4.98
4.56
4.57
6.91
7.54
5.92
GM
3.58
3.01
4.4
4
4.31
4.25
4.13
3.46
4.06
2.65
3.43
1.48
2.3
4.29
3.98
3.79
3.68
3.54
2.87
1.7
2.11
5.01
3
2.71
1.65
1.62
1.99
2.14
1.79
1.71
2.65
5.13
4.72
4.13
GSD
2.5
2.42
3
2.9
2.99
2.97
3.06
3.04
2.93
2.28
2.23
1.83
2.06
2.86
2.79
2.75
2.74
2.56
2.32
1.99
2.11
2.81
2.49
2.51
1.98
1.99
2.19
2.37
2.28
2.23
2.35
2.2
2.48
2.29
Number of 5-minute Daily Maximum
> 100 ppb
17
5
45
42
34
66
56
52
39
1
0
0
0
59
38
47
59
12
1
0
1
4
10
11
1
0
0
2
2
1
2
10
13
3
> 200 ppb
2
0
7
5
5
7
4
4
2
0
0
0
0
11
14
7
10
2
0
0
1
0
1
1
0
0
0
0
0
0
0
0
1
0
> 300 ppb
0
0
1
1
0
2
0
2
0
0
0
0
0
3
6
4
1
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
> 400 ppb
0
0
1
0
0
1
0
1
0
0
0
0
0
0
0
2
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
A-116
-------
State
NC
NC
NC
NC
NC
NC
NC
NC
NC
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
County
Forsyth
Forsyth
Forsyth
Forsyth
Forsyth
New Hanover
New Hanover
New Hanover
New Hanover
Billings
Billings
Billings
Billings
Billings
Billings
Billings
Billings
Billings
Billings
Billings
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burke
Burleigh
Burleigh
Monitor ID
370670022
370670022
370670022
370670022
370670022
371290006
371290006
371290006
371290006
380070002
380070002
380070002
380070002
380070002
380070002
380070002
380070002
380070002
380070002
380070003
380130002
380130002
380130002
380130002
380130002
380130002
380130002
380130004
380130004
380130004
380130004
380130004
380150003
380150003
Year
2000
2001
2002
2003
2004
1999
2000
2001
2002
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1999
2000
2001
2002
2003
2004
2005
2003
2004
2005
2006
2007
2005
2006
Days
(n)
266
361
362
363
259
360
335
358
352
143
276
248
283
275
26
164
128
106
43
167
297
347
338
346
353
340
263
63
315
244
302
99
60
294
Hours
(n)
5203
7634
7022
8075
4710
8208
7980
8168
8028
1940
3216
2724
2860
3113
341
1256
835
418
221
2657
3852
5268
5653
5367
6328
5229
3098
882
3198
2238
3152
1227
683
3686
Annual Hourly (ppb)
Mean
5.52
5.12
6.12
5.87
5.56
4.1
4.67
5.71
6.44
1.31
1.38
1.42
1.37
1.43
1.48
1.24
1.44
1.53
1.5
1.72
2.79
2.96
2.72
2.64
2.6
2.77
2.88
2.89
2.76
2.47
2.27
3.8
3.4
2.33
std
5.58
5.64
8.19
6.19
8.21
8.34
8.92
13.73
13.85
1.04
1.04
1.1
1.12
1.11
0.87
0.85
0.92
1.25
1.26
1.52
4.61
5.77
4.97
4.72
4.77
5.03
4.99
3.99
3.59
3.18
3.16
5.18
2.97
2.6
GM
3.77
3.46
3.87
4.17
3.37
1.92
2.13
2.08
2.61
1.16
1.21
1.24
1.2
1.26
1.32
1.13
1.27
1.29
1.29
1.43
1.65
1.77
1.62
1.58
1.62
1.65
1.67
1.84
1.83
1.72
1.59
2.27
2.47
1.68
GSD
2.39
2.38
2.51
2.24
2.55
2.73
2.87
3.09
3.12
1.48
1.53
1.56
1.51
1.53
1.54
1.41
1.55
1.64
1.6
1.7
2.31
2.27
2.25
2.24
2.16
2.26
2.33
2.26
2.21
2.09
2.02
2.53
2.2
2.04
Number of 5-minute Daily Maximum
> 100 ppb
1
5
15
11
6
54
76
109
127
0
0
0
0
0
0
0
0
0
0
0
3
7
3
4
7
6
4
0
0
0
1
1
0
0
> 200 ppb
0
0
3
0
1
8
6
54
39
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
> 300 ppb
0
0
0
0
0
4
3
10
7
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
> 400 ppb
0
0
0
0
0
3
0
3
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
A-117
-------
State
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
County
Burleigh
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Cass
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
Dunn
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
Monitor ID
380150003
380171003
380171003
380171004
380171004
380171004
380171004
380171004
380171004
380171004
380171004
380171004
380171004
380250003
380250003
380250003
380250003
380250003
380250003
380250003
380250003
380250003
380250003
380250003
380530002
380530002
380530002
380530002
380530002
380530002
380530002
380530002
380530002
380530104
Year
2007
1997
1998
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
2001
2002
2003
2004
2005
2006
2007
1998
Days
(n)
97
206
132
162
246
213
203
274
200
256
146
358
116
224
242
323
353
276
334
355
347
183
262
79
238
144
108
262
305
303
225
276
73
224
Hours
(n)
947
2254
2943
2501
3325
1868
1686
2476
1297
3140
928
7385
2256
3313
2688
5099
7455
3575
4484
7289
6019
1314
2213
667
2552
1989
754
3361
5345
4614
2515
2896
511
1525
Annual Hourly (ppb)
Mean
3.77
1.74
1.88
1.11
1.32
1.37
1.34
1.12
1.25
1.21
1.24
0.39
0.55
1.38
1.78
1.5
1.4
1.6
1.31
1.5
1.34
1.48
1.53
1.65
1.5
1.66
1.31
1.23
1.5
1.4
1.29
1.28
1.64
2.38
std
4.32
2.31
1.83
0.43
0.75
0.84
0.93
0.43
0.82
0.6
0.68
0.42
0.74
1.14
2.07
1.56
1.44
1.48
1.09
1.28
1.13
1.53
1.57
1.5
1.23
1.57
0.84
0.77
1.29
1.19
0.82
0.85
1.34
4.92
GM
2.49
1.32
1.5
1.07
1.2
1.23
1.2
1.08
1.15
1.13
1.15
0.28
0.33
1.2
1.39
1.26
1.2
1.34
1.16
1.29
1.17
1.23
1.26
1.37
1.28
1.36
1.18
1.13
1.28
1.22
1.17
1.16
1.38
1.59
GSD
2.36
1.79
1.8
1.27
1.46
1.5
1.49
1.27
1.41
1.37
1.41
2.19
2.6
1.54
1.79
1.62
1.55
1.66
1.48
1.58
1.51
1.62
1.65
1.69
1.61
1.7
1.47
1.4
1.6
1.55
1.46
1.45
1.67
2.04
Number of 5-minute Daily Maximum
> 100 ppb
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
4
> 200 ppb
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
> 300 ppb
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
> 400 ppb
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
A-118
-------
State
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
County
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
McKenzie
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Mercer
Morton
Morton
Morton
Monitor ID
380530104
380530104
380530104
380530104
380530104
380530104
380530104
380530104
380530104
380530111
380530111
380530111
380530111
380530111
380530111
380530111
380530111
380530111
380530111
380570001
380570001
380570001
380570004
380570004
380570004
380570004
380570004
380570004
380570004
380570004
380570004
380590002
380590002
380590002
Year
1999
2000
2001
2002
2003
2004
2005
2006
2007
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
Days
(n)
240
294
283
236
293
271
245
234
71
258
294
329
336
297
288
308
296
304
78
243
319
14
334
362
338
336
351
344
273
301
107
346
290
359
Hours
(n)
1500
2755
2281
1526
2333
2231
1900
1827
764
2063
2379
2805
3183
2255
2243
2857
2790
2896
722
2824
4735
320
5584
7348
4647
3701
5555
4678
3037
2755
1133
6547
4696
6837
Annual Hourly (ppb)
Mean
2.3
1.96
1.68
1.9
1.98
1.34
1.32
1.32
1.44
3.11
2.36
2.68
1.81
1.87
2.03
1.82
1.39
1.35
1.61
2.93
3.33
5.18
2.6
2.29
2.9
2.65
2.21
2.62
2.43
2.77
2.48
9.31
9.3
7.7
std
3.7
4.07
1.75
4.04
5.29
1.34
2.32
1.78
1.13
7.34
5.4
8.27
2.09
3.52
3.84
5.94
3.28
2.43
1.89
4.29
6.47
3.12
3.94
3.8
5.34
4.59
3.11
3.57
3.25
3.37
3.44
20.26
22.47
16.99
GM
1.66
1.44
1.38
1.34
1.3
1.19
1.14
1.14
1.26
1.8
1.56
1.65
1.4
1.38
1.44
1.27
1.14
1.16
1.3
1.87
2.09
4.43
1.66
1.55
1.76
1.73
1.55
1.73
1.68
1.86
1.7
2.93
2.78
2.53
GSD
1.97
1.85
1.72
1.83
1.84
1.49
1.46
1.46
1.56
2.29
2.02
2.1
1.81
1.8
1.87
1.72
1.5
1.48
1.69
2.26
2.28
1.73
2.2
2.06
2.26
2.17
2.01
2.19
2.08
2.21
2.1
3.67
3.75
3.55
Number of 5-minute Daily Maximum
> 100 ppb
3
5
1
9
15
1
2
4
0
7
7
7
0
8
7
3
5
4
1
0
5
0
3
3
8
2
1
1
0
0
0
102
75
90
> 200 ppb
3
2
0
2
3
0
0
1
0
2
2
5
0
3
2
1
2
1
1
0
2
0
1
1
0
1
0
0
0
0
0
19
8
4
> 300 ppb
1
1
0
0
1
0
0
0
0
0
1
4
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
> 400 ppb
0
1
0
0
0
0
0
0
0
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A-119
-------
State
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
County
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Morton
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Oliver
Steele
Steele
Steele
Steele
Williams
Williams
Williams
Williams
Williams
Monitor ID
380590002
380590002
380590002
380590002
380590002
380590002
380590003
380590003
380590003
380590003
380590003
380590003
380590003
380590003
380650002
380650002
380650002
380650002
380650002
380650002
380650002
380650002
380650002
380650002
380650002
380910001
380910001
380910001
380910001
381050103
381050103
381050103
381050103
381050103
Year
2000
2001
2002
2003
2004
2005
1998
1999
2000
2001
2002
2003
2004
2005
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
1997
1998
1999
2000
2002
2003
2004
2005
2006
Days
(n)
363
346
355
365
363
111
95
353
351
357
342
364
344
106
244
319
349
351
214
350
357
354
275
325
101
216
202
152
83
319
339
348
301
322
Hours
(n)
7964
5947
6258
8033
7532
1450
1924
6522
5984
6345
5245
7991
6338
1012
2356
4175
4856
4765
2404
4482
6953
6138
2443
3369
780
3134
2804
1845
805
2724
3323
3438
2331
2976
Annual Hourly (ppb)
Mean
6.47
7.48
6.26
6.25
6.74
4.85
3.71
5.06
4.71
4.94
4.41
3.55
4.44
3.84
4.28
3.92
3.47
3.14
3.42
2.71
2.37
2.76
3.86
2.85
4.12
1.41
2.22
1.25
1.11
3.18
2.48
2.52
3.51
1.88
std
14.58
13.57
12.03
13.66
13.2
6.08
7.47
8.84
8.04
8.17
7.53
6.34
7.03
5.1
7.23
7.23
6.94
5.54
5.86
4.75
5.58
5.16
6.7
4.32
6.99
0.74
2.1
0.79
0.4
7.56
3.71
5.21
8
2.32
GM
2.22
2.81
2.49
2.33
2.62
2.7
2.01
2.48
2.44
2.54
2.35
1.96
2.5
2.42
2.3
2.1
1.93
1.89
1.96
1.69
1.47
1.65
2.05
1.77
2.35
1.28
1.72
1.14
1.07
1.68
1.64
1.62
1.85
1.4
GSD
3.31
3.5
3.25
3.18
3.29
2.82
2.48
2.88
2.74
2.81
2.68
2.49
2.59
2.39
2.63
2.58
2.42
2.32
2.42
2.21
2.05
2.24
2.62
2.28
2.53
1.5
1.91
1.42
1.26
2.36
2.13
2.12
2.45
1.87
Number of 5-minute Daily Maximum
> 100 ppb
73
66
59
82
76
1
8
41
24
27
26
27
24
1
7
12
15
8
1
4
10
7
6
1
2
0
0
0
0
8
3
5
20
0
> 200 ppb
3
2
1
3
2
0
0
2
0
1
1
0
0
0
0
1
1
0
0
0
1
1
2
0
0
0
0
0
0
3
0
3
3
0
> 300 ppb
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
1
1
0
> 400 ppb
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
A-120
-------
State
ND
ND
ND
ND
ND
ND
ND
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Williams
Williams
Williams
Williams
Williams
Williams
Williams
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Monitor ID
381050103
381050105
381050105
381050105
381050105
381050105
381050105
420030002
420030002
420030002
420030021
420030021
420030021
420030021
420030031
420030031
420030031
420030032
420030032
420030032
420030064
420030064
420030064
420030064
420030067
420030067
420030067
420030116
420030116
420030116
420030116
420031301
420031301
420031301
Year
2007
2002
2003
2004
2005
2006
2007
1997
1998
1999
1997
1998
1999
2002
1997
1998
1999
1997
1998
1999
1997
1998
1999
2002
1997
1998
1999
1997
1998
1999
2002
1997
1998
1999
Days
(n)
86
302
342
346
349
262
24
357
3
325
355
3
362
313
362
3
360
364
3
210
361
3
355
350
364
3
257
361
3
299
232
363
3
363
Hours
(n)
834
2843
3523
4129
4492
2938
263
7821
72
6986
7830
72
8279
7291
8000
68
7443
7951
60
4326
7526
71
7232
8239
8231
72
5891
7767
70
5684
5403
7663
70
8161
Annual Hourly (ppb)
Mean
3.35
6.77
5.67
5.64
6.79
3.74
3.59
12.57
43.18
11.04
18.11
10.22
9
7.32
10.98
11.38
8.98
15.4
35.2
8.18
11.9
20.11
12.11
10.9
10.43
17.01
10.05
13.26
17
12.12
7
9.37
12.66
9.64
std
4.62
10.88
9.39
10.64
13
6.66
5.63
15.05
32.27
11.16
18.87
8.23
7.94
7.33
9.63
9.36
7.84
19.34
20.65
7.8
13.08
7.99
14.34
13.26
11.13
12.54
8.81
17.76
11.04
16.01
7.96
9.8
6.88
9.62
GM
2.07
2.93
2.55
2.55
2.49
1.91
1.99
7.69
31.63
7.36
11.07
7.48
6.64
4.49
8.05
8.2
6.43
9.39
27.51
5.66
7.16
18.41
7.35
5.91
6.69
12.63
7.35
8.33
12.59
7.82
4.56
6.25
11.29
6.57
GSD
2.4
3.34
3.12
3.1
3.46
2.62
2.53
2.68
2.43
2.53
2.93
2.27
2.2
2.85
2.24
2.3
2.33
2.73
2.26
2.41
2.86
1.56
2.78
3.15
2.62
2.25
2.22
2.6
2.46
2.54
2.5
2.48
1.58
2.44
Number of 5-minute Daily Maximum
> 100 ppb
0
35
13
19
52
14
1
70
3
31
87
0
3
3
12
0
1
84
2
2
17
0
18
18
12
0
1
60
0
50
3
21
0
21
> 200 ppb
0
4
1
2
12
1
0
8
1
2
19
0
0
0
1
0
0
15
0
0
2
0
3
5
2
0
0
19
0
26
0
4
0
3
> 300 ppb
0
1
0
2
1
0
0
2
0
0
5
0
0
0
0
0
0
6
0
0
0
0
2
1
1
0
0
12
0
13
0
1
0
1
> 400 ppb
0
0
0
1
0
0
0
0
0
0
2
0
0
0
0
0
0
4
0
0
0
0
1
0
1
0
0
8
0
8
0
1
0
1
A-121
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Allegheny
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Beaver
Berks
Berks
Berks
Cambria
Cambria
Cambria
Erie
Erie
Erie
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Monitor ID
420033003
420033003
420033003
420033003
420033004
420033004
420033004
420070002
420070002
420070005
420070005
420070005
420070005
420070005
420070005
420070005
420070005
420110009
420110009
420110009
420210011
420210011
420210011
420490003
420490003
420490003
421010022
421010022
421010022
421010022
421010022
421010048
421010048
421010048
Year
1997
1998
1999
2002
1997
1998
1999
1997
1998
1997
1998
2002
2003
2004
2005
2006
2007
1997
1998
1999
1997
1998
1999
1997
1998
1999
1997
1998
1999
2000
2001
1997
1998
1999
Days
(n)
356
2
350
316
362
3
361
351
270
359
277
361
365
364
362
361
324
350
365
119
361
356
120
363
363
120
364
363
137
179
98
365
356
178
Hours
(n)
7422
45
6998
7363
7461
66
7408
7889
6205
7447
6388
8491
8706
8656
8578
8457
7556
7805
8641
2790
8129
7908
2835
8169
8416
2778
8297
8065
2665
3630
2094
8456
7285
3939
Annual Hourly (ppb)
Mean
11.8
11.47
13.59
12.66
9.18
13.12
8.55
11.83
12.96
16.57
16.14
14.24
10.79
11.59
12.57
9.26
9.79
8.66
8.93
9.22
9.76
8.78
9.74
9.76
10.57
11.48
8.56
7.3
7.79
7.63
7.53
8.88
6.27
6.08
std
13.86
6.31
19.91
18.25
9.66
6.01
9.09
15.38
16.48
25.11
26.85
26.51
17.07
17.68
18.18
18.5
13.98
8.87
7.56
8.38
9.15
9.69
7.99
11.22
13.5
15.12
8.74
7.04
8.26
6.88
7.17
18.38
6.03
6.57
GM
7.01
9.35
7.86
6.32
6.17
11.71
5.79
6.83
7.8
8.65
8.36
5.28
4.38
5.55
6.82
3.49
4.94
5.87
7.11
6.86
6.72
5.65
7.61
6.68
7.09
7.46
5.63
4.82
4.76
5.05
5.16
4.96
4.21
3.95
GSD
2.85
2.09
2.86
3.29
2.47
1.65
2.47
2.84
2.71
3.14
3.01
4.12
3.83
3.39
3.04
3.78
3.26
2.44
1.92
2.17
2.47
2.62
1.99
2.33
2.35
2.48
2.57
2.56
2.79
2.62
2.44
2.74
2.48
2.53
Number of 5-minute Daily Maximum
> 100 ppb
27
0
37
53
12
0
6
91
74
98
92
113
75
74
75
71
45
35
33
9
8
16
1
60
60
26
7
2
4
1
0
59
0
1
> 200 ppb
1
0
2
8
2
0
3
11
6
39
39
49
16
22
26
30
12
4
3
1
0
1
0
9
12
7
1
0
1
0
0
40
0
1
> 300 ppb
0
0
2
5
0
0
2
1
2
17
21
23
3
10
12
11
4
0
0
0
0
0
0
1
1
3
0
0
0
0
0
27
0
0
> 400 ppb
0
0
2
3
0
0
0
1
0
11
13
13
2
3
7
5
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
23
0
0
A-122
-------
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
County
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Philadelphia
Warren
Warren
Warren
Warren
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Washington
Barnwell
Barnwell
Barnwell
Charleston
Charleston
Charleston
Charleston
Charleston
Charleston
Georgetown
Georgetown
Georgetown
Greenville
Greenville
Greenville
Monitor ID
421010136
421010136
421010136
421010136
421010136
421010136
421010136
421230003
421230003
421230004
421230004
421250005
421250005
421250005
421250200
421250200
421250200
421255001
421255001
450110001
450110001
450110001
450190003
450190003
450190003
450190046
450190046
450190046
450430006
450430006
450430006
450450008
450450008
450450008
Year
1997
1998
1999
2000
2001
2002
2003
1997
1998
1997
1998
1997
1998
1999
1997
1998
1999
1997
1998
2000
2001
2002
2000
2001
2002
2000
2001
2002
2000
2001
2002
2000
2001
2002
Days
(n)
360
339
337
351
266
359
119
346
89
355
89
364
362
120
364
365
120
365
277
100
267
202
114
344
201
100
269
189
71
241
140
113
356
212
Hours
(n)
7532
6491
7144
7044
5149
7271
2585
7157
2126
7022
1966
8374
8540
2821
8369
8656
2829
8425
6559
789
2625
2544
1703
4806
3509
1252
3497
2927
604
2218
1169
1987
6418
4679
Annual Hourly (ppb)
Mean
4.99
5.25
5.63
5.76
6.77
5.38
6.74
10.53
7.62
17.14
13.97
8.95
8.88
8.32
10.52
10.46
10.15
12.71
13.46
3.95
2.72
2.11
6.24
4.16
2.85
4.61
2.64
2.34
4.92
4.76
2.5
4.84
4.24
3.06
std
5.52
5.52
6.04
5.97
7.43
5.7
6.71
11.59
7.38
28.18
21.76
8.41
7.78
7.68
11.23
10.49
9.81
15.24
13.09
2.83
2.61
1.72
5.36
4.12
3.49
3.9
2.6
2.89
4.35
6.11
4.33
3.75
3.86
2.8
GM
3.29
3.5
3.71
3.74
4.38
3.57
4.62
6.64
5.41
7.47
6.8
6.45
6.68
6.36
6.99
7.18
7
8.39
10.28
3.39
2.13
1.67
4.77
2.95
1.97
3.71
1.99
1.68
3.97
3.13
1.67
3.95
3.18
2.29
GSD
2.43
2.44
2.48
2.54
2.55
2.47
2.42
2.68
2.26
3.66
3.18
2.25
2.14
2.02
2.45
2.37
2.4
2.36
1.97
1.66
1.93
1.88
2.02
2.22
2.16
1.84
2
2.02
1.82
2.33
2.08
1.82
2.1
2.09
Number of 5-minute Daily Maximum
> 100 ppb
1
2
2
0
2
3
2
26
0
148
30
7
4
1
17
15
3
57
42
0
1
0
1
1
0
0
0
0
0
3
1
0
3
1
> 200 ppb
0
0
1
0
0
0
0
3
0
44
6
0
0
0
0
1
1
5
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
> 300 ppb
0
0
0
0
0
0
0
0
0
14
2
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
> 400 ppb
0
0
0
0
0
0
0
0
0
8
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A-123
-------
State
SC
sc
SC
sc
sc
sc
sc
sc
sc
sc
sc
sc
sc
UT
UT
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
wv
County
Lexington
Lexington
Oconee
Oconee
Oconee
Richland
Richland
Richland
Richland
Richland
Richland
Richland
Richland
Salt Lake
Salt Lake
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wayne
Wood
Wood
Wood
Wood
Wood
Monitor ID
450630008
450630008
450730001
450730001
450730001
450790007
450790007
450790007
450790021
450790021
450790021
450791003
450791003
490352004
490352004
540990002
540990003
540990003
540990003
540990003
540990004
540990004
540990004
540990004
540990005
540990005
540990005
540990005
541071002
541071002
541071002
541071002
541071002
Year
2001
2002
2000
2001
2002
2000
2001
2002
2000
2001
2002
2001
2002
1997
1998
2002
2002
2003
2004
2005
2002
2003
2004
2005
2002
2003
2004
2005
2001
2002
2003
2004
2005
Days
(n)
263
211
89
288
188
110
365
210
109
283
202
193
212
335
354
365
361
362
366
365
362
365
366
363
365
365
366
365
92
365
365
366
266
Hours
(n)
3941
4242
1218
4304
3063
1808
6419
4335
911
2700
2505
3346
4323
4524
5792
8711
7417
8057
8659
8141
8560
8570
8673
8586
8283
7927
8681
8453
2152
8648
8641
8581
6219
Annual Hourly (ppb)
Mean
4.2
4.5
3.85
2.9
1.82
4.48
3.88
2.95
4.43
3.73
2.94
3.14
2.87
2.31
1.94
7.49
8.48
8.76
9.21
9.58
9.21
8.53
7.22
7.67
8.44
8.31
7.03
6.68
7.76
9.9
9.48
10.88
8.34
std
7.8
8.74
2.87
2.1
1.52
2.81
3.47
2.71
5.47
4.89
4.85
2.8
2.8
2.5
1.66
7.14
9.1
9.73
9.46
11.8
9.18
9.77
6.66
6.39
9.75
11.03
5.92
5.52
12.51
11.29
12.26
13.25
12.71
GM
2.37
2.33
3.26
2.35
1.43
3.86
2.99
2.23
3.4
2.64
1.92
2.46
2.16
1.76
1.58
5.13
5.21
5.56
6.38
5.96
6.37
5.84
5.36
5.97
5.38
5.02
5.25
4.89
4.04
6.21
5.8
7
4.07
GSD
2.44
2.61
1.7
1.89
1.95
1.69
2.02
2.04
1.85
2.1
2.16
1.96
2.04
1.94
1.78
2.42
2.75
2.58
2.31
2.61
2.37
2.35
2.12
2
2.58
2.7
2.16
2.26
3.04
2.63
2.61
2.55
3.23
Number of 5-minute Daily Maximum
> 100 ppb
26
22
0
0
0
0
0
0
0
0
0
0
0
6
0
1
7
8
5
6
22
26
6
7
67
52
2
4
9
42
53
57
42
> 200 ppb
3
3
0
0
0
0
0
0
0
0
0
0
0
1
0
0
2
0
1
0
1
4
0
0
3
20
0
1
3
7
9
13
12
> 300 ppb
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
3
0
0
0
5
0
0
2
1
2
3
1
> 400 ppb
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
1
0
0
0
0
0
0
1
0
0
1
1
A-124
-------
Appendix B: Supplement to the SO2 Exposure Assessment
B-l
-------
B.1 OVERVIEW
This appendix contains supplemental descriptions of the methods and data used in the
SC>2 exposure assessment, as well as detailed results from the exposure analyses performed.
First, a broad description of the exposure modeling approach is described (section B.2),
applicable to the two exposure modeling domains conducted: Greene County, Mo. and St. Louis,
MO. Supplementary input data used in AERMOD are provided in section B.3, as well as the
model predictions and ambient monitor measurements in each modeling domain. Section B.4
has additional input and output data for APEX.
A series of Attachments also follow, further documenting some of the data sources and
modeling approaches used, as well as previously conducted uncertainty analyses on selected
input parameters in APEX:
Attachment 1. Technical Memorandum on Meteorological Data Preparation for AERMOD for
SO2 REA for Greene County And St. Louis Modeling Domains, Year 2002.
Attachment 2. Technical Memorandum on the Analysis of NHIS Asthma Prevalence Data.
Attachment 3. Technical Memorandum on Estimating Physiological Parameters for the Exposure
Model
Attachment 4. Technical Memorandum on Longitudinal Diary Construction Approach
Attachment 5. Technical Memorandum on the Evaluation Cluster-Markov Algorithm
Attachment 6. Technical Memorandum on Analysis of Air Exchange Rate Data
Attachment 7. Technical Memorandum on the Uncertainty Analysis of Residential Air Exchange
Rate Distributions
Attachment 8. Technical Memorandum on the Distributions of Air Exchange Rate Averages
Over Multiple Days
B-2
-------
B.2 HUMAN EXPOSURE MODELING USING APEX
The Air Pollutants Exposure model (APEX) is a personal computer (PC)-based program
designed to estimate human exposure to criteria and air toxic pollutants at the local, urban, and
consolidated metropolitan levels. APEX, also known as TREVI.Expo, is the human inhalation
exposure module of EPA's Total Risk Integrated Methodology (TRIM) model framework (US
EPA, 1999), a modeling system with multimedia capabilities for assessing human health and
ecological risks from hazardous and criteria air pollutants. It is developed to support evaluations
with a scientifically sound, flexible, and user-friendly methodology. Additional information on
the TRIM modeling system, as well as downloads of the APEX Model, user's guide, and other
supporting documentation, are on EPA's Technology Transfer Network (TTN) at
http://www.epa.gov/ttn/fera.
B.2.1 History
APEX was derived from the National Ambient Air Quality Standards (NAAQS)
Exposure Model (NEM) series of models, developed to estimate exposure to the criteria
pollutants (e.g., carbon monoxide (CO), ozone Os). In 1979, EPA began by assembling a
database of human activity patterns that could be used to estimate exposures to indoor and
outdoor pollutants (Roddin et al., 1979). These data were then combined with measured outdoor
concentrations in NEM to estimate exposures to CO (Biller et al., 1981; Johnson and Paul,
1983). In 1988, OAQPS began to incorporate probabilistic elements into the NEM methodology
and use activity pattern data based on various human activity diary studies to create an early
version of probabilistic NEM for Os (i.e., pNEM/Os). In 1991, a probabilistic version of NEM
was extended to CO (pNEM/CO) that included a one-compartment mass-balance model to
estimate CO concentrations in indoor microenvironments. The application of this model to
Denver, Colorado has been documented in Johnson et al. (1992). Additional enhancements to
pNEM/Os in the early- to mid-1990's allowed for probabilistic exposure assessments in nine
urban areas for the general population, outdoor children, and outdoor workers (Johnson et al.,
1996a; 1996b; 1996c). Between 1999 and 2001, updated versions of pNEM/CO (versions 2.0
and 2.1) were developed that relied on activity diary data from EPA's Consolidated Human
Activities Database (CHAD) and enhanced algorithms for simulating gas stove usage, estimating
alveolar ventilation rate (a measure of human respiration), and modeling home-to-work
commuting patterns.
B-3
-------
The first version of APEX was essentially identical to pNEM/CO (version 2.0) except
that it was capable of running on a PC instead of a mainframe. The next version, APEX2, was
substantially different, particularly in the use of a personal profile approach (i.e., simulation of
individuals) rather than a cohort simulation (i.e., groups of similar persons). APEX3 introduced
a number of new features including automatic site selection from national databases, a series of
new output tables providing summary exposure and dose statistics, and a thoroughly reorganized
method of describing microenvironments and their parameters. Most of the spatial and temporal
constraints of pNEM and APEX1 were removed or relaxed by version 3.
The version of APEX used in this exposure assessment is APEX4.3, described in the
APEX User's Guide and the APEX Technical Support Document (US EPA, 2009a; 2009b) and
referred to here as the APEX User's Guide and TSD. This latest version has the added flexibility
of addressing user defined exposure timesteps within an hour.
B.2.2 APEX Model Overview A microenvironment is a three-
APEX estimates human exposure to criteria and toxic dimensional space in which human
contact with an environmental
air pollutants at the local, urban, or consolidated metropolitan pollutant takes place and which can
area levels using a stochastic, microenvironmental approach. "Q treated as a well-Characterized,
relatively homogeneous location
The model randomly selects data for a sample of hypothetical wjfh respect to pollutant
individuals from an actual population database and simulates concentrations for a specified time
period.
each hypothetical individual's movements through time and
space (e.g., at home, in vehicles) to estimate their exposure to a pollutant. APEX simulates
commuting, and thus exposures that occur at home and work locations, for individuals who work
in different areas than they live.
APEX can be conceptualized as a simulated field study that would involve selecting an
actual sample of specific individuals who live in (or work and live in) a geographic area and then
continuously monitoring their activities and subsequent inhalation exposure to a specific air
pollutant during a specific period of time.
The main differences between APEX and an actual field study are that in APEX:
• The sample of individuals is a virtual sample, not actual persons. However, the
population of individuals appropriately balanced according to various demographic
B-4
-------
variables and census data using their relative frequencies, in order to obtain a
representative sample (to the extent possible) of the actual people in the study area
• The activity patterns of the sampled individuals (e.g., the specification of indoor and
other microenvironments visited and the time spent in each) are assumed by the
model to be comparable to individuals with similar demographic characteristics,
according to activity data such as diaries compiled in EPA's Consolidated Human
Activity Database (or CHAD; US EPA, 2002; McCurdy et al., 2000)
• The pollutant exposure concentrations are estimated by the model using a set of user-
input ambient outdoor concentrations (either modeled or measured) and information
on the behavior of the pollutant in various microenvironments;
• Variation in ambient air quality levels can be simulated by either adjusting air quality
concentrations to just meet alternative ambient standards, or by reducing source
emissions and obtaining resulting air quality modeling outputs that reflect these
potential emission reductions, and
• The model accounts for the most significant factors contributing to inhalation
exposure - the temporal and spatial distribution of people and pollutant
concentrations throughout the study area and among microenvironments - while also
allowing the flexibility to adjust some of these factors for alternative scenarios and
sensitivity analyses.
APEX is designed to simulate human population exposure to criteria and air toxic
pollutants at local, urban, and regional scales. The user specifies the geographic area to be
modeled and the number of individuals to be simulated to represent this population. APEX then
generates a personal profile for each simulated person that specifies various parameter values
required by the model. The model next uses diary-derived time/activity data matched to each
personal profile to generate an exposure event sequence (also referred to as activity pattern or
diary) for the modeled individual that spans a specified time period, such as one year. Each
event in the sequence specifies a start time, exposure duration, geographic location,
microenvironment, and activity performed. Probabilistic algorithms are used to estimate the
pollutant concentration associated with each exposure event. The estimated pollutant
concentrations account for the effects of ambient (outdoor) pollutant concentration, penetration
factors, air exchange rates, decay/deposition rates, and proximity to emission sources, depending
B-5
-------
on the microenvironment, available data, and estimation method selected by the user. Because
the modeled individuals represent a random sample of the population of interest, the distribution
of modeled individual exposures can be extrapolated to the larger population. Additional
discussion regarding the five basic exposure modeling steps noted in the SC>2 REA are described
in sections that follow.
B.2.2.1 Study Area Characterization
The APEX study area has traditionally been on the scale of a city or slightly larger
metropolitan area, although it is now possible to model larger areas such as combined statistical
areas (CSAs). In the exposure analyses performed as part of this NAAQS review, the study area
is defined by either a single or a few counties. The demographic data used by the model to
create personal profiles is provided at the census block level. For each block the model requires
demographic information representing the distribution of age, gender, race, and work status
within the study population. Each block has a location specified by latitude and longitude for
some representative point (e.g., geographic center). The current release of APEX includes input
files that already contain this demographic and location data for all census tracts, block groups,
and blocks in the 50 United States, based on the 2000 Census. In this assessment, exposures
were evaluated at the block level.
Air Quality Data
Air quality data can be input to the model as measured data from an ambient monitor or
that generated by air quality modeling. This exposure analysis used modeled air quality data,
whereas the principal emission sources included both mobile and stationary sources as well as
fugitive emissions. Air quality data used for input to APEX were generated using AERMOD, a
steady-state, Gaussian plume model (US EPA, 2004). The following steps were performed using
AERMOD.
In APEX, the ambient air quality data are assigned to geographic areas called districts.
The districts are used to assign pollutant concentrations to the blocks/tracts and
microenvironments being modeled. The ambient air quality data are provided by the user as
hourly time series for each district. As with blocks/tracts, each district has a representative
location (latitude and longitude). APEX calculates the distance from each block/tract to each
district center, and assigns the block/tract to the nearest district, provided the block/tract
representative location point (e.g., geographic center) is in the district. Each block/tract can be
B-6
-------
assigned to only one district. In this assessment the district was synonymous with the receptor
modeled in the dispersion modeling.
Meteorological Data
Ambient temperatures are input to APEX for different sites (locations). As with districts,
APEX calculates the distance from each block to each temperature site and assigns each block to
the nearest site. Hourly temperature data are from the National Climatic Data Center Surface
Airways Hourly TD-3280 dataset (NCDC Surface Weather Observations). Daily average and 1-
hour maxima are computed from these hourly data.
There are two files that are used to provide meteorological data to APEX. One file, the
meteorological station location file, contains the locations of meteorological data recordings
expressed in latitude and longitude coordinates. This file also contains start and end dates for the
data recording periods. The temperature data file contains the data from the locations in the
temperature zone location file. This file contains hourly temperature readings for the period
being modeled for the meteorological stations in and around the study area.
B.2.2.2 Generate Simulated Individuals
APEX stochastically generates a user-specified number of simulated persons to represent
the population in the study area. Each simulated person is represented by a personal profile, a
summary of personal attributes that define the individual. APEX generates the simulated person
or profile by probabilistically selecting values for a set of profile variables (Table B.2-1). The
profile variables could include:
• Demographic variables, generated based on the census data;
• Physical variables, generated based on sets of distribution data;
• Other daily varying variables, generated based on literature-derived distribution data
that change daily during the simulation period.
APEX first selects demographic and physical attributes for each specified individual, and
then follows the individual over time and calculates his or her time series of exposure.
B-7
-------
Table B.2-1. Examples of profile variables in APEX.
Variable
Type
Demographic
Physical
Profile Variables
Age
Gender
Home block
Work tract
Employment status
Air conditioner
Gas Stove
Description
Age (years)
Male or Female
Block in which a simulated person lives
Tract in which a simulated person works
Indicates employment outside home
Indicates presence of air conditioning at home
Indicates presence of gas stove at home
Population Demographics
APEX takes population characteristics into account to develop accurate representations of
study area demographics. Specifically, population counts by area and employment probability
estimates are used to develop representative profiles of hypothetical individuals for the
simulation.
APEX is flexible in the resolution of population data provided. As long as the data are
available, any resolution can be used (e.g., county, census tract, census block). For this
application of the model, census block level data were used. Block-level population counts come
from the 2000 Census of Population and Housing Summary File 1 (SF-1). This file contains the
100-percent data, which is the information compiled from the questions asked of all people and
about every housing unit.
As part of the population demographics inputs, it is important to integrate working
patterns into the assessment. In the 2000 U.S. Census, estimates of employment were developed
by census information (US Census Bureau, 2007). The employment statistics are broken down
by gender and age group, so that each gender/age group combination is given an employment
probability fraction (ranging from 0 to 1) within each census tract. The age groupings used are:
16-19, 20-21, 22-24, 25-29, 30-34, 35-44, 45-54, 55-59, 60-61, 62-64, 65-69, 70-74, and >75.
Children under 16 years of age were assumed to be not employed.
Since this analysis was conducted at the census block level, block level employment
probabilities were required. It was assumed that the employment probabilities for a census tract
apply uniformly to the constituent census blocks.
B-8
-------
Commuting
In addition to using estimates of employment by tract, APEX also incorporates home-to-
work commuting data. Commuting data were originally derived from the 2000 Census and were
collected as part of the Census Transportation Planning Package (CTPP) (US DOT, 2007). The
data used contain counts of individuals commuting from home to work locations at a number of
geographic scales. These data were processed to calculate fractions for each tract-to-tract flow to
create the national commuting data distributed with APEX. This database contains commuting
data for each of the 50 states and Washington, D.C.
Commuting within the Home Tract
The APEX data set does not differentiate people that work at home from those that
commute within their home tract.
Commuting Distance Cutoff
A preliminary data analysis of the home-work counts showed that a graph of log(flows)
versus log(distance) had a near-constant slope out to a distance of around 120 kilometers.
Beyond that distance, the relationship also had a fairly constant slope but it was flatter, meaning
that flows were not as sensitive to distance. A simple interpretation of this result is that up to
120 km, the majority of the flow was due to persons traveling back and forth daily, and the
numbers of such persons decrease rapidly with increasing distance. Beyond 120 km, the
majority of the flow is comprised of persons who stay at the workplace for extended times, in
which case the separation distance is not as crucial in determining the flow.
To apply the home-work data to commuting patterns in APEX, a simple rule was chosen.
It was assumed that all persons in home-work flows up to 120 km are daily commuters, and no
persons in more widely separated flows commute daily. This meant that the list of destinations
for each home tract was restricted to only those work tracts that are within 120 km of the home
tract. When the same cutoff was performed on the 1990 census data, it resulted in 4.75% of the
home-work pairs in the nationwide database being eliminated, representing 1.3% of the workers.
The assumption is that this 1.3% of workers do not commute from home to work on a daily
basis. It is expected that the cutoff reduced the 2000 data by similar amounts.
Eliminated Records
A number of tract-to-tract pairs were eliminated from the database for various reasons. A
fair number of tract-to-tract pairs represented workers who either worked outside of the U.S.
B-9
-------
(9,631 tract pairs with 107,595 workers) or worked in an unknown location (120,830 tract pairs
with 8,940,163 workers). An additional 515 workers in the commuting database whose data
were missing from the original files, possibly due to privacy concerns or errors, were also
deleted.
Commuting outside the study area
APEX allows for some flexibility in the treatment of persons in the modeled population
who commute to destinations outside the study area. By specifying "KeepLeavers = No" in the
simulation control parameters file, people who work inside the study area but live outside of it
are not modeled, nor are people who live in the study area but work outside of it. By specifying
"KeepLeavers = Yes," these commuters are modeled. This triggers the use of two additional
parameters, called LeaverMult and LeaverAdd. While a commuter is at work, if the workplace is
outside the study area, then the ambient concentration is assumed to be related to the average
concentration over all air districts at the same point in time, and is calculated as:
Ambient Concentrat ion = LeaverMult xavg(t) + LeaverAdd equation (B-l)
where:
Ambient Concentration = Calculated ambient air concentrations for locations outside
of the study area (ppm or ppm)
LeaverMult = Multiplicative factor for city-wide average concentration,
applied when working outside study area
avg(t) = Average ambient air concentration over all air districts in
study area, for time t (ppm or ppm)
LeaverAdd = Additive term applied when working outside study area
All microenvironmental concentrations for locations outside of the study area are
determined from this ambient concentration by the same function as applies inside the study
area.
Block-level commuting
For census block simulations, APEX requires block-level commuting file. A special
software preprocesser was created to generate these files for APEX on the basis of the tract-level
B-10
-------
commuting data and finely-resolved land use data. The software calculates commuting flows
between census blocks for the employed population according equation (B-2).
Flowblock =Flow tractxFpopxFland equat.on (B_2)
where:
Flow biock = flow of working population between a home block and a work block.
Flow tmct = flow of working population between a home tract and a work tract.
Fpop = fraction of home tract's working population residing in the home block.
F land = fraction of work tract's commercial/industrial land area in the work
block
Thus, it is assumed that the frequency of commuting to a workplace block within a tract
is proportional to the amount of commercial and industrial land in the block.
Profile Functions
A Profile Functions file contains settings used to generate results for variables related to
simulated individuals. While certain settings for individuals are generated automatically by
APEX based on other input files, including demographic characteristics, others can be specified
using this file. For example, the file may contain settings for determining whether the profiled
individual's residence has an air conditioner, a gas stove, etc. As an example, the Profile
Functions file contains fractions indicating the prevalence of air conditioning in the cities
modeled in this assessment (Figure B.2-1). APEX uses these fractions to stochastically generate
air conditioning status for each individual. The derivation of particular data used in specific
microenvironments is provided below.
AC_Home
! Has air conditioning at home
TABLE
INPUT 1 PROBABILITY 2 "A/C probabilities"
0.850.15
RESULT INTEGER 2 "Yes/No"
12
#
Figure B.2-1. Example of a profile function file for A/C prevalence.
B-ll
-------
B.2.2.3 Longitudinal Activity Pattern Sequences
Exposure models use human activity pattern data to predict and estimate exposure to
pollutants. Different human activities, such as spending time outdoors, indoors, or driving, will
have varying pollutant exposure concentrations. To accurately model individuals and their
exposure to pollutants, it is critical to understand their daily activities.
The Consolidated Human Activity Database (CHAD) provides data for where people
spend time and the activities performed. CHAD was designed to provide a basis for conducting
multi-route, multi-media exposure assessments (McCurdy et al., 2000). The data contained
within CHAD come from multiple activity pattern surveys with varied structures (Table B.2-2),
however the surveys have commonality in containing daily diaries of human activities and
personal attributes (e.g., age and gender).
There are four CHAD-related input files used in APEX. Two of these files can be
downloaded directly from the CHADNet (http://www.epa.gov/chadnetl), and adjusted to fit into
the APEX framework. These are the human activity diaries file and the personal data file, and
are discussed below. A third input file contains metabolic information for different activities
listed in the diary file, these are not used in this exposure analysis. The fourth input file maps
five-digit location codes used in the diary file to APEX microenvironments; this file is discussed
in the section describing microenvironmental calculations (Section B.2.2.4.4).
Personal Information file
Personal attribute data are contained in the CHAD questionnaire file that is distributed
with APEX. This file also has information for each day individuals have diaries. The different
variables in this file are:
• The study, person, and diary day identifiers
• Day of week
• Gender
• Employment status
• Age in years
• Maximum temperature in degrees Celsius for this diary day
• Mean temperature in degrees Celsius for this diary day
• Occupation code
B-12
-------
• Time, in minutes, during this diary day for which no data are included in the database
Diary Events file
The human activity diary data are contained in the events file that is distributed with
APEX. This file contains the activities for the nearly 23,000 people with intervals ranging from
one minute to one hour. An individuals' diary varies in length from one to 15 days. This file
contains the following variables:
• The study, person, and diary day identifiers
• Start time of this activity
• Number of minutes for this activity
• Activity code (a record of what the individual was doing)
• Location code (a record of where the individual was)
B-13
-------
Table B.2-2. Summary of activity pattern studies used in CHAD.
Study Name
Baltimore
California
Adolescents
and Adults
(GARB)
California
Children
(GARB)
Cincinnati
(EPRI)
Denver
(EPA)
Los Angeles:
Elementary
School
Children
Los Angeles:
High School
Adolescents
National:
NHAPS-Air
National:
NHAPS-
Water
Washington,
D.C. (EPA)
Location
A single
building in
Baltimore
California
California
Cincinnati
MSA
Denver
MSA
Los
Angeles
Los
Angeles
National
National
Wash. DC
MSA
Study
time
period
01/1997-
02/1997,
07/1 998-
08/1 998
1 0/1 987-
09/1 988
04/1 989-
02/1 990
03/1 985-
04/1985,
08/1 985
11/1982-
02/1 983
10/1989
09/1 990-
10/1990
09/1 992-
10/1994
09/1 992-
10/1994
11/1982-
02/1 983
Ages
72-93
12-17
18-94
0-11
0-86
18-70
10-12
13-17
0-93
0-93
18-98
Persons
26
181
1,552
1,200
888
432
17
19
4,326
4,332
639
Person
-days
292
181
1,552
1,200
2,587
791
51
42
4,326
4,332
639
Diary type
/study
design
Diary
Recall
/Random
Recall
/Random
Diary
/Random
Diary
/Random
Diary
Diary
Recall
/Random
Recall
/Random
Diary
/Random
Reference
Williams et al. (2000)
Robinson et al.
(1989);
Wiley etal. (1991 a)
Wiley etal. (1991b)
Johnson (1989)
Johnson (1984);
Aklandetal. (1985)
Spier etal. (1992)
Spier etal. (1992)
Klepeisetal. (1996);
Tsang and Klepeis
(1996)
Klepeisetal. (1996);
Tsang and Klepeis
(1996)
Hartwell etal. (1984);
Aklandetal. (1985)
Construction of Longitudinal Activity Sequences
Typical time-activity pattern data available for inhalation exposure modeling consist of a
sequence of location/activity combinations spanning a 24-hour duration, with 1 to 3 diary-days
for any single individual. Exposure modeling requires information on activity patterns over
longer periods of time, e.g., a full year. For example, even for pollutant health effects with short
averaging times (e.g., SO2 5-minute average concentration) it may be desirable to know the
frequency of exceedances of a concentration over a long period of time (e.g., the annual number
of exceedances of a 24-hour average SC>2 concentration of 100 ppb for each simulated
individual).
B-14
-------
Long-term multi-day activity patterns can be estimated from single days by combining
the daily records in various ways, and the method used for combining them will influence the
variability of the long-term activity patterns across the simulated population. This in turn will
influence the ability of the model to accurately represent either long-term average high-end
exposures, or the number of individuals exposed multiple times to short-term high-end
concentrations.
A common approach for constructing long-term activity patterns from short-term records
is to re-select a daily activity pattern from the pool of data for each day, with the implicit
assumption that there is no correlation between activities from day to day for the simulated
individual. This approach tends to result in long-term activity patterns that are very similar
across the simulated population. Thus, the resulting exposure estimates are likely to
underestimate the variability across the population, and therefore, underestimate the high-end
exposure concentrations or the frequency of exceedances.
A contrasting approach is to select a single activity pattern (or a single pattern for each
season and/or weekday-weekend) to represent a simulated individual's activities over the
duration of the exposure assessment. This approach has the implicit assumption that an
individual's day-to-day activities are perfectly correlated. This approach tends to result in long-
term activity patterns that are very different across the simulated population, and therefore may
over-estimate the variability across the population.
Cluster-Markov Algorithm
A new algorithm has been developed and incorporated into APEX to represent the day-
to-day correlation of activities for individuals. The algorithms first use cluster analysis to divide
the daily activity pattern records into groups that are similar, and then select a single daily record
from each group. This limited number of daily patterns is then used to construct a long-term
sequence for a simulated individual, based on empirically-derived transition probabilities. This
approach is intermediate between the assumption of no day-to-day correlation (i.e., re-selection
for each time period) and perfect correlation (i.e., selection of a single daily record to represent
all days).
The steps in the algorithm are as follows.
1. For each demographic group (age, gender, employment status), temperature range, and
day-of-week combination, the associated time-activity records are partitioned into 3
B-15
-------
groups using cluster analysis. The clustering criterion is a vector of 5 values: the time
spent in each of 5 microenvironment categories (indoors - residence; indoors - other
building; outdoors - near road; outdoors - away from road; in vehicle).
2. For each simulated individual, a single time-activity record is randomly selected from
each cluster.
3. A Markov process determines the probability of a given time-activity pattern occurring
on a given day based on the time-activity pattern of the previous day and cluster-to-
cluster transition probabilities. The cluster-to-cluster transition probabilities are
estimated from the available multi-day time-activity records. If insufficient multi-day
time-activity records are available for a demographic group, season, day-of-week
combination, then the cluster-to-cluster transition probabilities are estimated from the
frequency of time-activity records in each cluster in the CHAD data base.
Details regarding the Cluster-Markov algorithm and supporting evaluations are provided in
Attachments 4 and 5.
B.2.2.4 Calculating Microenvironmental Concentrations
Probabilistic algorithms estimate the pollutant concentration associated with each
exposure event. The estimated pollutant concentrations account for the effects of ambient
(outdoor) pollutant concentration, penetration factor, air exchange rate, decay/deposition rate,
and proximity to microenvironments can use the transfer factors method while the others use the
mass balance emission sources, depending on the microenvironment, available data, and the
estimation method selected by the user.
APEX calculates air concentrations in the various microenvironments visited by the
simulated person by using the ambient air data for the relevant blocks, the user-specified
estimation method, and input parameters specific to each microenvironment. APEX calculates
hourly concentrations in all the microenvironments at each hour of the simulation for each of the
simulated individuals using one of two methods: by mass balance or a transfer factors method.
Mass Balance Model
The mass balance method simulates an enclosed microenvironment as a well-mixed
volume in which the air concentration is spatially uniform at any specific time. The following
processes are used estimate the concentration of an air pollutant in such a microenvironment:
• Inflow of air into the microenvironment
B-16
-------
• Outflow of air from the microenvironment
• Removal of a pollutant from the microenvironment due to deposition, filtration, and
chemical degradation
• Pollutant emissions inside the microenvironment.
Table B.2-3 lists the parameters required by the mass balance method to calculate
concentrations in a microenvironment. A proximity factor (/proximity) is used to account for
differences in ambient concentrations between the geographic location represented by the
ambient air quality data (e.g., a regional fixed-site monitor or modeled concentration) and the
geographic location of the microenvironment (e.g., near a roadway). This factor could take a
value either greater than or less than 1. Emission source (ES) represents the emission rate for the
emission source and concentration source (CS) is the mean air concentration resulting from the
source. Removal is defined as the removal rate of a pollutant from a microenvironment due to
deposition, filtration, and chemical reaction. The air exchange rate (Rair exchange) is expressed in
air changes per hour.
Table B.2-3. Mass balance model parameters.
Variable
' proximity
CS
" removal
p
" air exchange
V
Definition
Proximity factor
Concentration source
Removal rate due to deposition,
filtration, and chemical reaction
Air exchange rate
Volume of microenvironment
Units
unitless
ppb
1/hr
1/hr
mj
Value Range
' proximity — "
CS>0
"removal — "
" air exchange — U
V>0
The mass balance equation for a pollutant in a microenvironment is described by:
dt
where:
dCME(t)
ACm
- = ACm - ACout - ACremoval + ACsour
equation (B-3)
Change in concentration in a microenvironment at time t (ppb),
Rate of change in microenvironmental concentration due to influx
of air (ppb/hour),
Rate of change in microenvironmental concentration due to outflux
of air (ppb/hour),
B-17
-------
A Cremovai = Rate of change in microenvironmental concentration due to
removal processes (ppb/hour), and
A Csource = Rate of change in microenvironmental concentration due to an
emission source inside the microenvironment (ppb/hour).
Within the timestep selected, each of the rates of change, A. Cm, ACOMf, ACremova/, and
AC'source, is assumed to be constant. At each timestep of the simulation period, APEX estimates
the equilibrium, ending, and mean concentrations using a series of equations that account for
concentration changes expected to occur due to these physical processes. Details regarding these
equations are provided in the APEX TSD (US EPA, 2009b). The calculation continues to the
next timestep by using the end concentration for the previous timestep as the initial
microenvironmental concentration. A brief description of the input parameters estimates used
for microenvironments using the mass balance approach is provided below.
Factors Model
The factors method is simpler than the mass balance method. It does not calculate
concentration in a microenvironment from the concentration in the previous hour and it has
fewer parameters. Table B.2-4 lists the parameters required by the factors method to calculate
concentrations in a microenvironment without emissions sources.
Table B.2-4. Factors model parameters.
Variable
' proximity
i penetration
Definition
Proximity factor
Penetration factor
Units
unitless
unitless
Value Range
' proximity — "
U — l penetration — *
The factors method uses the following equation to calculate the timestep concentration in
a microenvironment from the user-provided hourly air quality data:
^ME = ^ambient X J proximity X J penetration equation (B-4)
where:
= Timestep concentration in a microenvironment (ppb)
= Timestep concentration in ambient environment (ppb)
/proximity = Proximity factor (unitless)
/penetration = Penetration factor (unitless)
B-18
-------
The ambient NC>2 concentrations are from the air quality data input file. The proximity
factor is a unitless parameter that represents the proximity of the microenvironment to a
monitoring station. The penetration factor is a unitless parameter that represents the fraction of
pollutant entering a microenvironment from outside the microenvironment via air exchange. The
development of the specific proximity and penetration factors used in this analysis are discussed
below for each microenvironment using this approach.
Microenvironments Modeled
In APEX, microenvironments represent the exposure locations for simulated individuals.
For exposures to be estimated accurately, it is important to have realistic microenvironments that
match closely to the locations where actual people spend time on a daily basis. As discussed
above, the two methods available in APEX for calculating pollutant levels within
microenvironments are: 1) factors and 2) mass balance. A list of microenvironments used in this
study, the calculation method used, and the parameters used to calculate the microenvironment
concentrations can be found in Table B.2-5.
Each of the microenvironments is designed to simulate an environment in which people
spend time during the day. CHAD locations are linked to the different microenvironments in the
Microenvironment Mapping File (see below). There are many more CHAD locations than
microenvironment locations (there are 113 CHAD codes versus 12 microenvironments in this
assessment), therefore most of the microenvironments have multiple CHAD locations mapped to
them.
B-19
-------
Table B.2-5. List of microenvironments and calculation methods used.
Microenvironment
No.
1
2
3
4
5
6
7
8
9
10
11
12
0
Name
Indoors - Residence
Indoors - Bars and restaurants
Indoors - Schools
Indoors - Day-care centers
Indoors -Office
Indoors -Shopping
Indoors - Other
Outdoors - Near road
Outdoors - Public garage - parking lot
Outdoors - Other
In-vehicle - Cars and Trucks
In-vehicle - Mass Transit (bus, subway, train)
Not modeled
Calculation
Method
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Mass balance
Factors
Factors
Factors
Factors
Factors
Parameter
Types used 1
AERandDE
AERand DE
AER and DE
AERandDE
AERandDE
AERandDE
AERandDE
PR
PR
None
PE and PR
PE and PR
1 AER=air exchange rate, DE=decay-deposition rate, PR=proximity factor, PE=penetration
factor
Mapping of APEX Microenvironments to CHAD Diaries
The Microenvironment Mapping file matches the APEX Microenvironments to CHAD
Location codes. Table B.2-6 gives the mapping used for the APEX simulations.
Table B.2-6. Mapping of CHAD activity locations to APEX microenvironments.
CHAD LOG.
U
X
30000
30010
30020
30100
30120
30121
30122
30123
30124
30125
30126
30127
30128
30129
30130
30131
30132
30133
30134
30135
Description
Uncertain of correct code
No data
Residence, general
Your residence
Other residence
Residence, indoor
Your residence, indoor
. . . , kitchen
. . . , living room or family room
. . . , dining room
. . . , bathroom
. . . , bedroom
. . . , study or office
. . . , basement
. . . , utility or laundry room
. . . , other indoor
Other residence, indoor
. . . , kitchen
. . . , living room or family room
. . . , dining room
. . . , bathroom
. . . , bedroom
APEX
-1
-1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
micro
Unknown
Unknown
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
Indoors -Residence
B-20
-------
30136 ..., study or office = 1
30137 ..., basement = 1
30138 ..., utility or laundry room = 1
30139 ..., other indoor = 1
30200 Residence, outdoor = 10
30210 Your residence, outdoor = 10
30211 ..., pool or spa = 10
30219 ..., other outdoor = 10
30220 Other residence, outdoor = 10
30221 ..., pool or spa = 10
30229 ..., other outdoor = 10
30300 Residential garage or carport = 7
30310 ..., indoor = 7
30320 ..., outdoor = 10
30330 Your garage or carport = 1
30331 ..., indoor = 1
30332 ..., outdoor = 10
30340 Other residential garage or carport = 1
30341 ..., indoor = 1
30342 ..., outdoor = 10
30400 Residence, none of the above = 1
31000 Travel, general = 11
31100 Motorized travel = 11
31110 Car = 11
31120 Truck = 11
31121 Truck (pickup or van) = 11
31122 Truck (not pickup or van) = 11
31130 Motorcycle or moped = 8
31140 Bus = 12
31150 Train or subway = 12
31160 Airplane = 0
31170 Boat = 10
31171 Boat, motorized = 10
31172 Boat, other = 10
31200 Non-motorized travel = 10
31210 Walk = 10
31220 Bicycle or inline skates/skateboard = 10
31230 In stroller or carried by adult = 10
31300 Waiting for travel = 10
31310 •••, bus or train stop = 8
31320 ..., indoors = 7
31900 Travel, other = 11
31910 ..., other vehicle = 11
32000 Non-residence indoor, general = 7
32100 Office building/ bank/ post office = 5
32200 Industrial/ factory/ warehouse = 5
32300 Grocery store/ convenience store = 6
32400 Shopping mall/ non-grocery store = 6
32500 Bar/ night club/ bowling alley = 2
32510 Bar or night club = 2
32520 Bowling alley = 2
32600 Repair shop = 7
32610 Auto repair shop/ gas station = 7
32620 Other repair shop = 7
32700 Indoor gym /health club = 7
32800 Childcare facility = 4
32810 ..., house = 1
32820 ..., commercial = 4
32900 Large public building = 7
32910 Auditorium/ arena/ concert hall = 7
32920 Library/ courtroom/ museum/ theater = 7
33100 Laundromat = 7
33200 Hospital/ medical care facility = 7
33300 Barber/ hair dresser/ beauty parlor = 7
33400 Indoors, moving among locations = 7
Indoors-Residence
Indoors-Residence
Indoors-Residence
Indoors-Residence
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Indoors-Other
Indoors-Other
Outdoors-Other
Indoors-Residence
Indoors-Residence
Outdoors-Other
Indoors-Residence
Indoors-Residence
Outdoors-Other
Indoors-Residence
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
Outdoors-Near_Road
In Vehicle-Mass_Transit
In Vehicle-Mass_Transit
Zero_concentration
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Other
Outdoors-Near_Road
Indoors-Other
In Vehicle-Cars_and_Trucks
In Vehicle-Cars_and_Trucks
Indoors-Other
Indoors-Office
Indoors-Office
Indoors-Shopping
Indoors-Shopping
Indoors-Bars_and_Restaurants
Indoors-Bars_and_Restaurants
Indoors-Bars_and_Restaurants
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Day_Care_Centers
Indoors-Residence
Indoors-Day_Care_Centers
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Other
Indoors-Other
B-21
-------
33500
33600
33700
33800
33900
34100
34200
34300
35000
35100
35110
35200
35210
35220
35300
35400
35500
35600
35610
35620
35700
35800
35810
35820
35900
36100
36200
36300
School
Restaurant
Church
Hotel/ motel
Dry cleaners
Indoor parking garage
Laboratory
Indoor, none of the above
Non-residence outdoor, general
Sidewalk, street
Within 10 yards of street
Outdoor public parking lot /garage
. . . , public garage
. . . , parking lot
Service station/ gas station
Construction site
Amusement park
Playground
. . . , school grounds
. . . , public or park
Stadium or amphitheater
Park/ golf course
Park
Golf course
Pool/ river/ lake
Outdoor restaurant/ picnic
Farm
Outdoor, none of the above
3
2
7
7
7
7
7
7
10
8
8
9
9
9
10
10
10
10
10
10
10
10
10
10
10
10
10
10
Indoors- Schools
Indoors-Bars and Restaurants
Indoors -Other
Indoors -Other
Indoors -Other
Indoors -Other
Indoors -Other
Indoors -Other
Outdoors -Other
Outdoors-Near Road
Outdoors-Near Road
Outdoors-Public Garage-Parking
Outdoors-Public Garage-Parking
Outdoors-Public Garage-Parking
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
Outdoors -Other
B.2.2.5 Exposure Calculations
APEX calculates exposure as a time series of exposure concentrations that a simulated
individual experiences during the simulation period. APEX determines the exposure using
hourly ambient air concentrations, calculated concentrations in each microenvironment based on
these ambient air concentrations (and indoor sources if present), and the minutes spent in a
sequence of microenvironments visited according to the composite diary. The hourly exposure
concentration at any clock hour during the simulation period is determined using the following
equation:
r~< timestep >
-ME(j) l (j)
7=1
where:
C,
N
/~i timestep
T
equation (B-5)
= Hourly exposure concentration at clock hour /' of the simulation period
(ppb)
= Number of events (i.e., microenvironments visited) in clock hour /' of the
simulation period.
= Timestep concentration in microenvironment y (ppm)
= Time spent in microenvironment y (minutes)
B-22
-------
T = Length of timestep (minutes)
From the timestep exposures, APEX calculates time series of 1-hour, 8-hour and daily average
exposures that a simulated individual would experience during the simulation period. APEX then
statistically summarizes and tabulates the timestep, hourly, 8-hour, and daily exposures. Note
that if the APEX timestep is greater than an hour, the 1-hour and 8-hour exposures are not
calculated and the corresponding tables are not produced. Exposures are calculated
independently for all pollutants in the simulation.
From the timestep exposures, APEX can calculate the time-series of 1-hour, 8-hour, and
daily average exposures that a simulated individual would experience during the simulation
period. APEX then statistically summarizes and tabulates the timestep (or hourly, daily, annual
average) exposures. In this analysis, the exposure indicator is 5-minute exposures above
potential health effect benchmark levels. From this, APEX can calculate two general types of
exposure estimates: counts of the estimated number of people exposed to a specified SO2
concentration level and the number of times per year that they are so exposed; the latter metric is
typically expressed in terms of person-occurrences or person-days. The former highlights the
number of individuals exposed at least one or more times per modeling period to the health
effect benchmark level of interest. APEX can also report counts of individuals with multiple
exposures. This person-occurrences measure estimates the number of times per season that
individuals are exposed to the exposure indicator of interest and then accumulates these
estimates for the entire population residing in an area.
APEX tabulates and displays the two measures for exposures above levels ranging from
any number of benchmark levels, by any increment (e.g., 0 to 800 ppb by 50 ppb increments for
5-minute exposures). These exposure results are tabulated for the population and subpopulations
of interest.
Exposure Model Output
All of the output files written by APEX are ASCII text files. Table B.2-7 lists each of the
output data files written for these simulations and provides descriptions of their content.
Additional output files that can produced by APEX are given in Table 5-1 of the APEX User's
Guide, and include hourly exposure, ventilation, and energy expenditures, and even detailed
event-level information, if desired. The names and locations, as well as the output table levels
B-23
-------
(e.g., output percentiles, cut-points), for these output files are specified by the user in the
simulation control parameters file.
Table B.2-7. Example of APEX output files.
Output File Type
Log
Profile Summary
Microenvironment
Summary
Sites
Output Tables
Description
The Log file contains the record of the APEX model simulation as it progresses.
If the simulation completes successfully, the log file indicates the input files and
parameter settings used for the simulation and reports on a number of different
factors. If the simulation ends prematurely, the log file contains error messages
describing the critical errors that caused the simulation to end.
The Profile Summary file provides a summary of each individual modeled in the
simulation.
The Microenvironment Summary file provides a summary of the time and
exposure by microenvironment for each individual modeled in the simulation.
The Sites file lists the tracts, districts, and zones in the study area, and identifies
the mapping between them.
The Output Tables file contains a series of tables summarizing the results of the
simulation. The percentiles and cut-off points used in these tables are defined
in the simulation control parameters file.
B-24
-------
B.3 Supplemental AERMOD Dispersion Modeling Data
B-25
-------
B.3-1 AERMOD Input data
Table B.3-1. Emission parameters by stack for all major facility stacks in Missouri.
Stack ID
5049
5050
5051
5054
5063
5064
5066
5068
5069
5070
5073
5074
City
LABADIE
LABADIE
LABADIE
LABADIE
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
Facility Name
AMERENUE-LABADIE
PLANT
AMERENUE-LABADIE
PLANT
AMERENUE-LABADIE
PLANT
AMERENUE-LABADIE
PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
NEI Site ID
NEI7514
NEI 7514
NEI 7514
NEI 7514
NEI 7525
NEI 7525
NEI 7525
NEI 7525
NEI 7525
NEI 7525
NEI 7525
NEI 7525
UTMX
(m)1
688,392
688,357
688,461
688,442
476,842
476,853
476,913
476,884
476,890
476,918
476,919
UTMY
(m)1
4,270,394
4,270,439
4,270,338
4,270,322
4,106,944
4,106,922
4,106,929
4,106,932
4,106,922
4,106,919
4,106,930
S02
Emissions
(tpy)
10,970
14,753
14,285
7,602
1,137
1,433
757
159
660
567
218
255
Stack
Height
(m)
213
213
213
213
107
107
61
61
61
61
60
60
Exit
Temp.
(K)
444
444
444
444
422
422
422
422
422
422
422
422
Stack
Diam.
(m)
6.2
6.2
8.8
8.8
2.5
2.5
3.7
3.7
3.7
3.7
3.7
3.7
Exit
Velocity
(mis)
28
28
28
28
15
15
6
6
5
5
6
6
Profile
Method2
Tier 1
Tier 1
Tier 1
Tier 1
Tier 2
TieM
Tier 1
Tier 1
TieM
Tier 1
Tier 1
Tier 1
B-26
-------
Stack ID
5076
5077
5084
5113
5114
5115
5131
5141
5145
5147
5148
5149
5244
5245
City
FIELD
SPRING-
FIELD
SPRING-
FIELD
SPRING-
FIELD
WEST
ALTON
WEST
ALTON
WEST
ALTON
HERCU-
LANEUM
HERCU-
LANEUM
HERCU-
LANEUM
FESTUS
FESTUS
FESTUS
STE. GENE-
VIE VE
STE. GENE-
VIE VE
Facility Name
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-JAMES
RIVER POWER PLANT
CITY UTILITIES OF
SPRINGFIELD
MISSOURI-
SOUTHWEST POWER
PLANT
AMERENUE-SIOUX
PLANT
AMERENUE-SIOUX
PLANT
AMERENUE-SIOUX
PLANT
DOE RUN COMPANY-
HERCULANEUM
SMELTER
DOE RUN COMPANY-
HERCULANEUM
SMELTER
DOE RUN COMPANY-
HERCULANEUM
SMELTER
AMERENUE-RUSH
ISLAND PLANT
AMERENUE-RUSH
ISLAND PLANT
AMERENUE-RUSH
ISLAND PLANT
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
NEI Site ID
NEI 7525
NEI 7525
NEI 12640
NEI 7516
NEI 7516
NEI 7516
NEI 34412
NEI 34412
NEI 34412
NEI 12618
NEI 12618
NEI 12618
NEI
M01 860001
NEI
MO1 860001
UTMX
(m)1
476,952
477,050
476,992
465,416
735,034
735,027
734,948
729,589
729,543
729,537
739,910
739,893
739,931
757,358
757,384
UTMY
(m)1
4,106,940
4,106,880
4,106,881
4,111,816
4,310,876
4,310,819
4,310,864
4,238,084
4,237,936
4,237,973
4,223,934
4,223,827
4,223,869
4,207,065
4,207,015
S02
Emissions
(tpy)
219
252
3,390
24,932
21,025
2
2
2
15,219
2
10,511
12,744
62
89
Stack
Height
(m)
60
60
117
183
183
65
3
9
168
76
213
213
23
23
Exit
Temp.
(K)
422
422
397
427
427
436
295
287
350
577
405
405
519
469
Stack
Diam.
(m)
3.7
3.7
3.4
5.8
5.8
1.4
0.0
0.3
6.1
1.5
8.8
8.8
3.2
3.4
Exit
Velocity
(mis)
6
6
21
29
29
15
0
6
18
9
25
25
4
6
Profile
Method2
Tier 1
Tier 1
Tier 2
TieM
TieM
TieM
Tier 2
TierS
Tier 2
Tier 1
Tier 1
Tier 1
TierS
TierS
B-27
-------
Stack ID
5246
5247
5248
5261
5262
5263
5264
5265
5267
5270
5271
5276
5277
5278
5279
5293
City
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
STE. GENE-
VIE VE
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
Facility Name
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
MISSISSIPPI LIME
COMPANY-
MISSISSIPPI LIME CO
AMERENUE-
MERAMEC PLANT
AMERENUE-
MERAMEC PLANT
AMERENUE-
MERAMEC PLANT
AMERENUE-
MERAMEC PLANT
ANHEUSER-BUSCH
INC-ST LOUIS
NEI Site ID
NEI
M01 860001
NEI
MO1 860001
NEI
MO1 860001
NEI
M01 860001
NEI
MO1 860001
NEI
M01 860001
NEI
MO1 860001
NEI
M01 860001
NEI
MO1 860001
NEI
M01 860001
NEI
MO1 860001
NEI 7515
NEI 7515
NEI 7515
NEI 7515
NEI 34732
UTMX
(m)1
757,697
757,666
757,697
757,561
757,735
757,727
757,550
757,524
757,633
757,627
757,540
732,584
732,631
732,677
732,714
742,736
UTMY
(m)1
4,206,939
4,206,950
4,206,981
4,206,988
4,206,971
4,206,997
4,206,964
4,206,924
4,206,999
4,206,989
4,206,931
4,253,799
4,253,790
4,253,784
4,253,779
4,275,786
S02
Emissions
(tpy)
103
106
105
1,290
1,394
1,505
67
77
2
1
1,199
5,195
6,463
2,359
2,430
2
Stack
Height
(m)
23
23
23
35
35
35
35
35
20
20
35
107
107
76
76
30
Exit
Temp.
(K)
469
469
469
343
343
344
346
346
367
362
343
463
447
436
436
371
Stack
Diam.
(m)
3.4
3.4
3.4
1.7
1.7
1.7
2.1
2.1
1.1
1.2
1.7
4.9
4.3
3.4
3.2
1.2
Exit
Velocity
(mis)
6
6
6
11
11
13
9
9
15
11
11
33
31
27
27
3
Profile
Method2
TierS
TierS
TierS
TierS
TierS
TierS
TierS
TierS
Tier 2
TierS
TierS
Tier 1
Tier 1
Tier 1
Tier 1
Tier 2
B-28
-------
Stack ID
5295
5296
5297
5298
5299
5302
5304
City
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
ST. LOUIS
Facility Name
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
ANHEUSER-BUSCH
INC-ST LOUIS
NEI Site ID
NEI 34732
NEI 34732
NEI 34732
NEI 34732
NEI 34732
NEI 34732
NEI 34732
UTMX
(m)1
742,775
742,750
742,781
742,800
742,759
742,739
742,711
UTMY
(m)1
4,275,743
4,275,704
4,275,753
4,275,764
4,275,714
4,275,677
4,275,740
S02
Emissions
(tpy)
176
256
249
158
3,066
2,339
4
Stack
Height
(m)
69
69
69
69
69
69
22
Exit
Temp.
(K)
450
450
450
450
461
439
486
Stack
Diam.
(m)
3.0
3.0
3.0
3.0
3.0
3.0
1.2
Exit
Velocity
(mis)
6
6
6
6
6
6
9
Profile
Method2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Notes:
1 UTM Zone 15 values in all cases.
2 Three methods were possible to convert annual total emissions data from the NEI into hourly temporal profiles required for AERMOD, based on availability of
data:
Tier 1 : CAMD hourly concentrations to create relative temporal profiles.
Tier 2: EMS-HAP seasonal and diurnal temporal profiles for source categorization codes (SCCs).
Tier 3: Flat profiles
B-29
-------
Table B.3-2. Emission parameters by stack for all major cross-border facility stacks in the St. Louis scenario.
Stack ID
1
2
3
4
5
9
10
11
City
East
Alton
East
Alton
Baldwin
Baldwin
Baldwin
Hartford
Hartford
Hartford
Facility Name
DYNEGY
MIDWEST
GENERATION
INC
DYNEGY
MIDWEST
GENERATION
INC
DYNEGY
MIDWEST
GENERATION
INC
DYNEGY
MIDWEST
GENERATION
INC
DYNEGY
MIDWEST
GENERATION
INC
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
NEI Site ID
NEI52119
NEI52119
NEI52781
NEI52781
NEI52781
NEI52159
NEI52159
NEI52159
UTMX
(m)1
748,654
748,654
775,316
775,316
775,316
753,003
752,783
752,886
UTM Y
(m)1
4,305,518
4,305,518
4,233,202
4,233,202
4,233,202
4,302,381
4,302,408
4,302,285
SO2
Emissions
(tpy)
1536.2
5725.8
9931.4
9053
7283
131.95786
907.24
132.9
Stack
Height
(m)
76.2
106.7
184.4
184.4
184.4
33.5
19.9
24.1
Exit
Temp.
(K)
427.6
416.5
425.4
428.7
424.8
533.2
502.0
519.3
Stack
Diam.
(m)
5.2
4.6
5.9
5.9
5.9
1.5
1.1
2.1
Exit
Velocity
(m/s)
8.5
34.6
39.7
38.3
38.4
3.1
6.5
7.0
Profile
Method2
TieM
TieM
TieM
TieM
TieM
Tier 2
Tier 2
Tier 2
B-30
-------
Stack ID
12
13
14
15
16
17
18
19
20
City
Hartford
Hartford
Hartford
Hartford
Hartford
Hartford
Sauget
Sauget
Roxana
Facility Name
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
Premcor
Refining Group
(prev. Clark Oil
and Refining
Corp.)
BIG RIVER
ZINC CORP
BIG RIVER
ZINC CORP
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
NEI Site ID
NEI52159
NEI52159
NEI52159
NEI52159
NEI52159
NEI52159
NEI53013
NEI53013
NEI55835
UTMX
(m)1
752,883
753,003
752,886
752,783
753,003
752,783
746,429
746,429
755,000
UTMY
(m)1
4,302,377
4,302,381
4,302,285
4,302,408
4,302,381
4,302,408
4,276,339
4,276,339
4,302,599
S02
Emissions
(tpy)
106.67
79
66.43
4.90219
171.36006
7.42
1.34
1377.28
15.38
Stack
Height
(m)
24.4
36.0
16.8
35.4
30.5
30.7
21.3
25.9
106.7
Exit
Temp.
(K)
533.2
533.2
677.6
570.4
533.2
513.2
317.6
422.0
472.0
Stack
Diam.
(m)
1.8
1.2
1.8
1.5
1.5
1.7
0.7
0.9
4.6
Exit
Velocity
(mis)
2.6
3.1
6.2
7.8
4.1
11.4
10.6
41.3
11.4
Profile
Method2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
B-31
-------
Stack ID
21
22
23
24
25
26
27
28
29
30
31
32
City
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Roxana
Facility Name
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
NEI Site ID
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
NEI55835
UTMX
(m)1
755,000
755,188
754,405
754,997
754,505
754,994
755,084
754,658
754,994
755,000
754,658
755,231
UTMY
(m)1
4,302,599
4,302,550
4,303,105
4,302,691
4,302,984
4,302,783
4,303,003
4,302,515
4,302,783
4,302,599
4,302,515
4,302,561
S02
Emissions
(tpy)
7.27
1.2
1.25
1.45
1.53
3.39
1.15
385.25
3.24
16.73
11677.82
212.41
Stack
Height
(m)
106.7
45.7
56.4
61.0
95.1
40.2
45.7
36.9
45.7
106.7
10.1
45.7
Exit
Temp.
(K)
463.7
628.2
432.6
672.0
483.7
491.5
699.8
754.8
431.5
483.7
293.7
699.8
Stack
Diam.
(m)
4.6
2.3
2.4
3.7
4.3
2.1
2.3
3.4
3.0
4.6
0.1
2.4
Exit
Velocity
(mis)
0.3
7.9
6.7
6.7
0.3
13.2
7.0
5.9
15.9
0.3
0.1
8.8
Profile
Method2
Tier 2
Tierl
Tier 2
Tier 2
Tier 2
Tier 2
Tierl
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
B-32
-------
Stack ID
33
34
35
36
37
38
39
40
41
42
City
Roxana
Roxana
Roxana
Granite
City
Granite
City
Granite
City
Granite
City
Granite
City
Granite
City
Granite
City
Facility Name
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
ConocoPhillips
Co. (prev.
Phillips 66 Co.)
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NEI Site ID
NEI55835
NEI55835
NEI55835
NEI55848
NEI55848
NEI55848
NEI55848
NEI55848
NEI55848
NEI55848
UTMX
(m)1
755,231
753,801
755,231
747,795
750,041
749,778
749,897
749,970
750,041
749,847
UTMY
(m)1
4,302,561
4,303,085
4,302,561
4,286,723
4,286,824
4,286,692
4,286,788
4,286,761
4,286,824
4,286,849
S02
Emissions
(tpy)
206.96
110.6
108.6
61.88
506.7
228.47
421.58883
375.19
351.93
264.95442
Stack
Height
(m)
45.7
38.1
45.7
30.5
46.3
24.5
68.6
15.4
46.3
61.0
Exit
Temp.
(K)
672.0
792.0
672.0
616.5
441.5
372.0
460.9
453.7
441.5
460.9
Stack
Diam.
(m)
2.4
2.2
2.4
2.1
2.1
1.5
4.3
0.9
2.1
3.4
Exit
Velocity
(mis)
8.2
5.4
4.3
17.9
10.6
6.2
4.5
9.9
1.2
3.1
Profile
Method2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
B-33
-------
Stack ID
43
44
46
47
50
51
City
Granite
City
Granite
City
Granite
City
Granite
City
Granite
City
Granite
City
Facility Name
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NATIONAL
STEEL CORP -
GRANITE CITY
DIV
NEI Site ID
NEI55848
NEI55848
NEI55848
NEI55848
NEI55848
NEI55848
UTMX
(m)1
750,280
749,828
747,842
750,180
748,053
750,255
UTMY
(m)1
4,286,925
4,286,663
4,286,755
4,286,983
4,287,055
4,286,924
S02
Emissions
(tpy)
923.52
501.19
85.86
20.99
8
959.82
Stack
Height
(m)
43.5
43.5
30.5
24.9
19.2
43.5
Exit
Temp.
(K)
538.7
538.7
616.5
335.9
323.7
538.7
Stack
Diam.
(m)
2.0
2.0
2.1
1.5
2.1
2.0
Exit
Velocity
(mis)
9.2
9.2
17.9
8.9
13.1
9.2
Profile
Method2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Notes:
1 UTM Zone 15 values in all cases.
2 Three methods were possible to convert annual total emissions data from the NEI into hourly temporal profiles required for AERMOD, based on
availability of data:
Tier 1 : CAMD hourly concentrations to create relative temporal profiles.
Tier 2: EMS-HAP seasonal and diurnal temporal profiles for source categorization codes (SCCs).
Tier 3: Flat profiles
B-34
-------
B.3.2 AERMOD Air Quality Evaluation Data
Table B.3-2. Measured ambient monitor SO2 concentration distributions and the
modeled monitor receptor and receptors within 4 km of the ambient monitors in
Greene County for year 2002.
Ambient
Monitor ID
290770026
290770032
290770037
290770040
290770041
Receptor(s)1
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
Percentile Concentration (ppb)
100
29
48
101
114
48
30
41
62
28
42
35
53
106
144
115
34
53
106
203
116
31
52
108
33
73
99
6
12
46
46
22
10
12
14
8
14
5
13
55
49
42
5
13
55
18
45
5
14
56
9
23
95
2
4
18
16
11
5
6
8
6
8
2
3
21
8
5
2
3
21
6
6
2
3
22
3
4
90
1
2
8
7
6
3
4
6
5
6
1
2
8
4
3
1
2
8
3
3
1
2
8
2
2
80
0
1
2
3
2
2
3
5
4
5
0
1
2
2
1
0
1
2
2
1
0
1
2
1
1
70
0
1
1
2
1
1
2
4
4
4
0
0
1
2
1
0
0
1
1
1
0
0
1
1
1
60
0
0
1
1
1
1
2
3
3
3
0
0
1
2
1
0
0
1
1
1
0
0
1
1
1
50
0
0
1
1
1
1
2
3
3
3
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
25
0
0
0
1
0
0
1
2
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Notes:
1 AERMOD concentrations are for the given percentile (p2.5 = 2.5th; p50 = 50th; p97.5 = 97.5th) of the
modeled distribution of all modeled air quality receptors within 4 km of ambient monitor. AERMOD
monitor is the concentration prediction at the ambient monitor location using AERMOD.
B-35
-------
Table B.3-3. Measured ambient monitor SO2 concentration diurnal profile and the
modeled monitor receptor and receptors within 4 km of the ambient monitors in
Greene County for year 2002.
Monitor ID
290770026
290770032
Hour
of
Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Annual Average SO2 Concentration (ppb) at Given Receptor
AERMOD
P2.5
0.2
0.2
0.2
0.1
0.1
0.1
0.2
0.5
0.7
0.8
0.7
0.5
0.6
0.6
0.5
0.6
0.6
0.5
0.3
0.2
0.2
0.2
0.2
0.2
1.5
1.4
1.4
1.3
1.1
1.1
1.4
1.9
1.7
1.5
1.3
1.2
1.1
1.1
1.0
1.1
1.4
1.6
1.8
AERMOD
P50
0.4
0.4
0.4
0.3
0.3
0.3
0.7
1.5
1.5
1.6
1.4
1.2
1.1
1.0
1.0
1.0
1.3
1.2
0.9
0.5
0.5
0.5
0.4
0.4
2.3
2.1
2.2
2.0
1.7
1.8
2.0
2.2
2.3
2.3
2.2
2.2
2.1
2.0
2.0
2.1
2.4
2.4
2.4
AERMOD
P97.5
1.6
1.8
2.0
1.6
1.5
1.7
2.3
4.2
5.6
5.6
6.0
6.3
6.1
6.0
5.8
5.1
4.6
3.7
2.3
1.7
1.7
1.9
1.9
1.9
3.2
3.0
3.4
2.8
2.4
2.5
2.5
2.7
3.4
3.9
4.0
4.1
4.3
4.1
4.0
4.1
3.6
3.3
3.2
Ambient
Monitor
2.7
2.4
2.5
2.5
2.9
2.9
3.6
4.2
4.6
5.3
5.0
4.7
4.5
4.1
3.7
3.8
3.7
2.9
2.6
3.0
2.8
3.0
2.9
2.9
2.8
2.6
2.5
2.5
2.3
2.2
2.3
2.7
3.0
3.3
3.2
3.2
3.3
3.2
3.2
3.1
3.2
3.1
3.1
AERMOD
Monitor
1.2
1.0
0.9
0.9
0.8
0.9
1.4
2.8
3.6
3.4
3.7
3.2
2.8
2.6
2.6
2.5
2.9
2.9
2.4
1.4
1.4
1.4
1.1
1.1
3.0
2.9
2.8
2.7
2.5
2.4
2.5
2.9
3.2
3.6
3.5
3.6
3.6
3.6
3.5
3.4
3.5
3.5
3.4
B-36
-------
Monitor ID
290770037
290770040
Hour
of
Day
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Annual Average SO2 Concentration (ppb) at Given Receptor
AERMOD
P2.5
1.7
1.8
1.7
1.6
1.4
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0.4
0.6
0.8
0.7
0.6
0.6
0.6
0.5
0.5
0.5
0.4
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0.4
0.6
0.8
0.7
0.5
0.6
0.6
0.5
0.5
0.5
0.4
AERMOD
P50
2.5
2.4
2.6
2.5
2.4
0.2
0.2
0.2
0.2
0.2
0.2
0.4
1.0
1.2
1.5
1.4
1.3
1.3
1.2
1.2
1.2
1.2
1.0
0.5
0.3
0.3
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.4
1.0
1.2
1.5
1.4
1.3
1.3
1.2
1.2
1.2
1.1
1.0
AERMOD
P97.5
3.2
3.3
3.6
3.4
3.2
1.4
1.5
1.7
1.5
1.4
1.6
2.4
4.5
6.2
6.4
7.0
6.9
7.0
7.2
6.5
5.7
5.2
4.2
2.5
1.4
1.5
1.7
1.8
1.6
1.4
1.5
1.7
1.5
1.4
1.6
2.4
4.5
6.2
6.4
7.0
6.9
7.0
7.2
6.5
5.7
5.2
4.2
Ambient
Monitor
3.1
3.1
3.1
3.1
2.9
1.6
1.5
1.5
1.9
1.9
1.8
1.9
2.3
3.1
3.8
4.1
4.8
5.0
5.2
5.3
4.9
4.2
3.0
2.2
2.2
1.9
1.8
1.9
2.1
1.0
0.8
1.0
1.0
1.0
1.0
1.0
1.2
1.6
2.2
2.4
2.9
2.4
3.0
2.8
2.2
1.8
1.8
AERMOD
Monitor
3.4
3.3
3.4
3.4
3.1
0.5
0.5
0.5
0.4
0.3
0.3
0.3
1.0
1.9
3.7
4.6
5.4
5.0
4.6
4.3
3.1
2.4
1.9
0.7
0.5
0.5
0.6
0.5
0.5
0.5
0.5
0.5
0.4
0.3
0.3
0.3
0.8
1.8
3.7
4.8
6.2
5.9
4.8
4.6
3.2
2.6
1.8
B-37
-------
Monitor ID
290770041
Hour
of
Day
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Annual Average SO2 Concentration (ppb) at Given Receptor
AERMOD
P2.5
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0.5
0.6
0.7
0.6
0.4
0.5
0.4
0.4
0.4
0.4
0.4
0.3
0.2
0.2
0.2
0.2
0.2
AERMOD
P50
0.5
0.3
0.3
0.3
0.2
0.2
0.3
0.3
0.3
0.2
0.2
0.2
0.6
1.2
1.4
1.7
1.5
1.4
1.4
1.3
1.2
1.3
1.2
1.1
0.5
0.3
0.3
0.4
0.3
0.3
AERMOD
P97.5
2.5
1.4
1.5
1.7
1.8
1.6
1.5
1.7
1.9
1.5
1.4
1.7
2.4
4.9
6.2
6.5
7.2
7.0
7.4
7.9
6.6
6.2
5.4
4.2
2.6
1.5
1.6
1.8
1.9
1.8
Ambient
Monitor
1.2
1.0
1.0
0.9
0.9
1.0
0.6
0.6
0.6
0.5
0.4
0.6
0.6
0.8
1.1
1.4
1.5
1.5
1.6
1.3
1.1
1.0
0.9
0.7
0.6
0.6
0.7
0.7
0.6
0.7
AERMOD
Monitor
0.8
0.5
0.5
0.6
0.5
0.5
0.6
0.5
0.6
0.5
0.4
0.4
0.6
1.5
1.7
1.8
2.4
2.1
2.4
2.0
2.1
2.3
2.1
1.8
0.8
0.6
0.6
0.7
0.6
0.6
B-38
-------
Table B.3-4. Measured ambient monitor SO2 concentration distributions and the
modeled monitor receptor and receptors within 4 km of the ambient monitors in
St. Louis for year 2002.
Ambient
Monitor ID
291890004
291890006
291893001
291895001
291897003
295100007
295100086
Receptor(s)1
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
AERMOD P2.5
AERMOD P50
AERMOD P97.5
Ambient Monitor
AERMOD Monitor
Percentile Concentration (ppb)
100
60
69
103
99
67
48
55
94
85
73
58
75
91
80
71
97
168
545
158
191
67
89
138
91
99
71
93
137
64
100
71
91
124
86
111
99
20
22
25
24
22
19
20
20
18
20
24
26
30
24
25
32
38
51
23
40
25
28
32
24
27
26
31
43
25
32
29
32
36
30
31
95
11
12
14
13
11
10
11
12
9
11
13
14
17
12
14
13
14
15
12
14
11
12
13
12
12
13
18
23
14
17
15
18
22
16
17
90
7
8
9
8
7
7
7
8
6
8
9
10
12
8
10
8
9
10
8
9
7
8
9
8
8
8
11
16
10
11
11
13
16
11
13
80
4
5
5
5
4
4
5
5
3
5
5
6
8
5
6
5
6
6
5
6
4
5
5
5
5
4
7
10
6
7
7
9
11
7
8
70
3
3
4
3
3
3
3
4
2
3
4
4
6
4
4
4
4
4
4
4
3
3
4
4
4
3
5
8
5
6
5
6
8
5
6
60
2
2
3
2
2
2
2
3
1
3
3
3
5
3
3
3
3
3
3
3
2
3
3
3
3
2
4
6
3
5
4
5
6
4
5
50
2
2
2
1
2
2
2
2
1
2
2
2
3
2
2
2
2
2
2
2
2
2
2
2
2
2
3
5
2
4
3
4
5
3
4
25
1
1
1
0
1
1
1
1
0
1
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
2
1
2
3
1
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
0
0
0
0
0
0
0
0
0
Notes:
1 AERMOD concentrations are for the given percentile (p2.5 = 2.5th; p50 = 50th; p97.5 = 97.5th) of the
modeled distribution of all modeled air quality receptors within 4 km of ambient monitor. AERMOD
monitor is the concentration prediction at the ambient monitor location using AERMOD.
B-39
-------
Table B.3-5. Measured ambient monitor SO2 concentration diurnal profile and the
modeled monitor receptor and receptors within 4 km of the ambient monitors in
St. Louis for year 2002.
Ambient
Monitor ID
291890004
291890006
Hour
of
Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Annual Average SO2 Concentration (ppb) at Given Receptor
AERMOD
P2.5
2.0
1.7
1.5
1.5
1.4
1.4
1.8
2.7
4.0
4.6
4.8
4.6
4.4
4.1
3.9
3.7
3.8
3.5
3.1
3.0
2.7
2.4
2.1
2.1
2.4
2.1
1.7
1.8
1.6
1.5
2.0
2.5
4.2
4.4
4.8
4.3
4.2
4.1
3.9
3.7
3.8
3.4
3.0
AERMOD
P50
2.6
2.2
1.9
1.9
1.8
1.8
2.2
3.2
4.5
5.0
5.0
5.2
4.9
4.7
4.3
4.2
4.4
4.2
3.6
3.5
3.4
2.9
2.7
2.6
2.6
2.3
1.9
2.0
1.8
1.6
2.0
2.7
4.3
4.6
4.9
4.5
4.3
4.2
4.0
3.8
4.0
3.7
3.3
AERMOD
P97.5
3.2
2.7
2.3
2.3
2.1
2.6
2.6
4.0
4.8
5.4
5.3
5.6
5.2
5.0
4.6
4.4
4.7
4.6
4.0
3.7
3.7
3.3
3.3
3.0
2.9
2.6
2.1
2.3
1.9
1.7
2.1
2.7
4.4
4.8
5.3
4.8
4.7
4.6
4.4
4.2
4.6
4.5
3.8
Ambient
Monitor
2.4
2.1
1.8
1.8
1.8
2.0
2.4
3.3
4.1
4.6
4.6
4.5
4.1
4.3
4.0
3.6
3.8
3.8
3.4
2.9
2.8
2.9
2.9
2.5
1.6
1.6
1.5
1.2
1.2
1.2
1.4
1.9
2.7
3.4
3.5
3.8
3.5
2.9
2.6
2.9
2.7
2.6
2.4
AERMOD
Monitor
2.2
1.8
1.7
1.6
1.7
1.8
2.0
3.0
4.2
4.7
4.8
4.8
4.5
4.2
4.0
3.7
3.9
3.5
3.1
3.1
2.9
2.6
2.4
2.2
2.6
2.5
2.0
2.1
1.8
1.5
2.0
2.5
4.3
4.7
5.1
4.6
4.6
4.6
4.3
4.2
4.3
4.1
3.4
B-40
-------
291893001
291895001
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
2.9
2.8
2.5
2.3
2.2
3.0
2.8
2.3
2.5
2.1
2.2
2.6
3.2
4.5
5.2
5.3
5.3
5.0
4.6
4.5
4.3
4.5
4.3
3.9
3.6
3.6
3.3
3.2
3.0
3.2
3.1
2.8
2.8
2.4
2.5
2.6
3.2
4.5
5.3
5.5
5.6
4.6
4.6
4.4
4.1
4.1
4.5
3.2
3.3
3.1
3.1
3.0
2.7
2.8
2.4
3.6
3.2
2.7
2.8
2.5
2.6
3.0
3.7
5.1
5.6
5.8
5.8
5.4
5.1
4.8
4.7
4.8
4.8
4.3
3.9
4.0
3.8
3.8
3.7
3.7
3.7
3.2
3.3
3.0
3.6
3.2
4.3
5.5
6.0
5.9
5.7
5.3
5.1
4.8
4.6
4.5
4.8
4.2
4.1
4.3
3.4
3.4
3.0
3.2
2.8
4.2
3.9
3.3
3.4
2.8
3.1
3.7
4.6
6.5
7.8
8.1
8.2
7.7
7.8
7.3
7.1
6.4
6.0
5.3
4.7
4.6
4.7
4.5
4.4
4.9
5.1
3.8
4.3
3.9
4.9
3.9
5.3
6.1
6.6
6.1
6.0
5.7
5.3
5.0
4.9
4.8
5.3
5.3
5.5
5.6
2.4
2.0
2.0
1.9
1.8
2.2
2.2
2.0
1.7
1.8
2.3
2.8
3.7
4.5
4.9
4.8
4.7
4.4
4.5
4.1
3.8
3.8
4.1
3.7
3.5
3.4
3.2
3.0
2.7
3.1
2.9
2.7
2.7
2.7
2.8
3.2
4.0
4.6
4.7
5.2
5.1
4.6
4.1
4.1
3.9
4.1
3.9
3.7
3.2
3.5
3.0
3.1
2.8
2.8
2.7
3.5
3.2
2.7
2.8
2.5
2.7
3.1
3.6
5.0
5.7
5.8
5.8
5.3
5.0
4.8
4.7
4.7
4.8
4.2
3.9
4.0
3.7
3.9
3.6
3.8
3.5
3.2
3.2
3.3
3.8
3.0
4.7
5.4
6.0
6.0
6.0
5.3
5.2
4.9
4.6
4.7
4.9
3.7
4.0
3.9
B-41
-------
291897003
295100007
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
3.5
3.7
3.8
2.3
2.6
2.1
2.2
1.7
2.1
1.8
2.9
4.1
4.9
4.8
4.5
4.3
4.0
3.9
3.6
3.7
3.2
3.1
2.8
2.8
3.4
2.7
3.0
2.2
2.1
1.6
1.7
1.4
1.6
2.9
3.6
5.0
5.1
5.1
4.9
4.6
4.5
4.0
4.0
4.4
4.2
3.6
3.5
3.2
2.7
2.6
4.4
4.4
4.3
2.8
3.0
2.6
2.5
2.2
2.5
2.1
3.4
4.4
5.3
5.2
5.0
4.7
4.4
4.4
4.1
4.3
4.2
3.8
3.2
3.4
3.7
3.3
3.4
3.7
3.4
3.0
3.1
3.0
3.1
4.5
5.2
6.6
6.8
7.1
7.0
6.9
6.8
6.3
6.2
6.2
6.0
5.7
5.1
4.8
4.4
4.0
5.6
6.3
5.1
3.3
3.2
2.9
3.0
2.7
3.2
2.6
3.9
5.0
5.8
5.6
5.6
5.1
4.8
4.7
4.4
4.6
4.7
4.5
4.0
3.9
4.0
4.2
3.8
5.8
5.6
5.3
5.2
5.1
5.2
7.6
7.8
8.4
8.2
8.5
8.2
8.1
8.1
7.6
7.8
8.2
8.7
10.0
7.4
7.1
6.6
5.9
3.3
3.2
2.9
2.6
2.4
2.5
2.4
2.3
2.5
3.2
4.1
4.7
4.8
5.1
4.7
4.5
4.4
4.2
3.8
3.7
3.9
3.6
3.5
3.3
3.1
3.1
2.9
3.4
3.2
3.2
3.0
2.9
3.1
3.7
4.1
5.2
5.7
5.5
4.9
4.7
4.6
4.4
4.2
4.2
3.9
3.8
4.2
4.1
4.0
4.1
4.7
5.6
4.3
2.7
3.2
2.7
2.4
2.2
2.5
2.1
3.2
4.5
5.4
5.4
5.1
4.9
4.6
4.6
4.3
4.5
4.4
3.8
3.2
3.2
3.9
3.6
3.6
4.0
3.5
3.3
3.3
3.1
3.2
4.6
5.2
6.6
7.0
7.5
7.2
7.1
7.1
6.6
6.6
6.6
6.3
6.5
5.2
4.9
4.5
4.0
B-42
-------
295100086
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2.3
3.8
3.3
3.1
3.0
3.0
2.3
3.7
4.2
5.8
6.7
6.6
6.5
6.2
6.1
5.7
5.6
5.3
5.2
5.0
4.7
4.6
4.3
4.1
3.8
3.9
4.9
4.7
4.3
4.0
3.8
3.9
4.4
5.5
7.0
8.1
8.0
7.8
7.7
7.5
7.1
7.2
7.2
7.2
6.5
6.2
6.3
5.6
5.4
5.1
5.8
6.5
6.7
5.5
5.6
5.8
5.9
7.0
7.0
8.0
8.4
8.4
8.3
8.1
7.9
7.4
7.5
7.9
8.5
9.2
8.0
7.7
7.9
7.3
6.9
3.7
4.3
4.2
3.7
3.9
3.9
4.3
4.4
5.4
6.3
6.0
6.0
5.4
5.1
4.9
4.6
4.7
4.9
4.4
4.6
4.7
4.9
4.8
4.3
4.2
3.9
4.8
4.4
4.1
4.0
3.6
4.0
4.3
5.3
6.9
8.1
8.0
7.8
7.6
7.4
7.1
7.2
6.9
7.0
6.2
6.0
6.1
5.1
5.1
4.8
B-43
-------
B.4 SUPPLEMENTAL APEX EXPOSURE MODELING
DATA
B.4.1 APEX Input Data Distributions for SO2 deposition
In recognizing the relationship between SO2 deposition rate and various surface
types within indoor microenvironments and that the presence of these surfaces would
vary in proportions dependent on the microenvironment, staff estimated the APEX input
SC>2 deposition rate distributions using a Monte Carlo sampling approach. First, 1,000
different hypothetical indoor microenvironments were simulated, each with a different
ratio of wall area to floor area and furniture area to floor area. Based on these ratios,
surface area to volume ratios were estimated in each sample indoor microenvironment.
Then, surface area to volume ratios were used to convert the deposition velocities to
deposition rates in hr"1 by dividing the velocities by the surface area to volume ratio and
then making an appropriate unit conversion. And finally, the deposition rate for each
surface type was combined using a weighted average to estimate an effective deposition
rate, as follows:
A A
r\ , r\ * ceiling j^. % furntiture
^floor ' ^ceiling . ' ^furniture .
Deff = *" *=- equation (B-6)
^ ceiling furniture
A A
•^ floor ^floor
where D denotes deposition rate, A denotes area of the indoor microenvironment,
and Deff is the effective deposition rate. If more than one surface type is present in the
sample indoor microenvironment (e.g. both carpet and non-carpeted floors), these values
were first averaged using the fraction of the room that contains each. Details regarding
the data used for estimating the SC>2 deposition rate within simulated indoor
microenvironments are provided in the following sections.
B.4.1.1 Surface deposition data and surface type mapping
Staff obtained SC>2 deposition velocities from a literature review conducted by
Grantoft and Raychaudhuri (2004). These authors categorized the data by several
relative humidities and considering several different surface types. Staff mapped the
B-44
-------
surface classes reported in Grantoft and Raychaudhuri (2004) to surface types typically
found within indoor microenvironments (Table B.4-1).
Table B.4-1. Classification of SO2 deposition data for several
microenvironmental surfaces.
Surface
Category
Floor
Ceiling
Wall
Furniture
Surface Type
Carpet
Floor
Ceiling Tile
Ceiling Wallboard
Wallpaper
Wall Wallboard
Wood paneling
Furniture
Surface Class1
Average of the wool
and synthetic carpet
values
Synthetic Floor
Covering - medium
worn
Coarse composite
panels
Treated gypsum
wallboard
Wall paper
Treated gypsum
wallboard
Surface treated wood
work and wall boards
Cloth
Deposition in cm/s1
50%
Relative
Humidity
0.0625
0.007
0.14
0.048
0.036
0.048
0.014
0.019
70%
Relative
Humidity
0.075
0.015
0.15
0.16
0.043
0.16
0.047
0.023
90%
Relative
Humidity
0.117
0.032
0.18
0.27
0.068
0.27
0.078
0.036
Notes:
1 Obtained from Table 6 of Grantoft and Raychaudhuri (2004).
B.4.1.2 Indoor Microenvironment Configurations
Because the configuration of rooms within a building will affect the wall area to
floor area ratio, staff first estimated the areas of several indoor microenvironments. Staff
had to make several assumptions due to the limited availability of data. The first broad
assumption was that a single room within the indoor microenvironment could represent
all potential rooms within the particular building type. Secondly, staff assumed all rooms
were square to calculate the area distributions. Additional assumptions specific to the
type of indoor microenvironment are provided below, along with the estimated indoor
microenvironment area distributions.
B-45
-------
Residential area distributions
In residences, the American Housing Survey (AHS, 2008) provides a matrix that
gives the number of survey homes within a given total square footage and a given
number of rooms category. Staff converted the data to probabilities using the total
number of homes in each category (Table B.4-2). In calculating the room area using
these distributions, a series of two independent random numbers were used to select a
square footage category and then to find the number of rooms within that square footage
category, accounting for the inherent correlation of the number of rooms in a given
building with the total square footage. Staff derived a representative room area by
dividing the square footage by the number of rooms.
Table B.4-2. Distributions used to calculate a representative room size in an
indoor residential microenvironment.
Cumulative
probability
for number
of rooms
within
each
square
footage
class
1
Cumulative
probability
for each
square
footage
class — *•
Rooms
1
2
3
4
5
6
7
8
9
10
Square Footage
250
0.01
750
0.10
1250
0.35
1750
0.60
2250
0.77
2750
1.00
0.07
0.13
0.40
0.64
0.81
0.90
0.97
0.99
0.99
1.00
0.00
0.00
0.07
0.47
0.80
0.94
0.98
0.99
1.00
0.00
0.00
0.01
0.13
0.54
0.86
0.96
0.99
1.00
0.00
0.00
0.00
0.04
0.28
0.65
0.90
0.98
0.99
1.00
0.00
0.00
0.00
0.02
0.15
0.41
0.71
0.92
0.98
1.00
0.00
0.00
0.00
0.01
0.08
0.23
0.46
0.71
0.87
1.00
Non-residential area distributions
An office can contain many different rooms, each with either one or two
occupants (usually a smaller office) or a collection of cubicles (usually a larger office).
B-46
-------
Staff used the Building Assessment Survey and Evaluation study (BASE; US EPA,
2008a) to generate representative office areas for simulated buildings. The BASE data
provided the mean, standard deviation, minimum, and maximum of the total square
footage and the number of people per square meter of occupied space (Table B.4-3).1
Based on this, staff represented the data as a normal distribution and set the lower and
upper limits using the minimum and maximum observations. BASE (US EPA, 2008a)
also provides the average number of occupants in private or semi-private work areas
(40%) compared to shared space (60%).2 Staff assumed that the private and semi-private
offices have an average of two people in each and the shared spaces have an average of
six people in each. In calculating the area, two independent random numbers were used
to select the total floor area and the number of occupants in that floor area. The total
square footage of the office was then divided by the number of rooms to obtain the
representative office area.
For schools, the distribution of the total building square footage is available from
the Commercial Building Energy Consumption Survey (CBECS; US DOE, 2003);
however, information on the number of rooms in each square footage class is not
available. As an alternate data source, information was available on the range of the
square footage of a typical school classroom (600 to 1,400 square feet) to generate a
uniform distribution bounded by these extremes (NCBG, 2008; US Army Corps of
Engineers, 2002). For restaurants and other buildings, staff assumed that the entire
building was one room; therefore, the CBECS (US DOE, 2003) provided data for this
building category to estimate square footage distributions (Table B.4-3).
Table B.4-3. Distributions used to calculate representative room size for
non-residential microenvironments.
Microenvironment
Office, Building size
(ft2)
Office, number of
people per m2.
Parameter
1a
16,632
4.0
Parameter
2b
8,035
1.5
Parameter
3C
4,612
1.5
Parameter
4d
69,530
8.5
Distribution
Type
Normal
Normal
http://www.epa.gov/iaq/base/pdfs/test_space_characteristics/tc-0.pdf
http://www.epa. gov/iaq/base/pdfs/test_space_characteristics/tc-l.pdf
B-47
-------
School (ft2)
Restaurant (ft2)
Other Buildings (ft2)
600
5,340
3,750
1,400
31
24
N/A
668
750
N/A
42,699
18,796
Uniform
Lognormal
Lognormal
Notes:
a Mean for normal, geometric mean for lognormal, lower limit for uniform distribution.
b Standard deviation for normal, geometric standard deviation for lognormal, upper limit for
uniform.
c Minimum value for normal and lognormal.
d Maximum value for normal and lognormal.
Additional specifications
Two additional specifications were required to calculate the room volumes and
surface areas: the ceiling heights and surface area of furniture within the rooms. Table
B.4-4 provides the data values and sources used to estimate each of these variables.
Table B.4-4. Ceiling heights and furniture surface area to floor ratios for
simulated indoor microenvironments.
Indoor
Microenvironment
Residence
Office
School
Restaurant
Other Buildings
Ceiling Height3
8ft
10ft
10ft
10ft
10ft
Furniture Surface Area
to Floor Ratio
2D
4C
4C
4C
4C
Notes:
3 Assumed by staff.
b Thatcher et al. (2002) and Singer et al. (2002).
c The surface area to volume ratio was assumed higher in the commercial
microenvironments than in residences. A value of 4 was selected since it
kept the range of total surface area to volume ratio within a typical range of 2
to 4 (Lawrence Berkeley National Laboratory., 2003).
B.4.1.3 Surface type probabilities
Following the calculation of the basic dimensions of the simulated room, staff
performed additional probabilistic sampling to specify the surface types present. In some
microenvironments, it is possible that only a single surface type be present (e.g., a public
access building likely contains only hard floors and no carpet). However, in other cases,
a typical building may have multiple surface types (e.g., a residence may have a mixture
B-48
-------
of both hard floor and carpet). Thus, in each microenvironment, staff estimated a
probability of occurrence for each surface type. If more than one surface type is possible
at the same time, then staff also approximated the fraction of each. Table B.4-5
summarizes both the probabilities and fractions assumed by staff for each
microenvironment.
B.4.1.4 Final SOi deposition distributions
Following the estimation of the room dimensions and surface types within each
simulated indoor microenvironment, an effective deposition rate was estimated for all
1,000 sample buildings. The geometric mean and geometric standard deviation were
calculated across all 1,000 samples and used to parameterize a lognormal distribution
(Table B.4-6). In applying these to the relative humidity conditions in the study areas,
staff assumed that the relative humidity is below 50% when the air conditioning or
heating unit is on. If the building has no air conditioner, the ambient summer humidity
was used (90 % in the morning, 50% in the afternoon). Staff also assumed that all non-
residential buildings had air-conditioning.
As far as mapping to the APEX microenvironments, residences, offices, and
restaurants are explicitely modeled microenvironments. The daycare microenvironment
used the school deposition distribution, while other indoor microenvironments (i.e.,
shopping or other) used the other building deposition distribution.
B-49
-------
Table B.4-5. Probability of occurrence and fractional quantity for surface types in indoor micronenvironments.
Indoor
Microenvironment
Residence
Office
School
Restaurant
Other Buildings
Floor
Carpet
P = 1
F =
N{0.52, 0.23}a
P = 1
F = 5/6 if hard
floor present0
P = 0C
P = 0.1d
P = 0.1d
Hard floor
P = 1
F =
1 - fraction
carpeted3
P = 0.34
F= 1/6 if hard
floor is present13
P = 1C
P = 0.9d
P = 0.9d
Ceiling
Wallboard
P = 1C
P = 0C
P = 0C
P = 0.55d
P = 0.19d
Ceiling Tile
P = 0C
P = 1C
P = 1C
P = 0.45d
P = 0.81d
Wall
Wallboard
P = 1
F=5/6
if wallpaper is
present0
P = 1
F is adjusted if
wallpaper and/or
wood paneling is
present0
P = 1C
P = 1
F is adjusted if
wallpaper and/or
wood paneling is
present0
P = 1
F is adjusted if
wallpaper and/or
wood paneling is
present0
Wallpaper
P = 0.225
F= 1/6 if
wallpaper is
presentd
P = 0.11
F= 1/6 if
wallpaper is
present13
P = 0C
P = 0.09
F = 1/2if
wallpaper is
presentd
P = 0.09
F = 1/2 if
wallpaper is
presentd
Wood Paneling
P = 0C
P = 0.13
F = 1/6 if wood
paneling is
present13
P = 0C
P = 0.25
F = 1/10 If wood
paneling is
presentd
P = 0.045
F=1/10 if wood
paneling is
presentd
Notes:
a US EPA, 2008b.
b BASE study, Table 4 (US EPA, 2008a); the fraction of 1/6 is based on professional judgment.
c Assumed by staff.
d Source Ranking Database (SRD, US EPA, 2004b). The fraction of buildings value in the database was used to specify a probability each surface type
occurs in the microenvironment. SRD names were matched to the APEX environments. Most categories in the SRD have the same fraction of building
values. To map to the necessary surface types: Carpet - Networx represented carpet; Ceiling tile represented ceiling tile; vinyl coated wallpaper
represented wallpaper; and Hardwood plywood paneling represented wood paneling. Fractions were assumed by staff. Then, probabilities in the
remaining surface types were calculated assuming either only one type could be present or multiple types could be present.
B-50
-------
Table B.4-6. Final parameter estimates of SO2 deposition distributions in several
indoor microenvironments modeled in APEX.
Microenv-
ironment
Residence
Office
School
Restaurant
Other
Buildings
Heating or Air Conditioning in Use
Geom.
Mean
(hr-1)
3.14
3.99
4.02
2.36
2.82
Geom.
Stand.
Dev.
(hr1)
1.11
1.04
1.02
1.28
1.21
Lower
Limit
(hr-1)
2.20
3.63
3.90
1.64
1.71
Upper
Limit
(hr-1)
5.34
4.37
4.21
4.17
4.12
Air Conditioning Not in Use
(Summertime Ambient Morning
Relative Humidity of 90%)
Geom.
Mean
(hr-1)
13.41
N/A
N/A
N/A
N/A
Geom
Stand.
Dev.
(hr1)
1.11
N/A
N/A
N/A
N/A
Lower
Limit
(hr-1)
10.31
N/A
N/A
N/A
N/A
Upper
Limit
(hr-1)
26.96
N/A
N/A
N/A
N/A
Notes:
N/A not applicable, assumed by staff to always have A/C in operation.
B-51
-------
B.4.2 APEX Exposure Output
Table B.4-7. APEX estimated SO2 exposures in Greene County (as is air quality
scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
309
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1821000
193
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
193
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7280
108
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-52
-------
Table B.4-8. APEX estimated SO2 exposures in Greene County (current standard
air quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
9598
1659
511
197
90
22
18
13
4
0
0
0
0
0
0
0
1821000
6393
1036
323
112
49
13
9
4
4
0
0
0
0
0
0
0
Number
of
Persons
21262
4322
982
323
139
67
18
13
13
4
0
0
0
0
0
0
0
7280
2609
569
188
72
40
9
4
4
4
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.20
0.04
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.36
0.08
0.03
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-53
-------
Table B.4-9. APEX estimated SO2 exposures in Greene County (99th
alternative standard scenario) while at moderate or greater exertion
%ile, 50 ppb
level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1821000
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7280
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-54
-------
Table B.4-10. APEX estimated SO2 exposures in Greene County (99th %ile, 100
ppb alternative standard scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
359
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1821000
229
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
229
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7280
139
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-55
-------
Table B.4-11. APEX estimated SO2 exposures in Greene County (99th %ile, 150
ppb alternative standard scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
1327
139
18
9
0
0
0
0
0
0
0
0
0
0
0
0
1821000
798
67
9
4
0
0
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
811
103
13
9
0
0
0
0
0
0
0
0
0
0
0
0
7280
466
49
4
4
0
0
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.04
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.06
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-56
-------
Table B.4-12. APEX estimated SO2 exposures in Greene County (99th %ile, 200
ppb alternative standard scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
2779
359
94
18
13
0
0
0
0
0
0
0
0
0
0
0
1821000
1757
229
54
9
4
0
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
1600
229
72
13
13
0
0
0
0
0
0
0
0
0
0
0
7280
955
139
45
4
4
0
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.07
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.13
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-57
-------
Table B.4-13. APEX estimated SO2 exposures in Greene County (99th %ile, 250
ppb alternative standard scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
4918
780
202
63
18
13
4
0
0
0
0
0
0
0
0
0
1821000
3201
457
117
40
9
4
4
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
2726
484
143
54
13
13
4
0
0
0
0
0
0
0
0
0
7280
1659
256
76
31
4
4
4
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.12
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.23
0.04
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-58
-------
Table B.4-14. APEX estimated SO2 exposures in Greene County (98th %ile, 200
ppb alternative standard scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Number
of
Person
Days
3218000
4138
632
161
45
18
13
0
0
0
0
0
0
0
0
0
0
1821000
2654
390
85
27
9
4
0
0
0
0
0
0
0
0
0
0
Number
of
Persons
21262
2304
386
117
40
13
13
0
0
0
0
0
0
0
0
0
0
7280
1390
220
58
22
4
4
0
0
0
0
0
0
0
0
0
0
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Fraction of
Total
Population
0.97
0.10
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.19
0.03
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-59
-------
Table B.4-15. APEX estimated SO2 exposures in St. Louis (as is air quality
scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
102436
24405
3866
789
229
69
23
8
8
0
0
0
0
0
0
0
0
41714
16938
2776
575
160
39
16
0
0
0
0
0
0
0
0
0
0
Number
of Person
Days
16677000
44100
4631
896
244
69
23
8
8
0
0
0
0
0
0
0
0
10560000
32800
3357
651
176
38
15
0
0
0
0
0
0
0
0
0
0
Fraction of
Total
Population
0.97
0.23
0.04
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.41
0.07
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-60
-------
Table B.4-16. APEX estimated SO2 exposures in St. Louis (current standard air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
93692
79422
63016
48211
36315
26363
19278
14181
10448
7853
5880
4431
3336
2631
1985
1556
41714
41607
40319
36287
30504
24386
18254
13539
9991
7547
5658
4237
3204
2376
1909
1426
1111
Number of
Person Days
16677000
2889400
793000
316400
153990
84540
49440
31700
20719
14242
10060
7229
5343
3972
3099
2253
1747
10560000
2158300
602800
239310
116260
63570
36830
23507
15304
10636
7420
5295
3901
2851
2231
1609
1240
Fraction of
Total
Population
0.97
0.89
0.75
0.60
0.46
0.34
0.25
0.18
0.13
0.10
0.07
0.06
0.04
0.03
0.02
0.02
0.01
1.00
1.00
0.97
0.87
0.73
0.58
0.44
0.32
0.24
0.18
0.14
0.10
0.08
0.06
0.05
0.03
0.03
B-61
-------
Table B.4-17. APEX estimated SO2 exposures in St. Louis (99th %ile, 50 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
14488
1595
298
69
16
8
0
0
0
0
0
0
0
0
0
0
41714
10229
1135
214
39
8
0
0
0
0
0
0
0
0
0
0
0
Number
of Person
Days
16677000
21379
1794
328
69
15
8
0
0
0
0
0
0
0
0
0
0
10560000
15835
1272
237
38
8
0
0
0
0
0
0
0
0
0
0
0
Fraction of
Total
Population
0.97
0.14
0.02
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.25
0.03
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-62
-------
Table B.4-18. APEX estimated SO2 exposures in St. Louis (99th %ile, 100 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
48725
14488
4654
1595
666
298
153
69
38
16
8
8
8
0
0
0
41714
30703
10229
3349
1135
491
214
99
39
31
8
0
0
0
0
0
0
Number
of Person
Days
16677000
158000
21379
5619
1794
742
328
152
69
38
15
8
8
8
0
0
0
10560000
119350
15835
4100
1272
551
237
99
38
31
8
0
0
0
0
0
0
Fraction of
Total
Population
0.97
0.46
0.14
0.04
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.74
0.25
0.08
0.03
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-63
-------
Table B.4-19. APEX estimated SO2 exposures in St. Louis (99th %ile 150 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
68830
33447
14488
6702
3212
1595
844
521
298
198
130
69
38
23
16
8
41714
38024
22721
10229
4843
2323
1135
621
376
214
138
76
39
31
16
8
0
Number
of Person
Days
16677000
429400
73000
21379
8403
3817
1794
958
582
328
198
130
69
38
23
15
8
10560000
325900
54890
15835
6177
2767
1272
705
422
237
137
76
38
31
15
8
0
Fraction of
Total
Population
0.97
0.65
0.32
0.14
0.06
0.03
0.02
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.91
0.54
0.25
0.12
0.06
0.03
0.01
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
B-64
-------
Table B.4-20. APEX estimated SO2 exposures in St. Louis (99th %ile, 200 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
79775
48725
27030
14488
8097
4654
2707
1595
1050
666
428
298
214
153
107
69
41714
40388
30703
18690
10229
5856
3349
1947
1135
773
491
314
214
145
99
61
39
Number
of Person
Days
16677000
813700
158000
51270
21379
10427
5619
3198
1794
1180
742
458
328
229
152
107
69
10560000
618700
119350
38210
15835
7718
4100
2292
1272
857
551
336
237
160
99
61
38
Fraction of
Total
Population
0.97
0.76
0.46
0.26
0.14
0.08
0.04
0.03
0.02
0.01
0.01
0.00
0.00
0.00
0.00
0.00
0.00
1.00
0.97
0.74
0.45
0.25
0.14
0.08
0.05
0.03
0.02
0.01
0.01
0.01
0.00
0.00
0.00
0.00
B-65
-------
Table B.4-21. APEX estimated SO2 exposures in St. Louis (99th %ile, 250 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
1 02436
85784
60235
39121
23681
14488
9180
5750
3696
2452
1595
1150
751
574
405
298
229
41714
41147
35351
25834
16477
10229
6686
4138
2637
1786
1135
849
536
422
298
214
160
Number
of Person
Days
16677000
1276000
278550
97390
42330
21379
12037
7061
4416
2843
1794
1287
858
643
435
328
244
10560000
967000
210680
73310
31530
15835
8975
5166
3173
2070
1272
941
613
475
321
237
176
Fraction of
Total
Population
0.97
0.81
0.57
0.37
0.22
0.14
0.09
0.05
0.04
0.02
0.02
0.01
0.01
0.01
0.00
0.00
0.00
1.00
0.99
0.85
0.62
0.39
0.25
0.16
0.10
0.06
0.04
0.03
0.02
0.01
0.01
0.01
0.01
0.00
B-66
-------
Table B.4-21. APEX estimated SO2 exposures in St. Louis (98th %ile 200 ppb air
quality scenario) while at moderate or greater exertion level.
5-minute
Exposure
Level
(ppb)
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Subpopulation
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMA,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
DMTS,ASTHMACHILD,MOD
Number
of
Persons
102436
84633
57867
36682
21576
12925
8014
5022
3174
2023
1387
913
666
474
314
229
198
41714
41070
34529
24576
15085
9168
5774
3648
2285
1464
1011
675
491
352
222
160
138
Number
of Person
Days
16677000
1159900
249490
85910
37060
18498
10304
6041
3772
2299
1539
1035
742
512
344
252
198
10560000
880000
188770
64600
27517
13677
7596
4446
2721
1655
1110
759
551
375
245
183
137
Fraction of
Total
Population
0.97
0.80
0.55
0.35
0.20
0.12
0.08
0.05
0.03
0.02
0.01
0.01
0.01
0.00
0.00
0.00
0.00
1.00
0.98
0.83
0.59
0.36
0.22
0.14
0.09
0.05
0.04
0.02
0.02
0.01
0.01
0.01
0.00
0.00
B-67
-------
B.5 REFERENCES
AHS. (2008). The American Housing Survey. U.S. Department of Housing and Urban
Development, http://www.census.gov/prod/2008pubs/hl50-07.pdf. Last accessed December
17, 2008.
Akland GG, Hartwell TD, Johnson TR, Whitmore RW. (1985). Measuring human exposure to
carbon monoxide in Washington, D. C. and Denver, Colorado during the winter of 1982-83.
Environ Sci Technol. 19:911-918.
Biller WF, Feagans TB, Johnson TR, Duggan GM, Paul RA, McCurdy T, Thomas HC. (1981).
A general model for estimating exposure associated with alternative NAAQS. Paper No. 81-
18.4 in Proceedings of the 74th Annual Meeting of the Air Pollution Control Association,
Philadelphia, PA.
Grontoft T, and Raychaudhuri MR. (2004). Compilation of tables of surface deposition
velocities for Os, NC>2 and SC>2 to a range of indoor surfaces. Atmos Environ. 38:533-544.
Hartwell TD, Clayton CA, Ritchie RM, Whitmore RW, Zelon HS, Jones SM, Whitehurst DA.
(1984). Study of Carbon Monoxide Exposure of Residents of Washington, DC and Denver,
Colorado. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
Research and Development, Environmental Monitoring Systems Laboratory. EPA-600/4-84-
031.
Johnson T, Capel J, Mozier J, McCoy M. (1996b). Estimation of Ozone Exposures
Experienced by Outdoor Children in Nine Urban Areas Using a Probabilistic Version of
NEM. Prepared for the Air Quality Management Division under Contract No. 68-DO-30094,
April.
Johnson T, Capel J, McCoy M, Mozier J. (1996c). Estimation of Ozone Exposures Experienced
by Outdoor Workers in Nine Urban Areas Using a Probabilistic Version of NEM. Prepared
for the Air Quality Management Division under Contract No. 68-DO-30094, April.
Johnson T, Capel J, McCoy M. (1996a). Estimation of Ozone Exposures Experienced by Urban
Residents Using a Probabilistic Version of NEM and 1990 Population Data. Prepared by IT
Air Quality Services for the Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina, September.
Johnson T, Capel J, Olaguer E, Wijnberg L. (1992). Estimation of Ozone Exposures
Experienced by Residents of ROMNET Domain Using a Probabilistic Version of NEM.
Prepared by IT Air Quality Services for the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
Johnson T. (1984). A Study of Personal Exposure to Carbon Monoxide in Denver, Colorado.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory. EPA-600/4-84-014.
Johnson T. (1989). Human Activity Patterns in Cincinnati, Ohio. Palo Alto, CA: Electric
Power Research Institute. EPRI EN-6204.
Johnson TR and Paul RA. (1983). The NAAQS Exposure Model (NEM) Applied to Carbon
Monoxide. EPA-450/5-83-003. Prepared for the U.S. Environmental Agency by PEDCo
Environmental Inc., Durham, N.C. under Contract No. 68-02-3390. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
B-68
-------
Klepeis NE, Tsang AM, Behar JV. (1996). Analysis of the National Human Activity Pattern
Survey (NHAPS) Respondents from a Standpoint of Exposure Assessment. Washington, DC:
U.S. Environmental Protection Agency, Office of Research and Development. EPA/600/R-
96/074.
Lawrence Berkeley National Laboratory (2003). Modeling shelter-in-place including sorption
on indoor surfaces. Available at:
http://repositories.cdhb.org/cgi/viewcontent.cgi?article=1450&context=lbnl.
McCurdy T, Glen G, Smith L, Lakkadi Y. (2000). The National Exposure Research
Laboratory's Consolidated Human Activity Database, JExp Anal Environ Epidemiol. 10:
566-578.
NCBG (2008). Design Capacity in the Chicago Public Schools. The Neighborhood Capital
Budget Group. Available at: http://www.ncbg.org/schools/design_capacity.htm .
Roddin MF, Ellis HT, Siddiqee WM. (1979). Background Data for Human Activity Patterns,
Vols. 1, 2. Draft Final Report. Prepared for Strategies and Air Standards Division, Office of
Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research
Triangle Park, N.C.
Singer, BC, AT Hodgson, and WW Nazaroff. (2002). Effect of sorption on exposures to
organic gases from environmental tobacco smoke. LBNL report number 49576. January
2002.
Spier CE, Little DE, Trim SC, Johnson TR, Linn WS, Hackney JD. (1992). Activity patterns in
elementary and high school students exposed to oxidant pollution. J Expo Anal Environ
Epidemiol. 2:277-293.
Thatcher, TL, Lai ACK, Moreno-Jackson R, Sextro RG, and Nazaroff WW. (2002). Effects of
room furnishings and air speed on particle deposition rates indoors. Atmos Environ.
36:1811-1819.
Tsang AM and Klepeis NE. (1996). Descriptive Statistics Tables from a Detailed Analysis of
the National Human Activity Pattern Survey (NHAPS) Data. U.S. Environmental Protection
Agency. EPA/600/R-96/148.
US Army Corps of Engineers. (2002). Unified Facilities Criteria (Ufc) Design: Child
Development Centers. Report no. UFC 4-740-14, (Table 2.6). Available at:
http ://www. wbdg. org/ccb/DOD/UFC/ufc_4_740_l 4.pdf.
US Census Bureau. (2007). Employment Status: 2000- Supplemental Tables. Available at:
http://www.census.gov/population/www/cen2000/phc-t28.html.
US DOE (2003). Commercial Building Energy Consumption Survey. U. S. Department of
Energy. Available at:
http://www.eia.doe.gov/emeu/cbecs/cbecs2003/public_use_2003/cbecs_pudata2003.html.
Last accessed December 17, 2008.
US DOT. (2007). Part 3-The Journey To Work files. Bureau of Transportation Statistics
(BTS). Available at: http://transtats.bts.gov/.
US EPA. (1999). Total Risk Integrated Methodology. Website:
http://www.epa.gov/ttnatw01/urban/trim/trimpg.html.
US EPA. (2002). Consolidated Human Activities Database (CHAD) Users Guide. Database
and documentation available at: http://www.epa.gov/chadnetl/.
US EPA. (2004a). AERMOD: Description of Model Formulation. Office of Air Quality
Planning and Standards. EPA-454/R-03-004. Available at:
http://www.epa.gov/scramOO l/7thconf/aermod/aermod_mfd.pdf
B-69
-------
US EPA. (2004b). Source Ranking Database. Available at:
http://www.epa.gov/opptintr/exposure/pubs/srddl.htm. Last accessed September 30, 2008.
US EPA. (2008a). Building Assessment Survey and Evaluation (BASE) Study.
http://www.epa.gov/iaq/base/summarized_data.html. Last accessed December 17, 2008.
US EPA. (2008b). The Approach Used for Estimating Changes in Children's IQ from Lead
Dust Generated During Renovation, Repair, and Painting in Residences and Child-Occupied
Facilities: Appendix A. Inputs for the Residential and Child-occupied Facilities (COFs) RRP
Models. Office of Pollution Prevention and Toxics. March 28, 2008.
US EPA. (2009a). Total Risk Integrated Methodology (TRIM) - Air Pollutants Exposure Model
Documentation (TRIM.Expo / APEX, Version 4.3) Volume I: User's Guide. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. January 2009. Available at:
http ://www. epa.gov/ttn/fera/human_apex.html.
US EPA. (2009b). Total Risk Integrated Methodology (TRIM) - Air Pollutants Exposure Model
Documentation (TRIM.Expo / APEX, Version 4.3) Volume II: Technical Support Document.
Office of Air Quality Planning and Standards, Research Triangle Park, NC. January 2009.
Available at: http://www.epa.gov/ttn/fera/human_apex.html.
Wiley JA, Robinson JP, Cheng Y-T, Piazza T, Stork L, Pladsen K. (1991b). Study of Children's
Activity Patterns: Final Report. California Air Resources Board, Sacramento, CA. ARB-R-
93/489.
Wiley JA, Robinson JP, Piazza T, Garrett K, Cirksena K, Cheng Y-T, Martin G. (1991a).
Activity Patterns of California Residents: Final Report. California Air Resources Board,
Sacramento, CA. ARB/R93/487. Available from: NTIS, Springfield, VA., PB94-108719.
Williams R, Suggs J, Creason J, Rodes C, Lawless P, Kwok R, Zweidinger R, Sheldon L.
(2000). The 1998 Baltimore particulate matter epidemiology-exposure study: Part 2.
Personal exposure associated with an elderly population. J Expo Anal Environ Epidemiol.
10(6):533-543.
B-70
-------
ATTACHMENT 1. TECHNICAL MEMORANDUM ON
METEOROLOGICAL DATA PREPARATION FOR AERMOD
FOR SO2 REA FOR GREENE COUNTY AND ST. LOUIS
MODELING DOMAINS, YEAR 2002
B-71
-------
Meteorological data preparation for AERMOD for SO2 REA for Greene County, MO and
St. Louis, MO
James Thurman and Roger Erode
U.S. EPA, OAQPS, AQAD
Air Quality Modeling Group
1. Introduction
National Weather Service (NWS) meteorological data are often used as the source of input
meteorological data for AERMOD (U. S. EPA, 2004a). For the SO2 Risk and Exposure
Assessment, two study areas were chosen: Greene County, Missouri, which includes the city of
Springfield, and St. Louis, Missouri. Tables 1 and 2 list the surface and upper air NWS stations
chosen for the two areas. Figure 1 shows the relationship between each surface station and its
paired upper air station.
For the St. Louis domain, two other stations were also considered: Spirit of St. Louis Airport
(SUS) and St. Louis Downtown Airport (CPS). SUS and CPS were used in the 1st draft REA
(U.S. EPA 2008a). The spatial relationship between the St. Louis area stations is shown in
Figure 2. Preliminary analysis of the three stations for the St. Louis domain revealed that CPS
and SUS contained significantly more calms and missing hours than STL. It was therefore
determined that STL would be more representative for the majority of emission sources for the
St. Louis modeling domain, and would be used for all of the St. Louis modeling. Given the
distances shown in Figures 2 and 3 between the stations, the choice was not unreasonable.
Table 1. Surface stations for the 862 study areas. Latitude and longitude are the best
approximation coordinates of the meteorological towers.
Area
Greene
County
St. Louis
Station
Springfield-
Branson
Regional
AP
Lambert-St.
Louis
International
AP
Identifier
SGF
STL
WMO
(WBAN)
724400 (13995)
724340 (13994)
Latitude
37.23528
38.7525
Longitude
-93.40028
-90.37361
Elevation
(m)
387
161
GMT
offset
6
6
Table 2. Upper air stations for the SC^ study areas.
Area
Greene
County
St. Louis
Station
Springfield-
Branson
Regional
AP
Lincoln-
Logan
County AP,
IL
Identifier
SGF
ILX
WMO
(WBAN)
724400 (13995)
724340 (4833)
Latitude
37.23
40.15
Longitude
-93.40
-89.33
Elevation
(m)
394
178
GMT
offset
6
6
B-72
-------
Figure 1. Location of surface stations (red dots) relative to upper air stations (crosses)
for Greene County and St. Louis, MO.
A potential concern related to the use of NWS meteorological data for dispersion modeling is the
often high incidence of calms and variable wind conditions reported for the Automated Surface
Observing Stations (ASOS) in use at most NWS stations since the mid-1990's. A variable wind
observation may include wind speeds up to 6 knots, but the wind direction is reported as missing.
The AERMOD model currently cannot simulate dispersion under these conditions. To reduce
the number of calms and missing winds in the surface data for each of the four stations, archived
one-minute winds for the ASOS stations were used to calculate hourly average wind speed and
directions, which were used to supplement the standard archive of winds reported for each
station in the Integrated Surface Hourly (ISH) database. Details regarding this procedure are
described below.
B-73
-------
Section 2 describes preparation of the surface and upper data from the ISH database and FSL
website including the preparation of data and calculation of hourly winds from one-minute
ASOS data, Section 3 describes AERSUKFACE processing for surface characteristics, and
Section 4 describes the AERMET processing. Section 5 provides a brief analysis of the
AERMET output for the stations. References are listed in Section 6.
Figure 2. Distance between STL and SUS and CPS.
B-74
-------
2. Surface and upper air data preparation
2.1 Surface data and hourly averaged wind calculations
One year of surface data for 2002 for each of the stations listed in Table 1 were downloaded
from the ISH archive at NCDC. Surface data from NWS locations often contain a large number
of calms and variable winds. This is due to the implementation of the ASOS program to replace
observer-based data beginning in the mid-1990's, and the adoption of the METAR standard for
reporting NWS observations in July 1996. Currently, the wind speed and direction used to
represent the hour in AERMOD is based on a single two-minute average, usually reported about
10 minutes before the hour. The METAR system reports winds of less than three knots as calm
(coded as 0 knots), and winds up to six knots will be reported as variable when the variation in
the 2-minute wind direction is more than 60 degrees. This variable wind is reported as a non-
zero wind speed with a missing wind direction. The number of calms and variable winds can
influence concentration calculations in AERMOD because concentrations are not calculated for
calms or variable wind hours. Significant numbers of calm and variable hours may compromise
the representativeness of NWS surface data for AERMOD applications. This is especially of
concern for applications involving low-level releases since the worst-case dispersion conditions
for such sources are associated with low wind speeds, and the hours being discarded as calm or
variable are biased toward this condition.
Recently, NCDC began archiving the two-minute average wind speeds for each minute of the
hour for most ASOS stations for public access. These values have not been subjected to the
METAR coding for calm and variable winds. Recent work in AQMG has focused on utilizing
these 1 minute winds to calculate hourly average winds to reduce the number of calms and
variable winds for a given station and year. For data input into AERMOD, one minute winds for
SGF and STL were used to calculate hourly average winds for 2002 (the 1-minute ASOS wind
data were not available for SUS or CPS for 2002). These hourly average winds are input to
AERMET and replace the winds reported for the hour from the ISH dataset. Following is the
methodology used to calculate the hourly average winds for this application:
One minute data files are monthly, so each month for 2002 was downloaded.
1. Each line of the data file was read and QA performed on the format of the line to check if
the line is valid data line. Currently, the one minute data files loosely follow a fixed
format, but there are numerous exceptions. The program performed several checks on
the line to ensure that wind direction and wind speed were in the correct general location.
If a minute was listed twice, the second line for that minute was assumed to be the correct
line. In the files, wind directions were recorded at the nearest whole degree and wind
speed to the nearest whole knot.
2. If the reported wind speed was less than 2 knots, the wind speed was reset to 1 knot. This
was done because anything less than 2 knots was considered below the instrument
threshold (if the anemometer is not a sonic anemometer, which was the case for SGF and
STL for 2002). This generally conforms to the meteorological monitoring guidance
recommendation of applying a wind speed of one half the threshold value to each wind
B-75
-------
sample below threshold when processing samples to obtain hourly averages. At the same
time, the x- and y-components of the wind direction were calculated using equations 1
and 2 below, which are the functions inside the summation of equations 6.2.17 and 6.2.18
of the meteorological guidance document (U.S. EPA, 2000). The components were only
calculated for minutes that did not require resetting.
vx=-sin0 (1)
vy=-cos0 (2)
where vx and vy are the x- and y-components of the one minute wind direction 9.
For all minutes that passed the QA check in step 1, the wind speeds were converted from
knots to m/s.
Before calculating hourly averages, the number of valid minutes (those with wind
directions) was checked for each hour. An hourly average would be calculated if the
there were at least two valid 2-minute averages reported for the hour. This could be even
minutes, odd minutes, or a mixture of non-overlapping even and odd minutes. Even
minutes were given priority over odd. If at least two valid minutes were found, then all
available (non-overlapping) minutes would be used to calculate hourly averages. The
most observations that could be used were 30 2-minute values (30 even or 30 odd).
For wind speed averages, all available non-overlapping minutes' speeds were used, even
those subject to resets as described in step 2. The hourly wind speed was an arithmetic
average of the wind speeds used.
For wind directions, the x- and y-components were summed according to equations
6.2.17 and 6.2.18 of the meteorological monitoring guidance (U.S. EPA, 2000),
summarized in equations 3 and 4 below with vxi and vyi calculated in equations 1 and 2.
The hourly wind direction was calculated based on a unit-vector approach, using equation
6.2.19 of the meteorological monitoring guidance (U.S.EPA, 2000), summarized in
equation 5. The one minute average wind directions do not use the flow correction as
shown in equation 6.2.19, since the calculated direction is the direction from which the
wind was blowing, not the direction in which it is blowing, as shown by the flow
correction in 6.2.19. Instead, the one minute program corrected for the direction from
which the wind was blowing.
6 = Arc\.mVx/v ] + CORK (5)
v / Ki-y
Where Vx and Vy are the hourly averaged x- and y-components of the wind, 9 is the
hourly averaged wind direction, N is the number of observations used for the hour, and
B-76
-------
= 180 for Vx > 0 and Vy > 0 or Vx < 0 and Vy > 0
CORK = 0 for Vx < 0 and Vy < 0
= 360forFx>OandFy<0
2.2 Upper air data
For AERMET processing, an upper air station must be paired with the surface station, as shown
in Table 2. Upper air data in the Forecast System Laboratory (FSL) format was downloaded
from the FSL, (currently named Global Systems Division) website, http://www.fsl.noaa.gov/.
The data period chosen was January 1, 2002 through December 31, 2002 for all times and all
levels. The selected wind speed units were chosen as tenths of a meter per second. Each station
was downloaded as a separate file.
B-77
-------
3. AERSURFACE
The AERSURFACE tool (U.S. EPA, 2008b) was used to determine surface characteristics
(albedo, Bowen ratio, and surface roughness) for input to AERMET. Surface characteristics
were calculated for the location of the ASOS meteorological towers. As noted in the
AERSURFACE User's Guide (U.S. EPA, 2008), AERSURFACE should be run for the location
of the actual meteorological tower to ensure accurate representation of the conditions around the
site. The approximate locations of the meteorological towers were determined using aerial
photos and the station history from NCDC. The coordinates used are listed in Table 1.
A draft version of AERSURFACE (08256) that utilizes 2001 NLCD was used to determine the
surface characteristics for this application since the 2001 land cover data will be more
representative of the meteorological data period than the 1992 NLCD data supported by the
current version of AERSURFACE available on EPA's SCRAM website. All stations were
considered "at an airport" for the low, medium, and high intensity developed categories. SGF
and STL did not have continuous snow cover as outlined in the 1st draft SO2 REA (U.S. EPA,
2008a). Monthly seasonal assignments did not follow the defaults as outlined in the
AERSURFACE User's Guide (U.S. EPA, 2008a) and the monthly seasonal assignments were
defined as shown in Table 3. Since the default seasonal assignments were not used, the surface
characteristics were output by month.
Table 3. Seasonal monthly assignments.
Station
SGF
STL
Winter (no snow)
December, January, February,
March
December, January, February
Spring
April, May
March, April, May
Summer
June, July, August
June, July, August
Autumn
September, October,
November
September, October,
November
Seasonal definitions
Winter (no snow)
Spring
Summer
Autumn
Late autumn after frost and harvest, or winter with no snow
Transitional spring with partial green coverage or short annuals
Midsummer with lush vegetation
Autumn with unharvested cropland
Moisture conditions (average, dry or wet) for Bowen ratio were based on annual precipitation
using the methodology outlined in the AERSURFACE User's Guide (U.S. EPA, 2008b): Years
in the top 30% of the 30-year precipitation distribution are considered wet. Those in the bottom
30% of the distribution are dry. Otherwise, a given year is considered average. For the two
surface stations, the 2007 local climatological database was used to look at 30 years (1978-2007)
annual precipitation. For SGF, 2002 was considered dry while STL was considered average.
The ranked 30 year distributions are shown in Table 4 with time series of the annual precipitation
in Figure 3.
B-78
-------
Table 4. Annual precipitation (inches) for Springfield and St. Louis. Years in green are
top 30% of distribution (wettest), years in brown are the bottom 30% of the distribution
(driest) and years in white are the middle 40%. 2002 is denoted in bold. 30 year
averages are denoted by yellow rows.
Springfield
Year
1990
1985
1993
1987
1994
1979
1998
1988
1992
1982
1984
2001
1983
1996
2007
1978
1981
2004
2003
1995
1999
1986
2006
1997
2002
1991
2000
2005
1989
1980
30-year
average
Precipitation
(inches)
63.19
56.50
55.78
55.49
49.02
48.94
48.47
48.46
48.04
47.67
45.78
45.29
45.05
44.86
44.27
43.95
43.72
43.23
42.61
41.86
41.53
40.19
38.87
38.48
37.82
37.59
35.36
35.32
31.50
27.36
40.21
St. Louis
Year
1982
1993
1984
1985
2003
1981
1990
1983
1996
1998
2004
1995
2002
1987
2005
1978
2000
2001
1986
1994
1999
1988
1992
1991
1997
2007
2006
1979
1989
1980
30-year
average
Precipitation
(inches)
54.97
54.76
51.65
50.73
46.06
45.52
45.09
44.80
43.67
43.62
42.27
41.68
40.95
38.38
37.85
37.71
37.37
35.29
34.88
34.70
34.06
33.93
33.49
33.48
31.23
30.57
29.93
29.48
28.60
27.48
39.14
B-79
-------
I
-------
a jj
Legend
Open water
Developed, low intensty
Oveloped. medium intensity
Cevfiloped. Ngh miensir,1
Barren land
D?cnjuous fores
I MRSd (Wist
--•.IIj.M .".I,
I ef*ssia
-------
Legend
|^^| t^jen water
I I Developed , open s.pac*
H Developed . lew ineensry
^^| Developed . met* urn rterorty
^H Developed , high nttnWy
•:-jT«it r s
K Ctecduws forest
^H Evergreen Forra
V -!•: ••- ;
] Grassland herbaceous
j Pistuffeihsy
Figure 5. Surface roughness sectors for STL with a) 2001 NLCD landuse and b) 2002
aerial photograph.
B-82
-------
4. AERMET
The meteorological data files for each station (upper air, ISH data, one minute data) were
processed in AERMET (U.S. EPA, 2004), which includes three "Stages" for processing of
meteorological data. Stage 1 was used to read in all the data files and perform initial QA. The
upper air data was processed via the UPPERAIR pathway. The ISH data was processed via the
SURFACE pathway, and the one minute hourly average winds were processed via the ONSITE
pathway. Hourly averaged winds were read into AERMET for the one minute hourly average
winds. For the hourly averaged one minute winds, the threshold was set to 0.01 m/s. The lowest
wind speeds for SGF and STL, including one minute data, was around 0.54 m/s.
In Stage 2, the upper air, ISH surface data, and hourly averaged winds were merged together for
each station. After Stage 2, Stage 3 was run to create the input files for AERMOD. When
hourly averaged winds were available, those winds would be used for the hour and all other data
would come from the ISH data (temperature, cloud cover, precipitation, etc.) If no hourly
averaged winds were available for the hour, all surface data came from the ISH data via the
SUBNWS keyword in the Stage 3 input file. As noted in Section 3, surface characteristics from
AERSURFACE are input into Stage 3. The resulting output from Stage 3 were the .SFC and
.PFL files input into AERMOD.
An AERMOD run, using a single source and receptor, was used to determine the number of
calms and missing hours for each station. Missing hours can be due to missing winds,
temperatures or soundings. Missing hours can also result from variable winds. The number of
calms and missing hours for each station are shown in Table 5. Also shown in Table 5 are the
number of calms and missing when using only the ISH winds for surface winds, no hourly
averaged one minute winds included. Note that including the hourly averaged winds
dramatically reduces the number of calms and missing hours.
Table 5. Number of calms and missing hours for each station. Totals reflect the use of
hourly averaged one minute winds.
Station
SGF
STL
With hourly
averaged winds
Calms
116
67
Missing
135
98
Without hourly
averaged winds
Calms
830
648
Missing
448
401
B-83
-------
5. Analysis
Wind roses for 2002 for the two stations are shown in Figure 6. For SGF, the wind was
predominantly from the south and south-southeast. For STL, winds were predominantly from
the south but strong components of the wind were from the westerly direction (northwest, west,
and southwest).
'WEST
& '
EAST
20%
~"-% 16%
~-, 12%
8%,
'SOUTH .
VWNO SPEED
(m/s)
n -11.1
• 8-8-11,1
• 57- 3,8
• 3.8 5.7
I I 31. 36
• 0.6- 2.1
Calms: 1 33*
Figure 6. 2002 wind roses after AERMET processing for a) SGF and b) STL.
For SGF and STL, 2002 was compared against 30-year climatology for precipitation and
temperature. Precipitation has been discussed in Section 3 (Table 4 and Figure 3). A
distribution of the annual mean temperatures from 1978 to 2007 is shown in Table 6 with time
series of mean temperatures shown in Figure 7. For Springfield, 2002 was drier than the 30-year
average (Table 4 and Figure 3) and about average for the mean temperature (Table 6 and Figure
7). For St. Louis, the precipitation was slightly above the 30-year average (Table 4 and Figure 3)
with the mean temperature about one degree above the 30-year average (Table 6 and Figure 7).
B-84
-------
Table 6. 30-year distribution of mean annual temperatures (Fahrenheit) for Springfield
and St. Louis. 2002 is denoted in bold.
Springfield
Year
2006
2007
1998
2005
1990
1999
1991
1987
1986
1980
1981
1994
1984
2004
2001
1982
1992
1995
2002
1983
2000
2003
1988
1985
1993
1997
1996
1989
1978
1979
30-year
average
Temperature
58.9
58.1
57.9
57.9
57.8
57.6
57.5
57.3
57.2
57.1
57.1
57.1
56.7
56.7
56.6
56.3
56.3
56.2
56.2
56.0
55.9
55.8
55.3
55.1
54.9
54.5
54.4
54.0
53.9
53.5
56.3
St. Louis
Year
1991
1990
1998
2006
2007
1987
1999
2005
2002
1994
2001
1986
2004
1992
1988
1995
2003
1980
1984
2000
1981
1983
1989
1993
1985
1997
1996
1982
1979
1978
30-year
average
Temperature
59.2
59.0
58.7
58.5
58.3
58.2
58.0
58.0
57.9
57.7
57.7
57.6
57.6
57.2
57.0
57.0
56.5
56.4
56.4
56.2
56.1
55.9
55.7
55.6
55.2
55.1
54.9
54.8
54.1
53.2
56.8
B-85
-------
SpnnglieW Temperature (F) Fw 197810 20S7
e !F}For H3?81o 200?
-4 . 11.
Figure 7. 30 year time series of mean annual temperatures (Fahrenheit) for a)
Springfield, and b) St. Louis. Annual averages are in red, 30-year averages in blue, and
2002 denoted by asterisk.
B-86
-------
6. References
U.S. EPA, 2000: Meteorological Monitoring Guidance for Regulatory Modeling Applications.
EPA-454/R-99-005. U.S. Environmental Protection Agency, Research Triangle Park, NC
27711.
U.S. EPA, 2004: User's Guide for the AERMOD Meteorological Preprocessor (AERMET).
EPA-454/B-03-002. U.S. Environmental Protection Agency, Research Triangle Park,
NC 27711.
U.S. EPA, 2008a: Risk and Exposure Assessment to Support the Review of the SCh Primary
National Ambient Air Quality Standard: First Draft. EPA-452/P-08-003. U.S.
Environmental Protection Agency, Research Triangle Park, NC 27711.
U.S. EPA, 2008b. AERSURFACE User's Guide. EPA-454/B-08-001. U.S. Environmental
Protection Agency, Research Triangle Park, NC.
http://www.epa.gov/scram001/7thconf/aermod/aersurface_userguide.pdf
B-87
-------
Appendix A. Surface characteristics.
Tables Al and A2 show the surface characteristics for Springfield and St. Louis for 2002 based
on 2001 landuse.
Table Al. Springfield monthly surface characteristics by sector.
Month
January
February
March
April
May
June
Sector
1
2
3
4
1
2
o
5
4
1
2
o
J
4
1
2
3
4
1
2
o
3
4
1
2
3
4
Albedo
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.18
0.18
0.18
0.18
Bowen
ratio
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
.2
.2
.2
.2
.2
.2
.2
.2
1.36
1.36
1.36
1.36
Surface
roughness
0.022
0.021
0.037
0.022
0.022
0.021
0.037
0.022
0.022
0.021
0.037
0.022
0.033
0.032
0.055
0.034
0.033
0.032
0.055
0.034
0.102
0.146
0.206
0.147
Month
July
August
September
October
November
December
Sector
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Albedo
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
Bowen
ratio
1.36
1.36
1.36
1.36
1.36
1.36
1.36
1.36
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
2.06
Surface
roughness
0.102
0.146
0.206
0.147
0.102
0.146
0.206
0.147
0.095
0.145
0.205
0.145
0.095
0.145
0.205
0.145
0.095
0.145
0.205
0.145
0.022
0.021
0.037
0.022
B-88
-------
Table A2. St. Louis monthly surface characteristics by sector.
Month
January
February
March
April
May
June
Sector
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Albedo
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.18
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.17
0.17
0.17
0.17
0.17
Bowen
ratio
.02
.02
.02
.02
.02
.02
.02
.02
.02
.02
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.76
0.81
0.81
0.81
0.81
0.81
Surface
roughness
0.036
0.077
0.059
0.036
0.041
0.036
0.077
0.059
0.036
0.041
0.043
0.079
0.063
0.041
0.047
0.043
0.079
0.063
0.041
0.047
0.043
0.079
0.063
0.041
0.047
0.048
0.081
0.065
0.046
0.051
Month
July
August
September
October
November
December
Sector
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Albedo
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.17
0.18
0.18
0.18
0.18
0.18
Bowen
ratio
0.81
0.81
0.81
0.81
0.81
0.81
0.81
0.81
0.81
0.81
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
1.02
Surface
roughness
0.048
0.081
0.065
0.046
0.051
0.048
0.081
0.065
0.046
0.051
0.043
0.079
0.063
0.041
0.047
0.043
0.079
0.063
0.041
0.047
0.043
0.079
0.063
0.041
0.047
0.036
0.077
0.059
0.036
0.041
B-89
-------
ATTACHMENT 2. TECHNICAL MEMORANDUM ON THE
ANALYSIS OF NHIS ASTHMA PREVALENCE DATA
B-90
-------
INTERNATIONAL
DRAFT MEMORANDUM
To: John Langstaff
From: Jonathan Cohen, Arlene Rosenbaum
Date: September 30, 2005
EPA 68D01052, Work Assignment 3-08. Analysis of NHIS Asthma Prevalence
Ke: Data
This memorandum describes our analysis of children's asthma prevalence data from the National
Health Interview Survey (NHIS) for 2003. Asthma prevalence rates for children aged 0 to 17
years were calculated for each age, gender, and region. The regions defined by NHIS are
"Midwest," "Northeast," "South," and "West." For this project, asthma prevalence was defined
as the probability of a Yes response to the question C ASHMEV: "Ever been told that... had
asthma?" among those that responded Yes or No to this question. The responses were weighted
to take into account the complex survey design of the NHIS survey. Standard errors and
confidence intervals for the prevalence were calculated using a logistic model, taking into
account the survey design. Prevalence curves showing the variation of asthma prevalence
against age for a given gender and region were plotted. A scatterplot smoothing technique using
the LOESS smoother was applied to smooth the prevalence curves and compute the standard
errors and confidence intervals for the smoothed prevalence estimates. Logistic analysis of the
prevalence curves shows statistically significant differences in prevalence by gender and by
region. Therefore we did not combine the prevalence rates for different genders or regions.
Logistic Models
NHIS survey data for 2003 were provided by EPA. One obvious approach to calculate
prevalence rates and their uncertainties for a given gender, region, and age is to calculate the
proportion of Yes responses among the Yes and No responses for that demographic group,
weighting each response by the survey weight. Although that approach was initially used, two
problems are that the distributions of the estimated prevalence rates are not well approximated by
normal distributions, and that the estimated confidence intervals based on the normal
approximation often extend outside the [0, 1] interval. A better approach is to use a logistic
transformation and fit a model of the form:
Prob (asthma) = exp(beta) / (1 + exp(beta)),
B-91
-------
where beta may depend on the explanatory variables for age, gender, or region. This is
equivalent to the model:
Beta = logit (prob (asthma)} = log { prob (asthma) / [1 - prob (asthma)] }.
The distribution of the estimated values of beta is more closely approximated by a normal
distribution than the distribution of the corresponding estimates of prob (asthma). By applying a
logit transformation to the confidence intervals for beta, the corresponding confidence intervals
for prob (asthma) will always be inside [0, 1]. Another advantage of the logistic modeling is that
it can be used to compare alternative statistical models, such as models where the prevalence
probability depends upon age, region, and gender, or on age and region but not gender.
A variety of logistic models for asthma prevalence were fit and compared, where the transformed
probability variable beta is a given function of age, gender, and region. SAS's
SURVEYLOGISTIC procedure was used to fit the logistic models, taking into account the NHIS
survey weights and survey design (stratification and clustering).
The following Table G-l lists the models fitted and their log-likelihood goodness-of-fit
measures. 16 models were fitted. The Strata column lists the four possible stratifications: no
stratification, by gender, by region, by region and gender. For example, "4. region, gender"
means that separate prevalence estimates were made for each combination of region and gender.
As another example, "2. gender" means that separate prevalence estimates were made for each
gender, so that for each gender, the prevalence is assumed to be the same for each region. The
prevalence estimates are independently calculated for each stratum.
Table G-l. Alternative logistic models for asthma prevalence.
Model
1
2
3
4
5
6
7
8
9
10
11
12
13
Description
1 . logit(prob) = linear in age
1 . logit(prob) = linear in age
1 . logit(prob) = linear in age
1 . logit(prob) = linear in age
2. logit(prob) = quadratic in age
2. logit(prob) = quadratic in age
2. logit(prob) = quadratic in age
2. logit(prob) = quadratic in age
3. logit(prob) = cubic in age
3. logit(prob) = cubic in age
3. logit(prob) = cubic in age
3. logit(prob) = cubic in age
4. logit(prob) = f(age)
Strata
1. none
2. gender
3. region
4. region, gender
1. none
2. gender
3. region
4. region, gender
1. none
2. gender
3. region
4. region, gender
1. none
- 2 Log Likelihood
54168194.62
53974657.17
54048602.57
53837594.97
53958021.20
53758240.99
53818198.13
53593569.84
53849072.76
53639181.24
53694710.66
53441122.98
53610093.48
DF
2
4
8
16
3
6
12
24
4
8
16
32
18
B-92
-------
Model
14
15
16
Description
4. logit(prob) = f(age)
4. logit(prob) = f(age)
4. logit(prob) = f(age)
Strata
2. gender
3. region
4. region, gender
- 2 Log Likelihood
53226610.02
53099749.33
52380000.19
DF
36
72
144
The Description column describes how beta depends upon the age:
Linear in age:
Quadratic in age:
Beta = a + P x age, where a and P vary with the strata.
Beta = a + P x age + y x age2 where a P and y vary with the
strata.
Cubic in age: Beta = a + P x age + y x age2 + 5 x age3 where a P, y, and 5 vary with the
strata.
f(age) Beta = arbitrary function of age, with different functions for different strata
The category f(age) is equivalent to making age one of the stratification variables, and is also
equivalent to making beta a polynomial of degree 16 in age (since the maximum age for children
is 17), with coefficients that may vary with the strata.
The fitted models are listed in order of complexity, where the simplest model (1) is an
unstratified linear model in age and the most complex model (16) has a prevalence that is an
arbitrary function of age, gender, and region. Model 16 is equivalent to calculating independent
prevalence estimates for each of the 144 combinations of age, gender, and region.
Table G-l also includes the -2 Log Likelihood, a goodness-of-fit measure, and the degrees of
freedom, DF, which is the total number of estimated parameters. Two models can be compared
using their -2 Log Likelihood values; lower values are preferred. If the first model is a special
case of the second model, then the approximate statistical significance of the first model is
estimated by comparing the difference in the -2 Log Likelihood values with a chi-squared
random variable with r degrees of freedom, where r is the difference in the DF. This is a
likelihood ratio test. For all pairs of models from Table G-l, all the differences are at least
70,000 and the likelihood ratios are all extremely statistically significant at levels well below 5
percent. Therefore the model 16 is clearly preferred and was used to model the prevalences.
The SURVEYLOGISTIC model predictions are tabulated in Table G-2 below and plotted in
Figures 1 and 3 below. Also shown in Table G-2 and in Figures 2 and 4 are results for smoothed
curves calculated using a LOESS scatterplot smoother, as discussed below.
The SURVEYLOGISTIC procedure produces estimates of the beta values and their 95 %
confidence intervals for each combination of age, region, and gender. Applying the inverse logit
transformation,
Prob (asthma) = exp( beta) / (1 + exp(beta)),
converted the beta values and 95 % confidence intervals into predictions and 95 % confidence
intervals for the prevalence, as shown in Table G-2 and Figures 1 and 3. The standard error for
the prevalence was estimated as
B-93
-------
Std Error (Prob (asthma)} = Std Error (beta) x exp(- beta) / (1 + exp(beta) )2,
which follows from the delta method (a first order Taylor series approximation).
Loess Smoother
The estimated prevalence curves shows that the prevalence is not a smooth function of age. The
linear, quadratic, and cubic functions of age modeled by SURVEYLOGISTIC were one strategy
for smoothing the curves, but they did not provide a good fit to the data. One reason for this
might be due to the attempt to fit a global regression curve to all the age groups, which means
that the predictions for age A are affected by data for very different ages. We instead chose to
use a local regression approach that separately fits a regression curve to each age A and its
neighboring ages, giving a regression weight of 1 to the age A, and lower weights to the
neighboring ages using a tri-weight function:
Weight = {1 - [ |age - A / q ]3}, where | age - A| <= q.
The parameter q defines the number of points in the neighborhood of the age a. Instead of calling
q the smoothing parameter, SAS defines the smoothing parameter as the proportion of points in
each neighborhood. We fitted a quadratic function of age to each age neighborhood, separately
for each gender and region combination. We fitted these local regression curves to the beta
values, the logits of the asthma prevalence estimates, and then converted them back to estimated
prevalence rates by applying the inverse logit function exp(beta) / (1 + exp(beta)). In addition to
the tri-weight variable, each beta value was assigned a weight of
1 / [std error (beta)]2, to account for their uncertainties.
The SAS LOESS procedure was applied to estimate smoothed curves for beta, the logit of the
prevalence, as a function of age, separately for each region and gender. We fitted curves using
the choices 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0 for the smoothing parameter in an effort to
determine the optimum choice based on various regression diagnostics.3,4
3 Two outlier cases were adjusted to avoid wild variations in the "smoothed" curves: For the West region, males,
age 0, there were 97 children surveyed that all gave No answers to the asthma question, leading to an estimated
value of -15.2029 for beta with a standard error of 0.14. For the Northeast region, females, age 0, there were 29
children surveyed that all gave No answers to the asthma question, leading to an estimated value of -15.2029 for
beta with a standard error of 0.19. In both cases the raw probability of asthma equals zero, so the corresponding
estimated beta would be negative infinity, but SAS's software gives -15.2029 instead. To reduce the impact of these
outlier cases, we replaced their estimated standard errors by 4, which is approximately four times the maximum
standard error for all other region, gender, and age combinations.
4 With only 18 points, a smoothing parameter of 0.2 cannot be used because the weight function assigns zero
weights to all ages except age A, and a quadratic model cannot be uniquely fitted to a single value. A smoothing
parameter of 0.3 also cannot be used because that choice assigns a neighborhood of 5 points only (0.3 x 18 = 5,
rounded down), of which the two outside ages have assigned weight zero, making the local quadratic model fit
exactly at every point except for the end points (ages 0, 1, 16 and 17). Usually one uses a smoothing parameter
below one so that not all the data are used for the local regression at a given x value.
B-94
-------
Quantities predicted in these smoothing parameter tests were the predicted value, standard error,
confidence interval lower bound and confidence interval upper bound for the betas, and the
corresponding values for the prevalence rates.
The polygonal curves joining values for different ages show the predicted values with vertical
lines indicating the confidence intervals in Figures 3 and 4 for smoothing parameters 0 (i.e., no
smoothing) and 0.5, respectively. Note that the confidence intervals are not symmetric about the
predicted values because of the inverse logit transformation.
Note that in our application of LOESS, we used weights of 1 / [std error (beta)]2, so that a2 = 1
for this application. The LOESS procedure estimates a2 from the weighted sum of squares. Since
in our application we assume a2 = 1, we multiplied the estimated standard errors by 1 /
estimated a, and adjusted the widths of the confidence intervals by the same factor.
Additionally, because the true value of a equals 1, the best choices of smoothing parameter
should give residual standard errors close to one. Using this criterion the best choice varies with
gender and region between smoothing parameters 0.4 (3 cases), 0.5 (2 cases), 0.6 (1 case), and
0.7 (1 case).
As a further regression diagnostic the residual errors from the LOESS model were divided by
std error (beta) to make their variances approximately constant. These approximately studentized
residuals, 'student,' should be approximately normally distributed with a mean of zero and a
variance of a2 = 1. To test this assumption, normal probability plots of the residuals were
created for each smoothing parameter, combining all the studentized residuals across genders,
regions, and ages. The plots for smoothing parameters seem to be equally straight for each
smoothing parameter.
The final regression diagnostic is a plot of the studentized residuals against the smoothed beta
values. Ideally there should be no obvious pattern and an average studentized residual close to
zero. The plots indeed showed no unusual patterns, and the results for smoothing parameters 0.5
and 0.6 seem to showed a fitted LOESS close to the studentized residual equals zero line.
The regression diagnostics suggested the choice of smoothing parameter as 0.4 or 0.5. Normal
probability plots did not suggest any preferred choices. The plots of residuals against smoothed
predictions suggest the choices of 0.5 or 0.6. We therefore chose the final value of 0.5. These
predictions, standard errors, and confidence intervals are presented in tabular form below as
Table G-2.
B-95
-------
Figure 1. Raw asthma prevalence rates by age and gender tor each region
region = Mi dwes t
gender
Figure 1. Raw asthma prevalence rates by age and gender tor each region
region=Northeast
1 I '
10
1 I '
11
1 I '
13
14
1 I '
15
1 I '
16
17
Male
B-96
-------
Figure 1. Raw asthma prevalence rates by age and gender for each region
regi on = South
1 I '
10
11
13 14
17
gender
Figure 1. Raw asthma prevalence rates by age and gender for each region
regi on=We st
Male
B-97
-------
Figure 2. Smoothed asthma prevalence rates by age for each region and gender
region = Mi dwes t
1 I ' ' ' I '
14 15
gender
Fema1e
Figure 2. Smoothed asthma prevalence rates by age for each region and gender
region=Northeast
Fema1e
Male
B-98
-------
Figure 2. Smoothed asthma prevalence rates by age for each region and gender
regi on = South
gender
Fema1e
Figure 2. Smoothed asthma prevalence rates by age for each region and gender
regi on=We st
predprob
0.34-
0.32-
0 . 30 :
0.28-
0.26-
0.24-
0.22
0.20-
0.18-
0.16-
0.14-
0.12-
0.10-
Fema1e
Male
B-99
-------
Figure 3. Raw asthma prevalence rates and confidence intervals
region = Mi dwes t
0 1
1 I ' ' ' ' I
1 I '
10
11
13 14
17
gender
age
Fema1e
Figure 3. Raw asthma prevalence rates and confidence intervals
region=Northeast
1 I '
10
1 I '
11
1 I '
13
14
1 I '
15
1 I '
16
17
Male
B-100
-------
Figure 3. Raw asthma prevalence rates and confidence intervals
regi on = South
gender
age
Fema1e
Figure 3. Raw asthma prevalence rates and confidence intervals
regi on=We st
i I i i i i I i i i i I i
345
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
10 11 12 13 14 15 16
17
Male
B-101
-------
Rgure 4. Smoothed asthma prevalence rates and confidence intervals
region = Mi dwes t
1 I ' ' ' ' I
1 I '
10
11
13 14
17
gender
age
Fema1e
Rgure 4. Smoothed asthma prevalence rates and confidence intervals
region=Northeast
0 00-
1 I '
10
1 I '
11
1 I '
13
14
1 I '
15
1 I '
16
17
Male
B-102
-------
Rgure 4. Smoothed asthma prevalence rates and confidence intervals
regi on = South
1 I ' ' ' ' I ' ' ' ' I '
345
1 I '
10
11
13 14
17
gender
age
Fema1e
Rgure 4. Smoothed asthma prevalence rates and confidence intervals
regi on=We st
pr e v
0.32-
0.30-
0.28-
0.26-
0.24-
0.22-
0 . 20 :
0.18-
0.16-
0.14-
0.12-
0.10-
0.08-
0.06-
0 . 04 :
0.02-
0 00-
1 I '
10
1 I '
11
1 I '
13
14
1 I '
15
1 I '
16
17
Male
B-103
-------
Table G-2. Raw and smoothed prevalence rates, with confidence intervals, by region,
gender, and age.
Obs
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Region
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Gender
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Age
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
12
12
13
Smoothed
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.04161
0.06956
0.10790
0.07078
0.05469
0.07324
0.06094
0.07542
0.09049
0.08100
0.08463
0.09540
0.14869
0.09210
0.04757
0.09032
0.10444
0.08612
0.09836
0.11040
0.10916
0.16190
0.27341
0.19597
0.10055
0.21214
0.22388
Std
Error
0.02965
0.03574
0.04254
0.01995
0.02578
0.01778
0.03474
0.01944
0.03407
0.02163
0.03917
0.02613
0.08250
0.02854
0.02927
0.02563
0.03638
0.02181
0.04283
0.02709
0.04859
0.03486
0.06817
0.03920
0.04780
0.03957
0.05905
95 % Conf
Interval -
Lower
Bound
0.01001
0.02143
0.04840
0.03736
0.02131
0.04228
0.01936
0.04205
0.04233
0.04417
0.03317
0.05106
0.04643
0.04534
0.01389
0.04728
0.05160
0.04842
0.04062
0.06298
0.04400
0.09838
0.16112
0.12296
0.03816
0.13724
0.12907
95 % Conf
Interval -
Upper
Bound
0.15717
0.20330
0.22336
0.13008
0.13325
0.12395
0.17579
0.13163
0.18298
0.14393
0.19942
0.17131
0.38520
0.17808
0.15051
0.16571
0.19997
0.14857
0.21943
0.18643
0.24600
0.25484
0.42437
0.29763
0.23952
0.31309
0.35959
B-104
-------
Obs
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Region
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Gender
Female
Female
Female
Female
Female
Female
Female
Female
Female
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Age
13
14
14
15
15
16
16
17
17
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.16966
0.10511
0.14020
0.12026
0.13341
0.13299
0.14040
0.17497
0.16478
0.06419
0.03134
0.02824
0.06250
0.05102
0.10780
0.18650
0.15821
0.24649
0.21572
0.11609
0.17822
0.14158
0.12788
0.09726
0.12145
0.16718
0.12757
0.13406
0.14718
0.13986
Std
Error
0.03371
0.04233
0.02603
0.03805
0.02266
0.03933
0.02235
0.04786
0.04037
0.03612
0.01537
0.01694
0.01751
0.02343
0.02078
0.04864
0.02705
0.05823
0.03661
0.04818
0.03525
0.05280
0.02799
0.03614
0.02642
0.05814
0.02700
0.04783
0.02976
0.04422
95 % Conf
Interval -
Lower
Bound
0.10716
0.04637
0.09164
0.06327
0.09056
0.07288
0.09764
0.09970
0.09320
0.02068
0.01042
0.00859
0.03321
0.02040
0.06960
0.10898
0.10696
0.15035
0.14543
0.04973
0.11280
0.06576
0.07751
0.04588
0.07391
0.08134
0.07864
0.06458
0.09254
0.07331
95 % Conf
Interval -
Upper
Bound
0.25807
0.22104
0.20857
0.21670
0.19226
0.23037
0.19777
0.28884
0.27468
0.18227
0.09046
0.08879
0.11457
0.12189
0.16328
0.30057
0.22775
0.37686
0.30774
0.24793
0.27003
0.27873
0.20375
0.19448
0.19317
0.31276
0.20031
0.25769
0.22603
0.25050
B-105
-------
Obs
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
Region
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Midwest
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Gender
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Age
10
11
11
12
12
13
13
14
14
15
15
16
16
17
17
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.17728
0.23907
0.18961
0.13660
0.19487
0.18501
0.16939
0.16673
0.16795
0.14583
0.17953
0.24965
0.20116
0.21152
0.23741
0.00000
0.06807
0.12262
0.07219
0.07217
0.07522
0.08550
0.07709
0.08704
0.08171
0.07597
0.11603
0.19149
0.16106
0.22034
Std
Error
0.02996
0.05031
0.03044
0.04784
0.03078
0.04498
0.02841
0.05094
0.02631
0.04241
0.02561
0.06037
0.03048
0.06481
0.05816
0.00000
0.06565
0.07443
0.03765
0.03707
0.02212
0.03991
0.02021
0.03804
0.02252
0.03754
0.03012
0.06960
0.03737
0.07076
95 % Conf
Interval -
Lower
Bound
0.12020
0.15449
0.13100
0.06668
0.13541
0.11230
0.11528
0.08886
0.11734
0.08054
0.12951
0.15033
0.14187
0.11131
0.13243
0.00000
0.00670
0.03476
0.02088
0.02561
0.03764
0.03324
0.04162
0.03596
0.04269
0.02801
0.06258
0.08937
0.09219
0.11195
95 % Conf
Interval -
Upper
Bound
0.25366
0.35075
0.26639
0.25946
0.27221
0.28945
0.24195
0.29104
0.23459
0.24967
0.24347
0.38489
0.27721
0.36490
0.38835
0.00000
0.44174
0.35164
0.22109
0.18713
0.14468
0.20269
0.13840
0.19592
0.15080
0.18998
0.20515
0.36372
0.26629
0.38783
B-106
-------
Obs
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
Region
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Gender
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Male
Male
Male
Male
Male
Male
Male
Male
Male
Age
7
8
8
9
9
10
10
11
11
12
12
13
13
14
14
15
15
16
16
17
17
0
0
1
1
2
2
3
3
4
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.18503
0.11002
0.17054
0.17541
0.14457
0.12980
0.13487
0.15128
0.14072
0.11890
0.16615
0.22638
0.17374
0.15807
0.15137
0.07460
0.14564
0.13603
0.14601
0.19074
0.15662
0.03904
0.04768
0.05533
0.04564
0.05525
0.05161
0.03842
0.06766
0.07436
Std
Error
0.04087
0.05128
0.04039
0.07488
0.03538
0.04964
0.03098
0.05287
0.03068
0.04426
0.03375
0.06285
0.03402
0.05513
0.02946
0.03409
0.02761
0.05328
0.03095
0.07382
0.05374
0.03829
0.03299
0.03425
0.01831
0.03119
0.01505
0.02923
0.01784
0.02906
95 % Conf
Interval -
Lower
Bound
0.10844
0.04241
0.09628
0.07159
0.08042
0.05930
0.07799
0.07366
0.08367
0.05568
0.10211
0.12650
0.10861
0.07694
0.09519
0.02971
0.09279
0.06081
0.08805
0.08451
0.06784
0.00547
0.00991
0.01596
0.01850
0.01781
0.02680
0.00840
0.03734
0.03393
95 % Conf
Interval -
Upper
Bound
0.29764
0.25654
0.28407
0.36981
0.24618
0.26087
0.22319
0.28547
0.22704
0.23597
0.25877
0.37158
0.26626
0.29719
0.23220
0.17506
0.22127
0.27686
0.23241
0.37568
0.32151
0.23095
0.20023
0.17461
0.10821
0.15872
0.09709
0.15853
0.11955
0.15522
B-107
-------
Obs
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
Region
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
Northeast
South
South
South
Gender
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Female
Female
Female
Age
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
12
12
13
13
14
14
15
15
16
16
17
17
0
0
1
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.09964
0.17601
0.14854
0.23271
0.20731
0.13074
0.22820
0.33970
0.22240
0.13761
0.21238
0.21785
0.17652
0.11448
0.16617
0.17736
0.18279
0.19837
0.17078
0.16201
0.17033
0.11894
0.18246
0.24306
0.20406
0.22559
0.24185
0.02459
0.03407
0.08869
Std
Error
0.02330
0.04519
0.02948
0.09319
0.04235
0.05195
0.04524
0.08456
0.04298
0.05024
0.04071
0.06659
0.03731
0.05849
0.03516
0.05489
0.03589
0.05450
0.03078
0.04973
0.02889
0.04584
0.02858
0.05798
0.03216
0.06980
0.06066
0.01116
0.01282
0.03373
95 % Conf
Interval -
Lower
Bound
0.05859
0.10393
0.09428
0.09832
0.12875
0.05785
0.14338
0.19726
0.14157
0.06507
0.13589
0.11464
0.10824
0.04005
0.10200
0.09349
0.11611
0.11222
0.11288
0.08618
0.11547
0.05417
0.12740
0.14759
0.14187
0.11748
0.13291
0.01002
0.01465
0.04118
95 % Conf
Interval -
Upper
Bound
0.16441
0.28234
0.22623
0.45756
0.31640
0.26922
0.34311
0.51855
0.33157
0.26785
0.31617
0.37465
0.27460
0.28601
0.25907
0.31067
0.27581
0.32635
0.25000
0.28386
0.24408
0.24139
0.25438
0.37326
0.28447
0.38930
0.39898
0.05906
0.07723
0.18067
B-108
-------
Obs
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
Region
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
Gender
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Age
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
12
12
13
13
14
14
15
15
16
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.05182
0.05097
0.07110
0.08717
0.08759
0.11010
0.09897
0.09409
0.11870
0.15318
0.12150
0.09608
0.11192
0.09955
0.09287
0.07477
0.09117
0.10602
0.10821
0.14411
0.13237
0.12646
0.12346
0.11376
0.09653
0.02915
0.09469
0.11985
0.09988
0.14183
Std
Error
0.01167
0.02373
0.01386
0.03240
0.01718
0.03209
0.01914
0.02943
0.02157
0.04317
0.02282
0.03538
0.02171
0.03288
0.01897
0.02719
0.01786
0.03214
0.02026
0.04267
0.02251
0.02981
0.02004
0.03270
0.01717
0.01339
0.01619
0.03357
0.01586
0.03685
95 % Conf
Interval -
Lower
Bound
0.03127
0.02012
0.04584
0.04122
0.05624
0.06113
0.06387
0.05015
0.07855
0.08611
0.07925
0.04565
0.07204
0.05111
0.05850
0.03606
0.05855
0.05750
0.07077
0.07875
0.08989
0.07860
0.08543
0.06365
0.06458
0.01174
0.06436
0.06801
0.06978
0.08366
95 % Conf
Interval -
Upper
Bound
0.08472
0.12319
0.10869
0.17500
0.13394
0.19035
0.15025
0.16968
0.17548
0.25777
0.18182
0.19105
0.16985
0.18493
0.14436
0.14864
0.13929
0.18732
0.16201
0.24907
0.19071
0.19723
0.17519
0.19510
0.14190
0.07054
0.13721
0.20259
0.14099
0.23028
B-109
-------
Obs
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
Region
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
South
Gender
Female
Female
Female
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Age
16
17
17
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
12
12
13
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.11501
0.13141
0.14466
0.01164
0.04132
0.10465
0.06981
0.11644
0.10189
0.10794
0.12852
0.08480
0.14393
0.22243
0.16450
0.13908
0.16386
0.10695
0.13329
0.13660
0.13818
0.15978
0.16839
0.21482
0.17848
0.15078
0.16247
0.13727
0.14480
0.14136
Std
Error
0.01620
0.03007
0.02946
0.00852
0.01867
0.03216
0.01623
0.03486
0.01672
0.03253
0.02139
0.02973
0.02379
0.04227
0.02373
0.03392
0.02460
0.04272
0.02322
0.03841
0.02276
0.03742
0.02450
0.04702
0.02453
0.03440
0.02224
0.03260
0.01976
0.03119
95 % Conf
Interval -
Lower
Bound
0.08365
0.08280
0.09067
0.00275
0.01487
0.05629
0.04125
0.06353
0.07024
0.05874
0.08793
0.04190
0.09861
0.15052
0.11821
0.08485
0.11613
0.04747
0.08951
0.07712
0.09484
0.09920
0.12062
0.13676
0.13021
0.09492
0.11881
0.08489
0.10610
0.09049
95 % Conf
Interval -
Upper
Bound
0.15612
0.20226
0.22291
0.04790
0.10956
0.18635
0.11576
0.20382
0.14557
0.19005
0.18405
0.16410
0.20534
0.31592
0.22430
0.21964
0.22617
0.22347
0.19392
0.23049
0.19702
0.24720
0.23012
0.32086
0.23972
0.23112
0.21820
0.21438
0.19453
0.21409
B-110
-------
Obs
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
Region
South
South
South
South
South
South
South
South
South
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
Gender
Male
Male
Male
Male
Male
Male
Male
Male
Male
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Age
13
14
14
15
15
16
16
17
17
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.14318
0.16110
0.15339
0.16172
0.15088
0.15836
0.14038
0.11156
0.12247
0.00983
0.01318
0.02367
0.03105
0.08097
0.05440
0.07528
0.07444
0.09263
0.07696
0.01976
0.07737
0.15792
0.07298
0.06955
0.08146
0.07753
0.09062
0.13440
0.10215
0.06573
Std
Error
0.01928
0.03444
0.01875
0.03519
0.01746
0.03879
0.01773
0.02737
0.02596
0.00990
0.00987
0.01862
0.01312
0.03759
0.01482
0.03851
0.01842
0.03196
0.02064
0.01347
0.02123
0.07301
0.01985
0.02567
0.01987
0.02825
0.01994
0.04481
0.02347
0.03719
95 % Conf
Interval -
Lower
Bound
0.10537
0.10438
0.11612
0.10394
0.11598
0.09614
0.10533
0.06810
0.07537
0.00135
0.00248
0.00497
0.01204
0.03170
0.02948
0.02679
0.04257
0.04621
0.04194
0.00513
0.04157
0.06009
0.03947
0.03321
0.04691
0.03731
0.05507
0.06802
0.06061
0.02102
95 % Conf
Interval -
Upper
Bound
0.19165
0.24037
0.19992
0.24291
0.19398
0.24974
0.18467
0.17746
0.19286
0.06802
0.06700
0.10522
0.07769
0.19166
0.09825
0.19404
0.12701
0.17703
0.13701
0.07302
0.13949
0.35487
0.13107
0.13989
0.13776
0.15417
0.14558
0.24832
0.16709
0.18736
B-lll
-------
Obs
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
Region
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
Gender
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Female
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Age
10
11
11
12
12
13
13
14
14
15
15
16
16
17
17
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Prevalence
0.12152
0.15354
0.12719
0.10120
0.13054
0.14759
0.11968
0.08748
0.11063
0.10099
0.11236
0.12538
0.12224
0.14672
0.14371
0.00000
0.03075
0.05457
0.04584
0.07833
0.06254
0.05897
0.07844
0.07267
0.09122
0.19732
0.11262
0.13335
0.12119
0.08881
Std
Error
0.02660
0.04584
0.02688
0.03594
0.02498
0.04125
0.02369
0.03284
0.02132
0.03841
0.02051
0.04343
0.02210
0.04582
0.03992
0.00000
0.02534
0.02662
0.01889
0.02789
0.01442
0.02530
0.01913
0.03354
0.02482
0.10033
0.02937
0.04859
0.02916
0.03493
95 % Conf
Interval -
Lower
Bound
0.07376
0.08329
0.07852
0.04934
0.08440
0.08346
0.07629
0.04105
0.07145
0.04674
0.07428
0.06188
0.08108
0.07743
0.07558
0.00000
0.00437
0.02056
0.01729
0.03833
0.03627
0.02500
0.04398
0.02870
0.04765
0.06632
0.06021
0.06322
0.06799
0.04015
95 % Conf
Interval -
Upper
Bound
0.19374
0.26584
0.19950
0.19631
0.19650
0.24769
0.18284
0.17675
0.16744
0.20471
0.16645
0.23755
0.18021
0.26052
0.25621
0.00000
0.18642
0.13695
0.11595
0.15342
0.10573
0.13281
0.13607
0.17208
0.16763
0.45969
0.20092
0.25970
0.20680
0.18508
B-112
-------
Obs
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
Region
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
West
Gender
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Male
Age
7
8
8
9
9
10
10
11
11
12
12
13
13
14
14
15
15
16
16
17
17
Smoothed
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
Prevalence
0.12691
0.15183
0.13161
0.17199
0.15079
0.12897
0.16356
0.19469
0.16965
0.13214
0.17494
0.19947
0.16217
0.10759
0.16487
0.18459
0.17018
0.19757
0.17888
0.18078
0.19218
Std
Error
0.02806
0.05484
0.02705
0.05164
0.02837
0.03747
0.02584
0.04002
0.02623
0.04542
0.02738
0.04814
0.02773
0.03838
0.02644
0.05348
0.02480
0.04862
0.02540
0.04735
0.04291
95 % Conf
Interval -
Lower
Bound
0.07464
0.07210
0.08037
0.09260
0.09590
0.07151
0.11192
0.12785
0.11699
0.06547
0.12002
0.12127
0.10747
0.05220
0.11214
0.10138
0.11996
0.11892
0.12718
0.10548
0.11118
95 % Conf
Interval -
Upper
Bound
0.20758
0.29200
0.20811
0.29715
0.22915
0.22159
0.23279
0.28505
0.23956
0.24865
0.24792
0.31029
0.23732
0.20880
0.23582
0.31235
0.23578
0.30993
0.24569
0.29227
0.31153
B-113
-------
ATTACHMENT 3. TECHNICAL MEMORANDUM ON
ESTIMATING PHYSIOLOGICAL PARAMETERS FOR THE
EXPOSURE MODEL
B-114
-------
TECHNICAL MEMORANDUM
TO: Tom McCurdy, WA-COR, NERL WA 10
FROM: Kristin Isaacs and Luther Smith, Alion Science and Technology
DATE: December 20, 2005
SUBJECT: New Values for Physiological Parameters for the Exposure Model
Input File Physiology.txt.
Table of Contents
List of Figures 116
1. Introduction 117
2. Evaluation of the Current Physiology File Data 117
2.1 Normalized Maximal Oxygen Uptake (nvo2max) 117
2.2 Body Mass 118
2.3 Resting Metabolic Rate 118
2.4 Hemoglobin Content and Blood Volume Factor 118
2.5 Summary of Findings 118
3. Derivation of New Distributions for Body Mass 119
3.1 The NHANES Body Mass Dataset 119
3.2 Calculation of the New Sampling Weights for the Combined NHANES Dataset.
120
3.3 Fitting the Body Mass Data 120
4. Derivation of New Distributions for Normalized Vo2max 126
4.1 The Nvo2max Data 126
4.2 Determining the NVo2max Distributions 130
5. Derivation of New Distributions for Hemoglobin Content (Hemoglobin Density)
138
6. Blood Volume as a Function of Height and Weight 141
References 142
Appendix A. SAS Code for Estimating the Body Mass Distributions 156
Appendix B. SAS Code for Estimating the Normalized Vo2Max Distributions 156
Appendix C. SAS Code for Estimating the Hemoglobin Content Data 157
Appendix D. The New Physiology.txt file 159
Appendix E. All Derived Physiological Parameters 170
B-115
-------
LIST OF FIGURES
Figure 1. Geometric Means for the Best-fit Lognormal Distributions for Body Mass as a
Function of Age, Derived from NHANES 1999-2004 Study Data 122
Figure 2. Geometric Standard Deviations for the Best-fit Lognormal Distributions for
Body Mass as a Function of Age, Derived from NHANES 1999-2004 Study Data 123
Figure 3. Minimums (1st Percentile) for Body Mass as a Function of Age, Derived from
NHANES 1999-2004 Study Data 124
Figure 4. Maximums (99th Percentile) for Body Mass as a Function of Age, Derived
from NHANES 1999-2004 Study Data 125
Figure 5. Individual Nvo2max Measurements for Males and Females, Derived from
Literature Studies and Experimental Measurements 127
Figure 6. Grouped Mean Nvo2max Measurements for Males and Females, Derived from
Literature Studies 128
Figure 7. Nvo2max Standard Deviations for Males and Females, Derived from Literature
Studies 129
Figure 8. Combined Nvo2max Group Means for Males and Females 132
Figure 9. Combined Nvo2max Group Standard Deviations 133
Figure 10. Nvo2max Normal Distribution Fits: Raw Fit Means and Smoothed Fits.... 134
Figure 11. Nvo2max Normal Distribution Fits: Raw Fit Standard Deviations and
Smoothed Fits 135
Figure 12. Nvo2max Minimums. 1st Percentile of the Best-fit Normal Distribution... 136
Figure 13. Nvo2max Maximums. 99th Percentile of the Best-fit Normal Distribution. 137
Figure 14. Mean Values of Hemoglobin Content as Derived from the 1999-2002
NHANES Dataset, with Comparison to CurrentPhysiology.txt Values 139
Figure 15. Values of Hemoglobin Content Standard Deviation as Derived from the 1999-
2002 NHANES Dataset, with Comparison to Current Physiology.txt Values 140
B-116
-------
1. INTRODUCTION
The purpose of this memo is to present an updated version of the physiological
parameters input file (Physiology.txt) for the APEX model. Portions of this file are also
used as input for SHEDS-PM and SHEDS-AirToxics.
The physiology file contains age- and gender-based information for several physiological
parameters used in human exposure modeling. This information includes distributional
shapes and parameters for all age and gender cohorts from age 0 to 100 years for
normalized maximal oxygen uptake (nvo2max), body mass, resting metabolic rate
(RMR), and blood hemoglobin content. In addition, a parameter called blood volume
factor (BVF), which is a cohort-dependent parameter in the equation for blood volume as
a function of body mass, is present in the file as well.
New age- and gender-dependent distributions were developed based the best available
physiological data from the literature. In this report, a summary of the current state of the
physiology file is presented, followed by the derivation of new physiological data for
body mass, normalized vo2max, and hemoglobin content. Portions of the SAS code
used for analysis are included (Appendices A-C), as is the new Physiology.txt file
(Appendix D). The final appendix (Appendix E) contains tables of all the derived
physiological parameters.
2. EVALUATION OF THE CURRENT PHYSIOLOGY FILE
DATA
The physiology.txt file was originally generated for the PNEM model by T. Johnson. It
was last updated 6/11/1998, as documented in the report User's Guide: Software for
Estimating Ventilation (Respiration) Rates for Use in Dosimetry Models, (T. Johnson and
J. Capel). In that report, the original references for the data in the file were provided. An
evaluation of the data in the file was included in a previous memo to the WA-COR under
this work assignment. A summary of those findings is repeated here.
2.1 Normalized Maximal Oxygen Uptake (nvo2max).
The nvo2max data were derived from a number of sources. The data for males,
especially, were pieced together from a variety of studies (a total of 6), leading to
discontinuities in the distributional parameters. However, in each age and gender cohort,
the distributions parameters were derived from a single published study. Additionally,
much of the nvo2max data is quite old. The data for males at age 20 and at 28-69 came
from a study from 1960 [1]. Data for males aged 0-8 and 16-19, and females 0-19 came
from a figure in a textbook from 1977 [2], which in turn was based on limited earlier
data. An additional issue with the 1977 data is (according to the report mentioned above)
that values for certain ages (very young or elderly) were acquired by simple tangential
extrapolation of the data in the figure.
B-117
-------
In addition, in some cases it was not clear how the parameters were derived from the
referenced studies. For example, Heil et al. [3] was referenced as the source of the values
for females aged 66-100. However, an examination of that study provided no clues as to
how the values were actually determined. As far as can be determined, in no place did
the authors break down the means and SDs of their data into groups separated by both
gender and age simultaneously.
2.2 Body Mass.
The current body mass data were derived from an in-depth analysis [4, 5] of the second
CDC National Health and Nutrition Examination Survey (NHANES II) body mass data
[6]. The data were relatively comprehensive, and the methods used to generate the
lognormal distributions were sound. However, the NHANES II data were compiled for
the years 1976-1980, so an analysis of more recent data is necessary to accurately
account for changes in human activity patterns in adults and especially children.
2.3 Resting Metabolic Rate.
Not included for evaluation, per discussion with WA-COR.
2.4 Hemoglobin Content and Blood Volume Factor.
The original references for the hemoglobin content or blood volume factor values given
in the current physiology.txt file could not be identified. Therefore, their validity could
not be evaluated and it was desirable that new statistics be calculated.
2.5 Summary of Findings
• In some cases, especially for nvo2max, the data are unnecessarily and confusingly
disjointed across ages.
• It is also unclear how some of the nvo2max values were derived from the
referenced studies.
• With the exception of the Schofield equations for the BM/RMR regression,
parameter distributions at each age and gender cohort were derived from data from a
single study.
• Many of the studies used are very old (ex. 1960, 1977).
• Some the data is of questionable validity (for example, the extrapolation of a
textbook figure is used), although it may have been the best available at the time of
the compilation of the file.
• The original source of the hemoglobin content and blood volume factor data could
not be identified.
B-118
-------
Given these conclusions, we recommended a full review and update of the current
physiology.txt file data. Specifically, we recommended that where possible, new
distributions or equations should be developed based on thorough, compiled data
from appropriate studies.
3. DERIVATION OF NEW DISTRIBUTIONS FOR BODY
MASS
3.1 The NHANES Body Mass Dataset.
New body mass distributions were generated from data from the National Health and
Nutrition Examination Survey (NHANES). This survey is an ongoing study carried out
by the National Center for Health Statistics of the Centers for Disease Control. EPA
recognizes the utility of this dataset in characterizing the American population for risk
assessment and policy support purposes [7].
Older NHANES data (for the years 1976-1980) have been used previously to develop
population estimates of body mass distributions [4,5]. The current Physiology.txt file
body mass distributions are based on this work. However, the analysis presented here is
based on the most recent NHANES data, for the years 1999-2004 [8].
Demographic (Demo) and Body Measurement (BMX) datasets for each of the NHANES
studies were downloaded from the NHANES website. The files were downloaded as
SAS xpt datasets. The downloaded files were as follows:
1999-2000 2001-2002 2003-2004
BMX.xpt BMX b r.xpt BMX c.xpt
Demo.xpt Demo_b.xpt Demo_c.xpt
The Demographic datasets contained the age and gender values for each survey
participant, while the Body Measurement datasets contained the body weights for each
subject. The combined dataset comprised 31,126 individuals. This resulted in
approximately 400-500 persons in each age 0-18 year cohort, and approximately 80-150
persons in each age 19-85 year cohort (the NHANES studies more heavily sampled
children).
B-119
-------
3.2 Calculation of the New Sampling Weights for the Combined
NHANES Dataset.
In the analysis of the NHANES data, sampling weights must be used to ensure that the
data are weighted to appropriately represent the national population. Sampling weights
for the combined NHANES body mass dataset were derived as recommended by the
documentation provided with the most recent NHANES release [9]. Specifically, the
sampling weight for each subject was calculated as:
W combined ~ T W 2003-2004 \
2
T
J
combined ~ T ^1999-2002
where wcombined is the sampling weight for the combined dataset, W2003-2004 is the weight
for the subjects in the most recent study, and Wi999_2oo2 is the weight for subjects in
combined 4-year (1999-2000 and 2001-2002) NHANES dataset. (Both weights are
provided with the appropriate NHANES release. The combined 1999-2002 weight,
which is not a simply half of that for the corresponding 2-year periods, was explicitly
calculated for researcher use by CDC since the two 2-year periods use different census
data.)
By using the sampling weights, once can consider any 2-year NHANES dataset or any
combination of datasets as a nationally representative sample.
3.3 Fitting the Body Mass Data.
In the current physiology file, body mass is modeled as a two-parameter lognormal
distribution. The NHANES body mass data were fit to several types of distributions
(including normal, beta, and three-parameter lognormal distributions). It was determined
that overall, the distribution that provided the best combination of good behavior over
ages and good fit to the data was a two-parameter lognormal distribution.
The data were fit to the lognormal distributions using the SAS PROC UNIVARIATE
procedure. The FREQ option of the procedure was used to apply the sampling weights.
The SAS code used to generate the body mass distributions is provided in Appendix A.
As the NHANES 1999-2003 studies only covered persons up to age 85, linear forecasts
were made for ages 86-100, as based on the data for ages 60 and greater.
3. 4 Body Mass Results.
B-120
-------
Geometric means and standard deviations (SD) for the best-fit lognormal distributions for
body mass are given in Figures 1 and 2. The means behaved fairly smoothly across ages.
Note that for children age 0-18, the values of the new fits are similar, but slightly higher
than those in the current Physiology.txt file, which were derived from earlier NHANES
studies. The new means also capture the trend towards decreasing body weight in older
persons that was previously neglected in the Physiology.txt file.
The maximum and minimum values for the distributions are presented in Figures 3 and 4.
The minimums and maximums were calculated as the 1st and 99th percentile of the raw
body mass data for the cohort. (Note that these values differ from the 1st and 99th
percentiles of the fitted lognormals.) While the minimum value is consistent with the
current Physiology.txt (which was based on earlier NHANES studies), the new cohort
maximums are generally higher than before.
The behavior of several of the body mass parameters (especially the SD) is fairly noisy,
especially for adults. This is most likely due to the smaller number of samples for adults
as compared to children. Therefore, it may desirable to use age-grouped data or running
averages over years in these age ranges. While the attached prepared Physiology.txt file
uses the "raw" parameters, smoothed results using 5-year running averages are provided
in the attached data tables (Appendix E, plots not shown). These could be used at the
direction of EPA; changing the "official" release Physiology.txt file would be trivial.
B-121
-------
MALES: Body Mass Geometric Mean
an
100.000 -,
90.000 -
80.000 -
70.000 -
60.000 -
50.000 -
40.000 -
30.000 -
20.000 -
10.000
0.000
0
Current Physiology
File
20
40 60
Age (years)
80
100
FEMALES: Body Mass Geometric Mean
New Values
Current Physiology
File
20
40 60
Age (years)
80
100
Figure 8. Geometric Means for the Best-fit Lognormal Distributions for Body Mass as a
Function of Age, Derived from NHANES 1999-2004 Study Data.
B-122
-------
MALES: Body Mass GSD
New Values
Current Physiology
File
20
40 60
Age (years)
80
100
1.000
FEMALES: Body Mass GSD
New Values
Current Physiology
File
20
40 60
Age (years)
80
100
Figure 9. Geometric Standard Deviations for the Best-fit Lognormal Distributions for Body
Mass as a Function of Age, Derived from NHANES 1999-2004 Study Data.
B-123
-------
MALES: Body Mass Minimum
20
Current Physiology
File
40 60
Age (years)
80
100
ctf
FEMALES: Body Mass Minimum
20
New Values
Current Physiology File
40 60
Age (years)
80
100
Figure 10. Minimums (1 Percentile) for Body Mass as a Function of Age, Derived from
NHANES 1999-2004 Study Data.
B-124
-------
MALES: Body Mass Maximum
250 n
200 -
150 -
I 100
50 -
0
0
•New Values
• Current Physiology
File
20
40 60
Age (years)
80
100
FEMALES: Body Mass Maximum
200 n
180 -
160 -
140 -
'55 120 -
¥ loo-
sj
S 80 -
60 -
40 -
20 -
0 -
0
•New Values
• Current Physiology
File
20
nth
40 60
Age (years)
80
100
Figure 11. Maximums (99 Percentile) for Body Mass as a Function of Age, Derived from
NHANES 1999-2004 Study Data.
B-125
-------
4. DERIVATION OF NEW DISTRIBUTIONS FOR
NORMALIZED VO2MAX
4.1 The Nvo2max Data
The NHANES studies do report data for vo2max in individuals. However, the NHANES
vo2max values are estimated values, i.e. they are not measured directly. Such estimated
values are not appropriate for use in this context (as per discussion with the WA-COR).
Therefore, nvo2max distributional shapes were determined from a large database of
experimental and literature vo2max measurements for different age/gender cohorts.
A PubMed-based literature search located a number of studies in which vo2max was
directly measured. In addition, a large number of scientific papers (-350) reporting
vo2max were also provided to Alion by the WA-COR. All the studies were evaluated for
use by determining if: 1) any normalized vo2max data for individuals were reported or 2)
any group means for narrow age-gender cohorts were reported. Studies in which the
studied age group was very broad or contained both males and females were discarded.
Also discarded were any studies in which vo2max was not normalized by body mass, or
for which no age data were reported. Data for ill or highly-trained individuals were not
used; however, studies in which subjects underwent mild or moderate exercise training
were included. Two large databases, one of individual vo2max data and one of grouped
means and SDs, were constructed from the valid studies.
The database of individual data comprised age versus nvo2max data for 1949 men and
1558 women. The data were pulled from either tables or graphs in 20 published studies
[11-30]. Additional raw experimental data were provided by the WA-COR [31]. In the
case of the graphical data, the original source was digitized and the data points were
pulled from the digital figure using graphics software. (This was accomplished by
calibrating the pixels of the digitized image with the range of age and nvo2max values.)
The individual nvo2max data for males and females are shown in Figure 5.
The grouped mean and SD data were derived from 136 studies [32-167]. These data
comprised approximately 550 means and SDs for different age/gender cohorts. Single
age/gender cohort means and SD values for the Adams data [31] were also included in
this dataset. Only data for subject groups having an age SD of less than approximately 2-
3 years were considered. The grouped mean values for men and women are shown in
Figure 6, while the group SD values are shown in Figure 7.
B-126
-------
MALES :Nvo2max
oo
-------
MALES: Study Means, Nvo2max
80.0 n
70.0 -
oo 60.0 -
E 50-°"
£ 40.0 -
a
70.0 n
60.0 -
50.0 -
40.0 -
30.0 -
20.0 -
10.0 -
0.0 -
*
-------
MALES: Study Standard Deviations: Nvo2max
16.0 n
14.0 -
oj 12.0 -
1 10.0 -
•—^
* 8.0 -
6.0 -
4.0 -
2.0 -
0.0 -
(N
O
0.0
20.0
:
40.0
Age (Years)
60.0
80.0
FEMALES: Study Standard Deviations: Nvo2max
00
t
nvo2ma
14.0 n
12.0 -
10.0 -
8.0 -
6.0 -
4.0 -
2.0 -
o.o -
0
*
* • **
>^*J.
*
1 1 1 1 1 1 1
0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
Age (Years)
Figure 14. Nvo2max Standard Deviations for Males and Females, Derived from Literature
Studies.
B-129
-------
4.2 Determining the NVo2max Distributions
Both the grouped mean and the individual datasets were evaluated for use in deriving the
nvo2max parameters.
The group means and SD were combined into single age/gender cohort values. The
combined means were calculated as mean of the group means, weighted by the number of
subjects. The group SD were calculated by transforming each group SD to a group
variance, calculating the mean variance (weighted by the number of subjects in each
study) and retransforming the variances to SDs. The combined group means and SDs are
given in Figures 8 and 9.
The combined group means were fairly well-behaved across age and gender cohorts (see
Figure 8), while the SD data (Figure 9) were noisier. These data may be appropriate for
use in the Physiology.txt file; however, it was noted that the group mean data, while
plentiful for children, were not very well represented in the adult (30+ years) age range
(especially for women). This is mainly due to the fact that very few investigators use
narrow age cohorts when studying adults, rather, it was far more common for broader age
groups to be used. These data were not included in the grouped mean analysis, as the
mean nvo2max for a broad age group cannot be assumed valid for the cohort represented
by the study age mean. Therefore, we opted to use the database of individual nvo2max
measurements to develop new distributions for the Physiology.txt file.
The individual nvo2max data were fit to several types of distributions (including normal,
beta, and lognormal distributions). It was determined that the normal distribution fit the
data best. The parameters (means and standard deviations) of the best-fit distributions
were obtained using the SAS PROC UNIVARIATE procedure. The SAS code used to fit
the data is given in Appendix B.
Both raw and smoothed nvo2max fits were calculated. Calculating 5-year running
averages did not smooth the data considerably. Therefore, the smoothed fits were
determined by choosing a best-fit functional form for the nv02max data. The data were
fit to functions as follows:
Mean (Age 0-20): Linear function
Mean (Age 21-100): Parabolic function
SD (Age 0-26): Linear function
SD (Age 27-100): Parabolic function
Fitting the data in this manner also allowed for all age/gender cohorts to be represented.
Since only cohorts having N>10 were fit to distributions, there were some cohorts for
which no parameters were calculated. The raw and smoothed fits for means are given in
Figure 10; analogous data for SD is given in Figure 11. The raw nvo2max parameters
were not as clean across ages as the body mass data (probably due to the much smaller
sample size), and thus the smoothed fits were selected for use in the attached
Physiology.txt file. As with body mass, the raw fits may be used at the direction of EPA.
B-130
-------
The results for the nvo2max means were in fact quite close to those in the current file.
However, the values exhibited much more consistent behavior across ages, and the values
for elderly persons were lower than previously. The SD values were also in the same
range as the current values, yet they no longer demonstrate nonsensical discontinuities
across ages.
The minimum and maximum nvo2max values were assumed to be the 1st and 99th
percentile of the best-fit lognormal distribution. (Note: this is different from the method
used for estimating the body mass limits. In that case, the samples were large enough
that the percentiles of the raw data were appropriate for use as minimum and maximum.
As the nvo2max data cohorts had much smaller N than the NHANES studies, the raw
percentiles were less appropriate.) The maximum and minimum values are shown in
Figures 12 and 13.
B-131
-------
MALES: Nvo2max, Combined Group Means
60 -,
™ 50 -
1 40 -
I 3° "
C 10 -
o -
0
20
40
Age (Years)
60
80
50 -,
j^ 40 -
^ 30 -
H 20 -
o
10 -
0
FEMALES: Nvo2max, Combined Group Means
0
10
60
70
20 30 40 50
Age (Years)
Figure 15. Combined Nvo2max Group Means for Males and Females
B-132
-------
MALES: Nvo2max, Combined Group SD
16 -
14 -
UJJ
^ 12 -
£ 10-
^j>
8 -
6 -
4 -
2 -
0 -
c
0
20
40
Age (Years)
60
80
10 -i
^ 8 "
1 6-
I 4
O
« 2 H
FEMALES: Nvo2max, Combined Group SD
10
20
30 40
Age (Years)
50
60
70
Figure 16. Combined Nvo2max Group Standard Deviations.
B-133
-------
MALES: MEAN Nvo2max
x
cS
a so
-------
14 -
12-
10-
X
cS
P, 6-
2-
MALES: Nvo2max Standard Deviation
• Raw Values
• Fit Values
Current Physiology.txt
20
40 60
Age (Years)
100
FEMALES: Nvo2max Standard Deviation
Raw Values
Fit Values
Current Physiology.txt
20
40 60
Age (Years)
100
Figure 18. Nvo2max Normal Distribution Fits: Raw Fit Standard Deviations and Smoothed
Fits.
B-135
-------
oo
CN
60 -
50 -
40 -
30 -
20 -
10 -
0 -
0
MALES: Nvo2max Minimum
New Values
Current Physiology.txt
20
40 60
Age (Years)
80
100
45 n
40 -
oo 35 -
^ 25 "
I 20 -
-------
MALES: N vo2max Maximum
80 -,
70 -
j) 60 -
H 50"
& 40 -
^ 30-
a 20 -
10 -
0 -
0
•New Values
• Current Physiology.txt
20
40 60
Age (Years)
80
100
o
I 20-
10 -
0
FEMALES: Nvo2max Maximum
0
•New Values
Current Physiology.txt
20
40 60
Age (Years)
80
100
..th
Figure 20. Nvo2max Maximums. 99 Percentile of the Best-fit Normal Distribution.
B-137
-------
5. for
The new hemoglobin content values were derived from the combined NHANES 1999-
2000 and 2001-2002 datasets. As of December 2005, hemoglobin data had not yet been
released for the 2003-2004 study. The age data was provided in the Demographic
datasets (Demo.xpt and Demo_b.xpt, previously downloaded for the body mass analysis)
for the two survey periods, while hemoglobin content (in g/dL) was provided in the
Laboratory #25 (Complete Blood Count) datasets (Iab25.xpt and 125_b.xpt, which were
downloaded for this analysis). The dataset comprised 20,321 individuals; appropriate
sample weights were used for the combined 4-year (1999-2002) dataset as provided with
the NHANES 2001-2002 data release. Similarly to the body mass data, the hemoglobin
content values were analyzed in SAS. The age and hemoglobin datasets were merged
and fit to normal distributions using the SAS PROC UNIVARIATE procedure. The
FREQ option of the procedure was used to apply the sampling weights. The SAS code in
provided in the Appendix C.
Hemoglobin content statistics were estimated for single-year age and gender cohorts for
ages 1-19, as the behavior of the means were smooth in this age range. For persons 20
and over, the data were grouped in 5-year cohorts (20-24, 25-29, etc.) No blood count
data were available for subjects under 1 year of age or greater than 90. The age 0 mean
values were obtained by a linear regression of ages 1-20 (males) or 1 to 11 (females) back
to age 0. These were the ages for which the hemoglobin content demonstrated an
increase with age. The 91-95 and 96-100 mean values were obtained by a linear
regression of the 61-65 and older age groups. As the standard deviations did not appear
to behave as smoothly with age as did the mean values, the age 0 value was assumed
equal to the age 1 value, and the age 91-95 and 96-100 value was assumed equal to the
age 90-94 value.
The resulting means and standard deviations for the best-for normal distributions for
hemoglobin content are given in Figures 14 and 15. The current hemoglobin content
values are shown for comparison.
The main conclusions that can be made is that the current Physiology.txt input file
overestimates mean hemoglobin content in children and in older persons. The standard
deviation values in the current physiology.txt file are fairly close to those found in this
analysis. The new values are not very smooth over ages; EPA may elect to continue to
use the current values. It should be noted that the original reference for the current
hemoglobin statistics is unknown.
Note: In the current implementation of APEX, the hemoglobin content statistics affect
only the CO dose algorithm calculations.
B-138
-------
17 n
, , 16
5i is H
00
¥14"
I 13"
r^ 12
11
10
MALES: Mean Hemoglobin Content
* New Values
Current Physiology.txt
^r * • • *»»
•
>
20
40 60
Age (years)
80
100
17 n
16
^ 15 -
5 14 -
c
0^ -1 r-\
S 13 H
O 12 ^
ffi 11
10
FEMALES: Mean Hemoglobin Content
• *
0 20 40 60 80 100
Age (years)
Figure 21. Mean Values of Hemoglobin Content as Derived from the 1999-2002 NHANES
Dataset, with Comparison to Current Physiology.txt Values
B-139
-------
MALES: Hemoglobin Content Standard Deviation
"S
O
O
ffi
2 n
1.5 -
1 <
0.5 -
n
» » »
r
^^"^ 2 ^ ^
4 New Values
20 40 60 80 100
Age (years)
FEMALES: Hemoglobin Content Standard Deviation
O
O
0.5 -
» + « » /—•-
20
40 60
Age (years)
80
100
Figure 22. Values of Hemoglobin Content Standard Deviation as Derived from the 1999-
2002 NHANES Dataset, with Comparison to Current Physiology.txt Values
B-140
-------
6. BLOOD VOLUME AS A FUNCTION OF HEIGHT AND
WEIGHT
In APEX, blood volume is estimated as a function of height and weight by the following
equation:
vuood = BVF*Weight+ K*Height3 - 30
where Vbiood is the blood volume (ml), Weight is in pounds, and height is in inches. BVF
is the blood volume factor that is read in from the physiology file, and K is a gender-
dependent constant (0.00683 for males, 0.00678 for females). This is a modification of
Allen's equation [168] to include the age/gender dependent BVF and adjusted for the
given units.
As previously mentioned, the data upon which the BVF values in the physiology file
were based could not be identified. The available documentation for pNEM documents a
non-age-dependent use of these equations.
In addition, no appropriate data were found for deriving new estimates for the BVF
variable as a function of age and gender for use with the Allen equations. It should be
noted however, that these equations were modified by Nadler [169]. These equations
seem to be used somewhat more often than the originals in the literature.
In addition, other (more recent) equations exist for estimation of blood volume from
height and weight specifically in children [170,171] or body surface area [172]. In
particular, Linderkamp et al. [170] derived prediction equations for blood volume as a
function of a number of physiological parameters for children in three different age
groups. It is recommended that further analysis of this study and others be undertaken.
However, inclusion of new blood volume equations in APEX would require changes
beyond the current physiology file (i.e. other, more intensive, code changes would be
needed). Thus, at the present time, no specific improvements to the current BVF values
in the physiology file can be made.
B-141
-------
1. Astrand, I. 1960. Aerobic work capacity in men and women with special
reference to age. Acta. Phys. Scand. 49: Suppl. 169.
2. Astrand, PO and Rodhal K. 1977. Textbook of Work Physiology. 2nd addition,
McGraw-Hill, NY.
3. Heil DP, Freedson PS, Ahlquist LE, Price J, and Rippe JM. 1995. Nonexercise
regression models to estimate peak oxygen consumption. Med. Sci. Sports Exer 27:599-
606.
4. Brainard J and Burmaster D. 1992. Bivariate distributions for height and weight
of men and women in the United States. Risk Analysis, 12(2):267-275.
5. Burmaster DE and Crouch E. 1997. Lognormal distributions of body weight as a
function of age for males and females in the United States, 1976-1980. Risk Analysis,
17(4): 499-505.
6. Centers for Disease Control and Prevention (CDC). National Center for Health
Statistics (NCHS). National Health and Nutrition Examination Survey Data. Hyattsville,
MD: U.S. Department of Health and Human Services, Centers for Disease Control and
Prevention, 1976-1980. .
7. U.S. Environmental Protection Agency (EPA). (2003) Handbook for use of data
from the national health and nutrition examination surveys (NHANES). National Center
for Environmental Assessment, Washington, DC: EPA/600/R-02/044. Available from:
National Technical Information Service, Springfield, VA, and
.
8. Centers for Disease Control and Prevention (CDC). National Center for Health
Statistics (NCHS). National Health and Nutrition Examination Survey Data. Hyattsville,
MD: U.S. Department of Health and Human Services, Centers for Disease Control and
Prevention, 1999-2004. .
9. Center for Disease Control. National Center For Health Statistics. The National
Health and Nutriation Examination Survey (NHANES). Analytic and Reporting
Guidelines. December 2005. Available online at www.cdc.gov.
10. Andersen KL, Ilmarinen J, Rutenfranz J, Ottmann W, Berndt I, Kylian H. Leisure
time sport activities and maximal aerobic power during late adolescence. Eur J Appl
Physiol 1984; 52: 431-6.
11. Armstrong N, Kirby BJ, McManus AM, Welsman JR. Aerobic fitness of
prepubescent children. AnnHumBiol. 1995 Sep-Oct;22(5):427-41.
12. Armstrong N, Welsman JR. Assessment and interpretation of aerobic fitness in
children and adolescents.Exerc Sport Sci Rev. 1994;22:435-76.
B-142
-------
13. Astrand I. Aerobic work capacity in men and women with special reference to
age. ActaPhysiol Scand. 1960;49(Suppl 169): 1-92.
14. Eliakim A, Scheett T, Allmendinger N, Brasel JA, Cooper DM. Training, muscle
volume, and energy expenditure in nonobese American girls. J Appl Physiol. 2001
Jan;90(l):35-44.
15. Falgairette, G. 1989. Evolution de la puissance maximal aerobie de 1'enfance a
1'age adulte: influence de 1'activite physique et al. sportive. Revue Des Sciences et al.
Techniques en Activites Physiques et al. Sportives. 10:43-58.
16. Fleg JL, Lakatta EG. Role of muscle loss in the age-associated reduction in VO2
max. J Appl Physiol. 1988 Sep;65(3): 1147-51.
17. Ghesquiere J, Reybrouck T, Faulkner JA, Cattaert A, Fagard R, Amery A.
Anaerobic threshold for long-term exercise and maximal exercise performance. Ann Clin
Res. 1982; 14 Suppl 34:37-41.
18. Kamon E, Pandolf KB. Maximal aerobic power during laddermill climbing,
uphill running, and cycling. J Appl Physiol. 1972 Apr;32(4):467-73.
19. Katch VL, Park MW. Minute-by-minute oxygen requirement and work efficiency
for constant-load exercise of increasing duration. Res Q. 1975 Mar;46(l):38-47.
20. McMiken DF. Oxygen deficit and repayment in submaximal exercise. Eur J Appl
Physiol Occup Physiol. 1976 Jun 21;35(2): 127-36.
21. Nevill A, Rowland T, Goff D, Mattel L, Ferrone L. Scaling or normalising
maximum oxygen uptake to predict 1-mile run time in boys. Eur J Appl Physiol. 2004
Jul;92(3):285-8. Epub 2004 Apr 9.
22. Pivarnik JM, Fulton JE, Taylor WC, Snider SA. Aerobic capacity in black
adolescent girls. Res Q Exerc Sport. 1993 Jun;64(2):202-7.
23. Posner JD, Gorman KM, Klein HS, Cline CJ. Ventilatory threshold:
measurement and variation with age. J Appl Physiol. 1987 Oct;63(4):1519-25.
24. Reybrouck T, Heigenhauser GF, Faulkner JA. Limitations to maximum oxygen
uptake in arms, leg, and combined arm-leg ergometry. J Appl Physiol. 1975
May;38(5):774-9.
25. Ribisl PM, Kachadorian WA. Maximal oxygen intake prediction in young and
middle-aged males. J Sports Med Phys Fitness. 1969 Mar;9(l): 17-22.
26. Schiller BC, Casas YG, Desouza CA, Seals DR. Maximal aerobic capacity across
age in healthy Hispanic and Caucasian women. J Appl Physiol. 2001 Sep;91(3): 1048-54.
B-143
-------
27. Seals DR, Hagberg JM, Hurley BF, Ehsani AA, Holloszy JO. Endurance training
in older men and women. I. Cardiovascular responses to exercise. J Appl Physiol. 1984
Oct;57(4): 1024-9.
28. Sheel A, Seddon N, Knight A, McKenzie DC, R Warburton DE. Physiological
responses to indoor rock-climbing and their relationship to maximal cycle ergometry.
Med Sci Sports Exerc. 2003 Jul;35(7): 1225-31.
29. Talbot LA, Metter EJ, Fleg JL. Leisure-time physical activities and their
relationship to cardiorespiratory fitness in healthy men and women 18-95 years old. Med
Sci Sports Exerc. 2000 Feb;32(2):417-25.
30. Tanaka H, Desouza CA, Jones PP, Stevenson ET, Davy KP, Seals DR. Greater
rate of decline in maximal aerobic capacity with age in physically active vs. sedentary
healthy women. J Appl Physiol. 1997 Dec;83(6): 1947-53.
31. Adams. Personal Communication - T. McCurdy EPA.
32. Ahmad F, Kavey RE, Kveselis DA, Gaum WE, Smith FC. Responses of non-
obese white children to treadmill exercise. J Pediatr. 2001 Aug;139(2):284-90.
33. Andersen KL, Hermansen L. Aerobic work capacity in middle-aged Norwegian
men. J Appl Physiol. 1965 May;20(3):432-6.
34. Andersen KL, Ilmarinen J, Rutenfranz J, Ottmann W, Berndt I, Kylian H, Ruppel
M. Leisure time sport activities and maximal aerobic power during late adolescence. Eur
J Appl Physiol Occup Physiol. 1984;52(4):431-6.
35. Andersen KL, Seliger V, Rutenfranz J, Mocellin R. Physical performance
capacity of children in Norway. Part I. Population parameters in a rural inland
community with regard to maximal aerobic power. Eur J Appl Physiol Occup Physiol.
1974;33(3): 177-95.
36. Armstrong N and Welsman J. Daily physical activity estimated from continuous
heart rate monitoring and laboratory indices of aerobic fitness in pre-adolescent children.
(abs.) Res Q Exer Sport 64:A-24, 1993.
37. Armstrong N, Balding J, Gentle P, Kirby B. Estimation of coronary risk factors
in British schoolchildren: a preliminary report. Br J Sports Med. 1990 Mar;24(l):61-6.
38. Armstrong N, Balding J, Gentle P, Kirby B. Patterns of physical activity among
11 to 16 year old British children. BMJ. 1990 Jul 28;301(6745):203-5.
39. Armstrong N, Bray S. Physical activity patterns defined by continuous heart rate
monitoring. Arch Dis Child. 1991 Feb;66(2):245-7.
40. Armstrong N, Davies B, Heal M. The specificity of energy utilisation by trained
and untrained adolescent boys. Br J Sports Med. 1983 Sep; 17(3): 193-9.
B-144
-------
41. Armstrong N, Kirby BJ, McManus AM, Welsman JR. Aerobic fitness of
prepubescent children. Ann Hum Biol. 1995 Sep-Oct;22(5):427-41. Erratum in: Ann
HumBiol 1996 Mar-Apr;23(2): 188.
42. Armstrong N, Welsman JR, Kirby BJ. Submaximal exercise and maturation in
12-year-olds. J Sports Sci. 1999Feb;17(2):107-14.
43. Armstrong N, Welsman JR, Kirby BJ. Peak oxygen uptake and maturation in 12-
yr olds. Med Sci Sports Exerc. 1998 Jan;30(l): 165-9.
44. Armstrong N, Williams J, Balding J, Gentle P, Kirby B. The peak oxygen uptake
of British children with reference to age, sex and sexual maturity. Eur J Appl Physiol
Occup Physiol. 1991;62(5):369-75.
45. Atomi Y, Fukunaga T, Yamamoto Y, Hatta H. Lactate threshold and vo2max of
trained and untrained boys relative to muscle mass and composition. In: Rutenfranz J,
Mocellin R, Klimt F, eds. children and exercise XII. International series on sport
sciences, Vol 16. Champaign, IL: Human Kinetics, 1986:227-232.
46. Atomi Y, Iwaoka K, Hatta H, Miyashita M, Yamamoto Y. Daily physical activity
levels in preadolescent boys related to VO2max and lactate threshold. Eur J Appl Physiol
Occup Physiol. 1986;55(2):156-61.
47. Bailey, D.A., Ross, W.D., Mirwald, R.L. and Weese, C. Size dissociation of
maximal aerobic power during growth in boys. Medicine in Sport 1978; 11: 140-151.
48. Bale P. Pre- and post-adolescents' physiological response to exercise. Br J Sports
Med. 1981Dec;15(4):246-9.
49. Bar-Or O, Skinner JS, Bergsteinova V, Shearburn C, Royer D, Bell W, Haas J,
Buskirk ER. Maximal aerobic capacity of 6—15 year-old girls and boys with subnormal
intelligence quotients. ActaPaediatr Scand Suppl. 1971;217:108-13.
50. Becker DM, Vaccaro P. Anaerobic threshold alterations caused by endurance
training in young children. J Sports Med Phys Fitness. 1983 Dec;23(4):445-9.
51. Binkhorst 1984 (Ref??) From T. McCurdy spreadsheet, dated Nov 15, 2000.
52. Binkhorst RA, Saris WHM, Noordeloos AM, et al. Maximal oxygen consumption
of children (6 to 18 years) predicted from maximal and submaximal values in treadmill
and bicycle tests. In: Rutenfranz J, Mocellin R, Klimt F, eds. children and exercise XII.
International series on sport sciences, Vol 16. Champaign, IL: Human Kinetics,
1986:227-232.
53. Boileau RA, Bonen A, Heyward VH, Massey BH. Maximal aerobic capacity on
the treadmill and bicycle ergometer of boys 11-14 years of age. J Sports Med Phys
Fitness. 1977 Jun; 17(2): 153-62.
B-145
-------
54. Bonen A, Heyward VH, Cureton KJ, Boileau RA, Massey BH. Prediction of
maximal oxygen uptake in boys, ages 7-15 years. Med Sci Sports. 1979 Spring; 11(1):24-
9.
55. Brown, J.D., A.D. Mahon and D.M. Plank. Attainment of maximal exercise
criteria in boys and men. Journal of Sports Medicine and Physical Fitness, 42: 135-140,
2002.
56. Buskirk E, Taylor HL. Maximal oxygen intake and its relation to body
composition, with special reference to chronic physical activity and obesity. J Appl
Physiol. 1957 Jul;ll(l):72-8.
57. Chausow 1984 (Ref??) In T. McCurdy spreadsheet, dated Nov 15, 2000.
58. Cooper DM, Berry C, Lamarra N, Wasserman K. Kinetics of oxygen uptake and
heart rate at onset of exercise in children. J Appl Physiol. 1985 Jul;59(l):211-7.
59. Cooper DM, Weiler-Ravell D, Whipp BJ, Wasserman K. Aerobic parameters of
exercise as a function of body size during growth in children.! Appl Physiol. 1984
Mar;56(3):628-34.
60. Cumming GR, Borysyk LM. Criteria for maximum oxygen uptake in men over
40 in a population survey. Med Sci Sports. 1972 Spring;4(l): 18-22.
61. Cumming GR, Everatt D, Hastman L. Bruce treadmill test in children: normal
values in a clinic population. Am J Cardiol. 1978 Jan;41(l):69-75.
62. Cumming GR, Goulding D, Baggley G. Failure of school physical education to
improve cardiorespiratory fitness. Can Med Assoc J. 1969 Jul 26;101(2):69-73.
63. Cunningham DA, Paterson DH, Blimkie CJ, Donner AP. Development of
cardiorespiratory function in circumpubertal boys: a longitudinal study. J Appl Physiol.
1984Feb;56(2):302-7.
64. Cunningham DA, Stapleton JJ, MacDonald 1C, Paterson DH. Daily energy
expenditure of young boys as related to maximal aerobic power. Can J Appl Sport Sci.
1981Dec;6(4):207-ll.
65. Cunningham DA, Telford P, Swart GT. The cardiopulmonary capacities of young
hockey players: age 10. Med Sci Sports. 1976 Spring;8(l):23-5.
66. Cunningham DA, van Waterschoot BM, Paterson DH, Lefcoe M, Sangal SP.
Reliability and reproducibility of maximal oxygen uptake measurement in children. Med
Sci Sports. 1977 Summer;9(2): 104-8.
67. Cureton KJ, Sloniger MA, Black DM, McCormack WP, Rowe DA. Metabolic
determinants of the age-related improvement in one-mile run/walk performance in youth.
Med Sci Sports Exerc. 1997 Feb;29(2):259-67.
B-146
-------
68. Cureton KJ, Sloniger MA, O'Bannon JP, Black DM, McCormack WP. A
generalized equation for prediction of VO2peak from 1-mile run/walk performance. Med
Sci Sports Exerc. 1995 Mar;27(3):445-51.
69. Davies CT. Limitations to the prediction of maximum oxygen intake from
cardiac frequency measurements. J Appl Physiol. 1968 May;24(5):700-6.
70. Davies CT. Metabolic cost of exercise and physical performance in children with
some observations on external loading. Eur J Appl Physiol Occup Physiol. 1980;45(2-
3):95-102.
71. Dehn MM, Bruce RA. Longitudinal variations in maximal oxygen intake with
age and activity. J Appl Physiol. 1972 Dec;33(6):805-7.
72. Docherty D, Wenger HA, Collis ML. The effects of resistance training on aerobic
and anaerobic power of young boys.Med Sci Sports Exerc. 1987 Aug;19(4):389-92.
73. Eiberg S, Hasselstrom H, Gronfeldt V, Froberg K, Svensson J, Andersen LB.
Maximum oxygen uptake and objectively measured physical activity in Danish children
6-7 years of age: the Copenhagen school child intervention study. Br J Sports Med. 2005
Oct;39(10):725-30.
74. Eisenman PA, Golding LA. Comparison of effects of training on Vo2max in girls
and young women. Med Sci Sports. 1975 Summer;7(2): 136-8.
75. Fawkner S, Armstrong N. Modeling the vo2 kinetic response to moderate
intensity exercise in children. Acta Kinesiol Universit Tartu 7:80-84, 2002.
76. Fox EL, Bartels RL, Billings CE, O'Brien R, Bason RM, Mathews DK.
Frequency and duration of interval training programs and changes in aerobic power. J
Appl Physiol 1975; 38:481-84
77. Fredriksen PM, Ingjer F, Nystad W, Thaulow E. A comparison of VO2(peak)
between patients with congenital heart disease and healthy subjects, all aged 8-17 years.
Eur J Appl Physiol Occup Physiol. 1999 Oct;80(5):409-16.
78. Gastin PB, Costill DL, Lawson DL, Krzeminski K, McConell GK. Accumulated
oxygen deficit during supramaximal all-out and constant intensity exercise. Med Sci
Sports Exerc. 1995 Feb;27(2):255-63.
79. George JD. Alternative approach to maximal exercise testing and VO2 max
prediction in college students. Res Q Exerc Sport. 1996 Dec;67(4):452-7.
80. Gilliam TB, Sady S, Thorland WG, Weltman AL. Comparison of peak
performance measures in children ages 6 to 8, 9 to 10, and 11 to 13 years. Res Q. 1977
Dec;48(4):695-702.
B-147
-------
81. Gilliam, T. & Freedson, P. Effects of a twelve-week school physical fitness
program on peak VO2, body composition, and blood lipids in 7 to 9 year old children. Int
J of Sports Med 1980; 1:73-78.
82. Gutin B, Barbeau P, Owens S, Lemmon CR, Bauman M, Allison J, Kang HS,
Litaker MS. Effects of exercise intensity on cardiovascular fitness, total body
composition, and visceral adiposity of obese adolescents. Am J Clin Nutr. 2002
May;75(5):818-26.
83. Hansen, H. S., K. Froberk, J. R. Nielsen, and N. Hyldebrandt. A new approach to
assessing maximal aerobic power in children: the Odense School Child Study. Euro. J.
Appl. Physiol. 58: 618-624, 1989.
84. Harms C, Cordain L, Stager J, Sockler J, and Harris M. Body fat mass affects
postexercise metabolism in males of similar lean body mass. Med Exer Nutr Health 4:33-
39, 1995.
85. Hebestreit H, Bar-Or O, McKinty C, Riddell M, Zehr P. Climate-related
corrections for improved estimation of energy expenditure from heart rate in children. J
Appl Physiol. 1995 Jul;79(l):47-54.
86. Hebestreit H, Kriemler S, Hughson RL, Bar-Or O. Kinetics of oxygen uptake at
the onset of exercise in boys and men. J Appl Physiol. 1998 Nov;85(5):1833-41.
87. Hebestreit H, Mimura K, Bar-Or O. Recovery of muscle power after high-
intensity short-term exercise: comparing boys and men. J Appl Physiol. 1993
Jun;74(6):2875-80.
88. Hermansen L, Andersen KL. Aerobic work capacity in young Norwegian men
and women. J Appl Physiol. 1965 May;20(3):425-31.
89. Hossack KF, Bruce RA. Maximal cardiac function in sedentary normal men and
women: comparison of age-related changes. J Appl Physiol. 1982 Oct;53(4):799-804.
90. Humphrey LD, Falls HB. Circulatory and metabolic responses of young women
at HR = 180 and HR = max. Res Q. 1975 Dec;46(4):498-504.
91. Hunt B. R, George J.D., Vehrs P.R., Fisher A. G., Fellingham G. W. (2000),
Validity of a submaximal 1-mile track jog test in predicting VO2max in fit teenagers.
Pediatric exercise science 12(1):80-90.
92. Ikai M, Kitagawa K. Maximum oxygen uptake of Japanese related to sex and
age. Med Sci Sports. 1972 Fall;4(3): 127-31.
93. Kaminsky L, and Whaley M. Effect of interval-type exercise on excess
postexercise oxygen consumption in obese and normal-weight women. Med Exer Nutr
Health 2:106-111, 1993.
B-148
-------
94. Kaminsky LA, Wehrli KW, Mahon AD, Robbins GC, Powers DL, Whaley MH.
Evaluation of a shallow water running test for the estimation of peak aerobic power. Med
Sci Sports Exerc. 1993 Nov;25(ll): 1287-92.
95. Kanaley JA, Boileau RA. The onset of the anaerobic threshold at three stages of
physical maturity. J Sports Med Phys Fitness. 1988 Dec;28(4):367-74.
96. Katch FI, McArdle WD, Pechar GS. Relationship of maximal leg force and leg
composition to treadmill and bicycle ergometer maximum oxygen uptake. Med Sci
Sports. 1974 Spring;6(l):38-43.
97. Kemper H and Vershuur R. Maximal aerobic power in 13 and 14 year old
teenagers in relation to biologic age. Int J Sports Med. 2:97-100, 1981.
98. Koch G, Eriksson BO. Effect of physical training on pulmonary ventilation and
gas exchange during submaximal and maximal work in boys aged 11 to 13 years. Scand J
ClinLab Invest. 1973 Jan;31(l):87-94.
99. Kohrt WM, Malley MT, Coggan AR, Spina RJ, Ogawa T, Ehsani AA, Bourey
RE, Martin WH 3rd, Holloszy JO. Effects of gender, age, and fitness level on response of
VO2max to training in 60-71 yr olds. J Appl Physiol. 1991 Nov;71(5):2004-l 1.
100. Krahenbuhl GS, Pangrazi RP, Burkett LN, Schneider MJ, Petersen G. Field
estimation of VO2 max in children eight years of age. Med Sci Sports. 1977
Spring;9(l):37-40.
101. Krahenbuhl GS, Pangrazi RP, Petersen GW, Burkett LN, Schneider MJ. Field
testing of cardiorespiratory fitness in primary school children. Med Sci Sports. 1978
Fall;10(3):208-13.
102. Krahenbuhl GS, Pangrazi RP. Characteristics associated with running
performance in young boys. Med Sci Sports Exerc. 1983;15(6):486-90.
103. Loftin M, Strikmiller P, Warren B, Myers L, Schroth L, Pittman J, Harsha D, and
Sothern M. Comparison and relationship of Vo2peak and physical activity patterns in
elementary and high school females. Fed Exer Sci. 10:153-163, 1998.
104. Macek M and Vavra J. The adjustment of oxygen uptake at the onset of exercise:
a comparison between prepubertal boys and young adults. In J Sports Med 1:70-72, 1980.
105. Macek M, Vavra J, Novosadova J. Prolonged exercise in prepubertal boys. I.
Cardiovascular and metabolic adjustment. Eur J Appl Physiol Occup Physiol. 1976 Sep
23;35(4):291-8.
106. Macek M, Vavra J. Cardiopulmonary and metabolic changes during exercise in
children 6-14 years old. J Appl Physiol. 1971 Feb;30(2):200-4.
B-149
-------
107. Mahler DA, Mejia-Alfaro R, Ward J, Baird JC. Continuous measurement of
breathlessness during exercise: validity, reliability, and responsiveness. J Appl Physiol.
2001 Jun;90(6):2188-96.
108. Mahon AD, Anderson CS, Hipp MJ, Hunt KA. Heart rate recovery from
submaximal exercise in boys and girls. Med Sci Sports Exerc. 2003 Dec;35(12):2093-7.
109. Mahon AD, Gay JA, Stolen KQ. Differentiated ratings of perceived exertion at
ventilatory threshold in children and adults. Eur J Appl Physiol Occup Physiol. 1998
Jul;78(2):115-20.
110. Mahon AD, Vaccaro P. Ventilatory threshold and VO2max changes in children
following endurance training. Med Sci Sports Exerc. 1989 Aug;21(4):425-31.
111. Mahon, A.D., C.C. Cheatham, K.Q. Kelsey, and J.D. Brown. Plasma fibrinogen,
physical activity and aerobic fitness in children. In: Children and Exercise XIX: 19th
International Seminar on Paediatric Work Physiology, N. Armstrong, B. Kirby, and J.
Wellsman, Editors. London: Chapman & Hall Ltd., 1997. rs. London: Chapman &
Hall Ltd., 997
112. Maksud MG, Coutts KD, Hamilton LH. Oxygen uptake, ventilation, and heart
rate. Study in Negro children during strenuous exercise. Arch Environ Health. 1971
Jul;23(l):23-8.
113. Mandigout S, Lecoq AM, Courteix D, Guenon P, Obert P. Effect of gender in
response to an aerobic training programme in prepubertal children. Acta Paediatr. 2001
Jan;90(l):9-15.
114. Massicotte DR, Macnab RB. Cardiorespiratory adaptations to training at
specified intensities in children. Med Sci Sports. 1974 Winter;6(4):242-6.
115. Matsui H, Miyashita M, Miura M, Kobayashi K, Hoshikawa T. Maximum oxygen
intake and its relationship to body weight of Japanese adolescents.Med Sci Sports. 1972
Spring;4(l):27-32.
116. McArdle WD and Magel JR. Physical work capacity and maximum oxygen
uptake in treadmill and bicycle exercise. Med Sci Sports 2: 118-126, 1971 McCardle et
al. 1972
117. McDonough JR, Kusumi F, Bruce RA. Variations in maximal oxygen intake with
physical activity in middle-aged men. Circulation. 1970 May;41(5):743-51.
118. McManus AM, Armstrong N, Williams CA. Effect of training on the aerobic
power and anaerobic performance of prepubertal girls. Acta Paediatr. 1997
May;86(5):456-9.
B-150
-------
119. McMurray RG, Guion WK, Ainsworth BE, Harrell JS. Predicting aerobic power
in children. A comparison of two methods. J Sports Med Phys Fitness. 1998
Sep;38(3):227-33.
120. Mercier J, Varray A, Ramonatxo M, Mercier B, Prefaut C. Influence of
anthropometric characteristics on changes in maximal exercise ventilation and breathing
pattern during growth in boys. Eur J Appl Physiol Occup Physiol. 1991;63(3-4):235-41.
121. Murray TD, Walker JL, Jackson AS, Morrow JR Jr, Eldridge JA, Rainey DL.
Validation of a 20-minute steady-state jog as an estimate of peak oxygen uptake in
adolescents. Res Q Exerc Sport. 1993 Mar;64(l):75-82.
122. Nakagawa A, Ishiko T. Assessment of aerobic capacity with special reference to
sex and age of junior and senior highschool students in Japan.
123. Palgi Y, Gutin B, Young J, Alejandro D. Physiologic and anthropometric factors
underlying endurance performance in children. Int J Sports Med. 1984 Apr;5(2):67-73.
124. Paterson DH, Cunningham DA, Bumstead LA. Recovery O2 and blood lactic
acid: longitudinal analysis in boys aged 11 to 15 years. Eur J Appl Physiol Occup
Physiol. 1986;55(l):93-9.
125. Paterson DH, Cunningham DA, Donner A. The effect of different treadmill
speeds on the variability of VO2 max in children. Eur J Appl Physiol Occup Physiol.
1981;47(2): 113-22.
126. Paterson DH, McLellan TM, Stella RS, Cunningham DA. Longitudinal study of
ventilation threshold and maximal O2 uptake in athletic boys. J Appl Physiol. 1987
May;62(5):2051-7.
127. Petzl DH, Haber P, Schuster E, Popow C, Haschke F. Reliability of estimation of
maximum performance capacity on the basis of submaximum ergometric stress tests in
children 10-14 years old. Eur J Pediatr. 1988 Feb; 147(2): 174-8.
128. Pivarnik JM, Bray MS, Hergenroeder AC, Hill RB, Wong WW. Ethnicity affects
aerobic fitness in US adolescent girls. Med Sci Sports Exerc. 1995 Dec;27(12): 1635-8.
129. Pivarnik JM, Dwyer MC, Lauderdale MA. The reliability of aerobic capacity
(VO2max) testing in adolescent girls. Res Q Exerc Sport. 1996 Sep;67(3):345-8.
130. Pivarnik, J.M., W.C. Taylor, and S.S. Cummings. Longitudinal assessment of
aerobic fitness in middle-school African American girls. Pediatr Exerc Sci 10: 21-27,
1998.
131. Reybrouck T, Ghesquiere J, Weymans M, Amery A. Ventilatory threshold
measurement to evaluate maximal endurance performance. Int J Sports Med. 1986
Feb;7(l):26-9.
B-151
-------
132. Rivera-Brown, A.M. and Frontera, W.R. Achievement of plateau and reliability of
VO2max in active adolescents: Effects of testing ergometer. Fed. Exer. Sci., 10:164-175,
1998.
133. Rivera-Brown, A.M., Alvarez ,A., Rodriguez-Santana, J., Benetti, P. Anaerobic
power and achievement of VO2 plateau in pre-pubertal boys. Int. J. Sports Med. 22:111-
115,2001.
134. Rivera-Brown, AM, Rivera, MA, Frontera, WR (1992) Applicability of criteria
for VO2max in active children Pediatr Exerc Sci 4,331-339.
135. Rogers DM, Olson BL, Wilmore JH. Scaling for the VO2-to-body size
relationship among children and adults. J Appl Physiol. 1995 Sep;79(3):958-67.
136. Rogers, DM, Turley, KR, Kujawa, KI, Harper, KM, Wilmore, J. (1995)
Allometric scaling factors for oxygen uptake during exercise in children Pediatr Exerc
Sci 7,12-25.
137. Rowland T, Kline G, Goff D, Martel L, and Ferrone L. Physiological
determinants of maximal aerobic power in healthy 12-year-old boys. Fed Exer Sci
11:317-326, 1999.
138. Rowland TW, Green GM. Physiological responses to treadmill exercise in
females: adult-child differences. Med Sci Sports Exerc. 1988 Oct;20(5):474-8.
139. Rowland TW, Martel L, Vanderburgh P, Manos T, Charkoudian N. The
influence of short-term aerobic training on blood lipids in healthy 10-12 year old
children. Int J Sports Med. 1996 Oct;17(7):487-92.
140. Rowland, T, Miller K, Vanderburgh P, Goff D, Martel L, and Ferrone L.
Cardiovascular fitness in premenarcheal girls and young women. Int J Sports Med 20: 1-
5, 1999.
141. Rutenfranz J, Macek M, Lange Andersen K, Bell RD, Vavra J, Radvansky J,
Klimmer F, Kylian H. The relationship between changing body height and growth
related changes in maximal aerobic power. Eur J Appl Physiol Occup Physiol.
1990;60(4):282-7.
142. Sady, S. Transient oxygen uptake and heart rate responses at the onset of relative
endurance exercise in prepubertal boys and adult men. Int. J. Sports Med. 2: 240-244,
1981.
143. Sargent 1985 (Ref??) From T. McCurdy spreadsheet, dated Nov 15, 2000.
144. Saris WH, Noordeloos AM, Ringnalda BEM, et al. Binkhorst RA, Kemper HCG,
Saris WHM, editors. Children and Exercise XL Champaign: Human Kinetics Publisher;
1985; Reference values for aerobic power of healthy 4- to 18-year-old Dutch children:
Preliminary results, p. 151-60.
B-152
-------
145. Savage MP, Petratis MM, Thomson WH, Berg K, Smith JL, Sady SP. Exercise
training effects on serum lipids of prepubescent boys and adult men. Med Sci Sports
Exerc. 1986 Apr;18(2):197-204.
146. Shuleva K, Hunter G, Hester D, Dunaway D. Exercise oxygen uptake in 2 to 6
year olds. PedExer Sci 2:130-139, 1990.
147. Sobolova V, Seliger V, Grussova D, Machovcova J, Zelenka V. The influence of
age and sports training in swimming on physical fitness. Acta Paediatr Scand Suppl.
1971;217:63-7.
148. Spurr, G.B. & Reina, J.C. (1990). Daily pattern of %VO2 max and heart rates in
normal and undernourished school children. Med. Sci. Sports Exerc. 22:643-652.
149. Stewart KJ, Gutin B. Effects of physical training on cardiorespiratory fitness in
children. Res Q. 1976 Mar;47(l): 110-20.
150. Stoedefalke K, Armstrong N, Kirby BJ, Welsman JR. Effect of training on peak
oxygen uptake and blood lipids in 13 to 14-year-old girls. Acta Paediatr. 2000
Nov;89(ll): 1290-4.
151. Sundberg S, Elovainio R. Cardiorespiratory function in competitive endurance
runners aged 12-16 years compared with ordinary boys. Acta Paediatr Scand. 1982
Nov;71(6):987-92.
152. Sunnegardh J, Bratteby LE. Maximal oxygen uptake, anthropometry and physical
activity in a randomly selected sample of 8 and 13 year old children in Sweden. Eur J
Appl Physiol Occup Physiol. 1987;56(3):266-72.
153. Tolfrey K, Campbell IG, Batterham AM. Exercise training induced alterations in
prepubertal children's lipid-lipoprotein profile. Med Sci Sports Exerc. 1998
Dec;30(12): 1684-92.
154. W Bell (1997) Aerobic fitness and fat patterning in adolescent boys. In: Children
and Exercise (Eds N Armstrong, B Kirby, J Welsman) pp 63-68. London: E & F Spon.
155. W.D. McArdle et al. (1972) Reliability and interrelationships between maximal
oxygen uptake, physical work capacity and step test scores in college women. Medicine
and Science in Sports, Vol 4, pi82-186.
156. Washington RL, van Gundy JC, Cohen C, Sondheimer HM, Wolfe RR. Normal
aerobic and anaerobic exercise data for North American school-age children. J Pediatr.
1988Feb;112(2):223-33.
157. Welsman JR, Armstrong N, Nevill AM, Winter EM, Kirby BJ. Scaling peak VO2
for differences in body size. Med Sci Sports Exerc. 1996 Feb;28(2):259-65.
B-153
-------
158. Welsman JR, Armstrong N, Withers S. Responses of young girls to two modes of
aerobic training. Br J Sports Med. 1997 Jun;31(2):139-42.
159. Welsman JR, Armstrong N. Daily physical activity and blood lactate indices of
aerobic fitness in children. Br J Sports Med. 1992 Dec;26(4):228-32.
160. Williams CA, Armstrong N, Powell J. Aerobic responses of prepubertal boys to
two modes of training. Br J Sports Med. 2000 Jun;34(3): 168-73.
161. Williford HN, Blessing DL, Duey WJ, Barksdale JM, Wang N, Olson MS, Teel S.
Exercise training in black adolescents: changes in blood lipids and Vo2max. Ethn Dis.
1996 Autumn;6(3-4):279-85.
162. Wilmore JH, Sigerseth PO. Physical work capacity of young girls, 7-13 years of
age. J Appl Physiol. 1967 May;22(5):923-8.
163. Yamaji K, Miyashita M, Shephard RJ. Relationship between heart rate and
relative oxygen intake in male subjects aged 10 to 27 years. J Hum Ergol (Tokyo). 1978
Sep;7(l):29-39.
164. Yoshizawa S, Ishizaki T, and Honda H. Aerobic power and endurance running
in young children. In: Rutenfranz J, Mocellin R, Klimt F, eds. children and exercise XII.
International series on sport sciences, Vol 16. Champaign, IL: Human Kinetics,
1986:227-232.
165. Yoshizawa S, Honda H, NakamuraN. Effects of an 18-month endurance run
program on maximal aerobic power in 4- to 6- year old girls. Pediatr Exer Sci 9:33-43,
1997.
166. Yoshizawa S, Honda H, Urushibara M, Nakamura N. Effects of endurance run on
circulorespiratory system in young children. J Hum Ergol (Tokyo). 1990 Jun;19(l):41-
52.
167. Yoshizawa S, Ishizaki T, Honda H. Physical fitness of children aged 5 and 6
years. J Hum Ergol (Tokyo). 1977 Sep;6(l):41-51.
168. Allen TH, Peng MT, et al. Prediction of blood volume and adiposity in man from
body weight and cube of height. Metabolism. 1956; 5: 328.
169. Nadler SB, Hidalgo JU, Bloch T. Prediction of blood volume in normal human
adults. Surgery. 1962;51:224-232.
170. Linderkamp O, Versmold HT, et al. Estimation and prediction of blood volume in
infants and children. Europ J Pediat. 1977; 125: 227-234.
171. Cropp GJA. Changes in blood and plasma volumes during growth. J Pediatrics.
1971; 78: 220-229.
B-154
-------
172. Smetannikov Y, Hopkins D. Intraoperative bleeding: A mathematical model for
minimizing hemoglobin loss. Transfusion. 1996; 36: 832-835.
B-155
-------
- cllx ,.. ,'.•-.,'- '. '. L
Data weight;
merge Demo Demo_b Demo_c Bmx Bmx_b_r Bmx_c;
by SEQN;
mass=BMXWT;
gen=RIAGENDR;
ageyrs=RIDAGEYR;
agemonths=RIDAGEEX;
wt = (2/3)*WTMEC4YR;
if (SEQN>21004) THEN wt=(1/3)*WTMEC2YR;
if agemonths<12 and agemonths>0 THEN ageyrs=0,
keep SEQN mass gen ageyrs agemonths wt;
run;
proc sort data=weight;
by gen ageyrs;
run;
Proc univariate data=weight;
by gen ageyrs;
var mass;
freg wt;
histogram mass / lognormal;
run;
APPENDIX B. SAS CODE FOR ESTIMATING THE
NORMALIZED VO2MAX DISTRIBUTIONS
B-156
-------
Data alldata ;
infile 'H:\kki-05-PHYSIOLOGY_10\NVO2MAX\vo2max.csv' DLM="," END=eof;
input age nvo2max gender;
output alldata;
proc sort data=alldata;
by gender age;
run;
Proc univariate data=alldata;
by gender age;
var nvo2max;
histogram nvo2max / normal;
output out=outputdatal N=samplesize mean=Mean
std=StdDeviation ProbN=NormalFit;
run;
Proc export data=outputdatal outfile="H:\kki-05-PHYSIOLOGY_10\Alldata_vo2max.cs\
replace;
run;
APPENDIX C. SAS CODE FOR ESTIMATING THE
HEMOGLOBIN CONTENT DATA
Data Hb;
merge Demo Lab25 Demo b L25 b;
by SEQN;
Hb=LBXHGB;
gen=RIAGENDR;
ageyrs=RIDAGEYR;
agemonths=RIDAGEEX;
wt = WTMEC4YR;
if agemonths<' '-' and agemonthsX: THEN ageyra= ;
if ageyrsX'".: then ageyrs= ( floor ( ageyr? / )+ )*
B-157
-------
keep SEQN Hb gen ageyrs agemonths wt;
run;
proc sort data=Hb;
by gen ageyrs;
run;
Proc univariate data=Hb;
by gen ageyrs;
var Hb;
reg wt; * Apply sample weights;
histogram Hb / normal; ^ Fit to Normal;
output out=outputs N=samplesize mean=Mean
std=StdDeviation ProbN=NormalFit;
run;
Proc export data=outputs outfile="H:\kki-05-PHYSIOLOGY_10\Hemoglobin\HbFitswt.
replace;
run;
B-158
-------
APPENDIX D. THE NEW PHYSIOLOGY.TXT FILE
Note: The values contained in the file conform to the current APEX read formats. That
is, the number of decimal places for each parameter is dictated by the APEX code. It is
likely that this will change in the future, at which point more significant digits could be
added to the Physiology.txt file.
Males
age 0-100, then females age
NVO2max distribution
0-100 (last revised 12-20-05)
Age
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
Source Distr
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Mean
48
48
48
49
49
49
50
50
50
51
51
51
52
52
52
53
53
53
53
54
54
54
53
52
51
51
50
49
48
48
47
46
46
45
44
44
43
42
42
41
40
40
39
39
38
38
37
36
36
35
35
34
34
34
33
33
32
32
31
31
31
30
30
30
.3
.6
.9
.2
.5
.8
.1
.4
.8
.1
.4
.7
.0
.3
.6
.0
.3
.6
.9
.2
.5
.2
.4
.6
.8
.1
.3
.6
.8
.1
.4
.7
.0
.3
.6
.0
.3
.7
.1
.4
.8
.2
.7
.1
.5
.0
.4
.9
.4
.9
.4
.9
.5
.0
.6
.1
.7
.3
.9
.5
.1
.7
.4
.0
SD
1.7
2 .0
2 .4
2 .7
3 .0
3 .3
3 .7
4 .0
4 .3
4 .6
5.0
5.3
5.6
5.9
6 .2
6 .6
6 .9
7.2
7.5
7.9
8 .2
8 .5
8 .8
9.2
9.5
9.8
10.7
10.5
10.3
10.1
9.9
9.7
9.6
9.4
9.2
9.0
8 .9
8 .7
8 .6
7.3
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
4 .9
4 .9
4 .9
4 .9
4 .9
4 .9
4 .9
4 .9
4 .9
4 .9
5.3
5.3
5.3
5.3
Lower
44
43
43
43
42
42
41
41
40
40
39
39
39
38
38
37
37
36
36
35
35
34
32
31
29
28
25
25
24
24
24
24
23
23
23
23
22
22
22
25
28
28
28
28
28
28
28
28
28
28
23
23
23
23
23
23
23
23
23
23
21
21
21
21
.3
.8
.4
.0
.5
.1
.6
.2
.8
.3
.9
.4
.0
.6
.1
.7
.3
.8
.4
.9
.5
.5
.9
.4
.8
.3
.5
.2
.9
.6
.3
.0
.8
.5
.2
.0
.7
.4
.2
.5
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.0
.0
.0
.0
Upper
52
53
54
55
56
57
58
59
60
61
62
64
65
66
67
68
69
70
71
72
73
74
74
73
73
73
75
74
72
71
70
69
68
67
66
65
64
62
61
54
50
50
50
50
50
50
50
50
50
50
42
42
42
42
42
42
42
42
42
42
41
41
41
41
.2
.3
.4
.4
.5
.6
.6
.7
.8
.8
.9
.0
.0
.1
.2
.2
.3
.4
.4
.5
.6
.0
.0
.9
.9
.9
.2
.0
.8
.6
.4
.3
.2
.1
.0
.0
.0
.9
.9
.1
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.8
.8
.8
.8
Assumptions
B-159
-------
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
29
29
29
28
28
28
27
27
27
27
27
26
26
26
26
26
25
25
25
25
25
25
25
25
25
24
24
24
24
24
24
24
24
24
25
25
25
35
36
36
36
37
37
37
38
38
38
39
39
39
40
40
40
41
41
41
42
42
42
41
40
40
39
39
38
37
37
36
36
35
34
34
33
33
32
32
31
31
31
30
30
.7
.4
.1
.8
.5
.2
.9
.7
.4
.2
.0
.7
.5
.4
.2
.0
.8
.7
.6
.4
.3
.2
.1
.1
.0
.9
.9
.9
.8
.8
.8
.8
.9
.9
.0
.0
.1
.9
.2
.5
.9
.2
.5
.9
.2
.5
.9
.2
.5
.9
.2
.5
.9
.2
.5
.8
.2
.5
.1
.5
.8
.2
.6
.0
.4
.8
.2
.6
.0
.5
.9
.4
.9
.4
.9
.4
.9
.4
.0
.5
.1
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
7
7
7
7
7
7
6
6
6
6
6
6
6
5
5
.3
.3
.3
.3
.3
.3
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.9
.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
.0
.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
.0
.1
.2
.3
.4
.4
.1
.9
.7
.6
.4
.2
.0
.8
.7
.5
.4
.2
.1
.0
.8
.7
21
21
21
21
21
21
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
22
22
22
22
22
22
22
22
23
23
23
23
23
23
23
23
23
24
24
24
24
23
22
22
21
20
19
18
18
18
18
18
18
18
18
18
17
17
17
17
17
17
17
16
.0
.0
.0
.0
.0
.0
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.2
.3
.4
.5
.6
.7
.8
.9
.0
.1
.3
.4
.5
.6
.7
.8
.9
.0
.1
.2
.4
.7
.9
.0
.1
.3
.5
.9
.8
.7
.6
.5
.4
.2
.1
.0
.8
.7
.6
.4
.3
.1
.0
.8
41
41
41
41
41
41
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
38
49
50
50
51
51
52
52
53
54
54
55
55
56
56
57
57
58
59
59
60
60
60
60
59
59
58
58
57
56
55
54
53
52
51
50
49
48
48
47
46
45
44
44
43
.8
.8
.8
.8
.8
.8
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.6
.2
.7
.3
.8
.4
.9
.5
.0
.6
.1
.7
.2
.8
.3
.9
.5
.0
.6
.1
.7
.5
.1
.6
.2
.8
.4
.8
.7
.6
.6
.6
.6
.7
.7
.8
.9
.0
.2
.4
.6
.8
.0
.3
B-160
-------
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Males
Age
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
age 0-100
Body
Source
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
29.
29.
28 .
28 .
28 .
27.
27.
26 .
26 .
26 .
25.
25.
25.
24 .
24 .
24 .
24 .
23 .
23 .
23 .
23 .
22 .
22 .
22 .
22 .
21.
21.
21.
21.
21.
21.
21.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
, then females
6
2
8
4
0
6
2
8
5
1
8
5
2
9
6
3
0
7
5
2
0
7
5
3
1
9
7
6
4
3
1
0
9
8
7
6
5
4
3
3
3
2
2
2
2
2
2
2
3
3
4
4
5
6
7
8
9
age
mass distribution,
Distr
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
GM
7.8
11.
13 .
16 .
18 .
21.
23 .
27.
31.
34 .
38 .
44 .
48 .
55.
62 .
67.
72 .
73 .
75.
77.
78 .
4
9
0
5
6
1
1
7
7
3
1
0
4
8
7
5
1
1
2
0
5.
5.
5.
5.
5.
5.
5.
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
4 .
0-
kg
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
6
5
4
3
2
1
0
9
8
8
7
7
6
6
5
5
5
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
100
GSD
301
143
146
154
165
234
213
216
302
265
280
308
315
340
293
255
267
248
243
245
250
16 .6
16 .5
16 .3
16 .1
16 .0
15.8
15.6
15.4
15.2
15.1
14 .9
14 .7
14 .5
14 .3
14 .1
13 .9
13 .6
13 .4
13 .2
13 .0
12 .8
12 .5
12 .3
12 .1
11.9
11.7
11.5
11.4
11.2
11.1
10.9
10.8
10.7
10.6
10.4
10.4
10.3
10.2
10.1
10.1
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.1
10.1
10.2
10.2
10.3
10.4
10.5
10.6
10.7
(last
42 .6
41.9
41.2
40.6
40.0
39.4
38 .8
38 .3
37 .7
37.2
36 .7
36 .3
35.9
35.4
35.1
34 .7
34 .3
34 .0
33 .7
33 .4
33 .2
33 .0
32 .7
32 .5
32 .3
32 .1
32 .0
31.8
31.6
31.5
31.3
31.2
31.1
31.0
30.9
30.8
30.7
30.6
30.6
30.5
30.5
30.4
30.4
30.4
30.4
30.4
30.4
30.4
30.5
30.5
30.6
30.6
30.7
30.8
30.9
31.0
31.1
revised 12-20-05)
Lower Upper Assumptions
3 .6
8 .2
9.8
11.7
11.1
13 .7
16 .1
19.3
19.1
24 .0
24 .3
26 .2
27 .7
27.7
35.7
41.5
45.8
49.9
51.2
52 .6
50.5
11.8
16 .1
20.9
23 .7
28 .1
42 .4
41.1
46 .8
66 .2
69.9
72 .9
83 .8
94 .8
106 .6
121.0
117.9
139.1
136 .6
144 .2
134 .5
130.0
B-161
-------
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
0
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
78
83
80
81
84
81
85
84
82
81
81
84
88
81
87
83
85
84
84
90
87
88
88
88
87
88
86
84
86
84
88
89
89
90
88
84
87
85
84
87
89
84
89
90
89
86
86
85
87
82
79
82
85
83
84
78
79
79
77
79
75
76
74
75
71
74
73
72
72
71
70
70
69
69
68
67
67
66
65
65
7 .
.2
.8
.6
.7
.8
.8
.2
.3
.1
.6
.3
.7
.2
.2
.2
.4
.8
.1
.6
.1
.4
.3
.4
.5
.1
.2
.5
.8
.2
.7
.0
.9
.0
.1
.3
.8
.5
.1
.2
.0
.0
.8
.1
.0
.9
.8
.2
.2
.1
.8
.6
.0
.6
.0
.5
.7
.4
.9
.6
.9
.4
.8
.6
.3
.8
.0
.4
.7
.1
.5
.9
.3
.6
.0
.4
.8
.1
.5
.9
.3
4
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.297
.292
.222
.251
.206
.273
.249
.272
.236
.262
.249
.235
.231
.221
.251
.228
.241
.260
.196
.246
.173
.205
.233
.200
.205
.243
.229
.186
.240
.179
.208
.216
.228
.216
.222
.195
.253
.266
.182
.232
.207
.228
.262
.193
.215
.228
.207
.191
.222
.210
.240
.204
.196
.217
.185
.207
.170
.195
.155
.174
.157
.180
.158
.205
.191
.170
.170
.160
.160
.160
.160
.160
.150
.150
.150
.150
.140
.140
.140
.140
.304
46
53
50
50
50
48
50
51
50
52
48
49
64
53
61
45
59
52
61
58
61
62
54
56
60
54
49
56
47
53
57
55
58
64
55
45
58
51
58
57
49
56
56
59
58
54
43
61
50
46
51
51
56
53
56
55
58
41
56
56
55
54
53
41
46
50
50
50
50
49
49
49
49
49
48
48
48
48
48
47
3 .
.8
.3
.5
.6
.2
.9
.0
.0
.6
.5
.8
.7
.8
.1
.0
.8
.3
.8
.2
.5
.3
.2
.0
.6
.6
.2
.9
.3
.0
.4
.9
.2
.2
.1
.1
.0
.3
.6
.7
.3
.9
.0
.3
.1
.1
.0
.1
.2
.7
.5
.0
.9
.2
.3
.5
.9
.7
.1
.4
.0
.8
.4
.2
.5
.9
.6
.4
.2
.0
.8
.6
.4
.3
.1
.9
.7
.5
.3
.1
.9
7
199.2
155.4
137.6
132 .6
136 .1
164 .5
153 .9
167.2
147.2
139.0
170.6
135.8
146 .3
136 .9
193 .3
140.5
150.9
149.7
140.6
154 .0
117.7
144 .0
145.3
128 .9
160.2
154 .3
188 .3
128 .3
171.3
124 .4
143 .6
144 .9
143 .3
155.2
138 .6
110.3
160.0
179.0
112 .4
141.7
162 .8
152 .1
171.6
119.0
126 .3
150.1
127.5
163 .2
127.2
125.5
122 .8
132 .7
128 .3
120.0
133 .5
121.1
109.3
115.1
107.8
111.9
111.9
111.8
107.0
109.5
105.8
101.1
99.1
97.2
95.2
93 .2
91.3
89.3
87.4
85.4
83 .4
81.5
79.5
77 .6
75.6
73 .6
12 .1
B-162
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
11
13
15
18
20
22
26
30
35
40
46
50
56
57
60
61
61
64
66
67
67
66
69
70
66
73
70
74
69
70
73
72
72
69
73
73
70
75
72
72
73
73
73
75
76
77
72
74
72
75
72
74
74
72
76
77
72
74
80
75
77
73
72
75
72
73
75
73
74
69
69
69
71
70
70
69
70
66
67
62
65
.1
.3
.6
.0
.4
.5
.5
.5
.2
.6
.6
.7
.6
.2
.1
.6
.2
.6
.2
.0
.2
.8
.7
.3
.3
.0
.6
.4
.1
.6
.0
.9
.7
.8
.0
.5
.0
.6
.3
.9
.4
.7
.4
.7
.8
.5
.8
.6
.8
.2
.9
.5
.7
.4
.0
.3
.4
.5
.6
.8
.1
.3
.3
.4
.9
.1
.8
.2
.4
.0
.1
.9
.4
.4
.5
.5
.1
.4
.8
.2
.4
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.163
.158
.160
.171
.229
.194
.239
.315
.271
.304
.302
.274
.275
.248
.249
.255
.248
.281
.274
.262
.262
.273
.304
.289
.283
.281
.281
.312
.250
.305
.278
.281
.307
.230
.306
.289
.284
.295
.251
.289
.268
.270
.314
.266
.308
.304
.298
.303
.261
.292
.240
.283
.259
.281
.231
.315
.252
.267
.277
.260
.240
.198
.238
.281
.254
.242
.266
.250
.225
.188
.232
.240
.240
.277
.216
.199
.240
.211
.200
.255
.184
7 .
10
11
12
12
15
16
19
20
22
27
27
33
37
34
40
41
42
41
41
39
42
40
47
44
45
41
44
39
42
43
41
44
46
44
44
48
43
41
45
50
47
45
49
41
46
47
44
45
48
42
45
46
44
53
45
48
45
50
51
50
49
46
41
35
48
47
39
48
45
45
40
47
37
46
48
40
44
46
41
42
4
.1
.0
.8
.6
.9
.9
.8
.3
.7
.7
.8
.4
.7
.9
.9
.5
.4
.6
.5
.7
.0
.3
.5
.8
.3
.4
.3
.3
.1
.7
.5
.9
.6
.2
.6
.1
.7
.6
.5
.5
.1
.6
.5
.6
.6
.8
.2
.1
.4
.5
.7
.2
.3
.6
.6
.6
.0
.9
.3
.7
.7
.9
.1
.9
.4
.2
.3
.0
.9
.5
.7
.4
.4
.8
.8
.3
.1
.2
.2
.7
15.3
20.4
27.9
29.1
40.4
36 .7
51.0
60.8
58 .6
71.2
84 .6
93 .3
99.5
110.0
108 .4
113 .8
133 .1
123 .6
118 .5
122 .6
123 .7
123 .5
143 .0
144 .5
131.8
128 .9
140.9
142 .1
116 .3
151.5
125.9
139.7
135.2
115.3
138 .4
150.1
152 .1
151.7
123 .1
137.4
156 .9
146 .1
159.5
153 .0
141.5
145.8
130.6
166 .0
125.5
175.7
120.2
146 .6
176 .6
123 .1
125.6
134 .9
122 .6
117.7
133 .0
128 .3
125.6
121.1
119.9
132 .5
113 .7
113 .3
123 .8
120.7
118 .0
102 .8
108 .1
103 .8
127.6
106 .4
117.4
101.7
119.8
109.8
98 .4
121.4
91.4
B-163
-------
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Males
Age
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
CDC
age 0-100
Source
R47g
R47g
R47g
R47h
R47h
R47h
R47h
R47h
R47h
R47h
R47i
R47i
R47i
R47i
R47i
R47i
R47i
R47i
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47J
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
LN
64 .8
62 .9
62 .2
61.5
62 .4
61.8
61.3
60.7
60.2
59.6
59.1
58 .5
58 .0
57.4
56 .9
56 .3
55.8
55.2
54 .7
then females age
Regression
DV
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
1.260
1.196
1.216
1.209
1.210
1.210
1.210
1.210
1.210
1.200
1.200
1.200
1.200
1.200
1.200
1.200
1.190
1.190
1.190
0-100
40.6
44 .7
43 .5
42 .3
41.9
41.7
41.5
41.3
41.1
40.9
40.7
40.5
40.3
40.1
39.9
39.7
39.5
39.3
39.1
120.0
101.2
108 .4
93 .2
101.2
100.3
99.4
98 .4
97.5
96 .6
95.7
94 .8
93 .9
93 .0
92 .1
91.2
90.3
89.4
88 .5
(last revised 6-
equation Estimate
IV
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
Slope
0.244
0.244
0.244
0.095
0.095
0.095
0.095
0.095
0.095
0.095
0.074
0.074
0.074
0.074
0.074
0.074
0.074
0.074
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
for
RMR
Interc
-0
-0
-0
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
.127
.127
.127
110
110
110
110
110
110
110
754
754
754
754
754
754
754
754
896
896
896
896
896
896
896
896
896
896
896
896
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
653
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
SE
290
290
280
280
280
280
280
280
280
280
440
440
440
440
440
440
440
440
640
640
640
640
640
640
640
640
640
640
640
640
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
700
11-98)
Units med.
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
wgt
2
2
3
3
3
4
4
4
4
5
5
5
6
6
6
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
.1
.7
.2
.6
.8
.0
.3
.5
.8
.0
.4
.7
.0
.3
.9
.2
.7
.6
.3
.4
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
B-164
-------
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R47k
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R471
R47a
R47a
R47a
R47b
R47b
R47b
R47b
R47b
R47b
R47b
R47c
R47C
R47C
R47C
R47C
R47C
R47C
R47C
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47d
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
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
.048
.048
.048
.048
.048
.048
.048
.048
.048
.048
.048
.048
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.049
.244
.244
.244
.085
.085
.085
.085
.085
.085
.085
.056
.056
.056
.056
.056
.056
.056
.056
.062
.062
.062
.062
.062
.062
.062
.062
.062
.062
.062
.062
.034
.034
.034
.034
.034
.034
.034
.034
.034
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
-
-
-
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
.653
.653
.653
.653
.653
.653
.653
.653
.653
.653
.653
.653
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
.459
0.130
0.130
0.130
.033
.033
.033
.033
.033
.033
.033
.898
.898
.898
.898
.898
.898
.898
.898
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.538
.538
.538
.538
.538
.538
.538
.538
.538
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
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
.700
.700
.700
.700
.700
.700
.700
.700
.700
.700
.700
.700
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.690
.250
.250
.250
.290
.290
.290
.290
.290
.290
.290
.470
.470
.470
.470
.470
.470
.470
.470
.500
.500
.500
.500
.500
.500
.500
.500
.500
.500
.500
.500
.470
.470
.470
.470
.470
.470
.470
.470
.470
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
7
7
7
7
7
7
7
7
7
7
7
7
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
2
2
3
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
5
5
6
6
6
6
6
6
6
6
6
6
5
5
5
5
5
5
5
5
5
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.0
.5
.0
.3
.5
.7
.9
.1
.4
.7
.9
.2
.5
.7
.9
.0
.1
.2
.7
.8
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.7
.7
.7
.7
.7
.7
.7
.7
.7
B-165
-------
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Males
Blood
Age
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47e
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
R47f
age 0-100
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BMR
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
BM
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
then females age 0-
Volume factor and
BLDFAC
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
17.0
HGMN
11.
12 .
12 .
12 .
12 .
13 .
13 .
13 .
13 .
13 .
13 .
13 .
14 .
14 .
14 .
15.
9
2
4
7
8
0
2
5
4
6
6
7
0
3
7
1
Hemoglobin
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
034
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
038
100
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
3 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
2 .
(HG
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
538
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
755
last
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.470
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
.450
revised
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
MJ/day
12-20-05)
content
HGSTD
1
1
0
0
0
0
0
0
0
1
0
0
1
1
1
1
.0
.0
.8
.8
.8
.9
.9
.8
.8
.0
.9
.7
.0
.0
.0
.0
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
B-166
-------
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
17
17
17
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
.0
.0
.0
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
13
13
13
13
13
13
.4
.5
.7
.8
.8
.7
.7
.7
.7
.7
.7
.7
.7
.7
.7
.6
.6
.6
.6
.6
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.4
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.1
.1
.1
.1
.1
.0
.0
.0
.0
.0
.7
.7
.7
.7
.7
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.0
.0
.0
.0
.0
.8
.8
.8
.8
.8
.5
1
1
1
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.0
.0
.0
.8
.9
.9
.9
.9
.9
.9
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.0
.0
.0
.0
.0
.2
.2
.2
.2
.2
.2
.2
.2
.2
.2
.4
.4
.4
.4
.4
.5
.5
.5
.5
.5
.4
.4
.4
.4
.4
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
B-167
-------
97
98
99
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
20
20
20
20
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
14
.4
.4
.4
.4
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
.6
13
13
13
13
12
12
12
12
12
12
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
.5
.5
.5
.5
.2
.3
.6
.5
.8
.9
.0
.1
.3
.4
.6
.5
.6
.5
.6
.5
.5
.5
.5
.4
.5
.5
.5
.5
.5
.5
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.5
.5
.5
.5
.5
.5
.5
.5
.5
.5
.6
.6
.6
.6
.6
.7
.7
.7
.7
.7
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
.8
1
1
1
1
0
0
0
1
0
1
0
0
0
0
1
0
0
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
.8
.8
.8
.8
.7
.7
.8
.0
.8
.0
.8
.8
.8
.8
.0
.9
.9
.0
.0
.9
.1
.1
.2
.1
.1
.2
.2
.2
.2
.2
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.2
.2
.2
.2
.2
.3
.3
.3
.3
.3
.2
.2
.2
.2
.2
.1
.1
.1
.1
.1
.2
.2
.2
.2
.2
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.1
.3
B-168
-------
77
78
79
80
81
82
83
84
85
86
87
90
91
92
93
94
95
96
97
98
99
100
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
14 .6
13 .8
13 .8
13 .8
13 .8
13 .6
13 .6
13 .6
13 .6
13 .6
13 .4
13 .4
13 .4
13 .4
13 .4
13 .2
13 .2
13 .2
13 .2
13 .2
13 .0
13 .0
13 .0
13 .0
13 .0
1.3
1.3
1.3
1.3
1.2
1.2
1.2
1.2
1.2
1.6
B-169
-------
Table 7. Nv02max Values for Males: Raw and Smoothed Fits.
Age
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00
17.00
18.00
19.00
20.00
21.00
22.00
23.00
24.00
25.00
26.00
27.00
28.00
29.00
30.00
31.00
32.00
33.00
34.00
35.00
36.00
37.00
38.00
MALES
MEAN
Raw
Fit Values
51.37
53.46
51.10
51.28
50.13
50.70
52.74
52.93
53.18
49.46
49.77
51.98
59.88
56.80
54.60
54.61
53.76
57.23
50.90
50.06
46.38
48.32
51.02
45.59
45.86
46.90
42.08
44.48
38.63
42.63
40.41
39.70
MEAN
Smoothed
Fit Values
48.25
48.56
48.88
49.19
49.50
49.82
50.13
50.44
50.76
51.07
51.39
51.70
52.01
52.33
52.64
52.95
53.27
53.58
53.90
54.21
54.52
54.23
53.42
52.63
51.84
51.07
50.31
49.56
48.82
48.10
47.38
46.67
45.98
45.30
44.63
43.97
43.32
42.68
42.05
SD
Raw
Fit Values
2.86
2.86
6.26
5.87
6.04
7.13
5.13
4.72
5.57
6.06
6.93
7.48
9.65
9.31
8.17
8.40
9.60
10.44
10.63
9.66
8.95
10.47
12.31
9.91
10.14
11.03
9.08
8.95
10.10
7.11
8.81
6.22
SD
Smoothed
Fit Values
1.71
2.04
2.36
2.68
3.01
3.33
3.65
3.98
4.30
4.62
4.95
5.27
5.59
5.92
6.24
6.56
6.89
7.21
7.53
7.86
8.18
8.50
8.83
9.15
9.47
9.80
10.69
10.49
10.29
10.10
9.92
9.73
9.55
9.38
9.20
9.03
8.87
8.71
8.55
MIN
(IstPctl)
44.26
43.82
43.39
42.95
42.51
42.07
41.63
41.19
40.76
40.32
39.88
39.44
39.00
38.56
38.13
37.69
37.25
36.81
36.37
35.93
35.50
34.45
32.89
31.35
29.81
28.29
25.45
25.16
24.88
24.60
24.32
24.04
23.76
23.49
23.22
22.95
22.69
22.42
22.16
MAX
(99th Pctl)
52.24
53.30
54.37
55.43
56.50
57.56
58.63
59.70
60.76
61.83
62.89
63.96
65.02
66.09
67.16
68.22
69.29
70.35
71.42
72.48
73.55
74.01
73.95
73.91
73.88
73.86
75.17
73.96
72.77
71.59
70.44
69.31
68.20
67.10
66.03
64.98
63.95
62.94
61.94
B-170
-------
39.00
40.00
41.00
42.00
43.00
44.00
45.00
46.00
47.00
48.00
49.00
50.00
51.00
52.00
53.00
54.00
55.00
56.00
57.00
58.00
59.00
60.00
61.00
62.00
63.00
64.00
65.00
66.00
67.00
68.00
69.00
70.00
71.00
72.00
73.00
74.00
75.00
76.00
77.00
78.00
79.00
80.00
81.00
82.00
83.00
84.00
85.00
86.00
87.00
88.00
40.62
39.02
39.72
35.58
39.98
38.65
40.15
40.67
41.51
38.92
34.65
33.85
32.52
36.31
36.23
33.91
33.40
31.68
32.47
33.24
33.05
29.02
31.68
29.72
30.90
30.65
29.86
28.60
29.47
28.95
31.13
27.12
28.56
27.62
27.84
25.05
23.74
23.68
41.44
40.83
40.24
39.66
39.09
38.53
37.98
37.44
36.92
36.40
35.90
35.41
34.92
34.45
34.00
33.55
33.11
32.69
32.27
31.87
31.48
31.10
30.73
30.37
30.02
29.69
29.36
29.05
28.75
28.46
28.18
27.91
27.65
27.41
27.17
26.95
26.74
26.54
26.35
26.17
26.00
25.84
25.70
25.57
25.44
25.33
25.23
25.14
25.06
25.00
8.01
8.28
9.96
9.85
6.46
7.60
6.59
7.89
9.68
10.52
7.68
6.49
4.51
7.08
7.31
5.29
5.08
6.52
6.33
6.32
6.45
3.59
6.95
5.09
8.06
5.32
6.90
5.51
5.25
5.63
6.43
3.44
5.71
5.03
6.27
6.68
4.99
5.88
8.40
8.25
8.10
7.96
7.82
7.69
7.56
7.43
7.31
7.19
7.07
6.96
6.86
6.75
6.65
6.56
6.46
6.37
6.29
6.21
6.13
6.06
5.99
5.92
5.86
5.80
5.75
5.70
5.65
5.61
5.57
5.53
5.50
5.47
5.45
5.43
5.41
5.40
5.39
5.38
5.38
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
21.90
21.64
21.39
21.14
20.89
20.64
20.40
20.16
19.91
19.68
19.44
19.21
18.98
18.75
18.52
18.30
18.08
17.86
17.64
17.43
17.22
17.01
16.80
16.60
16.40
16.20
16.00
15.80
15.61
15.42
15.23
15.05
14.86
14.68
14.50
14.33
14.15
13.98
13.81
13.65
13.48
13.32
13.17
13.04
12.92
12.81
12.70
12.62
12.54
12.47
60.97
60.02
59.09
58.18
57.28
56.41
55.56
54.73
53.92
53.12
52.35
51.60
50.87
50.16
49.47
48.79
48.14
47.51
46.90
46.31
45.74
45.19
44.66
44.14
43.65
43.18
42.73
42.30
41.89
41.50
41.13
40.78
40.45
40.13
39.84
39.57
39.32
39.09
38.88
38.69
38.52
38.37
38.22
38.09
37.97
37.86
37.76
37.67
37.59
37.52
B-171
-------
89.00
90.00
91.00
92.00
93.00
94.00
95.00
96.00
97.00
98.00
99.00
100.00
24.94
24.90
24.86
24.84
24.83
24.83
24.84
24.87
24.90
24.95
25.00
25.07
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
5.39
12.42
12.37
12.34
12.32
12.31
12.31
12.32
12.34
12.37
12.42
12.48
12.54
37.47
37.42
37.39
37.37
37.36
37.36
37.37
37.39
37.43
37.47
37.53
37.60
Table 8. Nv02max Values for Females: Raw and Smoothed Fits
Age
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00
17.00
18.00
19.00
20.00
21.00
22.00
23.00
24.00
25.00
26.00
27.00
FEMALES
MEAN
Raw
Fit Values
30.56
45.53
43.88
43.03
42.00
37.57
39.57
35.51
38.22
45.67
43.87
42.52
43.45
43.22
43.87
41.14
38.20
38.98
34.94
MEAN
Smoothed
Fit
Values
35.88
36.21
36.54
36.87
37.20
37.54
37.87
38.20
38.53
38.86
39.19
39.52
39.85
40.18
40.51
40.85
41.18
41.51
41.84
42.17
42.50
42.10
41.45
40.81
40.18
39.56
38.95
38.35
SD
Raw
Fit
Values
9.90
6.27
5.26
6.88
7.48
6.79
5.43
5.36
8.86
8.53
7.83
7.69
8.51
7.59
10.13
8.22
7.09
11.12
8.02
SD
Smoothed
Fit
Values
5.90
6.00
6.09
6.19
6.28
6.38
6.47
6.57
6.66
6.76
6.85
6.95
7.04
7.14
7.23
7.33
7.42
7.52
7.61
7.71
7.80
7.90
7.99
8.09
8.18
8.28
8.37
8.35
MIN
(1st
Pctl)
22.15
22.26
22.37
22.48
22.59
22.70
22.81
22.92
23.03
23.14
23.25
23.36
23.47
23.58
23.69
23.80
23.91
24.02
24.13
24.24
24.35
23.73
22.86
21.99
21.14
20.30
19.47
18.93
MAX
(99th Pctl)
49.61
50.17
50.72
51.27
51.82
52.37
52.93
53.48
54.03
54.58
55.13
55.69
56.24
56.79
57.34
57.89
58.45
59.00
59.55
60.10
60.65
60.48
60.05
59.63
59.22
58.82
58.43
57.76
B-172
-------
28.00
29.00
30.00
31.00
32.00
33.00
34.00
35.00
36.00
37.00
38.00
39.00
40.00
41.00
42.00
43.00
44.00
45.00
46.00
47.00
48.00
49.00
50.00
51.00
52.00
53.00
54.00
55.00
56.00
57.00
58.00
59.00
60.00
61.00
62.00
63.00
64.00
65.00
66.00
67.00
68.00
69.00
70.00
71.00
72.00
73.00
74.00
75.00
76.00
77.00
38.08
35.13
35.79
35.22
36.06
34.95
38.13
32.63
33.59
31.11
33.12
28.80
29.06
29.54
30.90
27.60
29.33
28.53
29.41
30.49
27.92
26.48
29.80
27.49
28.95
23.77
25.34
26.05
26.30
26.06
23.67
24.70
21.63
26.64
23.84
20.26
20.38
20.49
22.05
21.92
20.38
25.30
21.21
20.46
20.63
20.60
20.91
22.27
37.75
37.17
36.60
36.04
35.48
34.94
34.41
33.88
33.37
32.87
32.37
31.89
31.42
30.95
30.50
30.05
29.62
29.19
28.78
28.37
27.97
27.59
27.21
26.84
26.49
26.14
25.80
25.48
25.16
24.85
24.55
24.27
23.99
23.72
23.46
23.21
22.97
22.74
22.52
22.31
22.11
21.92
21.74
21.57
21.41
21.26
21.12
20.99
20.87
20.76
9.80
6.30
9.10
7.89
6.93
9.51
7.08
4.88
6.17
5.13
3.76
5.14
5.74
8.00
6.82
4.32
4.17
4.90
6.00
7.15
6.05
5.36
5.13
3.66
5.83
3.56
4.61
4.29
4.91
4.07
4.81
4.65
4.99
7.38
3.77
3.83
3.90
4.56
4.15
4.59
3.80
8.14
7.94
7.74
7.55
7.37
7.19
7.01
6.84
6.68
6.52
6.37
6.22
6.08
5.95
5.82
5.70
5.58
5.47
5.36
5.26
5.16
5.07
4.99
4.91
4.83
4.77
4.70
4.65
4.60
4.55
4.51
4.48
4.45
4.43
4.41
4.40
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
18.82
18.71
18.59
18.47
18.35
18.23
18.10
17.97
17.83
17.70
17.55
17.41
17.26
17.11
16.96
16.80
16.64
16.48
16.31
16.14
15.97
15.79
15.61
15.43
15.24
15.06
14.86
14.67
14.47
14.27
14.06
13.85
13.64
13.43
13.21
12.99
12.76
12.53
12.31
12.10
11.90
11.71
11.53
11.36
11.20
11.05
10.91
10.78
10.66
10.55
56.69
55.64
54.61
53.60
52.62
51.66
50.72
49.80
48.91
48.04
47.19
46.37
45.57
44.79
44.03
43.30
42.59
41.90
41.24
40.60
39.98
39.38
38.81
38.26
37.73
37.23
36.74
36.29
35.85
35.44
35.05
34.68
34.33
34.01
33.71
33.44
33.18
32.95
32.73
32.52
32.32
32.13
31.95
31.78
31.62
31.47
31.33
31.20
31.08
30.97
B-173
-------
78.00
79.00
80.00
81.00
82.00
83.00
84.00
85.00
86.00
87.00
88.00
89.00
90.00
91.00
92.00
93.00
94.00
95.00
96.00
97.00
98.00
99.00
100.00
19.93
22.80
23.19
19.29
13.44
28.03
17.00
18.69
18.18
27.15
18.18
20.65
20.56
20.48
20.41
20.34
20.29
20.25
20.21
20.19
20.18
20.17
20.18
20.20
20.22
20.26
20.30
20.36
20.42
20.50
20.58
20.67
20.78
20.89
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
4.39
10.44
10.35
10.27
10.20
10.13
10.08
10.04
10.00
9.98
9.97
9.96
9.97
9.98
10.01
10.05
10.09
10.15
10.21
10.28
10.37
10.46
10.57
10.68
30.86
30.77
30.69
30.62
30.55
30.50
30.46
30.42
30.40
30.39
30.38
30.39
30.41
30.43
30.47
30.51
30.57
30.63
30.71
30.79
30.88
30.99
31.10
Table 3. Body Mass Raw Fits.
Age
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00
17.00
MALES
Geometric
Mean
7.767
1 1 .440
13.932
15.967
18.475
21.618
23.142
27.072
31.651
34.656
38.329
44.149
47.988
55.364
62.832
67.650
72.460
73.081
GSD
1.301
1.143
1.146
1.154
1.165
1.234
1.213
1.216
1.302
1.265
1.280
1.308
1.315
1.340
1.293
1.255
1.267
1.248
Min
3.6
8.2
9.8
11.7
11.1
13.7
16.1
19.3
19.1
24
24.3
26.2
27.7
27.7
35.7
41.5
45.8
49.9
Max
11.8
16.1
20.9
23.7
28.1
42.4
41.1
46.8
66.2
69.9
72.9
83.8
94.8
106.6
121
117.9
139.1
136.6
FEMALES
Geometric
Mean
7.429
11.119
13.258
15.587
18.005
20.353
22.454
26.483
30.534
35.235
40.550
46.579
50.673
56.649
57.214
60.091
61.582
61.229
GSD
1.304
1.163
1.158
1.160
1.171
1.229
1.194
1.239
1.315
1.271
1.304
1.302
1.274
1.275
1.248
1.249
1.255
1.248
Min
3.7
7.4
10.1
11
12.8
12.6
15.9
16.9
19.8
20.3
22.7
27.7
27.8
33.4
37.7
34.9
40.9
41.5
Max
12.1
15.3
20.4
27.9
29.1
40.4
36.7
51
60.8
58.6
71.2
84.6
93.3
99.5
110
108.4
113.8
133.1
B-174
-------
18.00
19.00
20.00
21.00
22.00
23.00
24.00
25.00
26.00
27.00
28.00
29.00
30.00
31.00
32.00
33.00
34.00
35.00
36.00
37.00
38.00
39.00
40.00
41.00
42.00
43.00
44.00
45.00
46.00
47.00
48.00
49.00
50.00
51.00
52.00
53.00
54.00
55.00
56.00
57.00
58.00
59.00
60.00
61.00
62.00
63.00
64.00
65.00
66.00
67.00
75.060
77.182
77.952
78.239
83.845
80.607
81.706
84.818
81.812
85.166
84.321
82.144
81.581
81.275
84.715
88.188
81.163
87.192
83.404
85.759
84.132
84.611
90.071
87.425
88.290
88.423
88.528
87.102
88.157
86.547
84.793
86.235
84.659
87.975
89.886
89.012
90.098
88.268
84.796
87.501
85.116
84.190
87.044
89.007
84.788
89.137
89.974
89.891
86.814
86.207
1.243
1.245
1.250
1.297
1.292
1.222
1.251
1.206
1.273
1.249
1.272
1.236
1.262
1.249
1.235
1.231
1.221
1.251
1.228
1.241
1.260
1.196
1.246
1.173
1.205
1.233
1.200
1.205
1.243
1.229
1.186
1.240
1.179
1.208
1.216
1.228
1.216
1.222
1.195
1.253
1.266
1.182
1.232
1.207
1.228
1.262
1.193
1.215
1.228
1.207
51.2
52.6
50.5
46.8
53.3
50.5
50.6
50.2
48.9
50
51
50.6
52.5
48.8
49.7
64.8
53.1
61
45.8
59.3
52.8
61.2
58.5
61.3
62.2
54
56.6
60.6
54.2
49.9
56.3
47
53.4
57.9
55.2
58.2
64.1
55.1
45
58.3
51.6
58.7
57.3
49.9
56.04
56.3
59.1
58.1
54
43.1
144.2
134.5
130
199.2
155.4
137.6
132.6
136.1
164.5
153.9
167.2
147.2
139
170.6
135.8
146.3
136.9
193.3
140.5
150.9
149.7
140.6
154
117.7
144
145.3
128.9
160.2
154.3
188.3
128.3
171.3
124.4
143.6
144.9
143.3
155.2
138.6
110.3
160
179
112.4
141.7
162.8
152.1
171.6
119
126.3
150.1
127.5
64.591
66.156
66.981
67.218
66.823
69.721
70.284
66.300
72.973
70.604
74.363
69.110
70.616
73.039
72.938
72.710
69.773
73.044
73.547
70.019
75.587
72.295
72.888
73.363
73.697
73.438
75.742
76.795
77.544
72.849
74.646
72.844
75.217
72.941
74.472
74.733
72.413
75.951
77.322
72.378
74.548
80.638
75.777
77.121
73.347
72.308
75.440
72.910
73.101
75.835
1.281
1.274
1.262
1.262
1.273
1.304
1.289
1.283
1.281
1.281
1.312
1.250
1.305
1.278
1.281
1.307
1.230
1.306
1.289
1.284
1.295
1.251
1.289
1.268
1.270
1.314
1.266
1.308
1.304
1.298
1.303
1.261
1.292
1.240
1.283
1.259
1.281
1.231
1.315
1.252
1.267
1.277
1.260
1.240
1.198
1.238
1.281
1.254
1.242
1.266
42.4
41.6
41.5
39.7
42
40.3
47.5
44.8
45.3
41.4
44.3
39.3
42.1
43.7
41.5
44.9
46.6
44.2
44.6
48.1
43.7
41.6
45.5
50.5
47.1
45.6
49.5
41.6
46.6
47.8
44.2
45.1
48.4
42.5
45.7
46.2
44.3
53.6
45.6
48.6
45
50.9
51.3
50.7
49.7
46.9
41.1
35.9
48.4
47.2
123.6
118.5
122.6
123.7
123.5
143
144.5
131.8
128.9
140.9
142.1
116.3
151.5
125.9
139.7
135.2
115.3
138.4
150.1
152.1
151.7
123.1
137.4
156.9
146.1
159.5
153
141.5
145.8
130.6
166
125.54
175.7
120.2
146.6
176.6
123.1
125.6
134.9
122.6
117.7
133
128.3
125.6
121.1
119.9
132.5
113.7
113.3
123.8
B-175
-------
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
73
73
72
85.172
87.116
82.775
79.630
82.011
85.590
83.001
84.465
78.733
79.376
79.909
77.629
79.866
75.405
76.798
74.611
75.325
71.776
986494
364276
742058
72.11984
71
70
70
69
497622
875404
253186
630968
69.00875
68
67
67
66
386532
764314
142096
519878
65.89766
65
275442
1.191
1.222
1.210
1.240
1.204
1.196
1.217
1.185
1.207
1.170
1.195
1.155
1.174
1.157
1.180
1.158
1.205
1.191
1.17
1.17
1.16
1.16
1.16
1.16
1.16
1.15
1.15
1.15
1.15
1.14
1.14
1.14
1.14
61.2
50.7
46.5
51
51.9
56.2
53.3
56.5
55.9
58.7
41.1
56.4
56
55.8
54.4
53.2
41.5
46.9
50.57
50.38
50.19
50
49.81
49.62
49.44
49.25
49.06
48.87
48.68
48.49
48.3
48.11
47.92
163.2
127.2
125.5
122.8
132.7
128.3
120
133.5
121.1
109.3
115.1
107.8
111.9
111.9
111.8
107
109.5
105.8
101.07
99.113
97.154
95.194
93.235
91.276
89.317
87.358
85.399
83.44
81.481
79.522
77.563
75.604
73.645
62
61
61
73.207
74.368
68.977
69.083
69.898
71.360
70.410
70.526
69.549
70.128
66.375
67.780
62.214
65.397
64.755
62.886
62.215
61.453
400356
847614
294872
60.74213
60
59
59
58
189388
636646
083904
531162
57.97842
57
56
56
55
425678
872936
320194
767452
55.21471
54
661968
1.250
1.225
1.188
1.232
1.240
1.240
1.277
1.216
1.199
1.240
1.211
1.200
1.255
1.184
1.260
1.196
1.216
1.209
1.21
1.21
1.21
1.21
1.21
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.19
1.19
1.19
39.3
48
45.9
45.5
40.7
47.4
37.4
46.8
48.8
40.3
44.1
46.2
41.2
42.7
40.6
44.7
43.5
42.3
41.85
41.66
41.47
41.27
41.08
40.88
40.69
40.49
40.3
40.1
39.91
39.71
39.52
39.32
39.13
120.7
118
102.8
108.1
103.8
127.6
106.4
117.4
101.7
119.8
109.8
98.4
121.4
91.4
120
101.2
108.4
93.2
101.16
100.26
99.351
98.445
97.538
96.632
95.726
94.82
93.914
93.008
92.102
91.195
90.289
89.383
88.477
**
Dark shading (age 86+) designates linear forecast.
Table 4. Body Mass Smoothed Fits (5-Year Running Averages).
Age
0.00
1.00
2.00
3.00
4.00
5.00
MALES
Geometric
Mean
7.767209794
1 1 .44008024
13.93227373
15.96664726
18.47458493
21.61756114
GSD
1.300901
1.143324
1.145566
1.153689
1.164972
1 .233822
Min
3.6
8.2
9.8
11.7
11.1
13.7
Max
11.8
16.1
20.9
23.7
28.1
42.4
FEMALES
Geometric
Mean
7.428916349
11.11947416
13.25797158
15.58684049
18.00506307
20.35285099
GSD
1 .304229
1.162608
1.158434
1.159883
1.17108
1 .229237
Min
3.7
7.4
10.1
11
12.8
12.6
Max
12.1
15.3
20.4
27.9
29.1
40.4
B-176
-------
6.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00
17.00
18.00
19.00
20.00
21.00
22.00
23.00
24.00
25.00
26.00
27.00
28.00
29.00
30.00
31.00
32.00
33.00
34.00
35.00
36.00
37.00
38.00
39.00
40.00
41.00
42.00
43.00
44.00
45.00
46.00
47.00
48.00
49.00
50.00
51.00
52.00
53.00
54.00
55.00
23.14243627
27.07246068
31.6505017
34.65600448
38.32939135
44.14863459
47.98795299
55.36374737
62.83159173
67.65031426
72.45980541
73.08089659
75.06031573
77.18236513
77.95205826
78.45564692
79.56489519
80.46958232
81 .84267254
82.55729313
82.82151847
83.56439112
83.65195203
83.00459482
82.89721864
82.80701235
83.58034187
83.38418057
84.50647805
84.9321819
85.14102649
84.32994666
85.01958212
85.59524544
86.39949423
86.90564401
87.76379051
88.54719729
87.95342484
88.09985934
87.751282
87.02523405
86.56661258
86.07815707
86.04175058
86.70964624
87.55345712
88.32616726
89.04784314
88.4120991
1.213499
1.215834
1.301873
1.265317
1.279707
1.30753
1.314848
1 .33952
1.292533
1.254999
1 .267468
1 .248405
1 .243204
1 .244928
1 .250326
1.265585
1.261251
1 .262527
1.253588
1 .248802
1 .240222
1.250399
1 .247428
1.258753
1.253937
1.251132
1 .242848
1.239735
1.237533
1.233184
1 .234298
1.240177
1.235131
1.233983
1 .223065
1.215924
1.210495
1.211458
1.203416
1.217379
1.222211
1.212835
1 .220669
1.215489
1.208607
1.206031
1.2144
1 .209663
1.218268
1.215526
16.1
19.3
19.1
24
24.3
26.2
27.7
27.7
35.7
41.5
45.8
49.9
51.2
52.6
50.5
50.88
50.74
50.34
50.28
50.7
50.04
50.14
50.14
50.6
50.58
50.52
53.28
53.78
55.48
54.88
56.8
54.4
56.02
55.52
58.62
59.2
59.44
58.52
58.94
57.52
55.06
55.52
53.6
52.16
52.9
53.96
54.34
57.76
58.1
55.52
41.1
46.8
66.2
69.9
72.9
83.8
94.8
106.6
121
117.9
139.1
136.6
144.2
134.5
130
152.66
151.34
150.96
152.18
145.24
144.94
150.86
153.78
154.36
155.58
151.96
147.78
145.72
156.58
150.56
153.58
154.26
155
147.14
142.58
141.2
140.32
137.98
139.22
146.54
155.4
152
160.48
153.32
151.18
142.5
145.5
142.28
145.12
138.46
22.45431948
26.48323788
30.53391399
35.23472141
40.54996835
46.57910267
50.67329267
56.64881107
57.21362103
60.09135575
61.58214656
61.22931022
64.59054256
66.15556407
66.98146906
66.35375002
67.37976393
68.20537834
68.06901959
69.21992781
69.97607936
70.90453453
70.66975978
71.53295767
71.54621552
72.01313142
71 .6826276
71.81523165
72.30094254
72.40264379
71.81884258
72.3941641
72.89859355
72.86733489
72.830387
73.56585153
73.13604869
73.82543503
74.60684165
75.44302619
75.27348935
75.51517243
74.93569966
74.62001355
73.69947055
74.02400492
74.04127315
73.95491798
74.10188224
74.97813364
1.194119
1 .23892
1.315137
1.271364
1.303997
1.302182
1 .273946
1 .275455
1 .24795
1 .24897
1.255162
1 .248057
1.281298
1 .274083
1.261822
1 .270386
1 .274844
1.277813
1.282127
1.285979
1.287735
1.289413
1.28161
1 .285847
1.285108
1 .28495
1.283915
1 .280002
1 .280205
1 .282492
1.283151
1.280709
1.284821
1.281407
1.277165
1 .274483
1 .278433
1.281514
1 .285327
1 .292445
1.29795
1 .295658
1 .29459
1.291492
1 .278764
1 .27574
1 .266893
1.270898
1.25873
1 .273724
15.9
16.9
19.8
20.3
22.7
27.7
27.8
33.4
37.7
34.9
40.9
41.5
42.4
41.6
41.5
41.44
41.02
42.2
42.86
43.98
43.86
44.66
43.02
42.48
42.16
42.18
42.3
43.76
44.18
44.36
45.68
45.44
44.44
44.7
45.88
45.68
46.06
47.64
46.86
46.08
46.22
45.94
45.06
46.42
45.6
45.18
45.58
45.42
46.46
47.08
36.7
51
60.8
58.6
71.2
84.6
93.3
99.5
110
108.4
113.8
133.1
123.6
118.5
122.6
122.38
126.26
131.46
133.3
134.34
137.82
137.64
132
135.94
135.34
135.1
133.72
133.52
130.9
135.74
138.22
141.52
143.08
142.88
144.24
143.04
144.6
150.58
151.4
149.18
146.08
147.38
141.888
148.728
143.608
146.808
148.928
148.44
138.42
141.36
B-177
-------
56.00
57.00
58.00
59.00
60.00
61.00
62.00
63.00
64.00
65.00
66.00
67.00
68.00
69.00
70.00
71.00
72.00
73.00
74.00
75.00
76.00
77.00
78.00
79.00
80.00
81.00
82.00
83.00
84.00
85.00
86.00
87.00
88.00
89.00
90.00
91.00
92.00
93.00
94.00
95.00
96.00
97.00
98.00
99.00
100.00
87.93495739
87.15584772
85.97418819
85.72952642
86.57173577
86.0292098
86.83331368
87.99005122
88.55927286
88.12051692
88.40439667
87.61146369
87.03986775
85.61667034
84.17987726
83.34052655
83.42413149
82.60108145
82.93914453
82.75981666
82.23282472
81.09670348
80.02242628
79.10265965
78.43698141
77.92142176
76.86173441
76.40090269
74.7828307
74.4992137
73.81250877
73.43871959
72.79767378
72.74205786
72.11983988
71.4976219
70.87540393
70.25318595
69.63096798
69.00875
68.38653203
67.76431405
67.14209607
66.5198781
66.20876911
1 .222906
1.230391
1.223617
1 .22558
1 .228074
1 .222944
1 .22222
1 .224363
1 .220869
1 .225034
1.220898
1.206714
1.212565
1.211601
1.213949
1.213444
1.214423
1.213508
1.208433
1.201809
1.19492
1.194698
1.182282
1.180128
1.170309
1.17229
1.164819
1.174796
1.17822
1.180574
1.177977
1.179377
1.170756
1.164535
1.162177
1.159818
1.157459
1.1551
1.152742
1.150383
1.148024
1.145665
1.143307
1.140948
1.139769
56.14
54.82
53.74
54.18
55.16
54.708
55.648
55.728
55.888
56.708
54.12
55.1
53.42
51.1
50.5
52.26
51.26
51.78
53.78
54.76
56.12
53.1
53.72
53.62
53.6
52.74
55.16
52.18
50.36
49.31362
48.50949
47.90759
49.60794
50.19052
50.00172
49.81292
49.62412
49.43532
49.24652
49.05772
48.86892
48.68012
48.49132
48.30252
48.20812
141.48
148.62
140.06
140.68
151.18
149.6
148.12
149.44
146.36
143.82
138.9
137.22
138.86
138.7
133.24
134.28
127.3
125.86
127.46
127.12
122.44
119.8
117.36
113.04
111.2
111.7
110.08
110.42
109.2
107.0343
104.4969
102.5276
99.66646
97.15354
95.19446
93.23538
91.27631
89.31723
87.35815
85.39908
83.44
81 .48092
79.52185
77.56277
76.58323
74.55937637
74.52242942
76.16748501
76.13252691
76.09221736
76.28599922
75.83796229
74.79845832
74.22522224
73.42130739
73.91902253
74.09892054
73.88448789
73.09783811
72.29417333
71.10679374
70.73734313
69.94568967
70.25540111
70.34868617
70.39465694
69.39757689
68.87151054
67.20924709
66.37891246
65.30424541
64.60647334
63.49351577
63.34131978
62.74189138
62.16041309
61 .84223228
61.5476723
61.29487188
60.74212987
60.18938785
59.63664584
59.08390383
58.53116181
57.9784198
57.42567778
56.87293577
56.32019375
55.76745174
55.49108073
1 .267547
1 .269258
1 .268509
1 .274205
1.259138
1.248419
1 .242552
1 .243365
1 .242246
1 .242692
1 .256261
1 .258602
1.247351
1 .234088
1.23216
1 .227009
1.225132
1 .235452
1.241
1 .23444
1 .234265
1.228511
1.213229
1.221048
1.217987
1 .222093
1.219074
1 .22226
1.213219
1.218822
1 .208932
1.211505
1.209819
1.209189
1.207753
1.206317
1 .20488
1 .203444
1 .202008
1.200571
1.199135
1.197699
1.196263
1.194826
1.194108
47.66
47.42
48.74
48.28
49.3
49.52
49.9
47.94
44.86
44.4
43.9
42.38
43.76
45.76
45.18
43.88
45.5
43.38
43.56
44.22
44.14
43.48
45.24
44.12
42.9
42.96
43.08
42.54
42.76
42.59095
42.80295
42.15598
41.71005
41.46516
41.27036
41.07556
40.88075
40.68595
40.49115
40.29634
40.10154
39.90674
39.71193
39.51713
39.41973
136.56
124.78
126.76
127.3
125.44
125.14
125.58
125.48
122.56
120.1
120.64
120.8
117.9
115.72
114.68
110.68
112.06
109.74
112.66
111.38
114.58
111.02
109.42
110.22
108.16
108.2
106.48
108.48
102.84
104.7926
100.844
100.4742
98.48308
99.35077
98.44462
97.53846
96.63231
95.72615
94.82
93.91385
93.00769
92.10154
91.19538
90.28923
89.83615
B-178
-------
Table 5. Hemoglobin Content.
Age
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
MALES
MEAN
11.927
12.20959
12.42075
12.69015
12.8006
12.95822
13.19574
13.46198
13.35161
13.59742
13.63062
13.66
13.9727
14.28293
14.70654
15.13583
15.36442
15.45945
15.7487
15.76812
15.79371
15.71703
15.70837
15.55635
15.43525
15.44038
15.41492
15.31983
15.27653
15.07274
14.96193
14.72786
14.51
14.52915
13.97647
13.801
13.534
STD
0.993545
1.013091
0.823171
0.83159
0.80152
0.878515
0.893008
0.836639
0.833121
0.971019
0.906785
0.726155
0.955869
1.036749
1 .020254
1 .04546
1.021623
0.979296
1.02514
0.831813
0.880956
0.91072
1.045808
0.959964
1.021741
1.105939
1.096952
1.123792
0.97796
1.192645
1 .24457
1.418355
1 .476879
1.352814
1.757686
1.757686
1.757686
FEMALES
12.209
12.27307
12.55018
12.4519
12.83442
12.87154
13.01866
13.09899
13.25291
13.36671
13.58919
13.52681
13.6273
13.46986
13.58878
13.47154
13.50562
13.49842
13.46091
13.35445
13.5016
13.47168
13.2967
13.34583
13.4881
13.48617
13.61113
13.67737
13.83717
13.76529
13.81911
13.79013
13.84426
13.57546
13.43767
13.2085
13.005
0.729499905
0.719158646
0.843436666
0.965868504
0.773409545
0.969254536
0.828912341
0.754370806
0.826349227
0.808377267
1.034306588
0.90041802
0.884271668
0.97623121
1.034527514
0.856131982
1 .088863466
1.117860417
1.18250671
1.090493585
1.072791517
1.170602542
1.145254677
1.134192006
1.163867696
1.348669176
1.193756618
1.106237392
1.237714453
1.093354796
1.093565513
1.056812752
1.30818261
1.238910845
1 .552685662
1 .552685662
1 .552685662
B-179
-------
ATTACHMENT 4. TECHNICAL MEMORANDUM ON
LONGITUDINAL DIARY CONSTRUCTION APPROACH
B-180
-------
INTERNATIONAL
TECHNICAL MEMORANDUM
TO: Stephen Graham and John Langstaff, US EPA
FROM: Arlene Rosenbaum
DATE: February 29, 2008
SUBJECT: The Cluster-Markov algorithm in APEX
Background
The goals of population exposure assessment generally include an accurate estimate of
both the average exposure concentration and the high end of the exposure distribution. One of
the factors influencing the number of exposures at the high end of the concentration distribution
is time-activity patterns that differ from the average, e.g., a disproportionate amount of time
spent near roadways. Whether a model represents these exposure scenarios well depends on
whether the treatment of activity pattern data accurately characterizes differences among
individuals.
Human time-activity data for population exposure models are generally derived from
demographic surveys of individuals' daily activities, the amount of time spent engaged in those
activities, and the ME locations where the activities occur. Typical time-activity pattern data
available for inhalation exposure modeling consist of a sequence of location/activity
combinations spanning a 24-hour duration, with 1 to 3 records for any single individual. But
modeling assessments of exposure to air pollutants typically require information on activity
patterns over long periods of time, e.g., a full year. For example, even for pollutant health
effects with short averaging times (e.g., ozone 8-hour average) it may be desirable to know the
frequency of exceedances of a threshold concentration over a long period of time (e.g., the
annual number of exceedances of an 8-hour average ozone concentration of 0.07 ppm for each
simulated individual).
Long-term activity patterns can be estimated from daily ones by combining the daily
records in various ways, and the method used for combining them will influence the variability
of the long-term activity patterns across the simulated population. This in turn will influence the
ability of the model to accurately represent either long-term average high-end exposures, or the
number of individuals exposed multiple times to short-term high-end concentrations.
A common approach for constructing long-term activity patterns from short-term records
is to re-select a daily activity pattern from the pool of data for each day, with the implicit
assumption that there is no correlation between activities from day to day for the simulated
individual. This approach tends to result in long-term activity patterns that are very similar
across the simulated population. Thus, the resulting exposure estimates are likely to
B-181
-------
underestimate the variability across the population, and therefore, underestimate the high-end
concentrations.
A contrasting approach is to select a single activity pattern (or a single pattern for each
season and/or weekday-weekend) to represent a simulated individual's activities over the
modeling period. This approach has the implicit assumption that an individual's day to day
activities are perfectly correlated. This approach tends to result in long-term activity patterns
that are very different across the simulated population, and therefore may over-estimate the
variability across the population.
The Cluster-Markov Algorithm
Recently, a new algorithm has been developed and incorporated into APEX that attempts
to more realistically represent the day-to-day correlation of activities for individuals. The
algorithms first use cluster analysis to divide the daily activity pattern records into groups that
are similar, and then select a single daily record from each group. This limited number of daily
patterns is then used to construct a long-term sequence for a simulated individual, based on
empirically-derived transition probabilities. This approach is intermediate between the
assumption of no day-to-day correlation (i.e., re-selection for each time period) and perfect
correlation (i.e., selection of a single daily record to represent all days).
The steps in the algorithm are as follows.
• For each demographic group (age, gender, employment status), temperature range,
and day-of-week combination, the associated time-activity records are partitioned into
3 groups using cluster analysis. The clustering criterion is a vector of 5 values: the
time spent in each of 5 microenvironment categories (indoors - residence; indoors -
other building; outdoors - near road; outdoors - away from road; in vehicle).
• For each simulated individual, a single time-activity record is randomly selected from
each cluster.
• Next the Markov process determines the probability of a given time-activity pattern
occurring on a given day based on the time-activity pattern of the previous day and
cluster-to-cluster transition probabilities. The cluster-to-cluster transition
probabilities are estimated from the available multi-day time-activity records. (If
insufficient multi-day time-activity records are available for a demographic group,
season, day-of-week combination, then the cluster-to-cluster transition probabilities
are estimated from the frequency of time-activity records in each cluster in the CHAD
data base.).
Figure 1 illustrates the Cluster-Markov algorithm in flow chart format.
B-182
-------
CHAD Data Base
Demographic
Group 1
Weekday
Season 1
Demographic
Group 1
Weekday
Season 2
Demographic
Group 1
Weekend
TRANSITION
PROBABILITIES
Annual Time-Activity Sequence
Figure 1. Flow chart of Cluster-Markov algorithm used for constructing longitudinal time-activity diaries.
B-183
-------
Evaluation of modeled diary profiles versus observed diary profiles
The Cluster-Markov algorithm is also incorporated into the Hazardous Air
Pollutant Exposure Model (HAPEM). Rosebaum and Cohen (2004) incorporated the
algorithm in HAPEM and tested modeled longitudinal profiles with multi-day diary data
sets collected as part of the Harvard Southern California Chronic Ozone Exposure Study
(Xue et al. 2005, Geyh et al. 2000). In this study, 224 children in ages between 7 and 12
yr were followed for 1 year from June 1995 to May 1996, for 6 consecutive days each
month. The subjects resided in two separate areas of San Bernardino County: urban
Upland CA, and the small mountain towns of Lake Arrowhead, Crestline, and Running
Springs, CA.
For purposes of clustering the activity pattern records were characterized
according to time spent in each of 5 aggregate microenvironments: indoors-home,
indoors-school, indoors-other, outdoors, and in-transit. For purposes of defining diary
pools and for clustering and calculating transition probabilities the activity pattern
records were divided by day type (i.e., weekday, weekend), season (i.e., summer or ozone
season, non-summer or non-ozone season), age (7-10 and 11-12), and gender.
Week-long sequences (Wednesday through Tuesday) for each of 100 people in
each age/gender group for each season were simulated. To evaluate the algorithm the
following statistics were calculated for the predicted multi-day activity patterns and
compared them with the actual multi-day diary data.
• For each age/gender group for each season, the average time in each
microenvironment
• For each simulated person-week and microenvironment, the average of the
within-person variance across all simulated persons. (The within-person
variance was defined as the variance of the total time per day spent in the
microenvironment across the week.)
• For each simulated person-week the variance across persons of the mean time
spent in each microenvironment.
In each case the predicted statistic for the stratum was compared to the statistic for
the corresponding stratum in the actual diary data. The mean normalized bias for the
statistic, which is a common performance measure used in dispersion model performance
and was also calculated as follows.
(predicted - observed)
j
N i observed
The predicted time-in-microenvironment averages matched well with the
observed values. For combinations of microenvironment/age/gender/season the
normalized bias ranges from -35% to +41%. Sixty percent of the predicted averages
have bias between -9% and +9%, and the mean bias across any microenvironment ranges
from -9% to +4%. Fourteen predictions have positive bias and 23 have negative bias.
B-184
-------
For the variance across persons for the average time spent in each
microenvironment, the bias ranged from -40% to +120% for any
microenvironment/age/gender/season. Sixty-five percent of the predicted variances had
bias between -22% and +24%. The mean normalized bias across any microenvironment
ranged from -10% to +28%. Eighteen predictions had positive bias and 20 had negative
bias.
For the within-person variance for time spent in each microenvironment, the bias
ranged from -47% to +150% for any microenvironment/age/gender/season. Seventy
percent of the predicted variances had bias between -25% and +30%. The mean
normalized bias across any microenvironment ranged from -11% to +47%. Twenty-eight
predictions had positive bias and 12 had negative bias, suggesting some tendency for
overprediction of this variance measure.
The overall conclusion was that the proposed algorithm appeared to be able to
replicate the observed data reasonably well. Although some discrepancies were rather
large for some of the "variance across persons" and "within-person variance" subsets,
about two-thirds of the predictions for each case were within 30% of the observed value.
A detailed description of the evaluation using HAPEM is presented in Attachment 5.
Comparison of Cluster-Markov approach with other algorithms
As part of the application of APEX in support of US EPA's recent review of the
ozone NAAQS several sensitivity analyses were conducted (US EPA, 2007). One of
these was to make parallel simulations using each of the three algorithms for constructing
multi-day time-activity sequences that are incorporated into APEX.
Table 1 presents the results for the number of persons in Atlanta population
groups with moderate exertion exposed to 8-hour average concentrations exceeding 0.07
ppm. The results show that the predictions made with alternative algorithm Cluster-
Markov algorithm are substantially different from those made with simple re-sampling or
with the Diversity-Autocorrelation algorithm ("base case"). Note that for the cluster
algorithm approximately 30% of the individuals with 1 or more exposure have 3 or more
exposures. The corresponding values for the other algorithms range from about 13% to
21%.
Table 2 presents the results for the mean and standard deviation of number of
days/person with 8-hour average exposures exceeding 0.07 ppm with moderate or greater
exertion. The results show that although the mean for the Cluster-Markov algorithm is
similar to the other approaches, the standard deviation is substantially higher, i.e., the
Cluster-Markov algorithm results in substantially higher inter-individual variability.
B-185
-------
Table 1. Sensitivity to longitudinal diary algorithm: 2002 simulated counts of Atlanta
general population and children (ages 5-18) with any or three or more 8-hour ozone
exposures above 0.07 ppm concomitant with moderate or greater exertion (after US EPA
2007).
Population
Group
General
Population
Children (5-18)
One or more exposures
Simple
re-sampling
979,533
411,429
Diversity-
Autocorrelation
939,663
(-4%)
389,372
(-5%)
Cluster-
Markov
668,004
(-32%)
295,004
(-28%)
Three or more exposures
Simple
re-sampling
124,687
71,174
Diversity-
Autocorrelation
144,470
(+16%)
83,377
(+17%)
Cluster-
Markov
188,509
(+51%)
94,216
(+32%)
Table 2. Sensitivity to longitudinal diary algorithm: 2002 days per person with 8-hour
ozone exposures above 0.07 ppm concomitant with moderate or greater exertion for
Atlanta general population and children (ages 5-18) (after US EPA 2007).
Population
Group
General
Population
Children (5-1 8)
Mean Days/Person
Simple
re-sampling
0.332
0.746
Base case
0.335
(+1%)
0.755
(+1%)
Cluster-
Markov
0.342
(+3%)
0.758
(+2%)
Standard Deviation
Simple re-
sampling
0.757
1.077
Base case
0.802
(+6%)
1.171
(+9%)
Cluster-
Markov
1.197
(+58%)
1.652
(+53%)
References
Geyh AS, Xue J, Ozkaynak H, Spengler JD. (2000). The Harvard Southern California
chronic ozone exposure study: Assessing ozone exposure of grade-school-age
children in two Southern California communities. Environ Health Persp. 108:265-
270.
Rosenbaum AS and Cohen JP. (2004). Evaluation of a multi-day activity pattern
algorithm for creating longitudinal activity patterns. Memorandum prepared for Ted
Palma, US EPA OAQPS, by ICF International.
US EPA. (2007). Ozone Population Exposure Analysis for Selected Urban Areas. EPA-
452/R-07-010. Available at:
http://www.epa.gov/ttn/naaqs/standards/ozone/data/2007-01_o3_exposure_tsd.pdf
Xue J, Liu SV, Ozkaynak H, Spengler J. (2005). Parameter evaluation and model
validation of ozone exposure assessment using Harvard Southern California Chronic
Ozone Exposure Study Data. J. Air & Waste Manage Assoc. 55:1508-1515.
B-186
-------
ATTACHMENT 5. TECHNICAL MEMORANDUM ON THE
EVALUATION CLUSTER-MARKOV ALGORITHM
B-187
-------
INTERNATIONAL
TECHNICAL MEMORANDUM
TO: Ted Palma, US EPA
FROM: Arlene Rosenbaum and Jonathan Cohen, ICF Consulting
DATE: November 4, 2004
SUBJECT: Evaluation of a multi-day activity pattern algorithm for creating longitudinal
activity patterns.
BACKGROUND
In previous work ICF reviewed the HAPEM4 modeling approach for developing
annual average activity patterns from the CHAD database and recommended an approach to
improve the model's pattern selection process to better represent the variability among
individuals. This section summarizes the recommended approach. (For details see
Attachment 4)
Using cluster analysis, first the CHAD daily activity patterns are grouped into either
two or three categories of similar patterns for each of the 30 combinations of day type
(summer weekday, non-summer weekday, and weekend) and demographic group (males or
females; age groups: 0-4, 5-11, 12-17, 18-64, 65+). Next, for each combination of day type
and demographic group, category-to-category transition probabilities are defined by the
relative frequencies of each second-day category associated with each given first-day
category, where the same individual was observed for two consecutive days. (Consecutive
day activity pattern records for a single individual constitute a small subset of the CHAD
data.)
To implement the proposed algorithm, for each day type and demographic group, one
daily activity pattern per category is randomly selected from the corresponding CHAD data
to represent that category. That is, if there are 3 cluster categories for each of 3 day types, 9
unique activity patterns are selected to be averaged together to create an annual average
activity pattern to represent an individual in a given demographic group and census tract.
The weighting for each of the 9 activity patterns used in the averaging process is
determined by the product of two factors. The first is the relative frequency of its day type,
i.e., 0.18 for summer weekdays, 0.54 for non-summer weekdays, and 0.28 for weekends.
The second factor in the weighting for the selected activity pattern is determined by
simulating a sequence of category-types as a one-stage Markov chain process using the
transition probabilities. The category for the first day is selected according to the relative
frequencies of each category. The category for the second day is selected according to the
category-to-category transition probabilities for the category selected for the first day. The
category for the third day is selected according to the transition probabilities for the category
B-188
-------
selected for the second day. This is repeated for all days in the day type (65 for summer
weekdays, 195 for non-summer weekdays, 104 for weekends), producing a sequence of daily
categories. The relative frequency of the category-type in the sequence associated with the
selected activity pattern is the second factor in the weighting.
PROPOSED ALGORITHM STEPS
The proposed algorithm is summarized in Figure 1. Each step is explained in this
section.
Data Preparation
Step 1: Each daily activity pattern in the CHAD data base is summarized by the
total minutes in each of five micro-environments: indoors - residence; indoors - other
building; outdoors - near road; outdoors - away from road; in vehicle. These five
numbers are assumed to represent the most important features of the activity pattern
for their exposure impact.
Step 2: All CHAD activity patterns for a given day-type and demographic
group are subjected to cluster analysis, resulting in 2 or 3 cluster categories. Each
daily activity pattern is tagged with a cluster category.
Step 3: For each day-type and demographic group, the relative frequency of each
day-type in the CHAD data base is determined.
Step 4: All CHAD activity patterns for a given day-type and demographic group
that are consecutive days for a single individual, are analyzed to determine the
category-to-category transition frequencies in the CHAD data base. These transition
frequencies are used to calculate category-to-category transition probabilities.
For example, if there are 2 categories, A and B, then
PAA = the probability that a type A pattern is followed by a type A pattern,
PAB = the probability that a type A pattern is followed by a type B pattern (PAB =
I-PAA),
PBB = the probability that a type B pattern is followed by a type B pattern, and
PBA = the probability that a type B pattern is followed by a type A pattern (PBA =
I-PBB).
Activity Pattern Selection
For each day-type and demographic group in each census tract:
Step 5: One activity pattern is randomly selected from each cluster category group
(i.e., 2 to 3 activity patterns)
Creating Weights for Day-type Averaging
For each day-type and demographic group in each census tract:
B-189
-------
Step 6: A cluster category is selected for the first day of the day-type sequence,
according to the relative frequency of the cluster category days in the CHAD data set.
Step 7: A cluster category is selected for each subsequent day in the day-type
sequence day by day using the category-to-category transition probabilities.
Step 8: The relative frequency of each cluster category in the day-type sequence is
determined.
Step 9: The activity patterns selected for each cluster category (Step 5) are
averaged together using the cluster category frequencies (Step 8) as weights, to create
a day-type average activity pattern.
Creating Annual Average Activity Patterns
For each demographic group in each census tract:
Step 10: The day-type average activity patterns are averaged together using the
relative frequency of day-types as weights, to create an annual average activity
pattern.
Creating Replicates
For each demographic group in each census tract:
Step 11: Steps 5 through 10 are repeated 29 times to create 30 annual average
activity patterns.
EVALUATING THE ALGORITHM
The purpose of this study is to evaluate how well the proposed one-stage Markov
chain algorithm can reproduce observed multi-day activity patterns with respect to
demographic group means and inter-individual variability, while using one-day selection.
In order to accomplish this we propose to apply the algorithm to observed multi-day
activity patterns provided by the WAM, and compare the means and variances of the
predicted multi-day patterns with the observed patterns.
Current APEX Algorithm
Because the algorithm is being considered for incorporation into APEX, we would
like the evaluation to be consistent with the approach taken in APEX for selection of activity
patterns for creating multi-day sequences. The APEX approach for creating multi-day
activity sequences is as follows.
Stepl: A profile for a simulated individual is generated by selection of gender,
age group, and home sector from a given set of distributions consistent with the
population of the study area.
B-190
-------
Step 2: A specific age within the age group is selected from a uniform distribution.
Step 3: The employment status is simulated as a function of the age.
Step 4: For each simulated day, the user defines an initial pool of possible diary
days based on a user-specified function of the day type (e.g., weekday/weekend) and
temperature.
Step 5: The pool is further restricted to match the target gender and employment
status exactly and the age within 2A years for some parameter A. The diary days
within the pool are assigned a weight of 1 if the age is within A years of the target age
and a weight of w (user-defined parameter) if the age difference is between A and 2A
years. For each simulated day, the probability of selecting a given diary day is equal
to the age weight divided by the total of the age weights for all diary days in the pool
for that day.
Approach to Incorporation of Day-to-Day Dependence into APEX Algorithm
If we were going to incorporate day-to-day dependence of activity patterns into the
APEX model, we would propose preparing the data with cluster analysis and transition
probabilities as described in Steps 1-4 for the proposed HAPEM 5 algorithm, with the
following modifications.
• For Step 2 the activity patterns would be divided into groups based on day-type
(weekday, weekend), temperature, gender, employment status, and age, with
cluster analysis applied to each group. However, because the day-to-day
transitions in the APEX activity selection algorithm can cross temperature bins,
we would propose to use broad temperature bins for the clustering and transition
probability calculations so that the cluster definitions would be fairly uniform
across temperature bins. Thus we would probably define the bins according to
season (e.g., summer, non-summer).
• In contrast to HAPEM, the sequence of activity patterns may be important in
APEX. Therefore, for Step 4 transition probabilities would be specified for
transitions between days with the same day-type and season, as in HAPEM, and
also between days with different day-types and/or seasons. For example,
transition probabilities would be specified for transitions between summer
weekdays of each category and summer weekends of each category.
Another issue for dividing the CHAD activity records for the purposes of clustering
and calculating transition probabilities is that the diary pools specified for the APEX activity
selection algorithm use varying and overlapping age ranges. One way to address this
problem would be to simply not include consideration of age in the clustering process, under
the assumption that cluster categories are similar across age groups, even if the frequency of
each cluster category varies by age group. This assumption could be tested by examination
of the cluster categories stratified by age group that were developed for HAPEM5. If the
assumption is found to be valid, then the cluster categories could be pre-determined for input
B-191
-------
to APEX, while the transition probabilities could be calculated within APEX during the
simulation for each age range specified for dairy pools.
If the assumption is found to be invalid, then an alternative approach could be
implemented that would create overlapping age groups for purposes of clustering as follows.
APEX age group ranges and age window percentages would be constrained to some
maximum values. Then a set of overlapping age ranges that would be at least as large as the
largest possible dairy pool age ranges would be defined for the purposes of cluster analysis
and transition probability calculation. The resulting sets of cluster categories and transition
probabilities would be pre-determined for input into APEX and the appropriate set used by
APEX for each diary pool used during the simulation.
The actual activity pattern sequence selection would be implemented as follows. The
activity pattern for first day in the year would be selected exactly as is currently done in
APEX, as described above. For the selecting the second day's activity pattern, each age
weight would be multiplied by the transition probability PAB where A is the cluster for the
first day's activity pattern and B is the cluster for a given activity pattern in the available pool
of diary days for day 2. (Note that day 2 may be a different day-type and/or season than day
1). The probability of selecting a given diary day on day 2 is equal to the age weight times
PAB divided by the total of the products of age weight and PAB for all diary days in the pool
for day 2. Similarly, for the transitions from day 2 to day 3, day 3 to day 4, etc.
Testing the Approach with the Multi-day Data set
We tested this approach using the available multi-day data set. For purposes of
clustering we characterized the activity pattern records according to time spent in each of 5
microenvironments: indoors-home, indoors-school, indoors-other, outdoors (aggregate of the
3 outdoor microenvironments), and in-transit.
For purposes of defining diary pools and for clustering and calculating transition
probabilities we divided the activity pattern records by day type (i.e., weekday, weekend),
season (i.e., summer or ozone season, non-summer or non-ozone season), age (6-10 and 11-
12), and gender. Since all the subjects are 6-12 years of age and all are presumably
unemployed, we need not account for differences in employment status. For each day type,
season, age, and gender, we found that the activity patterns appeared to group in three
clusters.
In this case, we simulated week-long sequences (Wednesday through Tuesday) for
each of 100 people in each age/gender group for each season, using the transition
probabilities. To evaluate the algorithm we calculated the following statistics for the
predicted multi-day activity patterns for comparison with the actual multi-day diary data.
For each age/gender group for each season, the average time in each
microenvironment
For each age/gender group, season, and microenvironment, the average of the
within-person variance across all simulated persons (We defined the within-
B-192
-------
person variance as the variance of the total time per day spent in the
microenvironment across the week.)
• For each age/gender group, season, and microenvironment, the variance across
persons of the mean time spent in that microenvironment
In each case we compared the predicted statistic for the stratum to the statistic for the
corresponding stratum in the actual diary data. 5
We also calculated the mean normalized bias for the statistic, which is a common
performance measure used in dispersion model performance and which is calculated as
follows.
* rr, T , o 100^ (predicted - observed) n /
NBIAS = > — %
N i observed
RESULTS
Comparisons of simulated and observed data for time in each of the 5
microenvironments are presented in Tables 1-3 and Figures 2-5.
Average Time in Microenvironment
Table 1 and Figure 2 show the comparisons for the average time spent in each of the
5 microenvironments for each age/gender group and season. Figure 3 shows the comparison
for all the microenvironments except indoor, home in order to highlight the lower values.
Table 1 and the figures show that the predicted time-in-microenvironment averages
match well with the observed values. For combinations of
microenvironment/age/gender/season the normalized bias ranges from -35% to +41%. Sixty
percent of the predicted averages have bias between -9% and +9%, and the mean bias across
any microenvironment ranges from -9% to +4%. Fourteen predictions have positive bias and
23 have negative bias. A Wilcoxon signed rank test that the median bias across the 40
combinations = 0 % was not significant (p-value = 0.40) supporting the conclusion of no
overall bias.
Variance Across Persons
Table 2 and Figure 4 show the comparisons for the variance across persons for the
average time spent in each microenvironment. In this case the bias ranges from -40% to
+120% for any microenvironment/age/gender/season. Sixty-five percent of the predicted
variances have bias between -22% and +24%. The mean normalized bias across any
microenvironment ranges from -10% to +28%. Eighteen predictions have positive bias and
20 have negative bias. Figure 4 suggests a reasonably good match of predicted to observed
5 For the diary data, because the number of days per person varies, the average of the within-person variances
was calculated as a weighted average, where the weight is the degrees of freedom, i.e., one less than the number
of days simulated. Similarly, the variance across persons of the mean time was appropriately adjusted for the
different degrees of freedom using analysis of variance.
B-193
-------
variance in spite of 2 or 3 outliers. A Wilcoxon signed rank test that the median bias across
the 40 combinations = 0 % was not significant (p-value = 0.93) supporting the conclusion of
no overall bias.
Within-Person Variance for Persons
Table 3 and Figure 5 show the comparisons for the within-person variance for time
spent in each microenvironment. In this case the bias ranges from -47% to +150% for any
microenvironment/age/gender/season. Seventy percent of the predicted variances have bias
between -25% and +30%. The mean normalized bias across any microenvironment ranges
from -11% to +47%. Twenty-eight predictions have positive bias and 12 have negative bias,
suggesting some tendency for overprediction of this variance measure. And indeed a
Wilcoxon signed rank test that the median bias across the 40 combinations = 0 % was very
significant (p-value = 0.01) showing that the within-person variance was significantly
overpredicted. Still, Figure 4 suggests a reasonably good match of predicted to observed
variance in most cases, with a few overpredicting outliers at the higher end of the
distribution. So although the positive bias is significant in a statistical sense (i.e., the variance
is more likely to be overpredicted than underpredicted), it is not clear whether the bias is
large enough to be important.
CONCLUSIONS
The proposed algorithm appears to be able to replicate the observed data reasonably
well, although the within-person variance is somewhat overpredicted.
It would be informative to compare this algorithm with the earlier alternative
approaches in order to gain perspective on the degree of improvement, if any, afforded by
this approach.
Two earlier approaches were:
1. Select a single activity pattern for each day-type/season combination from the
appropriate set, and use that pattern for every day in the multi-day sequence that
corresponds to that day-type and season.
2. Re-select an activity pattern for each day in the multi-day sequence from the
appropriate set for the corresponding day-type and season.
Goodness-of-fit statistics could be developed to compare the three approaches and find
which model best fits the data for a given stratum.
B-194
-------
Table 1. Average time spent in each microenvironment: comparison of predicted and observed.
Microenvironment
Indoor, home
Indoor, school
Indoor, other
Demographic
Group
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
Season
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Observed
(hours/day)
15.5
15.8
15.7
15.8
16.2
16.5
16.0
16.2
0.7
2.3
0.8
2.2
0.7
2.1
0.6
2.4
2.9
2.4
2.2
1.9
2.2
Predicted
(hours/day)
16.5
15.5
15.2
16.4
15.3
16.5
15.6
16.1
0.7
2.5
0.5
2.2
0.7
2.4
0.9
2.7
2.4
2.7
2.7
1.8
1.6
Normalized
Bias
6%
-2%
-3%
4%
-5%
0%
-3%
-1%
-1%
-9%
7%
-34%
0%
6%
13%
38%
11%
4%
-14%
13%
21%
-3%
-25%
B-195
-------
Outdoors
In-vehicle
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
2.2
2.3
1.9
3.7
2.5
4.1
3.1
3.7
2.3
3.9
2.6
1.1
1.0
1.1
1.0
1.2
0.9
1.1
0.9
2.1
2.2
2.0
3.5
2.5
4.3
2.7
5.2
2.1
4.3
2.4
0.9
0.9
1.3
0.9
1.1
0.8
1.0
0.8
-2%
-5%
4%
-2%
-6%
0%
4%
-12%
41%
-5%
9%
-7%
3%
-20%
-13%
13%
-16%
-12%
-15%
-5%
-7%
-9%
B-196
-------
Table 2. Variance across persons for time spent in each microenvironment: comparison of
predicted and observed.
Microenvironment
Indoor, home
Indoor, school
Indoor, other
Demographic
Group
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
Season
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Observed
(hours/day)2
70
67
54
35
56
42
57
39
6.0
9.5
5.6
5.3
4.9
5.4
5.6
9.2
46
44
34
23
21
Predicted
(hours/day)2
42
60
49
30
47
38
63
42
5.2
5.9
3.8
8.2
5.5
5.3
6.0
11
32
46.
33
16
18
Normalized
Bias
-40%
-9%
-9%
-12%
-17%
-10%
12%
8%
-10%
-13%
-38%
-32%
53%
11%
-1%
6%
23%
1%
-30%
6%
-4%
-27%
-15%
B-197
-------
Outdoors
In-vehicle
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
28
33
30
17
9.3
17
8.3
22
9.0
13
10
1.9
1.8
2.5
1.5
3.5
2.8
3.2
1.3
22
31
30
23
6.8
18
7.6
22
9.1
29
11
2.3
1.6
4.7
1.6
4.7
2.0
5.4
1.7
-22%
-6%
0%
-12%
37%
-27%
3%
-8%
0%
1%
120%
8%
17%
24%
-11%
93%
9%
34%
-28%
69%
35%
28%
B-198
-------
Table 3. Average within person variance for time spent in each microenvironment: comparison
of predicted and observed.
Microenvironment
Indoor, home
Indoor, school
Indoor, other
Demographic
Group
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
Season
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Observed
(hours/day)2
20
18
17
15
22
22
21
17
2.3
7.3
2.0
6.7
1.7
7.4
1.4
7.3
14
14
12
10
10
Predicted
(hours/day)2
29
23
30
24
42
25
24
24
2.4
6.4
1.5
5.8
2.1
7.6
2.9
7.8
14
18
17
13
10
Normalized
Bias
49%
25%
75%
64%
93%
13%
16%
38%
47%
5%
-12%
-25%
-14%
29%
3%
101%
6%
12%
-4%
30%
42%
26%
1%
B-199
-------
Outdoors
In-vehicle
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 8-10
Girls, 11-
12
Boys, 11-
12
MEAN
Girls, 6-10
Boys, 6-10
Girls, 11-
12
Boys, 11-
12
MEAN
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
Summer
Not
Summer
14
11
12
8.4
3.4
6.7
3.4
10
4.0
9.2
4.3
1.0
0.90
1.1
0.81
1.3
1.3
2.4
0.85
15
14
13
9.5
3.2
9.5
4.4
25
4.5
7.4
3.7
0.90
0.48
1.4
0.71
1.3
1.1
1.6
0.85
7%
26%
7%
17%
13%
-3%
42%
28%
150%
11%
-20%
-15%
26%
-13%
-47%
31%
-12%
4%
-16%
-34%
1%
-11%
B-200
-------
Consolidated Human Activity Database - CHAD (CHAD)
Winter Weekday
Pattern Group
Summer Weekday
Pattern Group
Cluster Analysis
ransition
nalysi
Weekend
Pattern Group
*•
Transition
Probabilities
i
r
Average Winter
Weekday Pattern
0.54
Markov Selection
Individual Annual Average Activity Pattern
Figure 1. Flow diagram of proposed algorithm for creating annual average activity patterns for HAPEM5.
B-201
-------
Of) -,
^3 H c
1
o
•= 10
•c
1 5-
£
0.
0 ]
*^
.^
-X'"^
^^^_
^^^
.^
.s^
.^
^^
X*
• Indoor, home
• Indoor, school
indoor, other
Outdoor
x In chicle
0 5 10 15 20
Observed (hours/day)
Figure 2. Comparison of predicted and observed average time in each of 5 microenvironments
for age/gender groups and seasons.
•I* 4
i s-
•c o
oj ^
0
•5 -|
i
n
^^
^^
^s^
T^^>
3>r
^^
3
-------
80
70
Rn
•o 'so
o ou
S 40
3 ^U
•- ^0
w JU
on
10
0
C
/
-2
/
-r * .
/'.
/*
+£•
S"
f
• girls, 6-10, summer
• girls, 6-10, winter
boys , 6-10, summer
x boys, 6-10, winter
X girls, 11-12, summer
• girls, 11-12, winter
+ boys, 11-12, summer
- boys, 11-12, winter
1 1 1 1
) 20 40 60 80
Observed
Figure 4. Comparison of predicted and observed variance across persons for time spent in each
of 5 microenvironments for age/gender groups and seasons.
30 | •
OR J
TO -IR
D ID
E
(/) in
/
X • S
\/
f
f.
f
• girls, 6-10, summer
• girls, 6-10, winter
boys , 6-10, summer
boys, 6-10, winter
X girls, 11-12, summer
• girls, 11-12, winter
+ boys, 11-12, summer
-boys, 11-12, winter
0 10 20 30
Observed
Figure 5. Comparison of predicted and observed the average within-person variance for time
spent in each of 5 microenvironments by age/gender groups and seasons.
B-203
-------
ATTACHMENT 6. TECHNICAL MEMORANDUM ON
ANALYSIS OF AIR EXCHANGE RATE DATA
B-204
-------
IGF
INTERNATIONAL
DRAFT MEMORANDUM
To: John Langstaff
From: Jonathan Cohen, Hemant Mallya, Arlene Rosenbaum
Date: September 30, 2005
Re: EPA 68D01052, Work Assignment 3-08. Analysis of Air Exchange Rate Data
EPA is planning to use the APEX exposure model to estimate ozone exposure in 12 cities /
metropolitan areas: Atlanta, GA; Boston, MA; Chicago, IL; Cleveland, OH; Detroit, MI;
Houston, TX; Los Angeles, CA; New York, NY; Philadelphia, PA; Sacramento, CA; St. Louis,
MO-IL; Washington, DC. As part of this effort, ICF Consulting has developed distributions of
residential and non-residential air exchange rates (AER) for use as APEX inputs for the cities to
be modeled. This memorandum describes the analysis of the AER data and the proposed APEX
input distributions. Also included in this memorandum are proposed APEX inputs for
penetration and proximity factors for selected microenvironments.
Residential Air Exchange Rates
Studies. Residential air exchange rate (AER) data were obtained from the following seven
studies:
Avol: Avol et al, 1998. In this study, ozone concentrations and AERs were measured at
126 residences in the greater Los Angeles metropolitan area between February and
December, 1994. Measurements were taken in four communities: Lancaster, Lake
Gregory, Riverside, and San Dimas. Data included the daily average outdoor
temperature, the presence or absence of an air conditioner (either central or room), and
the presence or absence of a swamp (evaporative) cooler. Air exchange rates were
computed based on the total house volume and based on the total house volume corrected
for the furniture. These data analyses used the corrected AERs.
RTF Panel: Williams et al, 2003a, 2003b. In this study particulate matter concentrations
and daily average AERs were measured at 37 residences in central North Carolina during
2000 and 2001 (averaging about 23 AER measurements per residence). The residences
belong to two specific cohorts: a mostly Caucasian, non-smoking group aged at least 50
years having cardiac defibrillators living in Chapel Hill; a group of non-smoking, African
Americans aged at least 50 years with controlled hypertension living in a low-to-
moderate SES neighborhood in Raleigh. Data included the daily average outdoor
temperature, and the number of air conditioner units (either central or room). Every
residence had at least one air conditioner unit.
RIOPA: Meng et al, 2004, Weisel et al, 2004. The Relationship of Indoor, Outdoor, and
Personal Air (RIOPA) study was undertaken to estimate the impact of outdoor sources of
air toxics to indoor concentrations and personal exposures. Volatile organic compounds,
B-205
-------
carbonyls, fine particles and AERs were measured once or twice at 310 non-smoking
residences from summer 1999 to spring 2001. Measurements were made at residences in
Elizabeth, NJ, Houston TX, and Los Angeles CA. Residences in California were
randomly selected. Residences in New Jersey and Texas were preferentially selected to
be close (< 0.5 km) to sources of air toxics. The AER measurements (generally over 48
hours) used a PMCH tracer. Data included the daily average outdoor temperature, and the
presence or absence of central air conditioning, room air conditioning, or a swamp
(evaporative) cooler.
TEACH: Chillrud at al, 2004, Kinney et al, 2002, Sax et al, 2004. The Toxic Exposure
Assessment, a Columbia/Harvard (TEACH) study was designed to characterize levels of
and factors influencing exposures to air toxics among high school students living in
inner-city neighborhoods of New York City and Los Angeles, CA. Volatile organic
compounds, aldehydes, fine particles, selected trace elements, and AER were measured at
87 high school student's residences in New York City and Los Angeles in 1999 and
2000. Data included the presence or absence of an air conditioner (central or room) and
hourly outdoor temperatures (which were converted to daily averages for these analyses).
Wilson 1984: Wilson et al, 1986, 1996. In this 1984 study, AER and other data were
collected at about 600 southern California homes with three seven-day tests (in March
and July 1984, and January, 1985) for each home. We obtained the data directly from Mr.
Wilson. The available data consisted of the three seven-day averages, the month, the
residence zip code, the presence or absence of a central air conditioner, and the presence
or absence of a window air conditioner. We matched these data by month and zip code to
the corresponding monthly average temperatures obtained from EPA's SCRAM website
as well as from the archives in www.wunderground.com (personal and airport
meteorological stations). Residences more than 25 miles away from the nearest available
meteorological station were excluded from the analysis. For our analyses, the
city/location was defined by the meteorological station, since grouping the data by zip
code would not have produced sufficient data for most of the zip codes.
Wilson 1991: Wilson et al, 1996. Colome et al, 1993, 1994. In this 1991 study, AER and
other data were collected at about 300 California homes with one two-day test in the
winter for each home. We obtained the data directly from Mr. Wilson. The available data
consisted of the two-day averages, the date, city name, the residence zip code, the
presence or absence of a central air conditioner, the presence or absence of a swamp
(evaporative) cooler, and the presence or absence of a window air conditioner . We
matched these data by date, city, and zip code to the corresponding daily average
temperatures obtained from EPA's SCRAM website as well as from the archives in
www.wunderground.com (personal and airport meteorological stations). Residences
more than 25 miles away from the nearest available meteorological station were excluded
from the analysis. For our analyses, the city/location was defined by the meteorological
station, since grouping the data by zip code would not have produced sufficient data for
most of the zip codes.
Murray and Burmaster: Murray and Burmaster (1995). For this article, Murray and
Burmaster corrected and compiled nationwide residential AER data from several studies
conducted between 1982 and 1987. These data were originally compiled by the Lawrence
Berkeley National Laboratory. We acknowledge Mr. Murray's assistance in obtaining
B-206
-------
these data for us. The available data consisted of AER measurements, dates, cities, and
degree-days. Information on air conditioner presence or absence was not available.
Table A-l summarizes these studies.
For each of the studies, air conditioner usage, window status (open or closed), and fan status (on
or off) was not part of the experimental design, although some of these studies included
information on whether air conditioners or fans were used (and for how long) and whether
windows were closed during the AER measurements (and for how long).
As described above, in the following studies the homes were deliberately sampled from specific
subsets of the population at a given location rather than the entire population: The RTF Panel
study selected two specific cohorts of older subjects with specific diseases. The RIOPA study
was biased towards residences near air toxics sources. The TEACH study focused on inner-city
neighborhoods. Nevertheless, we included all these studies because we determined that any
potential bias would be likely to be small and we preferred to keep as much data as possible.
B-207
-------
Table A-l. Summary of Studies of Residential Air Exchange Rates
Locations
Years
Months/Seasons
Number of
Homes
Total AER
Measurements
Average
Number of
Measurements
per Home
Measurement
Duration
Measurement
Technique
Min AER Value
Max AER Value
Mean AER
Value
Min
Temperature
(C)
Avol
Lancaster, Lake
Gregory,
Riverside, San
Dimas. All in
Southern CA
1994
Feb; Mar; Apr;
May; Jun; Jul;
Aug; Sep; Oct;
Nov
86
161
1.87
Not Available
Not Available
0.01
2.70
0.80
-0.04
RTF Panel
Research Triangle
Park, NC
2000; 2001
2000 (Jun; Jul;
Aug; Sep; Oct;
Nov), 2001 (Jan;
Feb; Apr; May)
37
854
23.08
24 hour
Perflourocarbon
tracer.
0.02
21.44
0.72
-2.18
RIOPA
CA; NJ; TX
1999; 2000; 2001
1999 (July to
Dec); 2000 (all
months); 2001
(Jan and Feb)
284
524
1.85
24 to 96 hours
PMCH tracer
0.08
87.50
1.41
-6.82
TEACH
Los Angeles, CA;
New York City, NY
1999; 2000
1999 (Feb; Mar; Apr;
Jul; Aug); 2000 (Jan;
Feb; Mar; Sep; Oct)
85
151
1.78
Sample time (hours)
reported. Ranges
from about 1 to 7
days.
Perflourocarbon
tracer.
0.12
8.87
1.71
-1.36
Wilson 1984
Southern CA
1984, 1985
Mar 1984, Jul 1984, Jan
1985
581
1,362
2.34
7 days
Perflourocarbon tracer.
0.03
11.77
1.05
11.00
Wilson 1991
Southern CA
1984
Jan, Mar, Jul
288
316
1.10
7 days
Perflourocarbon tracer.
0.01
2.91
0.57
3.00
Murray
and
Burmaster
AZ, CA, CO,
CT, FL, ID,
MD, MN, MT,
NJ
1982-1987
Various
1,884
2,844
1.51
Not available
Not available
0.01
11.77
0.76
Not available
B-208
-------
Max
Temperature
(C)
Air Conditioner
Categories
Air Conditioner
Measurements
Fan Categories
Fan
Measurements
Window Open/
Closed Data
Comments
Avol
36.25
No A/C; Central
or Room A/C;
Swamp Cooler
only; Swamp +
[Central or Room]
A/C use in
minutes
Not available
Time on or off for
various fan types
during sampling
was recorded, but
not included in
database provided.
Duration open
between times
6am- 12 pm; 12pm
- 6 pm; and 6pm -
6am
RTF Panel
30.81
Central or Room
A/C (Y/N)
Not Available
Fan (Y/N)
Not Available
Windows (open /
closed along with
duration open in
inch-hours units
RIOPA
32.50
Window A/C
(Y/N); Evap
Coolers (Y/N)
Duration
measurements in
Hrs and Mins
Fan (Y/N)
Duration
measurements in
Hrs and Mins
Windows (Open /
Closed) along with
window open
duration
measurements
CA sample was a
random sample of
homes. NJ and TX
homes were
deliberately
chosen to be near
to ambient
sources.
TEACH
32.00
Central or Room A/C
(Y/N)
Not Available
Not Available
Not Available
Not Available
Restricted to inner-
city homes with high
school students.
Wilson 1984
28.00
Central A/C (Y/N);
Room A/C (Y/N);
Not Available
Not Available
Not Available
Not Available
Contemporaneous
temperature data
obtained for these
analyses from SCRAM
and
www. wunderground. com
meteorological data.
Wilson 1991
25.00
Central A/C (Y/N);
Room A/C (Y/N);
Swamp Cooler(Y/N)
Not Available
Not Available
Not Available
Not Available
Contemporaneous
temperature data
obtained for these
analyses from SCRAM
and
www.wunderground.com
meteorological data.
Murray
and
Burmaster
Not available
Not available
Not available
Not available
Not available
Not available
B-209
-------
We compiled the data from these seven studies to create the following variables, of which some
had missing values:
• Study
• Date
• Time - Time of the day that the AER measurement was made
• House_ID - Residence identifier
• Measurement_ID - Uniquely identifies each AER measurement for a given study
• AER - Air Exchange Rate (per hour)
• AER_Duration - Length of AER measurement period
• Have_AC - Indicates if the residence has any type of air conditioner (A/C), either a room
A/C or central A/C or swamp cooler or any of them in combination. "Y" = "Yes." "N" =
"No."
• Type_of_ACl - Indicates the types of A/C or swamp cooler available in each house
measured. Possible values: "Central A/C" "Central and Room A/C" "Central or Room
A/C" "No A/C" "Swamp + (Central or Room)" "Swamp Cooler only" "Window A/C"
"Window and Evap"
• Type_of_AC2 - Indicates if a house measured has either no A/C or some A/C. Possible
values are "No A/C" and "Central or Room A/C."
• Have_Fan - Indicates if the house studied has any fans
• Mean_Temp - Daily average outside temperature
• Min_Temp - Minimum hourly outside temperature
• Max_Temp - Maximum hourly outside temperature
• State
• City
• Location - Two character abbreviation
• Flag - Data status. Murray and Burmaster study: "Used" or "Not Used." Other studies:
"Used"; "Missing" (missing values for AER, Type_of_AC2, and/or Mean_Temp);
"Outlier".
The main data analysis was based on the first six studies. The Murray and Burmaster data were
excluded because of the absence of information on air conditioner presence. (However, a subset
of these data was used for a supplementary analysis described below.).
Based on our review of the AER data we excluded seven outlying high AER values - above 10
per hour. The main data analysis used all the remaining data that had non-missing values for
AER, Type_of_AC2, and Mean_Temp. We decided to base the A/C type variable on the broad
characterization "No A/C" versus "Central or Room A/C" since this variable could be calculated
from all of the studies (excluding Murray and Burmaster). Information on the presence or
absence of swamp coolers was not available from all the studies, and, also importantly, the
corresponding information on swamp cooler prevalence for the subsequent ozone modeling cities
was not available from the American Housing Survey. It is plausible that AER distributions
B-210
-------
depend upon the presence or absence of a swamp cooler. It is also plausible that AER
distributions also depend upon whether the residence specifically has a central A/C, room or
window A/C, or both. However we determined to use the broader A/C type definition, which in
effect assumes that the exact A/C type and the presence of a swamp cooler are approximately
proportionately represented in the surveyed residences.
Most of the studies had more than one AER measurement for the same house. It is reasonable to
assume that the AER varies with the house as well as other factors such as the temperature. (The
A/C type can be assumed to be the same for each measurement of the same house). We expected
the temperature to be an important factor since the AER will be affected by the use of the
available ventilation (air conditioners, windows, fans), which in turn will depend upon the
outside meteorology. Therefore it is not appropriate to average data for the same house under
different conditions, which might have been one way to account for dependence between
multiple measurements on the same house. To simplify the data analysis, we chose to ignore
possible dependence between measurements on the same house on different days and treat all the
AER values as if they were statistically independent.
Summary Statistics. We computed summary statistics for AER and its natural logarithm
LOG_AER on selected strata defined from the study, city, A/C type, and mean temperature.
Cities were defined as in the original databases, except that for Los Angeles we combined all the
data in the Los Angeles ozone modeling region, i.e. the counties of Los Angeles, Orange,
Ventura, Riverside, and San Bernardino. A/C type was defined from the Type_of_AC2 variable,
which we abbreviated as "NA" = "No A/C" and "AC" = "Central or Room A/C." The mean
temperature was grouped into the following temperature bins: -10 to 0 °C, 0 to 10 °C, 10 to 20
°C, 20 to 25 °C, 25 to 30 °C, 30 to 40 °C.(Values equal to the lower bounds are excluded from
each interval.) Also included were strata defined by study = "All" and/or city = "All," and/or
A/C type = "All" and/or temperature bin = "All." The following summary statistics for AER and
LOG_AER were computed:
• Number of values
• Arithmetic Mean
• Arithmetic Standard Deviation
• Arithmetic Variance
• Deciles (Min, 10th, 20th ... 90th percentiles, Max)
These calculations exclude all seven outliers and results are not used for strata with 10 or fewer
values, since those summary statistics are extremely unreliable.
Examination of these summary tables clearly demonstrates that the AER distributions vary
greatly across cities and A/C types and temperatures, so that the selected AER distributions for
the modeled cities should also depend upon the city, A/C type and temperature. For example, the
mean AER for residences with A/C ranges from 0.39 for Los Angeles between 30 and 40 °C to
1.73 for New York between 20 and 25 °C. The mean AER for residences without A/C ranges
from 0.46 for San Francisco between 10 and 20 °C to 2.29 for New York between 20 and 25 °C.
The need to account for the city as well as the A/C type and temperature is illustrated by the
B-211
-------
result that for residences with A/C and between 20 and 25 °C, the mean AER ranges from 0.52
for Research Triangle Park to 1.73 for New York. Statistical comparisons are described below.
Statistical Comparisons. Various statistical comparisons were carried out between the different
strata, for the AER and its logarithm. The various strata are defined as in the Summary Statistics
section, excluding the "All" cases. For each analysis, we fixed one or two of the variables Study,
City, A/C type, temperature, and tested for statistically significant differences among other
variables. The comparisons are listed in Table A-2.
Table A-2. Summary of Comparisons of Means
Comparison
Analysis
Number.
1.
2.
O
4.
5.
6.
Comparison
Variable(s)
"Groups
Compared"
City
Temp. Range
Type of A/C
City
City
Type of A/C
AND Temp.
Range
Stratification
Variable(s)
(not missing in
worksheet)
Type of A/C AND
Temp. Range
Study AND City
Study AND City
Type of A/C
Temp. Range
Study AND City
Total
Comparisons
12
12
15
2
6
17
Cases with significantly
different means (5 %
level)
AER
8
5
5
2
5
6
Log AER
8
5
5
2
6
6
For example, the first set of comparisons fix the Type of A/C and the temperature range; there
are twelve such combinations. For each of these twelve combinations, we compare the AER
distributions across different cities. This analysis determines whether the AER distribution is
appropriately defined by the A/C type and temperature range, without specifying the city.
Similarly, for the sixth set of comparisons, the study and city are held fixed (17 combinations)
and in each case we compare AER distributions across groups defined by the combination of the
A/C type and the temperature range.
The F Statistic comparisons compare the mean values between groups using a one way analysis
of variance (ANOVA). This test assumes that the AER or log(AER) values are normally
distributed with a mean that may vary with the comparison variable(s) and a constant variance.
We calculated the F Statistic and its P-value. P-values above 0.05 indicate cases where all the
group means are not statistically significantly different at the 5 percent level. Those results are
summarized in the last two columns of the above table "Summary of Comparisons of Means"
which gives the number of cases where the means are significantly different. Comparison
analyses 2, 3, and 6 show that for a given study and city, slightly less than half of the
comparisons show significant differences in the means across temperature ranges, A/C types, or
both. Comparison analyses 1, 4, and 5 show that for the majority of cases, means vary
significantly across cities, whether you first stratify by temperature range, A/C type, or both.
B-212
-------
The Kruskal-Wallis Statistic comparisons are non-parametric tests that are extensions of the
more familiar Wilcoxon tests to two or more groups. The analysis is valid if the AER minus the
group median has the same distribution for each group, and tests whether the group medians are
equal. (The test is also consistent under weaker assumptions against more general alternatives)
The P-values show similar patterns to the parametric F test comparisons of the means. Since the
logarithm is a strictly increasing function and the test is non-parametric, the Kruskal-Wallis tests
give identical results for AER and Log (AER).
The Mood Statistic comparisons are non-parametric tests that compare the scale statistics for two
or more groups. The scale statistic measures variation about the central value, which is a non-
parametric generalization of the standard deviation. Specifically, suppose there is a total of N
AER or log(AER) values, summing across all the groups. These N values are ranked from 1 to
N, and the j'th highest value is given a score of (j - (N+l)/2}2. The Mood statistic uses a one
way ANOVA statistic to compare the total scores for each group. Generally, the Mood statistics
show that in most cases the scale statistics are not statistically significantly different. Since the
logarithm is a strictly increasing function and the test is non-parametric, the Mood tests give
identical results for AER and Log (AER).
Fitting Distributions. Based on the summary statistics and the statistical comparisons, the need
to fit different AER distributions to each combination of A/C type, city, and temperature is
apparent. For each combination with a minimum of 11 AER values, we fitted and compared
exponential, log-normal, normal, and Weibull distributions to the AER values.
The first analysis used the same stratifications as in the above "Summary Statistics" and
"Statistical Comparisons" sections. Results are not reported for all strata because of the
minimum data requirement of 11 values. Results for each combination of A/C type, city, and
temperature (i.e., A, C, and T) were analyzed. Each combination has four rows, one for each
fitted distribution. For each distribution we report the fitted parameters (mean, standard
deviation, scale, shape) and the p-value for three standard goodness-of-fit tests: Kolmogorov-
Smirnov (K-S), Cramer-Von-Mises (C-M), Anderson-Darling (A-D). Each goodness-of-fit test
compares the empirical distribution of the AER values to the fitted distribution. The K-S and C-
M tests are different tests examining the overall fit, while the Anderson-Darling test gives more
weight to the fit in the tails of the distribution. For each combination, the best-fitting of the four
distributions has the highest p-value and is marked by an x in the final three columns. The mean
and standard deviation (Std_Dev) are the values for the fitted distribution. The scale and shape
parameters are defined by:
• Exponential: density = a"1 exp(-x/a), where shape = mean = a
• Log-normal: density = {axV(2:r)}"1 exp{ -(log x - Q2 / (2a2)}, where shape = a and
scale = C,. Thus the geometric mean and geometric standard deviation are given by
exp(Q and exp(a), respectively.
• Normal: density = {aV(2:r)}"1 exp{ -(x - (j,)2 / (2a2)}, where mean = (j, and standard
deviation = a
• Weibull: density = (c/o) (x/a)0"1 exp{-(x/a)c}, where shape = c and scale = a
B-213
-------
Generally, the log-normal distribution was the best-fitting of the four distributions, and so, for
consistency, we recommend using the fitted log-normal distributions for all the cases.
One limitation of the initial analysis was that distributions were available only for selected cities,
and yet the summary statistics and comparisons demonstrate that the AER distributions depend
upon the city as well as the temperature range and A/C type. As one option to address this issue,
we considered modeling cities for which distributions were not available by using the AER
distributions across all cities and dates for a given temperature range and A/C type.
Another important limitation of the initial analysis was that distributions were not fitted to all of
the temperature ranges due to inadequate data. There are missing values between temperature
ranges, and the temperature ranges are all bounded. To address this issue, the temperature ranges
were regrouped to cover the entire range of temperatures from minus to plus infinity, although
obviously the available data to fit these ranges have finite temperatures. Stratifying by A/C type,
city, and the new temperature ranges produces results for four cities: Houston (AC and NA); Los
Angeles (AC and NA); New York (AC and NA); Research Triangle Park (AC). For each of the
fitted distributions we created histograms to compare the fitted distributions with the empirical
distributions.
AER Distributions for The First Nine Cities. Based upon the results for the above four cities
and the corresponding graphs, we propose using those fitted distributions for the three cities
Houston, Los Angeles, and New York. For another 6 of the cities to be modeled, we propose
using the distribution for one of the four cities thought to have similar characteristics to the city
to be modeled with respect to factors that might influence AERs. These factors include the age
composition of housing stock, construction methods, and other meteorological variables not
explicitly treated in the analysis, such as humidity and wind speed patterns. The distributions
proposed for these cities are as follows:
• Atlanta, GA, A/C: Use log-normal distributions for Research Triangle Park. Residences
with A/C only.
• Boston, MA: Use log-normal distributions for New York
• Chicago, IL: Use log-normal distributions for New York
• Cleveland, OH: Use log-normal distributions for New York
• Detroit, MI: Use log-normal distributions for New York
• Houston, TX: Use log-normal distributions for Houston
• Los Angeles, CA: Use log-normal distributions for Los Angeles
• New York, NY: Use log-normal distributions for New York
• Philadelphia, PA: Use log-normal distributions for New York
Since the AER data for Research Triangle Park was only available for residences with air
conditioning, AER distributions for Atlanta residences without air conditioning are discussed
below.
To avoid unusually extreme simulated AER values, we propose to set a minimum AER value of
0.01 and a maximum AER value of 10.
B-214
-------
Obviously, we would be prefer to model each city using data from the same city, but this
approach was chosen as a reasonable alternative, given the available AER data.
AER Distributions for Sacramento and St. Louis. For these two cities, a direct mapping to one
of the four cities Houston, Los Angeles, New York, and Research Triangle Park is not
recommended because the cities are likely to be too dissimilar. Instead, we decided to use the
distribution for the inland parts of Los Angeles to represent Sacramento and to use the aggregate
distributions for all cities outside of California to represent St. Louis. The results for the city
Sacramento were obtained by combining all the available AER data for Sacramento, Riverside,
and San Bernardino counties. The results for the city St. Louis were obtained by combining all
non-California AER data.
AER Distributions for Washington DC. Washington DC was judged likely to have similar
characteristics both to Research Triangle Park and to New York City. To choose between these
two cities, we compared the Murray and Burmaster AER data for Maryland with AER data from
each of those cities. The Murray and Burmaster study included AER data for Baltimore and for
Gaithersburg and Rockville, primarily collected in March. April, and May 1987, although there
is no information on mean daily temperatures or A/C type. We collected all the March, April,
and May AER data for Research Triangle Park and for New York City, and compared those
distributions with the Murray and Burmaster Maryland data for the same three months.
The results for the means and central values show significant differences at the 5 percent level
between the New York and Maryland distributions. Between Research Triangle Park and
Maryland, the central values and the mean AER values are not statistically significantly
different, and the differences in the mean log (AER) values are much less statistically significant
than between New York and Maryland. The scale statistic comparisons are not statistically
significantly different between New York and Maryland, but were statistically significantly
different between Research Triangle Park and Maryland. Since matching central and mean
values is generally more important than matching the scales, we propose to model Washington
DC residences with air conditioning using the Research Triangle Park distributions, stratified by
temperature:
• Washington DC, A/C: Use log-normal distributions for Research Triangle Park.
Residences with A/C only.
Since the AER data for Research Triangle Park was only available for residences with air
conditioning, the estimated AER distributions for Washington DC residences without air
conditioning are discussed below.
AER Distributions for Washington DC and Atlanta GA Residences With No A/C. For
Atlanta and Washington DC we have proposed to use the AER distributions for Research
Triangle Park. However, all the Research Triangle Park data (from the RTF Panel study) were
from houses with air conditioning, so there are no available distributions for the "No A/C" cases.
For these two cities, one option is to use AER distributions fitted to all the study data for
residences without A/C, stratified by temperature. We propose applying the "No A/C"
B-215
-------
distributions for modeling these two cities for residences without A/C. However, since Atlanta
and Washington DC residences are expected to be better represented by residences outside of
California, we instead propose to use the "No A/C" AER distributions aggregated across cities
outside of California, which is the same as the recommended choice for the St. Louis "No A/C"
AER distributions.
A/C Type and Temperature Distributions. Since the proposed AER distribution is conditional
on the A/C type and temperature range, these values also need to be simulated using APEX in
order to select the appropriate AER distribution. Mean daily temperatures are one of the
available APEX inputs for each modeled city, so that the temperature range can be determined
for each modeled day according to the mean daily temperature. To simulate the A/C type, we
obtained estimates of A/C prevalence from the American Housing Survey. Thus for each
city/metropolitan area, we obtained the estimated fraction of residences with Central or Room
A/C (see Table A-3), which gives the probability p for selecting the A/C type "Central or Room
A/C." Obviously, 1-p is the probability for "No A/C." For comparison with Washington DC and
Atlanta, we have included the A/C type percentage for Charlotte, NC (representing Research
Triangle Park, NC). As discussed above, we propose modeling the 96-97 % of Washington DC
and Atlanta residences with A/C using the Research Triangle Park AER distributions, and
modeling the 3-4 % of Washington DC and Atlanta residences without A/C using the combined
study No A/C AER distributions.
Table A-3. Fraction of residences with central or room A/C (from American Housing
Survey)
CITY
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington DC
Research Triangle Park
SURVEY AREA & YEAR
Atlanta, 2003
Boston, 2003
Chicago, 2003
Cleveland, 2003
Detroit, 2003
Houston, 2003
Los Angeles, 2003
New York, 2003
Philadelphia, 2003
Sacramento, 2003
St. Louis, 2003
Washington DC, 2003
Charlotte, 2002
PERCENTAGE
97.01
85.23
87.09
74.64
81.41
98.70
55.05
81.57
90.61
94.63
95.53
96.47
96.56
Other AER Studies
We recently became aware of some additional residential and non-residential AER studies that
might provide additional information or data. Indoor / outdoor ozone and PAN distributions were
studied by Jakobi and Fabian (1997). Liu et al (1995) studied residential ozone and AER
distributions in Toronto, Canada. Weschler and Shields (2000) describes a modeling study of
B-216
-------
ventilation and air exchange rates. Weschler (2000) includes a useful overview of residential and
non-residential AER studies.
AER Distributions for Other Indoor Environments
To estimate AER distributions for non-residential, indoor environments (e.g., offices and
schools), we obtained and analyzed two AER data sets: "Turk" (Turk et al, 1989); and "Persily"
(Persily and Gorfain 2004; Persily et al. 2005).
The earlier "Turk" data set (Turk et al, 1989) includes 40 AER measurements from offices (25
values), schools (7 values), libraries (3 values), and multi-purpose (5 values), each measured
using an SF6 tracer over two- or four-hours in different seasons of the year.
The more recent "Persily" data (Persily and Gorfain 2004; Persily et al. 2005) were derived
from the U.S. EPA Building Assessment Survey and Evaluation (BASE) study, which was
conducted to assess indoor air quality, including ventilation, in a large number of randomly
selected office buildings throughout the U. S. The data base consists of a total of 390 AER
measurements in 96 large, mechanically ventilated offices; each office was measured up to four
times over two days, Wednesday and Thursday AM and PM. The office spaces were relatively
large, with at least 25 occupants, and preferably 50 to 60 occupants. AERs were measured both
by a volumetric method and by a CO2 ratio method, and included their uncertainty estimates. For
these analyses, we used the recommended "Best Estimates" defined by the values with the lower
estimated uncertainty; in the vast majority of cases the best estimate was from the volumetric
method.
Another study of non-residential AERs was performed by Lagus Applied Technology (1995)
using a tracer gas method. That study was a survey of AERs in 16 small office buildings, 6 large
office buildings, 13 retail establishments, and 14 schools. We plan to obtain and analyze these
data and compare those results with the Turk and Persily studies.
Due to the small sample size of the Turk data, the data were analyzed without stratification by
building type and/or season. For the Persily data, the AER values for each office space were
averaged, rather using the individual measurements, to account for the strong dependence of the
AER measurements for the same office space over a relatively short period.
Summary statistics of AER and log (AER) for the two studies are presented in Table A-4.
Table A-4. AER summary statistics for offices and other non-residential buildings
Study
Persily
Turk
Persily
Turk
Variable
AER
AER
Log(AER)
Log(AER)
N
96
40
96
40
Mean
1.9616
1.5400
0.1038
0.2544
Std Dev
2.3252
0.8808
1.1036
0.6390
Min
0.0712
0.3000
-2.6417
-1.2040
25th %ile
0.5009
0.8500
-0.6936
-0.1643
Median
1.0795
1.5000
0.0765
0.4055
75th %ile
2.7557
2.0500
1.0121
0.7152
Max
13.8237
4.1000
2.6264
1.4110
B-217
-------
The mean values are similar for the two studies, but the standard deviations are about twice as
high for the Persily data. The proposed AER distributions were derived from the more recent
Persily data only.
Similarly to the analyses of the residential AER distributions, we fitted exponential, log-normal,
normal, and Weibull distributions to the 96 office space average AER values. The results are
shown in Table A-5.
Table A-5. Best fitting office AER distributions from the Persily et al. (2004, 2005)
Scale
1.9616
0.1038
1.9197
Shape
1.1036
0.9579
Mean
1.9616
2.0397
1.9616
1.9568
Std_Dev
1.9616
3.1469
2.3252
2.0433
Distribution
Exponential
Lognormal
Normal
Weibull
P-Value
Kolmogorov-
Smirnov
0.13
0.15
0.01
P-Value
Cramer-
von
Mises
0.04
0.46
0.01
0.01
P-Value
Anderson-
Darling
0.05
0.47
0.01
0.01
(For an explanation of the Kolmogorov-Smirnov, Cramer-von Mises, and Anderson-Darling P-
values see the discussion residential AER distributions above.) According to all three goodness-
of-fit measures the best-fitting distribution is the log-normal. Reasonable choices for the lower
and upper bounds are the observed minimum and maximum AER values.
We therefore propose the following indoor, non-residential AER distributions.
• AER distribution for indoor, non-residential microenvironments: Lognormal, with scale
and shape parameters 0.1038 and 1.1036, i.e., geometric mean = 1.1094, geometric
standard deviation = 3.0150. Lower Bound = 0.07. Upper bound = 13.8.
Proximity and Penetration Factors For Outdoors, In-vehicle, and Mass Transit
For the APEX modeling of the outdoor, in-vehicle, and mass transit micro-environments, an
approach using proximity and penetration factors is proposed, as follows.
Outdoors Near Road
Penetration factor = 1.
For the Proximity factor, we propose using ratio distributions developed from the Cincinnati
Ozone Study (American Petroleum Institute, 1997, Appendix B; Johnson et al. 1995). The field
study was conducted in the greater Cincinnati metropolitan area in August and September, 1994.
Vehicle tests were conducted according to an experimental design specifying the vehicle type,
road type, vehicle speed, and ventilation mode. Vehicle types were defined by the three study
vehicles: a minivan, a full-size car, and a compact car. Road types were interstate highways
(interstate), principal urban arterial roads (urban), and local roads (local). Nominal vehicle
B-218
-------
speeds (typically met over one minute intervals within 5 mph) were at 35 mph, 45 mph, or 55
mph. Ventilation modes were as follows:
• Vent Open: Air conditioner off. Ventilation fan at medium. Driver's window half open.
Other windows closed.
• Normal A/C. Air conditioner at normal. All windows closed.
• Max A/C: Air conditioner at maximum. All windows closed.
Ozone concentrations were measured inside the vehicle, outside the vehicle, and at six fixed site
monitors in the Cincinnati area.
The proximity factor can be estimated from the distributions of the ratios of the outside-vehicle
ozone concentrations to the fixed-site ozone concentrations, reported in Table 8 of Johnson et al.
(1995). Ratio distributions were computed by road type (local, urban, interstate, all) and by the
fixed-site monitor (each of the six sites, as well as the nearest monitor to the test location). For
this analysis we propose to use the ratios of outside-vehicle concentrations to the concentrations
at the nearest fixed site monitor, as shown in Table A-6.
Table A-6. Ratio of outside-vehicle ozone to ozone at nearest fixed site1
Road
Type1
Local
Urban
Interstate
All
Number
of cases1
191
299
241
731
Mean1
0.755
0.754
0.364
0.626
Standard
Deviation1
0.203
0.243
0.165
0.278
25th
Percentile1
0.645
0.585
0.232
0.417
50th
Percentile1
0.742
0.722
0.369
0.623
75th
Percentile1
0.911
0.896
0.484
0.808
Estimated
5th
Percentile2
0.422
0.355
0.093
0.170
1. From Table 8 of Johnson et al. (1995). Data excluded if fixed-site concentration < 40
ppb.
2. Estimated using a normal approximation as Mean - 1.64 x Standard Deviation
For the outdoors-near- road microenvironment, we recommend using the distribution for local
roads, since most of the outdoors-near-road ozone exposure will occur on local roads. The
summary data from the Cincinnati Ozone Study are too limited to allow fitting of distributions,
but the 25th and 75th percentiles appear to be approximately equidistant from the median (50th
percentile). Therefore we propose using a normal distribution with the observed mean and
standard deviation. A plausible upper bound for the proximity factor equals 1. Although the
normal distribution allows small positive values and can even produce impossible, negative
values (with a very low probability), the titration of ozone concentrations near a road is limited.
Therefore, as an empirical approach, we recommend a lower bound of the estimated 5th
percentile, as shown in the final column of the above table. Therefore in summary we propose:
• Penetration factor for outdoors, near road: 1.
B-219
-------
• Proximity factor for outdoors, near road: Normal distribution. Mean = 0.755. Standard
Deviation = 0.203. Lower Bound = 0.422. Upper Bound = 1.
Outdoors, Public Garage / Parking Lot
This micro-environment is similar to the outdoors-near-road microenvironment. We therefore
recommend the same distributions as for outdoors-near-road:
• Penetration factor for outdoors, public garage / parking lot: 1.
• Proximity factor for outdoors, public garage / parking lot: Normal distribution. Mean =
0.755. Standard Deviation = 0.203. Lower Bound = 0.422. Upper Bound = 1.
Outdoors, Other
The outdoors, other ozone concentrations should be well represented by the ambient monitors.
Therefore we propose:
• Penetration factor for outdoors, other: 1.
• Proximity factor for outdoors, other: 1.
In-Vehicle
For the proximity factor for in-vehicle, we also recommend using the results of the Cincinnati
Ozone Study presented in Table A-6. For this microenvironment, the ratios depend upon the road
type, and the relative prevalences of the road types can be estimated by the proportions of
vehicle miles traveled in each city. The proximity factors are assumed, as before, to be normally
distributed, the upper bound to be 1, and the lower bound to be the estimated 5th percentile.
• Proximity factor for in-vehicle, local roads: Normal distribution. Mean = 0.755. Standard
Deviation = 0.203. Lower Bound = 0.422. Upper Bound = 1.
• Proximity factor for in-vehicle, urban roads: Normal distribution. Mean = 0.754.
Standard Deviation = 0.243. Lower Bound = 0.355. Upper Bound = 1.
• Proximity factor for in-vehicle, interstates: Normal distribution. Mean = 0.364. Standard
Deviation = 0.165. Lower Bound = 0.093. Upper Bound = 1.
To complete the specification, the distribution of road type needs to be estimated for each city to
be modeled. Vehicle miles traveled (VMT) in 2003 by city (defined by the Federal-Aid
urbanized area) and road type were obtained from the Federal Highway Administration.
(http://www.fhwa.dot.gov/policy/ohim/hs03/htm/hm71.htm). For local and interstate road types,
the VMT for the same DOT categories were used. For urban roads, the VMT for all other road
types was summed (Other freeways/expressways, Other principal arterial, Minor arterial,
Collector). The computed VMT ratios for each city are shown in Table A-7.
B-220
-------
Table A-7. Vehicle Miles Traveled by City and Road Type in 2003 (FHWA, October 2004)
FEDERAL-AID URBANIZED
AREA
Atlanta
Boston
Chicago
Cleveland
Detroit
Houston
Los Angeles
New York
Philadelphia
Sacramento
St. Louis
Washington
FRACTION VMT BY ROAD TYPE
INTERSTATE
0.38
0.31
0.30
0.39
0.26
0.24
0.29
0.18
0.23
0.21
0.36
0.31
URBAN
0.45
0.55
0.59
0.45
0.63
0.72
0.65
0.67
0.65
0.69
0.45
0.61
LOCAL
0.18
0.14
0.12
0.16
0.11
0.04
0.06
0.15
0.11
0.09
0.19
0.08
Note that a "Federal-Aid Urbanized Area" is an area with 50,000 or more persons that at a
minimum encompasses the land area delineated as the urbanized area by the Bureau of the
Census. Urbanized areas that have been combined with others for reporting purposes are not
shown separately. The Illinois portion of Round Lake Beach-McHenry-Grayslake has been
reported with Chicago.
Thus to simulate the proximity factor in APEX, we propose to first select the road type according
to the above probability table of road types, then select the AER distribution (normal) for that
road type as defined in the last set of bullets.
For the penetration factor for in-vehicle, we recommend using the inside-vehicle to outside-
vehicle ratios from the Cincinnati Ozone Study. The ratio distributions were summarized for all
the data and for stratifications by vehicle type, vehicle speed, road type, traffic (light, moderate,
or heavy), and ventilation. The overall results and results by ventilation type are shown in Table
A-8.
Table A-8. Ratio of inside-vehicle ozone to outside-vehicle ozone1
Ventilation1
Vent Open
Normal
A/C
Maximum
A/C
All
Number
of
cases1
226
332
254
812
Mean1
0.361
0.417
0.093
0.300
Standard
Deviation1
0.217
0.211
0.088
0.232
25th
Percentile1
0.199
0.236
0.016
0.117
50th
Percentile1
0.307
0.408
0.071
0.251
75th
Percentile1
0.519
0.585
0.149
0.463
Estimated
5th
Percentile2
0.005
0.071
O.OOO3
O.OOO3
B-221
-------
1. From Table 7 of Johnson et al.(1995). Data excluded if outside-vehicle concentration <
20 ppb.
2. Estimated using a normal approximation as Mean - 1.64 x Standard Deviation
3. Negative estimate (impossible value) replaced by zero.
Although the data in Table A-8 indicate that the inside-to-outside ozone ratios strongly depend
upon the ventilation type, it would be very difficult to find suitable data to estimate the
ventilation type distributions for each modeled city. Furthermore, since the Cincinnati Ozone
Study was scripted, the ventilation conditions may not represent real-world vehicle ventilation
scenarios. Therefore, we propose to use the overall average distributions.
• Penetration factor for in-vehicle: Normal distribution. Mean = 0.300. Standard Deviation
= 0.232. Lower Bound = 0.000. Upper Bound = 1.
Mass Transit
The mass transit microenvironment is expected to be similar to the in-vehicle microenvironment.
Therefore we recommend using the same APEX modeling approach:
• Proximity factor for mass transit, local roads: Normal distribution. Mean = 0.755.
Standard Deviation = 0.203. Lower Bound = 0.422. Upper Bound = 1.
• Proximity factor for mass transit, urban roads: Normal distribution. Mean = 0.754.
Standard Deviation = 0.243. Lower Bound = 0.355. Upper Bound = 1.
• Proximity factor for mass transit, interstates: Normal distribution. Mean = 0.364.
Standard Deviation = 0.165. Lower Bound = 0.093. Upper Bound = 1.
• Road type distributions for mass transit: See Table A-6
• Penetration factor for mass transit: Normal distribution. Mean = 0.300. Standard
Deviation = 0.232. Lower Bound = 0.000. Upper Bound = 1.
References
American Petroleum Institute (1997). Sensitivity testing ofpNEM/O3 exposure to changes in the
model algorithms. Health and Environmental Sciences Department.
Avol, E. L., W. C. Navidi, and S. D. Colome (1998) Modeling ozone levels in and around
southern California homes. Environ. Sci. Technol. 32, 463-468.
Chilrud, S. N., D. Epstein, J. M. Ross, S. N. Sax, D. Pederson, J. D. Spengler, P. L. Kinney
(2004). Elevated airborne exposures of teenagers to manganese, chromium, and iron from steel
dust and New York City's subway system. Environ. Sci. Technol. 38, 732-737.
Colome, S.D., A. L. Wilson, Y. Tian (1993). California Residential Indoor Air Quality Study,
Volume 1, Methodology and Descriptive Statistics. Report prepared for the Gas Research
Institute, Pacific Gas & Electric Co., San Diego Gas & Electric Co., Southern California Gas Co.
B-222
-------
Colome, S.D., A. L. Wilson, Y. Tian (1994). California Residential Indoor Air Quality Study,
Volume 2, Carbon Monoxide and Air Exchange Rate: An Univariate andMultivariate Analysis.
Chicago, IL. Report prepared for the Gas Research Institute, Pacific Gas & Electric Co., San
Diego Gas & Electric Co., Southern California Gas Co. GRI-93/0224.3
Jakobi, G and Fabian, P. (1997). Indoor/outdoor concentrations of ozone and peroxyacetyl nitrate
(PAN). Int. J. Biometeorol. 40: 162-165..
Johnson, T., A. Pakrasi, A. Wisbeth, G. Meiners, W. M. Ollison (1995). Ozone exposures within
motor vehicles - results of a field study in Cincinnati, Ohio. Proceedings 88th annual meeting
and exposition of the Air & Waste Management Association, June 18-23, 1995. San Antonio,
TX. Preprint paper 95-WA84A.02.
Kinney, P. L., S. N. Chillrud, S. Ramstrom, J. Ross, J. D. Spengler (2002). Exposures to multiple
air toxics in New York City. Environ Health Perspect 110, 539-546.
Lagus Applied Technology, Inc. (1995) Air change rates in non-residential buildings in
California. Sacramento CA, California Energy Commission, contract 400-91-034.
Liu, L.-J. S, P. Koutrakis, J. Leech, I. Broder, (1995) Assessment of ozone exposures in the
greater metropolitan Toronto area. J. Air Waste Manage. Assoc. 45: 223-234.
Meng, Q. Y., B. J. Turpin, L. Korn, C. P. Weisel, M. Morandi, S. Colome, J. J. Zhang, T. Stock,
D. Spektor, A. Winer, L. Zhang, J. H. Lee, R. Giovanetti, W. Cui, J. Kwon, S. Alimokhtari, D.
Shendell, J. Jones, C. Farrar, S. Maberti (2004). Influence of ambient (outdoor) sources on
residential indoor and personal PM2 5 concentrations: Analyses of RIOPA data. Journal of
Exposure Analysis and Environ Epidemiology. Preprint.
Murray, D. M. and D. E. Burmaster (1995). Residential Air Exchange Rates in the United
States: Empirical and Estimated Parametric Distributions by Season and Climatic Region. Risk
Analysis, Vol. 15, No. 4, 459-465.
Persily, A. and J. Gorfain.(2004). Analysis of ventilation data from the U.S. Environmental
Protection Agency Building Assessment Survey and Evaluation (BASE) Study. National Institute
of Standards and Technology, NISTIR 7145, December 2004.
Persily, A., J. Gorfain, G. Brunner.(2005). Ventilation design and performance in U.S. office
buildings. ASHRAE Journal. April 2005, 30-35.
Sax, S. N., D. H. Bennett, S. N. Chillrud, P. L. Kinney, J. D. Spengler (2004) Differences in
source emission rates of volatile organic compounds in inner-city residences of New York City
and Los Angeles. Journal of Exposure Analysis and Environ Epidemiology. Preprint.
Turk, B. H., D. T. Grimsrud, J. T. Brown, K. L. Geisling-Sobotka, J. Harrison, R. J. Prill (1989).
Commercial building ventilation rates and particle concentrations. ASHRAE, No. 3248.
B-223
-------
Weschler, C. J. (2000) Ozone in indoor environments: concentration and chemistry. Indoor Air
10: 269-288.
Weschler, C. J. and Shields, H. C. (2000) The influence of ventilation on reactions among indoor
pollutants: modeling and experimental observations. Indoor Air. 10: 92-100.
Weisel, C. P., J. J. Zhang, B. J. Turpin, M. T. Morandi, S. Colome, T. H. Stock, D. M. Spektor,
L. Korn, A. Winer, S. Alimokhtari, J. Kwon, K. Mohan, R. Harrington, R. Giovanetti, W. Cui,
M. Afshar, S. Maberti, D. Shendell (2004). Relationship of Indoor, Outdoor and Personal Air
(RIOPA) study; study design, methods and quality assurance / control results. Journal of
Exposure Analysis and Environ Epidemiology. Preprint.
Williams, R., J. Suggs, A. Rea, K. Leovic, A. Vette, C. Croghan, L. Sheldon, C. Rodes, J.
Thornburg, A. Ejire, M. Herbst, W. Sanders Jr. (2003a). The Research Triangle Park particulate
matter panel study: PM mass concentration relationships. Atmos Env 37, 5349-5363.
Williams, R., J. Suggs, A. Rea, L. Sheldon, C. Rodes, J. Thornburg (2003b). The Research
Triangle Park particulate patter panel study: modeling ambient source contribution to personal
and residential PM mass concentrations. Atmos Env 37, 5365-5378.
Wilson, A. L., S. D. Colome, P. E. Baker, E. W. Becker (1986). Residential Indoor Air Quality
Characterization Study of Nitrogen Dioxide, Phase I, Final Report. Prepared for Southern
California Gas Company, Los Angeles.
Wilson, A. L., S. D. Colome, Y. Tian, P. E. Baker, E. W. Becker, D. W. Behrens, I. H. Billick,
C. A. Garrison (1996). California residential air exchange rates and residence volumes. Journal
of Exposure Analysis and Environ Epidemiology. Vol. 6, No. 3.
B-224
-------
ATTACHMENT 7. TECHNICAL MEMORANDUM ON THE
UNCERTAINTY ANALYSIS OF RESIDENTIAL AIR EXCHANGE
RATE DISTRIBUTIONS
B-225
-------
INTERNATIONAL
MEMORANDUM
To: John Langstaff, EPA OAQPS
From: Jonathan Cohen, Arlene Rosenbaum, ICF International
Date: June 5, 2006
Re: Uncertainty analysis of residential air exchange rate distributions
This memorandum describes our assessment of some of the sources of the uncertainty of city-
specific distributions of residential air exchange rates that were fitted to the available study data.
City-specific distributions for use with the APEX ozone model were developed for 12 modeling
cities, as detailed in the memorandum by Cohen, Mallya and Rosenbaum, 20056 (Appendix A of
this report). In the first part of the memorandum, we analyze the between-city uncertainty by
examining the variation of the geometric means and standard deviations across cities and studies.
In the second part of the memorandum, we assess the within-city uncertainty by using a
bootstrap distribution to estimate the effects of sampling variation on the fitted geometric means
and standard deviations for each city. The bootstrap distributions assess the uncertainty due to
random sampling variation but do not address uncertainties due to the lack of representativeness
of the available study data, the matching of the study locations to the modeled cities, and the
variation in the lengths of the AER monitoring periods.
Variation of geometric means and standard deviations across cities and studies
The memorandum by Cohen, Mallya and Rosenbaum, 2005 (Attachment 6 of this report)
describes the analysis of residential air exchange rate (AER) data that were obtained from seven
studies. The AER data were subset by location, outside temperature range, and the A/C type, as
defined by the presence or absence of an air conditioner (central or window). In each case we
chose to fit a log-normal distribution to the AER data, so that the logarithm of the AER for a
given city, temperature range, and A/C type is assumed to be normally distributed. If the AER
data has geometric mean GM and geometric standard deviation GSD, then the logarithm of the
AER is assumed to have a normal distribution with mean log(GM) and standard deviation
log(GSD).
Table D-l shows the assignment of the AER data to the 12 modeled cities. Note that for Atlanta,
GA and Washington DC, the Research Triangle Park, NC data for houses with A/C was used to
represent the AER distributions for houses with A/C, and the non-California data for houses
without A/C was used to represent the AER distributions for houses without A/C. Sacramento,
CA AER distributions were estimated using the AER data from the inland California counties of
Sacramento, Riverside, and San Bernardino; these combined data are referred to by the City
6 Cohen, I, H. Mallya, and A. Rosenbaum. 2005. Memorandum to John Langstaff. EPA 68D01052, Work
Assignment 3-08. Analysis of Air Exchange Rate Data. September 30, 2005.
B-226
-------
Name "Inland California." St Louis, MO AER distributions were estimated using the AER data
from all states except for California and so are referred to be the City Name "Outside
California."
Table D-1. Assignment of Residential AER distributions to modeled cities
Modeled city
Atlanta, GA, A/C
Atlanta, GA, no A/C
Boston, MA
Chicago, IL
Cleveland, OH
Detroit, MI
Houston, TX
Los Angeles, CA
New York, NY
Philadelphia, PA
Sacramento
St. Louis
Washington, DC, A/C
Washington, DC, no
A/C
AER distribution
Research Triangle Park, A/C only
All non-California, no A/C ("Outside
California")
New York
New York
New York
New York
Houston
Los Angeles
New York
New York
Inland parts of Los Angeles ("Inland
California")
All non-California ("Outside California")
Research Triangle Park, A/C only
All non-California, no A/C ("Outside
California")
It is evident from Table D-l that for some of the modeled cities, some potentially large
uncertainty was introduced because we modeled their AER distributions using available data
from another city or group of cities thought to be representative of the first city on the basis of
geography and other characteristics. This was necessary for cities where we did not have any or
sufficient AER data measured in the same city that also included the necessary temperature and
A/C type information. One way to assess the impact of these assignments on the uncertainty of
the AER distributions is to evaluate the variation of the fitted log-normal distributions across the
cities with AER data. In this manner we can examine the effect on the AER distribution if a
different allocation of study data to the modeled cities had been used.
Even for the cities where we have study AER data, there is uncertainty about the fitted AER
distributions. First, the studies used different measurement and residence selection methods. In
some cases the residences were selected by a random sampling method designed to represent the
entire population. In other cases the residences were selected to represent sub-populations. For
example, for the RTF study, the residences belong to two specific cohorts: a mostly Caucasian,
non-smoking group aged at least 50 years having cardiac defibrillators living in Chapel Hill; a
B-227
-------
group of non-smoking, African Americans aged at least 50 years with controlled hypertension
living in a low-to-moderate SES neighborhood in Raleigh. The TEACH study was restricted to
residences of inner-city high school students. The RIOPA study was a random sample for Los
Angeles, but was designed to preferentially sample locations near major air toxics sources for
Elizabeth, NJ and Houston TX. Furthermore, some of the studies focused on different towns or
cities within the larger metropolitan areas, so that, for example, the Los Angeles data from the
Avol study was only measured in Lancaster, Lake Gregory, Riverside, and San Dimas but the
Los Angeles data from the Wilson studies were measured in multiple cities in Southern
California. One way to assess the uncertainty of the AER distributions due to variations of study
methodologies and study sampling locations is to evaluate the variation of the fitted log-normal
distributions within each modeled city across the different studies.
We evaluated the variation between cities, and the variation within cities and between studies, by
tabulating and plotting the AER distributions for all the study/city combinations. Since the
original analyses by Cohen, Mallya and Rosenbaum, 2005 clearly showed that the AER
distribution depends strongly on the outside temperature and the A/C type (whether or not the
residence has air conditioning), this analysis was stratified by the outside temperature range and
the A/C type. Otherwise, study or city differences would have been confounded by the
temperature and A/C type differences and you would not be able to tell how much of the AER
difference was due to the variation of temperature and A/C type across cities or studies. In order
to be able to compare cities and studies we could not use different temperature ranges for the
different modeled cities as we did for the original AER distribution modeling. For these analyses
we stratified the temperature into the ranges <= 10, 10-20, 20-25, and >25 °C and categorized the
A/C type as "Central or Window A/C" versus 'No A/C," giving 8 temperature and A/C type
strata.
Table D-2 shows the geometric means and standard deviations by city and study. These
geometric mean and standard deviation pairs are plotted in Figure D-l through D-8. Each figure
shows the variation across cities and studies for a given temperature range and A/C type. The
results for a city with only one available study are shown with a blank study name. For cities
with multiple studies, results are shown for the individual studies and the city overall distribution
is designated by a blank value for the study name.
Table D-2. Geometric means and standard deviations by city and study.
A/C Type
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Temperature
<=10
<=10
<= 10
<=10
<= 10
<=10
<=10
<= 10
<=10
<= 10
10-20
10-20
City
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
Research Triangle Park
Sacramento
San Francisco
Stockton
Arcata
Bakersfield
Study*
Avol
RIOPA
Wilson 1991
N
2
5
2
1
2
20
157
3
2
7
1
2
Geo Mean
0.32
0.62
0.72
0.31
0.77
0.71
0.96
0.38
0.43
0.48
0.17
0.36
Geo Std Dev**
1.80
1.51
1.22
1.12
2.02
1.81
1.82
1.00
1.64
1.34
B-228
-------
Table D-2. Geometric means and standard deviations by city and study.
A/C Type
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
Temperature
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
>25
>25
>25
>25
>25
>25
>25
>25
>25
<=10
<= 10
<=10
<= 10
<= 10
<=10
<= 10
<=10
<= 10
City
Fresno
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Redding
Research Triangle Park
Sacramento
San Diego
San Francisco
Santa Maria
Stockton
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Red Bluff
Research Triangle Park
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Research Triangle Park
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Sacramento
Study*
Avol
RIOPA
TEACH
Wilson 1984
Wilson 1991
RIOPA
TEACH
Avol
RIOPA
Wilson 1984
RIOPA
TEACH
Avol
RIOPA
Wilson 1984
RIOPA
TEACH
Avol
RIOPA
Wilson 1991
RIOPA
TEACH
N
8
13
716
33
11
1
634
37
5
4
1
1
320
7
23
5
1
4
20
273
32
26
215
37
20
17
2
196
79
114
25
10
79
19
14
5
145
13
18
14
2
2
48
44
4
3
Geo Mean
0.30
0.42
0.59
0.48
0.60
0.68
0.59
0.64
1.36
1.20
2.26
0.31
0.56
0.26
0.41
0.42
0.23
0.73
0.47
1.10
0.61
0.90
1.23
1.11
0.93
1.37
0.61
0.40
0.43
0.72
0.37
0.94
0.86
1.24
1.23
1.29
0.38
0.66
0.54
0.51
0.72
0.60
1.02
1.04
0.79
0.58
Geo Std Dev**
1.62
2.19
1.90
1.87
1.87
1.89
2.11
2.34
2.53
1.91
1.67
1.55
1.25
1.42
1.94
2.36
1.95
2.42
2.33
2.74
2.91
2.52
3.20
1.89
2.17
2.60
3.10
1.71
2.33
2.18
2.28
2.04
1.71
1.68
3.09
3.60
1.11
1.00
2.14
2.20
1.28
1.30
B-229
-------
Table D-2. Geometric means and standard deviations by city and study.
A/C Type
NoA/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
No A/C
Temperature
<= 10
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
20-25
>25
>25
>25
>25
>25
>25
>25
>25
>25
City
San Francisco
Bakersfield
Fresno
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
Sacramento
San Diego
San Francisco
Santa Maria
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Red Bluff
Houston
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
Study*
Avol
RIOPA
TEACH
Wilson 1984
Wilson 1991
Avol
RIOPA
Wilson 1984
RIOPA
TEACH
Avol
RIOPA
TEACH
Wilson 1984
RIOPA
TEACH
N
9
1
4
28
390
23
87
9
241
30
59
1
49
15
2
10
148
19
38
91
26
19
7
1
2
25
6
4
3
12
6
3
3
Geo Mean
0.39
0.85
0.90
0.63
0.75
0.78
0.78
2.32
0.70
0.75
0.79
1.09
0.47
0.34
0.27
0.92
1.37
0.95
.30
.52
.62
.50
.99
0.55
0.92
0.99
1.56
1.33
0.86
0.74
1.54
1.73
1.37
Geo Std Dev**
1.42
2.42
2.92
2.09
2.55
1.96
2.05
2.06
1.82
2.04
1.95
3.05
1.23
2.41
2.28
1.87
2.11
2.40
2.24
2.30
2.11
3.96
1.97
1.36
1.37
1.02
2.29
1.65
2.00
1.38
* For a given city, if AER data were available from only one study, then the study name is missing.
for two or more studies, then the overall city distribution is shown in the row where the study name
distributions by study and city are shown in the rows with a specific study name.
** The geometric standard deviation is undefined if the sample size equals 1.
If AER data were available
is missing, and the
In general, there is a relatively wide variation across different cities. This implies that the AER
modeling results would be very different if the matching of modeled cities to study cities was
changed, although a sensitivity study using the APEX model would be needed to assess the
impact on the ozone exposure estimates. In particular the ozone exposure estimates may be
sensitive to the assumption that the St. Louis AER distributions can be represented by the
combined non-California AER data. One way to address this is to perform a Monte Carlo
analysis where the first stage is to randomly select a city outside of California, the second stage
picks the A/C type, and the third stage picks the AER value from the assigned distribution for the
city, A/C type and temperature range. Note that this will result in a very different distribution to
B-230
-------
the current approach that fits a single log-normal distribution to all the non-California data for a
given temperature range and A/C type. The current approach weights each data point equally, so
that cities like New York with most of the data values get the greatest statistical weight. The
Monte Carlo approach gives the same total statistical weight for each city and fits a mixture of
log-normal distributions rather than a single distribution.
In general, there is also some variation within studies for the same city, but this is much smaller
than the variation across cities. This finding tends to support the approach of combining different
studies. Note that the graphs can be deceptive in this regard because some of the data points are
based on very small sample sizes (N); those data points are less precise and the differences
would not be statistically significant. For example, for the No A/C data in the range 10-20 °C,
the Los Angeles TEACH study had a geometric mean of 2.32 based on only nine AER values,
but the overall geometric mean, based on 390 values, was 0.75 and the geometric means for the
Los Angeles Avol, RIOPA, Wilson 1984, and Wilson 1991 studies were each close to 0.75. One
noticeable case where the studies show big differences for the same city is for the A/C houses in
Los Angeles in the range 20-25 °C where the study geometric means are 0.61 (Avol, N=32), 0.90
(RIOPA, N=26) and 1.23 (Wilson 1984, N=215).
Bootstrap analyses
The 39 AER subsets defined in the Cohen, Mallya, and Rosenbaum, 2005 memorandum
(Appendix A of this report) and their allocation to the 12 modeled cities are shown in Table D-3.
To make the distributions sufficiently precise in each AER subset and still capture the variation
across temperature and A/C type, different modeled cities were assigned different temperature
range and A/C type groupings. Therefore these temperature range groupings are sometimes
different to those used to develop Table D-2 and Figure D-l through D-8.
Table D-3. AER subsets by city, A/C type, and temperature range.
Subset City
Name
Houston
Houston
Houston
Houston
Houston
Houston
Inland California
Inland California
Inland California
Inland California
Inland California
Study Cities
Houston
Houston
Houston
Houston
Houston
Houston
Houston
Sacramento, Riverside,
and San Bernardino
counties, CA
Sacramento, Riverside,
and San Bernardino
counties, CA
Sacramento, Riverside,
and San Bernardino
counties, CA
Sacramento, Riverside,
and San Bernardino
counties, CA
Sacramento, Riverside,
Represents
Modeled Cities:
Houston, TX
Houston, TX
Houston, TX
Houston, TX
Houston, TX
Houston, TX
Houston, TX
Sacramento, CA
Sacramento, CA
Sacramento, CA
Sacramento, CA
Sacramento, CA
A/C Type
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
No A/C
No A/C
No A/C
Central or Room A/C
Central or Room A/C
No A/C
No A/C
No A/C
Temperature
Range (°C)
<=20
20-25
25-30
>30
<=10
10-20
>20
<=25
>25
<=10
10-20
20-25
B-231
-------
Table D-3. AER subsets by city, A/C type, and temperature range.
Subset City
Name
Inland California
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
New York City
Study Cities
and San Bernardino
counties, CA
Sacramento, Riverside,
and San Bernardino
counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
Los Angeles, Orange,
Riverside, San
Bernardino, and
Ventura counties, CA
New York, NY
New York, NY
New York, NY
New York, NY
Represents
Modeled Cities:
Sacramento, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Los Angeles, CA
Boston, MA,
Chicago, IL,
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
Boston, MA,
Chicago, IL,
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
Boston, MA,
Chicago, IL,
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
Boston, MA,
Chicago, IL,
A/C Type
NoA/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
No A/C
No A/C
No A/C
No A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
No A/C
Temperature
Range (°C)
>25
<=20
20-25
25-30
>30
<=10
10-20
20-25
>25
<=10
10-25
>25
<=10
B-232
-------
Table D-3. AER subsets by city, A/C type, and temperature range.
Subset City
Name
New York City
New York City
Outside California
Outside California
Outside California
Outside California
Outside California
Outside California
Outside California
Outside California
Research Triangle Park
Research Triangle Park
Research Triangle Park
Research Triangle Park
Study Cities
New York, NY
New York, NY
Cities outside CA
Cities outside CA
Cities outside CA
Cities outside CA
Cities outside CA
Cities outside CA
Cities outside CA
Cities outside CA
Research Triangle
Park, NC
Research Triangle
Park, NC
Research Triangle
Park, NC
Research Triangle
Park, NC
Represents
Modeled Cities:
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
Boston, MA,
Chicago, IL,
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
Boston, MA,
Chicago, IL,
Cleveland, OH,
Detroit, MI,
New York, NY,
Philadelphia, PA
St. Louis, MO
St. Louis, MO
St. Louis, MO
St. Louis, MO
St. Louis, MO
St. Louis, MO
Atlanta, GA
Washington DC
St. Louis, MO
Atlanta, GA
Washington DC
St. Louis, MO
Atlanta, GA
Washington DC
Atlanta, GA
Washington DC
Atlanta, GA
Washington DC
Atlanta, GA
Washington DC
Atlanta, GA
Washington DC
A/C Type
NoA/C
No A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
No A/C
No A/C
No A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Central or Room A/C
Temperature
Range (°C)
10-20
>20
<=10
10-20
20-25
25-30
>30
<=10
10-20
>20
<=10
10-20
20-25
>25
The GM and GSD values that define the fitted log-normal distributions for these 39 AER subsets
are shown in Table D-4. Examples of these pairs are also plotted in Figures D-9 through D-19, to
be further described below. Each of the example figures D-9 through D-19 corresponds to a
single GM/GSD "Original Data" pair. The GM and GSD values for the "Original Data" are at
the intersection of the horizontal and vertical lines that are parallel to the x- and y-axes in the
figures.
B-233
-------
Table D-4. Geometric means and standard deviations for AER subsets by city, A/C type,
and temperature range.
Subset City
Name
Houston
Houston
Houston
Houston
Houston
Houston
Inland California
Inland California
Inland California
Inland California
Inland California
Inland California
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
New York City
New York City
New York City
New York City
New York City
New York City
Outside California
Outside California
Outside California
Outside California
A/C Type
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
No A/C
No A/C
No A/C
Central or Room
A/C
Central or Room
A/C
No A/C
No A/C
No A/C
No A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
No A/C
No A/C
No A/C
No A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
No A/C
No A/C
No A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Temperature
Range (°C)
<=20
20-25
25-30
>30
<=10
10-20
>20
<=25
>25
<=10
10-20
20-25
>25
<=20
20-25
25-30
>30
<=10
10-20
20-25
>25
<=10
10-25
>25
<=10
10-20
>20
<=10
10-20
20-25
25-30
N
15
20
65
14
13
28
12
226
83
17
52
13
14
721
273
102
12
18
390
148
25
20
42
19
48
59
32
179
338
253
219
Geometric
Mean
0.4075
0.4675
0.4221
0.4989
0.6557
0.6254
0.9161
0.5033
0.8299
0.5256
0.6649
1.0536
0.8271
0.5894
1.1003
0.8128
0.2664
0.5427
0.7470
1.3718
0.9884
0.7108
1.1392
1.2435
1.0165
0.7909
1.6062
0.9185
0.5636
0.4676
0.4235
Geometric
Standard
Deviation
2.1135
1.9381
2.2579
1.7174
1.6794
2.9162
2.4512
1.9210
2.3534
3.1920
2.1743
1.7110
2.2646
1.8948
2.3648
2.4151
2.7899
3.0872
2.0852
2.2828
1.9666
2.0184
2.6773
2.1768
2.1382
2.0417
2.1189
1.8589
1.9396
2.2011
2.0373
B-234
-------
Table D-4. Geometric means and standard deviations for AER subsets by city, A/C type,
and temperature range.
Subset City
Name
Outside California
Outside California
Outside California
Outside California
Research Triangle
Park
Research Triangle
Park
Research Triangle
Park
Research Triangle
Park
A/C Type
Central or Room
A/C
No A/C
No A/C
No A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Central or Room
A/C
Temperature
Range (°C)
>30
<=10
10-20
>20
<=10
10-20
20-25
>25
N
24
61
87
44
157
320
196
145
Geometric
Mean
0.5667
0.9258
0.7333
1.3782
0.9617
0.5624
0.3970
0.3803
Geometric
Standard
Deviation
1.9447
2.0836
2.3299
2.2757
1.8094
1.9058
1.8887
1.7092
To evaluate the uncertainty of the GM and GSD values, a bootstrap simulation was performed,
as follows. Suppose that a given AER subset has N values. A bootstrap sample is obtained by
sampling N times at random with replacement from the N AER values. The first AER value in
the bootstrap sample is selected randomly from the N values, so that each of the N values is
equally likely. The second, third, ..., N'th values in the bootstrap sample are also selected
randomly from the N values, so that for each selection, each of the N values is equally likely.
The same value can be selected more than once. Using this bootstrap sample, the geometric
mean and geometric standard deviation of the N values in the bootstrap sample was calculated.
This pair of values is plotted as one of the points in a figure for that AER subset. 1,000 bootstrap
samples were randomly generated for each AER subset, producing a set of 1,000 geometric mean
and geometric standard deviation pairs, which were plotted in example Figures D-9 through D-
19.
The bootstrap distributions display the part of the uncertainty of the GM and GSD that is entirely
due to random sampling variation. The analysis is based on the assumption that the study AER
data are a random sample from the population distribution of AER values for the given city,
temperature range, and A/C type. On that basis, the 1,000 bootstrap GM and GSD pairs estimate
the variation of the GM and GSD across all possible samples of N values from the population.
Since each GM, GSD pair uniquely defines a fitted log-normal distribution, the pairs also
estimate the uncertainty of the fitted log-normal distribution. The choice of 1,000 was made as a
compromise between having enough pairs to accurately estimate the GM, GSD distribution and
not having too many pairs so that the graph appears as a smudge of overlapped points. Note that
even if there were infinitely many bootstrap pairs, the uncertainty distribution would still be an
estimate of the true uncertainty because the N is finite, so that the empirical distribution of the N
measured AER values does not equal the unknown population distribution.
In most cases the uncertainty distribution appears to be a roughly circular or elliptical geometric
mean and standard deviation region. The size of the region depends upon the sample size and on
the variability of the AER values; the region will be smallest when the sample size N is large
B-235
-------
and/or the variability is small, so that there are a large number of values that are all close
together.
The bootstrap analyses show that the geometric standard deviation uncertainty for a given
CMSA/air-conditioning-status/temperature-range combination tends to have a range of at most
from "fitted GSD-1.0 hr"1" to "fitted GSD+1.0 hr"1", but the intervals based on larger AER
sample sizes are frequently much narrower. The ranges for the geometric means tend to be
approximately from "fitted GM-0.5 hr"1" to "fitted GM+0.5 hr"1", but in some cases were much
smaller.
The bootstrap analysis only evaluates the uncertainty due to the random sampling. It does not
account for the uncertainty due to the lack of representativeness, which in turn is due to the fact
that the samples were not always random samples from the entire population of residences in a
city, and were sometimes used to represent different cities. Since only the GM and GSD were
used, the bootstrap analyses does not account for uncertainties about the true distributional
shape, which may not necessarily be log-normal. Furthermore, the bootstrap uncertainty does not
account for the effect of the calendar year (possible trends in AER values) or of the uncertainty
due to the AER measurement period; the distributions were intended to represent distributions of
24 hour average AER values although the study AER data were measured over a variety of
measurement periods.
To use the bootstrap distributions to estimate the impact of sample size on the fitted distributions,
a Monte Carlo approach could be used with the APEX model. Instead of using the Original Data
distributions, a bootstrap GM, GSD pair could be selected at random and the AER value could be
selected randomly from the log-normal distribution with the bootstrap GM and GSD.
B-236
-------
4.0—
3.5—
Figure D-1
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: Central or Room A/C
Temperature Range: <= 10 Degrees Celsius
Q
•o
,0
^H
"5
o
3.0—
2.5—
2.0—
1.5—
1.0—
E
AG F
T
B
cn
H
III 1
0.5
1.0 1.5 2.0
Geometric Mean
2.5
3.0
AAAHouston
E E ENewYorkCity
I I I Stockton
B B BLosAngeles C C CLosAngeles-Avol
F F FResearchTnanglePark GGGSacramento
DDDLosAngeles-Wilson 1991
H H Ffs anFranci sco
B-237
-------
Figure D-2
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: Central or Room A/C
Temperature Range: 10-20 Degrees Celsius
4.0-
3.5-
-------
Figure D-3
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
Q
T3
-t—»
(Si
•c
"5
o
-------
Q
T3
-t—»
(Si
•c
"5
o
25 Degrees Celsius
A
D
H
0.0
0.5
1.0 1.5
Geometric Mean
2.0
2.5
3.0
AAAHouston B B BLosAngeles
E E ELosAngeles-Wilsonl984 F F FNewYorkCity
I I I ResearchTnanglePark
C C CLosAngeles-Avol D D DLosAngeles-RIOPA
GGGNewYorkCity-RIOPA HHHNewYorkCity-TEACH
B-240
-------
Q
T3
-t— »
(Si
•c
"5
o
-------
Figure D-6
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: No A/C
Temperature Range: 10-20 Degrees Celsius
Q
T3
-t—»
(Si
•c
"5
o
-------
Figure D-7
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: No A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5—
CD
Q 3.0H
.g 2.5—
"CD
o 2 0—
CD ^••V I
O
1.5—
1.0—
A
D
H
0.0
0.5
1.0 1.5
Geometric Mean
2.0
2.5
3.0
AAAHouston B B BLosAngeles
E E ELosAngeles-Wilson 1984 F F FNewYorkCity
C C CLosAngeles-Avol
G G GNewYorkCity-RIOP A
D D DLosAngeles-RIOPA
H H HNewYorkCity-TEACH
B-243
-------
Q
T3
-t— »
(Si
•c
"5
o
(U
O
4.0—
3.5—
3.0H
1.5—
1.0—
Figure D-8
Geometric mean and standard deviation of air exchange rate
For different cities and studies
Air Conditioner Type: No A/C
Temperature Range: > 25 Degrees Celsius
B
H
G
DI C
E
1 1 1
0.0 0.5 1.0 1.5 2.0 2
5 3.0
Geometric Mean
AAAHouston B B BLosAngeles C C CLosAngeles-Avol DDDLosAngeles-RIOPA
E E ELosAngeles-TEACH F F FLosAngeles-Wilsonl984 GGGNewYorkCity HHHNewYorkCity-RIOPA
I I iNewYorkCity-TEACH
B-244
-------
Figure D-9
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Houston
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5 —
Q 3.0—
"3
.g 2.5—1
•s
8 2.0—
O
1.5 —
1.0—
0.0
1
0.5
1
1.0
1.5
Geometric Mean
2.0
2.5
1
3.0
Bootstrapped Data +++Original Data
B-245
-------
Figure D-10
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Houston
Air Conditioner Type: No A/C
Temperature Range:
20 Degrees Celsius
4.0—
3.5 —
1 3.0-
Geometric Std
K> K>
o u<
1 1
1.5 —
1.0—
.
• "t"*2&
*
.
^id%---
Pp^'-M^.
1 1 1 1 1 1 1
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Geometric Mean
••Bootstrapped Data +++Original Data
B-246
-------
Figure D-11
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Inland California
Air Conditioner Type: Central or Room A/C
Temperature Range: <=25 Degrees Celsius
4.0—
3.5-
3.0H
"3
oo
| 2.5-
I
8 2.0— |
o
1.5 —
1.0—
0.5
\
1.0
1.5 2.0
Geometric Mean
2.5
3.0
••Bootstrapped Data +++Original Data
B-247
-------
Figure D-12
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Inland California
Air Conditioner Type: No A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5-
Q 3.0—
"3
oo
| 2.5-
I
8 2.0—1
o
1.5 —
1.0—
1
0.5
1
1.0
1.5 2.0
Geometric Mean
2.5
1
3.0
••Bootstrapped Data +++Original Data
B-248
-------
Figure D-13
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Los Angeles
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5-
3.0H
•3
w
| 2.5-
I
8 2.0— |
o
1.5 —
1.0—
\
0.5
\
1.0
1.5 2.0
Geometric Mean
2.5
3.0
••Bootstrapped Data +++Original Data
B-249
-------
Figure D-14
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Los Angeles
Air Conditioner Type: No A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5-
I 3.0H
•3
00
•I 2-5~
1
8 2.0—1
o
1.5 —
1.0—
\
0.5
\
1.0
1.5 2.0
Geometric Mean
2.5
\^
3.0
••Bootstrapped Data +++Origiiial Data
B-250
-------
Figure D-15
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: New York City
Air Conditioner Type: Central or Room A/C
Temperature Range: 10-25 Degrees Celsius
4.0—
3.5 —
1 3.0-
Geometric Std
K> K>
o Lo
1 1
1.5 —
1.0—
•
••• V'^£
"'• '-^'^^^
. -:;J^
••fc. :'
S^SjVjI^M.- " . •
isSC*>""iHv-.'-
^^•.•-•;. ••". ' •
"./".. ***•'•.•.
i i i i i i i
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Geometric Mean
••Bootstrapped Data +++Original Data
B-251
-------
Figure D-16
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: New York City
Air Conditioner Type: No A/C
Temperature Range: >20 Degrees Celsius
4.0—
3.5-
I 3.0H
•3
00
•c 2-5~
1
8 2.0—1
o
1.5 —
1.0—
\
0.5
\
1.0
1.5 2.0
Geometric Mean
2.5
\^
3.0
••Bootstrapped Data +++Original Data
B-252
-------
Figure D-17
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Outside California
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
4.0—
3.5-
3.0H
•3
oo
| 2.5-
I
8 2.0— |
o
1.5 —
1.0—
\
0.5
\
1.0
1.5 2.0
Geometric Mean
2.5
3.0
••Bootstrapped Data +++Original Data
B-253
-------
Figure D-18
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Outside California
Air Conditioner Type: No A/C
Temperature Range:
20 Degrees Celsius
4.0—
3.5-
I 3.0H
3
00
•c 2-5~
1
8 2.0—1
o
1.5 —
1.0—
\
0.5
\
1.0
1.5
Geometric Mean
2.0
2.5
\^
3.0
••Bootstrapped Data +++Original Data
B-254
-------
Figure D-19
Geometric mean and standard deviation of air exchange rate
Bootstrapped distributions for different cities
City: Research Triangle Park
Air Conditioner Type: Central or Room A/C
Temperature Range: 20-25 Degrees Celsius
Q
"2
Geometric £
4.0—
3.5 —
3.0—
2.5 —
2.0—
1.5 —
1.0—
i
'
i
f
1 1 1 1 1 1 1
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Geometric Mean
•Bootstrapped Data +++Original Data
B-255
-------
ATTACHMENT 8. TECHNICAL MEMORANDUM ON THE
DISTRIBUTIONS OF AIR EXCHANGE RATE AVERAGES
OVER MULTIPLE DAYS
B-256
-------
INTERNATIONAL
MEMORANDUM
To: John Langstaff, EPA OAQPS
From: Jonathan Cohen, Arlene Rosenbaum, ICF International
Date: June 8, 2006
Re: Distributions of air exchange rate averages over multiple days
As detailed in the memorandum by Cohen, Mallya and Rosenbaum, 20057 (Appendix A of this
report) we have proposed to use the APEX model to simulate the residential air exchange rate
(AER) using different log-normal distributions for each combination of outside temperature
range and the air conditioner type, defined as the presence or absence of an air conditioner
(central or room).
Although the averaging periods for the air exchange rates in the study databases varied from one
day to seven days, our analyses did not take the measurement duration into account and treated
the data as if they were a set of statistically independent daily averages. In this memorandum we
present some analyses of the Research Triangle Park Panel Study that show extremely strong
correlations between consecutive 24-hour air exchange rates measured at the same house. This
provides support for the simplified approach of treating all averaging periods as if they were 24-
hour averages.
In the current version of the APEX model, there are several options for stratification of time
periods with respect to AER distributions, and for when to re-sample from a distribution for a
given stratum. The options selected for this current set of simulations resulted in a uniform AER
for each 24-hour period and re-sampling of the 24-hour AER for each simulated day. This re-
sampling for each simulated day implies that the simulated AERs on consecutive days in the
same microenvironment are statistically independent. Although we have not identified sufficient
data to test the assumption of uniform AERs throughout a 24-hour period, the analyses described
in this memorandum suggest that AERs on consecutive days are highly correlated. Therefore, we
performed sensitivity simulations to assess the impact of the assumption of temporally
independent air exchange rates, but found little difference between APEX predictions for the two
scenarios (i.e., temporally independent and autocorrelated air exchange rates).
7 Cohen, I, H. Mallya, and A. Rosenbaum. 2005. Memorandum to John Langstaff. EPA 68D01052, Work
Assignment 3-08. Analysis of Air Exchange Rate Data. September 30, 2005.
B-257
-------
Distributions of multi-day averages from the RTF Panel Study
The RTF Panel study included measurements of 24-hour averages at 38 residences for up to four
periods of at least seven days. These periods were in different seasons and/or calendar years.
Daily outside temperatures were also provided. All the residences had either window or room air
conditioners or both. We used these data to compare the distributions of daily averages taken
over 1, 2, 3, .. 7 days.
The analysis is made more complicated because the previous analyses showed the dependence of
the air exchange rate on the outside temperature, and the daily temperatures often varied
considerably. Two alternative approaches were employed to group consecutive days. For the first
approach, A, we sorted the data by the HOUSE_ID number and date and began a new group of
days for each new HOUSE_ID and whenever the sorted measurement days on the same
HOUSE_ID were 30 days or more apart. In most cases, a home was measured over four different
seasons for seven days, potentially giving 38 x 4 = 152 groups; the actual number of groups was
124. For the second approach, B, we again sorted the data by the HOUSE_ID number and date,
but this time we began a new group of days for each new HOUSE_ID and whenever the sorted
measurement days on the same HOUSE_ID were 30 days or more apart or were for different
temperature ranges. We used the same four temperature ranges chosen for the analysis in the
Cohen, Mallya, and Rosenbaum, 2005, memorandum (Appendix A): <= 10, 10-20, 20-25, and >
25 °C. For example, if the first week of measurements on a given HOUSE_ID had the first three
days in the <= 10 °C range, the next day in the 10-20 °C range, and the last three days in the <=
10 °C range, then the first approach would treat this as a single group of days. The second
approach would treat this as three groups of days, i.e., the first three days, the fourth day, and the
last three days. Using the first approach, the days in each group can be in different temperature
ranges. Using the second approach, every day in a group is in the same temperature range. Using
the first approach we treat groups of days as being independent following a transition to a
different house or season. Using the second approach we treat groups of days as being
independent following a transition to a different house or season or temperature range.
To evaluate the distributions of multi-day air exchange rate (AER) averages, we averaged the
AERs over consecutive days in each group. To obtain a set of one-day averages, we took the
AERs for the first day of each group. To obtain a set of two-day averages, we took the average
AER over the first two days from each group. We continued this process to obtain three-, four-,
five-, six-, and seven-day averages. There were insufficiently representative data for averaging
periods longer then seven days. Averages over non-consecutive days were excluded. Each
averaging period was assigned the temperature range using the average of the daily temperatures
for the averaging period. Using Approach A, some or all of the days in the averaging period
might be in different temperature ranges than the overall average. . Using Approach B, every day
is in the same temperature range as the overall average. For each averaging period and
temperature range, we calculated the mean, standard deviation, and variance of the period
average AER and of its natural logarithm. Note than the geometric mean equals e raised to the
power Mean log (AER) and the geometric standard deviation equals e raised to the power Std
Dev log (AER). The results are shown in Tables E-l (Approach A) and E-2 (Approach B).
B-258
-------
Table E-l. Distribution of AER averaged over K days and its logarithm. Groups defined by Approach A.
Temperature
(°C)
<= 10
<=10
<= 10
<=10
<=10
<= 10
<=10
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-25
20-25
20-25
20-25
20-25
20-25
20-25
>25
>25
>25
>25
>25
>25
>25
K
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
Groups
35
30
28
28
24
24
29
48
55
51
50
53
49
34
32
28
27
17
17
17
14
9
11
12
23
23
23
17
Mean
AER
1.109
1.149
1.065
1.081
1.103
1.098
1.054
0.652
0.654
0.641
0.683
0.686
0.677
0.638
0.500
0.484
0.495
0.536
0.543
0.529
0.571
0.470
0.412
0.411
0.385
0.390
0.399
0.438
Mean
log(AER)
-0.066
-0.009
-0.088
-0.090
-0.082
-0.083
-0.109
-0.659
-0.598
-0.622
-0.564
-0.546
-0.533
-0.593
-1.005
-0.972
-0.933
-0.905
-0.905
-0.899
-0.889
-1.058
-1.123
-1.036
-1.044
-1.028
-1.010
-0.950
Std Dev
AER
0.741
0.689
0.663
0.690
0.754
0.753
0.704
0.417
0.411
0.416
0.440
0.419
0.379
0.343
0.528
0.509
0.491
0.623
0.672
0.608
0.745
0.423
0.314
0.243
0.176
0.175
0.193
0.248
Std Dev
log(AER)
0.568
0.542
0.546
0.584
0.598
0.589
0.556
0.791
0.607
0.603
0.619
0.596
0.544
0.555
0.760
0.623
0.604
0.652
0.649
0.617
0.683
0.857
0.742
0.582
0.429
0.425
0.435
0.506
Variance
AER
0.549
0.474
0.440
0.476
0.568
0.567
0.496
0.174
0.169
0.173
0.194
0.175
0.144
0.118
0.279
0.259
0.241
0.389
0.452
0.370
0.555
0.179
0.098
0.059
0.031
0.031
0.037
0.061
Variance
log(AER)
0.322
0.294
0.298
0.341
0.358
0.347
0.309
0.625
0.368
0.363
0.384
0.355
0.296
0.308
0.577
0.388
0.365
0.425
0.421
0.381
0.466
0.734
0.551
0.339
0.184
0.181
0.189
0.256
Using both approaches, Tables E-l and E-2 show that the mean values for the AER and its
logarithm are approximately constant for the same temperature range but different averaging
periods. This is expected if the daily AER values all have the same statistical distribution,
regardless of whether or not they are independent. More interesting is the observation that the
standard deviations and variances are also approximately constant for the same temperature
range but different averaging periods, except for the data at > 25 °C where the standard
deviations and variances tend to decrease as the length of the averaging period increases. If the
daily AER values were statistically independent, then the variance of an average over K days is
given by Var / K, where Var is the variance of a single daily AER value. Clearly this formula
does not apply. Since the variance is approximately constant for different values of K in the same
temperature range (except for the relatively limited data at > 25 °C), this shows that the daily
AER values are strongly correlated. Of course the correlation is not perfect, since otherwise the
AER for a given day would be identical to the AER for the next day, if the temperature range
were the same, which did not occur.
B-259
-------
Table E-2. Distribution of AER averaged over K days and its logarithm. Groups defined by Approach B.
Temperature
(°C)
<=10
<= 10
<= 10
<=10
<= 10
10-20
10-20
10-20
10-20
10-20
10-20
10-20
20-25
20-25
20-25
20-25
>25
>25
>25
>25
>25
>25
>25
K
1
2
3
4
5
1
2
3
4
5
6
7
1
2
3
4
1
2
3
4
5
6
7
Groups
62
41
32
17
5
109
81
63
27
22
12
6
107
63
23
3
54
32
23
12
12
6
6
Mean
AER
1.125
1.059
1.104
1.292
1.534
0.778
0.702
0.684
0.650
0.629
0.614
0.720
0.514
0.511
0.577
1.308
0.488
0.486
0.427
0.401
0.410
0.341
0.346
Mean
log(AER)
-0.081
-0.063
-0.040
0.115
0.264
-0.482
-0.532
-0.540
-0.626
-0.660
-0.679
-0.587
-0.915
-0.930
-0.837
-0.484
-0.949
-0.900
-0.970
-1.029
-1.003
-1.164
-1.144
Std Dev
AER
0.832
0.595
0.643
0.768
1.087
0.579
0.451
0.409
0.414
0.417
0.418
0.529
0.518
0.584
0.641
1.810
0.448
0.351
0.218
0.207
0.207
0.129
0.125
Std Dev
log(AER)
0.610
0.481
0.530
0.531
0.608
0.721
0.603
0.580
0.663
0.654
0.638
0.816
0.639
0.603
0.659
1.479
0.626
0.595
0.506
0.509
0.507
0.510
0.494
Variance
AER
0.692
0.355
0.413
0.590
1.182
0.336
0.204
0.167
0.171
0.174
0.175
0.280
0.269
0.341
0.411
3.277
0.201
0.123
0.048
0.043
0.043
0.017
0.016
Variance
log(AER)
0.372
0.231
0.281
0.282
0.370
0.520
0.363
0.336
0.440
0.428
0.407
0.667
0.409
0.364
0.434
2.187
0.392
0.354
0.256
0.259
0.257
0.261
0.244
These arguments suggest that, based on the RTF Panel study data, to a reasonable
approximation, the distribution of an AER measurement does not depend upon the length of the
averaging period for the measurement, although it does depend upon the average temperature.
This supports the methodology used in the Cohen, Mallya, and Rosenbaum, 2005, analyses that
did not take into account the length of the averaging period.
The above argument suggests that the assumption that daily AER values are statistically
independent is not justified. Statistical modeling of the correlation structure between consecutive
daily AER values is not easy because of the problem of accounting for temperature effects, since
temperatures vary from day to day. In the next section we present some statistical models of the
daily AER values from the RTF Panel Study.
Statistical models of AER auto-correlations from the RTF Panel Study
We used the MIXED procedure from SAS to fit several mixed models with fixed effects and
random effects to the daily values of AER and log(AER). The fixed effects are the population
average values of AER or log(AER), and are assumed to depend upon the temperature range.
The random effects have expected values of zero and define the correlations between pairs of
B-260
-------
measurements from the same Group, where the Groups are defined either using Approach A or
Approach B above. As described above, a Group is a period of up to 14 consecutive days of
measurements at the same house. For these mixed model analyses we included periods with one
or more missing days. For all the statistical models, we assume that AER values in different
Groups are statistically independent, which implies that data from different houses or in different
seasons are independent.
The main statistical model for AER was defined as follows:
AER = Mean(Temp Range) + A(Group, Temp Range)
+ B(Group, Day Number) + Error(Group, Day Number)
Mean(Temp Range) is the fixed effects term. There is a different overall mean value for each of
the four temperature ranges.
A(Group, Temp Range) is the random effect of temperature. For each Group, four error terms are
independently drawn from four different normal distributions, one for each temperature range.
These normal distributions all have mean zero, but may have different variances. Because of this
term, there is a correlation between AER values measured in the same Group of days for a pair
of days in the same temperature range.
B(Group, Day Number) is the repeated effects term. The day number is defined so that the first
day of a Group has day number 1, the next calendar day has day number 2, and so on. In some
cases AER's were missing for some of the day numbers. B(Group, Day Number) is a normally
distributed error term for each AER measurement. The expected value (i.e., the mean) is zero.
The variance is V. The covariance between B(Group g, day i) and B(Group h, day j) is zero for
days in different Groups g and h, and equals V x exp(d x i-j|) for days in the same Group. V and
d are fitted parameters. This is a first order auto-regressive model. Because of this term, there is a
correlation between AER values measured in the same Group of days, and the correlation
decreases if the days are further apart.
Finally, Error(Group, Day Number) is the Residual Error term. There is one such error term for
every AER measurement, and all these terms are independently drawn from the same normal
distribution, with mean 0 and variance W.
We can summarize this rather complicated model as follows. The AER measurements are
uncorrelated if they are from different Groups. If they are in the same Group, they have a
correlation that decreases with the day difference, and they have an additional correlation if they
are in the same temperature range.
Probably the most interesting parameter for these models is the parameter d, which defines the
strength of the auto-correlation between pairs of days. This parameter d lies between -1 (perfect
negative correlation) and +1 (perfect positive correlation) although values exactly equal to +1 or
-1 are impossible for a stationary model. Negative values of d would be unusual since they
would imply a tendency for a high AER day to be followed by a low AER day, and vice versa.
The case d=0 is for no auto-correlation.
B-261
-------
Table E-3 gives the fitted values of d for various versions of the model. The variants considered
were:
• model AER or log(AER)
• include or exclude the term A(Group, Temp Range) (the "random" statement in the SAS
code)
• use Approach A or Approach B to define the Groups
Since Approach B forces the temperature ranges to be the same for very day in a Group, the
random temperature effect term is difficult to distinguish from the other terms. Therefore this
term was not fitted using Approach B.
Table E-3. Autoregressive parameter d for various statistical models for the RTF Panel
Study AERs.
Dependent variable
AER
AER
AER
Log(AER)
Log(AER)
Log(AER)
Include A(Group,
Temp Range)?
Yes
No
No
Yes
No
No
Approach
A
A
B
A
A
B
d
0.80
0.82
0.80
0.87
0.87
0.85
In all cases, the parameter d is 0.8 or above, showing the very strong correlations between AER
measurements on consecutive or almost consecutive days.
Impact of accounting for daily average AER auto-correlation
In the current version of the APEX model, there are several options for stratification of time
periods with respect to AER distributions, and for when to re-sample from a distribution for a
given stratum. The options selected for this current set of simulations resulted in a uniform AER
for each 24-hour period and re-sampling of the 24-hour AER for each simulated day. This re-
sampling for each simulated day implies that the simulated AERs on consecutive days in the
same microenvironment are statistically independent. Although we have not identified sufficient
data to test the assumption of uniform AERs throughout a 24-hour period, the analyses described
in this memorandum suggest that AERs on consecutive days are highly correlated.
Therefore, in order to determine if bias was introduced into the APEX estimates with respect to
either the magnitudes or variability of exposure concentrations by implicitly assuming
uncorrelated air exchange rates, we re-ran the 2002 base case simulations using the option to not
re-sample the AERs. For this option APEX selects a single AER for each
microenvironment/stratum combination and uses it throughout the simulation.
The comparison of the two scenarios indicates little difference in APEX predictions, probably
because the AERs pertain only to indoor microenvironments and for the base cases most
exposure to elevated concentrations occurs in the "other outdoors" microenvironment. Figures E-
B-262
-------
1 and E-2 below present the comparison for exceedances of 8-hour average concentration during
moderate exertion for active person in Boston and Houston, respectively.
200
Figure E-l
Air Exchange Rate Resampling Sensitivity:
Days/Person with Exceedances of
8-Hour Average Exposure Concentration During Moderate Exertion
-Active Persons, Boston, 2002-
20 40 60
Cumulative Percentile
80
100
-base-. 01
-rsoff- .01
base -.02
rsoff- .02
-base -.03
-rsoff- .03
-base - .04
-rsoff- .04
-base - .05
rsoff- .05
Figure E-2
Air Exchange Rate Resampling Sensitivity:
Days/Person with Exceedances of
8-Hour Average Exposure Concentration During Moderate Exertion
-Active Persons, Houston, 2002-
re
20 40 60
Cumulative Percentile
80
100
-base-. 01
-rsoff-.01
base -.02
rsoff-.02
-base -.03
-rsoff- .03
-base -.04
-rsoff-.04
-base -.05
rsoff-.05
B-263
-------
Appendix C
Sulfur Dioxide Health Risk Assessment
Prepared for
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC
Prepared by
Ellen Post
Jin Huang
Andreas Maier
Hardee Mahoney
Abt Associates Inc.
Work funded through
Contract No. EP-W-05-022
Work Assignments 2-63 and 3-63, and
Contract No. EP-D-08-100
Work Assignment 0-04
Harvey Richmond, Work Assignment Manager
Catherine Turner, Project Officer, Contract No. EP-W-05-022
Nancy Riley, Project Officer, Contract No. EP-D-08-100
June 2009
-------
DISCLAIMER
This report is being furnished to the U.S. Environmental Protection Agency
(EPA) by Abt Associates Inc. in partial fulfillment of Contract No. EP-W-05-022, Work
Assignments 2-63 and 3-63, and Contract No. EP-D-08-100, Work Assignment 0-4. Any
opinions, findings, conclusions, or recommendations are those of the authors and do not
necessarily reflect the views of the EPA or Abt Associates. Any questions concerning
this document should be addressed to Harvey Richmond, U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, C504-06, Research Triangle Park,
North Carolina 27711 (email: richmond.harvey@epa.gov).
Abt Associates Inc. i June 2009
-------
Table of Contents
1 INTRODUCTION 1-1
2 PRELIMINARY CONSIDERATIONS 2-1
2.1 The Broad Empirical Basis for a Relationship Between SCh and Adverse
Health Effects 2-1
2.2 Basic Structure of the Risk Assessment 2-2
2.3 Air Quality Considerations 2-2
3 METHODS 3-1
3.1 Selection of health endpoints and target population 3-1
3.2 Development of exposure-response functions 3-3
3.2.1 Calculation of risk estimates 3-9
3.2.2 Selection of urban areas 3-11
3.2.3 Addressing variability and uncertainty 3-12
4 RESULTS 4-15
5 REFERENCES 5-1
Abt Associates Inc. ii June 2009
-------
List of Tables
Table 3-1. Study-Specific SOi Exposure-Response Data for Lung Function
Decrements 3-4
Table 3-2. Example: Calculation of Number of Occurrences of Lung Function
Response, Defined as an Increase in sRaw > 100%, Among Asthmatics in St.
Louis Engaged in Moderate or Greater Exertion Associated with Exposure to
SOi Concentrations that Just Meet an Alternative 1-Hour 99th Percentile 100
ppb Standard 3-10
Table 3-3. Example: Calculation of the Number of Asthmatics in St. Louis Engaged
in Moderate or Greater Exertion Estimated to Experience at Least One Lung
Function Response, Defined as an Increase in sRaw > 100%, Associated with
Exposure to SOi Concentrations that Just Meet an Alternative 1-Hour 99th
Percentile 100 ppb Standard 3-11
Table 4-1. Number of Occurrences (in Hundreds) of Lung Function Response
(Defined in Terms of sRaw) Among Asthmatics Engaged in Moderate or
Greater Exertion Associated with Exposure to "As Is" SOi Concentrations,
SOi Concentrations that Just Meet the Current Standards, and SOi
Concentrations that Just Meet Alternative Standards 4-16
Table 4-2. Number of Occurrences (in Hundreds) of Lung Function Response
(Defined in Terms of sRaw) Among Asthmatic Children Engaged in
Moderate or Greater Exertion Associated with Exposure to "As Is" SOi
Concentrations, SOi Concentrations that Just Meet the Current Standards,
and SOi Concentrations that Just Meet Alternative Standards 4-17
Table 4-3. Number of Occurrences (in Hundreds) of Lung Function Response
(Defined in Terms of FEVi) Among Asthmatics Engaged in Moderate or
Greater Exertion Associated with Exposure to "As Is" SOi Concentrations,
SOi Concentrations that Just Meet the Current Standards, and SOi
Concentrations that Just Meet Alternative Standards 4-18
Table 4-4. Number of Occurrences (in Hundreds) of Lung Function Response
(Defined in Terms of FEVi) Among Asthmatic Children Engaged in
Moderate or Greater Exertion Associated with Exposure to "As Is" SOi
Concentrations, SOi Concentrations that Just Meet the Current Standards,
and SOi Concentrations that Just Meet Alternative Standards 4-19
Table 4-5. Number of Asthmatics Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of sRaw) Associated with Exposure to "As Is" SOi Concentrations,
SOi Concentrations that Just Meet the Current Standards, and SOi
Concentrations that Just Meet Alternative Standards 4-20
Table 4-6. Number of Asthmatic Children Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of sRaw) Associated with Exposure to "As Is" SOi Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-21
Abt Associates Inc. iii June 2009
-------
Table 4-7. Number of Asthmatics Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-22
Table 4-8. Number of Asthmatic Children Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-23
Table 4-9. Percent of Asthmatics Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-24
Table 4-10. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-25
Table 4-11. Percent of Asthmatics Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-26
Table 4-12. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion
Estimated to Experience At Least One Lung Function Response (Defined in
Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations,
SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards 4-27
Abt Associates Inc. iv June 2009
-------
List of Figures
Figure 3-1. Components of SC>2 Health Risk Assessment Based on Controlled
Human Exposure Studies 3-2
Figure 3-2. Bayesian-Estimated Median Exposure-Response Functions: Increase in
sRaw > 100% for 5-Minute Exposures of Asthmatics Engaged in Moderate or
Greater Exertion 3-7
Figure 3-3. Bayesian-Estimated Median Exposure-Response Functions: Increase in
sRaw > 200% for 5-Minute Exposures of Asthmatics Engaged in Moderate or
Greater Exertion 3-7
Figure 3-4. Bayesian-Estimated Median Exposure-Response Functions: Decrease in
FEVi > 15% for 5-Minute Exposures of Asthmatics Engaged in Moderate or
Greater Exertion 3-8
Figure 3-5. Bayesian-Estimated Median Exposure-Response Functions: Decrease in
FEVi > 20% for 5-Minute Exposures of Asthmatics Engaged in Moderate or
Greater Exertion 3-8
Figure 4-1. Percent of Asthmatics Engaged in Moderate or Greater Exertion in St.
Louis Exhibiting Lung Function Response (Defined as an Increase in sRaw >
100%) Attributable to SO2 Within Given Ranges Under Different Air Quality
Scenarios 4-28
Figure 4-2. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion
in St. Louis Exhibiting Lung Function Response (Defined as an Increase in
sRaw > 100%) Attributable to SO2 Within Given Ranges Under Different Air
Quality Scenarios 4-29
Figure 4-3. Percent of Asthmatics Engaged in Moderate or Greater Exertion in St.
Louis Exhibiting Lung Function Response (Defined as a Decrease in FEVi >
15%) Attributable to SO2 Within Given Ranges Under Different Air Quality
Scenarios 4-30
Figure 4-4. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion
in St. Louis Exhibiting Lung Function Response (Defined as a Decrease in
FEVi > 15%) Attributable to SO2 Within Given Ranges Under Different Air
Quality Scenarios 4-31
Figure 4-5. Number of Occurrences of Lung Function Response (Defined as an
Increase in sRaw > 100%) Among Asthmatics Engaged in Moderate or
Greater Exertion in St. Louis Attributable to SO2 Within Given Ranges
Under Different Air Quality Scenarios 4-32
Figure 4-6. Number of Occurrences of Lung Function Response (Defined as an
Increase in sRaw > 100%) Among Asthmatic Children Engaged in Moderate
or Greater Exertion in St. Louis Attributable to SO2 Within Given Ranges
Under Different Air Quality Scenarios 4-33
Figure 4-7. Number of Occurrences of Lung Function Response (Defined as a
Decrease in FEVi > 15%) Among Asthmatics Engaged in Moderate or
Greater Exertion in St. Louis Attributable to SO2 Within Given Ranges
Under Different Air Quality Scenarios 4-34
Figure 4-8. Number of Occurrences of Lung Function Response (Defined as a
Decrease in FEVi > 15%) Among Asthmatic Children Engaged in Moderate
Abt Associates Inc. v June 2009
-------
or Greater Exertion in St. Louis Attributable to SOi Within Given Ranges
Under Different Air Quality Scenarios 4-35
Figure 4-9. Percent of Asthmatics Engaged in Moderate or Greater Exertion in
Greene Co. Exhibiting Lung Function Response (Defined as an Increase in
sRaw > 100%) Attributable to SO2 Within Given Ranges Under Different Air
Quality Scenarios 4-36
Figure 4-10. Percent of Asthmatic Children Engaged in Moderate or Greater
Exertion in Greene Co. Exhibiting Lung Function Response (Defined as an
Increase in sRaw > 100%) Attributable to SOi Within Given Ranges Under
Different Air Quality Scenarios 4-37
Figure 4-11. Percent of Asthmatics Engaged in Moderate or Greater Exertion in
Greene Co. Exhibiting Lung Function Response (Defined as a Decrease in
FEVi > 15%) Attributable to SO2 Within Given Ranges Under Different Air
Quality Scenarios 4-38
Figure 4-12. Percent of Asthmatic Children Engaged in Moderate or Greater
Exertion in Greene Co. Exhibiting Lung Function Response (Defined as a
Decrease in FEVi > 15%) Attributable to SOi Within Given Ranges Under
Different Air Quality Scenarios 4-39
Figure 4-13. Number of Occurrences of Lung Function Response (Defined as an
Increase in sRaw > 100%) Among Asthmatics Engaged in Moderate or
Greater Exertion in Greene Co. Attributable to SOi Within Given Ranges
Under Different Air Quality Scenarios 4-40
Figure 4-14. Number of Occurrences of Lung Function Response (Defined as an
Increase in sRaw > 100%) Among Asthmatic Children Engaged in Moderate
or Greater Exertion in Greene Co. Attributable to SOi Within Given Ranges
Under Different Air Quality Scenarios 4-41
Figure 4-15. Number of Occurrences of Lung Function Response (Defined as a
Decrease in FEVi > 15%) Among Asthmatics Engaged in Moderate or
Greater Exertion in Greene Co. Attributable to SOi Within Given Ranges
Under Different Air Quality Scenarios 4-42
Figure 4-16. Number of Occurrences of Lung Function Response (Defined as a
Decrease in FEVi > 15%) Among Asthmatic Children Engaged in Moderate
or Greater Exertion in Greene Co. Attributable to SOi Within Given Ranges
Under Different Air Quality Scenarios 4-43
Figure 4-17. Legend for Figures 4-1 - 4-16 4-44
Abt Associates Inc. vi June 2009
-------
Sulfur Dioxide Health Risk Assessment
1 INTRODUCTION
The U.S. Environmental Protection Agency (EPA) is presently conducting a
review of the national ambient air quality standards (NAAQS) for sulfur dioxide (802).
Sections 108 and 109 of the Clean Air Act (Act) govern the establishment and periodic
review of the NAAQS. These standards are established for pollutants that may
reasonably be anticipated to endanger public health and welfare, and whose presence in
the ambient air results from numerous or diverse mobile or stationary sources. The
NAAQS are to be based on air quality criteria, which are to accurately reflect the latest
scientific knowledge useful in indicating the kind and extent of identifiable effects on
public health or welfare that may be expected from the presence of the pollutant in
ambient air. The EPA Administrator is to promulgate and periodically review, at five-
year intervals, "primary" (health-based) and "secondary" (welfare-based) NAAQS for
such pollutants.1 Based on periodic reviews of the air quality criteria and standards, the
Administrator is to make revisions in the criteria and standards, and promulgate any new
standards, as may be appropriate. The Act also requires that an independent scientific
review committee advise the Administrator as part of this NAAQS review process, a
function performed by the Clean Air Scientific Advisory Committee (CASAC).
EPA's plan and schedule for this SC>2 NAAQS review is presented in the
"Integrated Plan for Review of the Primary National Ambient Air Quality Standards for
Sulfur Oxides" (U.S. EPA, 2007a). The plan discusses the preparation of two key
components in the NAAQS review process: an Integrated Science Assessment (ISA) and
risk/exposure assessments. The ISA critically evaluates and integrates scientific
information on the health effects associated with exposure to oxides of sulfur (SOX) in the
ambient air. The risk/exposure assessments develop, as appropriate, quantitative
estimates of human exposure and health risk and related variability and uncertainties,
drawing upon the information summarized in the ISA.
In May 2008 EPA's National Center for Environmental Assessment (NCEA)
released a draft version of the "Integrated Science Assessment for Oxides of Sulfur -
Health Criteria, henceforth referred to as the draft ISA (U.S. EPA, 2008a). In June 2008,
EPA's Office of Air Quality Planning and Standards (OAQPS) released a first draft of its
"Risk and Exposure Assessment to Support the Review of the SO2 Primary National
Ambient Air Quality Standard," henceforth referred to as the 1st draft REA (U.S. EPA,
2008b). Both of these documents were reviewed by the CASAC SO2 Panel on July 30-
31, 2008. Based on its review of the draft ISA, OAQPS decided to expand the health risk
assessment to include a quantitative assessment of lung function responses indicative of
1 Section 109(b)(l) [42 U.S.C. 7409] of the Act defines a primary standard as one "the attainment and
maintenance of which in the judgment of the Administrator, based on such criteria and allowing an
adequate margin of safety, are requisite to protect the public health."
Abt Associates Inc. 1-1 June 2009
-------
bronchoconstriction experienced by asthmatic subjects associated with 5 to 10 minute
exposures to SC>2 while engaged in moderate or greater exertion. In September 2008,
NCEA released the final version of the ISA, "Integrated Science Assessment for Oxides
of Sulfur - Health Criteria, henceforth referred to as the ISA (U.S. EPA, 2008c). A
second draft REA (EPA, 2009a) was made available to the CASAC and public in March
2009. The second draft REA was reviewed by the CASAC SO2 Panel on April 16-17,
2009. This final report has been informed by comments from CASAC and the public on
the second draft REA, as well as findings and conclusions contained in the final ISA.
SC>2 is one of a group of compounds known as sulfur oxides (SOX), which include
multiple chemicals (e.g., SO2, SO, SOs). However only SO2 is present at concentrations
significant for human exposures and the ISA indicates there is limited adverse health
effect data for the other gaseous compounds. Therefore, as in past NAAQS reviews, SO2
is considered as a surrogate for gaseous SOX species in this assessment, with the
secondarily formed particulate species (i.e., sulfate or 804) addressed as part of the
particulate matter (PM) NAAQS review.
In the previous review, concluded in 1996, it was clearly established that subjects
with asthma are more sensitive to the respiratory effects of SO2 exposure than healthy
individuals (ISA, section 3.1.3.2). Asthmatics exposed to SO2 concentrations as low as
0.2-0.3 ppm for 5-10 minutes during exercise have been shown to experience significant
bronchoconstriction, measured as an increase in specific airway resistance (sRaw)
(>100%) or a decrease in forced expiratory volume in one second (FEVi) (>15%) after
correction for exercise-induced responses in clean air.
The basic structure of the SO2 health risk assessment described in this document
reflects the fact that we have available controlled human exposure study data from
several studies involving volunteer asthmatic subjects who were exposed to SO2
concentrations at specified exposure levels while engaged in moderate or greater exertion
for 5- or 10-minute exposures.2 The risk assessment estimates lung function risks for (1)
recent ambient levels of SO2, (2) air quality adjusted to simulate just meeting the current
primary 24-hour and annual standards,3 and (3) air quality adjusted to simulate just
meeting selected alternative 1-hour standards in selected locations encompassing a
variety of SO2 emission source types in the Greene County and the St. Louis area within
Missouri.
The SO2 health risk assessment builds upon the methodology, analyses, and
lessons learned from the assessments conducted for the last SO2 NAAQS review in 1996,
as well as the methodology and lessons learned from the health risk assessment work
conducted for the recently concluded Os NAAQS review (Abt Associates, 2007a) - in
2 An additional characterization of risk may involve use of concentration-response functions, if sufficient
and relevant epidemiological data are identified in the ISA to support development of functions that are
related to ambient SO2 concentrations.
3 There is a 3-hr secondary standard as well. However, this risk assessment is taking into account only the
primary standards. The current primary SO2 standards include a 24-hour standard set at 0.14 parts per
million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc. 1-2 June 2009
-------
particular, the assessment of risk based on controlled human exposure studies described
in Chapter 3 of that document. The SC>2 risk assessment is based on our current
understanding of the SC>2 scientific literature as reflected in the evaluation provided in the
final ISA.
The goals of this SC>2 health risk assessment are: (1) to develop health risk
estimates of the number and percent of the asthmatic population in the selected study area
locations that would experience moderate or greater lung function decrements in response
to daily 5-minute maximum peak exposures while engaged in moderate or greater
exertion for a recent year of air quality and under a scenario in which the SC>2
concentrations are adjusted to simulate just meeting the current 24-hour standard; (2) to
develop a better understanding of the influence of various inputs and assumptions on the
risk estimates; and (3) to gain insights into the risk levels and patterns of risk reductions
associated with air quality simulating just meeting alternative 1-hour SC>2 standards. The
risk assessment is intended as a tool that, together with other information on lung
function and other health effects evaluated in the SO2 ISA, can aid the Administrator in
judging whether the current primary standards protect public health with an adequate
margin of safety, or whether revisions to the standards are appropriate.
Preliminary considerations and the basic structure of the risk assessment are
described in section 2. Section 3 describes the methods used, and section 4 presents the
results of the risk assessment.
Abt Associates Inc. 1-3 June 2009
-------
2 PRELIMINARY CONSIDERATIONS
The health risk assessment described in this document estimated lung function
decrements (measured as increases in sRaw or decreases in FEVi) associated with SC>2
exposures under several scenarios: (1) recent ambient levels of SO2, (2) air quality
adjusted to simulate just meeting the current 24-hour and annual standards, and (3) air
quality adjusted to simulate just meeting several alternative 1-hour standards. In this
section we address preliminary considerations. Section 2.1 briefly discusses the broad
empirical basis for a relationship between SO2 exposures and adverse health effects.
Section 2.2 describes the basic structure of the risk assessment. Finally, section 2.3
addresses air quality considerations.
2.1 The Broad Empirical Basis for a Relationship Between SOi and Adverse
Health Effects
The ISA concludes that the health evidence "is sufficient to infer a causal
relationship between respiratory morbidity and short-term exposure to 802" (ISA, p. 3-
33). In support of this conclusion, the ISA notes the following:
The strongest evidence for this causal relationship comes from
human clinical studies reporting respiratory symptoms and decreased lung
function following peak exposures of 5-10 min duration to SO2. These
effects have been observed consistently across studies involving
exercising mild to moderate asthmatics. Statistically significant
decrements in lung function accompanied by respiratory symptoms
including wheeze and chest tightness have been clearly demonstrated
following exposure to 0.4-0.6 ppm 862. Although studies have not
reported statistically significant respiratory effects following exposure to
0.2-0.3 ppm SC>2, some asthmatic subjects (5-30%) have been shown to
experience moderate to large decrements in lung function at these
exposure concentrations.
A larger body of evidence supporting this determination of
causality comes from numerous epidemiological studies reporting
associations with respiratory symptoms, ED visits, and hospital
admissions with short-term 862 exposures, generally of 24-h avg.
Important new multicity studies and several other studies have found an
association between 24-h avg ambient SC>2 concentrations and respiratory
symptoms in children, particularly those with asthma....
... Collectively, the findings from both human clinical and
epidemiological studies provide a strong basis for concluding a causal
relationship between respiratory morbidity and short-term exposure to
SO2.
Abt Associates Inc. 2-1 June 2009
-------
2.2 Basic Structure of the Risk Assessment
As noted above, this SC>2 health risk assessment is based on controlled human
exposure studies involving volunteer subjects who were exposed while engaged in
different exercise regimens to specified levels of SC>2 under controlled conditions for 5 or
10 minute periods. The responses measured in these studies were measures of lung
function decrements, including increases in sRaw and decreases in FEVi. We used
probabilistic exposure-response relationships, based on analysis of individual data, that
describe the relationships between a measure of personal exposure to SO2 and the
measure(s) of lung function recorded in these studies. These probabilistic exposure-
response relationships were combined with daily 5-minute maximum peak exposure
estimates associated with the air quality scenarios mentioned above for mild and
moderate asthmatics engaged in moderate or greater exertion. Estimates of personal
exposures to varying ambient concentrations associated with several air quality scenarios
including recent air quality levels, and air quality levels simulating just meeting the
current SC>2 primary standard and several alternative primary 1-hour standards were
derived through exposure modeling. The details of the exposure modeling are described
in Chapter 8 and Appendix B of the final REA (EPA, 2009b).
The characteristics that are relevant to carrying out a risk assessment based on
controlled human exposure studies can be summarized as follows:
• A risk assessment based on controlled human exposure studies uses exposure-
response functions, and therefore requires as input (modeled) personal
exposures to 862.
• Controlled human exposure studies, carried out in laboratory settings, are
generally not specific to any particular real world location. A controlled
human exposure studies-based risk assessment can therefore appropriately be
carried out for any location for which there are adequate air quality data on
which to base the modeling of personal exposures.
The methods for the SC>2 risk assessment are discussed in section 3 below. The risk
assessment was implemented within a new probabilistic version of TRIM.Risk, the
component of EPA's Total Risk Integrated Methodology (TRIM) model that estimates
human health risks.4
2.3 Air Quality Considerations
The SO2 health risk assessment estimates lung function risks associated with (1)
"as is" ambient levels of 862, (2) air quality simulating just meeting the current 24-hour
and annual standards, and (3) air quality simulating just meeting several alternative 1-
4 TRIM.Risk was most recently applied to EPA's O3 health risk assessment. A User's Guide for the
Application of TRIM.Risk to the O3 health risk assessment (Abt Associates, 2007b) is available online at:
http://epa.gov/ttn/fera/data/trim/trimrisk ozone ra userguide 8-6-07.pdf.
Abt Associates Inc. 2-2 June 2009
-------
hour standards in a recent year (2002) in two selected locations encompassing a variety of
SC>2 emission source types in Greene County, Missouri and St. Louis, Missouri.
In order to estimate health risks associated with just meeting the current 24-hour
and annual standards and alternative 1-hour 862 standards, it is necessary to estimate the
distribution of short-term (5-minute) SC>2 concentrations that would occur under any
given standard. Since compliance with the current SC>2 standards is based on a single
year, air quality data from 2002 were used to determine the change in SO2 concentrations
required to meet the current standards. Estimated design values were used to determine
the adjustment necessary to just meet the current 24-hour and annual standards. The
approach to simulating just meeting the current standards and alternative 1-hour
standards is described in section 8.8.1 of the final REA (EPA, 2009b).
The risk estimates developed for the recently concluded PM and Os NAAQS
reviews represented risks associated with PM and Os levels in excess of estimated policy-
relevant background (PRB) levels in the U.S. PRB levels have been historically defined
by EPA as concentrations of a pollutant that would occur in the U.S. in the absence of
anthropogenic emissions in continental North America (defined as the United States,
Canada, and Mexico). The ISA notes that PRB SC>2 concentrations are below 10 parts
per trillion (ppt) over much of the United States and are generally less than 30 ppt. With
the exception of a few locations on the West Coast and locations in Hawaii, where
volcanic SC>2 emissions cause high PRB concentrations, PRB contributes less than 1% to
present-day SC>2 concentrations in surface air. Since PRB is well below concentrations
that might cause potential health effects, there was no adjustment made for risks
associated with PRB concentrations in the current 862 health risk assessment.
Abt Associates Inc. 2-3 June 2009
-------
3 METHODS
The major components of the SC>2 lung function risk assessment are illustrated in
Figure 3-1. The air quality and exposure analysis components that are integral to the risk
assessment are discussed in Chapters 6 and 7, respectively, of the 2nd draft REA. As
described in the ISA and the 2nd draft REA, there are numerous controlled human
exposure studies reporting lung function decrements (as measured by increases in SRaw
and/or decreases in FEVi) among mild and/or moderate asthmatic adults associated with
short-term (5 or 10 minute) peak exposures to various levels of SO2 while engaged in
moderate or greater exercise. The 862 lung function risk assessment focuses on these
lung function responses among asthmatic children and adults.
3.1 Selection of health endpoints and target population
The ISA concluded that there is sufficient evidence to infer a causal relationship
between respiratory morbidity and short-term exposure to SC>2 (ISA, section 5.2). This
determination was based in large part on controlled human exposure studies
demonstrating a relationship between short-term (5- or 10-minute) peak SO2 exposures
and adverse effects on the respiratory system in exercising asthmatics. More specifically,
the ISA found consistent evidence from numerous controlled human exposure studies
demonstrating increased respiratory symptoms (e.g. cough, chest tightness, wheeze) and
decrements in lung function in a substantial proportion of exercising asthmatics
(generally classified as mild to moderate asthmatics) following short-term peak exposures
to SC>2 at concentrations > 0.4 ppm (400 ppb). As in previous reviews, the ISA also
concluded that at concentrations below 1.0 ppm (1,000 ppb), healthy individuals are
relatively insensitive to the respiratory effects of short-term peak SO2 exposures (ISA,
sections 3.1.3.2). Therefore, the SO2 lung function risk assessment focuses on
asthmatics. Exposure estimates for asthmatic children and adult asthmatics were
combined separately with probabilistic exposure-response relationships (described
below) for lung function response associated with daily 5-minute maximum peak
exposures while engaged in moderate or greater exertion.5
5 Only the highest 5-minute peak exposure (with moderate or greater exertion) on each day will be
considered in the lung function risk assessment, since the controlled human exposure studies have shown
an acute-phase response that was followed by a short refractory period where the individual was relatively
insensitive to additional SO2 challenges.
Abt Associates Inc. 3-1 June 2009
-------
Figure 3-1. Components of SO2 Health Risk Assessment Based on Controlled Human Exposure Studies
Air Quality
Ambient
Monitoring for
Selected Urban
Areas
Air Quality
Adjustment
Procedures
Current and
Alternative
Proposed
Standards
Recent
("As Is")
Ambient
S02
Levels
Exposure
Exposure
Model
Exposure Estimates
Associated with:
• Recent Air Quality
• Current Standard
• Alternative
Standards
Exposure-Response
Controlled Human
Exposure Studies
Probabilistic
Exposure -
Response
Relationships
Health
Risk
Model
Risk Estimates:
• Recent Air
Quality
• Current
Standard
• Alternative
Standards
Abt Associates Inc.
3-2
June 2009
-------
Two measures of lung function response - specific airway resistance (sRaw) and forced
expiratory volume in one second (FEVi) - have been used in the controlled human exposure
studies that have focused on the effects of exposure to SO2 on exercising asthmatics. Negative
effects are measured as the percent increase in sRaw or the percent decrease in FEVi. As
explained below, we estimated exposure-response relationships for four different definitions of
response:
• An increase in sRaw > 100%
• An increase in sRaw > 200%
• A decrease in FEVi > 15%
• A decrease in FEVi > 20%.
3.2 Development of exposure-response functions
We used a Bayesian Markov Chain Monte Carlo approach to estimate probabilistic
exposure-response relationships for lung function decrements associated with 5- or 10-minute
exposures at moderate or greater exertion, using the WinBUGS software (Spiegelhalter et al.
(1996)). For an explanation of these methods, see Gelman et al. (1995) or Gilks et al. (1996).
We treated both 5- and 10-minute exposures as if they were all 5-minute exposures.
The combined data set from Linn et al. (1987, 1988, 1990), Bethel et al. (1983, 1985),
Roger et al. (1985), and Kehrl et al. (1987) provide data with which to estimate exposure-
response relationships between responses defined in terms of sRaw and 5- or 10-minute
exposures to SC>2 at levels of 0.2, 0.25, 0.3, 0.4, 0.5, 0.6, and 1.0 ppm.6 As noted above, two
definitions of response were used: (1) an increase in sRaw > 100% and (2) an increase in sRaw
> 200%.
The combined data set from Linn et al. (1987, 1988, 1990) provide data with which to
estimate exposure-response relationships between responses defined in terms of FEVi and 5- or
10-minute exposures to 862 at levels of 0.2, 0.3, 0.4, and 0.6 ppm. As noted above, two
definitions of response were used: a decrease in FEVi > 15% and a decrease in FEVi > 20%.
Before being used to estimate exposure-response relationships for 5-minute exposures,
the data from these controlled human exposure studies were corrected for the effect of exercising
in clean air to remove any systematic bias that might be present in the data attributable to an
exercise effect.7 Generally, this correction for exercise in clean air is small relative to the total
effects measures in the SO2-exposed cases. The resulting study-specific results, based on the
corrected data, are shown in Table 3-1.
6 Data from Magnussen et al. (1990) were not used in the estimation of sRaw exposure-response functions because
exposures in this study were conducted using a mouthpiece rather than a chamber.
7 Corrections were subject-specific. A correction was made by subtracting the subject's percent change (in FEVi or
sRaw) under the no-SO2 protocol from his or her percent change (in FEVi or sRaw) under the given SO2 protocol,
and rounding the result to the nearest integer. For example, if a subject's percent change in sRaw under the no-SO2
protocol was 110.12% and his percent change in sRaw under the 0.6 ppm (600 ppb) SO2 protocol was 185.92%,
then his percent change in sRaw due to SO2 is 185.92% -110.12% = 75.8%, which rounds to 76%.
Abt Associates Inc. 3-3 June 2009
-------
Table 3-1. Study-Specific SO2 Exposure-Response Data for Lung Function Decrements
Study and SO2 Level
Increase in sRaw > 100%
Number
Exposed
Number
Responding
Increase in sRaw > 200%
Number
Exposed
Number
Responding
Decrease in FEV!>15%
Number
Exposed
Number
Responding
Decrease in FEVi>20%
Number
Exposed
Number
Responding
0.2ppmSO2
Linn etal. (1987)
40
2
40
0
40
5
40
2
0.25ppmSO2
Bethel etal. (1985)
Roger etal. (1985)
19
9
28
6
2
1
19
9
28
3
0
0
0.3ppmSO2
Linn etal. (1988)
Linn etal. (1990)
20
21
2
7
20
21
1
2
20
21
3
5
20
21
0
3
0.4ppmSO2
Linn etal. (1987)
40
9
40
3
40
12
40
9
0.5ppmSO2
Bethel etal. (1983)
Roger etal. (1985)
Magnussen et al. (1990)*
10
28
45
6
5
16
10
28
45
4
1
7
0.6ppmSO2
Linn etal. (1987)
Linn etal. (1988)
Linn etal. (1990)
40
20
21
14
12
13
40
20
21
11
7
6
40
20
21
21
11
9
40
20
21
19
11
7
1.0ppmSO2
Roger etal. (1985)
Kehrl etal. (1987)
28
10
14
6
28
10
7
2
*Data from Magnussen et al. (1990) were not used in the estimation of sRaw exposure-response functions because exposures in this study were conducted using a
mouthpiece rather than a chamber.
Abt Associates Inc.
3-4
June 2009
-------
We considered two different functional forms for the exposure-response functions: a
2-parameter logistic model and a probit model. In particular, we used the data in Table 3-1 to
estimate the logistic function,
and the probit function,
' -<2/2dt (3-2)
for each of the four lung function responses defined above, where x denotes the SC>2
concentration (in ppb) to which the individual is exposed, ln(x) is the natural logarithm of x, y
denotes the corresponding probability of response (increase in sRaw > 100% or > 200% or
decrease in FEVi > 15% or > 20%), and ft and y are the two parameters whose values are
estimated. 8
We assumed that the number of responses, st, out of TV, subjects exposed to a given
SC>2 concentration, xf, has a binomial distribution with response probability given by equation
(3-1) when we assume the logistic model and equation (3-2) when we assume the probit
model. The likelihood function is therefore
In some of the controlled human exposure studies, subjects were exposed to a given
SO2 concentration more than once. However, because there were insufficient data to estimate
subject-specific response probabilities, we assumed a single response probability (for a given
definition of response) for all individuals and treated the repeated exposures for a single
subject as independent exposures in the binomial distribution.
For each model, we derived a Bayesian posterior distribution using this binomial
likelihood function in combination with uniform prior distributions for each of the unknown
parameters.9 We used 4000 iterations as the "burn-in" period followed by a sufficient number
of iterations to ensure convergence of the resulting posterior density. Each iteration
corresponds to a set of values for the parameters of the logistic or probit exposure-response
function.
For ease of exposition, we use the same two Greek letters to indicate two unknown parameters in the logistic
and probit models; this does not imply, however, that the values of these two parameters are the same in the two
models.
9 We used the following uniform prior distributions for the 2-parameter logistic model: (3 ~ U(-10, 0); and y ~
U(-10,0); we used the following normal prior distributions for the probit model: (3 ~ N(0, 1000); and y ~
N(0,1000).
Abt Associates Inc. 3-5 June 2009
-------
For any SC>2 concentration, x, we could then derive the nih percentile response value,
for any n, by evaluating the exposure-response function at x using each of the 18,000 sets of
parameter values. The resulting median (50th percentile) logistic and probit exposure-
response functions are shown together, along with the data used to estimate these functions,
for increases in sRaw > 100% and > 200% and decreases in FEVi > 15% and > 20% in
Figures 3-2, 3-3, 3-4, and 3-5, respectively.
As can be seen in Figures 3-2 through 3-5, there were only limited data with which to
estimate the logistic and probit exposure-response functions, and in all cases it wasn't clear
that one function fit the data better than the other. In fact, for each of the four lung function
response definitions there was little difference between the estimated logistic and probit
models in the range of the data used to estimate the functions. However, most of the
exposures occur below the range of the data, where there are differences between the two
functions.10 We therefore estimated the risks associated with exposure to SC>2 under the
different air quality scenarios considered using both the logistic and the probit exposure-
response functions. The 2.5th percentile, median, and 97.5th percentile logistic and probit
exposure-response curves, along with the response data to which they were fit, are shown
separately for each of the four response definitions in Appendix A.
10 The differences are relatively small, as can be seen in Figures 3-2 through 3-2; however, even these relatively
small differences result in substantial differences in estimates of risk.
Abt Associates Inc. 3-6 June 2009
-------
Figure 3-2. Bayesian-Estimated Median Exposure-Response Functions: Increase in sRaw > 100% for 5-
Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
The data
probit
2-parameter logistic
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure 3-3. Bayesian-Estimated Median Exposure-Response Functions: Increase in sRaw > 200% for 5-
Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
90% -
80% -
70% -
The data
probit
2-parameter logistic
0.4 0.6
SO2 Concentration (ppm)
0.8
Abt Associates Inc.
5-7
June 2009
-------
Figure 3-4. Bayesian-Estimated Median Exposure-Response Functions: Decrease in FEVi > 15% for 5-
Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
90% -
80% -
70% -
A The data
probit
2-parameter logistic
0.4 0.6
SO2 Concentration (ppm)
0.8
Figure 3-5. Bayesian-Estimated Median Exposure-Response Functions: Decrease in FEVt > 20% for 5-
Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% -i
90% -
80% -
70% -
60% -
50% -
40% -
30% -
20% -
10% -
0%
0
A The data
probit
2-parameter logistic
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Abt Associates Inc.
June 2009
-------
3.2.1 Calculation of risk estimates
We generated two measures of risk for each of the lung function response definitions.
The first measure of risk is simply the number of occurrences of the lung function response in
the designated population (e.g., asthmatics) in a year associated with SO2 concentrations
under a given air quality scenario. To calculate this measure of risk we started with the
number of exposures among the population that are at or above each benchmark level (i.e., 0
ppb, 50 ppb, 100 ppb, etc.), estimated from the exposure modeling. From this we calculated
the number of exposures within each 50 ppb exposure "bin" (e.g., < 50 ppb, 50 - 100 ppb,
etc.). u We then calculated the number of occurrences of lung function response by
multiplying the number of exposures in an exposure bin by the response probability (given by
our logistic or probit exposure-response function for the specified definition of lung function
response) associated with the midpoint of that bin and summing the results across the bins.
Because response probabilities are calculated for each of several percentiles of a
probabilistic exposure-response distribution, estimated numbers of occurrences are similarly
percentile-specific. The kth percentile number of occurrences, Ok, associated with SO2
concentrations under a given air quality scenario is:
]X(Rk\ej) (3-3)
;=i
where:
6j = (the midpoint of) the jth category of personal exposure to SC>2;
Nj = the number of exposures to Cj ppb 862, given ambient 862 concentrations under
the specified air quality scenario;
Rk ej = the kth percentile response probability at SO2 concentration ef, and
n = the number of intervals (categories) of SC>2 personal exposure concentration.
An example calculation, using the logistic exposure-response function, is given in Table 3-2.
11 The final exposure bin was from 750 to 800 ppb SO2. In at least one of the alternative standard scenarios,
there were exposures greater than 800 ppb. For any exposures that exceeded 800 ppb, we assumed a final bin
from 800 to 850 ppb, and assigned them the midpoint value of that bin, 825 ppb. This will result in a slight
downward bias in the estimate of risk.
Abt Associates Inc. 3-9 June 2009
-------
Table 3-2. Example: Calculation of Number of Occurrences of Lung Function Response, Defined as an
Increase in sRaw > 100%, Among Asthmatics in St. Louis Engaged in Moderate or Greater Exertion
Associated with Exposure to SO2 Concentrations that Just Meet an Alternative 1-Hour 99th Percentile 100
ppb Standard*
SO2 Exposure Bin (ppb)
Lower
Bound
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
Upper
Bound
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Midpoint
(1)
25
75
125
175
225
275
325
375
425
475
525
575
625
675
725
775
Number of
Exposures
(2)
16519000
136621
15760
3826
1051
413
175
83
31
24
8
0
0
8
0
0
Probability of
Response at Midpoint
SO2 Level
(3)
0.00406
0.02334
0.05162
0.08563
0.12300
0.16220
0.20210
0.24190
0.28060
0.31830
0.35430
0.38850
0.42090
0.45150
0.46600
0.49380
Expected Number of
Occurrences of Lung
Function Response
=(2) x (3)
67067
3189
814
328
129
67
35
20
9
8
3
0
0
4
0
0
Expected Number
Total Number of Exposures: 16677000 of Occurrences: 71672
*Calculations were made using the logistic exposure-response function.
The second measure of risk generated for each lung function response definition is the
number of individuals in the designated population to experience at least one lung function
response in a year associated with 862 concentrations under a specified air quality scenario.
The calculation of this measure of risk is similar to the calculation of the first measure of risk
- however, here we started with estimates, from the exposure modeling, of the number of
individuals exposed at least once to x ppb 862 or higher, for x = 0, 50, 100, etc. From this we
calculated the number of individuals exposed at least once to 862 concentrations within each
SC>2 exposure bin defined above. We then multiplied the numbers of individuals in an
exposure bin by the response probability (given by our logistic or probit exposure-response
function for the specified definition of lung function response) corresponding to the midpoint
of the exposure bin, and summed the results across the bins.
Because response probabilities are calculated for each of several percentiles of a
probabilistic exposure-response distribution, estimated numbers of individuals with at least
one lung function response are similarly percentile-specific. The kth percentile number of
individuals, Yk, associated with SC>2 concentrations under a given air quality scenario is:
*,)
(3-4)
Where e^ Rk ej, and n are as defined above, and NIj is the number of individuals whose
highest exposure is to 6j ppb SC>2, given ambient SC>2 concentrations under the specified air
Abt Associates Inc.
3-10
June 2009
-------
quality scenario. An example calculation, using the logistic exposure-response function, is
given in Table 3-3.
Table 3-3. Example: Calculation of the Number of Asthmatics in St. Louis Engaged in Moderate or
Greater Exertion Estimated to Experience at Least One Lung Function Response, Defined as an
Increase in sRaw > 100%, Associated with Exposure to SO2 Concentrations that Just Meet an
Alternative 1-Hour 99th Percentile 100 ppb Standard*
SO2 Exposure Bin (ppb)
Lower
Bound
0
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
Upper
Bound
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
Midpoint
(D
25
75
125
175
225
275
325
375
425
475
525
575
625
675
725
775
Number of
Asthmatics with
At Least One
Exposure in Bin
(2)
53711
34236
9835
3059
929
368
145
84
31
22
8
0
0
8
0
0
Probability of
Response at Midpoint
SO2 Level
(3)
0.00406
0.02334
0.05162
0.08563
0.12300
0.16220
0.20210
0.24190
0.28060
0.31830
0.35430
0.38850
0.42090
0.45150
0.46600
0.49380
Estimated Number of
Asthmatics Experiencing
at Least One Lung
Function Response
=(2) x (3)
218
799
508
262
114
60
29
20
9
7
3
0
0
4
0
0
Total: 102436 Total: 2032
*Calculations were made using the logistic exposure-response function.
Note that this calculation assumes that individuals who do not respond at the highest
SC>2 concentration to which they are exposed will not respond to any lower SC>2
concentrations to which they are exposed.
Note also that, in contrast to the risk estimates calculated for the 63 health risk
assessment, the risk estimates calculated for the SO2 health risk assessment do not subtract out
risk given the personal exposures associated with estimated policy relevant background (PRB)
ambient 862 concentrations, because PRB 862 concentrations are so low (see section 2.3).
3.2.2 Selection of urban areas
Although it would be useful to characterize SCVrelated lung function risks associated
with "as is" 862 ambient concentrations and 862 concentrations that just meet the current and
alternative SC>2 standards nationwide, because the modeling of personal exposures is both
time and labor intensive, a regional and source-oriented approach was selected instead. The
selection of areas to include in the exposure analysis, and therefore the risk assessment, took
into consideration the availability of ambient monitoring, the desire to represent a range of
Abt Associates Inc.
3-11
June 2009
-------
geographic areas considering 862 emission sources, population demographics, general
climatology, and results of the ambient air quality characterization.
The first area of interest was initially identified based on the results of a preliminary
screening of the 5-minute ambient 862 monitoring data that were available. The state of
Missouri was one of only a few states having both 5-minute maximum and continuous 5-
minute SC>2 ambient monitoring, as well as having over 30 1-hour SC>2 monitors in operation
at some time during the period from 1997 to 2007. In addition, the air quality characterization,
described in Chapter 6 of the 1st draft REA (EPA, 2008b), estimated frequent exceedances
above the potential health effect benchmark levels at several of the 1-hour ambient monitors.
In a ranking of estimated SC>2 emissions reported in the National Emissions Inventory (NEI),
Missouri ranked 7th for the number of stacks with > 1000 tpy SOX emissions out of all U.S.
states. These stack emissions were associated with a variety of source types such as electrical
power generating units, chemical manufacturing, cement processing, and smelters. For all
these reasons, the current SC>2 lung function risk assessment focuses on Missouri and, within
Missouri, on those areas within 20 km of a major point source of SC>2 emissions in Greene
County and the St. Louis area.
3.2.3 Addressing variability and uncertainty
Any estimation of risks associated with "as is" SO2 concentrations or with SO2
concentrations that just meet the current or alternative 862 standards should address both the
variability and uncertainty that generally underlie such an analysis. Uncertainty refers to the
lack of knowledge regarding the actual values of model input variables (parameter
uncertainty) and of physical systems or relationships (model uncertainty - e.g., the shapes of
exposure-response and concentration-response functions). The goal of the analyst is to reduce
uncertainty to the maximum extent possible. Uncertainty can be reduced by improved
measurement and improved model formulation. In a health risk assessment, however,
significant uncertainty often remains.
The degree of uncertainty can be characterized, sometimes quantitatively. For
example, the statistical uncertainty surrounding the estimated SC>2 coefficients in the
exposure-response functions is reflected in confidence or credible intervals provided for the
risk estimates.
As described in section 3.2 above, we used a Bayesian Markov Chain Monte Carlo
approach to estimate exposure-response functions as well as to characterize uncertainty
attributable to sampling error based on sample size considerations. Using this approach, we
could derive the nl percentile response value, for any n, for any SC^concentration, x, as
described above (see section 3.2). Because our exposure estimates were generated at the
midpoints of 0.05 ppm intervals (i.e., for 0.025 ppm, 0.075 ppm, etc.), we derived 2.5th
percentile, 50th percentile (median), and 97.5th percentile response estimates for 862
concentrations at these midpoint values. The 2.5th percentile and 97.5th percentile response
estimates comprise the lower and upper bounds of the credible interval around each point
estimate (median estimate) of response.
Abt Associates Inc. 3-12 June 2009
-------
In addition to uncertainties arising from sampling variability, other uncertainties
associated with the use of the exposure-response relationships for lung function responses are
briefly summarized below. Additional uncertainties with respect to the exposure inputs to the
risk assessment are described in section 8.11 of the final REA (EPA, 2009b). The main
additional uncertainties with respect to the approach used to estimate exposure-response
relationships include:
• Length of exposure. The 5-minute lung function risk estimates are based on a combined
data set from several controlled human exposure studies, most of which evaluated
responses associated with 10-minute exposures. However, since some studies which
evaluated responses after 5-minute exposures found responses occurring as early as 5-
minutes after exposure, we used all of the 5- and 10- minute exposure data to represent
responses associated with 5-minute exposures. We do not believe that this approach would
appreciably impact the risk estimates.
• Exposure-response for mild/moderate asthmatics. The data set that was used to estimate
exposure-response relationships included mild and/or moderate asthmatics. There is
uncertainty with regard to how well the population of mild and moderate asthmatics
included in the series of controlled human exposure studies represent the distribution of
mild and moderate asthmatics in the U.S. population. As indicated in the ISA (p. 3-9), the
subjects studied represent the responses "among groups of relatively healthy asthmatics
and cannot necessarily be extrapolated to the most sensitive asthmatics in the population
who are likely more susceptible to the respiratory effects of exposure to 862."
• Extrapolation of exposure-response relationships. It was necessary to estimate responses
at SC>2 levels below the lowest exposure levels used in free-breathing controlled human
studies (i.e., 0.2 ppm or 200 ppb). We did not include alternative models that incorporate
hypothetical population thresholds, given the lack of evidence supporting the choice of
potential hypothetical threshold levels. As discussed later in this document, we have
presented information on the contribution of different exposure intervals to the total
estimated lung function risk. This information provides insights on how much of the
estimated risk is attributed to 862 exposures at the lower exposure levels (i.e., 0 to 50 ppb,
50 to 100 ppb, 100 to 150 ppb, etc.). One can use this information to get a rough sense of
the SC>2-related risk that would exist under alternative threshold assumptions.
• Reproducibility of SO2-induced responses. The risk assessment assumed that the SCV
induced responses for individuals are reproducible. We note that this assumption has
some support in that one study (Linn et al., 1987) exposed the same subjects on two
occasions to 0.6 ppm (600 ppb) and the authors reported a high degree of correlation (r >
0.7 for mild asthmatics and r > 0.8 for moderate asthmatics, p < 0.001), while observing
much lower and nonsignificant correlations (r = 0.0 - 0.4) for the lung function response
observed in the clean air with exercise exposures.
• Age and lung function response. Because the vast majority of controlled human exposure
studies investigating lung function responses were conducted with adult subjects, the risk
assessment relies on data from adult asthmatic subjects to estimate exposure-response
relationships that were applied to all asthmatic individuals, including children. The ISA
Abt Associates Inc. 3-13 June 2009
-------
(section 3.1.3.5) indicates that there is a strong body of evidence that suggests adolescents
may experience many of the same respiratory effects at similar SC>2 levels, but recognizes
that these studies administered SC>2 via inhalation through a mouthpiece rather than an
exposure chamber. This technique bypasses nasal absorption of SO2 and can result in an
increase in lung 862 uptake. Therefore, the uncertainty will be greater in the risk estimates
for asthmatic children.
• Exposure history. The risk assessment assumed that the SCVinduced response on any
given day is independent of previous SC>2 exposures.
• Interaction between SO? and other pollutants. Because the controlled human exposure
studies used in the risk assessment involved only SC>2 exposures, it was assumed that
estimates of SC>2-induced health responses would not be affected by the presence of other
pollutants (e.g., PM2.5, O3, NO2).
Variability refers to the heterogeneity in a population or parameter. Even if there is no
uncertainty surrounding inputs to the analysis, there may still be variability. For example,
there may be variability among exposure-response functions describing the relationship
between 862 and lung function in different locations. This variability does not imply
uncertainty about the exposure-response function in any location, but only that these functions
are different in the different locations, reflecting differences in the populations and/or other
factors that may affect the relationship between SO2 and the associated health endpoint. In
general, it is possible to have uncertainty but no variability (if, for instance, there is a single
parameter whose value is uncertain) or variability but little or no uncertainty (for example,
people's heights vary considerably but can be accurately measured with little uncertainty).
The SC>2 lung function risk assessment addresses variability-related concerns by using
location-specific inputs for the exposure analysis (e.g., location-specific population data, air
exchange rates, air quality and temperature data). The extent to which there may be
variability in exposure-response relationships for the populations included in the risk
assessment residing in different geographic areas is currently unknown.
Temporal variability is more difficult to address, because the risk assessment focuses
on some unspecified time in the future. To minimize the degree to which values of inputs to
the analysis may be different from the values of those inputs at that unspecified time, we are
using the most current inputs available.
Abt Associates Inc. 3-14 June 2009
-------
4 RESULTS
The results of the SO2 risk assessment are presented in Tables 4-1 through 4-12.
Each table includes results for both of the locations included in the risk assessment and
for all of the air quality scenarios considered, using both 2-parameter logistic and probit
exposure-response functions. Tables 4-1 and 4-2 show the numbers of occurrences of
lung function response in a year, defined in terms of sRaw, for asthmatics and for
asthmatic children, respectively, engaged in moderate or greater exertion associated with
SO2 concentrations under each of the different air quality scenarios considered in each of
the two locations. Tables 4-3 and 4-4 show the corresponding results when lung function
response is defined in terms of FEVi. Tables 4-5 and 4-6 show the numbers of
asthmatics and asthmatic children, respectively, engaged in moderate or greater exertion
estimated to experience at least one lung function response in a year, defined in terms of
sRaw, under each of the different air quality scenarios in each of the two locations.
Tables 4-7 and 4-8 show the corresponding results when lung function response is
defined in terms of FEVi. Finally, Tables 4-9 through 4-12 show results analogous to
those shown in Tables 4-5 through 4-8, only as percentages of all asthmatics (asthmatic
children) engaged in moderate or greater exertion.
In addition, responses attributable to exposure to SO2 within different
concentration ranges are shown in Figures 4-1 through 4-16. The exposure ranges are in
50 ppb increments - i.e., SO2 < 50 ppb, 50 ppb < SO2 < 100 ppb, 100 ppb < SO2 < 150
ppb, ... , SO2 > 500 ppb. Figures 4-la and b show the percent of asthmatics engaged in
moderate or greater exertion in St. Louis, MO, using the logistic and probit exposure-
response functions, respectively, estimated to experience at least one lung function
response in a year, defined as an increase in sRaw > 100%, attributable to exposure to
SO2 in each exposure "bin." Figures 4-2a and b show the corresponding percents for
asthmatic children engaged in moderate or greater exertion in St. Louis, MO, using the
logistic and probit exposure-response functions, respectively. Figures 4-3a and b, and 4-
4a and b, show the corresponding percents for asthmatics and asthmatic children,
respectively, in St. Louis, MO, when lung function response is defined as a decrease in
FEVi> 15%.
Figures 4-5a and b show the number of occurrences of lung function response,
defined as an increase in sRaw > 100%, among asthmatics engaged in moderate or
greater exertion in St. Louis, MO, using the logistic and probit exposure-response
functions, respectively, attributable to exposure to SO2 in each exposure "bin." Figures
4-6a and b show the corresponding numbers of occurrences among asthmatic children in
St. Louis, MO. Figures 4-7a and b and 4-8a and b show the corresponding numbers of
occurrences of lung function response for asthmatics and asthmatic children, respectively,
when lung function response is defined as a decrease in FEVi> 15%. Figures 4-9a and b
through 4-16a and b are the corresponding figures for Greene Co., MO. Figure 4-17
shows the legend that is used in Figures 4-1 through 4-16.
Abt Associates Inc. 4-15 June 2009
-------
Table 4-1. Number of Occurrences (in Hundreds) of Lung Function Response (Defined in Terms of sRaw) Among Asthmatics Engaged in Moderate or
Greater Exertion Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
125
(24 - 572)
16
(0 - 256)
127
(25 - 577)
18
(1 -261)
125
(24 - 572)
16
(0 - 256)
125
(24 - 572)
16
(0 - 256)
125
(24 - 573)
16
(1-257)
126
(24 - 573)
16
(1-257)
126
(24 - 575)
17
(1 -258)
126
(24 - 574)
17
(1 -258)
Sf. Louis, MO
2-Parameter Logistic
Probit
657
(128-2985)
90
(4 - 1 346)
1672
(663 - 4740)
933
(393-3107)
652
(125-2975)
86
(3-1336)
686
(141 -3041)
111
(11 -1402)
762
(176-3184)
170
(33-1543)
880
(234 - 3398)
264
(72-1756)
1036
(315-3673)
392
(128-2031)
997
(295 - 3604)
360
(114-1963)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
38
(4-310)
2
(0-123)
39
(4-312)
3
(0-124)
38
(4-310)
2
(0-122)
38
(4-310)
2
(0-123)
38
(4-310)
2
(0-123)
38
(4-310)
2
(0-123)
39
(4-311)
2
(0-123)
39
(4-311)
2
(0-123)
Sf. Louis, MO
2-Parameter Logistic
Probit
201
(21 -1614)
13
(0 - 643)
560
(165-2407)
258
(86-1388)
199
(20-1609)
12
(0 - 639)
211
(24-1639)
18
(1 - 666)
237
(32-1703)
33
(5 - 725)
278
(47-1799)
59
(12-814)
332
(68-1923)
95
(24 - 930)
319
(63-1892)
86
(21 -901)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03
ppm, calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-16
June 2009
-------
Table 4-2. Number of Occurrences (in Hundreds) of Lung Function Response (Defined in Terms of sRaw) Among Asthmatic Children Engaged in
Moderate or Greater Exertion Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards,
and SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
S02
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
71
(13-324)
9
(0-145)
72
(14-327)
10
(1 -148)
71
(13-324)
9
(0-145)
71
(14-324)
9
(0-145)
71
(14-324)
9
(0-145)
71
(14-325)
9
(0-146)
71
(14-325)
10
(0-146)
71
(14-325)
10
(0-146)
Sf. Louis, MO
2-Parameter Logistic
Probit
417
(81 -1893)
58
(3 - 855)
1179
(484 - 3209)
692
(296-2176)
413
(80-1885)
55
(2 - 847)
439
(91 -1935)
74
(8 - 896)
497
(118-2043)
118
(25-1004)
586
(162-2206)
189
(53-1166)
704
(222-2413)
286
(96-1373)
674
(207-2361)
262
(85-1321)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
22
(2-175)
1
(0-69)
22
(2-177)
2
(0-71)
22
(2-175)
1
(0 - 69)
22
(2-175)
1
(0 - 69)
22
(2-175)
1
(0-69)
22
(2-176)
1
(0 - 70)
22
(2-176)
1
(0 - 70)
22
(2-176)
1
(0 - 70)
Sf. Louis, MO
2-Parameter Logistic
Probit
128
(13-1023)
8
(0 - 408)
397
(122-1618)
192
(65 - 967)
126
(13-1019)
8
(0-405)
135
(15-1042)
12
(1-425)
155
(22-1091)
24
(4 - 470)
186
(33-1164)
43
(9 - 538)
227
(49-1257)
70
(18-625)
217
(45-1234)
63
(16-603)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-17
June 2009
-------
Table 4-3. Number of Occurrences (in Hundreds) of Lung Function Response (Defined in Terms of FEVi) Among Asthmatics Engaged in Moderate or
Greater Exertion Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FEV1 >= 15%
Greene County, MO
2-Parameter Logistic
Probit
69
(6 - 675)
6
(0-418)
71
(7 - 680)
8
(0 - 424)
69
(6 - 675)
6
(0-417)
69
(6 - 675)
6
(0-418)
69
(6 - 675)
6
(0-418)
69
(6 - 676)
6
(0-419)
70
(6 - 677)
7
(0-421)
70
(6 - 677)
6
(0 - 420)
Sf. Louis, MO
2-Parameter Logistic
Probit
366
(33 - 3520)
36
(1 -2189)
1341
(454 - 5632)
866
(322-4471)
361
(32 - 3507)
33
(0-2175)
391
(41 - 3587)
55
(5 - 2262)
461
(66 - 3759)
109
(20 - 2448)
570
(108-4016)
198
(49 - 2727)
718
(169-4346)
322
(94 - 3084)
681
(154-4264)
291
(82 - 2995)
Response = Decrease in FEV1 >= 20%
Greene County, MO
2-Parameter Logistic
Probit
3
(0-53)
0
(0-5)
3
(0 - 54)
0
(0-7)
3
(0 - 53)
0
(0-5)
3
(0 - 53)
0
(0-5)
3
(0-53)
0
(0-5)
3
(0 - 53)
0
(0-6)
3
(0 - 53)
0
(0-6)
3
(0 - 53)
0
(0-6)
Sf. Louis, MO
2-Parameter Logistic
Probit
15
(1 - 279)
1
(0 - 32)
310
(133-1045)
240
(120-697)
14
(0 - 276)
0
(0 - 30)
20
(2 - 299)
3
(1-47)
35
(7-351)
13
(5-89)
62
(17-435)
33
(14-158)
104
(34 - 550)
65
(29 - 256)
93
(30-521)
57
(25 - 232)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-18
June 2009
-------
Table 4-4. Number of Occurrences (in Hundreds) of Lung Function Response (Defined in Terms of FEVi) Among Asthmatic Children Engaged in
Moderate or Greater Exertion Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards,
and SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FEV1 >= 15%
Greene County, MO
2-Parameter Logistic
Probit
39
(3 - 382)
3
(0 - 236)
40
(4 - 386)
4
(0 - 240)
39
(3 - 382)
3
(0 - 236)
39
(3 - 382)
3
(0 - 236)
39
(3 - 382)
3
(0 - 237)
39
(3-383)
4
(0 - 237)
40
(4 - 384)
4
(0 - 238)
40
(4 - 383)
4
(0 - 238)
Sf. Louis, MO
2-Parameter Logistic
Probit
232
(21 -2231)
23
(1 -1389)
965
(338-3816)
648
(242-3101)
229
(20 - 2222)
21
(0-1379)
252
(27 - 2282)
38
(4-1444)
304
(46-2412)
79
(15-1585)
387
(77 - 2608)
146
(37-1797)
499
(123-2857)
239
(70 - 2066)
471
(112-2795)
216
(62-1999)
Response = Decrease in FEV1 >= 20%
Greene County, MO
2-Parameter Logistic
Probit
1
(0-30)
0
(0-3)
2
(0-31)
0
(0-4)
1
(0 - 30)
0
(0-3)
1
(0 - 30)
0
(0-3)
1
(0-30)
0
(0-3)
2
(0 - 30)
0
(0-3)
2
(0 - 30)
0
(0-3)
2
(0 - 30)
0
(0-3)
Sf. Louis, MO
2-Parameter Logistic
Probit
10
(0-178)
0
(0-21)
231
(99 - 753)
180
(90-521)
9
(0-175)
0
(0-19)
13
(1 -192)
2
(1 - 32)
24
(5 - 232)
10
(3-63)
45
(13-295)
25
(10-116)
76
(26 - 382)
49
(21 -190)
68
(22 - 360)
43
(18-171)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest whole number.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-19
June 2009
-------
Table 4-5. Number of Asthmatics Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response (Defined
in Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
90
(20 - 390)
10
(0-180)
210
(80 - 620)
110
(40-410)
80
(20 - 380)
10
(0-170)
90
(20 - 390)
10
(0-180)
100
(20 - 420)
20
(0-210)
120
(30 - 460)
40
(10-250)
160
(50 - 520)
70
(20-310)
140
(40 - 500)
60
(20 - 280)
Sf. Louis, MO
2-Parameter Logistic
Probit
1010
(340-3010)
500
(140-1990)
13460
(9740-18510)
13050
(9430-18100)
730
(220 - 2490)
290
(70-1470)
1990
(860 - 4690)
1340
(520 - 3690)
3650
(1900-7100)
2930
(1450-6200)
5520
(3230 - 9490)
4810
(2760-8710)
7500
(4770-11850)
6860
(4310-11190)
7050
(4410-11320)
6400
(3950-10640)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
30
(0-210)
0
(0-80)
70
(20-310)
30
(10-180)
30
(0-210)
0
(0 - 80)
30
(0-210)
0
(0 - 90)
30
(0 - 220)
10
(0-100)
40
(10-240)
10
(0-110)
50
(10-270)
20
(0-140)
50
(10-260)
10
(0-130)
Sf. Louis, MO
2-Parameter Logistic
Probit
330
(70-1520)
120
(20 - 880)
5520
(3400 - 8960)
5180
(3150-8570)
230
(40-1290)
60
(10-660)
670
(210-2270)
350
(90-1590)
1280
(510-3360)
870
(310-2680)
2010
(940 - 4470)
1560
(690 - 3820)
2830
(1470-5590)
2380
(1200-5000)
2640
(1340-5330)
2190
(1070-4730)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest ten.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-20
June 2009
-------
Table 4-6. Number of Asthmatic Children Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response
(Defined in Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and
SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
30
(10-130)
10
(0-60)
110
(40 - 270)
60
(20 - 200)
30
(10-130)
0
(0 - 60)
30
(10-140)
10
(0 - 60)
40
(10-150)
10
(0-80)
50
(20-180)
20
(10-100)
70
(30-210)
40
(10-140)
60
(20 - 200)
30
(10-130)
Sf. Louis, MO
2-Parameter Logistic
Probit
590
(220-1570)
340
(100-1150)
8020
(6080-10370)
7950
(6020-10320)
400
(130-1210)
190
(50 - 790)
1220
(560 - 2620)
890
(360 - 2220)
2240
(1240-4010)
1910
(1000-3690)
3370
(2090 - 5350)
3080
(1860-5110)
4560
(3060 - 6680)
4330
(2870-6510)
4290
(2840 - 6390)
4060
(2640-6210)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
10
(0-70)
0
(0-30)
40
(10-130)
20
(0-90)
10
(0 - 70)
0
(0 - 30)
10
(0 - 70)
0
(0 - 30)
10
(0-80)
0
(0-40)
20
(0 - 90)
10
(0 - 50)
20
(10-110)
10
(0 - 60)
20
(10-100)
10
(0 - 60)
Sf. Louis, MO
2-Parameter Logistic
Probit
190
(50 - 780)
80
(10-500)
3380
(2190-5070)
3290
(2110-5000)
130
(30-610)
40
(10-350)
410
(140-1240)
240
(60 - 950)
800
(340-1870)
580
(220-1590)
1250
(620 - 2500)
1030
(480 - 2250)
1750
(970-3140)
1560
(830 - 2940)
1640
(890 - 3000)
1440
(740 - 2790)
'Numbers are median (50th percentile) numbers of asthmatic children. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest ten.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-21
June 2009
-------
Table 4-7. Number of Asthmatics Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response (Defined
in Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FE\A, >= 15%
Greene County, MO
2-Parameter Logistic
Probit
50
(10-460)
10
(0 - 290)
170
(50 - 730)
100
(30 - 590)
50
(0 - 450)
0
(0 - 280)
50
(10-460)
10
(0 - 290)
60
(10-490)
20
(0 - 330)
80
(20 - 540)
30
(10-380)
110
(30-610)
50
(10-460)
100
(20 - 590)
50
(10-430)
Sf. Louis, MO
2-Parameter Logistic
Probit
750
(180-3580)
410
(80 - 2880)
15220
(10280-22530)
15040
(10140-22670)
510
(100-2950)
220
(30 - 2200)
1700
(580 - 5590)
1250
(370 - 5070)
3460
(1520-8500)
2970
(1230-8210)
5570
(2880-11400)
5130
(2580-11280)
7910
(4550-14280)
7550
(4280-14280)
7370
(4160-13640)
6990
(3880-13610)
Response = Decrease in FE\A, >= 20%
Greene County, MO
2-Parameter Logistic
Probit
0
(0 - 40)
0
(0-10)
30
(10-130)
20
(10-80)
0
(0 - 40)
0
(0-0)
0
(0 - 40)
0
(0-10)
0
(0-50)
0
(0-10)
10
(0 - 60)
0
(0 - 20)
20
(0 - 80)
10
(0 - 40)
10
(0 - 80)
10
(0 - 40)
Sf. Louis, MO
2-Parameter Logistic
Probit
100
(20 - 570)
40
(10-320)
9240
(6110-13840)
9260
(6200-13820)
50
(10-380)
20
(0-170)
350
(110-1290)
240
(80 - 960)
1020
(430 - 2680)
870
(390 - 2340)
2100
(1060-4450)
1950
(1020-4170)
3540
(1990-6540)
3430
(1980-6340)
3190
(1760-6050)
3070
(1740-5830)
'Numbers are median (50th percentile) numbers of asthmatics. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding the SO2
coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest ten.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-22
June 2009
-------
Table 4-8. Number of Asthmatic Children Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response
(Defined in Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and
SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FEV., >= 15%
Greene County, MO
2-Parameter Logistic
Probit
20
(0-160)
0
(0-100)
90
(30 - 320)
60
(20 - 280)
20
(0-150)
0
(0-100)
20
(0-160)
0
(0-100)
30
(0-180)
10
(0-120)
40
(10-210)
20
(0-160)
50
(10-250)
30
(10-200)
50
(10-240)
30
(10-180)
Sf. Louis, MO
2-Parameter Logistic
Probit
460
(120-1870)
280
(50-1630)
9310
(6620-12680)
9320
(6630-12800)
290
(60-1440)
150
(20-1160)
1080
(390-3130)
840
(260 - 2990)
2200
(1030-4810)
1970
(860 - 4800)
3510
(1930-6440)
3350
(1790-6510)
4950
(3030 - 8070)
4870
(2930-8190)
4630
(2780 - 7720)
4530
(2660 - 7830)
Response = Decrease in FEV! >= 20%
Greene County, MO
2-Parameter Logistic
Probit
0
(0-10)
0
(0-0)
20
(10-70)
10
(0 - 40)
0
(0-10)
0
(0-0)
0
(0-10)
0
(0-0)
0
(0-20)
0
(0-10)
0
(0 - 30)
0
(0-10)
10
(0 - 40)
0
(0 - 20)
10
(0 - 40)
0
(0 - 20)
Sf. Louis, MO
2-Parameter Logistic
Probit
70
(10-350)
30
(10-220)
6150
(4190-8700)
6210
(4280 - 8780)
30
(10-220)
10
(0-110)
240
(80 - 820)
170
(60 - 650)
700
(300-1710)
610
(280-1560)
1430
(740 - 2830)
1370
(730 - 2750)
2410
(1400-4160)
2380
(1410-4140)
2170
(1240-3850)
2140
(1240-3820)
'Numbers are median (50th percentile) numbers of asthmatic children. Numbers in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Numbers are rounded to the nearest hundred.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-23
June 2009
-------
Table 4-9. Percent of Asthmatics Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response (Defined
in Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
1%
(0.4% - 2.9%)
0.5%
(0.2% -1.9%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.5%
(0.1% -2%)
0.1%
(0% - 1 %)
0.6%
(0.2% -2.1%)
0.2%
(0%-1.2%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
0.7%
(0.2% - 2.3%)
0.3%
(0.1% -1.3%)
Sf. Louis, MO
2-Parameter Logistic
Probit
1%
(0.3% - 2.9%)
0.5%
(0.1% -1.9%)
13.1%
(9.5% -18.1%)
12.7%
(9.2% -17.7%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
1.9%
(0.8% - 4.6%)
1.3%
(0.5% - 3.6%)
3.6%
(1.9% -6.9%)
2.9%
(1.4% -6.1%)
5.4%
(3.2% - 9.3%)
4.7%
(2.7% - 8.5%)
7.3%
(4.7% - 1 1 .6%)
6.7%
(4.2% -10.9%)
6.9%
(4. 3% -11.1%)
6.2%
(3.9% -10.4%)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.2%
(0%-1%)
0%
(0% - 0.5%)
0.2%
(0%-1.1%)
0%
(0% - 0.5%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.2%
(0%-1.2%)
0.1%
(0% - 0.6%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.9%)
5.4%
(3.3% - 8.7%)
5.1%
(3.1% -8. 4%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.7%
(0.2% - 2.2%)
0.3%
(0.1% -1.6%)
1.3%
(0.5% - 3.3%)
0.8%
(0.3% - 2.6%)
2%
(0.9% - 4.4%)
1.5%
(0.7% - 3.7%)
2.8%
(1.4% -5.5%)
2.3%
(1.2% -4.9%)
2.6%
(1.3% -5.2%)
2.1%
(1%-4.6%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-24
June 2009
-------
Table 4-10. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response
(Defined in Terms of sRaw) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and
SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Increase in sRaw >= 100%
Greene County, MO
2-Parameter Logistic
Probit
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.9%)
1 .4%
(0.6% - 3.7%)
0.9%
(0.3% - 2.7%)
0.4%
(0.1% -1.8%)
0.1%
(0% - 0.8%)
0.4%
(0.1% -1.9%)
0.1%
(0% - 0.9%)
0.5%
(0.1% -2.1%)
0.2%
(0%-1.1%)
0.7%
(0.2% - 2.4%)
0.3%
(0.1% -1.4%)
1%
(0.3% - 2.9%)
0.5%
(0.2% - 1 .9%)
0.9%
(0.3% - 2.7%)
0.4%
(0.1% -1.7%)
Sf. Louis, MO
2-Parameter Logistic
Probit
1 .4%
(0.5% - 3.8%)
0.8%
(0.2% - 2.8%)
19.2%
(14.6% -24.9%)
19.1%
(14.4% -24.7%)
0.9%
(0.3% - 2.9%)
0.4%
(0.1% -1.9%)
2.9%
(1.3% -6.3%)
2.1%
(0.9% - 5.3%)
5.4%
(3% - 9.6%)
4.6%
(2.4% - 8.8%)
8.1%
(5% -12.8%)
7.4%
(4. 5% -12. 3%)
10.9%
(7.3% -16%)
10.4%
(6.9% -15.6%)
10.3%
(6.8% -15.3%)
9.7%
(6.3% -14.9%)
Response = Increase in sRaw >= 200%
Greene County, MO
2-Parameter Logistic
Probit
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.5%
(0.1% -1.8%)
0.2%
(0.1% -1.2%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.1%
(0%-1%)
0%
(0% - 0.4%)
0.2%
(0%-1.1%)
0%
(0% - 0.5%)
0.2%
(0% - 1 .3%)
0.1%
(0% - 0.6%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
0.3%
(0.1% -1.4%)
0.1%
(0% - 0.8%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.5%
(0.1% -1.9%)
0.2%
(0% - 1 .2%)
8.1%
(5.3% -12.2%)
7.9%
(5% -12%)
0.3%
(0.1% -1.5%)
0.1%
(0% - 0.8%)
1%
(0.3% - 3%)
0.6%
(0.2% - 2.3%)
1.9%
(0.8% - 4.5%)
1 .4%
(0.5% -3. 8%)
3%
(1.5% -6%)
2.5%
(1.2% -5.4%)
4.2%
(2.3% - 7.5%)
3.7%
(2% - 7%)
3.9%
(2.1% -7.2%)
3.4%
(1.8% -6. 7%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-25
June 2009
-------
Table 4-11. Percent of Asthmatics Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response (Defined
in Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and SO2
Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FEV1 >= 15%
Greene County, MO
2-Parameter Logistic
Probit
0.2%
(0%-2.1%)
0%
(0%-1.3%)
0.8%
(0.2% - 3.4%)
0.5%
(0.1% -2. 8%)
0.2%
(0%-2.1%)
0%
(0%-1.3%)
0.2%
(0%-2.1%)
0%
(0% - 1 .4%)
0.3%
(0% - 2.3%)
0.1%
(0% - 1 .5%)
0.4%
(0.1% -2. 5%)
0.1%
(0% - 1 .8%)
0.5%
(0.1% -2. 9%)
0.3%
(0.1% -2.1%)
0.5%
(0.1% -2.8%)
0.2%
(0% - 2%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.7%
(0.2% - 3.5%)
0.4%
(0.1% -2.8%)
14.9%
(10% -22%)
14.7%
(9.9% -22.1%)
0.5%
(0.1% -2.9%)
0.2%
(0%-2.1%)
1 .7%
(0.6% - 5.5%)
1 .2%
(0.4% - 4.9%)
3.4%
(1.5% -8.3%)
2.9%
(1.2% -8%)
5.4%
(2. 8% -11.1%)
5%
(2.5% -11%)
7.7%
(4.4% -13.9%)
7.4%
(4.2% -13.9%)
7.2%
(4.1% -13.3%)
6.8%
(3.8% -13.3%)
Response = Decrease in FEV1 >= 20%
Greene County, MO
2-Parameter Logistic
Probit
0%
(0% - 0.2%)
0%
(0% - 0%)
0.1%
(0% - 0.6%)
0.1%
(0% - 0.4%)
0%
(0% - 0.2%)
0%
(0% - 0%)
0%
(0% - 0.2%)
0%
(0% - 0%)
0%
(0% - 0.2%)
0%
(0%-0.1%)
0%
(0% - 0.3%)
0%
(0%-0.1%)
0.1%
(0% - 0.4%)
0%
(0% - 0.2%)
0.1%
(0% - 0.4%)
0%
(0% - 0.2%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.1%
(0% - 0.6%)
0%
(0% - 0.3%)
9%
(6% -13.5%)
9%
(6% -13. 5%)
0.1%
(0% - 0.4%)
0%
(0% - 0.2%)
0.3%
(0.1% -1.3%)
0.2%
(0.1% -0.9%)
1%
(0.4% - 2.6%)
0.8%
(0.4% - 2.3%)
2.1%
(1%-4.3%)
1.9%
(1%-4.1%)
3.5%
(1.9% -6.4%)
3.4%
(1.9% -6.2%)
3.1%
(1.7% -5.9%)
3%
(1.7% -5. 7%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
***The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-26
June 2009
-------
Table 4-12. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion Estimated to Experience At Least One Lung Function Response
(Defined in Terms of FEVi) Associated with Exposure to "As Is" SO2 Concentrations, SO2 Concentrations that Just Meet the Current Standards, and
SO2 Concentrations that Just Meet Alternative Standards*
Exposure-Response
Model
"As is" SO2
Concentrations**
SO2
Concentrations
that Just Meet the
Current
Standards***
SO2 Concentrations that Just Meet Alternative nth Percentile 1-Hr Daily Maximum Standards, with Levels (in
ppb) of m (Standard Denoted n/m):
99/50
99/100
99/150
99/200
99/250
98/200
Response = Decrease in FEV1 >= 15%
Greene County, MO
2-Parameter Logistic
Probit
0.2%
(0% - 2.2%)
0%
(0% - 1 .4%)
1 .2%
(0.4% - 4.4%)
0.8%
(0.2% - 3.8%)
0.2%
(0%-2.1%)
0%
(0%-1.3%)
0.2%
(0% - 2.2%)
0%
(0% - 1 .4%)
0.3%
(0.1% -2. 4%)
0.1%
(0% - 1 .7%)
0.5%
(0.1% -2. 9%)
0.2%
(0%-2.1%)
0.7%
(0.2% - 3.5%)
0.4%
(0.1% -2.8%)
0.6%
(0.2% - 3.2%)
0.3%
(0.1% -2. 5%)
Sf. Louis, MO
2-Parameter Logistic
Probit
1.1%
(0.3% - 4.5%)
0.7%
(0.1% -3.9%)
22.3%
(15.9% -30.4%)
22.3%
(15.9% -30.7%)
0.7%
(0.2% - 3.5%)
0.4%
(0.1% -2.8%)
2.6%
(0.9% - 7.5%)
2%
(0.6% - 7.2%)
5.3%
(2.5% -11. 5%)
4.7%
(2.1% -11. 5%)
8.4%
(4.6% -15.4%)
8%
(4.3% -15.6%)
11.9%
(7.3% -19.3%)
1 1 .7%
(7% -19.6%)
11.1%
(6.7% -18.5%)
10.9%
(6.4% -18.8%)
Response = Decrease in FEV1 >= 20%
Greene County, MO
2-Parameter Logistic
Probit
0%
(0% - 0.2%)
0%
(0% - 0%)
0.2%
(0.1% -0.9%)
0.1%
(0.1% -0.6%)
0%
(0% - 0.2%)
0%
(0% - 0%)
0%
(0% - 0.2%)
0%
(0% - 0%)
0%
(0% - 0.3%)
0%
(0%-0.1%)
0.1%
(0% - 0.4%)
0%
(0% - 0.2%)
0.1%
(0% - 0.5%)
0.1%
(0% - 0.3%)
0.1%
(0% - 0.5%)
0%
(0% - 0.3%)
Sf. Louis, MO
2-Parameter Logistic
Probit
0.2%
(0% - 0.8%)
0.1%
(0% - 0.5%)
14.7%
(10.1% -20.8%)
14.9%
(10. 3% -21%)
0.1%
(0% - 0.5%)
0%
(0% - 0.3%)
0.6%
(0.2% - 2%)
0.4%
(0.1% -1.6%)
1 .7%
(0.7% -4.1%)
1.5%
(0.7% - 3.7%)
3.4%
(1.8% -6.8%)
3.3%
(1.7% -6.6%)
5.8%
(3.4% -10%)
5.7%
(3.4% - 9.9%)
5.2%
(3% - 9.2%)
5.1%
(3% - 9.2%)
*Percents are median (50th percentile) percents of asthmatic children. Percents in parentheses below the median are 95% credible intervals based on statistical uncertainty surrounding
the SO2 coefficient in the logistic and probit exposure-response functions. Percents are rounded to the nearest tenth.
"The "as is" exposure scenario was based on monitoring and modeling using 2002 air quality information.
"The current primary SO2 standards include a 24-hour standard set at 0.14 parts per million (ppm), not to be exceeded more than once per year, and an annual standard set at 0.03 ppm,
calculated as the arithmetic mean of hourly averages.
Abt Associates Inc.
4-27
June 2009
-------
Figure 4-1. Percent of Asthmatics Engaged in Moderate or Greater Exertion in St. Louis Exhibiting
Lung Function Response (Defined as an Increase in sRaw > 100%) Attributable to SO2 Within Given
Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30%
=• 25% -
O
^
in
o>
1 20%
re
O)
c
•o
g 15% -
Q.
V
o:
I 10% -
U
0)
5% -
0% H
T
_ . .
31 j»
^++
tS5
S3
i1^
J-, ^
T
= = ii |S
T ,1, m ri rn
T ••!••• ^.-^^ .^.^. ^^.^ ^.^./
^L I-^H — -^- ^^^ -^-
^^^^ i i i i i i
.— _; o o o o o o
< >, ii in o in o in o
:SM S r r C! C! C!
(/) (Q +•* O O O O O 00
3 a g
: 3
O
b) Probit Exposure-Response Function
30%
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-28
June 2009
-------
Figure 4-2. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion in St. Louis
Exhibiting Lung Function Response (Defined as an Increase in sRaw > 100%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
J2 re
2 5
10
o>
o>
C!
o>
o>
in
C!
o>
o>
C!
CO
o>
o
b) Probit Exposure-Response Function
30% -
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-29
June 2009
-------
Figure 4-3. Percent of Asthmatics Engaged in Moderate or Greater Exertion in St. Louis Exhibiting
Lung Function Response (Defined as a Decrease in FEVi > 15%) Attributable to SO2 Within Given
Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
H H
C!
in
C!
o>
o>
C!
m
b) Probit Exposure-Response Function
30% -
O
,0 25% -
IO
O>
•a
TO 20% -
O)
c
•a
c 15% -
O
0.
(A
0)
10% -
c
0)
o
0)
Q- 5% -
0% -I
T
_L
1 L * i '
±i
J2 to
j? O
g
_ , _
~_ '_"
_ '_"
~_ '_"
Kif
JJJ
U ' J '
Vrt
V?
^
B
'c
£
=
O
I rrn
o o
in o
o> ^
o> o>
o>
=
M
c
U
T
a
a
S
M
9
9
9
9
4
?S
C
C
£
a
a
ft
3
3
M
r>
r>
B
X
&
C
U
£
c
c
as
V
3
n
>j
r>
r>
3m
m.
f,
'&
C
C
E
o
a
^
*w
V
5=S
9
9
4
3
»
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-30
June 2009
-------
Figure 4-4. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion in St. Louis
Exhibiting Lung Function Response (Defined as a Decrease in FEVi > 15%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
O
0 25% -
c£
m
•a
TO 20% -
0)
c 15% -
o
0.
(A
0
&.
^ 10% -
0)
0)
Q- 5% -
0% -I
I '
±i
4 o
_•
-;
-;
K
ft
^
//
ft
t
o>
c
c
T
c
c
9
9
>
>
\
p
^
C
V
T1
c
c
m
"%
>
>
i
»
CE
V
"S-
c
c
E
<
s
rt
•tf-i
9
9
M
>
\
~
_
Si
i*
$
C
u
£
c
c
•=•
_
a
tj
s
9
n
M
>
>
+
K
^
c
c
E
<
K
//
*
9
9
M
>
b) Probit Exposure-Response Function
30% -
O
- 25% -
0s
m
•a
TO 20% -
0)
^3
C 1 S%
0
Q.
(A
fl)
+. 10% -
c
0)
Q- 5% -
0% -I
^ >"l
*s
< 0
~-
_•
rj
^
^
V
s
w
C
*
!
1
t
"-
:-
rar
$
+
1
d"
?
W
9
^
D
5
_L
o
in
o>
o>
r*
c
c
T
c
c
*T
9
9
>
>
S
s?
c
h
T
C
c
IB?
9
)
»
»
ft
~
x
%•
C
C
E
<
<
V
0
^
9
9
M
>
>
Q
I!
H!
?
^;
c
u
£
c
c
a
il
•j
x
x^
9
>
M
>
>
If.
t
J
?^/
C
C
E
(
<
d
t
y.
^
9
9
M
)
>
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-31
June 2009
-------
Figure 4-5. Number of Occurrences of Lung Function Response (Defined as an Increase in sRaw >
100%) Among Asthmatics Engaged in Moderate or Greater Exertion in St. Louis Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600 -r
o
^
in
o>
•O
C
3.
in
•O
C
re
V)
o
500 -
400 -
300 -
u
u
O
200 -
100 -
b) Probit Exposure-Response Function
600 ^
o>
o>
o>
o>
o>
C!
o>
o>
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-32
June 2009
-------
Figure 4-6. Number of Occurrences of Lung Function Response (Defined as an Increase in sRaw >
100%) Among Asthmatic Children Engaged in Moderate or Greater Exertion in St. Louis
Attributable to SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600 -,
o
500 -
400 -
300 -
u
u
O
200 -
100 -
600 -i
o
in
o>
•D
500-
400 -
300 -
U
8
200 -
100 -
b) Probit Exposure-Response Function
. &
J2 "re
o> ~
o> o>
o>
o> o>
o> o>
o>
o>
CO
o>
o
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-33
June 2009
-------
Figure 4-7. Number of Occurrences of Lung Function Response (Defined as a Decrease in FEVi >
15%) Among Asthmatics Engaged in Moderate or Greater Exertion in St. Louis Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600
o
^
g
•O
C
3.
in
•O
500 -
400 -
c 300 -
200 -
u
u
O
100 -
« re
3°
B
v>
o
in
C!
o>
o>
C!
CO
o>
b) Probit Exposure-Response Function
600 i
o
g 500
•D
C
3.
1/1
•D
400 -
c 300 -
U
C
1
u
u
O
200 -
100 -
o>
C!
CO
en
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-34
June 2009
-------
Figure 4-8. Number of Occurrences of Lung Function Response (Defined as a Decrease in FEVi >
15%) Among Asthmatic Children Engaged in Moderate or Greater Exertion in St. Louis
Attributable to SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600 i
o
^
8 500 H
•O
400 -
300 -
200 -
o>
o>
o>
o>
C!
o>
o>
in
C!
o>
o>
b) Probit Exposure-Response Function
o
§ 500 -I
•O
400 -
300 -
200 -
100 -
re
2 a
V)
o
!S
m
m
o
o
a>
a>
a>
a>
a>
a>
a>
a>
g
o>
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-35
June 2009
-------
Figure 4-9. Percent of Asthmatics Engaged in Moderate or Greater Exertion in Greene Co.
Exhibiting Lung Function Response (Defined as an Increase in sRaw > 100%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
=• 25%
o
20% -
O)
c
15% -
10% -
.
= s
o
in
s
O
i
o
O
O)
O
O)
O
O)
O
O)
00
O)
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-36
June 2009
-------
Figure 4-10. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion in Greene Co.
Exhibiting Lung Function Response (Defined as an Increase in sRaw > 100%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
=• 25%
o
•D
O
Q.
V)
20% -
15% -
10% -
v> re
3 a
I
I
o
in
o>
o>
C!
o>
o>
in
C!
o>
o>
C!
CO
o>
b) Probit Exposure-Response Function
30% -
=• 25%
O
O)
c
20% -
15% -
10% -
5% -
0%
~~ ^
< O
O
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-37
June 2009
-------
Figure 4-11. Percent of Asthmatics Engaged in Moderate or Greater Exertion in Greene Co.
Exhibiting Lung Function Response (Defined as a Decrease in FEVi > 15%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
O
« as
B
v>
o
in
C!
o>
o>
C!
CO
o>
b) Probit Exposure-Response Function
30% -
" 25% -
&)
O>
•a
TO 20% -
O)
_c
T3
c 15% -
O
Q.
(A
0)
2: 10% -
(1)
o
0)
Q- 5% -
noi
U A) ~
T 1 T T T JL _L JL
.- _ ; o o o o o o
< >, 2 jo o to o to o
to 15 *^ ^ ^ ^ ^ ^ ^
~~ ^ C O) O) O) O) O)
30 g
o
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-38
June 2009
-------
Figure 4-12. Percent of Asthmatic Children Engaged in Moderate or Greater Exertion in Greene Co.
Exhibiting Lung Function Response (Defined as a Decrease in FEVi > 15%) Attributable to SO2
Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
30% -
25% -
20% -
0)
15% -
(A
0)
2:10%
0)
0)
Q- 5%
0%
< £
rf O
o
!S
m
m
o
o
a>
a>
b) Probit Exposure-Response Function
30% -
" 25% -
0s
m
O>
T3
TO 20% -
0)
c 15% -
(A
(1)
2: 10% -
0)
0)
Q- 5% -
0% -i
T 1 T T T _L -L -L
•— — ; o o o o o o
< >, ii IA o to o to o
(/) "J5 *^ o^ o) o) o) o) oo
^ ^ ^ ^* ^* ^* o o
So 2
o
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-39
June 2009
-------
Figure 4-13. Number of Occurrences of Lung Function Response (Defined as an Increase in sRaw >
100%) Among Asthmatics Engaged in Moderate or Greater Exertion in Greene Co. Attributable to
SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600
o
500 -
400 -
300 -
u
u
O
200 -
100 -
b) Probit Exposure-Response Function
o
•D
C
3.
1/1
•D
400 -
300 -
U
8
200 -
100 -
< £
-------
Figure 4-14. Number of Occurrences of Lung Function Response (Defined as an Increase in sRaw >
100%) Among Asthmatic Children Engaged in Moderate or Greater Exertion in Greene Co.
Attributable to SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
o
1 500-
•o
c
3.
in
c 400 -
re
U)
0
c 300 -
a)
u
c
1 200 -
u
u
O
'o
jj 100 -
E
3
Z
.- -j 0 0 0
< >, i! ;n o in
t/i T5 ** G) 0> o
~~ -3 c o o
3.0 g
O
b) Probit Exposure-Response
o
I 500-
•D
C
I/I
I 400 -
re
U)
0
!i 300 -
0)
u
c
1 200 -
u
1^
o
| 100 -
T T T T T
•— _; o o o
< >, S m o in
: £ W S C C
— 3 C 0>0>
o
o o o
o in o
0) 0) 00
0) 0) 0>
Function
T T T
o o o
o in o
515151
O) O) O)
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-41
June 2009
-------
Figure 4-15. Number of Occurrences of Lung Function Response (Defined as a Decrease in FEVi >
15%) Among Asthmatics Engaged in Moderate or Greater Exertion in Greene Co. Attributable to
SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
600 -i
o
§ 500 -
•D
C
3.
1/1
c 400 -
re
in
o
.c
c 300 -
in
u
0)
5 200 -
u
u
O
'o
« 100 -
z
0 -
IIIIIIII
.— _; o o o o o o
< >, 2 m o in o in o
. £ (/j 5 ^ ^ 51 51 51
(/) T5 ** 0> 0> 0> 0> 0> CO
^ ^ C 0) 0) 0) 0) 0)
30 g
O
b) Probit Exposure-Response Function
600 T
o
§ 500 -
•D
C
3.
1/1
c 400 -
re
in
o
c 300 -
in
u
0)
5 200 -
u
u
O
'o
« 100 -
|
z
0 -
I 1 1 I 1 1 I 1
.— _; o o o o o o
< >, 2 m o in o in o
(/)^Q4-i O O O O O 00
~~ ^ C O) O) O) O) O)
30 g
o
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-42
June 2009
-------
Figure 4-16. Number of Occurrences of Lung Function Response (Defined as a Decrease in FEVi >
15%) Among Asthmatic Children Engaged in Moderate or Greater Exertion in Greene Co.
Attributable to SO2 Within Given Ranges Under Different Air Quality Scenarios*
a) Logistic Exposure-Response Function
o
§ 500-
•D
-2-
c 400 -
0
c 300 -
a>
u
c
a>
3 200 -
u
8
'o
| 100 -
E
3
0 -
_L_L_L_L_L_L_L_L
.— _; o o o o o o
<>.£ in o in o in o
: S w « s: s: d d d
(/) (g *^ O) O) O) O) O) 00
^ 3 C O) O) O) O) O)
^ a £
o
b) Probit Exposure-Response Function
o
§ 500-
•D
-2-
c 400 -
0
c 300 -
.£ in o in o in o
r S w % :: :: C< C< d
(/) (g *^ O) O) O) O) O) 00
^ 3 C O) O) O) O) O)
S. a £
o
*For the legend for these figures see Figure 4-17.
Abt Associates Inc.
4-43
June 2009
-------
Figure 4-17. Legend for Figures 4-1 - 4-16.
Q Attributable to 500 ppb<=S02
E Attributable to 450 ppb<=S02<500 ppb
D Attributable to 400 ppb<=S02<450 ppb
C Attributable to 350 ppb<=S02<400 ppb
• Attribuable to 300 ppb<=S02<350 ppb
H Attributable to 250 ppb<=S02<300 ppb
D Attributable to 200 ppb<=S02<250 ppb
0 Attributable to 150 ppb<=S02<200 ppb
Q Attributable to 100 ppb<=S02<150 ppb
D Attributable to 50 ppb<=S02<100 ppb
• Attributable to S02<50 ppb
The current primary SC>2 standards include a 24-hour standard set at 0.14 parts per
million (ppm), not to be exceeded more than once per year, and an annual standard set at
0.03 ppm, calculated as the arithmetic mean of hourly averages. In St. Louis, 862
concentrations that are predicted to occur if the current standards were just met are
substantially higher than "as is" air quality (based on 2002 monitoring and modeling
data) and also substantially higher than they would be under any of the alternative 1-hr
standards considered in this analysis. Consequently, the levels of response that would be
seen if the current standard were just met are well above the levels that would be seen
under the "as is" air quality scenario or under any of the alternative 1-hr standards - for
asthmatics and for asthmatic children, and for all four definitions of lung function
response.
For example, of the estimated approximately 102,400 asthmatics engaged in
moderate or greater exertion in St. Louis, about 13,500 (or 13.1%) are estimated to have
at least one lung function response, defined as an increase in sRaw > 100%, if the current
standards were just met. Under "as is" air quality conditions, the corresponding number
is about 1,000 (1%). Only the most stringent alternative 99th percentile 1-hr standard, set
at 50 ppb (denoted "99/50" in the above tables of results), is predicted to lower the
numbers of responders below the levels estimated under the "as is" scenario. As the
Abt Associates Inc.
4-44
June 2009
-------
alternative 1-hr standards become less stringent (i.e., as the level is raised from 50 ppb to
100 ppb, to 150 ppb, etc.), the numbers responding correspondingly rise.
The pattern seen in St. Louis for lung function response, defined as an increase in
sRaw > 100%, is also seen for the other definitions of lung function response. For
example, of the estimated roughly 102,400 asthmatics engaged in moderate or greater
exertion, 750 are estimated to have at least one lung function response, defined as a
decrease in FEVi > 15%, under the "as is" air quality scenario; the corresponding number
(percent) if the current standards were just met is about 15,200 (14.9%); the
corresponding numbers for the alternative 1-hr standards denoted 99/50, 99/100, 99/150,
99/200, 99/250, and 98/200 are about 500 (0.5%), 1700 (1.7%), 3500 (3.4%), 5600
(5.4%), 7900 (7.7%), and 7400 (7.2%), respectively.
Although the basic pattern across air quality scenarios seen in St. Louis is
repeated in Greene County, the impact of changing from one air quality scenario to
another is substantially dampened in Greene County. This is because of the different
patterns of exposures in the two locations. In St. Louis there is a wide range of 862
concentrations to which asthmatics are exposed under the current standards scenario -
i.e., substantial percentages of asthmatics are exposed to relatively higher concentrations
of SC>2 under this scenario. There is thus much room for improvement. Under the most
stringent alternative 1-hr standard (99/50), much of that exposure is pushed down to the
lowest SC>2 concentration "bins." Under the current standards scenario, for example, only
about 22 percent of asthmatics in St. Louis have exposures no greater than 100 ppb;
under the most stringent alternative 1-hr standard (99/50), that increases to 98 percent.
In Greene County, in contrast, about 95 percent of asthmatics have exposures no
greater than 100 ppb under the current standards scenario. There is therefore little room
for improvement. Under the most stringent alternative 1-hr standard (99/50), that 95
percent becomes 100 percent. The situation is even more extreme for person days of
exposure. Under the current standards scenario, 99.9 percent of person days of exposure
are to < 100 ppb SC>2 in Greene County; the corresponding figure for St. Louis is 95.2
percent.
The generally lower levels of SC>2 to which asthmatics in Greene County are
exposed, relative to asthmatics in St. Louis, and the corresponding greater preponderance
of responses associated with the lowest 862 concentration "bins" in Greene County, can
be readily seen in Figures 4-1 through 4-8.12
Although the numbers are smaller for asthmatic children (because the underlying
populations are smaller), the patterns seen in St. Louis and in Greene County across the
different air quality scenarios, and the comparisons between the two locations, are fairly
similar for asthmatic children as for asthmatics for all lung function response definitions.
12 In several cases, responses associated with exposures in SO2 bins cannot be seen in the figures, because
the percent responding, or numbers of occurrences of lung function response are so small. We chose to
scale the y-axis the same on all comparable figures to facilitate comparisons between figures. This meant,
however, that some "response bars" essentially became visually undetectable.
Abt Associates Inc. 4-45 June 2009
-------
In general, however, the percentages of asthmatic children engaged in moderate or
greater exertion who experience at least one lung function response, for each of the
different lung function response definitions, tend to be greater than the corresponding
percentages of asthmatics. This presumably is a reflection of the greater amount of time
spent outdoors by asthmatic children relative to adults.
Finally, we note that, while in several air quality scenarios the great majority of
occurrences of lung function response are in the lowest exposure bin, the numbers of
individuals with at least one lung function response attributable to exposures in that
lowest bin are typically quite small. This is because the calculation of numbers of
individuals with at least one lung function response uses individuals' highest exposure
only. While individuals may be exposed mostly to low SO2 concentrations, many are
exposed at least occasionally to higher levels. Thus, the percentage of individuals in a
designated population with at least one lung function response associated with SC>2
concentrations in the lowest bin is likely to be very small, since most individuals are
exposed at least once to higher SO2 levels. For example, defining lung function response
as an increase in sRaw > 100%, under a scenario in which 862 concentrations just meet
an alternative 1-hour 99th percentile 100 ppb standard, about 93 percent of occurrences of
lung function response among asthmatics in St. Louis are associated with SC>2 exposures
in the lowest exposure bin (0 - 50 ppb). However, the lowest SO2 exposure bin accounts
for only about 0.2 percent of asthmatics estimated to experience at least 1 SCVrelated
lung function response. For this very small percent of the population, the lowest
exposure bin represents their highest SC>2 exposures under moderate exertion in a year.
Thus Figure 4-5b shows virtually all of the occurrences among asthmatics in St. Louis
associated with the lowest 862 exposure bin; however, Figure 4-lb shows a relatively
small proportion of asthmatics in St. Louis experiencing at least one response to be
experiencing those responses because of exposures in that lowest exposure bin.
Abt Associates Inc. 4-46 June 2009
-------
5 REFERENCES
Abt Associates Inc. 2007a. Ozone Health Risk Assessment for Selected Urban Areas.
Prepared for Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC., July 2007, Under Contract No. 68-D-
03-002, Work Assignment 3-39 and 4-56. Available electronically on the internet at:
http://www.epa.gOv/ttn/naaqs/standards/ozone/s o3crtd.html.
Abt Associates Inc. 2007b. TRIM: Total Risk Integrated Methodology. Users Guide for
TRIM.RiskHuman Health-Probabilistic Application for the Ozone NAAQS Risk
Assessment. Available online at:
http://epa.gov/ttn/fera/data/tri m/trimrisk_ozone_ra_userguide_8-6-07.pdf
Bethel RA, Epstein J, Sheppard D, Nadel JA, Boushey HA. 1983. Sulfur dioxide-induced
bronchoconstriction in freely breathing, exercising, asthmatic subjects. Am Rev Respir
Dis 128:987-90.
Bethel RA, Sheppard D, Geffroy B, Tarn E, Nadel JA, Boushey HA. 1985. Effect of 0.25
ppm sulfur dioxide on airway resistance in freely breathing, heavily exercising, asthmatic
subjects. Am Rev Respir Dis 131:659-61.
Gelman, A, Carlin, J. C., Stern, H., and Rubin, D. B. 1995. Bayesian Data Analysis.
Chapman and Hall, New York.
Gilks, W. R., Richardson, S., and Spiegelhalter, D. J. (Eds.) 1996. Markov Chain Monte
Carlo in Practice. Chapman and Hall, London, UK.
Kehrl HR, Roger LJ, Hazucha MJ, Horstman DH. 1987. Differing response of asthmatics
to sulfur dioxide exposure with continuous and intermittent exercise. Am Rev Respir Dis
135:350-5.
Linn WS, Avol EL, Peng RC, Shamoo DA, Hackney JD. 1987. Replicated dose-response
study of sulfur dioxide effects in normal, atopic, and asthmatic volunteers. Am Rev Respir
Dis 136:1127-34.
Linn WS, Avol EL, Shamoo DA, Peng RC, Spier CE, Smith MN, Hackney JD. 1988.
Effect of metaproterenol sulfate on mild asthmatics' response to sulfur dioxide exposure
and exercise. Arch Environ Health 43:399-406.
Linn WS, Shamoo DA, Peng RC, Clark KW, Avol EL, Hackney JD. 1990. Responses to
sulfur dioxide and exercise by medication-dependent asthmatics: effect of varying
medication levels. Arch Environ Health 45:24-30.
Abt Associates Inc. 5-1 June 2009
-------
Magnussen H, Torres R, Wagner HM, von Nieding G. 1990. Relationship between the
airway response to inhaled sulfur dioxide, isocapnic hyperventilation, and histamine in
asthmatic subjects. IntArch Occup Environ Health 62:485-91.
Roger LJ, Kehrl HR, Hazucha M, Horstman DH. 1985. Bronchoconstriction in
asthmatics exposed to sulfur dioxide during repeated exercise. J ApplPhysiol 59:784-
91.Spiegelhalter, D., Thomas, A., Best, N. and Gilks, W. 1996. Bugs .5 Bayesian
inference using Gibbs sampling. Manual, version ii. MRC Biostatistics Unit, Institute of
Public Health. Cambridge, U.K.
U.S. EPA. 2007a. Integrated Plan for Review of the Primary National Ambient Air
Quality Standards for Sulfur Oxides. October 2007. Available online at:
http://www.epa.gov/ttn/naaqs/standards/nox/s_nox_cr_pd.html.
U.S. EPA, National Center for Environmental Assessment. 2008a. Integrated Science
Assessment for Oxides of Sulfur - Health Criteria (Second External Review Draft).
EPA/600/R-08/047. May 2008. Available online at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 90346.
U.S. EPA, National Center for Environmental Assessment. 2008c. Integrated Science
Assessment (ISA) for Sulfur Oxides - Health Criteria (Final Report). EPA/600/R-
08/047F. September 2008. Available online at:
http ://cfpub. epa. gov/ncea/cfm/recordisplay. cfm?deid= 198843.
U.S. EPA, Office of Air Quality Planning and Standards. 2008b. Risk and Exposure
Assessment to Support the Review of the SO2 Primary National Ambient Air Quality
Standard: First Draft. EPA-452/P-08-003. June 2008. Available online at:
http://www.epa.gOv/ttn/naaqs/standards/so2/s so2crrea.html.
U.S. EPA, Office of Air Quality Planning and Standards. 2009a. Risk and Exposure
Assessment to Support the Review of the SO2 Primary National Ambient Air Quality
Standards: Second Draft. EPA-452/P-09-003. March 2009. Available online at:
http://www.epa.gov/ttn/naaqs/standards/so2/s_so2_cr_rea.html.
U.S. EPA, Office of Air Quality Planning and Standards. 2009b. Risk and Exposure
Assessment to Support the Review of the SO2 Primary National Ambient Air Quality
Standards. EPA-452/R-09-007. July 2009. Available online at:
http://www.epa.gOv/ttn/naaqs/standards/so2/s so2 cr rea.html.
Abt Associates Inc. 5-2 June 2009
-------
Appendix A: Bayesian-Estimated Logistic and Probit Exposure-Response
Functions: Median, 2.5th Percentile, and 97.5th Percentile Curves
Abt Associates Inc. A-l June 2009
-------
Figure A-la. Bayesian-Estimated Logistic Exposure-Response Function: Increase in sRaw > 100%
for 5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
v
I
v
U)
I
U)
£
70%
60%
20%
0%
A The data
median curve
2.5th percentile curve
97.5th percentile curve
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure A-lb. Bayesian-Estimated Probit Exposure-Response Function: Increase in sRaw > 100% for
5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% i
90%
80%
70%
« 60%
50%
40%
30%
20%
10%
0%
A The data
median curve
2.5th percentile curve
97.5th percentile curve
A-'
0.4 0.6
SO2 Concentration (ppm)
0.8
Abt Associates Inc.
A-2
June 2009
-------
Figure A-2a. Bayesian-Estimated Logistic Exposure-Response Function: Increase in sRaw > 200%
for 5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
70%
« 60%
8.
a>
U)
I
U)
&
20%
0%
A The data
median curve
2.5th percentile curve
97.5th percentile curve
0.4 0.6
SO2 Concentration (ppm)
0.8
Figure A-2b. Bayesian-Estimated Probit Exposure-Response Function: Increase in sRaw > 200% for
5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% i
80%
70%
« 60%
£
-------
Figure A-3a. Bayesian-Estimated Logistic Exposure-Response Function: Decrease in FEVi > 15%
for 5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% i
80%
70%
S> 60%
£
3!
I
V)
A The data
median curve
2.5th percentile curve
97.5th percentile curve
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure A-3b. Bayesian-Estimated Probit Exposure-Response Function: Decrease in FEVi > 15% for
5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% -,
0.4 0.6
SO2 Concentration (ppm)
0.8
Abt Associates Inc.
A-4
2009
-------
Figure A-4a. Bayesian-Estimated Logistic Exposure-Response Function: Decrease in FEVi > 20%
for 5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100%
80%
70%
« 60%
£
0)
V)
c
o
Q.
V)
V
£
20%
A The data
median curve
2.5th percentile curve
97.5th percentile curve
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Figure A-4b. Bayesian-Estimated Probit Exposure-Response Function: Decrease in FEVi > 20% for
5-Minute Exposures of Asthmatics Engaged in Moderate or Greater Exertion
100% i
£
V
V)
c
o
Q.
V)
& 40%
20%
10%
A The data
median curve
2.5th percentile curve
97.5th percentile curve
0.2 0.4 0.6
SO2 Concentration (ppm)
0.8
Abt Associates Inc.
A-5
June 2009
-------
APPENDIX D: SUPPLEMENT TO THE POLICY ASSESSMENT
D-l
-------
Table D-1. 99th percentile 24-hour average SO2 concentrations for 2005 given just
meeting the alternative 1-hour daily maximum standards analyzed in the risk
assessment (concentrations in ppb).
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
MI
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
99th percentile
50
7
14
10
11
10
8
5
14
9
17
16
13
14
8
12
19
18
25
9
12
17
19
12
15
14
10
11
16
19
19
17
7
9
21
13
14
11
43
100
14
27
20
22
20
15
9
27
17
35
32
26
28
17
25
38
36
49
18
25
33
37
24
30
29
20
22
32
38
38
35
13
18
43
25
27
21
87
150
20
41
31
33
29
23
14
41
26
52
48
39
43
25
37
57
55
74
28
37
50
56
35
44
43
29
33
48
57
56
52
20
26
64
38
41
32
130
200
27
55
41
44
39
31
19
54
34
70
64
52
57
34
50
76
73
98
37
50
66
74
47
59
58
39
45
65
76
75
70
27
35
86
51
54
42
173
250
34
69
51
55
49
38
24
68
43
87
79
65
71
42
62
96
91
123
46
62
83
93
59
74
72
49
56
81
95
94
87
33
44
107
64
68
53
217
98th percentile
100
18
33
28
26
28
20
11
35
21
40
36
29
37
24
29
48
45
57
22
28
39
45
27
34
37
24
36
41
44
43
41
19
21
48
32
32
27
97
200
36
66
56
51
56
40
21
71
43
80
72
58
73
48
59
97
90
113
44
56
78
89
53
67
73
47
71
81
87
87
83
38
42
96
64
64
54
194
D-2
-------
Table D-2. 99th percentile 24-hour average SO2 concentrations for 2006 given just
meeting the alternative 1-hour daily maximum standards analyzed in the risk
assessment (concentrations in ppb).
State
DE
IL
IN
IN
IN
IA
IA
MI
MO
MO
NH
NY
NY
NY
OH
OH
OH
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
County
New Castle
Madison
Floyd
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
99th percentile
50
11
9
8
5
5
11
15
17
17
7
14
23
7
7
14
11
12
12
9
16
15
11
16
16
8
11
17
12
14
10
100
23
18
15
10
11
23
31
34
33
13
28
46
13
15
28
23
24
23
19
32
29
22
32
31
17
23
35
24
28
21
150
34
28
23
14
16
34
46
51
50
20
41
69
20
22
43
34
35
35
28
48
44
33
48
47
25
34
52
36
42
31
200
46
37
30
19
22
45
62
68
66
26
55
92
27
29
57
46
47
46
38
63
59
45
65
62
34
45
70
49
56
42
250
57
46
38
24
27
56
77
85
83
33
69
115
33
36
71
57
59
58
47
79
73
56
81
78
42
56
87
61
70
52
98th percentile
100
27
22
20
12
13
26
35
38
43
19
33
53
16
16
34
28
27
30
23
50
37
26
37
37
24
26
39
31
33
27
200
55
43
40
25
27
52
70
76
86
37
66
106
32
33
67
55
53
59
46
101
74
51
75
74
49
53
78
61
66
53
D-3
-------
>nd
Table D-3. 2 highest 24-hour average SO2 concentrations (i.e. the current 24-
hour standard) for 2005 given just meeting the alternative 1-hour daily maximum
standards analyzed in the risk assessment (concentrations in ppb)
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
MI
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
99th percentile
50
7
18
13
12
8
9
5
19
12
19
18
17
18
10
18
22
18
29
12
14
26
22
12
16
18
11
11
17
23
23
22
9
10
22
14
16
12
48
100
15
36
26
24
20
18
11
38
23
38
37
34
37
20
35
45
45
57
19
27
53
44
24
31
36
21
23
33
46
46
43
19
20
49
28
32
23
95
150
22
54
38
36
30
28
16
56
35
57
55
50
55
29
53
67
68
86
37
41
79
66
36
47
55
32
35
50
69
69
65
28
30
74
42
48
35
143
200
29
72
51
48
41
37
22
75
46
76
73
67
73
39
71
89
90
115
49
54
105
88
49
63
73
42
47
66
92
92
87
37
39
98
56
64
47
190
250
37
90
64
60
51
46
27
94
58
95
92
84
92
49
88
111
113
144
62
68
132
110
61
79
91
53
58
83
115
115
108
46
49
123
70
80
58
238
98th percentile
100
19
43
35
28
29
24
12
49
29
44
41
37
47
28
42
56
56
66
23
30
63
53
27
36
46
26
37
42
53
53
51
27
23
55
35
38
30
106
200
39
86
71
56
58
49
25
98
57
88
83
75
95
55
84
113
112
132
59
61
125
106
55
72
93
52
74
84
106
107
103
54
46
110
71
76
59
213
D-4
-------
>nd
Table D-4. 2 highest 24-hour average SO2 concentrations (i.e. the current 24-
hour standard) for 2006 given just meeting the alternative 1-hour daily maximum
standards analyzed in the risk assessment (concentrations in ppb)
State
DE
IL
IN
IN
IN
IA
IA
MI
MO
MO
NH
NY
NY
NY
OH
OH
OH
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
County
New Castle
Madison
Floyd
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
99th percentile
50
18
10
12
6
7
16
18
24
20
7
19
25
8
12
21
16
13
13
12
50
19
14
21
20
10
13
20
14
15
11
100
37
20
23
11
14
32
36
48
40
14
38
49
15
24
43
33
26
27
24
101
38
29
41
41
21
26
41
28
31
22
150
55
31
35
17
21
48
54
72
60
32
57
74
23
36
64
49
39
40
35
151
57
43
62
61
31
39
61
42
46
34
200
73
41
47
27
28
63
72
96
80
43
76
99
30
47
85
65
52
53
47
202
76
58
83
82
42
52
82
56
61
45
250
92
51
58
34
35
79
90
120
100
54
95
124
38
59
106
82
65
66
59
252
95
72
104
102
52
65
102
70
76
56
98th percentile
100
44
24
31
18
17
36
41
54
52
20
45
57
18
27
51
39
29
34
29
161
48
33
48
48
30
31
46
35
36
29
200
88
48
61
36
35
73
82
107
103
61
90
114
36
54
101
79
58
68
57
321
96
66
96
97
60
61
91
71
72
57
D-5
-------
Table D-5. Annual average SO2 concentrations for 2005 given just meeting the
alternative 1-hour daily maximum standards analyzed in the risk assessment
(concentrations in ppb).
State
AZ
DE
FL
IL
IL
IN
IN
IN
IN
IA
IA
MI
MO
MO
NH
NJ
NJ
NY
NY
NY
OH
OH
OH
OK
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
WV
County
Gila
New Castle
Hillsborough
Madison
Wabash
Floyd
Gibson
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Hudson
Union
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Tulsa
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
Wayne
99th percentile
50
1.5
2.3
1.5
1.8
1.5
2.3
0.8
1.8
1.9
2.0
2.3
2.4
1.9
1.4
2.4
6.4
6.2
6.9
2.1
2.3
4.6
2.8
2.7
3.6
3.6
2.8
2.9
3.2
4.7
2.9
2.9
1.5
1.4
7.8
4.5
4.3
2.6
6.0
100
2.9
4.6
2.9
3.7
3.0
4.5
1.7
3.6
3.7
4.1
4.6
4.9
3.8
2.8
4.8
12.9
12.3
13.7
4.3
4.5
9.3
5.7
5.4
7.2
7.1
5.5
5.9
6.5
9.3
5.8
5.7
3.0
2.8
15.5
8.9
8.6
5.2
12.0
150
4.3
6.9
4.4
5.5
4.5
6.8
2.5
5.4
5.5
6.1
6.9
7.3
5.7
4.2
7.2
19.3
18.4
20.6
6.4
6.8
13.9
8.5
8.1
10.7
10.7
8.3
8.8
9.7
14.0
8.7
8.6
4.5
4.2
23.2
13.4
13.0
7.8
18.0
200
5.8
9.2
5.8
7.4
6.0
9.0
3.4
7.1
7.4
8.2
9.1
9.7
7.6
5.6
9.5
25.7
24.6
27.4
8.6
9.1
18.6
11.3
10.8
14.3
14.2
11.0
11.7
13.0
18.7
11.7
11.5
6.1
5.6
31.0
17.9
17.3
10.3
24.0
250
7.2
11.5
7.3
9.2
7.5
11.3
4.2
8.9
9.2
10.2
11.4
12.1
9.5
7.1
11.9
32.1
30.7
34.3
10.7
11.3
23.2
14.1
13.5
17.9
17.8
13.8
14.6
16.2
23.3
14.6
14.4
7.6
7.1
38.7
22.4
21.6
12.9
30.0
98th percentile
100
3.8
5.5
4.0
4.3
4.3
5.9
1.9
4.7
4.6
4.7
5.2
5.4
4.9
4.0
5.7
16.3
15.2
15.8
5.1
5.1
11.0
6.8
6.1
8.2
9.0
6.7
9.3
8.2
10.7
6.7
6.8
4.4
3.3
17.3
11.3
10.2
6.6
13.4
200
7.7
11.0
8.0
8.6
8.6
11.9
3.8
9.3
9.1
9.4
10.3
10.8
9.8
8.0
11.4
32.5
30.4
31.6
10.3
10.2
22.1
13.6
12.1
16.3
18.1
13.4
18.7
16.3
21.5
13.5
13.6
8.7
6.6
34.7
22.6
20.5
13.2
26.8
D-6
-------
Table D-6. Annual average SO2 concentrations for 2006 given just meeting the
alternative 1-hour daily maximum standards analyzed in the risk assessment
(concentrations in ppb).
State
DE
IL
IN
IN
IN
IA
IA
MI
MO
MO
NH
NY
NY
NY
OH
OH
OH
PA
PA
PA
PA
PA
TN
TN
TN
TX
VA
WV
WV
WV
County
New Castle
Madison
Floyd
Lake
Vigo
Linn
Muscatine
Wayne
Greene
Jefferson
Merrimack
Bronx
Chautauqua
Erie
Cuyahoga
Lake
Summit
Allegheny
Beaver
Northampton
Warren
Washington
Blount
Shelby
Sullivan
Jefferson
Fairfax
Brooke
Hancock
Monongalia
99th percentile
50
2.2
1.7
1.6
1.7
1.4
1.8
1.7
2.2
2.0
1.5
2.1
6.5
1.6
1.5
4.1
2.4
2.2
2.7
2.0
3.7
2.5
4.3
3.0
3.7
1.8
1.4
6.9
3.9
4.1
2.0
100
4.4
3.5
3.2
3.3
2.8
3.6
3.4
4.4
4.0
3.0
4.3
13.0
3.1
3.1
8.2
4.8
4.3
5.5
4.0
7.3
4.9
8.5
6.0
7.5
3.6
2.9
13.9
7.7
8.2
3.9
150
6.7
5.2
4.8
5.0
4.2
5.4
5.2
6.6
6.1
4.5
6.4
19.5
4.6
4.6
12.4
7.2
6.5
8.2
6.0
11.0
7.4
12.8
8.9
11.2
5.3
4.3
20.8
11.6
12.3
5.8
200
8.9
6.9
6.3
6.6
5.6
7.2
6.9
8.8
8.1
5.9
8.5
26.0
6.2
6.1
16.5
9.6
8.7
10.9
8.0
14.6
9.9
17.1
11.9
14.9
7.1
5.7
27.7
15.5
16.3
7.8
250
11.1
8.6
7.9
8.3
7.0
9.1
8.6
10.9
10.1
7.4
10.7
32.5
7.7
7.6
20.6
12.0
10.9
13.7
10.0
18.3
12.3
21.3
14.9
18.6
8.9
7.2
34.6
19.3
20.4
9.7
98th percentile
100
5.3
4.0
4.2
4.3
3.4
4.2
3.9
4.9
5.2
4.2
5.1
15.0
3.7
3.4
9.8
5.8
4.9
7.0
4.9
11.6
6.2
9.8
6.9
8.8
5.1
3.4
15.5
9.8
9.7
5.0
200
10.6
8.1
8.4
8.7
6.9
8.3
7.8
9.8
10.4
8.4
10.1
29.9
7.4
6.9
19.6
11.6
9.8
13.9
9.7
23.3
12.4
19.6
13.8
17.7
10.3
6.7
31.0
19.5
19.4
9.9
D-7
-------
United States Office of Air Quality Planning and Standards EPA-452/R-09-007
Environmental Protection Air Quality Strategies and Standards Division July 2009
Agency Research Triangle Park, NC
------- |