External Review Draft No. 1
April 1980
Draft
Do Not Quote or Cite
Air Quality
for Participate Matter
and Sulfur Oxides
Volume II
Air Quality
NOTICE
This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its technical accuracy and
policy implications.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
External Review Draft No. 1
April 1980
Draft
Do Not Quote or Cite
Air Quality Criteria
for Particulate Matter
and Sulfur Oxides
Volume II
Air Quality
NOTICE
This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
Preface to Volume II
This volume discusses the chemical and physical characteristics of sulfur
oxides and particulate matter, the methods used to measure them, their sources,
emission rates, atmospheric concentrations and exposure, transport, transformation
and removal. Thus, this volume provides background information for discussion
of welfare effects in Volume III (Chapters 7, 8, 9, 10) and health effects in
Volume IV (Chapters 11, 12, 13, 14).
The link between human activities and air pollution was recognized at
least as far back as the 13th century, when King Edward I of England banned
the burning of "sea coal," under penalty of death, in an attempt to clear the
smoky sky of London. In today's industrial world, the problem is much more
profound, and a quantitative link between emissions and overall air quality is
often difficult to determine. Key factors include measurement methods, source,
emissions, mixing and dilution in the atmosphere, chemical transformations
that occur during transport, the efficiency of removal processes for pollutant
gases and particles, and the actual concentrations to which people are exposed.
An indirect approach to evaluating the impact of man-made sulfur emissions
on atmospheric composition is to compare measurements with other continental
areas similar geographically to the United States but affected to a much
lesser degree by industrial activity. Measurements in the South American
continent show fine particulate sulfur concentrations frequently in the range
q
0.05 to 0.1 ug/m of air. These concentrations are 10 to 100 times lower than
those often observed in the United States, which are 1 to 10 ug/m . This
difference suggests that the atmosphere of the United States is 10 to 100
m
-------
times above natural levels. It is reasonable to believe that man-made emissions
of SCL and its conversion to sulfates are principally responsible.
The assessment of man-made emissions on air quality is complex because of
the varied nature of particulate matter. Some effects are associated with the
mass of particulate matter, while other effects are related to particle size. In
addition, certain elemental constituents of particulate matter may be toxic to
humans, animals, and plants. Thus, the chemical composition of particulate
matter, especially in different particle size ranges, is of considerable
importance in evaluating the impact of particulate matter on air quality.
Some particles are emitted to the atmosphere from local pollution sources,
while others are generated by the action of wind on soils, the movement of
vehicular traffic, and other processes. Particulate matter may be reentrained
in the atmosphere by primary processes such as dispersion of dusts and oceanic
aerosol formation. Particulate matter may also be formed in the atmosphere by
secondary processes, such as the condensation of vapors or the chemical reaction
of atmospheric trace gases to form compounds stable in liquid or solid form.
The particle size characteristics and chemical composition of particulate
emissions may vary considerably with pollution control technology and the
types of fuels used in combustion sources. As an example, particle collectors
may greatly reduce the emissions of coarse particles but may be less effective
in reducing emissions of fine particulate matter. The combustion of unleaded
gasoline in automobiles will reduce ambient lead concentrations. The control
of combustion and exhaust gas temperatures in stationary sources may affect
the content of certain heavy metals in fly ash emissions. Many methods have
been used to measure sulfur oxides and particulate matter. Early methods were
not as accurate or as precise as current methods. By evaluating these methods,
IV
-------
more accurate assessments of historical ambient air data can be made. For
example, information of this type is useful in reevaluating concentrations and
exposures reported is epidemiological studies.
Monitoring data provide information on total mass concentrations of
particulate matter but, for the most part, do not adequately reflect particle
size and chemical composition.
The fossil fuels used in energy production include fuel oil, coal, and
natural gas. In 1975 almost two-thirds of the total sulfur oxide emissions
in the United States were attributed to coal-fired power plants, almost entirely
in the form of sulfur dioxide. Because some of the gaseous sulfur dioxide may
be oxidized to sulfate in the ambient air before removal at the surface, the
ambient air sulfate concentrations stem mainly from S0? emissions.
Nationwide, man-made emissions of sulfur oxides during the 1970's have
declined somewhat, even though the amount of fossil fuels burned has remained
almost the same. This result is attributed to the use of low-sulfur oil,
controls on the combustion process, and the use of pollution-control techniques.
In order to determine the impact of particulate matter and sulfur oxide
emissions on health and welfare, the concentrations actually delivered to man
and his environment must be determined. For both sulfur oxides and particulate
matter this determination is complex, because exposure to these pollutants is
influenced by many variables, such as geographic location, occupation, and
individual activity patterns.
In summary, the readers of this volume will find that Volume II
presents an analysis of ambient air concentrations on a wide variety of time
and geographical scales to provide a broad background for later assessments of
effects. In many cases the adverse effects on visibility and climate, vegetation,
acid rain, and other materials covered in Volume III involve wide area, long-term
-------
analysis while human health effects covered in Volume IV may involve either
short-term high-level or long-term low-level exposures. In reading the effects
chapters in these later volumes, the reader is cautioned to be aware of the
exposure variables and to refer to the appropriate sections of Volume II for
background data.
-------
CONTENTS
2. PHYSICAL AND CHEMICAL PROPERTIES OF SULFUR OXIDES AND PARTICULATE
MATTER AND ANALYTICAL TECHNIQUES FOR THEIR MEASUREMENT 2-1
2.1 INTRODUCTION 2-1
2.2 PHYSICAL AND CHEMICAL PROPERTIES OF SULFUR OXIDES AND
PARTICULATE MATTER 2-2
2.2.1 Sulfur oxides 2-2
2.2.2 Particulate matter 2-5
2.3 MEASUREMENT TECHNIQUES FOR SULFUR DIOXIDE 2-10
2.3.1 Introduction 2-10
2.3.2 Integrated methods 2-11
2.3.2.1 Sample collection 2-11
2.3.2.1.1 Direct sampling 2-12
2.3.2.1.2 Absorption 2-12
2.3.2.1.3 Adsorption 2-13
2.3.2.2 Calibration 2-14
2.3.2.3 Analytical methods 2-15
2.3.2.3.1 Colorimetric method: pararosaniline 2-15
2.3.2.3.2 Titrimetric method: hydrogen
peroxide 2-18
2.3.2.3.3 Titrimetric method: barium
perchlorate - thorin 2-18
2.3.2.3.4 Turbidimetric method: hydrogen
peroxide 2-19
2.3.2.3.5 Colorimetric method: barium
chloranilate 2-19
2.3.2.3.6 Colorimetric method: iodine 2-19
2.3.2.3.7 Titrimetric method: iodine 2-20
2.3.2.3.8 Titrimetric method: iodine-
thiosulfate 2-20
2.3.2.3.9 Colorimetric method: ammonium
molybdate 2-21
2.3.2.3.10 Ion-exchange chromatographic method 2-21
2.3.2.3.11 Sulfation method: lead peroxide 2-22
2.3.3 Continuous methods 2-23
2.3.3.1 Sample collection 2-23
2.3.3.2 Calibration 2-23
2.3.3.3 Analytical methods 2-25
2.3.3.3.1 Conductimetric analyzers 2-25
2.3.3.3.2 Colorimetric analyzers 2-26
2.3.3.3.3 Flame photometric analyzers 2-27
2.3.3.3.4 Coulometric (Amperometric)
analyzers 2-28
2.3.3.3.5 Fluorescence analyzers 2-30
2.3.3.3.6 Second-derivative spectrometric
analyzers 2-31
2.3.3.3.7 Other methods: electrochemical
transducer 2-32
2.3.4 EPA reference and equivalent methods 2-33
2.3.5 Method comparison studies 2-38
2.3.5.1 Comparisons involving sulfation methods and
relationship between sulfation and SQy
concentration 2-38
vii
-------
LIST OF TABLES
Number
11-1 Summary of the respiratory deposition and clearance of inhaled
aerosols.
11-82
-------
2.5.2.4.1.5 Air volume measure-
ment 2-131
2.5.2.4.2 Mass concentration determination... 2-132
2.5.2.4.2.1 Gravimetric determina-
tion 2-133
2.5.2.4.2.2 Nongravimetric methods 2-133
2.5.2.4.2.3 Optical determination. 2-134
2.5.2.4.2.4 Measurement by beta ray
attenuation 2-134
2.5.2.4.3 Particle sizing determination of
filtered particles 2-136
2.5.2.4.4 Chemical analysis. Analysis of
trace elements 2-137
2.5.2.4.4.1 Atomic absorption
spectrometry 2-137
2.5.2.4.4.2 Optical emission
spectrometry 2-138
2.5.2.4.4.3 Spark source mass
spectrometry 2-138
2.5.2.4.4.4 Neutron activation
analysis 2-139
2.5.2.4.4.5 X-ray fluorescence
spectrometry 2-139
2.5.2.4.4.6 Electrochemical methods 2-140
2.5.2.4.4.7 Chemical methods 2-140
2.5.2.4.4.8 Analysis of sulfates
and NH. 2-140
2.5.2.4.4.9 Analysis of organics.. 2-140
2.5.3 Continuous sampling methods for particulate matter 2-141
2.5.3.1 Introduction 2-141
2.5.3.2 Integrating nephelometer 2-142
2.5.3.3 Condensation nuclei counter 2-142
2.5.3.4 Particle size distribution counters 2-144
2.5.3.5 Visibility monitors 2-145
2.5.4 Summary-Measurement Techniques for Particulate Matter... 2-146
2.6 SUMMARY 2-151
2.7 REFERENCES 2-153
IX
-------
LIST OF FIGURES
FIGURE Page
2-1. Schematic of a trimodal atmospheric aerosol size distribution.... 2-6
2-2. Sulfur dioxide concentrations between 1200 and 1500 hours on
November 24, 1970 2-47
2-3. Relationship between Phillips coulometric analyzer and average
of reference method analyzers 2-54
2-4. Relationship between Meloy flame photometric analyzer and average
of reference method analyzers 2-55
2-5. The effects of spurious sulfate formation on clean filters is
given for extreme exposures 2-76
2-6. Effects of filter temperature on sulfate distribution 2-77
2-7. Sampling effectiveness of 11 1/2" x 14" Hi-Vol 2-85
2-8. Mass distribution data emphasizing large (>10 pm) particles 2-86
2-9. Influence of particle size upon effectiveness of prototype
dichotomous sampler 2-88
2-10. The size separation characteristics and wall loss measurement of
a 2.5 pm cutpoint virtual impactor as a function of particle size
dp 2-89
2-11. TSP Hi-Vol used in IP network 2-94
2-12. Effect of flow rate on performance of 11 1/2" x 14" Hi-Vol 2-96
2-13. Deposition of standby mode, 11 1/2" x 14" Hi-vol 2-98
2-14. British smoke shade sampler 2-101
2-15. Penetration of aerosol through inlet alone and through entire
flow system to filter 2-102
2-16. A.I.S.I. sampler with built-in evaluator and recorder 2-104
2-17. Cascade impactor sampling system 2-106
2-18. Cross section schematic of the CHAMP aerosol sampler 2-108
-------
2-19. Example of computer printout and corresponding size distribution
curve 2-109
2-20. Experimentally derived collection efficiency curve of the
CHAMP fractionator inlet operated 2-110
2-21. Experimental results for the calibration tests 2-112
2-22. Cyclone sampler and shelter assembly 2-113
2-23. "Los Alamos" curve for fine particulates 2-114
2-24. Assembly for sampling with a total filter and cyclone in parallel 2-115
2-25. Fraction of methylene blue particles deposited in the cyclone
as a function of the aerodynamic particle diameter 2-117
2-26. Single stage centripeter 2-118
2-27. The size separation characteristics and wall loss measurement of
a 2.5 urn cutpoint virtual impactor as a function of particle size
dp 2-119
2-28. Inlets for the dichotomous virtual impactor 2-121
2-29. Sampling effectiveness of the sierra model 244E inlet vs. particle
diameter for three wind speeds 2-122
2-30. SSI Hi-Vol used in IP network 2-123
2-31. Aerosol collection performance for the size selective inlet high
volume sampler 2-124
2-32. Effect of wind speed upon cutpoint size of SSI with domed roof... 2-125
2-33. Variation of aerosol penetration with flow rate 2-127
2-34. OECD proposed international standard calibration curve 2-135
2-35. Light scattering per unit volume of aerosol material as a function
of particle size 2-143
-------
LIST OF TABLES
TABLE Page
2-1. Physical properties of Sulfur Dioxide 2-3
2-2. Some examples of shapes of fine particles 2-9
2-3. Performance specifications for automated methods 2-35
2-4. List of designated equivalent methods (automated analyzers) 2-36
2-5. Results from a correlation analysis of the thorin method versus
the pararosani 1 ine method 2-43
2-6. Sulfur dioxide correlation coefficients and mean concentrations
for measurements made between September 11 and November 24, 1970
(1) 2-46
2-7. Sulfur Dioxide correlation coefficients and mean concentrations
for measurements made November 24, 1970, 1200 to 1355 hours (1).. 2-48
2-8. Test concentration ranges, number of measurements required and
maximum discrepancy specification 2-50
2-9. Frequency distribution of sulfur dioxide measurements 2-51
2-10. Comparison of EPA designated equivalent methods for S0?
(continuous analyzers) 2-53
2-11. Fractional aerosol penetration for selected substrates as a
function of face velocity and particle size 2-91
2-12. Sampling effectiveness of Hi-volume sample at 15 ft/s 2-95
-------
TABLE OF CONTENTS
3. CRITICAL ASSESSMENT OF PRACTICAL APPLICATIONS OF SULFUR OXIDES AND
PARTICULATE MATTER MEASUREMENT TECHNIQUES 3-1
3.1 INTRODUCTION 3-1
3. 2 HISTORICAL PERSPECTIVE 3-2
3.3 CRITICAL ASSESSMENT OF SULFUR OXIDES MEASUREMENT APPLICATIONS. 3-15
3.3.1 British Approaches 3-15
3.3.1.1 Comparisons Between the Lead Dioxide Gauge and
Hydrogen Peroxide Method - The National Survey 3-16
3.3.1.2 Daily Sulfur Dioxide Measurements of the
United Kingdom National Survey 3-18
3.3.1.3 Summary 3-26
3. 3.2 American Approaches 3-28
3.3.2.1 Sulfur Oxides Measurements of the U.S. EPA
Chess Program 3-29
3.3.2.1.1 The West-Gaeke Method for
Measurement of Ambient SO,., 3-30
3.3.2.1.2 CHESS Program Suspended SOI fate
Measurements 3-38
3.3.2.2 Summary 3-41
3.3.3 Summary of Assessment of Sulfur Oxides Measurement
Applications 3-42
3.4 CRITICAL ASSESSMENT OF PARTICULATE MATTER MEASUREMENT
APPLICATIONS 3-47
3.4.1 British Approaches 3-48
3.4.1.1 Daily Smoke Measurements of the United Kingdom
National Survey 3-48
3.4.1.1.1 Sources of Imprecision in
Measurement Physical Layout 3-49
3.4.1.1.2 Reflectance Measurements 3-53
3.4.1.1.3 Form of the Reflectance Curve 3-62
3.4.1.1.4 Averaging Error 3-62
3.4.1.1.5 Filter Handling 3-63
3.4.1.1.6 Filter Weighing Procedure 3-64
3.4.1.1.7 Filter Shipment Losses 3-65
3.4.1.1.8 Choice of Clamp Size 3-65
3.4.1.2 Summary 3-68
3.4.2 American Approaches 3-69
3.4.2.1 Hi-Volume TSP Measurements of the U.S. EPA
Chess Program 3-70
3.4.2.2 Other CHESS Program Particulate Matter
Measurements 3-73
3.4.2.3 Additional Sources of Uncertainty in Hi-Volume
TSP Measurements 3-75
3.4.3 Summary for American Hi-Volume TSP Sampling 3-76
3. 5 COMPARISONS OF PARTICULATE MATTER MEASUREMENTS 3-84
3.5.1 Theoretical Considerations 3-85
3.5.1.1 Boundary Conditions and Dependent Variable
Selection 3-85
3.5.1.2 Dimensional Analysis 3-91
3.5.2 Comparison Between TSP and Other Gravimetric
Measurements 3-92
xm
-------
3.5.3 Comparison of Gravimetric Methods (TSP) with Light
Scattering Nephelometry (B ) 3-96
3.5.4 Comparison of Hi-Vol TSP Me^Red with ASTM (CoH) Method. 3-99
3.5.5 Comparison of American Hi-Volume (TSP) and British Smoke
(BS) Measurements 3-102
3.5.5.1 Introduction 3-102
3.5.5.2 Development of a Bounded Nonlinear (BNL) Model 3-108
3.5.5.3 Physical Phenomena Underlying TSP-BS Relation-
ships 3-128
3.5.5.4 Summary and Conclusions for BS-TSP Comparisons. 3-139
3.7 REFERENCES 3-142
APPENDICES
APPENDIX A A-l
APPENDIX B B-l
APPENDIX C C-l
XIV
-------
LIST OF FIGURES
Figure Page
3-1 A comparison of lead dioxide and hydrogen peroxide methods for
sulfur dioxide showing wide variations between simultaneous
measurements. The solid line is the regression line, and the
dotted lines are the 95 percent confidence limits. From WSL
(1967) 3-17
3-2 Comparison of smoke calibration curves for Eel reflectometer,
Whatman No. 1 paper and a 1-in. diameter filter. From WSL
(1967) 3-60
3-3 Comparison of simultaneous BS measurements using 1-in. and 2-in.
filter clamps, showing that the reflectance ratio (1-in./2-in.)
decreases below approximately 45 percent reflectance of the
1-inch filter clamp. From Moulds (1962) 3-61
3-4 Surface concentration of particulate matter on filter vs
percent transmittance showing lack of fit at the origin 3-87
3-5 Surface concentration of particulate matter on filter vs percent
transmittance showing a family of related curves 3-88
3-6 CoH vs particle count showing deviation from Beers Law 3-90
3-7 CoH/Np vs Np showing evidence of a quadratic relation 3-92
3-8 Measured mass concentration and light-scattering coefficient
for atmospheric aerosols at different locations. Showing
frequent variations of more than 50% 3-98
3-9 Measurements of British Smoke vs Hi-vol TSP, showing a consistent
relation between these measures over the entire range of reported
observations. Most points shown are annual mean values; see text
for discussion 3-111
3-10 Daily gravimetric and smokeshade data for two representative
periods as measured at the London County Hall. On December 9th,
the two curves cross, and BS is greater than TSP. From Ball and
Hume (1977) 3-112
3-11 Comparison of weekly average LIB Hi-vol and smokeshade from
Alborg, Denmark, showing negligible smoke and appreciable TSP-
From Dal ager (1975) 3-114
xv
-------
Figure
3-12 Comparison of hi-volume sampler TSP vs British smoke (BS)
relationship defined by data from United States sampling sites
(AISI, 1977) in contrast to analogous comparison relationship
defined by data from England (Lee, et al., 1972). From Pashel
and Egner (1980). Note that the Lee et al. (1977) line, as
plotted here from the original Pashel and Egner (1980) paper,
is incorrect. 3-117
3-13 Pressure drop on Wahtman #1 filter vs flow rate showing effect
of use of lower flow rate than recommended by IM (1966) 3-122
3-14 Frequency distributions of British Smoke and Hi-volume sampler
data for site MOG in Bethlehem, PA, reported by Pashel and
Egner (1980) 3-124
3-15 Pashel and Egner (1980) British Smoke vs Hi-volume sampler
TSP data from sixteen locations in the USA plotted in
relation to BNLM model of Mage (1980) and showing possible
effect of filter mailing. Note that six of eight data points
(MOR, MOG, NOR, HAJ, LAB, ISS) from Pennsylvania sites fall
closer to the BNLM curve than data points indicated as being
mailed over long distances from Ohio, Colorado, and Kansas
City, Missouri 3-127
3-16 Smokeshade Hi-vol correlations for Salford, England showing
occurrences of a significant number of days with smokeshade
greater than total suspended particulates. From Lee et al.
(1972) 3-134
3-17 Correlation between duplicate reflectometric measurements showing
deviations of up to 25 ug/m . From Muylle et al. (1978) 3-136
3-18 Correlation between duplicate gravimetric measurements showing
deviations greater than 25 ug/m . From Muylle et al. (1978) 3-137
2
3A-1 1/v variation of Hi-vol flow with time showing flow pattern
predicted by Darcy's Law. Data from Waller by personal
communication (1978) A-5
3A-2 Test for decrease in flow rate proportional to mass collected
showing deviation from the expected exponential relationship A-6
3C-1 Penetration of British Smoke sampler showing fit to the Johnson SB
distribution C-5
XVI
-------
LIST OF TABLES
Table Page
3-1. Summary of Evaluation of Sources, Magnitudes, and
Directional Biases of Errors Associated with British
S0? Measurements 3-43
3-2. Summary of Evaluation of Sources, Magnitudes, and
Directional Biases of Errors Assoicated with American
S0_ Measurements 3-44
3-3. Concentrations By Direct Weighing Compared with
Concentrations of Standard Smoke - 10 urn Clamp 3-58
3-4. Summary of Evaluation of Sources, Magnitudes, and
Directional Biases of Errors Associated with British
Smoke (Particulate) Measurements 3-77
3-5. Summary of Evaluation of Sources, Magnitudes, and
Directional Biases of Errors Associated with American
Total Suspended Particulate (TSP) Measurements 3-79
3-6. Particle Counts (Np) 0.3 - 2.0 urn 3-89
3-7. Comparison of TSP Results Obtained By Different Hi-Volume
Samplers Tested in German Cities By Laskus (1977) 3-94
3-8. Effect of Wind Speed on Collection Efficiency. From
Laskus (1977) 3-95
3-9. Arithmetic Mean Concentrations of Smoke (BS Method) and
Total Suspended Particulate (HV Method) at Two Sites
In Berlin, New Hampshire 3-110
3-10. Summary of Pashel and Egner (1980) BS and TSP Data from
Representative Locations in the U.S.A 3-126
3-11. Chi-square Test Summary Table for Pashel and Egner Points
Plotted in Figure 3-15 3-128
3-12. Comparison Between Filter Paper Stains with and without
Cyclone Pre-filters Using Two-inch Diameter Holder and
Sampoing Rate of 44 Liters Per Minute 3-130
3-13. Comparison Between Filter Paper Stains With and Without
Three Stages of Case!la Cascade Impactor in Series as
Pre-filter Using One-inch Diameter Holders and Sampling
Rate of 17.5 Liters Per Minute 3-130
xvn
-------
Table Page
3-14. Comparison Between Filter Paper Stains With and Without
Four Stages of Cascade Impactor in Series as Pre-filter
Using One-Inch Diameter Holders and Flow Rate of 17.5
Liters Per Minute 3-131
3-15. Concentration of Smoke and Sulfur Dioxide on January 19/20,
1955. Compatison of Methods 3-133
xv m
-------
TABLE OF CONTENTS
Page
4. SOURCES AND EMISSIONS 4-1
4.1 INTRODUCTION 4-1
4.2 NATURAL SOURCES 4-2
4.2.1 Terrestrial Dust 4-2
4.2.2 Radioactive Aerosols 4-8
4.2.3 Sea Spray 4-10
4.2.4 Volcanic Emissions 4-14
4.2.5 Biosphere Emanations 4-16
4.2.6 Biomass Burning 4-17
4.3 MANMADE SOURCES OF SULFUR OXIDES AND PARTICULATE MATTER . . . 4-19
4.3.1 Sulfur Oxide Emissions 4-19
4.3.1.1 National and Regional Overview 4-19
4.3.1.1.2 National Sulfur Oxide Emissions
and Fuel Use: Historial Trends . . 4-19
4.3.1.1.2.1 Utilities 4-23
4.3.1.1.2.2 Industrial
facilities 4-23
4.3.1.1.2.3 Residential and
commercial establish-
ments 4-26
4.3.1.1.3 Future Trends 4-26
4.3.1.1.3.1 Utilities 4-27
4.3.1.1.3.2 Industrial
facilities 4-27
4.3.1.1.3.3 Residential and
commercial establish-
ments 4-28
4.3.1.1.4 Regional SO Emissions 4-28
4.3.1.1.4.1 Current emissions
distribution 4-28
4.3.1.1.4.2 Industrial process
sources 4-31
4.3.1.1.4.3 Future regional
emission patterns . . 4-34
4.3.1.2 SO Source Emissions Characteristics 4-34
4.5.1.2.1 Utility and Industrial Fuel
Combustion Sources 4-36
4.3.1.2.1.1 Coal 4-36
4.3.1.2.1.2 Oil 4-38
4.3.1.2.2 Residential and Commercial
Space Heating 4-40
4.3.1.2.3 Industrial Process Sources .... 4-41
4.3.1.2.3.1 Primary metals
industries 4-41
4.3.1.2.3.2 Petroleum industry . . 4-42
4.3.1.2.3.3 Chemical production
industries 4-43
xix
-------
Page
4.3.2 Participate Matter Emissions 4-44
4.3.2.1 Major Manmade Sources of Participate 4-44
4.3.2.1.1 Geographic Distribution 4-44
4.3.2.1.2 Source Identification and Attribution . 4-55
4.3.2.1.3 Particulate Matter Emission Trends . . . 4-55
4.3.2.1.4 Discussion of Major Source Categories . 4-60
4.3.2.1.4.1 Stationary fuel
combustion 4-60
4.3.2.1.4.2 Mineral products
industries 4-70
4.3.2.1.4.3 Primary metals 4-79
4.3.2.1.4.4 Industrial process
fugitive particulate
emissions (IPFPE) 4-84
4.3.2.1.4.5 Non-industrial fugitive
particulate emissions . . . 4-94
4.3.2.2 Mobile Source Particulate Matter 4-99
4.4 REFERENCES 4-111
xx
-------
CH4B11/E 3-21-80
LIST OF FIGURES
Figure Page
4-1 Median concentrations of Si, Fe, and Zn are plotted
as a function of particle size 4-6
4-2 Bursting of a bubble, 500 urn in diameter, at the sea-
air interface produces atmospheric particles 4-9
4-3 Average particle-size spectra for chloride concentration,
bromide concentration, and bromide/chloride ratio were
obtained with an Andersen impactor with a minimum Stage E
cutoff diameter of 0.5 urn 4-12
4-4 Average particle-size spectra of particulate chloride
concentration, particulate iodine concentration, and
particulate iodide/chloride ratio were obtained with an
Andersen impactor 4-13
4-5 Percentage of 1975 national sulfur oxide emissions is
shown by source category 4-21
4-6 Bar chart representing nationwide estimates of sulfur
oxide emissions are shown for 1940-1990 4-22
4-7 Bar chart representing stationary-source fossil fuel
consumption by consuming sector for 1960, 1975, and
1990 illustrates data presented in Table 4-5 4-25
4-8 Distribution of 1975 sulfur oxide emissions and
percentages of U.S. total sulfur oxide emissions are
shown for EPA regions 4-29
4-9 Sulfur oxide emission density by county 4-30
4-10 Locations are shown for existing primary metal
smelters (midyear 1973) 4-33
4-11 Characterization of U.S. population density is shown
by state 4-53
4-12 Characterization of U.S. particulate emission density
is shown by state 4-54
4-13 Bar chart depicts the trend in national particulate
emissions since 1940 4-57
4-14 Percent fractionation of stationary fuel combustion
particulate emissions is shown for states emitting
>_ 1.0 percent of U.S. mineral products total 4-61
4-15 Percentage of fractionation of mineral products
industry particulate emissions is shown for states
emitting > 1.0 percent of U.S. mineral products total . . 4-71
4-16 Map shows geographic distribution (as percent of total)
of particulate emissions from primary metal production
by state 4-72
4-17 U.S. raw steel production is characterized by process,
1950-77 4-85
4-18 Relative particulate emissions are shown for vehicle
categories (total and major components) 4-106
4-19 Total suspended particulate composition is projected,
based on best diesel estimate without particulate
regulations 4-108
xxi
-------
LIST OF TABLES
Table Page
4-1 Aerosol enrichment factors relative to Al 4^4
4-2 Comparison of aerosol composition in remote regions 4-7
4-3 Sulfur emitted by volcanoes and fumaroles 4-15
4-4 National sulfur oxides emissions inventory summary 4-20
4-5 U.S. fossil fuel consumption for stationary sources by year by
consuming sector 4-24
4-6 Percentage of EPA regional SO emissions by source category. . 4-32
4-7 Current and projected sulfur oxides emissions by EPA region. . 4-35
4-8 Comparison of calculated and inventoried emissions for the
United States, 1975 4-45
4-9a Major national sources of particulate matter (1975 NEDS
inventory) 4-46
4-9b Major national sources of particulate matter (1977 SURE
inventory) 4-46
4-10 Summary of major source categories and particle characteriza-
tion data 4-47
4-11 State-by-state listing of total particulate emissions,
population, and density factors 4-48
4-12 Trends in national particulate emissions 1940-1977 4-56
4-13 Stationary fuel combustion particulate emissions 4-62
4-14 Total fuel combustion particulate emissions: rank by state
and cumulative percent of U.S. total 4-64
4-15 Fuel combustion: percent of emissions less than 3 urn for four
source categories and by fuel type 4-67
4-16 Trace element air emissions: percent of total from conventional
stationary combustion system categories 4-68
4-17 Trace elements in solid waste (ash): percent of total generated
by conventional stationary combustion system categories. . . . 4-69
4-18 Production summary for major mineral products industries . . . 4-73
4-19 Typical composition of cement kiln dust 4-76
4-20 Typical values of trace elements in particulate emissions from
crushed and broken stone processes 4-78
4-21 Distribution of emissions among primary metal industries . . . 4-80
4-22 Production of primary metals, 1975 4-80
4-23 Industrial process fugitive particulate emissions 4-87
4-24 Toxic components of fugitive (and stack) particulate
emissions 4-90
4-25 Estimated annual suspended particulate emission rates from
open-dust sources in the United States 4-96
4-26 Parameters affecting emissions factors for open dust sources . 4-97
4-27 Summary of inhalable particulate (IP) fractions from open-dust
sources 4-100
4-28 Motor vehicle engine particulate emissions 4-102
xxi i
-------
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF TABLES
LIST OF ABBREVIATIONS AND SYMBOLS
5. ENVIRONMENTAL CONCENTRATION AND EXPOSURE 5-1
5.1 DIMENSION OF EXPOSURES TO PARTICLES
AND SULFUR DIOXIDE 5-1
5.2 AMBIENT MEASUREMENTS 5-2
5.2.1 Ambient TSP Concentrations 5-2
5.2.1.1 Composition and Range of
Concentrations 5-17
5.2.1.2 TSP Concentrations by
Location 5-25
5.2.1.2.1 Industrial-commercial-
residential 5-28
5.2.1.2.2 Intracity comparisons 5-30
5.2.1.2.3 Regional differences in
background concentrations 5-36
5.2.1.2.4 Special local problems
with TSP 5-36
5.2.1.3 Diurnal and Seasonal Variation
in TSP Concentrations 5-40
5.2.1.3.1 Diurnal 5-40
5.2.1.3.2 Weekly patterns 5-41
5.2.1.3.3 Seasonal 5-41
5.2.1.4 Trends in TSP concentrations 5-43
5.2.1.5 National Status of NAAQS
for TSP 5-53
5.2.1.6 Severity of Peak TSP
Concentrations 5-58
5.2.2 Composition of Ambient Particulate
Matter 5-71
5.2.2.1 Ambient sulfates 5-71
5.2.2.1.1 Spatial and temporal
distribution 5-72
5.2.2.1.2 Urban variations 5-85
5.2.2.2 Ambient Nitrate Aerosols 5-92
5.2.2.3 Airborne Organic Particulate
Matter 5-105
5.2.3 Ambient Sulfur Dioxide Concentrations 5-106
5.2.3.1 Sulfur Dioxide Monitoring 5-108
5.2.3.2 Distribution of Sulfur Dioxide
Concentrations 5-112
5.2.3.3 Sulfur Dioxide Concentration by Location 5-113
5.2.3.3.1 1978-Highest Annual Average
Concentrations 5-113
5.2.3.3.2 1978-Highest Daily Average
Concentrations 5-115
5.2.3.3.3 Sulfur Dioxide Concentrations
by Site 5-121
xxiii
-------
5.2.3.4 Sulfur Dioxide Cycles 5-131
5.2.3.4.1 Diurnal Patterns 5-131
5.2.3.4.2 Seasonal Patterns 5-136
5.2.3.5 National Status of Sulfur Dioxide
Concentrations 5-139
5.2.3.5.1 National View 5-139
5.2.3.5.2 Regional Distribution of Sulfur
Dioxide 5-150
5.2.3.6 Severity of Ambient Exposures 5-157
5.2.3.7 Sulfur Dioxide Concentrations and
Historic Trends 5-157
5.2.3.7.1 National Trends 5-157
5.2.3.7.2 Urban Trends 5-166
5.2.3.8 Sunwiarv e-171
-------
LIST OF FIGURES
Figure Page
5-1 Distribution of the number of observations in 1978
per valid site is for a total of 2882 sites 5-6
5-2 The 95 percent confidence intervals about the annual primary
standard for TSP is shown for various sampling frequencies 5-8
5-3 In this scattergram of maximum TSP concentration
against number of observations in 1978, the total
number of sites is 1448 5-11
5-4 In this scattergram of median TSP concentration
against elevation of monitor above grade, the total
number of sites is 1448 5-13
5-5 Distribution of mean and 90th percentile TSP
concentrations is shown for valid 1978 sites 5-19
5-6 Histogram of number of sites against concentration
shows that over one-third of the sites had-annual
mean concentrations between 40 and 60 (jg/m 5-22
5-7 Histogram of mean TSP levels by neighborhood shows
lowest levels in residential areas, higher levels in
commercial areas, and highest levels in
industrial areas 5-29
5-8 Average estimated contributions to nonurban levels in the
East, Midwest, and West are most variable for transported
secondary and continental sources 5-37
5-9 Average diurnal variation of coarse particles and
Plymouth Avenue traffic in Fall River during April
shows large contributions from suspended particles 5-38
5-10 Average diurnal variation of APM concentrations at Fall
River trailer during April is less for coarse particles
than near road 5-39
5-11 Seasonal variations in urban, suburban, and rural areas are
shown for four size ranges of particles 5-42
5-12 Monthly mean TSP concentrations are shown for the
Northern Ohio Valley Air Monitoring Headquarters,
Steubenville, OH 5-44
xxv
-------
Figure
5-13 Annual geometric mean TSP trends are shown for NASN
sites 5-46
5-14 Nationwide trends in annual mean total suspended
particulate concentrations from 1972 to 1977 are
shown for 2707 sampling sites 5-48
5-15 U.S. Environmental Protection Agency Air Quality Control
Regions, and regional trends of annual mean TSP concentrations,
are shown for Eastern States, 1972-77 5-49
5-16 U.S Environmental Protection Agency Air Quality Control
Regions, and regional trends of annual mean TSP
concentrations, are shown for Western States, 1972-77 5-50
5-17 Map of the United States shows the AQCR's violating the
primary annual TSP standard (>75 ug/m ), 1977 5-56
5-18 Maximum annual average of TSP concentrations for 1974-76 are
shown by county 5-57
5-19 Map of the United States shows the AQCR's violating the primary
24-hr TSP standard (second highest, >260 ug/m ), 1977 5-59
5-20 Shaded areas indicate AQCR's in which at least one monitoring
site in 1977 had a 90th percentile TSP concentration that exceeded
260 ug/m 5-61
5-21 Severity of TSP peak exposures is shown on the basis of the 90th
percentile concentration. 5-62
5-22 Contour maps of sulfate concentrations for 1974 are shown for:
(a) annual average; (b) winter average; (c) summer average 5-73
5-23 Intensive Sulfate Study area in Eastern Canada shows the
geometric mean of the concentration of particulate soluble
sulfate during the study period 5-75
5-24 Map of SURE regions shows locations and numbers of ground
measurement stations 5-77
5-25 Cumulative plots show the frequency of sulfate concentrations
in the SURE region on the basis of the 1974-75 historical data 5-78
5-26 Plot shows the number of grid cells in which sulfate con-
centration equaled or exceeded 10 and 20 pg/m for 24-hr
mean in the SURE region for August 1977, October 1977, and
January-February 1978 5-79
xxvi
-------
Figure Page
5-27 Map shows the spatial distribution of number of days per month
that the sulfate concentration equaled or exceeded 10 ug/m 5-81
5-28 1977 seasonal patterns of SO^ emissions and 24-hr average SO^
and SO. ambient levels in the New York area are normalized to
the annual average values 5-82
5-29 Monthly variation in monthly mean of 24-hr average sulfate
concentration at downtown Los Angeles is compared with monthly
mean 1973 Los Angeles County power plant S0_ emissions 5-83
5-30 Regional and seasonal distribution of sulfate concentrations are
shown for the Northeast quadrant of the United States 5-84
5-31 Spatial and temporal distributions of daily mean sulfate
concentrations over the Northeast quadrant of the United States
are shown for July 1974 5-86
5-32 Spatial and temporal distributions of daily mean sulfate
concentrations over the Northeast quadrant of the United States
are shown for January 1975 5-87
5-33 Map shows annual mean 24-hr average sulfate levels in micrograms
per cubic meter in the New York area, based on 1972 data from
Lynn et al. (1975) 5-89
5-34 Distribution of annual average sulfate concentration in micro-
grams per cubic meter in the greater Los Angeles area is based
on 1972-74 data 5-90
5-35 Mean annual nitrate aerosol concentrations in urban and nonurban
air were obtained by the National Air Surveillance Network 5-96
5-36 Map shows U.S. mean annual ambient nitrate levels in micrograms
per cubic meter 5-97
5-37 Mean nitrate concentrations in micrograms per cubic meter were
measured at nonurban sites by the U.S. Environmental Protection
Agency (unpublished data) 5-98
5-38 Map shows U.S. ambient emission densities in grams per square
meter per year 5-99
5-39 Histogram shows monthly mean and extremes of 24-hr particle
concentrations for August and October 1977 at three sites
in the SURE region 5-103
xxvn
-------
Relative locations for sites measuring concentrations represent
several spatial scales of measurement in an urban complex,
with respect to annual averaging times 5-107
5-41 Monthly means of hourly sulfur dioxide concentrations are
shown for Watertown, MA, for December 1978 5-132
5-42 Monthly means of hourly sulfur dioxide concentrations are shown
for Kingston (TVA site 44-1714-003, "Laddie Village") for
January 1975 and 1978 5-134
5-43 Monthly means of hourly sulfur dioxide concentrations are shown
for St. Louis (city site no. 26-4280-007, "Broadway & Hurck")
for February 1977 and 1978 5-135
5-44 Monthly means of hourly sulfur dioxide concentrations are shown
for Steubenville, OH (NOVAA site 36-6420-012) for June
1976 and 1977 5-137
5-45 Seasonal variations in sulfur dioxide levels are shown for
Steubenville, St. Louis, and Watertown 5-138
5-46 National status is shown for average annual (1977) sulfur dioxide
concentrations; less than 100, 100 to 200, 200 to 365, and
greater than 365 ug/m 5-142
5-47 Statistical characterization of 1977 national sulfur dioxide
status is shown by 90th percentile concentrations 5-143
5-48 Characterization of 1974-76 national SO, status is shown by
second highest 24-hr average concentration 5-144
5-49 Distribution of annual average sulfur dioxide concentration
in 1978 is shown for continuous and bubbler methods 5-146
5-50 Histogram shows annual average sulfur dioxide concentrations
for valid bubbler sites, 1978 5-147
5-51 Histogram shows annual average sulfur dioxide concentrations
for valid continuous sites, 1978 5-148
5-52 Cumulative frequency distribution of the 90th percentile is
shown for valid bubblers and continuous monitors 5-149
5-53 Characterization of 1977 national sulfur dioxide status is shown
by 24-hr average concentrations 5-165
5-54 Average sulfur dioxide concentrations are shown for 32 urban
NASN stations 5-167
xxviii
-------
Page
5-55 Nationwide trends in annual average sulfur dioxide concentrations
from 1972 to 1977 are shown for 1233 sampling sites 5-168
X*
-------
LIST OF TABLES
TABLE Pa9e
5-1. Valid monitoring sites operating in 1978 5-4
5-2. Probability of selecting two or more days when site exceeds
standard 5~9
5-3. Height above grade for 1978 TSP monitors 5-14
5-4. Cross-tabulation of median concentration of TSP by height of
monitor above grade 5-15
5-5 Cross-tabulation of 90th percentile concentration of TSP by
height of monitor above grade 5-16
5-6. Distribution of valid sites by 1978 annual average TSP
concentrations 5-20
5-7. Thirty sites for highest TSP annual means in 1978 5-23
5-8. Thirty monitoring sites with the highest 24-hour concentrations
in 1978 5-26
5-9. Cross-tabulation of TSP sites by median concentration of site
descriptor 5-31
5-10. Cross-tabulation of TSP sites by 90th percentile values and site
descriptor 5-32
5-11. Cross-tabulation of TSP median value by site descriptor 5-33
5-12. Cross-tabulation of TSP 90th percentile values by site descriptor 5-34
5-13. Range of annual geometric mean concentrations in areas with high
TSP concentrations in 1977 5-34
5-14. Regional summaries of TSP values from valid monitors 5-51
5-15. National summary of total stations reporting data and number
reporting violations of air quality standards for TSP, 1977 and
1978 5-54
5-16. 1977 nationwide compliance status for total suspended particulate
matter and severity of TSP exposures, based on the 90th
percentile concentration of 24-hour values 5-63
5-17. Some characteristics of pollution in the New York and Los Angeles
areas 5-91
xxx
-------
Table Page
5-18. Primary ranking of variables for correlating airborne sulfate
in two cities based on a stepwise linear regression of 15
variables from CHAMP and related monitoring stations 5-93
5-19. Aerosol characterization experiment data for nitrate and sulfate
in the California south coast basin 1972-73 5-101
5-20. NASN network data for nitrate and sulfate in the California
south coast basin for 1968 5-102
5-21. Total monitors reporting SCL, by method, 1976-1977 5-110
5-22. Number of bubbler and continuous S02 monitoring sites in
operation in 1978 and valid for 1978 annual means 5-111
5-23. Statistics on the valid S02 monitoring sites for 1978 5-114
5-24. SOp monitoring sites with the highest 11 annual mean concen-
trations in 1978 (valid continuous sites only) 5-116
5-25. The 45 monitoring sites with the highest 24 hour SO- concen-
trations in 1978 (continuous methods) 5-117
5-26. Cross-tabulation of annual mean S02 concentration by method
(bubbler or continuous) for population-oriented and for
source-oriented center city sites 5-122
5-27. Cross-tabulation of annual mean SO- concentration by method
(bubbler or continuous) for population-oriented and for
source-oriented suburban sites 5--125
5-28. Cross-tabulation of 90th percent!le of daily average concen-
trations of S0? by method (bubbler or continuous) for
source-orientea sites in center city, suburban, rural,
and remote locations 5-127
5-29. Cross-tabulation of 90th percent!le of daily average concen-
trations of SO- by method (bubbler or continuous) for popula-
tion-oriented monitors center city, suburban, and rural
locations 5-129
5-30. National summary for sulfur dioxide total stations reporting
data, number and percentage of sites reporting concentrations
above the national ambient air quality standards 1977-1978 5-140
5-31. Bubbler S02 monitor results by region 5-151
5-32. Continuous S02 monitor results by region 5-153
5-33. Comparison of bubbler and continuous S02 monitor results by
region 5-156
xx xi
-------
Table Page
5-34 Ninetieth-percent!le sulfur dioxide levels, 1977 5-158
5-35 Comparison of frequency distribution of SO,, concentration (ppm)
duri ng 1962-67 and duri ng 1977 5-170
xxxn
-------
CONTENTS
TRANSMISSION THROUGH THE ATMOSPHERE 6-1
6.1 INTRODUCTION 6-1
6.2 TRANSFORMATIONS 6-4
6.2.1 Transformation of S02 6-4
6.2.1.1 Atmospheric investigations 6-5
6.2.1.2 Laboratory investigation and modeling 6-9
6.2.1.2.1 Smog chamber studies 6-9
6.2.1.2.2 Plume simulation with a chemical
kinetics model 6-10
6.2.1.3 Homogeneous gas-phase reactions 6-14
6.2.1.4 Heterogeneous aqueous reactions 6-16
6.2.1.5 Heterogeneous surface reactions 6-21
6.2.2 Physical transformations 6-24
6.2.2.1 Aerosol size distribution 6-24
6.2.2.2 Aerosol transformation 6-32
6.2.3 Water uptake and release 6-36
6.2.4 Carbon-containing aerosols 6-42
6.2.5 Particulate nitrate 6-46
6.3 WET AND DRY REMOVAL OF S02 AND AEROSOLS 6-47
6.3.1 Dry deposition 6-48
6.3.2 Wet removal 6-54
6.4 DISPERSION IN THE ATMOSPHERE 6-55
6.4.1 Spatial and temporal scales 6-57
6.4.2 Planetary boundary layer 6-62
6.4.3 Horizontal transport 6-68
6.5 SOURCE-RECEPTOR RELATIONSHIPS 6-70
6.5.1 Aerosol chemical composition 6-71
6.5.2 Historical analysis of trends in emission
concentrations 6-80
6.5.3 Pollution roses and sector analysis 6-80
6.5.4 Direct measurements of source impacts: plumes and
regional hazes 6-86
6.5.4.1 Power plant plume studies 6-86
6.5.4.2 Urban plumes 6-87
6.5.4.3 Regional-scale episodes of haziness 6-93
6.5.5 Diagnostic models 6-104
6.6 SUMMARY 6-110
6.7 REFERENCES 6-112
xxx i-ii
-------
LIST OF FIGURES
Figure
6-1. The daytime conversion rate in the MISTT study was quite
variable, ranging from 1 to 4 percent per hour, whereas night-
time values were consistently below 0.5 percent per hour 6-6
6-2. The ratio of particulate sulfur, Sp, to total sulfur concen-
trations (particulate S in sulfate plus gaseous S in S02, stat
average for traverse) 6-7
6-3. The initial NMHC and NO concentrations affect (A) the maximum
rates of S02 oxidation §nd (B) the total conversion of S02 to
sulfate aerosol in a 6-hr, irradiation interval. (C) the
maximum rate of S02 oxidation is plotted as a function of
initial NMHC/NOx ratio. . 6-11
6-4. The rate of oxidation of S02 is simulated by kinetic modeling . . 6-12
6-5. The theoretical rate of reaction (percent per hour) of various
free-radical species on S02 is shown for a simulated sunlight-
irradiated (solar zenith angle of 40°) polluted atmosphere. . . . 6-15
6-6. Schematic of an atmospheric aerosol size distribution showing
the three modes 6-26
6-7. Average urban model aerosol distribution plotted in five
different ways 6-27
6-8. (A) Condensable molecules (monomer) produced by gas-phase
reactions may either produce new particles or condense on
larger pre-existing particles. (B) Three time domains for
aerosol formation in an irradiated chamber 6-33
6-9. (A) Volumetric conversion rate is shown as a function of
initial S02 concentration for an ultraviolet light intensity
of 30.5 W/m2 and an NMHC concentration of 1.57 ppm 6-35
6-10. Solute mole fraction is shown as a function of water vapor
pressure (as a fraction of saturation) for three sulfate com-
pounds 6-37
6-11. Theoretical growth curves based on vapor pressure data are
shown for three sulfate aerosols exposed to increasing humidity . 6-39
6-12. Humidograph data for pure, laboratory-generated sulfate
compounds show deliquescence steps for some compounds 6-40
6-13. (A) Diurnal variation of hexanedioic acid, a presumed secondary
component. (B) Comparison of the diurnal variation of total
amides (a presumed secondary component) with ozone 6-44
xxxiv
-------
Page
6-14. Data of Sehmel and Hodgson (1976) for dry deposition of
particles on various solid surfaces was correlated by SI inn
et al. (1979) 6-53
6-15. The spatial and temporal scales relevant for trophospheric air
pollutants range from meters to hundreds of kilometers and from
seconds to years 6-58
6-16. This flow diagram of sulfur transmission through the atmosphere
shows that over half of the S02 is removed or transformed to
sulfate within the first day of its atmospheric residence .... 6-61
6-17. The wind data obtained from hourly releases by the RAPS balloon
sounding network are averaged here for July 1976 6-63
6-18. This nomogram permits estimation of the depth of the boundary
layer in the absence of marked advective effects or basic
changes in weather conditions 6-64
6-19. Sulfates evidently contribute only about 25 percent of the total
mass but cause about half of the light scattering 6-73
6-20. Compositions of fine and coarse particles collected in
Charleston, West Virginia 6-74
6-21. Source resolution of the St. Louis aerosol shows that less than
10 percent of the fine-particle mass is due to primary automo-
bile emissions 6-76
6-22. In the St. Louis area, monitoring data for about a 2-month
period shows that sulfur was about 10 to 12 percent (calculated
as elemental sulfur) of the fine-particle mass 6-78
6-23. Composition of fine and coarse particles collected during
flights in the four corners region is presented in pie diagrams . 6-79
6-24. This plot of the percentage of daylight observations with RH<60
percent for which visual range was >121 km as a function
of wind 6-82
6-25. Air mass trajectories to Whiteface, Schoharie and Holland on
July 10 and 22, 1976, indicate that elevated sulfate at non-
urban New York sites originated in the Ohio River Valley 6-84
6-26. Sulfate trajectory roses in Europe reveal that most of the
sulfate in southern Scandinavia is from emissions in other
countries lying to the south 6-85
6-27. The amount of particulate sulfur formed increases when the
plume is removed from the surface by dilution or by decoupling
from the surface layer 6-88
xxxv
-------
Figure Page
6-28. Horizontal profiles of S02 were made during selected constant-
altitude traverses on July 9 and July 18, 1976 6-89
6-29. Vertical profiles show the cross wind integrals of excess plume
concentrations of gaseous sulfur (Sg), particulate sulfur (Sp),
and light-scattering coefficient (B t) for the transport of
the Labadie plume of July 9 and 18, 1976 6-90
6-30. The three-dimensional flow of aerosols and ozone in the St.
Louis plume is plotted for a distance of 240 km 6-92
6-31. Sequential contour maps of noon visibility for the period
June 25 to July 5, 1975; illustrate the evolution and transport
of a large-scale hazy air mass 6-97
6-32. On June 30, 1975, the hazy air mass covered large parts of
Arkansas, Missouri, Kansas, Iowa, and Minnesota 6-98
6-33. Local monitoring data in the St. Louis, MO, area during the
June-July 1975 haziness episode 6-99
6-34. Noon extinction coefficient and daily mean sulfate concentrations
are compared for June 23 and July 5, 1975 6-101
6-35. Isopleths of total number of forecast days of high meteorological
potential for air pollution in a 5-year period 6-103
6-36. Maps of dry and wet deposition of sulfur compounds from the
Federal Republic of Germany 6-107
6-37. Extracted conversion rates of S02 to sulfate in the St. Louis
urban plume show a peak at about 5 percent per hour at 2 p.m. . . 6-108
xxxvi
-------
LIST OF TABLES
Table Page
6-1 Rate expressions for the Mn-, Fe-, Cu-, and V-Catalyzed
oxidation of dissolved S02 in water 6-18
6-2 Classification of major chemical species associated with
atmospheric aerosols 6-31
6-3 S02 deposition velocities over vegetation 6-50
6-4 S02 deposition velocities over soil 6-51
6-5 Representative annual average sulfur wet and dry deposition
rates 6-56
6-6 Three-year running average of sulfate by geographical region
for nonurban sites in the United States 6-95
6-7 Values applied in calculations with the lagrangian dispersion
model in the OECD project 6-105
xxxvn
-------
CONTRIBUTORS AND REVIEWERS
Mr. John Acquavella
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Roy E. Albert
Institute of Environmental Medicine
New York University Medical Center
New York, New York 10016
Dr. Martin Alexander
Department of Agronomy
Cornell University
Ithaca, New York 14850
Dr. A. P. Altshuller
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. David S. Anthony
Department of Botany
University of Florida
Gainesville, Florida 32611
Mr. John D. Bachmann
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Allen C. Basala
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Neil Berg
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Michael A. Berry
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xxxviii
-------
Mr. Francis M. Black
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Joseph Blair
Environmental Division
U. S. Department of Energy
Washington, D.C. 20545
Dr. Edward Bobalek
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Ms. F. Vandiver P. Bradow
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Ronald L. Bradow
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Bruce
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Angelo Capparella
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Chapman
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert J. Charlson
Department of Environmental Medicine
University of Washington
Seattle, Washington 98195
Dr. Peter Coffey
New York State Department of Environmental Conservation
Division of Air Resources
Albany, New York 12233
Mr. Chatten Cowherd
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
xxx ix
-------
Dr. Ellis B. Cowling
School of Forest Resources
North Carolina State University
Raleigh, North Carolina 27650
Mr. William M. Cox
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. T. Timothy Crocker
Department of Community and Environmental Medicine
Irvine, California 92664
Mr. Stanley T. Cuffe
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas C. Curran
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Michael Davis
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Gerrold A. Demarrais
National Oceanic and Atmospheric Administration
U. S. Department Of Commerce
Dr. Jerrold L. Dodd
Natural Resources Ecology Laboratory
Colorado State University
Fort Collins, Colorado 80523
Dr. Thomas G. Dzubay
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas G. Ellestad
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John Evans
School of Public Health
Harvard University
Boston, Massachusetts 02115
xl
-------
Dr. Lance Evans
Department of Energy and Environment
Brookhaven National Laboratory
Upton, New York 11973
Mr. Douglas Fennel!
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Benjamin G. Ferris, Jr.
School of Public Health
Harvard University
Boston, Massachusetts 02115
Mr. Patrick Festa
New York Department of Environmental Conservation
Division of Fish and Wildlife
Albany, New York 12233
Mr. Terrence Fitz-Simmons
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Christopher R. Fortune
Northrop Services, Inc.-Environmental Sciences
P. 0. Box 12313
Research Triangle Park, North Carolina 27709
Dr. Robert Frank
Department of Environmental Health
University of Washington
Seattle, Washington 98195
Dr. Warren Galke
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Phil Galvin
New York Department of Environmental Conservation
Division of Air Resources
Albany, New York 12233
Dr. Donald Gardner
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. J.H.B. Garner
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xli
-------
Dr. Donald Gillette
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Judy Graham
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Lester D. Grant
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Armin Gropp
Department of Chemistry
University of Miami
Miami, Florida 33124
Dr. Jack Hackney
Rancho Los Amigos Hospital
Downey, California 90242
Mr. Bertil Hagerhall
Ministry of Agriculture
Pack
S-163 20 Stockholm
Sweden
Dr. Douglas Hammer
2910 Wycliff Road
Raleigh, North Carolina 27607
Mr. R. P. Hangebrauck
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas A. Hartlage
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Victor Hasselblad
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas R. Hauser
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xlii
-------
Dr. Carl Hayes
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Fred H. Haynie
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Walter Heck
Department of Botany
North Carolina State University
Raleigh, North Carolina 27650
Dr. Howard Heggestad
USDA-SAE
The Plant Stress Laboratory
Plant Physiology Institute
Beltsville, Maryland 20705
Dr. George R. Hendrey
Department of Energy and Environment
Brookhaven National Laboratory
Upton, New York 11973
Dr. Ian Higgins
Department of Epidemiology
School of Public Health
University of Michigan
Ann Arbor, Michigan 48109
Mrs. Patricia Hodgson
Editorial Associates
Chapel Hill, North Carolina 27514
Mr. George C. Holzworth
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Horton
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Steven M. Horvath
Institute of Environmental Stress
University of California
Santa Barbara, California 93106
xliii
-------
Dr. F- Gordon Hueter
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Janja Husar
CAPITA
Washington University
St. Louis, Missouri 63130
Dr. Rudolf Husar
Department of Mechanical Engineering
Washington University
St. Louis, Missouri 63130
\
Dr. William T. Ingram
Consulting Engineer
7 North Drive
Whitestone, New York 11357
Dr. Patricia M. Irving
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
Dr. Jay Jacobson
Boyce Thompson Institute
Cornell University
Ithaca, New York 14850
Mr- James Kawecki
Biospherics, Inc.
4928 Wyaconda Road
Rockville, Maryland 20852
Dr. Sagar V. Krupa
Department of Plant Pathology
University of Minnesota
St. Paul, Minnesota 55108
Dr. Edmund J. LaVoie
Section of Metabolic Biochemistry
American Health Foundation
Dana Road
Valhalla, New York 10592
Dr. Michael D. Lebowitz
Arizona Health Sciences Center
1501 North Campbell
Tucson, Arizona 85724
xliv
-------
Dr. Robert E. Lee
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Allan H. Legge
Environmental Science Center
University of Calgary
Calgary, Alberta, Canada T2N 1N4
Ms. Peggy Le Sueur
Atmospheric Environment Service
Downsview, Ontario, Canada M3H5T4
Dr. Morton Lippmann
Institute of Environmental Medicine
New York University
New York, New York 10016
Dr. James P. Lodge
385 Broadway
Boulder, Colorado 80903
Dr. Gory J. Love
Institute of Environmental Studies
University of North Carolina
Chapel Hill, North Carolina 27514
Dr. David T. Mage
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North(Carolina 27711
Dr. Delbert McCune
Boyce Thompson Institute
Cornell University
Ithaca, New York 14850
Mr. Frank F. McElroy
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. David J. McKee
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas McMullen
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xl v
-------
Dr. Daniel B. Menzel
Department of Pharmacology
Duke University Medical Center
Durham, North Carolina 27710
Dr. Edwin L. Meyer
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Fred Mi 1ler
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John 0. Milliken
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Jarvis Moyers
Department of Chemistry
University of Arizona
Tucson, Arizona 85721
Dr. Thaddeus J. Murawski
New York State Department of Health
Empire State Plaza
Albany New York 12337
Dr. David S. Natusch
Department of Chemistry
Colorado State University
Fort Collins, Colorado 80523
Dr. Stephen A. Nielsen
Environmental Affairs
Joyce Environmental Consultants
414 Live Oak Boulevard
Casselberry, Florida 32707
Dr. Kenneth Noll
Department of Environmental Engineering
Illinois Institute of Technology
Chicago, Illinois 60616
Mr. John R. O'Connor
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xlvi
-------
Mr. Thompson G. Pace
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Jean Parker
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Nancy Pate
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas W. Peterson
Department of Chemical Engineering
University of Arizona
Tucson, Arizona 85721
Mr. Martin Pfeiffer
New York State Department of Environmental Conservation
Bureau of Fisheries
Raybrook, New York 12977
Dr. Marlene Phillips
Atmospheric Chemistry Division
Environment Canada
Downsview, Ontario, Canada M3H5T4
t
Dr. Charles Powers
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Mr. Larry J. Purdue
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. John C. Puzak
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xlvii
-------
Dr. Otto Raabe
Radiobiology Laboratory
University of California
Davis, California 95616
Mr. Danny Rambo
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Mr. Kenneth A. Rehme
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Elmer Robinson
Department of Chemical Engineering
Washington State University
Pullman, Washington 99163
Mr. Charles E. Rodes
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Douglas R. Roeck
GCA Corporation
Technology Division
Burlington Road
Bedford, Massachusetts 01730
Mr. J. C. Romanovsky
Environmental Sciences Research Laboratory
U.S. Environmental Protection'Agency
Research Triangle Park, North Carolina 27711
Dr. August Rossano
University of Washington
Seattle, Washington 98195
Mr. Joseph D. Sableski
Control Programs Development Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Dallas Safriet
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xlviii
-------
Dr. Victor S. Salvin
University of North Carolina at Greensboro
Greensboro, North Carolina 27408
Dr. Shahbeg Sandhu
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Joseph P. Santodonato
Life and Material Sciences Division
Syracuse Research Corporation
Merrill Lane
Syracuse, New York 13210
Dr. Herbert Schimmel
Neurology Department
Albert Einstein Medical College
26 Usonia Road
Pleasantville, New York 10570
Dr. Carl L. Schofield
Department of Natural Resources
Cornell University
Ithaca, New York 14850
Dr. David Shriner
Environmental Sciences Division
Oak Ridge National Laboratory
Ms. Donna Sivulka
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John M. Skelly
Department of Plant Pathology and Physiology
Virginia Polytechnic Institute
Blacksburg, Virginia 24061
Mr. Scott Smith
Biospherics, Inc.
4928 Wyaconda Road
Rockviell, Maryland 20852
Ms. Elaine Smolko
Department of Pharmacology
Duke University Medical Center
Durham, North Carolina
xlix
-------
Dr. Frank Speizer
School of Public Health
Harvard University
Boston, Kassachusetts 02115
Dr. John 0. Spengler
School of Public Health
Harvard University
Boston, Massachusetts 02115
Mr. Robert K. Stevens
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. George E. Taylor, Jr.
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
Dr. Larry Thibodeau
School of Public Health
Harvard University
Boston, Massachusetts 02115
Dr. W. Gene Tucker
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. D. Bruce Turner
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. James B. Upham
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Waller
Toxicology Unit
St. Bartholomew's Hospital
London, England
Mr. Stanley Wall in
Warren Spring Laboratory
Department of Industry
Stevenage, Hertfordshire SGI 2BX
England
-------
Dr. Joseph F. Walling
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. James Ware
School of Public Health
Harvard University
Boston, Massachusetts 02115
Dr. David Weber
Office of Air, Land, and Water Use
U.S. Environmental Protection Agency
Washington, D. C. 20460
Dr. Jean Weister
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. R. Murray Wells
Radian Corporation
8500 Shaol Creek Boulevard
Austin, Texas 78766
Dr. Kenneth T. Whitby
Mechanical Engineering Department
University of Minnesota
Minneapolis, Minnesota 55455
Dr. Warren White
CAPITA
Washington, University
St. Louis, Missouri 63130
Dr. Raymond Wilhour
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Dr. William E. Wilson
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John W. Winchester
Department of Oceanography
Florida State University
Tallahassee, Florida 32306
li
-------
Mr. Larry Zaragoza
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. William H. Zoller
Chemistry Department
University of Maryland
College Park, Maryland 20742
We wish to thank everyone who contributed their efforts to the preparation of
this document, including the following staff members of the Environmental
Criteria and Assessment Office, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina:
Mrs. Dela Bates
Ms. Hope Brown
Ms. Diane Chappell
Ms. Deborah Doerr
Ms. Mary El ing
Ms. Bettie Haley
Mr. Allen Hoyt
Ms. Susan Nobs
Ms. Evelynne Rash
Ms. Connie van Oosten
Ms. Donna Wicker
The final draft of this document will cite the many persons outside of the
Environmental Criteria and Assessment Office who have assisted in its pre-
paration.
lii
-------
CHAPTER 2. PHYSICAL AND CHEMICAL PROPERTIES OF SULFUR OXIDES AND
PARTICULATE MATTER AND ANALYTICAL TECHNIQUES FOR THEIR
MEASUREMENT
2.1 INTRODUCTION
The 1970 Air Quality Criteria documents for Sulfur Oxides and
Particulate Matter provided a reasonably thorough review of measurement techniques
at the date of their publication. Since that time, considerable new information
and advancements in the measurement technology for these pollutants has resulted
in a substantial amount of new information. This chapter provides a review of
both the earlier techniques, which are still important in evaluating earlier
monitoring and effects data, as well as newer methods and technology.
Present measurements of ambient sulfur oxides and particulate matter are
oriented toward a variety of objectives, the most salient of which are:
Trend Monitoring to Support Epidemilogical Studies,
Compliance monitoring to determine conformance to air quality
standards,
0 Long-term trend monitoring to determine effectiveness of control
strategies, and
Various research-related studies.
Research activities include such diverse subjects as studies of sources of
emissions; methodology development; transport and transformation; influence
on visibility and local or regional climate; effects on human health and
vegetation damage, soiling and corrosion of materials.
Compliance and long-term trend monitoring requires well defined and
2-1
-------
characterized methods for mass concentrations having sensitivities and averaging
times compatible with the form and level of existing regulatory standards.
Monitoring for research purposes involves broader areas of concern and
entails more complex, less well defined requirements. In addition to mass
concentration, particulate measurement techniques frequently must include
fractionation of the particulate matter into particle size ranges and the
particles subjected to chemical analysis to determine their composition.
Measurement technology for sulfur oxides and particulate matter is complex
and highly technical. Descriptions of the various methods in this chapter are
necessarily very brief, leaving the bulk of technical detail to the numerous
references cited. Emphasis is given primarily to those methods that are
pertinent to the discussions in Chapter 14 and to the most commonly used
methodology that figures prominently in current or imminent monitoring for
compliance and research measurement purposes.
2.2 PHYSICAL AND CHEMICAL PROPERTIES OF SULFUR OXIDES AND PARTICULATE MATTER
2.2.1 Sulfur Oxides
Sulfur dioxide (S02) is a colorless gas that can be detected by taste at
concentrations of 1000 to 3000 ug/m (0.38 to 1.15 ppm). At higher concentra-
3
tions (above about 10,000 ug/m or 3.8 ppm), it has a pungent, irritating
odor. It has an absorption maximum in the ultraviolet region of the spectrum
at around 254 nanometers (nm). The physical properties of S0? are summarized
in Table 2-1.
Chemically, SOp acts both as an oxidizing agent and as a reducing agent.
The reducing properties are more important, particularly those involving
aqueous solutions of SO^. The gas dissolves readily in water or alkaline
2-2
-------
TABLE 2-1. PHYSICAL PROPERTIES OF SULFUR DIOXIDE
Molecular weight 64.06
Density (gas), g/liter 2.927 at 0°C; 1 atm
Specific gravity (liquid) 1.434 at -10°C
Molecular volume (liquid), ml 44.7
Melting point, °C -75.46
Boiling point, °C -10.02
Critical temperature, °C 157.2
Critical pressure, atm 77.7
Heat of fusion, Kcal/mole 1.769
Heat of vaporization, Kcal/mole 5.96
Dielectric constant (practical units) 13.8 at 14.5°C
Viscosity, dyne sec/cm 0.0039 at 0°C
Molecular boiling point constant, °C/1000 g 1.45
Dipole moment, Debye units 1.61
2-3
-------
solution to form sulfurous acid (H2S03). In solution there is a series of
equilibria between S02 and H20, H2S03 and its ions:
HUSO, 5 H+ + HSO ~
23 3
HSO " 5 H+ + SO"
O w
An aqueous solution of S0? has the properties of a weak acid and is slowly
oxidized to sulfuric acid (HLSO.) by dissolved oxygen (02). Oxidation to
HLSO, is more rapid in the presence of catalyzing impurities such as manganese
and iron ions. Sulfur dioxide also reacts in the gas phase with oxygen or
other air pollutants to form sulfur trioxide (SO.,), HUSO-, and metal and
organic sulfates.
Sulfur trioxide (S03) is a highly reactive gas derived either directly
from combustion sources or from oxidation of atmospheric S0?. In the presence
of moisture in the air, it rapidly hydrates to H2S04 aerosol. In the atmosphere,
therefore, it occurs as HUSO, aerosol rather than as S0~.
Sulfuric acid (H2S04) is a strong, corrosive acid with a high boiling
point. When heated, it initially dissociates into SO., and HUO. Further heating
produces a constant boiling acid consisting of 98 percent HUSO.. It is commonly
formed in the atmosphere from hydration of SO., and oxidation of hydrogen sulfide
(H£S).
Sulfuric acid readily reacts with many metals and metal oxides to form
metal sulfates. Common sulfate species in the atmosphere include solid particulate
metal salts and liquid droplets containing either dissolved metal sulfates or
2-4
-------
H^SCL. Because HpSO. and many of its salts are strongly hygroscopic, droplets
of this type can readily take up moisture from the air until reaching equili-
brium. If ammonia is present, it will react with H^SO. to form ammonium
sulfate [(NH.)2S04] or ammonium bisulfate (NH.HSO.) which will continue to
exist as aerosols (in droplet or crystalline form, depending on the relative
humidity).
2.2.2 Particulate Matter
Atmospheric particulate matter is often very complex. Individual particles
can be composed of different chemical species, can be homogeneous or hetero-
geneous in structure, and may vary in size and shape. Particles dispersed in
a gas can be characterized by many different physical and chemical properties.
The single most important characteristic of airborne particulate matter is
particle size.
The basic definitions of particle size are related to the velocity at
which particles fall as a result of their aerodynamic behavior in the earth's
gravitational field. The aerodynamic equivalent diameter of a particle does
not refer to an actual measurement of the particle but is the size of a spherical
particle of unit density (1 g/cm ) which falls at the same speed. The Stokes
equivalent diameter, in a similar fashion, refers to the physical diameter of
a spherical particle of the same average density and the same falling velocity.
The aerodynamic diameter is the most useful definition relative to particle
sampling and deposition.
Recent studies of atmospheric particles have been conducted which have
led to important advances in the understanding of their size distribution
(Whitby, 1975; Wilson et al., 1976; Will eke and Whitby, 1975). A schematic
2-5
-------
CHEMICAL CONVERSION
OF GASES TO LOW
VOLATILITY VAPORS
CONDENSATION
LOW-
VOLATILITY
VAPOR
PRIMARY PARTICLES
• •• • • 9
HOMOGENEOUS
NUCLEATION
COAGULATION
CONDENSATION
GROWTH
OF NUCLEI
CHAIN AGGREGATES
COAGULATION
COAGULATION
WINDBLOWN DUST
+
EMISSIONS
+
SEA SPRAY
+
VOLCANOS
+
PLANT PARTICLES
RAINOUT
AND
WASHOUT
0.002 0.01
0.1 1 2
PARTICLE DIAMETER, pirn
10
100
.TRANSIENT NUCLEI OR .
AtTKEN NUCLEI RANGE
.ACCUMULATION.
RANGE
•FINE PARTICLES •
^_MECHANICALLY GENERATED.
AEROSOL RANGE
-COARSE PARTICLES"
Figure 2-1. Schematic shows an atmospheric aerosol surface area distribution showing the three
modes, main source of mass for each mode, the principal processes involved in inserting mass into each
mode, and the principal removal mechanisms.
Source: K.T. Whitby (1977)
2-6
-------
diagram of the size distribution of a typical atmospheric particles as described
by Whitby (1977) is shown in Figure 2-1. The distribution is trimodal in that
three distinct peaks are apparent. The figure shows the three principal
modes, the main source of mass within each mode, and the principal processes
involved in inserting and removing mass from each mode. Particles in the
Aitken or nuclei mode,from 0.005 to 0.05 pm diameter, are formed by conden-
sation of vapors produced by processes such as fossil fuel combustion. Par-
ticles in the accumulation mode, from 0.05 to 2 pm, are formed by coagulation
or agglomeration of nuclei mode particles and by growth of particles in the
nuclei mode through vapor condensation. Accumulation mode particles normally
do not continue to grow into the coarse mode. Typically, 80 percent or more
of the atmospheric sulfate particulate matter is found in the accumulation
mode. The upper size limit of fine particles is in the range of 1 to 3 pm and
is commonly stated as being 2.0 pm. Coarse particles, those in the > 2.0 pm
aerosol range, include reintrained surface dust and particles formed by pro-
cesses such as grinding and by the evaporation of liquid droplets.
Fine particles often serve as condensation nuclei, are associated mechani-
cally with liquid particles of rain, mist, or snow, and are often classified
into two major sources. Primary particles are a direct result of the dis-
charge of fine particles from industrial and other man-made or natural sources;
secondary particles are formed by chemical and physical reactions in the
atmosphere. Most of the reactants that form secondary particles are emitted
to the atmosphere as gaseous pollutants.
The actual density of ambient aerosols is difficult to determine precisely.
The use of the density reported for bulk chemicals can cause significant
problems in the interpretation of data. The density derived from chemical
2-7
-------
composition cannot be used in the relations between aerodynamic diameter and
volume or between aerodynamic diameter and area without challenge. Many fine
particles are flocculi and agglomerates having densities well below the "true
density" based on bulk chemical composition. Often the density of flocculent
aerosols is 10 to 30 times lower than the density of the substances of origin
(Hesketh, 1974).
Water is often exempted from the definition of particulate matter although
moisture can be combined with a particle chemically, physicochemically and
mechanically (Falkowski, 1951). From the aerodynamic point of view, the
manner in which water is combined with a particle is not important. However,
the effect of water vapor on the properties of fine particulate matter is an
important factor. Hygroscopic particles can grow in size when exposed to
relative humidity above 70 percent, and in humid weather appreciably contribute
to visibility degradation.
Morphology relates to particle shape and surface features. The concept
of the aerodynamic diameter does not accommodate particle shape or morphological
features such as porosity and roughness. Some examples of fine particles and
their shapes are shown in Table 2-2. Particles larger than 0.5 urn can be
examined with a conventional light microscope; particles in the size range
0.001 to 0.5 urn must be examined with an electron microscope. The shape of a
particle will determine the ratio between its surface area and its volume.
The absorption surface comprises the total particle surface, including the
surface of pores to which there is an external access. The area of the absorp-
tion surface determines the ability of the particle to adsorb gases or electric
charges.
The electric charge of a particle depends on the state of its surface and
2-8
-------
Shape
TABLE 2-2. SOME EXAMPLES OF SHAPES OF FINE PARTICLES
Nature of aerosols (fine particles)
Spherical
Rectangular
(irregular polyhedral)
Celliform
Fiber
Splinter
Flocculent
Platelet
Rodlet
carbon black, iron oxide, plastics,
glass, pollen
iron, quartz, mineral ash
epidermal minerals
textiles, asbestos, plants (cellulose)
cement, organic dyes
coal fumes
mica, graphite, bronze
talc, fluorine
2-9
-------
on the value of its dielectric constant and/or its size (Andrezejewski, 1968).
The electrical properties of atmospheric particles can affect the collection
capability of ambient samples. In principle, large particles greater than 3
pm carry a negative charge and fine particles less than 0.01 [jm carry a posi-
tive charge. The electric charges of particles can also affect the coagu-
lation rate and the speed of the dry deposition of particles (Brock and Marlow,
1975).
Another aspect of the physical properties of fine particles is their
ability to scatter or absorb light. The absorption of light will generally
increase with particle size. However, the ability of a particle to scatter
light will vary in a complex way with particle size. Particle scattering will
usually show a maximum when the particle diameter is approximately equal to
the wavelength of light. The scattering and absorption by individual particles
and the number of such particles per unit of gas volume can be related to bulk
optical properties, i.e., the bulk extinction coefficient and the scattering
cross-section.
Additional information concerning the physical and chemical properties of
particulate matter can be found in Aerosol Measurement (Lundgren et al., ed.
1979), Aerosol Science (Davies, ed. , 1966), Smoke, Dust, and Haze (Friedlander,
1977), Fine Particles, Aerosol Generation, Measurement and Analysis (Liu,
1976), Handbook mi Aerosols (Dennis/1976), and Airborne Particles (National
Academy of Sciences, 1979).
2.3 MEASUREMENT TECHNIQUES FOR SULFUR DIOXIDE
2.3.1 Introduction
This section discusses the methods used to measure atmospheric concen-
trations of sulfur dioxide. Each method is either integrated or continuous.
2-10
-------
The measurement of SCL by an integrated method represents the concentration
averaged over a specified time period. With a continuous method, the measure-
ment represents the instantaneous (or very nearly instantaneous) concentration
of SCL in the air sample. Most integrated methods consist of separate procedures
for sample collection and analytical measurement. In continuous methods,
sample collection and analysis are performed constantly and automatically, and
such methods are generally referred to as continuous analyzers.
This section will focus on a brief description of each method with emphasis
on measurement principle and method characteristics such as range, sensitivity,
and interferences. Sample collection and method calibration will be discussed
for integrated and for continuous methods. Among the integrated methods,
those that have experienced the most widespread use are discussed first. The
sulfation methods are presented last only because they measure "sulfation
rate" rather than S0? concentration per se. Among the continuous methods, the
discussion tends to follow a more or less chronological order with the earlier
analyzers described first.
Included in this section is a discussion of various continuous analyzers
designated by EPA as equivalent methods for the measurement of SO^ in the
atmosphere to determine compliance with the National Ambient Air Quality
Standards (NAAQS). Information from various comparison studies is also pre-
sented.
2.3.2 Integrated Methods
2.3.2.1 Sample Collection—Early research focused on methods to estimate
sulfur dioxide after it had been collected from a given volume of air. The
following critique of existing methods for the collection of sulfur dioxide
addresses their efficiency, selectivity, and ability to stabilize the collected
material.
2-11
-------
2.3.2.1.1 Direct Sampling. Several of the integrated methods use direct
sampling, and determination of the amount of gas present in a flowing stream
of ambient air that passes through or around the sampling device. In general,
such methods suffer from a lack of sensitivity, an inability to discriminate
between different sulfur-containing species, or insufficient ruggedness for
long-term use in the field. Such shortcomings are of little consequence for
industrial emissions sampling when concentrations are high and the sulfur
species are known.
A variation of direct sampling involves collection of a small volume of
air in a gas-tight syringe or other suitable container for later analysis.
The reactivity of sulfur dioxide is a major problem, however. Natusch and
Bauer (1978) have reported extensive adsorption losses of SO^ on thick-walled
®
mylar laminates, tygon, Teflon , and stainless steel container walls.
2.3.2.1.2 Absorption. A number of integrated methods use aqueous solutions
for collection of sulfur dioxide. The efficiency of mass transfer of sulfur
dioxide from air to the solution phase depends on the gas-liquid contact time,
the diffusion coefficients of sulfur dioxide in the gas and liquid phases, the
bubble size, the concentration of sulfur dioxide, and the solubility of sulfur
dioxide in solution. Calvert and Workman (1960) have produced a method to
predict the efficiency of various bubbler designs in collecting sulfur dioxide.
Their method is predominantly qualitative, but it can serve as a useful guide.
The more efficient designs include that of Wartburg et al. (1969); the Petticoat
bubbler the Greenberg-Smith impinger (Smith et al., 1961); midget impingers
(Jacobs et al., 1957); Dreschel bottles (Wartburg et al., 1969); and packed
columns (Bostrom, 1966), which are useful where low flow rates (2 to 4
liters/minute) are involved. In using such devices, care must be taken to
2-12
-------
prevent carryover of solution at high flow rates and to compensate for solvent
losses by evaporation.
Collection efficiency depends in part on the solution in which sulfur
dioxide is actively dissolved and stabilized. One current method involves
stabilization of the sulfite anion in a solution of potassium tetrachloromer-
curate, with which the sulfite anion complexes. A second approach involves
collection in aqueous solution and direct oxidation to the sulfate anion,
which is then determined analytically. In the latter case, a number of oxi-
dizing agents such as hydrogen peroxide are used.
Although stabilization of sulfur dioxide as the sulfate anion can be
effective, any soluble sulfate in the atmospheric aerosol will be collected
(unless removed by a particle filter) and added to the sample; thus, discrimi-
nation between sulfur dioxide and sulfate may be impossible. Stabilization of
the sulfite anion with potassium tetra-chloromercurate is currently the basis
of the EPA reference method to determine sulfur dioxide in ambient air. To
prevent conversion of sulfite to sulfate, the temperature of the collecting
solution must be maintained below 20°C. Failure to maintain temperature
control of samples during collection, shipment, and storage will lead to
underestimation of the amount of sulfur dioxide present in the atmosphere,
particularly during summer months.
Several analytical procedures employ alkaline solutions for the absorption
of sulfur dioxide and their collection efficiency is quite high (Hochheiser,
1964). However, alkaline solutions rapidly oxidize the collected sulfite
anion to sulfate unless some means are available for the direct complexation
and stabilization of the sulfite anion.
2.3.2.1.3 Adsorption. There are several methods to collect sulfur dioxide by
adsorption. The Stratmann method (Katz, 1968; Stratman, 1954) utilizes a quartz
2-13
-------
tube packed with silica gel which removes sulfur dioxide from the air that
passes through the tube. The exact form of the collected sulfur dioxide has
not been determined, but it seems probable that stabilization as either a
sulfite or a sulfate occurs. Since the subsequent analytical procedure involves
desorption at elevated temperatures with a stream of hydrogen, any sulfate
species collected from the atmospheric aerosol may also be desorbed, resulting
in poor discrimination between sulfur dioxide and sulfate species. Compara-
tive tests show that this method yields highly erratic results (Organization
for European Economic Cooperation, European Productivity Agency, 1961).
The second adsorption procedure involves the use of filter papers or
tapes impregnated with an alkaline reagent such as potassium hydroxide, tri-
ethanolamine, or potassium carbonate, together with small amounts of glycerine
as an humectant. The adsorbed sulfur dioxide is supposed to be maintained as
a sulfite, but it may become oxidized to sulfate. Although laboratory tests
have shown that such oxidation can be negligible, field tests have produced
very erratic results. Particulate matter collected from air passing through
the filter contains traces of transition metal ions, which promote rapid
oxidation of sulfite to sulfate.
The third adsorption method involves the reaction of sulfur dioxide with
lead peroxide to form lead sulfate (Katz, 1968; Hochheiser, 1964). The sulfur
dioxide is stabilized in the form of a sulfate, eliminating the problem of
oxidative conversion. However, any particulate material containing sulfate
species which come in contact with the collection surface will lead to errors.
2.3.2.2 Calibration—The relationship between true pollutant concentration
and its measurement by any method is determined by calibration. For those
2-14
-------
integrated methods that measure relative exposure to sulfur species (e.g.,
sulfation methods), no calibration is usually attempted. The use of uniform
reagents, equipment, and procedures is recommended with these methods in order
to compare exposure data over time and space.
Those integrated methods that involve collection of an air sample for
later analysis or collection of the S0? in an air sample by absorption or
adsorption require calibration of both the sample volume measurement and the
analytical measurement.
Devices used for sample volume measurement are generally calibrated
against reliable volume standards. The analytical measurement is often cali-
brated statically using a known amount of the sulfite or sulfate anion in
solution. Static calibration is a rapid and simple method for checking the
analytical procedure, but does not subject the overall measurement method to
scrutiny since the process of S02 collection is circumvented.
Dynamic calibration of integrated methods has an advantage over the
static approach by subjecting the total measurement to scrutiny, but it is
time consuming and therefore is not used routinely. This approach, described
in more detail in a later section on continuous methods, uses synthetic atmo-
spheres containing the pollutant in known concentrations to define the response
of the method.
2.3.2.3 Analytical Methods—The principal methods of determining sulfur
dioxide in the air are discussed in this section.
2.3.2.3.1 Colorimetric Method: Pararosaniline. The West-Gaeke method is
probably the most widely used colorimetric procedure for SO,, determination in
ambient air (West and Gaeke, 1956). It is also the basis of the EPA reference
method for manual measurement of S0? in the atmosphere (U.S. Environmental
Protection Agency, 1979).
2-15
-------
In the West-Gaeke method, air is bubbled into fritted bubblers containing
0.1 M sodium tetrachloromercurate (TCM) solution, which forms the stable
dichlorosulfitomercurate ion with SOp. The TCM complex is reacted with bleached
pararosaniline and formaldehyde to form red-purple pararosaniline methylsulfonic
acid. The optical absorbance of the solution is measured spectrophotometrically
at 560 nm and is, within limits, linearly proportional to the concentration of
sulfur dioxide. The method is capable of measuring SO^ concentrations from 13 to
•3
13,000 ug/m (0.005 to 5 ppm). Ozone, nitrogen dioxide, and heavy metals were
negative interferences in early versions of this method.
An improved version of the West-Gaeke method was adopted by EPA in 1971 as
the reference method for the determination of S0? in the atmosphere. Several
important parameters were optimized, resulting in greater sensitivity and
reproducibi1ity, as well as adherence to Beer's Law throughout a greaterworking
range.
In the EPA method, SO^ is collected in impingers containing 0.04 M potassium
tetrachloromercurate. A 20-minute wait before analysis allows ozone, a potential
interferent, to decompose. Sulfamic acid is then added, followed by a 10-
minute wait to remove interference from nitrogen oxides. Interference by
heavy metals is eliminated by the use of phosphoric acid in the dye reagent
and the disodium salt of ethylenediaminetetraacetic acid (EDTA) in the TCM
absorbing solution. The complex is then reacted with a purified pararosaniline
dye reagent and formaldehyde to form the colored pararosaniline methylsulfonic
acid. Absorbance is measured at 548 nm.
Accuracy depends on rigid control of many critical variables: pH,
temperature, reagent purity, color development time, age of solutions, and
2-16
-------
concentrations of some atmospheric interferents (Scaringelli et al., 1967).
The final color can be stabilized with phosphoric acid. Because temperature
affects the rate of color formation and color fading, a constant-temperature
bath is recommended for maximum precision. Highly purified reagents, especially
the pararosaniline dye, are vital for acceptable reproducibility.
The precision of the analytical procedure in the EPA reference method was
estimated using standard sulfite samples (Scaringelli et al., 1967) and was
reported as 4.6 percent at the 95 percent confidence level. The lower limit
of detection of S02 in 10 ml of TCM absorbing solution was 0.75 pg, representing
3
a concentration of 25 pg SO^/m (0.01 ppm) in an air sample of 30 liters. A
collaborative study (McCoy et al., 1973) of the 24-hour EPA reference method
procedure indicated the following: method repeatability (day-to-day variability
within an individual laboratory) varies linearly with S02 concentration
from ± 18 (jg/m (0.007 ppm) at concentration levels of 100 |jg/m (0.04 ppm)
3 3
to ± 51 ug/m (0.019 ppm) at concentration levels of 400 pg/m (0.15 ppm);
method reproducibility (day-to-day variability between two or more laboratories)
3
varies linearly with SOp concentration from ± 37 pg/m (0.014 ppm) at 100
33 3
pg/m to ± 104 pg/m (0.040 ppm) at 400 pg/rn .
A recent investigation (Fuerst et al., 1976) showed that collected SOp
TCM samples decay at a temperature dependent rate. Significant decay can
occur during collection of ambient samples and during shipment and storage of
collected samples when the TCM solutions are exposed to temperatures above
20°C. The effect of this decay is to underestimate the concentration of SO,,
in the air sample. Measures to minimize these effects have been investigated
by Martin (1977), who recommends the use of thermostatted shelters to house the
2-17
-------
sampling equipment during sample collection. The temperature of samples during
shipment can be controlled with cold-pack shipping containers. When samples
are stored before being analyzed, refrigeration at 5°C minimizes further
decay. Temperature control procedures are currently being incorporated in the
EPA reference method.
2.3.2.3.2 Titrimetric Method: Hydrogen Peroxide. In this British Standard
Method, atmospheric SCL is absorbed in 0.03 N (1-Vol)hydrogen peroxide and
adjusted to pH 4.5 (British Standard 1747, Part 3, 1963). It is oxidized to
sulfuric acid by the absorbing solution and titrated with standard alkali.
Only simple equipment is required, and samples may be stored for long periods
o
before titration. With an 850-liter air sample, 25 ug SCL/m (0.01 ppm) may
be detected (Katz, 1968).
This method has several weaknesses; its sensitivity is insufficient for
instantaneous analysis, long sampling times (24-hr) are needed and accuracy is
reduced at low concentrations. Since it measures total acidity rather than
SOp, any atmospheric acid will produce positive errors. Similarly, ammonia
will neutralize absorbed SO, to give low results. If a particle filter is in
line before the bubbler, SO,, may be absorbed; this would also give a low
result.
2.3.2.3.3 Titrimetric Method: Barium Perchlorate-Thorin. Atmospheric SO,., is
absorbed in 1 percent hydrogen peroxide adjusted to pH 9 with ammonia. Inter-
fering aerosol sulfate is removed by first passing the air through phosphoric
acid. The sulfate formed from S02 is titrated in isopropanol with barium
perchlorate, using thorin as the indicator. A color change from yellow to
pink when excess barium ions are present indicates the end point. The titratior
is followed spectrophotometrically at 550 nm (Fritz and Yamamura, 1955). The
o
limit of detection with an air sample of 1000-liters is 50 pg S0?/m (0.02
ppm). Heavy metals, which complex with thorin, may be removed initially with
a particulate filter.
2-18
-------
2.3.2.3.4 Turbidimetric Method: Hydrogen Peroxide. Hydrogen peroxide (0.03
N) is the absorbent solution for removing S0? from air. Dilute HC1 and a
glycerol-alcohol solution is added to an aliquot of the resultant sulfate
solution. The optical absorbance of this solution is measured against a water
blank at 500 nm. Solid barium chloride crystals are added, and the solution
is shaken and permitted to stand for 40 minutes. The absorbance of the turbid
mixture is measured against water and corrected for the previous reading.
Factors that must be controlled in this method are pH, sulfate concentration,
stability of the colloidal barium sulfate suspension, barium ion concentration,
and aging of the barium chloride solution (Volmer and Frohlich, 1944).
2.3.2.3.5 Colorimetric Method: Barium Chloranilate. This method is based
upon the reaction of solid barium chloranilate with sulfate ions to form
barium sulfate and colored acid chloranilate ion (Bertolacini and Barney,
1957). SOp is scrubbed by a 0.5 percent hydrogen peroxide solution. The
solution is buffered at pH 4 with potassium acid phthalate; solid barium
chloranilate and 95 percent ethyl alcohol are then added. Excess barium chlo-
ranilate and precipitated barium sulfate are removed by filtration or centri-
fugation, and the optical absorbance of the liberated red-purple acid chlo-
ranilate ion is measured at 530 nm. A limit of detection of 130 ug SO^/m
o
(0.05 ppm) with a 1-m sample has been noted. Phosphate, fluoride, and chloride
interfere and require preliminary separation.
2.3.2.3.6 Colorimetric Method: Iodine. An absorbing solution containing
soluble starch, potassium iodide, dilute sulfuric acid, and standard 0.01 N
iodine solution is prepared (Katz, 1950). S02 in the air sample reacts with
2-19
-------
this 8 x 10~5 N iodine solution to decolorize the blue iodine-starch complex.
The reduction in color intensity is measured spectrophotometrically. The
range of applicability is 25 to 2600 |jg S02/m (0.01 to 1 ppm), depending upon
the volume and concentration of absorbent solution and the volume of air
sampled.
Oxidizing gases interfere to give low results; reducing agents interfere
to give high results. Interference from high concen trations of nitrogen
oxides or ozone can be removed by introducing hydrogen into the air sample and
passing the mixture over a platinum catalyst at 100°C (Bokhaven and Niessin,
1966).
2.3.2.3.7 Titrimetric Method: Iodine. Air is bubbled through a sodium
hydroxide solution which absorbs S02 (Jacobs, 1960). After acidification of
the solution, the liberated sulfurous acid is titrated with standard iodine
solution, using starch as an indicator. Because sulfite oxidizes to sulfate
in the alkaline absorbent solution, samples cannot be stored. Oxidizing
agents, nitrogen dioxide, and ozone interfere by underestimating the analysis.
Hydrogen sulfide and reducing agents overestimate the analysis. For an 850-
3
liter air sample collected at 30 liters/min, sensitivity is 25 ug S0?/m (0.01
ppm) (Terraglio and Manganelli, 1962).
2.3.2.3.8 Titrimetric Method: lodine-Thiosulfate. In this method, S02 is
absorbed in a solution of iodine, potassium iodide, starch, and sulfuric acid.
Some of the iodine is reduced by the S02; the excess iodine is titrated with a
standard thiosulfate solution (Katz, 1969). Interferences are the same as in
the previously described iodimetric methods.
2-20
-------
2.3.2.3.9 Col or1 metric Method: Ammonium Molybdate. Stratmann (Katz, 1968)
developed a method that uses silica gel in a quartz tube to adsorb SOp from
air that is passed through the tube. The collected sample in the tube is
placed in a 500°C furnace, and a stream of hydrogen is passed through. Sulfur
dioxide is desorbed and then reduced to hydrogen sulfide by passing it over a
platinum wire mesh catalyst heated to 600°C. The hydrogen sulfide is absorbed
in a solution containing ammonium molybdate, sulfuric acid, and urea. Absor-
bance of the resulting molybdenum blue complex is measured at 570 nm. A limit
of detection of 25 |jg SCL/m (0.01 ppm) can be obtained with a 40-liter air
sample (Katz, 1968).
Sulfuric acid, sulfur trioxide, and hydrogen sulfide interfere. H^SO.
and S03 may be eliminated by prior passage through phosphoric acid. Samples
containing more than 300 mg of water must be dried by preliminary passage
through phosphorous pentoxide, although some SCL is lost in the process. In a
series of comparative tests, this method has yielded erratic results (Organi-
zation for European Economic Cooperation, European Productivity Agency, 1961).
2.3.2.3.10 Ion-Exchange Chromatographic Method. Ion-exchange resins can be
used to provide excellent separation of ions. However, the continuous determv
nation of ionic species as they are eluted from an ion-exchange column has
been difficult, due principally to the electrolyte solution used as a eluant.
Recently, Small and co-workers (1975) devised a novel combination of resins
which effectively neutralize the eluant and suppress its conductivity without
affecting the analyte, permitting the use of a conductivity cell as a monitor
for even ppm levels of either cationic or anionic species.
A method for collection and ion Chromatographic analysis of atmospheric
SO has recently been developed by EPA (Mulik et al., 1978). The method uses
2-21
-------
dilute (0.6 percent) hydrogen peroxide to collect the ambient SO,,. The resul-
tant sulfate ion is analyzed by ion-exchange chromatography. When a prefilter
is used in the sampling train to remove aerosol sulfates, there are no apparent
interferences. Collection efficiency is approximately TOO percent over the
range of the method, 25 to 1300 ug S02/m3 (0.01 to 0.5 ppm).
2.3.2.3.11 Sulfation Method: Lead Peroxide. Sulfation methods are based on
the reaction of gaseous S0? in air with lead peroxide (PbCL) paste to form
lead sulfate (PbSO,). They are cumulative methods for estimating average
concentrations over extended time periods.
In the lead candle method (Wilsdon and McConnell, 1934), the paste is
prepared by mixing Pb02, gum tragacanth, alcohol, and water. The paste is
applied to a piece of cotton gauze wrapped around a cylinder 10 cm round and
10 cm high. After drying, the cylinder is exposed to the atmosphere in a
sheltered location. After exposure, the sulfated lead candle is treated with
sodium carbonate solution to dissolve the lead sulfate. Sulfate is then
determined gravimetrically or turbidimetrically. Measurements with the method
o
are reported as sulfation rates (mg SO.,/100 cm /day).
In the sulfation plate method (Huey, 1968), a similar paste containing
glass fiber filters is poured into a petri dish and, after drying, is exposed
to the atmosphere. Sulfation methods have the advantage of being inexpensive,
but their accuracy is subject to many physical and chemical variables and
interferents.
The rate of sulfate formation is proportional to atmospheric S0? concen-
tration up to 15 percent conversion of the lead peroxide (Wilsdon and McConnell,
1934). Reaction rate increases with temperature and with humidity. Other
2-22
-------
factors affecting rate of sulfation are wind velocity, purity of lead peroxide,
and shape of the shelter (Bowden, 1964). The consistency of lead peroxide,
both within batch and between batches, is difficult to control. Consequently,
the area of any sulfation surface could vary, giving different rates. Factors
affecting surface area are particle size of lead peroxide and the density and
thickness of the layer. Positive errors are contributed by hydrogen sulfide
and sulfate aerosols. Methyl mercaptan is a potential negative interferent.
o
The relationship between sulfation rates (mg SO^/100 cm /day) and SO-
concentration (ppm) is discussed in detail in section 2.3.5.1.
2.3.3 Continuous Methods
2.3.3.1 Sample Col 1ection--In continuous sulfur dioxide analyzers sample
collection is an integral part of the total automated measurement process.
The sample line leading from the sample manifold to the inlet of the analyzer
®
should be constructed of an inert material such as Teflon . The sample line
should be as short as is practically possible and its internal diameter should
assure short residence times and insignificant pressure drops between the
sample manifold and the analyzer inlet. The use of a particle filter at the
inlet of the analyzer should depend on the analyzer's susceptibility to inter-
ference, malfunction, or damage due to particulate matter. Heavy loading of
particulate matter on the filter may lead to erroneous S0? measurements;
therefore, it may be necessary to change the filter frequently.
2.3.3.2 Calibration—The relationship between true pollutant concentration
and the response of a continuous analyzer is best determined by dynamic calibra-
tion. In dynamic calibration, standard atmospheres containing SO- in known
concentrations are introduced into the analyzer. The analyzer is adjusted
accordingly to produce the proper response. Dynamic calibration provides
evidence that all components of the instrument are functioning properly.
2-23
-------
Standard gas for calibration purposes may be obtained from permeation
®
tubes (O'Keeffe and Ortman, 1966), which are sealed Teflon tubes containing
liquified gas. Gas diffuses through the walls at a low, known constant rate
at constant temperature. By diluting the emitted gas with varying flows of
clean air in a mixing chamber, a wide range of concentrations may be generated.
Permeation tubes with certified permeation rates are available from the
National Bureau of Standards (NBS) as Standard Reference Materials (SRMs).
Tubes with certified rates traceable to NBS are also available from commercial
suppliers.
When enclosed in an inert plastic tube, a liquefiable gas such as SOp
escapes at a constant temperature-dependent rate that can be determined by
measuring the weight loss of the tube over a long period of time. The concen-
tration of S02 in air flowing over the permeation tube at a given flow rate is
given by
r _ R x K
L ~ F
where C = S02 concentration, ppm
R = rate of weight loss of permeation tube at constant
temperature, ug/min
K = 0.382 ul S02/ug S02 (at 25°C and 760 mm Hg)
F = air flowrate (corrected to 25°C and 760 mm Hg),
liter/in in
A variation of this procedure allows pure S02 from a gas cylinder to diffuse
at a constant rate through a ceramic frit into the airstream. Permeation
tubes provide a simple and accurate method for preparing known S0? concen-
trations in air. However, adsorption losses restrict the lower limits at
which S02 concentrations can be considered reliable. To achieve precise
2-24
-------
calibrations, the temperature of the permeation tube must be carefully con-
trolled. Permeation tubes presently provide the simplest and most reliable
method of calibration.
Dynamic calibration may also be carried out using known mixtures of SOp
gas in high pressure cylinders. For stability purposes they are usually
prepared in high concentrations and dynamically diluted to the desired level.
Traceability of such standards to NBS SRMs may be established by the gas
standard manufacturer or by the user.
Static calibration techniques are possible for several of the continuous
SOp analyzers described below. Static calibration introduces a stimulus to
measure instrumental response under no sample air flow conditions. Typical
stimuli are electrical signals, solutions chemically equivalent to the pollutant,
or solutions producing comparable physical effects upon properties by which
the pollutant is detected such as optical density or electrical conductivity.
Static calibration is a rapid and simple method for checking various components
of the instrument, but does not necessarily subject total instrument performance
to scrutiny.
2.3.3.3 Analytical Methods
2.3.3.3.1 Conductimetric analyzers. Conductimetric analyzers were the first
commercially available instruments for continuously monitoring SCL in the
atmosphere and are still used today. In their operation, air is brought into
contact with an absorbing solution, which dissolves SOp. The ions formed by
SOp dissolution increase the conductivity, which is proportional to the
concentration. The absorbent may be either deionized water or acidified
hydrogen peroxide solution. When water is used, conductance is increased by
formation and dissociation of sulfurous acid:
SO, + H90 -» H9SO,
c. t. £5
2-25
-------
Hydrogen peroxide solution oxidizes S02 to form sulfuric acid:
S02 + H202 -> H£S04
Conductance is measured by a pair of inert (platinum) electrodes within
the cell. To increase accuracy, comparison is made to a reference cell,
which measures conductance of unused absorbent. Typical features of conducti-
3 3
metric analyzers are detection limits of 25 ug S02/m to about 5000 ug/m
(0.01 to 2 ppm), lag time (time interval from change in input concentration to
change in output signal) of about 20 seconds, and response time (time interval
from change in input concentration to 90 percent of maximum output signal) of
2 minutes.
The major disadvantage of conductimetric analyzers stems from their
operation; i.e., any species that either forms or removes ions from solution
and changes conductivity can interfere. The degree of interference depends on
humidity, temperature, SOp concentration, and the particular instrument. The
worst interferents are chlorine, hydrochloric acid, and ammonia; nitrogen dioxide
and carbon dioxide interfere to a lesser extent (Roberts and Friedlander,
1976). Airborne particles, especially ocean-borne salt aerosols, are potentially
damaging. Several methods have been used to minimize these problems (Intertech
JJP
Corp/j. Chemical scrubbers which selectively remove gaseous interferents have
been incorporated into some conductimetric analyzers. Particulate filters
have also been employed.
2.3.3.3.2 Colorimetric analyzers. Colorimetric analyzers are based upon the
reaction of S02 with solutions of organic dyes t,o form colored species.
Optical absorbance of the ensuing solution, measured spectrophotometrically,
is within limits linearly proportional to the concentration of the colored
species in accordance with Beer's Law.
2-26
-------
Most instruments utilize improved versions of the manual pararosaniline
method developed by West and Gaeke. Automation of the West-Gaeke method per se
does not ensure a practical continuous monitoring instrument since some solutions
require daily preparation.
Typical features for some commercially available instruments are detection
2
limits of M3 to 5000 ug SCL/m (0.005 to 2 ppm), lag time of 2 minutes, and
response time of 8 minutes (-Instrument-alion fur Eiiy^ruiiineiiLdl Mojm.-oring, Air,-
^9-72). Advantages of these instruments include simplicity, high sensitivity,
and, with proper control, good specificity. Interferences by nitrogen oxides
may be controlled by using a sulfamic acid reagent. Heavy metals may be
complexed with EDTA in the scrubbing solution or with phosphoric acid in the
dye solution. Ozone interference may be controlled by the use of a delay coil
downstream from the absorber to allow time for ozone to decay, but this results
in longer lag and response times. Major disadvantages of these instruments
are the need for reagent and pump tubing replacement and frequent recalibration.
2.3.3.3.3 Flame photometric analyzers. The flame photometric detector (FPD)
*
is based on the measurement of the band emission of excited S? molecules
during passage of sulfur-containing compounds through a hydrogen-hyperventilated
(reducing) flame. The emitted light passes through a narrow-pass optical
filter, which isolates the 394 nm S? band, and is detected by a photomultiplier
tube (PMT). Application of the FPD to the detection of S02 was first made by
Crider (1965) and analyzers using FPD have been widely accepted for ambient
S02 monitoring.
Typical operational features of continuous flame photometric S02 analyzers
are a sensitivity of 0.005 to 1 ppm, lag time of 3 seconds, and response time
of 10 seconds (Ift£4yumentation for .Environmental Monitoring,—A4-P-, 1972).
2-27
-------
Although the FPD is insensitive to nonsulfur species, it will detect
sulfur compounds other than SOp. Since the concentration of H2$, mercaptans,
and other organic sulfur compounds in the atmosphere is usually less than 10
percent of the SOp level, their potential interference is normally minimal.
Particulate filters will remove troublesome aerosol sulfates along with other
particulates to eliminate clogging and light scattering. Selective filters
may be used to reduce interference from other gaseous sulfur compounds; e.g.,
an HpS filter is used on most commercial instruments. Interference by carbon
dioxide can be minimized by maintaining identical COp concentrations in the
calibration and sample matrices.
Gas chromatographs with flame photometric detectors (GC-FPD) are also
available commercially. GC-FPD can separate individual sulfur compounds and
measure them individually (Stevens et al., 1971). However, the temporal
resolution of GC-FPD data is limited by the chromatographic elution time of
SCL and other gaseous sulfur compounds.
Disadvantages of FPD systems include the need for a compressed hydrogen
source and sensitivity to all sulfur compounds. Photomultiplier tube output
is proportional to the square of SO^ concentration; hence, an electronic
system to "linearize" output is a desirable feature. Advantages of FPD systems
include low maintenance, high sensitivity, very fast response, and good selec-
tivity for sulfur compounds. No reagents are necessary other than compressed
hydrogen.
2.3.3.3.4 Coulometric (amperometric) analyzers. Coulometric analyzers are
based on the reaction of S02 with a halogen, formed directly by electrolysis
of a halide solution:
S02 + 2H20 + Br2 •* H2S04 + 2HBr
(I9) (2HI)
2-28
-------
The current necessary to replace the depleted halogen is proportional to the
amount of S02 absorbed in the solution, and hence to the S0? concentration in
the air.
In one common coulometric system, an inner chamber, into which air is
introduced, is contiguous with an outer chamber (Treon and Crutchfield, 1942).
Both contain a solution of potassium bromide and bromine in dilute sulfuric
acid. Potential difference between both chambers, relative to a reference
potential, is measured by the reference electrodes. As absorbed SO,, reduces
the Br2 concentration in the inner chamber, the amplifier produces a current
to restore the Br2 content in the inner chamber until the potential difference
is again zero.
In a second system, the change in halogen concentration is detected as a
current change rather than a potential difference. The cell is filled with a
potassium iodide solution, buffered to pH 7. At the platinum anode, a constant
current source continuously generates iodine, which is subsequently reduced at
the cathode. An equilibrium concentration of iodine is established, and no
current is generated at an activated-carbon bipolar reference electrode,
connected in parallel. Reaction with S02 decreases the equilibrium concentration
of iodine, which cannot transport the charge generated by the constant-current
source. Part of the current is diverted through the reference electrode; this
flow is proportional to the S02 in the air sample.
Interferent species are those able to oxidize halides, reduce halogens,
or complex with either. They consist primarily of sulfur compounds (hydrogen
sulfide, mercaptans, and organic sulfides and disulfides) with sensitivities
comparable to that of SCL. Other potential interferents, at lower sensitivities,
are ozone, nitrogen oxides, chlorine, ethylene, aldehydes, benzene, chloroform,
2-29
-------
a-pyrene, other nitrogen or halogen-containing compounds, and other hydrocarbons
(DeVeer et al., 1969; Thoen et al., 1968). Interferences can be minimized by
selective filters, which are sometimes built into the instrument or offered as
optional features. For example, a heated silver gauze filter is reported to
remove hydrogen sulfide, ozone, chlorine, nitrogen oxides, carbon disulfide,
ethylene, aldehydes, benzene, and chloroform, but will not remove mercaptans
^
(Philips Electronic Instruments/).
The major advantage of a coulometric analyzer is minimal maintenance
(reagent may need only monthly replacement; electrodes may require annual
cleaning). Also, reagent consumption is negligible because of halide regenera-
tion, and evaporated water is replaced by condensation from air or from a
reservoir.
2.3.3.3.5 Fluorescence analyzers. Fluorescence analyzers are based on detection
of the characteristic fluorescence released by the sulfur dioxide molecule
when it is irradiated by ultraviolet light (Okabe et al., 1973). This fluorescent
light is also in the ultraviolet region of the spectrum, but at a different
wavelength than the incident radiation. The fluorescent wavelengths usually
monitored are between 190 and 230 nm. In this region of the spectrum, there
is relatively little quenching of the fluorescence by other molecules occurring
in ambient air. The light is detected by a PMT that, through the use of
electronics, produces a voltage proportional to the light intensity and S0?
concentration.
The fluorescent light reaching the PMT is usually modulated to facilitate
the high degree of amplification necessary. Some analyzers mechanically
"chop" the incident irradiation before it enters the reaction chamber. This
process is accomplished by a fan-blade-like chopper rotating at a constant
2-30
-------
speed, which alternately blocks and passes the light to the chamber. Other
instruments electronically pulse the incident light source at a constant rate.
Potential interferences to the fluorescence technique include any species
that either quenches or exhibits fluorescence. Both water vapor and oxygen
strongly quench the fluorescence of S0~ at some wavelengths. Water vapor can
be removed by a dryer within the instrument or the water interference can be
minimized by careful selection of the incident radiation wavelength. The
effect of oxygen quenching can be minimized by maintaining identical oxygen
concentrations in the calibration and sample matrices.
Aromatic hydrocarbons such as naphthalene exhibit strong fluorescence in
the same spectral regions as S02 and are major interferents. These aromatics
must be removed from the sample gas stream by an appropriate scrubber upstream
of the reaction chamber. The scrubbers may operate at ambient or elevated
temperature. Certain elevated-temperature scrubbers, however, have the potential
for converting ambient hydrogen sulfide (which normally does not inerfere with
the fluorescence technique) into SCL. In these cases, the hydrocarbon scrubber
must be preceded by a scrubber for H?S.
2.3.3.3.6 Second-derivative spectrometric analyzers. The second-derivative
spectrometer processes the transmission-versus-wavelength function of a spectrum
to produce a signal proportional to the second-derivative of this function
(Hager and Anderson, 1970). The signal is proportional to the concentration
of the gas in the absorption path. These instruments center on the shape
characteristics rather than basic intensity changes of molecular band spectral
absorption. The slope and curvature characteristics are often large, specific,
and independent of intensity. Because these shape characteristics are large but
specific to individual compounds, complex separations of component gases are
possible.
2-31
-------
In the operation of a second-derivative spectrometer, radiant energy from
a UV or visible source is directed into a monochromator, where it is dispersed
by a grating to provide monochromatic light to the sample cell. The wavelength
of this light is modulated with respect to time in a sinusoidal fashion by an
oscillating entrance slit. The angular position of the grating sets the
center wavelength coming out of the monochromator into the multipass cell. The
sample is continuously drawn through the cell by a pump. Output from the
photomultiplier tube is electronically analyzed to develop the second derivative
of the absorbance.
Sensitivity is greatly enhanced over ordinary spectrometers because the
output is an AC signal of known wavelength and phase, adaptable to high-gain
electronic amplification. Uniqueness of the curvature of a given molecular
band enables this type of instrument to be-highly specific. Measurements are
independent of sample flow rate, but relatively high flow rates (4 liters/min)
are necessary to achieve reasonable response time.
2.3.3.3.7 Other Methods. Electrochemical transducer (ECT) analyzers measure
the current generated by electrochemical oxidation of SO^ at a sensing electrode.
All the chemical reactions take place within a sealed transducer module,
eliminating the wet chemistry of conductimetric, colorimetric, and coulometric
analyzers.
Correlation spectrometry is based upon absorption spectrometry of molecular
S02 in a longpath sample of air. Either natural or artificial light may be
used. Light is collected by a telescope, collimated, and dispersed by a
grating or prism. The spectrum is then focused on the plane of an exit mask,
which is a replica of the absorption spectrum of SO^. The difference in light
intensity reaching the detector is related to the S02 concentration in the
2-32
-------
path between the light source and the instrument (Barringer Research, Ltd,
1969; Moffat et al., 1971).
Condensation nuclei formation instruments consist of two units: a converter
for changing S0~ molecules into condensation nuclei, and a monitor for detecting
the nuclei. Sample air is drawn through a glass fiber filter, which removes
existing condensation nuclei. S02 molecules, in the presence of water vapor
and UV radiation, are converted into H?SO. aerosol. The aerosol, consisting
of condensation nuclei, is introduced into the monitor. Constant volume
expansion in the presence of water vapor produces a cloud; transmission of
this cloud is directly related to the initial S0? concentration.
Differential lidar is a remote sensing system which provides range-resolved
measurements of concentration by reflecting pulses of laser light at two wave-
lengths with different absorption coefficients from particles along the line
of sight. The amount of light that is scattered and received at the detector
determines the detection limits. Thus, a high-energy laserpulse operating
under conditions of good visibility provides the bes sensitivity.
2.3.4 EPA Reference and Equivalent Methods
Under the provisions of EPA's "Ambient Air Monitoring Reference and
Equivalent Methods" regulations codified as 40 CFR Part 53 (U.S. Environmental
Protection Agency, 1979), several commercial analyzers have been designated as
equivalent methods for determining compliance with the NAAQS for sulfur dioxide.
These analyzers have been subjected to the required testing and have met EPA's
performance specifications for automated methods. These specifications are
given in Table 2-3, and a list of the S0~ analyzers, which have been designated
as of March 1, 1980, is given in Table 2-4. Information about the designation
of these analyzers as equivalent methods may be obtained by writing the Environ-
mental Monitoring Systems Laboratory, Methods Standardization Branch (MD-77),
2-33
-------
United States Environmental Protection Agency, Research Triangle Park, North
Carolina 27711.
A review of the performance data submitted in support of the designations
listed in Table 2-4 indicates that these analyzers exhibit performance better
than that specified in Table 2-3. For the analyzers tested, noise levels were
typically 3 ppb or less. The zero drift results (12-and 24-hour) were all
less than 5 ppb and typically less than 3 ppb. The span drift results (at 20
and 80 percent of the full scale range of 0 to 0.5 ppm) were all less than 5
percent and typically 2 to 3 percent. The precision results (at 20 and 80
percent of the full scale range of 0 to 0.5 ppm) indicate a typical precision
of 1 to 2 ppb. The response times (lag, rise, and fall times) for the various
types of analyzers were typically as follows: flame photometric, 1 minute or
less; fluorescence, 5 minutes; coulometric, 3 minutes; conductimetric, 0.5
minute; second derivative spectrometric, 8 minutes. For analyzers of the same
type (e.g., flame photometric), interference test results for a given potential
interferent were somewhat variable. Interference equivalents of 5 ppb or less
were obtained in each case except for the following: flame photometric, negative
COp interference equivalents of about 10 ppb were typical; coulometric, positive
Oo interference equivalents of about 8 ppb were typical.
Two manual methods have also been designated as equivalent methods
(Federal Register, 1975a). These methods are identified as:
-WI
h, "Pararosanil ine Method for the Determination of Sulfur
Dioxide in the Atmosphere Technicon I Automated Analysis System."
&QS-0-11f'602-
£QS0775Qeg-. "Pararosanil ine Method for the Determination of Sulfur
Dioxide in the Atmosphere Technicon II Automated Analysis System."
These methods employ the same sample collection procedure used in the EPA
reference method and an automated analytical measurement based on the colori-
metric pararosaniline method.
2-34
-------
TABLE 2-3. PERFORMANCE SPECIFICATIONS FOR AUTOMATED METHODS
Performance parameter
Range
Noise
Lower detectable limit
Interference equivalent
Each interferent
Total interferent
Zero drift, 12-and 24-hour
Span drift, 24-hour
20 percent of upper range limit
80 percent of upper range limit
Lag time
Rise time
Fall time
Precision
20 percent of upper range limit
80 percent of upper range limit
Units
ppm
ppm
ppm
ppm
ppm
ppm
percent
percent
minutes
minutes
minutes
ppm
ppm
Sulfur dioxide
0-0.5
0.005
0.01
±0.02
0.06
±0.02
±20.0
±5.0
20
15
15
0.01
0.015
Source: Federal Register (1975)
2-35
-------
TABLE 2-4. LIST OF DESIGNATED EQUIVALENT METHODS (AUTOMATED ANALYZERS)
Designation
number
EQSA-1275-005
EQSA-1275-006
EQSA-0276-009
EQSA-0676-010
Identification and source
Lear Siegler Model SM1000
S0? Ambient Monitor
Lear Siegler, Inc.
74 Inverness Drive East
Englewood, Colorado 80112
Meloy Model SA185-2A ,
Sulfur Dioxide Analyzer
Meloy Laboratories, Inc.
6715 Electronic Drive •
Springfield, Virginia 22151
Thermo Electron Model 43
Pulsed Fluorescent S02
Analyzer
Thermo Electron Corp.
108 South Street
Hopkinton, Massachusetts 01748
Philips PW9755 S02 Analyzed
Federal
vol .
41
41
42
45
41
43
41
41
42
44
45
41
41
42
Register
page
3893
32946
13044
1147
3893
38088
8531
15363
20490
21861
2700
26252
46019
28571
Notices
date
1/27/76
8/06/76
3/08/77
1/04/80
1/27/76
8/25/78
2/27/76
4/12/76
4/20/77
4/12/79
1/14/80
6/25/76
10/19/76
6/03/77
Philips Electronic Instruments, Inc.
85 McKee Drive
Mahwah, New Jersey 07430
EQSA-0876-011
EQSA-0876-013
Philips PW9700 S0£ Analyzer0
Philips Electronic Instruments,
85 McKee Drive
Mahwah, New Jersey 07430
Monitor Labs Model 8450
Sulfur Monitor
41
Inc.
41
44
34105
36245
33476
8/12/76
8/27/76
6/11/79
Monitor Labs, Inc.
10180 Scripps Ranch Blvd.
San Diego, California 92131
2-36
-------
TABLE 2-4. LIST OF DESIGNATED EQUIVALENT METHODS (AUTOMATED ANALYZERS) (Contd)
Designation
number
Identification and source
Federal Register Notices
vol. page date
EQSA-0877-024
EQSA-0678-029
EQSA-1078-030
EQSA-1078-032
EQSA-0779-039
ASARCO Model 500 and Model 600 42
e *
Sulfur Dioxide Monitors ' 44
ASARCO Incorporated
3422 South 700 West
Salt Lake City, Utah 84119
Beckman Model 953 Fluorescent 43
Ambient SO,, Analyzer0
Beckman Instruments, Inc.
2500 Harbor Boulevard
Fullerton, California 92634
Bendix Model 8303 43
Sulfur Analyzer
The Bendix Corporation
P. 0. Box 831
Lewisburg, West Virginia 24901
Meloy Model SA285E 43
Sulfur Dioxide Analyzer
Meloy Laboratories, Inc.
6715 Electronic Drive
Springfield, Virginia 22151
Monitor Labs Model 8850 44
Fluorescent S02 Analyzer0
Monitor Labs, Inc.
10180 Scripps Ranch Blvd.
San Diego, California 92131
44264
67522
35995
9/02/77
11/26/79
8/14/78
50733 10/31/78
50733 10/31/78
44616
7/30/79
Second derivative spectrometric measurement principle
Flame photometric measurement principle
Fluorescence measurement principle
Coulometric measurement principle
Q
Conductimetric measurement principle
*These analyzers are not offered for sale
2-37
-------
2.3.5 Method Comparison Studies
2.3.5.1 Comparisons Involving Sulfation Methods and Relationship Between Sulfation
and S02 Concentration—Stalker et al., (1963) compared the lead
candle method and the pararosaniline method to measure sulfur dioxide at 123
stations in Nashville, Tennessee. The lead candle method used to collect
monthly sulfation samples was a modified version of the method developed by
Wilsdon and McConnell (1934). After exposure, the rate of sulfation was
determined by the gravimetric procedure in which barium chloride precipitates
the sulfate as barium sulfate. The results of 2800 samples were reported in
p
milligrams of sulfur trioxide per 100 cm of exposed gauze per day. Sulfur
dioxide concentration was determined by a modified version of the pararosaniline
method described by West and Gaeke. The dye concentration was changed from
the recommended 0.04 percent pararosaniline hydrochloride-6 percent hydrochloric
acid mixture to a 0.005 percent-5 percent mixture in order to use an automated
analytical system. The sensitivity of the revised method was 0.006 ppm by
volume for a 120-liter air sample scrubbed through 15 ml of collecting solution.
Colorimetric results were based on 13,000 24-hr samples and 30,000 2-hr samples
(Welch and Terry, 1960).
The lead candle method was considered good for estimating mean SO,, levels
in communities during months with arithmetic mean concentrations of at least
0.025 ppm. The reliability of these mean estimates was estimated to be within
±25 percent. The authors concluded that the relationships between results
obtained with the two different methods of measuring S0? were best during the
fall and winter, when atmospheric concentrations were highest. During the
p
winter season of high S02 levels, 1 mg of S03/100 cm /day (sulfation rate)
measured by the lead candle method was equivalent to 0.042 ppm S0? measured by
the pararosaniline method. During the fall season of moderate S02 levels,
2-38
-------
the 1-mg equivalent was 0.035 ppm SO,,. During the spring season, the 1-mg
equivalent decreased to only 0.015 ppm S02. Using an average conversion factor
of 0.031, the lead candle measurements of S02 during the spring season of low
S0? levels would thus be about twice as high as simultaneous 24-hr colorimetric
measurements of S0?.
Huey et al., (1969) compared the sulfation plate method with the lead
candle method to determine S02 concentrations at some 250 sampling sites
nationwide. Both sulfation plates and lead candles were used at each sampling
site. Twenty-five thousand paired samples were analyzed by linear regression.
The correlation coefficient was found to be 0.95; the regression equation
obtained was X = 0.03 + 1.1Y, where Y = sulfation plate result in mg of SO.,/100
2 2
cm /day and X = predicted lead candle result in mg of S03/100 cm /day.
Obtaining a correlation coefficient of 0.95 confirmed that both methods
are measuring the same species. Because no values of less than 0.1 mg of
2
SO.,/100 cm /day were measured, the constant 0.03 in the regression equation
can be considered to be zero. The authors concluded from the coefficient of Y
that sulfation plates are 10 percent less reactive than lead candles.
Huey and coworkers also compared S02 conductivity monitors with sulfation
results. They concluded that sulfation data could be converted to S02 concen-
trations by multiplying by 0.03, and that 95 percent of the time this approxi-
mation from a single value will lie within a factor of about 3 of any single
instrument measurement.
Turner and Sholtes (1971) investigated atmospheric S02 levels by exposing
a series of lead peroxide cylinders and sulfation plates in parallel to several
known concentrations of S0? for specific times in a 10,000-liter stainless-
steel tank (test chamber). Lead peroxide cylinders and sulfation plates were
2-39
-------
contained within two different types of shelters. For each experimental run,
most sulfation plates were analyzed photometrically, while several were analyzed
gravimetrically. Most lead peroxide cylinders were analyzed gravimetrically,
while a few were analyzed photometrically. In several experimental runs, the
same lead peroxide cylinder was analyzed both by gravimetric and photometric
methods.
The authors concluded that gravimetric analysis of lead peroxide cylinders
was a long and laborious procedure subject to many errors. A typical problem
2
was the care required to analyze only the lead included in the 100 cm area,
excluding any that might be outside this zone. The photometric analysis was
conparatively quick and simple but less consistent and less reproducible than
the gravimetric procedure.
The sulfation plates were easier to prepare and analyze (photometrically)
than cylinders (gravimetrically), requiring only about half as much time for
preparation and analysis.
The following observations and conclusions were also noted: (1) Lead
peroxide cylinders are more sensitive to the effects of SCL than sulfation
plates. This may be due in part to the difference in the geometry of the two
devices, since the cylinder presents a more exposed surface than the plate.
(2) The shape, or absence, of the shelter appears to affect the sulfation
rate of the plates and cylinders. Samples in round shelters give sulfation
results that are closer to open samples than do samples in square shelters.
(3) The factor 0.03 proposed by Huey (1969), which gives a prediction of
S02 concentration in pprn when multiplied by cylinder sulfation in mg of SO.V100
2
cm /day, appears to be low. The results of this study indicate the factor
would be 0.06-0.09 for cylinders in round shelters and 0.09-0.12 for cylinders
in square shelters.
2-40
-------
(4) Exposure of cylinders and plates to varying concentrations of SCL
produced curved lines, showing a nonlinear relationship between sulfation
rates and SO^ concentration.
(5) Prediction of atmospheric S0? concentrations in ppm from open
sulfation plate analyses can be made by multiplying the sulfation value of the
2
plate in mg of S03/100 cm /day by a factor of 0.11.
Rider et al., (1977) studied the performance of Huey sulfation plates and
the effects of humidity at high S0? concentrations. Tests were conducted in a
6.25-liter desiccator for various SCL concentrations (800 to 3200 ppm) at
different relative humidities (0 percent, 50 percent, and 100 percent) and
exposure times (1 to 12 days).
Taking the data for 0 percent relative humidity, all concentrations
except the highest (3200 ppm) showed a linear increase in plate sulfation with
time over a period of several days. For the highest concentration the sulfation
rate decreased after a few days, indicating that the plates were saturated on
their surfaces and the process became nonlinear in time.
The sulfation rate did not appear to be appreciably affected by changes
in humidity, although one concentration (2400 ppm) appeared to exhibit an
effect. Rider et al. attribute this result to experimental error, since there
was no reason to suspect that such an effect would occur only for this concen-
tration. At 3200 ppm the rate did not change with humidity but the capacity
of the plate appeared to increase, and the effects of saturation occurred
later.
The authors' analysis of the relationship between sulfation and SO^
concentration using the model
[Sulfation (ug SO^/cm /day) x constant = S09 (ppm)n]
2-41
-------
yielded a value for n = 3/2 and a value for the constant = 309 for the high
SCL concentrations tested. A true linear relationship would yield an n value
of 1.
Rider et al.'s studies using the 6.25-liter desiccator suggest that
ambient S02 concentration (ppm) could be estimated using a linear correlation
2
factor by multiplying sulfation rate (mg S03/100 cm /day) by a value of approxi-
mately 800. Preliminary studies by the authors using a 20.7-liter carboy
suggested a factor of 260. From these results, those of Huey (1969), and
those of Turner and Sholtes (1971) using a 10,000-1iter chamber, the authors
concluded that (1) the factor depends upon the volume of the container used in
the studies, and (2) the factor decreases as the volume of the container
increases. Calculations by Rider et al. indicate that the amount of sulfation
was roughly proportional to the number of moles of S02 present in the container.
Considering the atmosphere to be approaching an infinite volume, the Huey
factor of 0.03 might represent a limiting value that is appropriate for very
large volumes.
For low ambient concentrations, the assumption of a linear relationship
(n = 1 instead of n = 3/2) appears to be reasonable and the Huey factor appears
to be appropriate for converting sulfation rate to SOp concentration.
2.3.5.2 Comparison of Pararosaniline and Barium Perch!orate-Thorin Methods—The
Rijksinstituut voor de Volksgezondheid in the Netherlands made a comparison in
1972-73 of the pararosaniline method and the barium perchlorate-thorin method
for determining sulfur dioxide in air (Thrane, 1978). The sampling devices
for the two methods were placed side by side at three monitoring stations.
The results of a correlation analysis of the data are presented in Table 2-5
2-42
-------
ro
r
CJ
TABLE 2-5. RESULTS FROM A CORRELATION ANALYSIS OF THE THORIN METHOD VERSUS
THE PARAROSANILINE METHOD
Mean SO^ concentration,
yg/m
No. of
Station Samples Thorin Pararosaniline
Den Helder 252 19.7 18.2
Wageningen 239 32.3 27.0
Witteveen 223 19.1 23.7
Correlation
Coefficient Regression equation
0.937 S02(Th) = 0.937
0.888 S02(Th) = 0.897
0.912 S02(Th) = 0.946
S02 (Para)
S02 (Para)
S02 (Para)
-0.187
-0.200
+0.511
SOURCE: Thrane (1978)
-------
and show reasonably good agreement between the two methods even at these low
SCL concentrations.
2.3.5.3 Comparisons of Continuous Methods—Palmer et al. (1969) evaluated the
field performance of eleven commercially available sulfur dioxide monitors
over a 3-month period in New York City. Among the characteristics investigated
were calibration drift, maintenance requirements, and comparability of data
for each analyzer. The study was conducted during the winter when S02 concen-
trations were relatively high. The mean concentration, calculated from the
means of the eleven analyzers (8 conductimetric, 2 colorimetric, 1 coulometric)
over the course of the study, was 0.206 ppm.
A statistical analysis of the data by the authors revealed the following:
(1) There were significant differences at the 0.001 probability level
among instrument means. However, at the 0.01 level, the only significant
differences were between one of the colorimetric analyzers and two of the
conductimetric analyzers.
(2) There were significant differences between instrument variances
(variability from day-to-day as compared with the means of all instruments for
each hour of that day). The highest variability was observed with one of the
colorimetric analyzers and two of the conductimetric analyzers.
(3) The correlation coefficients for all eleven instruments paired with
each other ranged from a minimum of 0.40 (colorimetric vs conductimetric) to a
maximum of 0.96 (conductimetric vs conductimetric).
A re-evaluation of the Palmer study reveals that, with the exception of
the one colorimetric analyzer and the two conductimetric analyzers, the corre-
lation between monitors was reasonably good (0.80-0.96).
2-44
-------
Concern over whether the reasonable agreement found in the Palmer study
would hold at lower mean concentrations than those observed in New York City
led to a study by Stevens et al. (1972). Ambient S0? concentrations were
measured over an 83-day period by five continuous monitors in Los Angeles.
The measurement principles of the five analyzers were: (1) colorimetric
(West-Gaeke), (2) conductimetric, (3) coulometric, (4) flame photometric, and
(5) gas chromatographic-flame photometric.
A correlation matrix of 24-hourly averages selected over the monitoring
period is presented in Table 2-6. Mean values, along with correlation
coefficients, can be used to compare measurements on an absolute basis. The
authors concluded that the conductimetric and coulometric analyzers (1)
exhibited poor correlations with the colorimetric, FPD, and GC-FPD analyzers,
and (2) recorded SCL concentrations as much as twice the values measured with
the other analyzers.
Measurements obtained in downtown Los Angeles on a day when the SO-
concentration exceeded 0.100 ppm are shown in Figure 2-2. The corresponding
correlation coefficients and mean concentrations for the analyzers operational
on that day are given in Table 2-7. The SGy concentrations recorded by the
conductimetric and coulometric analyzers were consistently higher than those
observed with the FPD analyzers.
The authors suggest that substances such as fine particulate matter,
ozone, oxides of nitrogen, and certain organics, which are generated during
the daytime hours by industrial and automotive activity, can produce positive
interference in the conductimetric and coulometric analyzers.
The conductimetric analyzer in the Stevens study was not equipped with a
chemical scrubber to remove potential interferences. Instead, it relied on
t
2-45
-------
TABLE 2-6. SULFUR DIOXIDE CORRELATION COEFFICIENTS AND MEAN CONCENTRATIONS
FOR MEASUREMENTS MADE BETWEEN SEPTEMBER 11 AND NOVEMBER 24, 1970 (1)
Technicon
Instrument Colorimetric
Technicon
Colorimetric
West-Gaeke
Me Toy
FPD
Philips
Coulometric
Leeds & Northrup
Conducti metric
Tracer
GC-FPD
1
0.805
0.721
0.140
0.870
Philips Leeds &
Meloy Coulo- Northrop Tracer
FPD metric Conductimetric GC-FPD
1
0.810 1
0.663 0.572 1
0.743 0.432 0.139 1
Mean Concentration,
ppm
0.036
0.033 0.050 0.063 0.040
(1)
Data selected for this correlation matrix were restricted to hourly average
SO- concentrations exceeding 0.02 ppm.
SOURCE: Stevens et al. (1972)
2-46
-------
250
200
.150
esj
o
CO
100
50
NO DATA j *
FPD
COUL-—-
COND
GC-FPD
1130
1230 1330
TIME OF DAY
1430
FIGURE 2-2. Sulfur Dioxide Concentrations Between 1200 and 1500 Hours
on November 24, 1970.
SOURCE:
Stevens et al. (1972)
2-47
-------
TABLE 2-7. SULFUR DIOXIDE CORRELATION COEFFICIENTS AND MEAN CONCENTRATIONS
FOR MEASUREMENTS MADE NOVEMBER 24, 1970, 1200 to 1355 HOURS^
(2")
Technicorr Meloy Philips
Instrument Colorimetric FPD Coulometric
Technicon —
Colorimetric
West-Gaeke
Meloy — 1
FPD
Philips -— 0.967 1
Coulometric
Leeds & Northrup — 0.978 0.962
Conductimetric
Tracer — 0.887 0.809
GC-FPD
Mean Concentration, — 0.084 0.177
ppm
Leeds & Northrop Tracer
Conductimetric GC-FPD
1
0.927 1
0.118 0.082
(1)
5-Minute Intervals
(2)
Colorimetric data not available.
SOURCE: Stevens et al. (1972)
2-48
-------
control of absorbing solution pH to derive specificity for SO,,. The coulometric
analyzer was equipped with a scrubber to remove only HLS. Current analyzers
based on these principles are generally equipped with complex chemical and
physical scrubbers to minimize interference problems.
2.3.5.4 Comparison of EPA Reference and Equivalent Methods—As part of the
required equivalency testing by the manufacturer, all of the continuous SCL
analyzers designated by EPA as equivalent methods have demonstrated a consistent
relationship with the reference method. A consistent relationship is demonstrated
when the differences between (1) measurements made by the test analyzer and
(2) measurements made by the reference method are less than or equal to the
maximum discrepancy specification in the last column of Table 2-8, when both
methods simultaneously measure SOp concentrations in a real atmosphere. All
of the equivalent methods listed in Table 2-4 have demonstrated this consistent
relationship with the reference method. The differences between simultaneous
measurements were generally well within the specifications listed in Table 2-8.
A comparison study using EPA designated equivalent methods for S02 was
recently conducted by EPA in an urban-industrial-commercial area of Durham,
North Carolina (Environmental Monitoring Systems Laboratory, 1979) Eight
continuous S0? analyzers were compared over a period of 150 days under more or
less typical air monitoring conditions. During the study, the analyzers
simultaneously measured ambient air sampled from a common manifold. The
ambient sample was occasionally augmented with artifically generated pollutant
to allow for analyzer comparisons at higher concentrations. Concentration
levels at which "spiked" testing occurred were 75, 125, and 250 ppb. A frequency
distribution of the 1-hour average SOp concentrations measured (ambient and
ambient + spike) is shown in Table 2-9.
2-49
-------
TABLE 2-8. TEST CONCENTRATION RANGES, NUMBER OF MEASUREMENTS
REQUIRED, AND MAXIMUM DISCREPANCY SPECIFICATION
S02 concentration Simultaneous measurements Maximum discrepancy
range, ppm required specification, ppm
1-hr 24-hr
low (0.02-0.05)
med (0.10-0.15)
high (0.30-0.50)
3
2
7 2
0.02
0.03
0.04
SOURCE: Federal Register (1975)
E fa
2-50
-------
TABLE 2-9. FREQUENCY DISTRIBUTION OF SULFUR DIOXIDE MEASUREMENTS(1)
Concentration Percent of Measurements
ppm in Concentration Range
<0.020
0.020 -
0.040 -
0.060 -
0.080 -
0.100 -
0.120 -
0.140 -
0.160 -
0.180 -
0.200 -
0.220 -
0.240 -
0.260 -
0.280 -
0.300 -
0.320 -
0.340 -
0.360 -
0.039
0.059
0.079
0.099
0.119
0.139
0.159
0.179
0.199
0.219
0.239
0.259
0.279
0.299
0.319
0.339
0.359
0.379
44.6
7.8
9.5
2.7
1.1
7.8
12.7
2.7
0.2
2.7
0.3
2.0
3.0
0.8
0.3
0.6
0.5
0.5
0.1
(1)
Ambient and Ambient + Spike Measurements.
SOURCE: Environmental Monitoring Systems Laboratory, 1979.
2-51
-------
A statistical comparison of the hourly averages from each test analyzer
with the average of the hourly averages (for corresponding hours) from the
other test analyzers is presented in Table 2-10. Each test analyzer is
identified in the table by manufacturer, model number, and measurement principle.
The data clearly indicate that these continuous S02 analyzers are capable of
excellent performance (high correlation with one another, small mean differences)
Two scatter diagrams (for Philips coulometricjpd Melcni' flame^)hptometric
analyzers) are shown in Figures 2-3 and 2-4.y\ TnefTgures al^o indicate the
absolute relationship of the subject analyzers to the average of the other
test analyzers and are fairly typical of all the analyzers tested.
2.3.6 Summary
Methods for the measurement of sulfur dioxide can be classified as
integrated methods, which involve collection of the sample over a specified
time period and subsequent analysis by a variety of analytical techniques, or
continuous methods, in which sample collection and analysis are performed
constantly and automatically.
The sampling procedures used in integrated methods can be direct (e.g.,
collection in an inert container) or based on techniques involving absorption
and stabilization in aqueous solution (e.g., collection in dilute hydrogen
peroxide) or adsorption and stabilization on a solid substrate (e.g., collection
on silica gel). Techniques used for the analysis of the collected sample are
commonly based on colorimetric, titrimetric, turbidimetric, gravimetric, and
ion-chromatographic measurement principles.
The most widely used integrated method for the determination of atmospheric
sulfur dioxide is the pararosaniline method developed by West and Gaeke. An
improved version of this colorimetric method, adopted as the EPA reference
2-52
-------
TABLE 2-10. COMPARISON OF EPA DESIGNATED EQUIVALENT METHODS FOR S02 (CONTINUOUS ANALYZERS)
Measurement
principle
Correlation
coefficient
Mean
difference
ppb
Std. dev.
of diff.
ppb
Max. abs.
diff.
ppb
No. of abs.
diff.
>20 ppb
No of data
pai rs
Meloy
SA185-2A
Flame
photometry
0.999
-3.695
3.925
19.3
0
3302
Monitor Labs Meloy
8450 SA285E
Flame Flame
photometry photometry
0.999 0.999
-0.006 -0.251
4.555 3.243
19.2 15.5
0 0
3186 3306
Thermo Electron Beckman Lear Siegler Philips
43 953 SM 1000 PW 9755
Second
Derivative
Fluorescence Fluorescence spectrometry Coulometry
0.997 0.998 0.936 0.998
-0.177 5.108 4.924 5.775
8.300 6.901 20.712 4.631
29.9 25.4 100.9 25.6
49 21 427 13
2170 1594 1820 3070
Bendix
8303
Flame
photometry
0.998
-3.278
4.392
21.0
1
1984
rn
33
O
D
T|
Z
T|
-4
33
H
Z
n
D
/•)
n
ppTFn
i
5
£
SOURCE: Environmental Monitoring Systems Laboratory, 1979.
-------
PHILIPS PW9755
r\s
en
400--
m
o.
. 300
en
LJ
M
< 200 ••
CJ
LU
m
C/> 100
0
N= 10.0000
s
X"
'•s
xX
..
SLOPE
1. 0050
INTERCEPT
5. 4250
100 200 300 400
REFERENCE ANALYZER AVERAGE. PPB
500
2-3. Relationship Between Phillips Coulometric Analyzer and Average
of-reference method analyzers.
-------
MELOY SA285E
fO
I
en
500 T
400--
co
Q.
QL
. 300--
a:
LJ
< 200--
CJ
UJ
i
m
en 100..
0
0
N= 10. 0000
SLOPE
1. 0157
INTERCEPT- -0. 8039
h
100 200 300 400
REFERENCE ANALYZER AVERAGE, PPB
500
2-4. Relationship between Meloy flame photometric analyzer and average
of reference method analyzers.
-------
method in 1971, is capable of measuring ambient S02 concentrations as low as
25 [jg/rn3 (0.01 ppm) with sampling times ranging from 30 minutes to 24 hours.
The method has acceptable specificity for SO-, but collected samples are
subject to a temperature-dependent decay which can result in an underestimation
of the ambient S0~ concentration. Temperature control during sample collection,
shipment, and storage effectively minimizes this decay problem.
A titrimetric method based on collection of SO- in dilute hydrogen peroxide
followed by titration of the resultant HUSCL with standard alkali is the
standard method used extensively in Great Britain. Although simple to perform,
the method requires long sampling times (24-hours) and is subject to interference
from atmospheric acids and bases.
One of the most sensitive methods now available is a promising new technique
that uses ion chromatography to determine ambient levels of SOp which have been
absorbed into dilute hydrogen peroxide and oxidized to sulfate. This method
does not distinguish between sulfur dioxide and sulfate actually present in
the ambient air unless sampling is conducted in such a way as to separate
these two sulfur species. The course of research suggests that ion chromatography
will probably become one of the most useful analytical evaluation methods for
the sequential determination of both sulfur dioxide and sulfate.
Sulfation methods, based on the reaction of S0? with lead peroxide paste
to form lead sulfate, have commonly been used to estimate ambient S0? concen-
tration over extended time periods. After exposure, the collection device
(lead candle or sulfation plate) is treated with sodium carbonate solution to
dissolve the lead sulfate, which is then determined gravimetrically or
turbidimetrically. The accuracy of sulfation methods is subject to many
physical and chemical variables and interferents. Sulfation rate (ing SCL/100
2-56
-------
2
cm /day) is commonly converted to 50^ concentration (ppm) by multiplying the
rate by the Huey factor (0.03).
Continuous methods for the measurement of ambient levels of sulfur dioxide
have gained widespread use in the air monitoring community. Some of the
earliest continuous SOp analyzers were based on conductivity and coulometry.
These first generation analyzers were subject to interference by a wide variety
of substances present in typical ambient atmospheres. However, the more
recent commercially available analyzers using these measurement principles
exhibit improved specificity for 502 through the incorporation of sophisticated
chemical and physical scrubbers. The early continuous colorimetric analyzers
using West-Gaeke type reagents and having high sensitivity and acceptable
specificity for S0? were wrought with a variety of mechanical problems, required
frequent calibration, and thus never gained widespread acceptance.
Continuous sulfur dioxide analyzers using the techniques of flame
photometric detection, fluorescence, and second derivative spectrometry have
been developed over the past ten years and are commercially available from a
number of air monitoring instrumentation companies.
The flame photometric detection of ambient S02 is based on the measurement
*
of the band emission of excited S? molecules formed from sulfur species in a
hydrogen-rich flame. The FPD analyzers exhibit high sensitivity and fast
response, but must be used with selective scrubbers or coupled with gas
chromatographs when high specificity is desired.
Fluorescence analyzers are based on detection of the characteristic
fluorescence of the S0? molecule when it is irradiated by UV light. These
analyzers have acceptable sensitivity and response times, are insensitive to
sample flowrate, and require no support gases. They are subject to interference
2-57
-------
by water vapor (due to quenching effects) and certain aromatic hydrocarbons,
and therefore must incorporate a means to eliminate or minimize these species
or their effects.
Second derivative spectrometry is a highly specific technique for the
measurement of SCL in the atmosphere and a continuous analyzer based on this
principle is commercially available. The analyzer is insensitive to sample
flowrate and requires no support gases. Sample flowrates of 4 liters/min are
required to achieve reasonable response times, and frequent realignment of the
optics is necessary when the analyzer is used under typical field conditions.
Analyzers based on many of the above measurement principles (conductivity,
coulometry, flame photometry, fluorescence, and second derivative spectrometry)
have been designated by EPA as equivalent methods for the measurement of SO-
in the atmosphere. Testing of these analyzers by the manufacturers prior to
designation have demonstrated adequate performance for use when an EPA reference
or equivalent method is desired or required. Testing of these methods by EPA
has verified their performance and has also demonstrated excellent comparability
between these designated methods under typical monitoring conditions.
2.4 MEASUREMENT TECHNIQUES FOR SULFATES
2.4.1 Introduction
This section discusses methods for determining sulfates in airborne
particulates. The methodology involves collecting particulate matter by
various filtration techniques and subsequently analyzing for soluble sulfates,
total sulfur or specific sulfate species such as sulfuric acid. The methods
are discussed with primary emphasis on their applicability and limitations.
Specific details of each method can be obtained from the referenced articles;
techniques for the collection of particulates are covered in section 2.5.
2-58
-------
2.4.2 Total Soluble Sulfate
Although the gravimetric determination of sulfate has been a laboratory
experiment in elementary quantitative analysis courses for over 50 years, the
determination of sulfate in aqueous solution at concentrations of 1 |jg/ml or
below remains a tedious and nontrivial task. A comprehensive and up-to-date
review of wet chemical methods has recently been compiled by Hoffer and Kothny
(1974). The principal methods for determination of trace sulfate that are
applicable to aqueous leachates of airborne particle samples are reviewed
here.
2.4.2.1 Direct Methods Via Sulfate Precipitation—Accurate analyses for
sulfate in aqueous solution are still performed by preconcentration and barium
Jb^Zf^zj^^iJij,
sulfate gravimetry (Methods of Air Sampling and Analysis-, 1972). Simplified
procedures for small amounts of sulfate samples at lower concentrations have
accrued from measurement of light-scattering properties of colloidal precipitates
of barium sulfate. Both nephelometric (90-degree scattering) and turbidimetric
(180-degree source attenuation) measurement techniques for sulfate have been
developed (Kolthoff et al., 1969). Other advances include automation of the
turbidimetric procedure and its extension to organic sulfates. A generally
accepted lower detection limit for an automated turbidimetric sulfate
determination is 3 pg/ml aqueous extract. Any source of turbidity other than
colloidal sulfate (e.g., other anions forming insoluble barium salts) will
positively interfere with the determination (Technicon Corp., 1959).
Precipitation of radioactively labeled BaSO. with addition of a surfactant
allowing complete removal of BaCl^ has been reported (Dimitt and Graham,
1976). This procedure appears to be useful for sulfate concentrations as low
2-59
-------
as 1 (jg/ml. A related technique for sulfate analysis by BaSO^ determination
has been reported by Bogen and Welford (1977). This technique uses standard
35
addition of high specific activity S-sulfate to monitor the degree of
precipitation, and exceptionally low concentrations of sulfate (to 0.01 M9/ml)
may be determined, apparently with no significant interferents except lead
(II).
A popular direct titrimetric method for sulfate is titration with millimolar
barium ion with thorin indicator and visual or spectrophotometric endpoint
detection. Optimization of the thorin method has been reported, and a recommenda-
tion of dimethysulfonazo III indicator as best for visual endpoint detection
has been made by Scroggins (1974). Recently, Brossot and FornV (1978)
have described a modification of the thorin method that permits rapid determina-
3
tion of sulfate in aqueous leachates with a limit of detection of <0.2 ug/m .
2.4.2.2 Indirect Methods Via Sulfate Precipitation—Several indirect analytical
methods based on the exchange of a more easily determined anion bound to
barium with sulfate ion have been devised. Mixtures of barium ion and chloranilate
or iodate ion are added to the aqueous sulfate solution with the exchanged
chloranilate ion determined spectrophotometrically or the exchanged iodate ion
determined by redox conversion to iodine (Bertolacini and Barney, 1957; Klockow
and Ronicke, 1973). The precision of the more sensitive and widely used
chloranilate method has been improved, and the method has been adapted for
automated analysis (Gales et al., 1968).
Hoffer and Kothny (1974) have reported a dye-exchange method illustrated
by the following equation:
Ba(dye) + S042" •» BaS04 + dye
2-60
-------
in which the reduction of absorbance of the barium-nitrochromazo dye complex is
measured spectrophotometrically in acetonitrile solvent. As little as 3.5 ^jg
of sulfate may be determined by this method. A more sensitive and widely
used application of dye-exchange reactions to sulfate analysis was reported by
Lazarus et al. (1966) and uses the metallochromic indicator methyl thymol blue
(MTB). In the refinement of the MTB method reported by Adamski and Villard
2+
(1975), equimolar Ba and MTB in aqueous ethanol are mixed with sulfate
solution at about pH 3 to precipitate BaSO,; then, after adjustment of pH to
about 12.5, released MTB is measured colorimetrically. An intercomparison
study sponsored by EPA demonstrated that the MTB method is superior in range
and the most reliable of the methods tested for long-term, low-volume sampling
(Appel et al., 1976). A related technique has been reported in which sulfate
is precipitated with 2-aminopyrimidine and the excess reagent is determined
colorimetrically. Conversion of the amino reagent to a stronger absorbing
derivative by treatment with nitrate/base has reduced the lower detection
limit of this method to about 1 [jg/ml sulfate based on 0.03 absorbance change
(Archer, 1975).
2.4.2.3 Direct Enthalpimetric Methods—Thermometric methods are potentially
applicable to the determination of several oxyanions of environmental interest,
(%
-------
The direct addition of concentrated BaCl2 results in a temperature rise of
about 0.0032°C/umol, from which, by reference to a standard curve, the amount
of sulfate in solution may be determined.
2.4.2.4 Ion-exchange Chromatography Methods--Ion-exchange resins can be used to
provide excellent separation of ions but, until recently, the continuous
determination of ionic species as they are eluted from the ion-exchange column
has been difficult, due principally to the presence of the electrolyte solution
used as eluant. Recently, Small and co-workers (1975) have devised a novel
combination of resins which effectively neutralize the eluant and suppress its
conductivity without affecting the analyte, thus permitting the use of a
conductivity cell as a monitor for microgram levels of either cationic or
anionic species. A commercial instrument based on this system is now available
for use in trace anion analysis including sulfate (Dionex Corp., 1975). In
this system, a strong-base anion exchanger of low capacity, agglomerated onto a
surface-sulfonated DVB resin, is used as the analytical column. This is followed
by a high-capacity, strong acid exchange column which converts the eluant,
typically 0.003 M Na2C03 + 0.024 M NaHC03, into a non-conducting high-sensitivity,
multirange conductivity meter.
Stevens et al. (1978) have described the application of ion exchange
chromatography to analysis of sulfate, sulfite and nitrate from aerosols
collected on teflon filters with dichotomous samplers at several urban areas
in the United States. Because the ion chromatograph method is unambiguous,
exceptionally accurate, sensitive (<1 pg/ml), and applicable to determination
of several anions in the same sample, it may become the preferred method for
sulfate in airborne particulate samples. The only significant disadvantages
2-62
-------
are the relatively high initial capital cost of the equipment and the relatively
high level of operator technique required to fully realize the method's advantages
2.4.2.5 Flash volatilization-FPD Methods—The continuing need for more
sensitive determination of sulfate in large numbers of aerosol samples has led
to the development of an analytical technique based on flash pyrolysis and
flame photometric determination (FPD). This technique was first devised by
Roberts and Fried!ander (4£^4% 1976), who collected samples on stainless-steel
impactor slides, then passed a capacitative discharge through the slide heating
it to >1000°C and converting aerosol sulfate to SOp for subsequent detection
by FPD. Filter samples and evaporated extracts transferred to the slides were
also analyzed, but the short lifetimes of the slides limited the utility of
the method. Modifications were subsequently made by Husar et al. (1975,
1976), who substituted a tungsten boat for the stainless-steel slide and obtained
precise recoveries exceeding 90 percent for sample concentrations of 1 to 10
ug/ml S with boat lifetimes <100 discharges. Refractory sulfate compounds not
volatile at about 1100°C are not determined.
Other modifications of the method have been reported, including using
sustained heating of evaporated samples at 500°C instead of an instantaneous
discharge (Tanner et al., 1977). This approach avoids problems of recovery
from boats, but in effect analyzes only those compounds, HUSO., (NH.^SO,,
CuSO., and a few mixed ammonium salts, which decompose at the chamber
temperature used. Another modification has been reported by Tanner et al.,
M
(1977a), who introduced the use of platinum sample containers and, with use of
reduced chamber and transfer line volumes and a modified FPD, have achieved
reliable determinations of <0.5 ug/ml sulfur. However, since the method does
2-63
-------
not distinguish between sulfate and other water-soluble sulfur compounds, does
not determine refractories such as K2SO., and is subject to irreproducible
recoveries when certain constituents are present in the extract, applications
of the method may be limited to relatively clean matrices where very low
extract volumes are feasible or required.
2.4.3 Total Aerosol Sulfur
Several techniques for determination of total sulfur in airborne particle
samples will be discussed in this section. It is generally found that nearly
100 percent of aerosol sulfur mass is present in the form of sulfate, and most
of the data on airborne sulfur concentrations can be accurately described as
total sulfur calculated as sulfate or total soluble sulfate.
2.4.3.1 X-Ray Fluorescence—One of the most promising techniques under
development in recent years for analyzing elements including sulfur in airborne
particles is X-ray flourescence (Goulding and Jaklevic, 1973). The
sample, collected on an appropriate filter, is irradiated with X-rays, gamma
rays, protons, or other charged particles, and the intensity of X-ray fluores-
cence induced is measured as a function of wavelength or energy to determine
the amounts of the constituent elements present. Qualitative and quantitative
analysis can be obtained when the system is properly calibrated for each
element. This calibration step is difficult, since few standards of known
r
elemental composition are available in disks of known thickness in an appropriate
matrix (Adams and Van Grieken, 1975).
The most extensive set of aerosol sulfur data were reported by Stevens et
al. (1978) and Loo et al. (1978) using a nondispersive x-ray fluorescence
spectrometer designed by Goulding and Jaklevic (1973). Stevens et al. (1978)
2-64
-------
reported sulfur and 18 other elements from dichotomous samplers operated in
New York City, New York, Philadelphia, Pennsylvania, Charleston, West Virginia,
St. Louis, Missouri, Portland, Oregon and Glendora, California. Loo et al.
(1978) reported sulfur concentrations collected over a 2-year period from a
network of 10 automated dichotomous samplers operated in St. Louis, Missouri
during the Regional Air Pollution Study (RAPS). They reported a detection
2
limit of 0.034 ug/cm of filter, which corresponds to a concentration value of
3
<0.1 ug/m sulfur for a 2-hr sample collected at 50 liters/minute on a 37-mm
filter and is adequately sensitive for a 1-hr time discrimination at ambient S
levels. The advantages of proton-induced XRF--lower bremsstrahlung background
and higher secondary X-ray cross-sections—may lead to its preferential use
when short-time resolution of ambient S levels is desired (Johansson et al.,
1975; Lochmuller et al., 1974). A related approach to nondestructive aerosol
sulfur analysis based on cyclotron in-beam approach gamma-ray spectroscopy has
been reported by Macias (1977). Gamma rays induced by proton or a-irradiation
are detected by a Li drifted, Ge detector and used to determine S and other
light elements such as Mg and C in aerosol samples. It is clear that induced
X-ray fluorescence methods will continue to be important tools in determining
total sulfur in large numbers of ambient aerosol samples.
2.4.3.2 Electron Spectroscopy for Chemical Analysis (ESCA)--The ESCA method
utilizes measurement of the kinetic energy of photoelectrons ejected from a
sample irradiated with monochromatic X-rays. The kinetic energy, KE, of an
ejected photoelectron is given by the equation:
KE = hv + Ei
where hv is the X-ray photon energy and E- is the binding energy of electrons
in the orbital subshell of the bombarded atom. Since these binding energies
2-65
-------
are characteristic of atoms of a particular element, the kinetic energy spectrum
can be used to identify kinds and amounts of elements present. Novakov (1973)
and Novakov et al. (1974) have pioneered in the use of this technique for
analysis of sulfur in airborne particle samples.
The most common X-ray photon sources used in ESCA, i.e., Mg or Al ka X-
rays, have relatively low energies and resultant sample penetration depths of
only 25 A. Thus, this method is particularly sensitive to the surface composition
of samples, which is an advantage for surface-oriented studies but a handicap
if bulk analysis is desired for samples whose elemental composition is likely
to be nonhomogeneous.
2.4.3.3 Direct Flame Photometric Detection—The flame photometric detector
(FPD) is based on the measurement of the band emission of excited $2 molecules
(sulfur mode) or HPO molecules (phosphorus mode) during passage of S- or
P-containing compounds through a hydrogen-hyperventilated (reducing) flame.
It has been widely used an an S,P-specific detector in gas chromatography and
in the ambient monitoring of gaseous sulfur compounds (Lucero and Paljug,
lf#
1973-; Newman et al., 1975 a and b). Using filtered ambient air as the flame
oxidant, commercially available instruments can routinely monitor as little as
2 ppb (v/v) of gaseous SO,,.
Application of the FPD to the detection of aerosol sulfur (as sulfate)
was first made by Crider et al. (1969), but its use in ambient aerosol sulfur
monitoring has been limited by three considerations: sensitivity, selectivity,
and specie-variable response. Ambient levels of sulfur range from 1 to 50
3
ug/m as sulfate (equivalent to 0.2 to 10 ppb SO^); hence the conventional FPD
has not been sensitive enough for the lower range of sulfur concentrations.
However, the use of digital signal-averaging techniques can reduce the
2-66
-------
signal-to-noise ratio by a factor of 10 and, used with recently introduced
commercial instrumentation in which noise reduction through hydrogen flow
regulation is substantial, a lower limit of detection for H?SO. of 0.25 ppb
o
(1.1 ug/m ) is now obtainable.
Since the flame photometric detector responds to all sulfur compounds, it
is necessary to remove gaseous sulfur compounds (principally S0?) in order to
monitor only particulate sulfur. Several groups have shown that laminar flow
of sample air through a tube whose inside walls are an efficient sink for S02
will reduce the S0? concentration (by "diffusion denuding" the air) to less
than 1 percent of ambient levels without significant loss of 0.1 urn-diameter
particles (Tanner et al., 1978). Further investigation of appropriate wall
coatings and sample flows for diffusion-denuding of SOp and concomitant
continuous particulate sulfur monitoring is in progress.
An alternate approach to ambient particulate sulfur monitoring using an
FPD has been reported by Kittelson et al. (1977). This system combines the
FPD with a pulsed electrostatic precipitator and a lock-in amplifier tuned to
the pulse frequency of 0.2 Hz. Since it is reported that particulate sulfur
in the size range 0.03 to 1.0 urn is removed with essentially 100 percent
efficiency, the signal output from the FPD contains a DC component resulting
from the sulfur gases (not affected by the precipitator) and an AC component
dependent on the particulate sulfur concentration which can be selectively
amplified by the lock-in amplifier. Aerosol sulfur concentrations at ambient
levels may be measured in the presence of relatively steady SCL concentrations
by this method, although the accuracy and specificity are not well established.
Since the response of the FPD varies depending on the chemical form of
aerosol sulfate (or reduced forms of sulfur), a calibration based on auxiliary
2-67
-------
speciation of the sulfate or thermal pretreatment of the aerosol may be necessary
for accurate determinations (Huntzicker et a!., 197^ 1977).
2.4.4 Specific Sulfate Species
2.4.4.1 Speciation by Thermal Volatilization Techniques—The determination of
sulfuric acid by its selective thermal volatilization from filters was first
reported by Scaringelli and Rehme (1969). Particulate samples collected on
glass fiber filters were heated to 400°C, and the volatilized H^SO^ (determined
by three different techniques) was calculated as percent of total sulfate,
which varied from 20 to 100 percent in samples from several urban areas. In
this procedure NH.HSO., (NH.)pSO., and probably CuSO. were also volatilized.
Thus, the reported HUSO, fractions more accurately indicate HUSO, and its
ammonia neutralization products as percent of total sulfate. Dubois et al.
(1969) concurrently developed a microdiffusion method for HUSO, which involved
thermal volatilization of HUSO, at 200°C from a filter sample in a closed
petri dish and its subsequent immobilization on the dish cover, which had been
coated with NaOH. The immobilized HUSO, chemically present as Na-SO., was
removed from the cover and determined by the thorin method. Reproducibility
of the method was adequate only for standard samples, and a potential inter-
ference from ammonium sulfate was reported. Reported values for HUSO, as
percent of total sulfate were about an order of magnitude lower than values
reported by Scaringelli and Rehme (1969).
The microdiffusion method was modified by Maddalone et al. (1975), who
jk
used perimidylammonium bromide as the immobilizing agent for HUSO, volatilized
from filters at 125°C. The perimidylammonium sulfate is decomposed quantita-
tively at 400°C, and the released S02 is determined by the FPD or the West-
Gaeke method. Low recoveries (based on H2S04 samples) were reported for glass
fiber filters, but >85 percent for samples containing 10 pg of HUS04 was
recovered from Teflon filters (Mitex and Fluoropore). Neither Scaringelli
2-68
-------
and Rehme nor Maddalone was consistently successful in removing H^SO. from
ambient particulate-loaded filters by thermally induced microdiffusion.
Benarie and coworkers (1973) have applied the microdiffusion technique of
Dubois during an air sampling campaign around Rouen, France, from which it was
found that 7 percent of airborne sulfur was in the form of HLSO..
The peri midylammonium bromide (PDA-Br) method, without the microdiffusion
step, has been suggested by Thomas et al. (1976) for HUSO., sulfur trioxide,
and total soluble sulfate determinations in the 10-ug sulfur range. Sulfuric
acid droplets are stabilized at the time of collection on PDA-Brimpregnated
glass fiber filters, preventing topochemical reactions of H^SO. with previously
collected particles. The resulting perimidylammonium sulfate (PDA)2(S04) is
then thermally decomposed at 400°C to produce S0?, which is determined by the
West-Gaeke method. The authors report that the impregnated glass fiber filters
are approximately 90 percent more efficient for H?SO» aerosol than Fluoropore
filters and are much easier to work with. Ammonium sulfate (and presumably
ammonium bisulfate) interferes with the method by large amounts, a serious
disadvantage since these sulfates are frequently present in excess of H?SCL in
ambient aerosol samples.
A more sophisticated thermal speciation technique is the controlled
pyrolysis method reported by Leahy et al. (1975). Sulfuric acid samples (both
evaporated solutions and laboratory-generated aerosols) were removed from
®
Teflon filters or quartz boats at 190°C, ammonium bisulfate samples likewise
at 275°C, and sodium bisulfate samples at 325°C by passing premoistened N?
through a heated quartz tube. The volatilized samples were rinsed from the
tube after each temperature increase and analyzed for sulfate by turbidimetry.
Experiments duplicating these results with mixed aerosols generated in a 90-m
2-69
-------
environmental chamber were less successful: only 60 to 70 percent of 100-pg
H?S04 samples were recovered, and nonreproducible recovery of H^SO^ from
NaHSO. decomposition at 325°C was observed. This was due principally to
reactions between sulfuric acid and particulate matter and emphasizes the
extreme difficulty inherent in transferring sulfuric acid in the presence of
airborne particles in any system at elevated temperatures. Although not
reported, ammonium sulfate should be decomposed and transferred with ammonium
bisulfate. Subsequent work with dry Np in the volatilization procedure did
not improve the collection efficiency (Tanner and Cordova, 1978).
The most successful approach to thermal volatilization of H^SO, in ambient
aerosol samples was reported by Mudgett et al. (1974). Aerosol samples are
collected on Fluoropore filters, the H?SO. is subsequently volatilized by
passage of heated (ca. 150°C), dry N~ in the reverse direction through the
filter, and released ^SO. is determined with a flame photometric detector.
Laboratory aerosol samples of as little as 0.25 pg HpSO. may be determined
with reasonable precision (Lamothe and Stevens, 1976). A prototype,
semi automated instrument designed with monitoring applications in mind has
been developed under EPA contract (Harris, 1975). Although no results have
been published, the prototype instrument has been tested in several field
monitoring applications. Based on the experience of other workers in removing
H2S04 from particulate filters at elevated temperatures, difficulty in quantita-
tively removing H2$04 has been encountered. In fact, serious difficulties are
encountered in removing H2S04 quantitatively in the presence of ammonium
bisulfate: H2$04 is totally removed at 180°C, but NH4HS04 is also partially
volatilized; at 140°C, NH4HS04 is not volatilized but H2S04 is incompletely
volatilized. Finally, the use of Fluoropore filters with attendant low flow
2-70
-------
3
rates (1 m /hr) will require sampling times of 1 hour or more for most ambient
applications.
2.4.4.2 Speciation by Solvent Extraction—A solvent extraction procedure to
selectively remove collected H^SO. aerosol in the presence of other aerosol
sulfates was first reported by Barton and McAdie (1971). They concluded that
aerosol collection on Nucleopore filters followed by extraction with isopropanol
for subsequent analysis by the chloranilate procedure was selective for airborne
H?SO«. Subsequent work reported reduction of interference by buffer control
of the isopropanol extract and development of an automated instrument for the
extraction procedure (Barton and McAdie, 1973). Leahy et al. (1975) have
shown, however, that isopropanol will also extract ammonium bisulfate quantita-
tively and partially extract other bisulfates. Since ammonium bisulfate is a
documented constituent of ambient aerosol, isopropanol cannot be considered as
a selective extractant for H^SO It has been demonstrated that benzaldehyde
is a selective extractant for H^SO, in the presence of bisulfates and sulfates.
Subsequent radiochemical experiments with H?SO. have established that HLSO.
may be reproducibly removed from a variety of filter media (Mitex, Fluoropore,
H3PO.-treated quartz) with recoveries varying from 75 percent for 10-ug H^SO.
samples to 95 percent for 100-ug samples. This procedure may thereby be used
to quantitate the presence of H?SO« in ambient particulate samples with time
resolution of 1 hour or less.
2.4.4.3 Speciation by Gas-Phase Ammonia Titration--A method has been proposed
for determination of strong acid aerosol (assumed by the authors to be H^SO.)
by gas-phase reaction with ammonia in excess from a permeation source (Dzubay
et al., 1974). The reduction in ammonia concentration produced by reaction
with acid aerosol is determined by conversion to NO over gold wool at 1000°C
2-71
-------
and measurement by an 03 chemiluminescent NO monitor. The method as reported
has inadequate sensitivity (5 ppb NH3 = 10 ug m3 H2$04) for ambient monitoring
but with state-of-the-art NO monitors, a lower detection limit of 2 ug/m
HpSO. would be possible. The method assumes that other than I^SO^ strong
acids reacting with NH~ are absent from ambient aerosol sample; it also requires
complete reaction of NH3 with acid aerosol particles and complete conversion
of unreacted NH3 to NO.
A more useful, albeit still qualitative, approach to sulfate speciation
in aerosol particles by gas-phase NH3 titration has been developed and widely
used by Charlson et al. (1974 a and b). An apparatus called a humidograph
measures the dependence of the light-scattering coefficient (bscat) °f airborne
particles as a function of relative humidity (RH). Particles consisting
mainly of hygroscopic materials such as HUSO, or NH.HSO, exhibit a monotonic
plot of b . versus RH (the plot is called a humidogram). The presence of a
deliquescent substance as a major constituent is denoted by a rapid rise in
b , to an inflection point in the humidogram after a relatively flat curve
at lower RH. For example, (NH.^SO. in aerosol particles exhibits an inflection
point at about 80 percent RH. The humidograph may also add ppm quantities of
NH3 before recording the humidogram. Production of (NH,)?SO. from the addition
of NH3 to acid sulfate aerosol can be detected by appearance of the inflection
point near 80 percent RH if at least 30 to 50 percent of the aerosol mass is
present as sulfate.
The humidographic technique has the advantages of i_n situ measurement
with short time resolution (15 min). Particle growth into the optical-scattering
region during recording of the humidogram does not appear to be a serious
problem. There is a serious interference from aerosol nitrate since most
2-72
-------
nitrate salts (including NH.NO.J are hygroscopic. The original method does
not distinguish between H^SO. and ammonium bisulfate, and attempts to do so by
stoichiometric NH~ gas phase titration have only been marginally successful.
2.4.4.4 Speciation by Cation Measurement—Since atmospheric sulfate can be
associated with various cations, the compounds of sulfate can sometimes be
characterized by measuring the cation. Brosset and Perm (1978) and Stevens et
al. (1978) describe a Gran titration procedure for H and a procedure using an
ion selective electrode for NH. in aqueous extracts of aerosols collected on
Teflon filters. Stevens et al. (1978) applied such techniques to aerosols
collected at Research Triangle Park, North Carolina during the summers of 1977
and 1978 and found a stoichiometric balance between S0.~ and the sum of H and
NH. ion concentrations. The acidity was found to range from none (ammonium
sulfate) to that of ammonium bisulfate.
2.4.4.5 Infrared Spectroscopy--The use of infrared (IR) spectroscopy for
identification of species such as ammonium and sulfate in airborne particles
was first described by Blanco and coworkers (1968, 1972). Anderson samplers
were used to obtain size-segregated samples with each fraction analyzed by an
IR microspectrophotometric technique after incorporation into a KBr pellet.
Bands attributed to ammonium and sulfate dominated the submicrometer fractions
in urban samples but were not found in desert samples.
In more recent work by Cunningham and Johnson (1976), data were reported
from an "acid sulfate" episode during the fall of 1973 in Chicago. Successive
3-hr samples from Lundgren stage four showed reduced sulfate band intensity
and appearance of a series of bands at 1205, 1063, 869, and 600 cm (and
other weaker bands), which the authors attributed to changes in the acid
content of the samples with bands assigned to bisulfate and "sulfate more
2-73
-------
acidic than bisulfate." Although mixed-acid sulfate phases are known [e.g.,
(NH4)3H(S04)2-lavoviciteJ, such complexities do make more difficult the
quantitative assessment of the degree of acidity in submicrometer sulfate-
containing particles by IR. Application of IR analysis to quantification of
other aerosol constituents, especially polynuclear aromatic hydrocarbons, is
an especially fruitful area of investigation (Mamantov et al., 1977).
2.4.4.6 Comp1exation-Spectrophotometry--Huygen (1975) has described a specific
method for HUSO, whereby dry diethylamine is passed through the aerosol sample
after collection on a teflon filter, immobilizing any HUSO, present as the
bis(diethylammonium) salt. The latter is reacted with carbon disulfide and
cupric ion to form the bis-(diethyldithiocarbamate)-copper(II) complex, which
may be colorimetrically determined for sulfuric acid samples from 25 to 100
ug. Major interferences are reported for phosphoric acid and ammonium bisulfate
under anhydrous conditions and for most weak acids (including ammonium sulfate)
under moist conditions. Tanner and Cordova (1978) have used the Huygen method
to produce linear calibration curves for 25 to 100 ug hLSO., but even under
anhydrous conditions a 70 mol percent interference by ammonium bisulfate was
observed. The interference problems make this method of marginal value for
specific H2S04 analysis, although it might be useful as a method for estima-
tion of "total acidic sulfate" in airborne particles.
Dzubay et al. (1979) recently developed a new approach to this method.
®
Aerosol samples collected on Teflon membrane filters were analyzed by exposing
14 +
them to C labeled trimethylamine. The amount of acidity (H ) in the samples
is determined by counting the radioactivity and comparing the result with that
of a standard. No interferences were observed due to ammonium sulfate or
carbon. For replicate samples collected at Research Triangle Park, North Carolina
2-74
-------
during the summer of 1979, the results from the radiochemical method were
found to be equivalent to those by Gran titration.
2.4.4.7 Heated Filter Co11ection/FPD--Benarie et al. (1976) have described
®
a method for determining HUSO^ by a different technique. A Teflon filter
heated to ca. 70°C was found to pass HUSO, reasonably quantitatively to a
flame photometric detector (FPD), where it was measured along with S0? and
any other gaseous sulfur compounds, while other aerosol sulfate compounds are
collected by the filter. By comparing this FPD output with that for an unheated
FeSO.-impregnated glass fiber filter (where HUSO, was collected as aerosol
sulfate), the HUSO, content can be determined. This technique was used for
HUSO, determinations during the air sampling campaign around Rouen, France.
3
Concentrations varied over the range from zero to 13.3 ug/m during winter and
summer seasons. It was concluded that oxidation rates of SCL to sulfate
varied from 0.9 percent to 7.5 percent/hr depending on the amount and composi-
tion of suspended particles. Sulfuric acid, averaging 3 percent of total
3
atmospheric sulfur, was measured with a precision of ± 0.9 ug/m HUSO This
promising technique has not been confirmed by other workers, and it is difficult
®
to explain the observation that HUSO, aerosol is not collected on Teflon at
70°C when other workers report difficulty in removing HUSO, from filters
containing airborne particles at 150°C (Lamothe and Stevens, 1976). Figure 2-5
shows general effects of spurious sulfate formation on filter, while Figure 2-6
shows the effects of filter temperature on sulfate distribution.
2.4.4.8 Electron Microscopy and Related Techniques—The use of electron
microscopy for identification and semi-quantitative determination of submicrometer
ammonium sulfate aerosol particles has been demonstrated by Heard and Wiffen
(1969). The techniques for preparing samples collected on membrane filters
2-75
-------
§001—
ac
u
GLASS —
NUCLEPORE.,
MILLIPORE
SILVER
POLYCARBONATE.
QUARTZ
TEPLON
10
20
30
TIME, hours
Figure 2-5. The affocts of Curious sulfsta formation on dan fil-
ters is grvwi for extra™ exposures (4000 ppm SO2 and 100% r.h.)
Sourta: Loo «t at. (1978).
2-76
-------
•0
u
i
o
o
u
UJ
in
O
5
£
c
H-
to
O
70
90
so
»
10
ARTICUUATESULFATE
FILTER
0 t3 150 20S
FILTER TEMPERATURE. °C
Figur* 2-6. Effects of filter t»mp*nturt on «ulfrt« dirtribution
and Homofyi (1979).
Soorca:
2-77
-------
and the electron microscopic determination itself are tedious and time consuming,
but this method does provide an unambiguous identification of a specific
aerosol sulfate species. The collection of acid aerosol particles on various
types of films and the counting of number and size of reactive spots with the
aid of an electron or optical microscope has been reported by several groups,
including Lodge et al. (1960). Differentiation of aerosol particles consisting
mainly of (NH.^SO. or H^SO. by electron microscopic examination of the particles
collected on BaC-coated carbon grids has been detailed by Mamane and dePena
(197f). This procedure is not highly specific for HpSO./HSO..
2.4.4.9 Acid-Base Indicators—A method has been described for spectrophoto-
metrically determining acidity in airborne particles from the displacement of
equilibrium and resultant color change of acid-base indicators (Whittaker Co.).
The collected sample is washed from the filter with bromophenol solution, and
the change in absorbance at 587 nm is noted. The method determines acidity
nonspecifically, and its value of assessing the amount of airborne acid sulfate
in the absence of other particulate species is quite limited.
2.4.5 Ammonium and Gaseous Ammonia Determination
An important supplementary measurement aiding speciation of sulfate in
airborne particles is the measurement of ammonium ion. Ammonium ion is found
predominantly in the optical-scattering size range or below and is presumed to
be secondary in origin, being formed in the neutralization of acidic sulfate
particles. The high correlation of ammonium content with soluble sulfate in
both urban and rural aerosol samples and the identification by X-ray diffrac-
tion of (NH4)2S04 in dried aqueous extracts of airborne particles would tend
to confirm the above hypothesis (Brosset et al., 1975).
2-78
-------
Ammonium ion in particulate matter is nearly universally determined by
collection on filters, extraction into an appropriate leach solution, and
determination by one of the two methods described below. The first is a
concentration measurement by an ion-selective electrode sensitive to either
ammonium ion (Beckman electrode) or ammonia (Orion or Markson electrodes).
The limit of detection is determined by the equilibration time of the electrode,
a representative value being 5 to 7 minutes for 20 ppb ammonium concentration
(Eagan and DuBois, 1974; Gilbert and Clay, 1973). This is marginally sensitive
for high-volume samples of rural ambient air where [NH.] may be as low as 0.3
ug/m . A more recent development, the air gap electrode, eliminates the problems
of electrode contamination by sensing of the ammonia-water equilibrium across
an air gap between the anolyte solution and the electrode surface. This
system has not yet been applied to a significant extent by groups in other
laboratories.
The second commonly used method for ammonium traces in aqueous solution
is the indophenol colorimetric method based on the color-producing reaction of
phenol and hypochlorite in the presence of ammonia. Modifications most
analytically useful for determination of ammonium in aqueous leaches were
reported by Bolleter et al. (1961) and by Tetlow and Wilson (1964). An
automated procedure has been proposed by Keay and Menage (1970). The
procedure can have a lower detection limit as low as 0.05 ng/ml (as nitrogen)
requires only a few minutes of analysis time, and has a minimum sample volume
of 2 ml.
There is still a striking need for development of new techniques for
real-time and/or short-time-averaged measurement of atmospheric concentrations,
especially in view of the important role of NH~ in converting aerosol H?SO. to
less harmful ammonium salts.
2-79
-------
2.4.6 Summary
Methods for the determination of soluble sulfates, total sulfates, and
specific sulfate species involve the collection of particulate matter and its
subsequent analysis by a variety of techniques. Direct methods for water
soluble sulfates based on barium sulfate precipitation include gravimetric,
nepholometric and turbidimetric, and titrimetric methods. A recent modifica-
tion of the Thorin method (titrimetric) permits the rapid determination of
sulfate in aqueous extracts with a limit of detection of <0.2 pg/m . Indirect
methods based on the exchange of barium with a more easily determined anion
include chloroanilate, iodate and dye exchange methods. A dye exchange method
based on methyl thymol blue (MTB) indicator has been demonstrated by EPA to be
superior to other frequently used methods in range and reliability for long-
term, low-volume sampling. Other anions that form insoluble barium salts will
cause positive interference with methods based on barium suHfate precipitation.
A commercial instrument based on ion-exchange chromatography is now availabl
for trace soluble sulfate determinations. This method is specific, exceptionally
accurate and sensitive (<1 ug/ml). The instrument is relatively expensive and
requires a high level of operator technique. A method based on flash volatiliza-
tion (pyrolysis) of sulfate to SCL with flame photometric detection of SCL has
been developed. This method offers increased sensitivity (<0.5 ug/ml sulfur)
but has several disadvantages including lack of specificity for sulfate (other
water soluble sulfur species are detected), irreproducible recoveries, and not
determining refractory species such as K?SO..
Methods for determining total aerosol sulfur are important because nearly
100 percent of aerosol sulfur mass is present in the form of sulfate. Methods
for total sulfur include x-ray flourescence, electron spectroscopy for chemical
2-80
-------
analysis (ESCA), and direct flame photometry. Induced X-ray fluorescence
methods are non-destructive and applicable to the determination of large numbers
of ambient aerosol samples. Although induced X-ray fluorescence methods are
difficult to calibrate and the instrumentation is expensive, the method is
non-destructive and applicable to a large number of aerosol samples. This
method will continue to be an important tool for total sulfur determinations.
The detection limit for sulfur for an induced X-ray fluorescence method has
2
been reported as 0.1 ug/cm of filter which corresponds to a concentration
3
value of <0.1 ug/m sulfur for a 2-hr sample collected at 50 liters/minute.
ESCA methods are sensitive to surface composition of samples, which is an
advantage for surface-oriented studies but not for ambient aerosol samples
whose elemental composition is likely to be nonhomogeneous. Direct flame
photometry has potential as a sensitive total aerosol sulfur analyzer, but is
complicated because SO^ must be removed and the FPD response varies with the
chemical form of the aerosol sulfates.
Procedures for determining specific sulfate species include thermal
volatilization and solvent extraction techniques, gas phase ammonia titration,
infrared and visible spectrometry, flame photometry and electron microscopy.
The determination of HUSO, by its selective thermal volatilization from filters
has been reported by several workers. This technique generally suffers from
poor HpSO. recoveries, poor reproducibility, and interferences from ammonium
sulfate salts. A prototype, semi automated instrument has been developed which
collects aerosol samples on Fluoropore filters, passes heated dry N? in the
reverse direction through the filter and detects the volatilized H^SO. with a
flame photometric detector. Laboratory aerosols containing 0.25 (jg sulfur may
be determined with reasonable precision with this instrument. However, serious
2-81
-------
difficulties are encountered in removing HUSO^ in the presence of ammonium
bisulfate and sampling times of 1-hour or more for most ambient applications
are required.
Solvent extraction procedures to remove hUSO. selectively in the presence
of other aerosol sulfates have been investigated. Isopropanol as an extraction
solvent was found to be inadequate because ammonium bisulfate is quantitatively
extracted. However, benzaldehyde is a selective extractant for ^SO. and can
be used to quantify the presence of H^SO. in ambient particulate samples.
Several semiquantitative methods for estimating sulfate species have been
investigated. These include gas phase ammonia titration techniques methods
based on infrared spectroscopy and electron microscopy techniques. Finally,
methods for measuring ammonia and gaseous ammonia are reviewed.
2.5 MEASUREMENT TECHNIQUES FOR PARTICULATE MATTER
2.5.1 Introduction
Measurements of atmospheric particulate matter are complicated by the
complex and diverse nature of individual particles. Unlike gaseous pollutants
which are typically composed of single compounds, suspended ambient particulate
matter includes a wide variety of particle sizes, shapes, and other physical
and chemical properties. The collection and examination of particles by their
aerodynamic equivalent diameter provides a means of examining mass concentrations
that inherently accounts for particle size, shape and other physical properties.
Without a detailed understanding of the characteristics of sampling methods
for collecting particles, it is impossible to draw conclusions from sampling
data and to compare sampling methods. Classical ambient particulate sampling
2-82
-------
consists of extracting the suspended particles from a measured volume
of ambient air by filtration. This technique is directly applicable for
routine monitoring and inherently provides a sample integrated over a specific
time interval long enough to collect sufficient mass for subsequent analysis.
The particle mass collected is determined by weighing the filter before and
after exposure, and mass concentration is obtained by dividing the mass of
particulate matter collected by the total air volume sampled to give mass per
unit volume. This relatively simple particulate measurement concept is the
basis for the high-volume sampling method, which is the current EPA reference
method for total suspended particulates (TSP) (U.S. Environmental Protection
Agency, 1979 a). This method has been widely used in this country and others,
and has provided many of the concentration data .used to establish health and
other effects. The high-volume sampler also collects sufficient particulate
material to allow subsequent chemical analysis.
Alternative filtration techniques, such as the optically-based British
Smoke Shade and AISI (American Iron and Steel Institute) tape sampler, have
been used for routine monitoring because of their simplicity of analysis. The
British Smoke Shade sampler is commonly used in Europe but has been rarely
used in the United States. The AISI Tape Sampler is more universally utilized.
The dustfall method, an entirely passive collector for settleable particles,
is rarely used in current studies but has been used in the past to collect
data relative to soiling and material effects. Because such methods do not
provide mass concentration directly, many comparison studies in the literature
attempt to relate these methods to methods based on direct gravimetric measurement
2-83
-------
A significant problem with the simple filtration methods used in earlier
studies is that little or no information was available about the size range of
collected particles. The knowledge of the size range of particles reaching
the collection surface relative to the true ambient distribution — as well as
the total mass concentration--is very important when utilizing the data to
meet specific objectives such as support of epidemiology studies or development
of source control strategies.
The particle size range collected by a filtration sampler depends on such
parameters such as inlet geometry, internal wall losses, and the efficiency of
the filter material. For example, the high-volume sampler defined by EPA in
the previous Air Quality Criteria for Particulate Matter (National Air Pollution
Control Administration, 1969) and in the reference method for TSP was considered
to have captured all sizes up to 100 urn (aerodynamic diameter). However,
recent sampler characterization testing by Stevens and Dzubay (1975) and
Wedding et al. (1977) has shown that the gable roof used as a weather shield
on this sampler produces inefficient collection of particles larger than 50
urn. As shown in Figure 2-7, the sampling efficiency of the high-volume sampler
for large particles is also affected by wind speed. Lundgren (1973) has
examined the true distribution of large particles up to 200 urn in the atmosphere,
and several typical mass distributions are shown in Figure 2-8. After comparing
the high-volume sampler collection efficiency data in Figure 2-7 with these
particle size distributions it is clear that the high-volume sampler does not
2-84
-------
100
80
LU
o
o:
UJ
CL
to
CO
LJ
60
£40
O
LJ
LJ
20
^u
0
o
a
o e
o 2 Km/hr
O 8 Km/hr
e> 24 Km/hr
I 3 5 7 ICT 30
AERODYNAMIC PARTICLE DIAMETER,
FIGURE 2-7. SAMPLING EFFECTIVENESS OF
11V x 14" HI-VOL.
Source: McFarland and Rodes (1979)
2-85
-------
180
165
150
135
120
O
O
_J 90
0
W 75
60
' MASS CONCENTRATION
-O— 112'
"-0- 13
"-O- 60
•FUGITIVE OUST EPISODE
0.1
1.0 10
PARTICLE DIAMETER (Dp). pm
100
Fl gure 2-8. Mass distribution data emphasizing large OlOpm) particles.
SOURCE: LUNDGREN. 1973
1000
2-86
-------
provide a true measure of the large particles in the atmosphere. Because the
mass of a particle increases as a cubic function of the diameter for particles
with constant density, the sampling of large particles must be treated carefully
when considering a broad size distribution. McFarland (1979) recently determined
that the British Smoke Shade sampler collects only a small portion of the
particles greater than about 4 (jm primarily because of the small diameter of
the inlet tubing. Knowledge of such characteristics is essential in making
comparisons between two particulate measurement methods.
Size-specific sampler inlets designed to limit the particles collected to
a certain size range are a relatively new technology, especially for particles
larger than 10 urn. Because these larger particles are difficult to transport
quantitatively, a sharp cutoff for large particles is not easily obtained
except at high sampler flowrates. The efficiency of a prototype inlet designed
3
to provide a 15 urn cutoff for a low flowrate sampler operating at 1.0 m /hr is
described by Wedding et al. (1977) and is shown in Figure 2-9. The D5Q or 50%
diameter -- the particle size at which 50 percent of the particle mass is
passed on to the filter -- for this inlet is approximately 15 \im. The geometric
standard deviation of the effective particle distribution (a measure of the
sharpness of the size cut-off and denoted as a ) for this inlet is approximately
9
2.0, as compared to an ideal step-function inlet with a a of 1.0.
technology size selective inlets such as the one developed for the high-volume
sampler, (ES&T Outlook, 197ft) ftave- a values/(from 1.3 to 1.7. The conventional
TSP high-volume sampler inlet has a a between 2.2 and 2.5, depending on the
2-87
-------
UJ
OL
in
VI
UJ
z
UJ
100
80
60
40
20
U. • 15 fp», 127 mm FILTER
O WITH < I "/. TURBULENCE
D WITH 8% TURBULENCE
U^ « 5 »ps. 127 mm FILTER
® WITH* 1% TURBULENCE
WITH 8% TURBULENCE
D
10
20
30
40
50
PARTICLE DIAMETER, Dp (
Figure 2-9 . Influence of particle size on effectiveness of prototype dichotomous sampler.
SOURCE: WEDDING, et. al. 1977
2-88
-------
w
u
100 -
80 -
60 -
40 -
20 -
T I I IT
LIQUID PARTICLE
LOSS * 10
i i i 4
34
10
Dp
20
791-8038
Figure 2-10. The size separation characteristics and wdU loss measurement of a
2.5 vim cutpolnt virtual Impactor as a function of particle size Dp.
C and F represent particle collections on the coarse and fine filters
respectively.
Source: Loo et al. (1979)
2-89
-------
windspeed (McFarland and Rodes, 1979). When used to size-segregate smaller
particles (<10 pro), size separation techniques such as cyclones and inertial
impactors often have a values less than 1.3 (John et al., 1978; Gushing et
a!., 1978). Figure 2-10 shows an example of the sharpness of a virtual impactor
(dichotomous sampler) cutpoint at 2.5 urn as reported by Loo et al. (1979).
After particles pass through the sampler inlet, they can still be lost
from the flowstream before collection or measurement by attraction to or
impaction on the internal surfaces of the sampler. Minimizing internal loss,
especially for larger particles, requires careful design of the sample transport
system geometry as well as consideration of factors such as surface charge
dissipation. Wedding et al. (1977) reported internal wall losses in a prototype
size-specific sampler to exceed 40 percent for particles greater than 15 urn,
which can be compared with the minimal losses shown in Figure 2-10.
The efficiency of the filter media used can influence the total mass
collected if very small particles are not retained on the filter, or if very
large particles bounce off the filter to the sampler walls. The collection
efficiencies over a range of particle sizes for a wide variety of filter
materials, face velocities, and effective porosities have been determined by
Liu et al. (1978). The efficiencies for a majority of the commonly used glass
fiber and membrane filters were shown to be nearly 100 percent for particles
as small as 0.03 urn. Table 2-11 tabulates selected fractional efficiency data
for a commonly used high-volume sampler glass fiber filter, a Teflon® membrane
filter, and the cellulose fiber filter material (Whatman No. 1) used in the
British Smoke Shade Sampler. Note that the latter filter shows some inefficiency
2-90
-------
TABLE 2-11. FRACTIONAL AEROSOL PENETRATION FOR SELECTED SUBSTRATES
AS A FUNCTION OF FACE VELOCITY AND PARTICLE SIZE
FILTER: Gelman Type A, glass fiber
AP, cm Hg
V, cm/sec
1
11.2
1.5
16.9
3
32.7
10
108
Dp, pm
0.035
0.10
0.30
1.0
FILTER: Ghia S2
AP, cm Hg
V, cm/sec
PENETRATION
<0.0001
<0.0001
<0.0001
<0.0001
37PJ 02,
1
23.4
<0.0001
<0.0001
<0.0001
<0.0001
teflon membrane,
3
64.1
<0.0001
<0.0001
<0. 0001
<0.0001
2.0 urn pore
10
187
0.0008
0.00054
<0. 00007
<0. 00002
Dp, Mm
0.035
0.10
0.30
1.0
FILTER: Whatman
AP, cm Hg
V, cm/sec
PENETRATION
<0.0002
<0. 00006
<0. 00007
<0. 00007
0.0011
0.00008
<0. 00007
<0. 00009
0.0005
<0. 00024
<0. 00022
<0. 00008
No.l, cellulose fiber
1
6.1
3
17.4
10
47.6
30
102
Dp, urn
0.035
0.10
0.30
1.0
PENETRATION
0:56
0.46
0.16
0.019
0.52
0.43
0.044
0.034
0.34
0.13
0.0049
0.0044
0.058.
0.0071
0.00051
0.00042
Source: Liu et al., 1978
2-90a
-------
(D)
at the smallest particle sizes, while the nominal 2 urn porosity Teflon filter
is highly efficient for small particles due to collection mechanisms other
than interception.
Non-filtration techniques for size selective particle sampling have been
used to determine mass distributions by particle size. The multistage inertial
impactors (more commonly called cascade impactors) provide discrete samples
associated with selected particle size ranges that can be analyzed for mass or
other constituents. Typically, these samplers have had uncharacterized inlets
when used for ambient monitoring, which often results in a misinterpretation
of the mass median diameter (MMD) with respect to total mass collected. Dzubay
et al. (1976) and Cahill (1979) have also noted that allowing entry of particles
much larger than the first stage cutpoint can cause particles to bounce to
lower stages, shifting the calculated MMD. Cyclone separators have also been
used to collect discrete ambient samples primarily for size fractions less
than 10 urn, but these devices are more difficult to use in serial operation.
Methods for continuous measurement of ambient particles provide real
time measurements, but with current technology do not provide information
directly related to mass concentration. Most of these i_n situ measurements
are optically based and provide information such as particle number, surface
area, and volume distributions by particle size or parameters related to
ambient particles such as light scatter or visible contrast change. These
measurements are valuable to researchers and provide input to control strategy
assessments, but in general are not currently viable as routine monitoring
methods for demonstration of compliance.
2-91
-------
2.5.2 Integrated Sampling Methods
2.5.2.1 Introduction—Most particle samplers collect samples integrated over
a specified time interval. EPA's National Air Surveillance Network operated
high-volume samplers for 24-hour intervals to support the current standard.
The British Smoke Shade sampler is also operated over a 24-hour period.
Inertial impactors often require operation for shorter periods, depending on
the ambient particle concentrations, to eliminate collection stage overloading.
Even though some tape samplers can provide a continuous analytical readout,
the measurements are reported as an integrated average for a selected time
interval, usually 1 to 4 hours per sample.
Sampling methods that were either used to collect aerometry data described
in Chapter 14 of this document, or are currently being used extensively in
routine monitoring programs are described below.
2.5.2.2 Methodology
2.5.2.2.1 TSP High-Volume Sampler—As noted earlier, the high-volume sampler
is the current EPA reference method for total suspended particulates (TSP). It
o
is intended to operate at flowrates from 1.1 to 1.7 m /min, drawing air through
a 200 x 250 mm glass fiber filter. The collected mass of particles on the
filter is determined by the difference between gravimetric measurements before
and after exposure. The mass concentration is interpreted over the sampling
2-92
-------
interval and is normally expressed in ug of mass collected per m of air
sampled (ug/m ). As shown in Figure 2-11, the inlet is formed by the overhang
of a gable roof above the filter. The inlet efficiency, as noted in Figure
2-7, does not provide a sharp particle size cutoff and is sensitive to wind
speed. Also shown on Figure 2-7 are the estimated particle sizes for 50 percent
collection (D™) for each of the wind speeds tested. The collection efficiency
also is affected by sampler orientation, i.e., it is sensitive to wind
direction, as shown in Table 2-12 (Wedding et al. 1977). The sampler
flowrate is measured either before and after collection using an external
flowmeter or continuously with a flow recorder. The effect of sampler flowrate
on particle collection as shown in Figure 2-12 is not substantial. The commonly
used filter media for this sampler are nearly 100 percent efficient for 0.3 urn
particles (Liu et al. , 1978). As noted by Friedlahder (1977), this size
particle is the most difficult to capture, since the collection of smaller and
larger particles is assisted by diffusion and interception, respectively.
The overall accuracy of ambient particulate measurements such as those
made by the high volume sampler cannot be quantified with current state-of-the-
art technology. Alternatively, estimates of the components of the overall
accuracy must be determined, including collection efficiency of the sampler
inlet and filter media and accuracy of the flow measurement system. Two
commonly used flow measurement devices on high volume samplers are the rotameter
(visi float) and the flow recorder. The rotameter is used to measure only the
initial and final flowrates from which an average is calculated. The flow
2-93
-------
.52
FLOW
FILTER
FLOW
CONTROLLER
FLOW
RECORDER
INLET COVER
Figure 2-11. TSP Hi-Vol Used in IP Network.
2-94
-------
TABLE 2-12. SAMPLING EFFECTIVENESS OF HI-VOLUME SAMPLER AT 15 FT/S
Particle diameter, (jm
15 30 50
o
Sampler orientation,
0 45 0 45 0 45 0 45
97% 100% 35% 55% 18% 41% 7% 34%
SOURCE: Wedding et al. (1977)
2-95
-------
100
LJ
o
(T
LJ
£L
80
60
CO
CO
LJ
| 40
H
O
LJ
LJ
0 I ' J
i i
i i i
0 0.5 10 1.5 2.0
VOLUMETRIC FLOW RATE, m3/min
Figure 2-12. EFFECT OF FLOW RATE ON PERFORMANCE
OF 11V x 14" HI-VOL.
AERODYNAMIC PARTICLE DIAMETER = 29 urn.
Source: McFarland and Rodes (1979)
2-96
-------
recorder provides a continuous trace that can be integrated for a more
accurate measurement. Smith et al. (1978) using high volume samplers with
both types of devices noted that the flow recorder produced smaller errors (2
to 4 percent) when compared with a reference flow device than the rotameter (6
to 11 percent). The precision of the high volume sampler as determined from
collocated sampler measurements under field conditions and expressed by the
coefficient of variation (CV) have been reported by several investigators.
McKee et al. (1971) determined the CV for a measurement by a single analyst to
be 3.0 percent, while the same measure among multiple analysts in a
collaborative test was 3.7 percent. The plenum design of the gable roof
provides a settling chamber for larger particles blown in during periods when
the sampler is not operational. McFarland and Rodes (1979) have quantified
this deposition experimentally as shown in Figure 2-13 as a function of
particle size and ambient wind speed. Interpreting these relationships
requires knowledge of the existing ambient size distribution. For a typical
distribution as shown in Figure 2-1, the amount of mass added to a high-volume
sampler filter during 5 days of exposure when it was not operational could
amount to 6 to 8 percent. This effect has also been measured in a field
situation by Sides and Saiger (1976) and Lizarraga-Rocha (1976), who measured
weight increases from 3 to 12 percent. Errors from this effect may be reduced
by equipping the sampler with one of the commercially available automatic
mechanical devices that keeps the filter covered during nonsampling periods.
2-97
-------
1000
10
Ejoo
o
o
•E
of
u
o:
0-
co
O
0_
LJ
O
10
1.0
m
c.
T - 1 - 1 I MM! - 1 - 1 - 1 I I I
deposition rate on Hi-Vol filter, vg/day
mean aerosol concentration collected by Isoklnetic
sampler,
O 2Km/hr
• 8Km/hr
I J i t I i I i
i i i i iu_
1.0 3 10 30
AERODYNAMIC PARTICLE DIAMETER,
100
Figure 2-13. Deposition of standby mode, 11 1/2" x 14" hi-vol.
Source: McFarland and Rodes (1979)'
2-98
-------
Alternatively, installation and retrieval of filters relative to the scheduled
sample period will also minimize the problem.
As shown by Coutant (1977) and Spicer and Schumacher (1979), artifact
participate matter can be formed on the surfaces of alkaline filter materials
such as glass fiber by oxidation of the acid gases present in the sample air.
The effect is a surface-limited reaction usually occurring early in the sample
period, and is dependent on factors such as the length of the sample period,
the filter pH, and the presence of acid gases in the sampled air. The magnitude
and the significance of artifact mass errors are variable and dependent on
local conditions. The majority of artifact mass is the sum of the sulfates
and nitrates formed by filter surface reactions with ambient sulfur dioxide
and nitric acid, respectively. The ambient concentration of sulfur dioxide is
primarily dependent on fossil fuel combustion, while nitric acid concentration
is dependent on atmospheric photochemistry. A laboratory study by Coutant
(1977) has reported sulfate artifact mass for 24 hour samples from 0.3-3
ug/m3. This is supported by the field data of Rodes and Evans (1977) who
noted a 0.5 ug/m3 artifact in Los Angeles, California. Spicer and Schumacher
(1979) have reported nitrate artifacts of approximately 5 |jg/m3 for field
measurements in Upland, California. These data would support a typical
influence of 2-5 ug/m3 and a worst case influence of 5-8 pg/m3 on the total
mass concentrations determined using glass fiber filters.
2.5.2.2.2 British smoke shade sampler. The design of the British Smoke Shade
sampler is based upon early work by Hill (1936), who used transmitted light to
assess the darkness of the stain resulting from particle collection on the
filter paper. This sampler draws an air stream upward through an inverted
funnel and 3 meters of nominal one-quarter inch diameter plastic tubing to an
2-99
-------
inverted filter holder containing a Whatman Number 1 cellulose fiber filter.
As noted earlier, this filter medium has been shown to be somewhat inefficient
when sampling very small particles (Liu et al., 1978). A schematic diagram of
a version of this sampler designed to sequentially collect samples for 8 days
is shown in Figure 2-14. A bubbler is often used down stream of the filter
holder for subsequent S0? measurements. The sampler is operated at approximately
1.5 liters/min, which is verified by a dry test meter built into the sampler.
The filter holder can be 25, 50, or 100 mm in diameter to collect a spot of
the proper darkness range for subsequent measurements by reflectance. A 1964
study supported by the Organization for Economic Cooperation and Development
(OECD) established relationships between smoke shade reflectance measurements
and gravimetric total suspended particulate concentrations, based on collocated
sampler tests performed in Britain with the high volume sampler. These data
were accepted by the World Health Organization (WHO, 1976) and compiled into a
standard operating procedure, which permitted reporting of smoke shade measure-
ments in equivalent ug/m .
This sampler was used extensively in early monitoring studies, and its
aerosol collection properties have been examined by McFarland (1979). This
examination produced the efficiency plot shown in Figure 2-15, which shows
that the D5Q for particles reaching the filter is only 4.5 urn. Many large
particles are either rejected at the inlet or lost in the inlet line. Since
the typical ambient size distribution contains only a small amount of mass
between 3.5 and 4.5 pm (see Figure 2-1), the size fraction collected by the
British Smoke Shade sampler is very nearly equivalent to the 0-3.5 urn respirable
suspended particulate (RSP) fraction defined by .the American Conferenc&jOf
J/tM^w(^fti^£>/-M&i^t-' Wntwuf&&
Governmental and Industrial Hygienists (ACGIH) fl!968) and described by Lippman
2-100
-------
I
I-J
o
•OT**Y srout NCIIVG VAIVI
Figure 2-14. British Smoke Shade Sampler.
-------
eoo
80
u
a:
u
tL 60
40
LJ
Z
iL)
20
t T I I I Tl
SNLET ALONE
e 2 km/hr
® 8 km/hr
ENTIRE SYSTEM
& Z km/hr
10
30
AERODYNAMIC PARTICLE DIAMETER, fim
Figure 2-15. PENETRATION OF AEROSOL THROUGH INLET ALONE
AND THROUGH ENTIRE FLOW SYSTEM TO FILTER
Source: McFarland (1979)
2-102
-------
(1970). Since the size range of particles collected by the smoke shade sampler
is substantially less than that collected by the high volume sampler, comparisons
between the methods could be expected to be variable. This varibility is
primarily influenced by the proportions of fine and coarse particles in the
atmosphere (e.g., see Figure 2-8). and to a lesser degree by the inherent
color of the particles. The reproducibility (CV) of collocated smoke shade
sampler measurements was reported by the British Standards Institution (1964)
as 6 percent.
2.5.2.2.3 Tape sampler. A variation of the optical measurement of spot
darkness is the use of a continuous filter tape and an automatic tape advancing
system. A sampler using this approach, developed by Hemeon (1953), samples at
a flowrate of approximately 7 liters/min using Whatman Number 4 filter paper
and collects particles on a 25 mm filter spot. The spot darkness is read by
either a transmittance or reflectance measurement. Transmittance measurement
is by far the most popular.
The sampler, often referred to as the AISI tape sampler, typically collects
particles in selected time intervals of 1 to 4 hours, and then advances to an
unexposed clean portion of the tape. Optical measurements are referenced to
an unexposed filter area and can be made external to the sampler after sample
collection or with a real-time continuous readout self-contained in the sampler.
Reflectance measurements are converted through a simple Beer's Law relation-
ship to CoH (Coefficient of Haze) units per 1000 linear feet of air sampled.
A CoH is defined as the quantity of particulate matter on the paper tape that
produces a change in optical density of 0.01. The less popular RUDS (Reflectance
of Dirt Shade) is equivalent to 0.1 CoH units.
The tape sampler, shown in Figure 2-16, utilizes a small diameter plastic
2-103
-------
H01AVE 7[
»US- 10 TEST WAIVE
T H B f f V. t V V i VI
11.25
46LBS.
Figure 2-16. A.I.S.I. Sampler with Built-in evaluator and recorder.
2-104
-------
inlet tube and funnel nearly identical to the British Smoke Shade sampler.
Even though the two samplers operate at different flowrates, the particles
reaching the filter tape will have a size range similar to that illustrated in
Figure 2-15. This size range restriction plus the capability for direct
real-time readout can be useful for providing a continuous measurement of fine
particles during air pollution episodes. The utility of the sampler to estimate
mass concentrations has been investigated by many researchers, usually in
comparison with the standard high-volume sampler. Since these two samplers do
not collect similar particle size ranges, such comparisons are difficult
unless very few coarse particles are present. Regan et al. (1979) have shown
with field data that the correlation improves when the tape sampler data are
compared with smaller particle fractions.
2.5.2.2.4 Inertial impactors. Often referred to as cascade impactors when
used in multiple stages, inertial impactors provide a means of collecting an
ambient mass fraction which is divided into subfractions of specific particle
sizes. This method involves acceleration of the ambient air stream by drawing
it through one or more converging nozzles or slots. As shown in Figure 2-17,
upon leaving the nozzle the jet of air is directed perpendicularly toward a
collection surface. Large, high inertia particles are unable to turn with the
air stream and consequently impact against the collection surface. Smaller
particles follow the air stream and can be directed either to another stage of
impaction or collected on a filter. Utilization of multiple stages, each with
a different nozzle velocity, provides collection of particles in several size
ranges.
Cascade impactors typically have 2 to 6 stages, and commercial low volume
sampler flowrates range from about 0.01 to 0.04 m /min. Lee and Goranson
2-105
-------
CSJ
01
en
1C
+J
(/I
O)
cr>
IQ I
Absolute
Filter
Flow
Flowmeter
Vacuum Gauge
Figure 2-17. Cascade Impactor Sampling System
2-106
-------
(1972) utilizing a modified commercially available 0.03 m An in low volume
o
impactor, operated it at 0.14 m /min to obtain larger mass collections on each
stage. Cascade impactors have also been designed to mount on a high-volume
sampler and operate at much higher flowrates, from approximately 0.6 to 1.1
m /min. A single stage high volume impactor as shown in Figure 2-18 was
utilized in the Community Health Air Monitoring Program (CHAMP) program and
operated at 1.1 m /min.
The particle size cutpoints for each stage are dependent primarily on the
sampler geometry factors and flowrate. The smallest practical size cutpoint
is approximately 0.3 urn. Below this size, the high vacuums and small nozzle
diameters required to attain the necessary jet velocities can result in changes
in particle character and rapid clogging of the smaller stage nozzles. A
filter is normally used after the last stage of inertial impactors to collect
particles not collected on the impactor stages. The masses collected on each
stage plus the backup filter mass collection are often reported as shown in
Figure 2-19 from data by Lee and Goranson (1972). This cumulative distribution
format permits determination of the MMD, at which point 50 percent of the mass
is smaller than the indicated size.
The inlet cutoff characteristics of most impactors have not been determined
resulting in uncertainty about the size range of particles sampled. McFarland
o
(1980) has examined the inlet of the NASN low volume (0.03 m /min) cascade
impactor and determined that particles larger than 10 urn were unlikely to
reach the collection stages. Willeke and McFeters (1975) characterized the
CHAMP high volume sampler inlet under static wind speed conditions as shown in
Figure 2-20. If the impactor inlet DrQ is known, the total mass collected by
the sampler can be used for comparison with other size specific measurements.
2-107
-------
HANDLE
SPACER
FLOW SENSOR
COVER
LARGE
PARTICLE
FRACTIONATOR N.
essss
IMPACTOR
UNIT
C
\
f~~
IMPACTOR
PLATE
\ |
•*• *— *s= =
A ^
1
f
J > 1
1 : — . .«..'..'
^
]l
^"
HI]
INLET
WATER DRAIN
GLASS FIBER
IMPACTION SURFACE
FINAL FILTER
VACUUM PUMP
Figure 2-18. Cross Section Schematic of the CHAMP Aerosol Sampler.
Source: Ranade and Osdell (1978)
2-108
-------
Mtnu
IIUI
ff.f
M LMb
• TVT
•*
™*
•1
in >LU v
Ml*
M
™
-------
CJ
t:
K
•»•>
c
&)
o
t-
o
o.
o
2.
b.
6.
£-
O
tJ
^a
o
100
90
So
70
60
30
1
10.0 3«j.O 40,0
(nicrometerr,)
20.0 25.0
Aerodynamic
Figure 2-20. Experimentally derived collection efficiency
curve o<" the CHAM? fracticnator if^et operated
at 1132.7 1/min (*»0 SCFM% Dpaero s 25^tnf (T f
Source: W111eke and McFeters (1975).
2-110
-------
Cascade impactors use removable impaction surfaces for collecting particles
Impaction substrates are weighed before and after exposure and are typically
glass, metal foil or glass fiber filters. The selection and preparation of
these substrates can have a significant effect on the impactor performance.
Improperly coated metal surfaces can cause particle bounce to lower stages
resulting in substantial cutpoint shifts (Dzubay et al., 1976). Glass fiber
substrates can also cause particle bounce and are subject to the formation of
artifact particles similar to high-volume sampler filters. The single impaction
stage of the CHAMP high volume sampler designed to be 3.5 urn was characterized
by Ranade and Van Osdell (1978) as shown in Figure 2-21.
Multiple stage impactors are more commonly used to collect discrete mass
fractions in specific size ranges, from which particle mass size distributions
such as Figure 2-8 are constructed. These size -distributions can provide
information valuable to the understanding of sources and transport of suspended
particles. Cascade impactors are not normally operated in routine monitoring
networks because of the manual labor requirements for sampling and analysis.
2.5.2.2.5 Cyclone sampler. Ambient particle size separation using cyclones
has primarily been used to provide cutpoints in the 2 to 4 urn range. A cyclone
sampler as used in the Community Health Environmental Surveillance Studies
(CHESS) (Bernard, 1976) is shown in Figure 2-22. This cyclone, as charac-
terized in Figure 2-23, provides a relatively sharp 3.5 pm separation. The
inlet of the sampler is the cyclone inlet, and a single 0 to 3.5 |jm particle
fraction is collected on the filter. The filter medium used in the CHESS
network was glass fiber. Collection of the larger particles excluded by a
cyclone on a removable substrate is difficult, but alternative approaches such
as that designed by John et al. (1978) and shown in Figure 2-24, are available
2-111
-------
ro
i
100
90
80
~ 70
o
z
UJ
U
u.
u.
UJ
Z
O
o
U
60
50
40
30
20
10
CALCULATED FROM MARPLE'S
THEORY (1970)
CUT OFF SIZE
4 /im
AMMONIUM FLUORESCEIN
(SOLID)
O DI-OCTLYL PHTHALATE
(LIQUID)
X *
3 4 56789
AERODYNAMIC DIAMETER (urn)
10
15
20
Figure 2-21. Experimental Results for The Calibration Tests.
Source: Ranade ard Van Osdell (1978)
-------
SHELTER
MAST SUPPORT
AND VACUUM LINE
RSP FILTER CASSETTE
CRFTICAL ORIFICE.
0.0 UTER/MIN
RUBBER
VACUUM HOSE
CONNECTIONS
CYCLONE SEPARATOR
Figure 2-22. Cyclone sampler and shelter assembly.
Source: Bernard (1976).
2-113
-------
•
§
§
<
FAOEOFS
NGS.
09 -*
fee
c z
^1 ^^
•J" flr
o t
u»z
Z UJ
tu o
cz
Z Z
GO
§g
eoc
u.
100-
90-
80-
70-
80-
60-
40-
30-
20-
10-
0
^^^^^
^
FRACTION OF PARTICLES jr
PASSED TO FILTER CASSETTE X
/ FRACTION OF PARTICLES
/ SEPARATED OUT BY CYCLONE
/
i
1
1
V
1 1 1 1 1 I 1 1 1 1
012 34 66786 K
AERODYNAMIC PARTICLE SIZE AT UNIT DENSITY, MICRONS
Figure 2-23. "Los Alamos" Curve for Fine Participates.
Source: Bernard (1976).
2-114
-------
I r
FOOT
47mm
TOTAL FILTER
AIR 1HLFT
TO PUMP
4 TO PUMP
AFTER-FILTER
-CYCLONE
Figure 2-24. Assembly for sampling with a total filter and cyclone in parallel
Source: John et al. (1978)
2-115
-------
to provide a total sample and a smaller subtraction sample. The efficiency
data for this cyclone as a function of sampler flowrate are shown in Figure 2-25
and indicate that sharp cutpoints with current state-of-the-art units are
possible. A neutral pH Teflon filter medium was recommended to minimize
artifact mass formation.
2.5.2.2.6 Dichotomous sampler. The dichotomous samplers currently available
commercially collect two particle size fractions, 0 to 2.5 urn and 2.5 to about
15 urn, the latter cutoff depending on the inlet. This bimodal collection
effectively separates the coarse particles from the smaller particles to
assist in the identification of particle sources. Since the fine and coarse
fractions collected in many locations tend to be acidic and basic, respectively,
this separation also minimizes potential particle interaction after collection.
The principle of "virtual impaction" applied to particle separation was
described by Hounam and Sherwood (1965) and Conner (1966). As shown in a
simplified version in Figure 2-26, the separation principle involves accelera-
tion of the particles through a nozzle, after which 90 percent of the flow
stream is drawn off at right angles. The small particles follow the right
angle flow stream, while the larger particles, because of their inertia,
continue toward the collection nozzle. A separate filter is used for each
fraction. The sharpness of separation is shown in Figure 2-27 from data by
Loo et al. (1979) for a design cutpoint of 2.5 urn. Inherent in this separation
technique is a contamination of the coarse particle fraction with a small
percentage of the fine particles in the total flow stream. This is not considered
a substantial problem for mass measurements, as a simple mathematical correction
as described by Dzubay et al. (1977) can be applied.
2-116
-------
3 4
AERODYNAMIC DIAHTTEE, ju>
Figure 2-25.
of Methylene Blue Particles Deposited in the Cyclone as a Function of
nunic Particle Diameter. Ohe CurveB are Labelled with the Flow Rate,
Source: John et al. (1978)
2-117
-------
TOTAL FLOW, 0
SEALED HOUSING
ACCELERATION
NOZZLE
FRACTIONATION
ZONE
COLLECTION
NOZZLE""
LARGE PARTICLE
FLOW, fQ
LARGE PARTICLE
COLLECTION
FILTER
SMALL
PARTICLE
FLOW
(l-f)Q
SMALL
PARTICLE
COLLECTION
FILTER
TO
FLOWMETER
AND
PUMP
TO
FLOWMETER
AND
PUMP
Figure 2-2C. Single Stage Centripeter
Source: Loo et al. (1979)
2-118
-------
100 -
I I I I I
c
C + F
LIQUID PARTICLE
LOSS x10
I I IL
A
I
34
10
Dp
20
XBL 791-8036
Figure 2-27. The size separation characteristics and wall loss measurement of a
2.5 urn cutpolnt virtual Impactor as a function of particle size Dp.
C and F represent particle collections on the coarse and fine filters
respectively.
Source: Loo et al. (1979)
2-119
-------
Dzubay et al. (1977) recommend that the filter substrate be Teflon , and
nominal porosities as large as 2.0 urn have been shown to have essentially 100
percent collection efficiency for 0.3 |jm particles (Liu et al. 1978). The
O
sampler operates at a flowrate of 1 m/hr and collects approximately 1/100 of
the particle mass collected by a high-volume sampler. Removal of the stickier
fine particles causes the particles on the coarse filter (2.4 to 15 urn) to
have a greater tendency to fall off the filter if care is not exercised during
I1?6
filter handling (Shaw, 1979).
The inlet typically used for this sampler has been a vertical elutriator
design by Wedding et al. (1977) shown in Figure 2-28. More extensive testing
by the designers has recently shown that this simple inlet is significantly
wind speed sensitive, as shown in Figure 2-29.
Current automated versions of this sampler can automatically change the
sample filters to provide unattended operation. Depending on atmospheric
concentrations, short-term samples of 2 to 4 hours are possible with the
automatic samplers to provide diurnal pattern information.
2.5.2.2.7 High-volume sampler with size selective inlet. To meet the monitoring
p^t/A-
requirements for a flew inhalable particulate (IP) standard (Miller et al.
1979), EPA commissioned the design of a size-selective inlet suitable for
existing TSP high-volume samplers to provide a single 0 to 15 ym particle size
fraction. This new inlet, shown mounted on a conventional high-volume sampler
in Figure 2-30, has been tested by the designers (McFarland and Ortiz, 1979)
and has an inlet efficiency as shown in Figure 2-31 and a sensitivity to wind
speed as shown in Figure 2-32. Dry particle bounce and re-entrainment are
3
insignificant at the designed sampler flowrate of 1.1 m /min.
2-120
-------
I
M
IVJ
• I !•»
Figure 2-28. Inlets Tor the flichotomous virtual input-tor f.i) Sierra Model
fh) flrflm.in ^utom.^teH Pichotomou?; Pnrticnl.itc Snmplinjj System.
Source: Wedding et al. (1980)
-------
ro
r\s
(SIERRA 244E INLET)
AVERAGE OF ALL TESTS
D 5 Km/Mr
O 15 Km/Hr
A 40 Km/rfr
10 I5~ 20 25
AERODYNAMIC PARTICLE DIAMETER , p.m
Figure 2-29.
Samp 1 ini? effectiveness of the Sierra Model ?44e inlet vs. particle
diameter for three wind speeds. The lines are arbitrary fits.
Source: Wedding et al. (1980)
-------
ro
i
r\>
FILTER
FLOW
CONTROLLER
RECORDER
WL~96ttM.
INLET
FLOW
STANDARD HI-VOL
SAMPLER
Figure 2-30. SSI hi-vol used in IP network.
-------
100 •
2 4 6 8 10 20
AERODYNAMIC PARTICLE DIAMETER,
Figure 2-31. AEROSOL COLLECTION PERFORMANCE FOR THE
SIZE SELECTIVE INLET HIGH VOLUME SAMPLER
Source: McFarland and Ortiz (1979).
2-124
-------
20
ro
en
C?
u 10
M
h-
5
0
0
8 12 16
WIND VELOCITY, Km/hr
20
24
Figure 2-32. Cffpct of Wind Speed upon Cutpoint
Size of SSI with Domed Roof
Source: McFarland and Ortiz (1979)
-------
The glass fiber filter material is the same as that used for a TSP
high-volume sampler, thereby presenting the same potential for artifact particu-
late formation. This size selective inlet (SSI) sampler, as with any size
fractionating device, is somewhat sensitive to sampler flowrate for larger
particles as shown in Figure 2-33. However, these data suggest that special
flow controlling measures are not necessarily required to maintain consistent
collection efficiencies over a range of sampler flowrates. The replacement of
the gable roof of the TSP high-volume sampler with a size selective inlet
effectively removes the directional sensitivity as well as the problem of
large particle deposition when the sampler is not operational.
2.5.2.2.8 Dustfall sampling. Since the largest suspended particles have
appreciable settling velocities, they are usually determined gravimetrically
following collection by deposition in a dustfall jar as described by ASTM
(1966). Although a cylindrical jar might be expected to collect the equivalent
of the dust content of an air column of its own diameter extending to the top
of the atmosphere, in fact the aerodynamic effects of the jar, the mounting
brackets for the jar, and adjacent structures tend to complicate the collection
pattern. As noted by Nadel (1958), only relative significance may be attached
to the resulting data, and then only if conditions are carefully standardized.
There is no definitive study of the effect of the height of the collector
above ground on measured dustfall.
2.5.2.3 Calibration of Integrated Sampling Methods—Dynamic calibration of
particle samplers by exposing the sampler inlet to known concentrations of
ambient particulate matter is currently not technically feasible. Because of
the variable nature of ambient particle size and shape, it is virtually impossible
to accurately simulate a real situation.
2-126
-------
100
80
LJ
U
60
LJ
20-
20
30
40
50
60
FLOW RATE, cfm
Figure 2-33. Variation of Aerosol Penetration
with Flow Rate. Particle Size =
14.1 urn. Wind Speed * 2 km/hr.
Source: McFarland and Ortiz (1979)
2-127
-------
Calibration of the sampler flow or volume measurement device is the only
type of system calibration that can be performed in the field. Since all
samplers used to measure particle concentration require a knowledge of the
total volume of air sampled, the accuracy of this measurement bears directly
on the concentration measurement. Typically, NBS-traceable fixed volume
flowmeter standards such as spirometers are used as reference devices against
which flowrate transfer standards such as orifices, mass flowmeters, or rotameters
are certified. These transfer standards are then used in the field to calibrate
the flow indicator device on the sampler. Correction for ambient temperature
and pressure must be made properly as described by Smith et al. (1978). The
EPA Quality Assurance Handbook, Volumes I and II, (von Lehmden and Nelson,
1977 a and b) provide additional guidance on quality assurance procedures.
The performance characteristics of particle samplers can be tested with
monodispersed (single size) particles which are usually spherical in shape and
near unit density. The sampler inlet is placed in the particle-laden air
stream of a wind tunnel, and the sampler is operated to collect the artificial
particles. Isokinetic nozzles in the wind tunnel are used as reference samplers
against which the sampler results are compared. Wedding et al. (1977) and
McFarland and Rodes (1979) describe such test systems. The results for various
windspeeds, sampler orientations, and turbulence levels can be determined
under controlled conditions to produce plots of particle collection efficiency
versus particle size for the sampling system or its individual components.
Testing with both wet and dry particles may be important to examine worst case
collection characteristics such as wall loss and particle bounce.
Additional particle sampler testing needed to characterize the nature and
quality of collected samples consists of side-by-side sampler tests to determine
2-128
-------
reproducibility, and tests to determine the sensitivity of the sample collection
to factors such as manufacturing tolerances and sampler flowrate.
2.5.2.4 Analysis of Particulate Samples
2.5.2.4.1 General considerations. Following collection of a particulate
sample at the monitoring site, the sample must be analyzed to determine the
desired quantitative (or qualitative in some cases) measurement result. For
simple analytical techniques such as optical density, optical reflectance or
beta density, the analysis may be carried out at the monitoring site. More
commonly, the exposed filter or collection surface is returned to a laboratory
for analysis. During this transfer, care must be exercised to avoid loss of
fibers or particulate matter from the filter or collection surface and to
protect the sample from damage or conditions that may affect the analytical
result. Sometimes special filter cartridges or filter holders are used to
safeguard the sample. Also transferred with the sample is an information
record containing the site and sampler identification, sample air flow readings
or total sample air volume, quality assurance data, and other information
pertinent to the sample.
2.5.2.4.1.1 Filter conditioning. Most filter materials used for particulate
sampling are reasonably insensitive to changes in humidity; however, collected
particles can be hygroscopic. For samples to be analyzed gravimetrically, the
weight of absorbed moisture could have a significant effect (Tierney and
Conner, 1967). Therefore, such filters are normally conditioned prior to
weighing. Conditioning consists of maintaining the filter in an environment
with constant humidity until moisture equilibration is achieved—usually about
24 hours. The humidity level used is somewhat arbitrary, but should be less
than about 50 percent relative humidity to avoid significant weight error due
2-129
-------
to moisture (Tierney and Conner, 1967). Excessively low humidities may
aggravate weighing errors due to static charges. Best results are obtained
when the humidity of the weighing environment is about the same as that of the
conditioning environment. For uniformity, the unexposed filters are also
conditioned prior to the preexposure weighing. No conditioning is usually
necessary for optical analyses, but analysis by beta attenuation may require
moisture conditioning.
2.5.2.4.1.2 Artifact mass. Artifact particulate matter, which can be
formed on the surfaces of alkaline filter materials such as glass fiber by
oxidation of acid gases in the sample air, results in an artifically high
particulate measurement. This surface-limited effect usually occurs early in
the sample period and is a function of the filter pH and the presence of acid
gases. While usually accounting for only a small part of the collected
particulate sample, the error can be more significant for short sampling
periods or where the total amount of particulate matter collected is very
®
small. Use of a neutral pH filter material such as Teflon minimizes the
formation of artifact material on the filter; however, as noted by Pierson et
®
al. (1980), particulate nitrates collected on inert surfaces such as Teflon
and quartz can exhibit significant metathetical losses in the presence of acid
aerosols, causing an underestimation of the true concentration.
Reactions between various particulates collected on the filter are also
possible. While such reactions may not significantly affect gravimetric
determinations, they could affect chemical analysis of the sample.
2.5.2.4.1.3 Loss of volatiles. Volatile particles collected on the
filter may be lost or partially lost during subsequent sampling, during
transport to a laboratory for analysis, or during storage before the
2-130
-------
post-exposure weighing. Filters are normally analyzed as soon after collection
as is practical, but some loss of volatiles is inevitable. For gravimetric
analyses, these losses may tend to compensate to some extent for artifact
particulate formation.
2.5.2.4.1.4 Airflow variation. The amount of particulate material
collected on the filter represents the (integrated) sum of the instantaneous
products of the sample air flowrate times the particulate concentration during
the sample period. Dividing the total quantity of particulate collected by
the averaging sample flowrate over the sample period yields the true particulate
concentration only when the sample flowrate is constant over the period. The
error resulting from a nonconstant flowrate depends on the magnitude of the
instantaneous changes in the flowrate and the particulate concentration.
Many high-volume samplers used in recent years are equipped with constant
flow regulators to avoid air flow variation. However, measurements from older
high-volume samplers, which do not have flow recorders or use flow regulators,
may suffer significant error due to flow variation. Significant flow variation
can be expected from heavy filter particulate loading, high relative humidity,
substantial ambient temperature and pressure changes, and sample pump variation.
Flowrates in low-volume samplers tend to be somewhat more stable because of
characteristically higher impedance flow systems.
2.5.2.4.1.5 Air volume measurement. Another potential source of error
from flowrate variation in samplers not equipped with a flow regulator is in
the estimation of the total air volume sampled. Typically, an average flowrate
over the sampling period is determined from measurements of the actual flowrate
before and after the sample period. When the flowrate changes nonlinearly
during the sample period, as may be characteristic during conditions
2-131
-------
of filter loading, malfunction of the flow regulator, etc., the linear estimation
of the average flowrate may not be accurate. The addition of a flow recorder
permits exact integration of the high-volume sampler flowrate during the
sampling period.
2.5.2.4.2 Mass concentration determination. Simplest of the particulate
sample analyses is the estimation of the mass concentration of the particulate
in the ambient air. This may be achieved directly by a gravimetric determina-
tion or indirectly by one of several nongravimetric techniques. Where the
particulate sample has been fractionated by particle size, mass concentration
for each size range may be determined.
Sample air volumes are determined either by measuring the total volume of
air actually sampled or by estimating the total sample air volume as the
product of the sampling time and the average sample flow ra"te. For samplers
equipped with a constant flow regulator, the average flowrate is identical to
the constant, controlled flowrate. However, for unregulated samplers, the
average flowrate itself may have to be estimated from pre- and post-exposure
flow measurements. Since an unregulated flow may be likely to change during
the sample period, some error in the final computed particulate concentration
may be expected, as described previously. Some samplers have a flowrate
recorder to provide a continuous record of the flowrate over the sampling
period, from which a better estimate of the average flowrate can be derived.
In all cases, air volume measurements and particularly flowrate measurements
are determined from the sampler flow or volume indicator readings via the
calibration relationship for the indicator established during sampler calibra-
tion. These measurements may be further corrected for the actual temperature
and pressure at the sampling point during the sampling period.
2-132
-------
Sampled air volumes are usually corrected to a reference temperature 25°C
and a pressure of 760 mm Hg (U.S. Environmental Protection Agency, 1979 b).
Accordingly, mass concentration is typically reported as mass (micrograms) per
standard cubic meter of air (ug/std m ; however, the "std" is often omitted).
If desired, the actual, physical mass concentration (ug/m ) can be computed
from the reported concentration by correcting back to the actual temperature
and pressure that prevailed at the time and place the sample was collected.
2.5.2.4.2.1 Gravimetric determination. Mass concentration is determined
directly by dividing the mass (weight) of the collected particulate by the
total (standard) volume of air sampled. Filters are conditioned and weighed
before and after exposure to determine the particulate mass as the net weight
gain of the filter or other collection surface. In addition to potential
errors described earlier, errors are possible in the weighing operation.
High-volume samplers generally collect sufficient particulate matter to be
weighed with a semi-micro balance and commensurate weighing techniques.
However, lower volume samples, and particularly size-fractionated samples,
often require weighing with a micro balance and more sophisticated weighing
techniques to achieve adequate precision.
2.5.2.4.2.2 Nongravimetric methods. Several nongravimetric analytical
methods are or have been used to estimate the mass concentration of particulate
matter in the atmosphere. In general, such methods rely on the measurement of
a property of the collected sample (as it is thinly and presumably uniformly
deposited on a filter matrix) that is closely related to mass. Common examples
include reduction in reflected light (British Smoke Shade sampler and tape
sampler-RUDS), attenuation of transmitted light (tape sampler-CoHs), and
attenuation of transmitted beta radiation. Typically, these techniques are
2-133
-------
either used as a relative indicator of participate levels or are "calibrated"
in terms of mass units by empirical, statistical comparisons with samples
collected concurrently at the same site and analyzed gravimetrically.
2.5.2.4.2.3 Optical determination. Optical analyses of integrated
samples are made on samples collected by the British Smoke Shade sampler and
AISI tape sampler. The reflectance measurement procedures for the British
Smoke Shade filters and AISI tapes are described in detail in the World Health
Organization's "Selected Methods of Measuring Air Pollutants" (WHO, 1976).
The reflectance smoke shade measurements, using specific laboratory reflectance
meters, are delineated along with the empirical relationships permitting
3
conversion to equivalent |jg/m . The smoke shade reflectance meter is calibrated
at zero percent against a standard black tile and 100 percent for a standard
white tile. An example of one of these calibration plots is shown in Figure
2-34. Note that this curve applies to only one spot size. As an alternative
to using the standard plots, a specific site relationship can be obtained by
operation of a standard high-volume sampler together with a smoke shade sampler.
Transmittance measurements of tape sampler filter spots are made with a
simple white light transmissometer which is scaled from zero at total blackness
to 100 percent for a clean portion of the filter. A conversion equation using
the percent transmittance values, flowrates, sampling times, and spot size is
used to calculate the quantitative result in CoHs/1000 linear feet (305 linear
meters).
2.5.2.4.2.4 Measurement by beta ray attenuation. For projects in which
an automated instrument is needed to measure large numbers of samples (Goulding
et al. 1978; Loo et al. 1978) or in applications where real time mass measure-
ments (Macias and Husar, 1976) are needed, mass measurement by beta-ray attenua-
tion is attractive (Liliennfeld, 1970). In a device called a beta gauge, one
2-134
-------
_i
I
120
110
1M
M
1 T
I I I I
40
o
Ul
K
VI
30
„
20
10
10
20
30
40
§0
10
7(
DARKNESS INDEX (4) • 110 • REFLECTANCE. %
Figure 2-34. OECD proposed international standard
calibration curve; Photovolt reflectometer • Whatman
No. 1 filter paper, 25-mm diameter.
Source: WHO (1976)
2-135
-------
measures the extent to which a sample attenuates beta rays from a radioactive
source, such as 63Ni, 14C, or 147Pm. For aerosol measurements, the beta ray
attenuation must be measured on each membrane filter before sampling and again
after an aerosol deposit is collected. A set of gravimetrically prepared
standards are needed to relate the results to units of mass. Investigators
(Macias and Husar, 1976; Goulding et al. 1978) have studied the dependence of
the beta ray absorption coefficient on elemental composition of the sample.
Goulding et al. (1978) have found the dependence on composition to be very
slight for the ranges of average compositions that occur in aerosol samples.
In a recent interlaboratory comparison of aerosol sampling and measurement
methods (Camp et al., 1978), it was demonstrated that laboratory beta gauge
®
measurements of ambient aerosols collected by dichotomous samplers on Teflon
filters compared favorably in precision and accuracy with gravimetric analyses.
Further studies to compare beta gauge and gravimetric measurements are in
progress. Principal sources of error in the method are: (1) possible changes
in the geometry of the beta source, filter substrate, or detector between the
pre- and post-sampling measurements, and (2) absorption of water from the
atmosphere by the filter material or the collected particulate (Lawrence
Berkeley Laboratory, 1975).
2.5.2.4.3 Particle sizing determination of filtered particles. Examination of
polydispersed particles on filter and inertial impaction substrates to provide
size distribution information is typically performed either by microscopy or
by using electrical resistance counters after resuspension in a liquid medium.
Microscopy is the most straightforward., and sizing is practical with a
reticle-equipped light microscope for particles larger than about 0.5 pm;
smaller particles cannot be differentiated with a light microscope, and must
2-136
-------
be examined with an electron microscope. Statistical counting techniques must
be applied when examining only a portion of the particles on a filter to
obtain reliable number distributions by particle size. The collection substrate
is also critical; a non-fibrous filter or impactor plate which maintains the
particles on its surface is needed for accurate counting.
Another technique for evaluating particles collected on a surface is
®
electrical resistance counting using the Coulter Counter . This method requires
resuspension of the particles from the collection surface in a suitable liquid
medium. The liquid suspension is passed through a small orifice in a detector
cell where the change in electrical resistance of the liquid suspension is
measured. The change in resistance is proportional to the volumetric particle
size, and use of discrete channel analysis provides a size distribution. A
conversion factor is necessary to convert from volumetric size to aerodynamic
diameter. The particle size range capability extends from 0.5 to 200 urn.
2.5.2.4.4 Chemical analysis. Analysis of trace elements. Most methods for
trace element analysis of atmospheric particulate material utilize spectroscopic
detection. They respond only to the presence of an element and provide no
information about chemical compounds. Most tell nothing of the oxidation
state of the element.
2.5.2.4.4.1 Atomic absorption spectrometry. Atomic absorption spectrometry
(Ahearn, 1972) is widely employed for quantitative elemental analysis of
airborne particles. It involves, usually, an acid extraction and excitation
of the solution by a flame. Light of a wavelength characteristic of the one
element of interest is made to traverse the flame. The amount of light absorbed
is related to the quantity of the element present. Special high temperature
furnaces can be employed instead of a flame.
2-137
-------
Individual elements must be determined sequentially. Thus, although
any element can be determined for which a lamp is available to produce
the characteristic light, airborne particulate samples are often large enough
for only half a dozen determinations to be made. Moreover, some trace elements
present in airborne particles (including antimony and arsenic) may require the
application of special methodology.
Atomic absorption is subject to significant interferences and can lead to
substantial errors. If recognized, these errors can generally be overcome or
eliminated to produce good quantitative analyses at the expense of additional
effort on each sample. Despite its drawbacks, atomic absorption spectrometry
still ranks as a useful method for elemental analysis of airborne particles.
It seems likely, however, to be superceded in the future by techniques capable
of multielement determinations.
2.5.2.4.4.2 Optical emission spectrometry. A variety of ways can be
used to excite rather loosely bound electrons in elements and observe character-
istic emissions as de-excitation occurs. The wavelength is characteristic of
the element, and the intensity is an indication of the quantity of the element
present. The most desirable excitation technique is argon plasma excitation
of an acid extract of the particulate matter (Lambert and Wilshire, 1979).
Plasma spectrometry can be more advantageous than atomic absorption,
since sample preparation and analysis rates are essentially equal. However,
emission techniques can simultaneously determine up to perhaps 50 elements,
with detection limits about the same as those of flame atomic absorption. It
is also more interference-free than atomic absorption, although interferences
are not absent.
2.5.2.4.4.3 Spark source mass spectrometry. Spark source mass spectrometry
(Morrison, 1965) can analyze airborne particulates if they can be separated
2-138
-------
from the filter and oxidized, or they can be extracted with an acid. Spectral
interferences are important, but can generally be overcome by employing a
spectrometer with high resolution.
The precision obtained depends on the way the analysis is performed but
can be about 30 percent (relative standard deviation) in careful work. Accuracy,
as with any multielement technique, may depend on the element and the matrix
in question. The advantage of this technique is that it can simultaneously
estimate the quantity of every nonvolatile element in the periodic table and
do so with roughly equal sensitivity. Large numbers of samples, however,
cannot be analyzed in a day.
2.5.2.4.4.4 Neutron activation analysis. Neutron activation analysis
(Morrison, 1965; Gootd, 1976) implies a variety of distinct methods all of
which produce unstable nuclei which emit y-radiation. Gamma-ray energy and
intensity are indicators of the element and its quantity. Instrumental thermal
neutron activation analysis is most commonly used. With this approach, a
nuclear reactor is used to produce unstable nuclei. The method can
simultaneously determine up to perhaps 25 elements in particulate samples
collected by high-volume samplers.
An advantage of instrumental activation analysis is that airborne particles
can be analyzed as received directly on the filter surface.
2.5.2.4.4.5 X-ray fluorescence spectrometry. X-ray fluorescence spectrometry
(Morrison, 1965; Fassel, 1978) involves excitation of tightly bound electrons
and observation of the x-ray emission as de-excitation occurs. Excitation may
be done by a variety of techniques, but use of an x-ray generator is most
common. Detection may either be multielement (up to perhaps 30) energy disper-
sive detection or only one to a few (up to perhaps 10 elements) by wavelength
dispersive detection. Only elements with atomic numbers greater than that of
2-139
-------
magnesium can be analyzed. Particles can be analyzed nondestructively directly
on a filter; however, samples must be thin and of uniform surface texture or
uncertain corrections must be made. Interferences are common and must be
considered. Adequate calibration can be a problem, but a reasonable sample
analysis rate can be achieved.
2.5.2.4.4.6 Electrochemical methods. Electrochemical methods have been
used to a limited extent to determine a small number of elements in airborne
particles. These methods include potentiometry with ion selective electrodes
and polarography and anodic stripping voltammetry (Morrison, 1965). Electro-
chemical methods have few advantages for airborne particle analysis, aside
from their low initial capital equipment costs when compared with other techniques.
2.5.2.4.4.7 Chemical methods. Many wet chemical procedures constitute
the classical methods employed for trace element analysis of airborne particles.
In general, a colorforming reagent is involved, and the amount of a given
element present is determined by the extent of color development. Probably
the best known of these procedures is based on the use of dithiocarbazone
(dithizone) (Snell, 1978) as the colorimetric reagent for lead. Wet chemical
procedures are labor intensive and slow compared with spectral techniques,
particularly since only one element can be determined at a time. Interferences
can also be a problem.
2.5.2.4.4.8 Analysis of sulfates and NH^. Analysis of sulfates and NH^
is covered in Section 2.4 of this chapter.
2.5.2.4.4.9 Analysis of organics. Procedures to estimate the total mass
of benzene extractable organic material in TSP have been employed occasionally.
A portion of a high-volume filter is placed in a soxhlet extractor and refluxed
with benzene for several hours. The benzene is then volatilized, and the mass
2-140
-------
of the residue is measured and reported. Utilization of this procedure now
presents very special problems because of the requirements for handling benzene.
Methods to identify and determine individual organic species present in
airborne particles abound. These schemes utilize different sequences of
solvent extractions which separate groups of different organic species based
on their solubility. Solutions are often subjected to chromatographic separation
with mass spectral detection. For organic compounds volatile at temperatures
up to about 300°C, gas chromatography-mass spectrometry (GC-MS) (McFadden,
1973) can be employed. For organic species with lower volatility, liquid
chromatography might be used. High performance liquid chromatography (HPLC)
(Kirkland, 1971) is typically used, but none of these procedures permit a high
rate of anaysis.
For one species of longstanding interest, benzo-crpyrene (BaP), thin
layer chromatography (TLC) with fluorescence detection is often used, and HPLC
procedures have been proposed. The TLC procedure requires a cyclohexane
extraction, spotting, and development of a TLC plate and fluorescence detector.
This procedure is more interference-free than some HLC methods and has a
higher production rate (Swanson, 1978).
2.5.3 Continuous Sampling Methods for Particulate Matter
2.5.3.1 Introduction--Continuous or J_n situ measurement methods for ambient
particles do not provide mass concentration directly as integrated sampling
methods can. These devices examine one or more aspects of atmospheric particles,
such as the particle size distribution or light scattering ability. This
information is valuable in research efforts to qualify the sources, transforma-
tion, and transport of particles in the atmosphere, but may not be directly
usable for regulatory monitoring to support a particle standard.
2-141
-------
The integrating nephelometer, because of its simplicity of operation, may
prove to be useful for real time determination or prediction of fine particle
episode situations. Its utility as a visibility monitor, especially in areas
concerned with prevention of significant deterioration (PSD), is being investigated.
2.5.3.2 Integrating Nephelometer--The most commonly used nephelometer for
continuous particle monitoring is the integrating nephelometer, which measures
the differences in light scatter caused by varying particle concentrations in
the air stream. These differences can be related to the scattering coefficient
(b . with units of m ) and the particle size distribution present, but are
normally expressed in terms of visible range in kilometers (Friedlander,
1977). The initial designs for this technique were made by Buttell and Brewer
(1949) and subsequently improved by Ahlquist and Charlson (1967). Light
scattering is at a maximum for particles in the 0.3 to 0.8 urn range, as shown
in Figure 2-35, and hence the response of techniques based on this principle
must be corrected for particle size effects. Light scattering is primarily
caused by accumulation mode particles and only slightly affected by particles
in the nucleation or coarse particle modes. Comparisons between nephelometer
and high-volume sampler mass concentrations, such as those by Charlson et al.
(1968), showed that a reasonable correlation existed under the conditions
tested. Comparison with any integrated sampling method in which the coarse
particle mode is variable will provide erratic, site-dependent results. As
noted in Airborne Particles (National Academy of Sciences, 1979), b .is
SCd l>
also significantly affected by relative humidity and the particle refractive
index.
2.5.3.3 Condensation Nuclei Counter—The condensation nuclei counter measures
the total light scatter of submicron particles by increasing their size by
condensing vapor from a cloud chamber onto their surface. This device is of
2-142
-------
o
I
E
O
E
a.
(0
O
2-35. Light scattering per unit volume of aerosol material as a function of particle size, integrated over all
wavelengths for a refractive index, m=1.5.
SOURCE: BELZ and TUVE, 197Q51
Source: Bolz and Turve (1970)
2-143
-------
interest in examining the number of particles in the nuclei mode, but is not
useful for particle sizes above about 0.5 urn (Perera and Ahmed, 1978).
2.5.3.4 Particle Size Distribution Counters—Single particle counters provide
discrete size-distribution information, but most are classed as research
measurement devices, not routine monitors, and cover only selected portions of
the size spectrum. The Electrical Aerosol Analyser, as originally described
by Whitby and Clark (1966), measures the electrical mobility of particles as
related to their size. Within several minutes, this device provides a detailed
size distribution over the range of approximately 0.01 to 0.5 urn electrical
diameter (Liu and Pui, 1975). The unit must be empirically calibrated to
obtain the relationship to aerodynamically sized particles. The Diffusion
Battery, as described by Sinclair et al. (1979), is a set of tubes through
which the air stream flows to produce selected differential particle removal
by diffusion to the walls as a function of particle size. A condensation
nuclei counter is used as the particle counter. This diffusion separation
principle is useful only in the range from 0.01 to 0.1 pm.
Optical versions of single particle counters direct the flow stream
through a small nozzle such that the light scatter from single particles can
be measured. This scatter produces a signal which is related to the size of a
spherical particle which scatters an equal amount of light. These devices
produce a size spectrum of particle sizes from about 0.5 to 10 (jm (Whitby and
Willeke, 1979). Calibration with monodispersed particles is required. For
sampling of particles larger than 10 urn, modification of commercially available
devices would be required. Mass concentrations for specific size ranges can
be estimated by selecting an appropriate particle density.
2-144
-------
2.5.3.5 Visibility Monitors—Visibility is usually interpreted to mean visual
range; however, this quantity is not directly measurable as described by Malm
(1979). The measurements of light scattering, color, and contrast changes can
be used individually to provide a measure of visibility change, but a combination
of measurements may be required to provide reliable correlation with human eye
observation. In urban areas, the reduction in visual range may also be aggravated
by N0? absorption. Friedlander (1977) shows that the extinction coefficient
(b) is equal to the sum of the coefficients for scattering (b .) and absorption
(b . ). Therefore, the nephelometer, which measures only b ., may not
provide a true indication of visible range at all locations. Another drawback
to visual range estimation with the nephelometer is the fact that it is a
point measurement integrated over a viewing angle range rather than an integrated
measurement over a measured distance. The interrelation of b . to visibility
is very complex, and is discussed in detail by Malm (1979) and in Airborne
Particles (National Academy of Sciences, 1979).
Long path optical measurement devices for ambient air are available to
examine an aspect of visibility over a defined distance. Transmissometers
measure the attenuation of transmitted light resulting from scattering and
absorption in the atmosphere. These devices are similar to their in-stack
counterparts, requiring either a light source and receptor or light source and
retroreflector at separate locations. Telephotometers measure the contrast
caused by brightness differences between a distant object and its surroundings.
The advantages and disadvantages of these long path devices compared with
other visibility monitors is currently under investigation by EPA.
2-145
-------
2.5.4 Summary - Measurement Techniques for Particulate Matter
Measurement of atmospheric particulate matter includes quantification of
the mass concentration, categorization of the particulate mass into various
particle size ranges of interest, quantitative chemical analysis, and to a
lesser extent, determination of direct effects such as reduction in visibility.
The majority of particulate measurements are made by the integrated collection
of particles from a measured volume of air over a significant time interval
using either filtration or impaction on a substrate material. The collected
particulate samples are usually analyzed gravimetrically to provide a direct
measure of mass concentration when related to the volume of air sampled.
Non-gravimetric techniques must be calibrated in mass units by empirical
comparison with gravimetrically analyzed samples. Most of the mass measurement
techniques are non-destructive, such that the particulate sample is available
for chemical analysis if the quantity of sample permits.
Current ambient particle size discrimination is typically accomplished by
specific geometrical design of the sampler inlet. Following or in conjunction
with this mechanical separation, particles of one or more size ranges are
collected on separate substrates. Most samplers have an inherent upper limit
on the size of particles collected, since large particles are very difficult
to transport to the collection surface without substantial losses. This limit
may be as low as 4 to 5 pm or as high as 50 to 60 |jm, depending on the sampler
configuration.
The high-volume sampler as specified in the current EPA reference method
is widely used for gravimetric analysis of mass concentrations. Operating for
3
periods up to 24-hours at a flowrate of over 1 m /min, the sample is large
enough to be suitable for a wide variety of chemical analyses. The sampler
2-146
-------
design is unsophisticated and a number of shortcomings have been reported.
The particle collection characteristics have only recently been quantified. A
recently-designed size-selective inlet head is available for the high-volume
sampler to restrict the collected particulates to the 0 to 15 urn inhalable
particulate range.
The optically based British Smoke Shade and AISI tape samplers are also
simple in design. These techniques have been widely used in early monitoring
studies. Efforts to compare them with high-volume sampler mass concentration
measurements are hampered by the differences in particulate size range collected.
Multi-stage cascade impactors segregate various particle size ranges by
controlled jet velocities and impaction of the particles on collection surfaces
for gravimetric analysis. Both high flow (0.6 to 1.1 m /min) and low flow
3
(0.01 to 0.04 m /min) impactors have been used to develop particle mass size
distributions, but the inlet cutoff characteristics of most impactors have not
been adequately determined. The devices are not widely used in monitoring
networks because of manual labor requirements for sampling and analysis.
Cyclone samplers are also used for ambient particle size fractionation, usually
providing a cut point in the 2 to 4 urn range.
The dichotomous sampler is of recent, more sophisticated design and
provides particulate samples in two size ranges: 0 to 2.5 pro and 2.5 to 15 urn.
The 15 urn cutoff coincides with the range currently defined by EPA for Inhalable
Particulates (IP) and is obtained by inlet geometry. The 2.5 pm cut is achieved
by the principle of "virtual impaction," and provides coarse and fine particle
fractions helpful in separating the atmospheric bimodal particle distribution
and identification of particulate sources. The low sample flow (about 1
2-147
-------
m3/hour) results in a small mass of collected material as compared with the
high-volume sampler and requires micro-balance weighing techniques.
Dynamic aerosol calibration of particulate samplers is not feasible
except under laboratory conditions, but the sampler's flow indicator can be
calibrated against a reference volumetric standard. The performance character-
istics of particulate samplers can be determined under laboratory conditions
with artifically generated monodispersed particles using a wind tunnel.
Gravimetric analysis of particulate samples requires filter conditioning
to minimize interference from moisture and static change. Other sources of
error include formation of artifact particulate matter, loss of volatiles, air
flow variation during sampling and weighing errors. Optical analysis involves
the use of an appropriate reflectance or transmittance meter with suitable 0
and 100 percent references as prescribed in the respective procedures.
Manual wet-chemistry analyses of particulate samples are being rapidly
replaced with automated spectroscopic methods. Some of the methods such as
atomic absorption require transfer of the particulate matter to a liquid
media. Other methods such as X-ray fluorescence spectrometry provide non-
destructive analyses. Most of the methods provide elemental, cation, or aniori
identification and not information about chemical compounds or oxidation
state. Atomic absorption is widely used for quantitative analysis for a
variety of elements with each element analyzed sequentially. Optical emission
spectrometry can also determine a variety of elements, but simultaneously and
with fewer interferences than atomic absorption. Spark source mass spectrometry
can be used for almost every nonvolatile element, but is not suitable for the
rapid analysis of large numbers of samples.
2-148
-------
Similarly neutron activation analysis is not rapid and some elements of
particular interest such as lead and sulfur are difficult to determine. One
of the more common non-destructive multi-element analysis techniques for
elements heavier than potassium is x-ray spectrometry. Particle samples must
be thinly deposited and preferably collected on a membrane filter substrate.
Other less commonly used elemental analysis techniques include electrochemical
methods such as potentiometry and anodic stripping voltammetry.
Analysis of a particulate sample for organics may include an estimate of
the fraction of solvent (e.g., benzene) extractable organic material or identi-
fication of individual organic species. Benzo-crpyrene (BaP) is a commonly
determined organic analyzed by thin layer chromatography with fluorescence
detection.
Methods for continuous or i_n situ monitoring of atmospheric particulates
are available. Such measurements are not as closely related to mass concentra-
tion as the integrated methods, but provide useful information for studying
particulate sources, transport, episodes, and effects such as visibility. The
integrating nephelometer measures light scattering by atmospheric particles
and is commonly used as an indicator of visibility. Its measurements are most
sensitive to fine particles and show reasonable correlation with integrated
samples of fine particles. Research monitoring devices which have been used
to provide discrete particle number, volume, and/or surface area size distri-
bution information include the electrical aerosol analyzer, the diffusion
battery, and single-particle optical counters.
Visibility as detected by the human observer is actually a combination of
factors including light scattering, color change, and loss of contrast.
Visibility monitors, such as the integrating nephelometer, measure only light
2-149
-------
scatter and do not provide an accurate indication of visual range. Long-path
instruments, now under development, may provide better measurements of visibility.
Historical interpretations of total suspended particulate data as collected
by the high volume sampler have often been clouded because this sampler is
sensitive to ambient wind speed and direction. In situations where the mass
contribution of large particles in the size distribution was very small, the
effect of these errors was minimal. During periods of elevated winds or near
fugitive dust sources, however, the high volume sampler can underestimate the
true mass concentration substantially.
Extremely sharp inlet cutoff for size selective particle samplers are
currently not a reality, and the acceptability of less than perfect size
separation must be considered. Historical development work in particle
separation techniques has concentrated on particle sizes less than 10 urn and
specifically in the 2 to 4 urn range. Separation techniques for this latter
range, including virtual impaction, inertial impaction, and cyclone separation,
provide relatively sharp cutpoints. Since this size range tends to be the
minimum in the size distribution between the accumulation and coarse particle
modes, small differences in sharpness are often academic when considering
total mass collected.
Particle separation for particles above 10 urn has been ignored until only
recently when EPA defined the 0-15 urn inhalable particulate fraction. Techniques
are currently commercially available to provide a 15 urn cutoff, such as the
dichotomous sampler and a size selective inlet for high volume samplers. The
current state-of-the-art, however, does not permit the sharpness of separation
for large particles as is possible with the smaller particles. An assessment
must therefore be made as to the utility of these less-than-ideal efficiencies
for collection of samples to support specific monitoring objectives.
2-150
-------
2.6 SUMMARY
Methods for the measurement of sulfur dioxide can be classified as
integrated or continuous. Integrated methods can be direct or based on
techniques involving absorption and stabilization in aqueous solution or
adsorption and stabilization on a solid substrate. The analysis of the
collected sample is commonly based on colorimetric, titrimetric, turbidi-
metric, gravimetric, and ion chromatographic measurement principles. The most
widely used integrated method for the determination of atmospheric sulfur
dioxide is an improved version of the colorimetric method developed by West
and Gaeke and adopted as the EPA reference method in 1971. Sulfation methods,
based on the reaction of SO^ with lead peroxide paste to form lead sulfate,
have commonly been used to estimate ambient SCL concentration over extended
time periods. Continuous methods for the measurement of ambient levels of
sulfur dioxide have gained widespread use in the air monitoring community.
Continuous sulfur dioxide analyzers using the techniques of flame photometric
detection, fluorescence, and second derivative spectrometry have been developed
over the past ten years and are commercially available.
Methods for determining soluble sulfates, total sulfates and specific
sulfate species involve the collection of particulate matter and its subsequent
analysis by direct or indirect methods. For trace soluble sulfate determinations,
a commercial method based on ion exchange chromatography exists that is specific,
exceptionally accurate, and sensitive. Methods for analysis of total sulfur
include x-ray fluorescence, electron spectroscopy, and flame photometry.
X-ray fluorescence methods are non-destructive and applicable to the determina-
tion of large numbers of ambient aerosol samples. Procedures for determining
specific sulfate species include thermal volatilization and solvent extraction
2-151
-------
techniques, gas phase ammonia titration, infrared and visible spectrometry,
flame photometry and electron microscopy. Several semiquantitative methods
for estimating sulfate species include gas phase ammonia titration technique
methods based on infrared spectroscopy and electron microscopy techniques.
The majority of particulates in the atom are collected using either
filtration or impaction methods. The collected particulate samples are usually
analyzed gravimetrically to provide a direct measure of mass concentration
when related to the volume of air sampled. The most widely used method for
gravimetric analysis of mass concentrations is the high-volume sampler, the
current EPA reference method. Other methods include the British Smoke Shade
and AISI tape samplers, multi-stage cascade impactors, cyclone samplers, and
the dichotomous sampler. Chemical analyses consist of manual wet-chemistry,
atomic absorption, x-ray fluorescence, optical emission spectrometry, spark
source mass spectrometry, neutron activation analysis, and thin layer chromato-
graphy with fluorescence detection. Methods for continuous or i_n situ monitoring
of atmospheric particulates are also available. They are not as closely
related to mass concentration as the integrated methods, but they provide
useful information for studying particulate sources, transport, episodes, and
effects such as visibility. The integrating nephelometer measures light
scattering by atmospheric particles and is commonly used as an indicator of
v i s i b i 1 i ty.
2-152
-------
2.7 REFERENCES
Adams, F. C., and R. E. Van Grieken. Absorption correction for X-ray
fluorescence analysis of aerosol loaded filters. Anal. Chem. 47:1767-1773,
1975.
Adamski, J. M., and S. P. Villard. Application of the methylthymol blue
sulfate method to water and wastewater analysis. Anal. Chem. 47:1191-1194,
1975.
Ahearn, A. I. Trace Analysis by Mass Spectrometry. Academic Press, New York,
NY, 1972.
Ahlquist, N. C., and R. J. Charlson. A new instrument for evaluating the
visual quality of air. J. Air Pollut. Control Assoc. 17:467-469,
1967.
American Society for Testing and Materials. Standard method for collection
and analysis of dustfall. In: Book of ASTM Standards. American Society
for Testing and Materials, Philadelphia, PA, 1966. pp. 785-788.
Andrezejewski, R. Les Proprietes physiques des poussiered, Ed. Slask, Katowice,
1968.
Appel, B. R. , E. L. Kothny, E. M. Hoffer, and J. J. Wesolowski. Comparison of
Wet Chemical and Instrumental Methods for Measuring Airborne Sulfate.
EPA-600/2-76-059, U.S. Environmental Protection Agency, Research Triangle
Park, NC, March 1976.
Archer, A. W. The indirect colorimetric determination of sulphate with
2-aminoperimidine. Analyst (London) 100:755-757, 1975.
Barringer Research, Ltd., A report to Department of Health, Education and
Welfare of optical measurements of S0? and N02 air pollution using Barringer
Correlation Spectrometers. Barringer Researcn Technical Report TR-69-113.
Canada, 1969.
Barton, S. C., and H. G. McAdie. A specific method for the automatic determina-
tion of ambient H^SO. aerosol. In: Proceedings of the 2nd International
Clean Air Congress. H. M. Englund and W. T. Berry, eds. , Academic Press,
New York, NY, 1971. pp. 379-382.
Barton, S. C., and H. G. McAdie. An automated instrument for monitoring
ambient HLSO. aerosol. In: Proceedings of the 3rd International Clean
Air Congress? VDI-Verlag GmbH, Dusseldorf, West Germany, 1973. pp. C25-C27.
Barnard, W. F. Community Health Environmental Surveillance Studies (CHESS) Air
Pollution Monitoring Handbook. EPA-600/1-76-011, U.S. Environmental
Protection Agency, Research Triangle Park, NC, January 1976.
Benarie, M., B. T. Chuong, and A. Nonat. Essai d'une Methode pour le Dosage
Selectif de 1'Acide Sulfurique dans 1'Atmosphere de Stransbourg et Defini-
tion du Role des Particules Solides dans 1'Oxydation du SCL dans 1'Atomsphere.
Report of IRCHA Study 7106, Centre de Recherche, Vert-le-Petit, France,
January 16, 1976.
2-153
-------
TARI E 2-11 FRACTIONAL AEROSOL PENETRATION FOR SELECTED SUBSTRATES
TABLE 2 1L ^jj^ QF FACE VELOCITY AND PARTICLE SIZE
•
FILTER: Gelman Tvpe A, gl
AP, cm Hg
V, cm/sec
Dp, Mm
0.035
0.10
0.30
1.0
FILTER: Ghia S2
AP, cm Hg
V, cm/sec
1
11.2
_— — ^— — — — —
ass fiber
1.5
16.9
_
3 10
32.7 108
PENETRATION
<0.0001
<0.0001
<0.0001
<0.0001
37PJ 02,
1
23.4
<0.0001
<0.0001
<0.0001
<0.0001
teflon membrane,
3
64.1
<0.0001 0.0008
<0.0001 0.00054
<0.0001 <0. 00007
<0.0001 <0. 00002
2.0 Mil pore
10
187
Dp, Mm
0.035
0.10
0.30
1.0
FILTER: Whatman
AP, cm Hg
V, cm/sec
Dp, Mm
0.035
0.10
0.30
1.0
PENETRATION
<0.0002
<0. 00006
<0. 00007
<0. 00007
No.l, eel
1
6.1
0.0011
0.00008
<0. 00007
<0. 00009
lulose fiber
3
17.4
0.0005
<0. 00024
<0. 00022
<0. 00008
10 30
47.6 102
PENETRATION
0.56
0.46
0.16
0.019
0.52
0.43
0.044
0.034
0.34 0.058
0.13 0.0071
0.0049 0.00051
0.0044 0.00042
Source: Liu et al. , 1978
2-90a
-------
Benarie, M., T. Menard, and A. Nonat. Etude de la transformation de 1'a sulfureux
en acide sulfurique en relation avec les donnees climatologiques, dans un
ensemble urban a caractere industriel, Rouen. Atmos. Environ. 7:403-421,
1973.
Bertolacini, R. J., and J. E. Barney II. Colorimetric determination of sulfate
with barium chloranilate. .Anal Chem. 29:281-283, 1957.
Blanco, A. J., and G. B. Hoidale. Microspectrophotometric technique for
obtaining the infrared spectrum of microgram quantities of atmospheric
dust. Atmos. Environ. 2:327-330, 1968.
Blanco, A. J., and R. G. Mclntyre. An Infra-red spectroscopic view of atmo-
spheric particulates over El Paso, Texas. Atmos. Environ. 6:557-562, 1972.
Bogen, D. C., and G. A. Welford. Radiometric determination of sulfate. J.
Radioanal. Chem., 1977.
Bokhaven, C., and H. J. Niessin. Int. J. Air Water Pollut. 10:223, 1966.
Bolleter, W. T., C. J. Bushman, and P. W. Tidwell. Anal. Chem. 33:592, 1961.
Bolz, R. E., and G. E. Tuve, eds. Handbook of Tables for Applied Engineering
Science. Chemical Rubber Co., Cleveland, OH, 1970. p. 159.
Bostrom, C. E. Air Water Pollut. 10:435, 1966.
Bowden, S. R. Improved lead dioxide method of assessing sulphurous pollution
of the atmosphere. Air Water Pollut. 8:101-106, 1964.
British Standards Institution. Methods for the measurement of Air Pollution.
Part 3. Determination of Sulphur Dioxide. B.S. 1747, Part 3, British
Standards Institution, London, England, 1963.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 2. Determination of Concentration of Suspended Matter. B.S. 1747,
Part 2, British Standards Institution, London, England, 1964.
Brock, J. R., and Marlow. Charged aerosol particles and air pollution.
Environ. Letters, No. 10, 1975.
Brosset, C., K. Andreason, and M. Perm. Atmos. Environ. 9:631- , 1975.
Brosset, C., and M. Perm. Man-made airborne acidity and its determination.
Atmos. Environ. 12:909-916, 1978.
Buttell, R. G., and Brewer J. Sci. Inst. 26:357, 1949.
Cahill, T. A. Comments on surface coatings for Lundgren - Type impactors.
In: Aerosol Measurement. University Presses of Florida, Gainesville,
PL, 1979. p. 131.
2-154
-------
Calvert, S., and W. Workman. Talanta 4:89, 1960.
Camp, D. C., A. L. Van Lehn, and B. W. Loo. Intercomparison of Samplers Used
in the Determination of Aerosol Composition. EPA-600/7-78-118, U.S.
Environmental Protection Agency, Research Triangle Park, NC, 1978.
Charlson, R. J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. H9SO./(NH,.)? SO. background aerosol: Optical detection in
St. Louis regiort. AlmCs. Environ. 8:1257-1267, 1974.
Charlson, R. J., N. C. Ahlquist, and N. Horvath. On the Generality of Correlation
of Atmospheric Aerosol Mass Concentration and Light Scatter. Atmos.
Environ. 2:455-464, 1968.
Charlson, R.J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. Sulfuric acid-ammonium sulfate aerosol: Optical detection in
the St. Louis region. Science 184:156-158, 1974b.
Cheney, J. L., and J. B. Homolya. Sampling parameters for sulfate measurement
and characterization. Environ. Sci. Technol. 13:584-588, 1979.
Conner, W. D. An Inertial-Type Particle Separator for Collecting Large Samples.
J. Air Pollut. Control Assoc.. 16:35-38, 1966.
Coutant, R. W. Effects of environmental variables on collection of atmospheric
sulfate. Environ. Sci. Technol. 11(9):873-878, September 1977.
Crider, W. L. Anal. Chem. 37:1770, 1965.
Crider, W. L., N. P. Barkley, M. J. Knott, and R. W. Slater. Hydrogen flame
chemiluminescence detector for sulfate in aqeous solutions. Anal. Chim.
Acta 47:237-241, 1969.
Cunningham, P. T., and S. A. Johnson. Spectroscopic observation of acid
sulfate in atmospheric particulate samples. Science 191:77-79, 1976.
Gushing, K. M., G. E. Lacey, J. D. McCain, and W. B. Smith. Particulate
Sizing Techniques for Control Device Evaluation: Cascade Impactor Calibrations.
EPA-600/2-76-280, U.S. Environmental Protection Agency, Research Triangle
Park, NC, February 1978.
Davies, C. N., ed. Aerosol Science. Academic Press, New York, NY, 1966.
Dennis, E. R., ed. Handbook on Aerosols. National Technical Information
Service, TID-26608, Arlington, VA, 1976.
deVeer, S. M., H. J. Bronwer, and H. Zeedijk. Presented at the 62nd Annual
Meeting, Air Pollution Control Association, New York, New York, June
1969. Paper No. 69-6.
Dimitt, R. L., and E. R. Graham. Anal. Chem. 48:604, 1976.
2-155
-------
Dionex Corporation. Dionex Analytical Ion Chromatographs. Technical Brochure,
Dionex Corp., Palo Alto, California, 1975.
Dubois, L., R. S. Thomas, T. Teichman, and J. L. Monkman. A general method of
analysis for high volume air samples. I. Sulphate and sulphuric acid.
Mikrochim. Acta 269:1268-1275, 1969.
Dzubay. T. G., H. L. Rook, and R. K. Stevens. A chemiluminescent approach to
measurement of strong acid aerosols. Jji: Analytical Methods Applied to
Air Pollution Measurements. R. K. Stevens and W. F. Herget, eds., Ann Arbor
Science Publishers, Inc., Ann Arbor, MI, 1974. pp. 71-83.
Dzubay, T. G., R. K. Stevens, and C. M. Paterson. Application of the dichotomous
sampler to the characterization of ambient aerosols. J_n: X-ray Fluo-
resence of Environmental Samples. T. G. Dzubay, ed., Ann Arbor Science
Publishers, Inc., Ann Arbor, MI, 1977. pp. 95-105.
Dzubay, T. G., G. K. Snyder, D. J. Reutter, and R. K. Stevens. Aerosol acidity
determination by reaction with C labeled amine. Atmos. Environ.
13:1209-1212, 1979.
Dzubay, T. G., L. E. Mines, and R. K. Stevens. Particle bounce errors in
cascade impactors. Atmos. Environ. 10:229-234, 1976.
Eagan, M. L. , and L. Dubois. Anal. Chim. Acta 70:157, 1974.
Falkowski, J. M. Suszka i uwlaznisnije lubovoloknistych matierijalow. Moscow,
GNTI, 1951.
Fassel, V. A. Quantitative elemental analyses by plasma emission spectroscopy.
Science. 202:183-191, 1978.
Federal Register, 40(157):34024, 1975.
Friedlander, S. K. Smoke, Dust, and Haze. Wiley Interscience Publication,
New York, NY, 1977. p. 154.
Fritz, S. J., and S. S. Yamamura. Anal. Chem. 27:1461, 1955.
Fuerst, R. G., F. P. Scaringelli, and J. H. Margeson. Effect of Temperature
on Stability of Sulfur Dioxide Samples Collected by the Federal Reference
Method. U.S. Environmental Protection Agency, Research Triangle Park,
NC, EPA-600/4-76-024, May 1976.
Gales, M. E. , Jr., W. H. Kaylor, and J. E. Longbottom. Determination of sulphate
by automatic colorimetric analysis. Analyst (London) 93:97-100, 1968.
Gilbert, T. R., and A. M. Clay. Determination of ammonia in aquaria and in
sea water using the ammonia electrode. Anal. Chem. 45:1757-1759, 1973.
Goold, R. W., C. S. Barrett, J. B. Newkirk, and C. 0. Ruud. Advances in X-Ray
Analysis. Kendall/Hunt, Dubuque, IA, 1976.
2-156
-------
Goulding, F. S., and J. M. Jaklevic. X-ray fluorescence spectrometer for
airborne participate monitoring. EPA Report No. EPA-R2-73-182, U.S.
Environmental Protection Agency, Washington, DC, April, 1973.
Goulding, F. S., J. M. Jaklevic, and B. W. Loo. Aerosol Analysis for the
Regional Air Pollution Study. EPA-600/4-78-034, U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1978.
Hager, R. N., Jr., and R. C. Anderson. J. Opt. Soc. Am. 60:1444, 1970.
Hansen, L. D., L. Whiting, D. J. Eatough, T. E. Jensen, and R. M. Izatt. Anal.
Chem. 48:634, 1976.
Harris, B. Second Stage Development of an Automated Field Sulfuric Acid
Sampler and the Development of a Sulfuric Acid Analyzer. Monthly Technical
Progress Narrative No. 6, Cabot Corporation, Bill erica, MO, December
1975.
Heard, M. J., and R. D. Wiffen. Electron microscopy of natural aerosols
and the identification of particulate ammonium sulphate. Atmos. Environ.
3:337-340, 1969.
Hemeon, W. C. L., G. F. Haines, Jr., and H. M. Ide. Determination of haze and
smoke concentrations by filter paper samples. Air Repair 3:22-28, 1953.
Hesketh, H. E. Understanding and Controlling Air Pollution, Ann Arbor Science
Publishers, Michigan, 1974.
Hill, A. S. G. Measurement of the optical densities of smokestains on filter
papers. Trans. Farad. Soc. 32:1125-1131, 1936.
Hochheiser, S. Methods of Measuring Atmospheric Sulfur Dioxide. 999-AP-6,
U.S. Department of Health, Education, and Welfare, Public Health Service,
Cincinnati, OH, August 1964.
Hoffer, E., and E. L. Kothny. A Micromethod for Sulfate in Atmospheric
Particulate Matter. Report No. 163, Air and Industrial Hygiene Labora-
tory, Berkeley, CA, July 1974.
Hounam, R. E., and R. J. Sherwood. The cascade centripeter: a device for
determining the concentration and size distribution of aerosols. J. Am.
Ind. Hyg. Assoc. 26:122-131, 1965.
Huey, N. A. The lead dioxide estimation of sulfur oxide pollution. J. Air
Pollut. Control Assoc. 18:610-611, 1968.
Huey, N. A., M. A. Wallar, and C. D. Robson. Field evaluation of an improved
sulfation measurement system. Presented at the 62nd Annual Meeting, Air
Pollution Control Association, New York, NY, June 22-26, 1969. paper
no. 69-133.
2-157
-------
Huntzicker, J. J., L. J. Isabella, and J. G. Watson. The continuous monitoring
of particulate sulfate by flame photometry. JJK International Conference
on Environmental Sensing and Assessment. Volume 2, Institute of Electrical
and Electronics Engineers, Inc., New York, NY, 1976. paper 23-4.
Huntzicker, J. J. , L. M. Isabelle, and J. G. Watson. The continuous monitoring
of particulate sulfur compounds by flame photometry. Presented at the
173rd National Meeting, American Chemical Society, New Orleans, LA,
March 20-25, 1977. paper ENVR-39.
Husar, J. D., R. B. Husar, and P. K. Stubits. Determination of submicrogram
amounts of atmospheric particulate sulfur. Anal. Chem. 47:2062-2065, 1975.
Husar, R. B., J. D. Husar, N. V. Gillani, S. B. Fuller, W. H. White, J. A.
Anderson, W. M. Vaughan, and W. E. Wilson, Jr. Pollutant flow rate
measurement in large plumes: sulfur budget in power plant and area source
plumes in the St. Louis region. Presented at the 171st Meeting. American
Chemical Society, New York, NY, April 4-9, 1976. paper ENVR-14.
Huygen, C. A simple photometric determination of sulphuric acid aerosol.
Atmos. Environ. 9:315-319, 1975.
Intersociety Committee. Methods of Air Sampling and Analysis. American
Public Health Association, Washington, DC, 1972.
Intertech Corp. Descriptive brochure.
Jacobs, M. B. The Chemical Analysis of Air Pollutants. Wiley-Interscience,
New York, NY, 1960.
Jacobs, M. B., M. M. Braverman, and S. Hochheiser. Anal. Chem. 29:1349-
1957.
Johansson, T. B., R. E. Van Grieken, J. W. Nelson, and J. W. Winchester. Anal.
Chem. 47:855, 1975.
John, W., G. Reischl, and J. Wesolowski. Size Selective Monitoring Techniques
for Particulate Matter. Report No. ARB A5-00487 California Department of
Health, Air and Industrial Hygiene Laboratory, Berkeley, CA, February
1978.
Katz, M. Anal. Chem. 22:1040, 1950.
Katz, M. In: Analysis of Inorganic Gaseous Pollutants. Air Pollution. Vol. II
2nd ed. A. C. Stern, ed., Academic Press, New York, NY, 1968. pp. 53-114.
Katz, M. Measurement of Air Pollutants—Guide to Selection of Methods. World
Health Organization, Geneva, 1969.
Keay, J., and P. M. A. Menage. Analyst (London) 95:379, 1970.
2-158
-------
Kirkland, J. J. Modern Practice of Liquid Chromatography. John Wiley, New
York, NY, 1971.
Kittleson, D. B., R. McKenzie, M Vermeersch, M. Linne, F. Dorman, D. Pui, B.
Liu, and K. Whitby. Total sulfur aerosol detection with an electrostatically
pulsed flame photometric detector system. Presented at the 173rd National
Meeting, American Chemical Society, New Orleans, LA, March 20-25, 1977.
paper ENVR-42.
Klockow, D., and G. Ronicke. An amplification method for the determination
of particle-sulphate in background air. Atmos. Environ. 7:163-168, 1973.
Kolthoff, I. M., E. B. Sandell, E. J. Meehan, and S. Bruckenstein. Quantitative
Chemical Analysis, 4th ed., MacMillan, Toronto, Canada, 1969. pp. 992-994.
Lambert, J. P. F. , and F. W. Wilshire. Neutron activation analysis for
simultaneous determination of trace elements in ambient air collected on
glass-fiber filters. Anal. Chem. 51:1346-1353, 1979.
Lamothe, P. J., and R. K. Stevens. Sulfuric acid analysis using low temperature
volatilization. Presented at the 171st National Meeting, American Chemical
Society, New York, NY, April 4-9, 1976. paper ENVR-38.
Lawrence Berkeley Laboratory. Instrumentation for Environmental Monitoring,
Air. LBL-1, Vol. 1, University of California, Berkeley, CA, May 1972.
Lawrence Berkeley Laboratory. Instrumentation for Environmental Monitoring,
AIR Part 2, University of California, CA, 1975.
Lazarus, A. L., K. C. Hill, and J. P. Lodge. In: Proceedings of the 1966
Technicon Symposium on Automation in Analytical Chemistry. Mediad, Inc.,
New York, NY, 1966. pp. 291.
Leahy, D., R. Siegel, P. Klotz, and L. Newman. The separation and characteri-
zation of sulfate aerosol. Atmos. Environ. 9:219-229, 1975.
Lee, R. E., and S. Goranson. National air surveillance cascade impactor
network. I. Size distribution measurements of suspended particulate
matter in air, Environ. Sci. Techno!., 6:1019-1024, 1972.
Lilienfield, P. J. Am. Ind. Hyg. Assoc. 31:722, 1970.
Lippman, M. "Respirable" dust sampling. Am. Ind. Hyg. Assoc. J. 31:138-159,
1970. ~~
Liu, B. Y. H., and D. Y. H. Pui. On the Performance of the Electrical Aerosol
Analyzer. J. Aerosol Sci. 6:249-264, 1975.
Liu, B. Y. H. , ed. Fine Particles, Aerosol Generation, Measurement, Sampling,
and Analysis. Academic Press, New York, NY, 1976.
2-159
-------
Liu, B. Y. H, D. Y. H. Pui, K. L. Rubow, and G. A. Kuhlmey. Progress Report -
Research on Air Sampling Filter Media. Grant Report R804600, University
of Minnesota, Minneapolis, MN, May 1978.
Lizarraga-Rocha, J. A. The Effect of Exposure of Filters During Non-operating
Periods in the High Volume Sampling Method. M.S. Thesis, University of
North Carolina, Chapel Hill, NC, 1976.
Lochmuller, C. H., J. W Galbraith, and R. L. Walter. Trace metal analysis in
water by proton-induced x-ray emission analysis of ion-exchange membranes.
Anal. Chem. 46:440-442, 1974.
Lodge, J. P., J. Ferguson, and B. R. Havlik. Analysis of micron-sized
particles: Determination of sulfuric acid aerosol. Anal. Chem. 32:
1206-1207, 1960.
Loo, B. W. , R. S. Adachi, C. P. Cork, F. S. Goulding, J. M. Jaklevic, D. A.
Landis, and W. L. Searles. A second generation dichotomous sampler for
large scale monitoring of airborne particulate matter. Presented at 86th
Annual Meeting, American Institute of Chemical Engineers, Houston, Texas,
April 1, 1979.
Loo, B. W., W. R. French, R. C. Gatti, F. S. Goulding, J. M. Jaklevic, J.
Llacer, and A. C. Thompson. Large scale measurement of airborne particulate
sulfur. Atmos. Environ. 12:759-771, 1978.
Lucero, D. P., and J. W. Pal jug. Monitoring sulfur compounds by flame photometry.
Trib. CEBEDEAU 27:139-147, 1974.
Lundgren, D. A. Mass Distribution of Larger Atmospheric Particles. Ph.D.
Thesis, University of Minnesota, Minneapolis, MN, 1973.
Lundgren, D. A., M. Lippman, F. S. Harris, W. E. Clark, W. H. Marlow, and
M. D. Durham, eds. Aerosol Measurement. University Press of Florida,
Gainesville, FL, 1979.
Macias, E. S. Instrumental analysis of light element composition of atmospheric
aerosols. _In: Methods and Standards for Environmental Measurement. W.
Kirchhoff, ed., National Bureau of Standards Special Publication 464.
Department of Commerce, Washington, DC, 1977. pp. 179-188.
Macias, E. S., and R. B. Husar. Atmospheric particulate mass measurement with
beta attenuation mass monitor. Environ. Sci. Technol. 10:904-907, 1976.
Maddalone, R. F., G. L. McClure, and P. W. West. Anal. Chem. 47:316, 1975.
Malm, W. Consideration in the measurement of visibility. J. Air Pollut.
Control Assoc. 29:1042-1052, 1979.
Mamane, Y., and R. G. de Pena. A quantitative method for the detection of
individual submicron size sulfate particles. In: Sulfur in the Atmosphere,
Dubrovnik, Proceeding of an International Symposium, Yugoslavia, September
7-14, 1977. Atmos. Environ. 12:69-82, 1978.
2-160
-------
Mamantov, G. , E. L. Wehry, R. R. Kemerer, and E. R. Hinton. Anal. Chem.
49:86, 1977.
Martin, B. E. Sulfur Dioxide Bubbler Temperature Study. EPA-600/4-77-040,
U.S. Environmental Protection Agency, Research Triangle Park, NC, August
1977.
McCoy, R. A., D. E. Camann, and H. C. McKee. Collaborative Study of Reference
Method for Determination of Sulfur Dioxide in the Atmosphere (Pararosaniline
Method) (24-Hour Sampling). EPA-650/4-74-027, U.S. Environmental Protection
Agency. Research Triangle Park, NC, December 1973.
McFadden, W. H. Techniques of Combined Gas Chromatography/ Mass Spectrometry.
John Wiley, New York, NY, 1973.
McFarland, A. R. Test Report - Wind Tunnel Evaluation of British Smoke Shach
Sampler. Air Quality Laboratory Report, 3565/05/79/ARM, Texas A&M
University, College Station, TX, May 1979.
McFarland, A. R. and C. E. Rodes. Characteristics of Aerosol Samplers Used in
Ambient Air Monitoring. Presented at 86th National Meeting, American
Institute of Chemical Engineers, Houston, TX, April 2, 1979.
McFarland, A. R., and C. A. Ortiz. Progress Report - Aerosol Characterization
of Ambient Particulate Samplers Used in Environmental Monitoring Studies.
Texas A&M University Research Foundation, College Station, TX, October
1979.
McFarland, A. R., C. A. Ortiz, and C. E. Rodes. Evaluation of the NASN Cascade
Impactor Ambient Air Monitor. Report 3565/07/78/ARM, Texas A&M, Air
Quality Laboratory, College Station, TX, 1978. Accepted Publication Atm.
Environ., 1980.
McKee, H. C., R. E. Childers, and 0. Saenz, Jr. Collaborative Study of Reference
Method for the Determination of Suspended Particulates in the Atmosphere
(High Volume Method). APTD-0904, U.S. Environmental Protection Agency,
Research Triangle Park, NC, June 1971.
Miller, F. J., D. E. Gardner, J. A. Graham, R. E. Lee, Jr., W. E. Wilson, and
J. D. Bachmann. Size considerations for establishing a standard for
inhalable particles. J. Air Pollut. Control Assoc. 29:610-615, 1979.
Moffat, A. J., J. R. Robbins, and A. R. Barringer. Electro-optical sensing
of environmental pollutants. Atmos. Environ. 5:511-525, 1971.
Morrison, G. H. Trace Analysis Physical Methods. Wiley Interscience, New
York, NY, 1965. pp. 245-263, 271-310, 325-360, 377-425.
Mudgett, P. S., L. W. Richards, and J. R. Roehrig. A new technique to measure
sulfuric acid in the atmosphere. In: Analytical Methods Applied to Air
Pollution Measurement. R. K. Stevens and W. F- Herget, eds., Ann Arbor
Science Publishers, Inc., Ann Arbor, MI, 1974.
2-161
-------
Mulik, J. D., G. Todd, E. Estes, R. Puckett, E. Sawicki, and D. Williams. In:
Ion Chromatographic Analysis of Environmental Pollutants. E. Sawicki, J.
D. Mulik, and E. Wittgenstein, eds., Ann Arbor Science Publishers, Inc.,
Ann Arbor, MI, 1978.
Nader, J. S. Dust retention efficiencies of dustfall collectors. J. Air
Pollut. Control Assoc. 8:35-38, 1958.
National Academy of Sciences. Airborne Particles. University Park Press,
Baltimore, MD, 1979.
National Air Pollution Control Administration Publication. Air Quality Criteria
for Particulate Matter. AP-49, U.S. Department of Health, Education, and
Welfare, Washington, DC, January 1969.
Natusch, D. F. S. , C. F. Bauer, and A. Loh. Collection and analysis of trace
elements in the atmosphere. In: Air Pollution Control, Part III.
W. Strauss, ed., John Wiley and Sons, Inc., 1978.
Newman, L., J. Forrest, and B. Manowitz. The application of an isotopic
ratio technique to a study of the atmospheric oxidation of sulfur dioxide
in the plume from an oil-fired power plant. Atmos. Environ. 9:959-968,
1975.
Newman, L., J. Forrest, and B. Manowitz. Atmos. Environ. 9:969, 1975b.
Novakov, T. Chemical characterization of atmospheric pollution particulates by
photoelectron spectroscopy. In: Proceedings of the 2nd Joint Conference
on Sensing of Environmental Pollutants, Instrument Society of America,
Washington, DC, December 10-12, 1973. p. 197.
Novakov, T., S. G. Chang, and A. B. Marker. Sulfates as pollution particulates:
Catalytic formation on carbon (soot) particles. Science 186:259-261, 1974.
Okabe, H., P. L. Splitstone, and J. J. Ball. J. Air Pollut. Control Assoc.
23:514, 1973.
O'Keeffe, A. E., and G. C. Ortman. Primary standards for trace gas analysis.
Anal. Chem. 38:760-763, 1966.
Organization for European Economic Cooperation. Paper EPA/AR/4283, European
Productivity Agency, Paris, France, 1961.
/•-•
Palmer, H. F., C. E. Rodes, and C. J. Nelson. Performance characteristics of
instrumental methods for monitoring sulfur dioxide. II. Field Evaluation.
J. Air Pollut. Control Assoc. 19:778, 1969.
Perera, F. P., and A. K. Ahmed. Respirable Particles: Impact of Airborne
Fine Particulates on Health and the Environment. National Resources
Defense Council, Washington, DC, October 1978.
2-162
-------
Philips Electronic Instruments. Coulometric amperometric analyzers. Bull.
WA-4-C13.
Pierson, W. R., W. W. Prachaczek, T. J. Korniski, T. J. Truex, and J. W.
Butler. Artifact formation of sulfate, nitrate, and hydrogen ion on
backup filters: Allegheny Mountain experiment. J. Air Pollut. Control
Assoc. 30:30-34, 1980.
Ranade, M. B., and D. VanOsdell. Quality Assurance Evaluation of the CHAMP
Aerosol Sampler, Project Report 43U-1487-59, Research Triangle Institute,
Research Triangle Park, NC, July 1978.
Regan, G. F., S. K. Goranson, and L. L. Lawson. Use of tape samplers as fine
particulate monitors. J. Air Pollut. Control Assoc. 29:1158-1160, 1979.
Rider, P. E., C. C. Rider, and R. N. Corning. Studies of Huey sulfation plates
at high ambient S09 concentrations. J. Air Pollution Control Assoc.
27:1011-1013, 1977?
Roberts, P. T., and S. K. Friedlander. Analysis of sulfur in deposited aerosol
particles by vaporization and flame photometric detection. Atmos. Environ.
10:403-408, 1976.
Rodes, C. E., and G. F. Evans. Summary of LACS Integrated Measurements.
EPA-600/4-77-034, U.S. Environmental Protection Agency, Research Triangle
Park, NC, June 1977.
Ross, J. W., Jr., and M. S. Frant. Potentiometric titrations of sulfate using
an ion-selective lead electrode. Anal. Chem. 41:967-969, 1969.
Sajo, I., and B. Sipas. Talanta 14:203, 1967.
Scaringelli, F. P., and K. A. Rehme. Determination of atmospheric concentrations
of sulfuric acid aerosol by spectrophotometry, coulometry, and flame
photometry. Anal. Chem. 41:707-713, 1969.
Scaringelli, F- P., B. E. Slatzman, and S. A. Frey. Spectrophotometric
determination of atmospheric sulfur dioxide. Anal. Chem. 39:1709-1719,
1967.
Scroggins, L. H. Collaborative study of the microanalytical oxygen flask
sulfur determination with dimethylsulfonazo III as indicator. J. Assoc.
Off. Anal. Chem. of the AOAC 57:22-25, 1974.
Shaw, R. Personal Communication to C. E. Rodes, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1980.
Sides, J. 0., and H. F. Saiger. Effects of Prolonged Static Exposure of
Filters to Ambient Air on High Volume Sampling Results. Kansas State
Department of Health and Environment, Bureau of Air Quality and
Occupational Health, Topeka, KS, May 1976.
2-163
-------
Sinclair, D. , R. J. Countess, B. Y. H. Liu, and D. Y. H. Pui. Automatic
analysis of submicron particles. In: Aerosol Measurement. University of
Florida Presses, Gainsville, FL, 1979. p. 544-563.
Small, H., T. S. Stevens, and W. C. Bauman. Anal. Chem. 47:1801, 1975.
Smith, A. F., D. G. Jenkins, and D. E. Cunningworth. J. Appl. Chem. 11:317,
1961. ~~
Smith, F., P. S. Wohlschlegel, R. S. C. Rogers, and Mulligan, D. J. Investigation
of Flow Rate Calibration Procedures Associated with the High Volume
Method for Determination of Suspended Particulates. EPA-600/4-78-047,
U.S. Environmental Protection Agency, Research Triangle Park, NC, June
1978.
Snell, F. D. Photometric and Fluorometric Methods of Analysis, Metals. Part
1. John Wiley, New York, NY, 1978. p. 1-25.
Spicer, C. W., and P. M. Schumacher. Particle Nitrate: Laboratory and Field
Studies of Major Sampling Interferences. Atmos. Environ. 13:543-552,
1979. ~~
Stalker, W. W. , R. C. Dickerson, and G. D. Kramer. Atmospheric sulfur dioxide
and particulate matter: A comparison of methods of measurements. Am.
Ind. Hyg. Assoc. J. 24:68-7-9, 1963.
Stevens, R. K., and T. G. Dzubay. Recent Development in Air Particulate
Monitoring. IEEE Trans. Nucl. Sci. 22: April 1975.
Stevens, R. K. , J. A. Hodgeson, L. F. Ballard, and C. E. Decker. Jjn: Determination
of Air Quality. G. Mamantov and W. D. Shults, eds., Plenum Publishing
Corporation, New York, NY, 1972.
Stevens, R. K., J. D. Mulik, A. E. O'Keeffe, and K. J. Krost. Gas chromato-
graphy of reactive sulfur gases in air at the parts-per-billion level.
Anal. Chem. 43:827-831, 1971.
Stevens, R. K. , T. G. Dzubay, G. Russwurm, and D. Rickel. Sampling and analysis
of atmospheric sulfates and related species. Atmos. Environ. 12:55-68,
1978.
Stratmann, H. Mikrochim Acta 6:668- , 1954.
Swanson, D., C. Morris, R. Hedgecoke, R. Jungers, R. J. Thompson, and J. E.
Bumgarner. A rapid analytical procedure for the analysis of BaP in
environmental samples. Trends Fluoresc. 1:22-27, 1978.
Tanner, R. L., and Cordova. U.S. Energy Research Development Administration.
Contract No." EY-76-C-02-0016, 1978.
2-164
-------
Tanner, R. L., J. Forrest, and L. Newman. Determination of atmospheric
gaseous and participate sulfur compounds. U.S. Energy Research Development
Administration. Contract No. EY-76-C-02-0016, 1978.
Tanner, R. L., R. Cederwall, R. Garber, D. Leahy, W. Marlow, R. Meyers, M.
Phillips, and L. Newman. Separation and analysis of aerosol sulfate
species at ambient concentrations. Atmos. Environ. 11:955-966, 1977.
Tanner, R. L., R. W. Garber, and L. Newman. Speciation of sulfate in ambient
aerosols by solvent extraction with flame photometric detection. Presented
at the 173rd National Meeting, American Chemical Society, New Orleans,
LA, March 20-25, 1977a. paper ENVR-41.
Tanner, R. L., R. W. Garber, W. H. Marlow, B. Leaderer, D. Eatough, and M. A.
Leyko. Chemical composition and size distribution of sulfate as a function
of particle size in New York City aerosol. Presented at the Industrial
Hygiene Conference, New Orleans, Louisiana, May 22-27, 1977b. paper No.
140.
Technicon Corporation. Sulfate method VIb via turbidimetry. Technicon
Corporation, Tarrytown, NY, 1959.
Terraglio, F. P., and R. M. Maganelli. Laboratory evaluation of sulfur dioxide
methods and the influence of ozone-oxides of nitrogen mixtures.
Anal. Chem. 34:675-677, 1962.
Tetlow, J. A., and A. L. Wilson. Analyst (London) 89:453, 1964.
Thoen, G. N., G. G. Dehaas, and R. R. Austin. Tappi 51:246, 1968.
Thomas, R. L., V. Dharmarajan, G. L. Lundquist, and P. W. West. Measurement
of sulfuric acid aerosol, sulfur trioxide, and the total sulfate content
of the ambient air. Anal. Chem. 48:639-642, 1976.
Thrane, K. E. Intercomparison of analytical methods within the European
monitoring and evaluation program (EMEP). In: Studies in Environmental
Science, Vol. 2, Air Pollution Reference Measurement Methods and Systems.
T. Schneider, H. W. de Koning, and L. J. Brasser, eds., Elsevier Scientific
Publishing Company, Amsterdam, 1978. pp. 139-149.
Threshold Limits Committee. Threshold Limit Values of Air Borne Contaminants
for 1968. American Conference of Governmental Industrial Hygienists,
Cincinnati, OH, 1968.
Tierney, G. P. , and W. D. Conner. Hydroscopic effects of weight determinations
of particulates collected on glass fiber filters. J. Am. Ind. Hyg.
Assoc., 1967.
Treon, J. F., and W. E. Crutchfield, Jr. Rapid turbidimetric method for
determination of sulfates. Ind. Eng. Chem. Anal. Ed. 14:119-121, 1942.
2-165
-------
Turner, J. M., and R. S. Sholtes. Laboratory evaluation of the sulfation
plate for estimation of atmospheric sulfur dioxide. Presented at the
64th Annual Meeting, Air Pollution Control Association, Atlantic City, NJ,
June 27-July 2, 1971. Paper No 71-39.
U.S. Environmental Protection Agency. Ambient Air Monitoring Reference and
Equivalent Methods. 40 CFR 53:952-986, July 1, 1979c.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. Appendix A - Reference method for the determina-
tion of sulfur dioxide in the atmosphere (pararosaniline method). 40 CFR
50:6-11, July 1, 1979.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. Appendix B - Reference method for the determina-
tion of suspended particulates in the atmosphere (high volume method).
40 CFR 50:12-16, July 1, 1979a.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. 40 CFR 50:3-5, July 1, 1979b.
U.S. Environmental Protection Agency. Summary of Performance Test Results and
Comparative Data for Designated Equivalent Methods for S02- Document No.
QAD/M-79.12, U.S. Environmental Protection Agency, Research Triangle
Park, NC, December 1979.
Volmer, W., and F. Z. Frohlich. Z. Anal. Chem. 126:401, 1944.
Von Lehmden, D. J., and C. Nelson. Quality Assurance Handbook for Air Pollution
Measurement Systems. Volume I: Principles. EPA-600/9-76-005, U.S.
Environmental Protection Agency, Research Triangle Park, NC, January 1976.
Von Lehmden, D. J., and C. Nelson. Quality Assurance Handbook for Air
Pollution Measurement System. Volume II. Ambient Air Specific Methods.
EPA-600/4-77-027a, U.S. Environmental Protection Agency, Research Triangle
Park, NC, May 1977b.
Wartburg, A. F., J. B. Pate, and J. P. Lodge, Jr. Environ. Sci. Technol.
3:767, 1969.
Wedding, J. B., A. R. McFarland, and J. E. Cernak. Large particle collection
characteristics of ambient aerosol samplers. Environ. Sci. Technol.
11:387-390, 1977.
Wedding, J. B., M. Weigand, W. John, and S. Wall. Sampling Effectiveness of
the Inlet to the Dichotomous Sampler. Aerosol Science Laboratory, Fort
Collins, CO, 1980.
Welch, A. F., and J. P. Terry. Developments in the measurement of atmospheric
sulfur dioxide. Am. Ind. Hyg. Assoc. J. 21:316-321, 1960.
2-166
-------
West, P. W., and G. C. Gaeke. Anal. Chem. 28:1816, 1956.
Whitaker Company Electrochemical Transducer.
Whitby, K. T. Modeling of Atmospheric Aerosol Size Distribution, Progress
Report, EPA Grant No. R800971, 1975.
Whitby, K. T. The Physical Characteristics of Sulfur Aerosols. Paper Prepared
for Presentation at the International Symposium on Sulfur in the Atmosphere,
7-14 Sept., 1977, Dubrovnik, Yugoslavia.
Whitby, K. T., and K. Willeke. Single particle optical counters: principles
and field use. Jn: Aerosol Measurement. University of Florida Presses,
Gainesville, FL, 1979. p. 145-180.
Whitby, K. T., and W. E. Clark. Electrical aerosol particle counting and size
distribution measuring system for the 0.015 to 1.0 mm size range. Tellus.
18:573-586, 1966.
Willeke, K., and J. J. McFeters. Calibration of the CHAMP Fractionator,
Particle Technology Publication No. 252, University of Minnesota,
Minneapolis, MN, March 1975.
Willeke, K., and Whitby, K. T. Atmospheric aerosols size distribution
interpretation. J. Air Pollut. Control Assoc. 25:529-534, 1975.
Wilsdon, B. H., and F. J. McConnell. J. Soc. Chem. Ind. 53:385, 1934.
Wilson, W., et al. Sulphates in the Atmosphere. Paper No. 76-30.06, 69th
Annual Meeting of the Air Pollution Control Assoc., Portland, Ore.,
U.S.A., June 27, 1976.
World Health Organization Selected Methods of Measuring Air Pollutants. WHO
Offset Publication No. 24, World Health Organization, Geneva, Switzerland,
1976.
Inhaled Particles. Environ. Sci. Technol. 12(13):1353-1355, 1978.
2-167
-------
2.7 REFERENCES
Adams, F. C. , and R. E. Van Grieken. Absorption correction for X-ray
fluorescence analysis of aerosol loaded filters. Anal. Chem. 47:1767-1773,
1975. ~~
Adamski, J. M., and S. P. Villard. Application of the methylthymol blue
sulfate method to water and wastewater analysis. Anal. Chem. 47:1191-1194,
1975.
Ahearn, A. I. Trace Analysis by Mass Spectrometry. Academic Press, New York,
NY, 1972.
Ahlquist, N. C. , and R. J. Charlson. J. Air Pollut. Control Assoc. 17:467-,
1967.
American Society for Testing and Materials. Standard method for collection
and analysis of dustfall. In: Book of ASTM Standards. American Society
for Testing and Materials, Philadelphia, PA, 1966. p. 785-788.
Andrezejewski, R. Les Proprietes physiques des poussiered, Ed. Slask, Katowice,
1968.
Appel, B. R. , E. L. Kothny, E. M. Hoffer, and J. J. Wesolowski. Comparison of
Wet Chemical and Instrumental Methods for Measuring Airborne Sulfate.
EPA-600/2-76-059, U.S. Environmental Protection Agency, Research Triangle
Park, NC, March 1976.
Archer, A. W. The indirect colorimetric determination of sulphate with
2-aminoperimidine. Analyst (London) 100:755-757, 1975.
Barringer Research, Ltd. , A report to Department of Health, Education and
Welfare of optical measurements of SO,, and NO^ air pollution using Barringer
Correlation Spectrometers. Barringer Researcn Technical Report TR-69-113.
Canada, 1969.
Barton, S. C. , and H. G. McAdie. A specific method for the automatic determina-
tion of ambient HLSO. aerosol. In: Proceedings of the 2nd International
Clean Air Congress. H. M. Englund and W. T. Berry, eds. , Academic Press,
New York, NY, 1971. p. 379-382.
Barton, S. C. , and H. G. McAdie. An automated instrument for monitoring
ambient H?SO, aerosol. In: Proceedings of the 3rd International Clean
Air Congress: VDI-Verlag 6 mbH, Dusseldorf, West Germany, 1973. p. c25.
Benard, W. F. Community Health Environmental Surveillance Studies (CHESS) Air
Pollution Monitoring Handbook. EPA-600/1-76-011, U.S. Environmental
Protection Agency, Research Triangle Park, NC, January 1976.
Benarie, M., B. T. Chuong, and A. Nonat. Essai d'une Methode pour le Dosage
Selectif de 1'Acide Sulfurique dans 1'Atmosphere de Stransbourg et Defini-
tion du Role des Particules Sol ides dans 1'Oxydation du SO,, dans VAtomsphere
Report of IRCHA Study 7106, Centre de Recherche, Vert-le-Petit, France,
January 16, 1976.
2-153
-------
Benarie, M., T. Menard, and A. Nonat. Atmos. Environ. 7:403, 1973.
Bertolacini, R. J., and J. E. Barney. Anal Chem. 29:281, 1957.
Blanco, A. J., and G. B. Hoidale. Atmos. Environ. 2:327, 1968.
Blanco, A. J., and R. G. Mclntyre. Atmos. Environ. 6:557, 1972.
Bogen, D. C., and G. A. Welford. Radiometric determination of sulfate. J.
Radioanal. Chem., 1977.
Bokhaven, C., and H. J. Niessin. Int. J. Air Water Pollut. 10:223, 1966.
Bolleter, W. T., C. J. Bushman, and P. W. Tidwell. Anal. Chem. 33:592, 1961.
Bolz, R. E., and G. E. Tuve, eds. Handbook of Tables for Applied Engineering
Science. Chemical Rubber Co., Cleveland, OH, 1970. p. 159.
Bostrom, C. E. Air Water Pollut. 10:435, 1966.
Bowden, S. R. Int. J. Air Water Pollut. 8:101, 1964.
British Standard 1747, Part 3, 1963.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 2. Determination of Concentration of Suspended Matter. B.S. 1747,
Part 2, British Standards Institution, London, England, 1964.
Brock, J. R., and Marlow. Charged aerosol particles and air pollution.
Environ. Letters, No. 10, 1975.
Brosset, C., and M. Perm. Man-made airborne acidity and its determination.
Atmos. Environ. 12:909-916, 1978.
Buttell, R. G. , and Brewer J. Sci. Inst. 26:357, 1949.
Cahill, T. A. Comments on surface coatings for Lundgren - Type impactors.
jjy. Aerosol Measurement. University Presses of Florida, Gainesville,
FL, 1979. p. 131.
Calvert, S., and W. Workman. Talanta4:89, 1960.
Camp, D. C., A. L. Van Lehn, and B. W. Loo. Intercomparison of Samplers Used
in the Determination of Aerosol Composition. EPA-600/7-78-118, U.S.
Environmental Protection Agency, Research Triangle Park, NC, 1978.
Charlson, R. J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. Atmos. Environ. 8:1257, 1974a.
2-154
-------
Charlson, R. J. , N. C. Ahlquist, and N. Horvath. On the Generality of Correlation
of Atmospheric Aerosol Mass Concentration and Light Scatter. Atmos.
Environ. 2:455-464, 1968.
Charlson, R.J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. Science 184:156, 1974b.
Cheney, J. L. , and J. B. Homolya. Sampling parameters for sulfate measurement
and characterization. Environ. Sci. Techno!. 13:584-588, 1979.
Conner, W. D. An Inertial-Type Particle Separator for Collecting Large Samples.
J. Air Pollut. Control Assoc.. 16:35-38, 1966.
Coutant, R. W. Effects of environmental variables on collection of atmospheric
sulfate. Environ. Sci. Technol. 11(9):873-878, September 1977.
Crider, W. L. Anal. Chem. 37:1770, 1965.
Crider, W. L. , N. P. Barkley, M. J. Knott, and R. W. Slater. Hydrogen flame
chemilurimescence detector for sulfate in aqeous solutions. Anal. Chim.
Acta 47:237-241, 1969.
Cunningham, P. T. , and S. A. Johnson. Science 191:77-79, 1976.
Cashing, K. M. , G. E. Lacey, J. D. McCain, and W. B. Smith. Particulate
Sizing Techniques for Control Device Evaluation: Cascade Impactor Calibration
EPA-600/2-76-280, U.S. Environmental Protection Agency, Research Triangle
Park, NC, February 1978.
Davies, C. N. , ed. Aerosol Science. Academic Press, New York, NY, 1966.
Dennis, E. R. , ed. Handbook on Aerosols. National Technical Information
Service, TID-26608, Arlington, VA, 1976.
deVeer, S. M. , H. J. Bronwer, and H. Zeedijk. Presented at the 62nd Annual
Meeting, Air Pollution Control Association, New York, New York, June
1969. Paper No. 69-6.
Dimitt, R. L. , and E. R. Graham. Anal. Chem. 48:604, 1976.
Dionex Corporation. Dionex Analytical Ion Chromatographs. Technical Brochure,
Dionex Corp., Palo Alto, California, 1975.
Dubois, L. , C. J. Baker, T. Teichman, A. Zdrojewski, and J. L. Monkman. Mikrochim.
Acta 269:1268, 1969.
Dzubay. T. G. , H. L. Rook, and R. K. Stevens. A chemiluminescent approach to
measurement of strong acid aerosols. In: Analytical Methods Applied to
Air Pollution Measurement. R. Stevens and W. Herget, eds. , Ann Arbor
Science Publishers, Inc., Ann Arbor, MI, 1974. p. 71.
2-155
-------
Dzubay, T. G., G. K. Snyder, D. J. Reutter, and R. K. Stevens. Aerosol acidity
determination by reaction with C labeled amine. Atmos. Environ.,
13:1209-1212, 1979.
Dzubay, T. G., L. E. Mines, and R. K. Stevens. Particle bounce errors in
cascade impactors. Atmos. Environ. 10:229-234, 1976.
Eagan, M. L., and L. Dubois. Anal. Chim. Acta 70:157, 1974.
Edited by Lundgren, D. A., M. Lippman, F. S. Harris, W. E. Clark, W. H. Marlow,
and M. D. Durham. Aerosol Measurement. University Press of Florida,
Gainesville, FL, 1979.
Environmental Monitoring Systems Laboratory, Summary of Performance Test
Results and Comparative Data for Designated Equivalent Methods for SO..
Document No. QAD/M-79.12. U.S. Environmental Protection Agency, Research
Triangle Park, NC, December 1979.
Falkowski, J. M. Suszka i uwlaznisnije lubovoloknistych matierijalow. Moscow,
GNTI, 1951.
Fassel, V. A. Quantitative elemental analyses by plasma emission spectroscopy.
Science. 202:183-191, 1978.
Federal Register, 40(33):7044, 1975.
Friedlander, S. K. Smoke, Dust, and Haze. Wiley Interscience Publication,
New York, NY, 1977. p. 154.
Fritz, S. J., and S. S. Yamamura. Anal. Chem. 27:1461, 1955.
Fuerst, R. G., F. P. Scaringelli, and J. H. Margeson. Effect of Temperature
on Stability of Sulfur Dioxide Samples Collected by the Federal Reference
Method. U.S. Environmental Protection Agency, Research Triangle Park,
NC, EPA-600/4-76-024, May 1976.
Gales, M. E., Jr., W. H. Kaylor, and J. E. Longbottom. Analyst (London)
93:97, 1968.
Gilbert, T. R., and A. M. Clay. Anal. Chem. 45:1757, 1973.
Goold, R. W. , C. S. Barrett, J. B. Newkirk, and C. 0. Ruud. Advances in X-Ray
Analysis. Kendall/Hunt, Dubuque, IA, 1976.
Goulding, F. S., and J. M. Jaklevic. X-ray fluorescence spectrometer for
airborne particulate monitoring. EPA Report No. EPA-R2-73-182, U.S.
Environmental Protection Agency, Washington, DC, April, 1973.
Goulding, F. S., J. M. Jaklevic, and B. W. Loo. Aerosol Analysis for the
Regional Air Pollution Study. EPA-600/4-78-034, U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1978.
2-156
-------
Hager, R. N., Jr., and R. C. Anderson. J. Opt. Soc. Am. 60:1444, 1970.
Hansen, L. D., L. Whiting, D. J. Eatough, T. E. Jensen, and R. M. Izatt. Anal.
Chem. 48:634, 1976.
Harris, B. Second Stage Development of an Automated Field Sulfuric Acid
Sampler and the Development of a Sulfuric Acid Analyzer. Monthly Technical
Progress Narrative No. 6, Cabot Corporation, Billerica, MO, December
1975.
Heard, M. J. , and R. D. Wiffen. Atmos. Environ. 3:337, 1969.
Hemeon, W. C. L. , G. F. Haines, and H. M. Ide. Determination of haze and
smoke concentrations by filter paper samples. J. Air Pollut. Control
Assoc. 3:22-28, 1953.
Hesketh, H. E. Understanding and Controlling Air Pollution, Ann Arbor Science
Publishers, Michigan, 1974.
Hill, A. S. G. Measurement of the optical densities of smokestains on filter
papers. Trans. Farad. Soc. 32:1125-1131, 1936.
Hochheiser, S. Methods of Measuring Atmospheric Sulfur Dioxide. AP-6, U.S.
Department of Health, Education, and Welfare, Cincinnati, OH, August
1964.
Hoffer, E. , and E. L. Kothny. A Micromethod for Sulfate in Atmospheric
Particulate Matter. Report No. 163, Air and Industrial Hygiene Labora-
tory, Berkeley, CA, July 1974.
Hounam, R. E. , and R. J. Sherwood. The cascade centripeter: a device for
determining the concentration and size distribution of aerosols. J. Am.
Ind. Hyg. Assoc. 26:122-131, 1965.
Huey, N. A. The lead dioxide estimation of sulfur oxide pollution. J. Air
Pollut. Control Assoc. 18:610, 1968.
Huey, N. A., M. A. Wallar, and C. D. Robson. Field evaluation of an improved
sulfation measurement system. Paper No. 69-133 Presented at Air Pollution
Control Association Annual Meeting, June, 1969. Paper No. 69-133.
Huntzicker, J. J. , L. J. Isabelle, and J. G. Watson. The continuous monitoring
of particulate sulfate by flame photometry. In: International Conference
on Environmental Sensing and Assessment. Volume 2, Institute of Electrical
and Electronics Engineers, Inc., New York, NY, 1976. Paper 23-4.
Huntzicker, J. J. , L. M. Isabelle, and J. G. Watson. The continuous monitoring
of particulate sulfur compounds by flame photometry. Presented at the
173rd National Meeting, American Chemical Society, New Orleans, Louisiana,
March 20-25, 1977. Paper ENVR-39.
2-157
-------
Husar, J. D., R. B. Husar, and P. K. Stubits. Anal. Chem. 47:2062, 1975.
Husar, R. B., J. D. Husar, N. V. Gillani, S. B. Fuller, W. H. White, J. A.
Anderson, W. M. Vaughn, and W. E. Wilson, Jr. Pollutant flowrate measurement
in large plumes: sulfur budget in power plant and area source plumes in
the St. Louis region. Presented at the 171st Meeting. American Chemical
Society, New York, New York, April 4-9, 1976. Paper ENVR-14.
Huygen, C. Atmos. Environ. 9:315, 1975.
Inhaled particles. Environ. Sci. Technol. 12(13):1353-1355, 1978.
Intersociety Committee. Methods of Air Sampling and Analysis. American
Public Health Association, Washington, DC, 1972.
Intertech Corp. Descriptive brochure.
Jacobs, M. B. The Chemical Analysis of Air Pollutants. Wiley-Interscience,
New York, NY, 1960.
Johansson, T. B., R. E. Van Grieken, J. W. Nelson, and J. W. Winchester. Anal.
Chem. 47:855, 1975.
John, W., G. Reischl, and J. Wesolowski. Size Selective Monitoring Techniques
for Particulate Matter. Report No. ARB A5-00487 California Department of
Health, Air and Industrial Hygiene Laboratory, Berkeley, CA, February
1978.
Katz, M. Anal. Chem. 22:1040, 1950.
Katz, M. In: Air Pollution. Vol. II, 2nd ed. A. C. Stern, ed., Academic
Press, New York, NY, 1968.
Katz, M. Measurement of Air Pollutants—Guide to Selection of Methods. World
Health Organization, Geneva, 1969.
Keay, J., and P. M. A. Menage. Analyst (London) 95:379, 1970.
Kirkland, J. J. Modern Practice of Liquid Chromatography. John Wiley, New
York, NY, 1971.
Kittleson, D. B., R. McKenzie, M Vermeersch, M. Linne, F. Dorman, D. Pui, B.
Liu, and K. Whitby. Total sulfur aerosol detection with an electrostatic
pulsed flame photometric detector system. Presented at the 173rd National
Meeting, American Chemical Society, New Orleans, Louisiana, March 20-25,
1977. Paper ENVR-42.
Klockow, D., and G. Ronicke. Atmos. Environ. 7:163, 1973.
Kolthoff, I. M. , E. B. Sandell, E. J. Meehan, and S. Bruckenstein. Quantitative
Chemical Analysis, 4th ed., MacMillan, Toronto, Canada, 1969. pp. 992-994.
2-158
-------
Lambert, J. P. F. , and F. W. Wilshire. Neutron activation analysis for
simultaneous determination of trace elements in ambient air collected on
glass-fiber filters. Anal. Chem. 51:1346-1353, 1979.
Lamothe, P. J. , and R. K. Stevens. Sulfuric acid analysis using low temperature
volatilization. Presented at the 171st National Meeting, American Chemical
Society, New York, New York, April 4-9, 1976. Paper ENVR-38.
Lawrence Berkeley Laboratory. Instrumentation for Environmental Monitoring,
Air. LBL-1, Vol. 1, University of California, Berkeley, CA, May 1972.
Lawrence Berkeley Laboratory. Instrumentation for Environmental Monitoring,
AIR Part 2, University of California, CA, 1975.
Lazarus, A. L. , K. C. Hill, and J. P. Lodge. In: Proceedings of the 1966
Technicon Symposium on Automation in Analytical Chemistry. Mediad, Inc.,
New York, NY, 1966. pp. 291.
Leahy, D. , R. Siege!, P. Klotz, and L. Newman. Atmos. Environ. 9:219, 1975.
Lee, R. E., and S. Goranson. National air surveillance cascade impactor
network. I. Size distribution measurements of suspended particulate
matter in air, Environ. Sci. Technol., 6:1019-1024, 1972.
Lilienfield, P. J. Am. Ind. Hyg. Assoc. 31:722, 1970.
Lippman, M. Respirable dust sampling. J. Am. Ind. Hyg. Assoc. 1970. p. 146.
Liu, B. Y. H. , and D. Y. H. Pui. On the Performance of the Electrical Aerosol
Analyzer. J. Aerosol Sci. 6:249-264, 1975.
Liu, B. Y. H. , ed. Fine Particles, Aerosol Generation, Measurement, Sampling,
and Analysis. Academic Press, New York, NY, 1976.
Liu, B. Y. H, D. Y. H. Pui, K. L. Rubow, and G. A. Kuhlmey. Progress Report -
Research on Air Sampling Filter Media. Grant Report R804600, University
of Minnesota, Minneapolis, MN, May 1978.
Lizarraga-Rocha, J. A. The Effect of Exposure of Filters During Non-operating
Periods in the High Volume Sampling Method. M.S. Thesis, University of
North Carolina, Chapel Hill, NC, 1976.
Lochmuller, C. H. , J. W Galbraith, and R. L. Walter. Anal. Chem. 46:440,
1974.
Lodge, J. P., J. Ferguson, and B. R. Haylik. Anal. Chem. 32:1206, 1960.
Loo, B. W., R. S. Adachi, C. P. Cork, F. S. Goulding, J. M. Jaklevic, D. A.
Landis, and W. L. Searles. A second generation dichotomous sampler for
large scale monitoring of airborne particulate matter. Presented at 86th
Annual Meeting, American Institute of Chemical Engineers, Houston, Texas,
April 1, 1979.
2-159
-------
Loo, B. W., W. R. French, R. C. Gatti, F. S. Goulding, J. M. Jaklevic, J.
Uacer, and A. C. Thompson. Large Scale Measurement of Airborne Participate
Sulfur. Atmos. Environ. 12:759-771, 1978.
Lucero, D. P., and J. W. Paljug. Monitoring sulfur compounds by flame photometry.
ASTM Committee D-22 Conference on Instrumental Monitoring for Ambient
Air, Boulder, Colorado, August 14-16, 1973.
Lundgren, D. A. Mass Distribution of Larger Atmospheric Particles. Ph.D.
Thesis, University of Minnesota, Minneapolis, MN, 1973.
Macias, E. S. Instrumental analysis of light element composition of atmospheric
aerosols. In: Methods and Standards for Environmental Measurement. W.
Kirchhoff, ed., National Bureau of Standards Special Publication 464.
Department of Commerce, Washington, DC, 1977. pp. 179-188.
Macias, E. S., and R. B. Husar. Atmospheric particulate mass measurement with
beta attenuation mass monitor. Environ. Sci. Techno!. 10:904-907, 1976.
Malm, W. Consideration in the measurement of visibility. J. Air Pollut.
Control Assoc. 29:1042-1052, 1979.
Mamane, Y. , and R. G. de Pena. A quantitative method for the detection of
individual submicron size sulfate particles. In: Sulfur in the Atmosphere,
Dubrovnik, Proceeding of an International Symposium, Yugoslavia, September
7-14, 1977. Atmos. Environ. 12:69-82, 1978.
Mamantov, G., E. L. Wehry, R. R. Kemerer, and E. R. Hinton. Anal. Chem.
49:86, 1977.
Martin, B. E. Sulfur Dioxide Bubbler Temperature Study. EPA-600/4-77-040,
U.S. Environmental Protection Agency, Research Triangle Park, NC, August
1977.
McCoy, R. A., D. E. Camann, and H. C. McKee. Collaborative Study of Reference
Method for Determination of Sulfur Dioxide in the Atmosphere (Pararosaniline
Method) (24-Hour Sampling). EPA-650/4-74-027, U.S. Environmental Protection
Agency, Research Triangle Park, NC, December 1973.
McFadden, W. H. Techniques of Combined Gas Chromatography/ Mass Spectrometry.
John Wiley, New York, NY, 1973.
McFarland, A. R. Test Report - Wind Tunnel Evaluation of British Smoke Shade
Sampler. 3565/05/79/ARM, Texas A&M University, College Station, TX, May,
1979.
McFarland, A. R. and C. E. Rodes. Characteristics of Aerosol Samplers Used in
Ambient Air Monitoring. Presented at 86th National Meeting, American
Institute of Chemical Engineers, Houston, TX, April 2, 1979.
2-160
-------
McFarland, A. R., and C. A. Ortiz. Progress Report - Aerosol Characterization
of Ambient Participate Samplers Used in Environmental Monitoring Studies.
Texas A&M University Research Foundation, College Station, TX, October
1979.
McFarland, A. R. , C. A. Ortiz, and C. E. Rodes. Evaluation of the NASN Cascade
Impactor Ambient Air Monitor. Report 3565/07/78/ARM, Texas A&M, Air
Quality Laboratory, College Station, TX, 1978. Accepted Publication Atm.
Environ., 1980.
McKee, H. C. , R. E. Childers, and 0. Saenz. Collaborative Study of Reference
Method for the Determination of Suspended Particulates in the Atmosphere
(High Volume Method). Southwest Research Institute, Houston, TX, June
1971.
Miller, F. J. , D. E. Gardner, J. A. Graham, R. E. Lee, W. E. Wilson, and J. D.
Bachmann. Size considerations for establishing a standard for inhalable
particles. J. Air Pollut. Control Assoc. 29:610, 1979.
Moffat, A. J. , J. R. Robbins, and A. R. Barringer. Atmos. Environ. 5:511,
1971.
Morrison, G. H. Trace Analysis Physical Methods. Wiley Interscience, New
York, NY, 1965. pp. 245-263, 271-310, 325-360, 377-425.
Mudgett, P. S., L. W. Richards, and J. R. Roehrig. A new technique to measure
sulfuric acid in the atmosphere. In: Analytical Methods Applied to Air
Pollution Measurement. R. Stevens and W. Herget, eds., Ann Arbor Science
Publishers, Inc., Ann Arbor, MI, 1974.
Mulik, J. D., G. Todd, E. Estes, R. Puckett, E. Sawicki, and D. Williams. In:
Ion Chromatographic Analysis of Environmental Pollutants. E. Sawicki, J.
D. Mulik, and E. Wittgenstein, eds., Ann Arbor Science Publishers, Inc.,
Ann Arbor, MI, 1978.
Nadel, J. A. Dust retention efficiencies of dustfall collectors. J. Air
Pollut. Control Assoc. 8:35-38, 1958.
National Academy of Sciences. Airborne Particles. University Park Press,
Baltimore, MD, 1979.
National Air Pollution Control Administration Publication. Air Quality Criteria
for Particulate Matter. AP-49, U.S. Department of Health, Education, and
Welfare, Washington, DC, January 1969.
Natusch, D. F. S. , and C. F. Bauer. Collection and analysis of trace elements
in the atmosphere. Jji: Air Pollution Control, Part III. W. Strauss,
ed., John Wiley and Sons, Inc., 1978.
Newman, L. , J. Forrest, and B. Manowitz. Atmos. Environ. 9:959, 1975a.
2-161
-------
Newman, L., J. Forrest, and B. Manowitz. Atmos. Environ. 9:969, 1975b.
Novakov, T. Chemical characterization of atmospheric pollution particles by
photoelectron spectroscopy. In: Proceedings of the 2nd Joint Conference
on Sensing of Environmental Pollutants, ISA, Washington, DC, December
1973. p. 197.
Novakov, T. , S. G. Chang, and A. B. Marker. Science 186:259, 1974.
Okabe, H., P. L. Splitstone, and J. J. Ball. J. Air Pollut. Control Assoc.
23:514, 1973.
O'Keeffe, A. E. , and G. C. Ortman. Anal. Chem. 38:760, 1966.
Organization for European Economic Cooperation. Paper EPA/AR/4283, European
Productivity Agency, Paris, France, 1961.
Palmer, H. F., C. E. Rodes, and C. J. Nelson. Performance characteristics of
instrumental methods for monitoring sulfur dioxide. II. Field Evaluation.
J. Air Pollut. Control Assoc. 19:778, 1969.
Perera, F. P., and A. K. Ahmed. Respirable Particles: Impact of Airborne
Fine Particulates on Health and the Environment. National Resources
Defense Council, New York, NY, 1978.
Philips Electronic Instruments. Coulometric amperometric analyzers. Bull.
WA-4-C13.
Pierson, W. R., W. W. Prachaczek, T. J. Korniski, T. J. Truex, and J. W.
Butler. Artifact formation of sulfate, nitrate, and hydrogen ion on
backup filters: Allegheny Mountain experiment. J. Air Pollut. Control
Assoc. 30:30-34, 1980.
R. F. Maddalone, G. L. McClure, and P. W. West. Anal. Chem. 47:316, 1975.
Ranade, M. B., and D. VanOsdell. Quality Assurance Evaluation of the CHAMP
Aerosol Inlet, Project Report 43U-1487-59, Research Triangle Institute,
Research Triangle Park, NC, July 1978.
Regan, G. F. , S. K. Goranson, and L. L. Lawson. Use of tape samplers as fine
particulate monitors. J. Air Pollut. Control Assoc. 29:1158-1160, November
1979. ~~
Rider, P. E., C. C. Rider, and R. N. Corning. J. Air Pollution Control Assoc.
27:1011, 1977.
Roberts, P. T. , and S. K. Friedlander. Conversion of S0? to ambient particulate
sulfates in the Los Angeles atmosphere. Atmos. Environ. 10:403, 1976.
Rodes, C. E., and G. F. Evans. Summary of LACS Integrated Measurements.
EPA-600/4-77-034, U.S. Environmental Protection Agency, Research Triangle
Park, NC, June 1977.
2-162
-------
Ross, J. W., Jr., and M. S. Frant. Potentiometric titrations of sulfate using
an ion-selective lead electrode. Anal. Chem. 41:967-969, 1969.
Sajo, I., and B. Sipas. Talanta 14:203, 1967.
Scaringelli, F. P., and K. A. Rehme. Anal. Chem. 41:707, 1969.
Scaringelli, F. P., B. E. Slatzman, and S. A. Frey. Anal. Chem. 39:1709,
1967.
Scroggins, L. H. Collaborative study of the microanalytical oxygen flask
sulfur determination with dimethylsulfonazo III as indicator. J. Assoc.
Off. Anal. Chem. of the AOAC 57:22-25, 1974.
Shaw, R. Personal Communication to C. E. Rodes, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1980.
Sides, J. 0., and H. F. Saiger. Effects of Prolonged Static Exposure of
Filters to Ambient Air on High Volume Sampling Results. State of Kansas
Report, Department of Health and Environment, Topeka, KS, May, 1979.
Sinclair, D., R. J. Countess, B. Y. H. Liu, and D. Y. H. Pui. Automatic
analysis of submicron particles. .In: Aerosol Measurement. University of
Florida Presses, Gainsville, FL, 1979. p. 544-563.
Small, H., T. S. Stevens, and W. C. Bauman. Anal. Chem. 47:1801, 1975.
Smith, A. F. , D. G. Jenkins, and D. E. Cunningworth. J. Appl. Chem. 11:317,
1961.
Smith, F., P. S. Wohlschlegel, R. S. C. Rogers, and Mulligan, D. J. Investigation
of Flow Rate Calibration Procedures Associated with the High Volume
Method for Determination of Suspended Particulates. EPA-600/4-78-047,
U.S. Environmental Protection Agency, Research Triangle Park, NC, June
1978.
Snell, F. D. Photometric and Fluorometric Methods of Analysis, Metals. Part
1. John Wiley, New York, NY, 1978. p. 1-25.
Spicer, C. W. , and P. M. Schumacher. Particle Nitrate: Laboratory and Field
Studies of Major Sampling Interferences. Atmos. Environ. 13:543-552,
1979.
Stalker, W. W., R. C. Dickerson, and G. D. Kramer. J. Am. Ind. Hyg. Assoc.
24:68, 1963.
2-163
-------
Stevens, R. K. , and T. G. Dzubay. Recent Development in Air Participate
Monitoring. IEEE Trans. Nucl. Sci. 22: April 1975.
Stevens, R. K. , J. A. Hodgeson, L. F. Ballard, and C. E. Decker. In: Determination
of Air Quality. G. Mamantov and W. D. Shults, eds., Plenum Publishing
Corporation, New York, NY, 1972.
Stevens, R. K., J. D. Mulik, A. E. O'Keeffe, and K. J. Krost. Anal. Chem.
43:827, 1971.
Stevens, R. K., T. G. Dzubay, G. Russwurm, and D. Rickel. Sampling and analysis
of atmospheric sulfates and related species. Atmos. Environ. 12:55-68,
1978.
Stratman, H. Mikrochim Acta 6:668- , 1954.
Swanson, D., C. Morris, R. Hedgecoke, R. Jungers, R. J. Thompson, and J. E.
Bumgarner. A rapid analytical procedure for the analysis of BaP in
environmental samples. Trends Fluoresc. 1:22-27, 1978.
Tanner, R. L. , and Cordova. U.S. Energy Research Development Administration.
Contract No. EY-76-C-02-0016, 1978.
Tanner, R. L., J. Forrest, and L. Neurnan. Determination of atmospheric
gaseous and particulate sulfur compounds. U.S. Energy Research Development
Administration. Contract No. EY-76-C-02-0016, 1978.
Tanner, R. L., R. Cederwall, R. Garber, D. Leahy, W. Marlow, R. Meyers, M.
Phillips, and L. Newman. Atmos. Environ. 11, 1977.
Tanner, R. L., R. W. Garber, and L. Newman. Speciation of sulfate in ambient
aerosols by solvent extraction with flame photometric detection. Presented
at the 173rd National Meeting, American Chemical Society, New Orleans,
Louisiana, March 20-25, 1977a. Paper ENVR-41.
Tanner, R. L., R. W. Garber, W. H. Marlow, B. Leaderer, D. Eatough, and M. A.
Leyko. Chemical composition and size distribution of sulfate as a function
of particle size in New York City aerosol. Presented at the Industrial
Hygiene Conference, New Orleans, Louisiana, May 22-27, 1977. Paper No.
140.
Technicon Corporation. Sulfate method VIb via turbidimetry. Technicon
Corporation, Tarrytown, NY, 1959.
Terraglio, F. P., and R. M. Maganelli. Anal. Chem. 34:675, 1962.
Tetlow, J. A., and A. L. Wilson. Analyst (London) 89:453, 1964.
Thoen, G. N. , G. G. Dehaas, and R. R. Austin. Tappi 51:246, 1968.
Thomas, R. L. , V. Dharmarajan, G. L. Lundquist, and P. W. West. Measurement
of sulfuric acid aerosol, s^fur trioxide, and the total sulfate content
of the ambient air. Anal. Chem. 48:639-642, 1976.
2-164
-------
Thrane, K. E. In: Studies in Environmental Science, Vol. 2, Air Pollution
Reference Measurement Methods and Systems. T. Schneider, H. W. de Koning,
and L. J. Brasser, eds. , Elsevier Scientific Publishing Company, Amsterdam,
1978.
Threshold Limits Committee. Threshold Limit Values of Air Borne Contaminants
for 1968. American Conference of Governmental Industrial Hygienists,
Cincinnati, OH, 1968.
Tierney, G. P-, and W. D. Conner. Hydroscopic effects of weight determinations
of particulates collected on glass fiber filters. J. Am. Ind. Hyg.
Assoc., 1967.
Treon, J. F. , and W. E. Crutchfield. Ind. Eng. Chem. Anal. Ed. 14:119, 1942.
Turner, J. M., and R. S. Shotes. Presented at the 64th Annual Meeting, Air
Pollution Control Association, 1971. Paper No 71-39.
U.S. Environmental Protection Agency. Ambient Air Monitoring Reference and
Equivalent Methods. 40 CFR 53:952-986, July 1, 1979c.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. Appendix A - Reference method for the determina-
tion of sulfur dioxide in the atmosphere (pararosanilin'e method). 40 CFR
50:6-11, July 1, 1979.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. Appendix B - Reference method for the determina-
tion of suspended particulates in the atmosphere (high volume method).
40 CFR 50:12-16, July 1, 1979a.
U.S. Environmental Protection Agency. National Primary and Secondary Ambient
Air Quality Standards. 40 CFR 50:3-5, July 1, 1979b.
Volmer, W., and F. Z. Frohlich. Z. Anal. Chem. 126:401, 1944.
Von Lehmden, D. J., and C. Nelson. Quality Assurance Handbook for Air Pollution
Measurement Systems. Volume I. Principles. EPA-600/9-76-005, U.S.
Environmental Protection Agency, Research Triangle Park, NC, May 1977.
Von Lehmden, D. J., and C. Nelson. Quality Assurance Handbook for Air
Pollution Measurement System. Volume II. Ambient Air Specific Methods.
EPA~600/4-77-027a, U.S. Environmental Protection Agency, Research Triangle
Park, NC, May 1977b.
Wartburg, A. F. , J. B. Pate, and J. P. Lodge, Jr. Environ. Sci. Technol.
3:767, 1969.
Wedding, J. B. , A. R. McFarland, and J. E. Cernak. Large particle collection
characteristics of ambient aerosol samplers. Environ. Sci. Technol.
11:389, 1977.
2-165
-------
Wedding, J. B., M. Weigand, W. John, and S. Wall. Sampling Effectiveness of
the Inlet to the Dichotomous Sampler. Aerosol Science Laboratory, Fort
Collins, CO, 1980.
Welch, A. F., and J. P. Terry. Developments in the measurement of atmospheric
sulfur dioxide. J. Am. Ind. Hyg. Assoc. 21:316-321, 1960.
West, P. W., and G. C. Gaeke. Anal. Chem. 28:1816, 1956.
Whitaker Company Electrochemical Transducer.
Whitby, K. T. Modeling of Atmospheric Aerosol Size Distribution, Progress
Report, EPA Grant No. R800971, 1975.
Whitby, K. T. The Physical Characteristics of Sulfur Aerosols. Paper Prepared
for Presentation at the International Symposium on Sulfur in the Atmosphere,
7-14 Sept., 1977, Dubrovnik, Yugoslavia.
Whitby. K. T., and K. Willeke. Single particle optical counters: principles
and field use. In: Aerosol Measurement. University of Florida Presses,
Gainesville, FL, 1979. p. 145-180.
Whitby, K. T., and W. E. Clark. Electrical aerosol particle counting and size
distribution measuring system for the 0.015 to 1.0 mm size range. Tellus.
18:573-586, 1966.
Willeke, K., and J. J. McFeters. Calibration of the CHAMP Fractionator,
Particle Technology Publication No. 252, University of Minnesota,
Minneapolis, MN, March 1975.
Willeke, K., and Whitby, K. T. Atmospheric Aerosols Size Distribution
Interpretation, J. Air Pollution Control Assoc., 25, No. 5, 1975.
Wilsdon, B. H. , and F. J. McConnell. J. Soc. Chem. Ind. 53:385, 1934.
Wilson, W., et al. Sulphates in the Atmosphere. Paper No. 76-30.06, 69th
Annual Meeting of the Air Pollution Control Assoc., Portland, Ore.,
U.S.A., June 27, 1976.
World Health Organization Selected Methods of Measuring Air Pollutants. World
Health Organization, Geneva, Switzerland, 1976.
2-166
-------
7/9/80
Chapter 3 Critical Appraisal of Air Quality Measurement Applications Corrigenda
Before listing specific minor errata (insertions/deletions) for text contained
in Chapter 3 of the April, 1980, External Review Draft, several major changes to
be made in the chapter should be noted. The proposed changes are based in part
on additional literature review and comments received since finalization and
release of the April, 1980, external review draft of the chapter.
On pg. 3-84, after the first paragraph, new text is to be inserted discussing
the fact that difficulties in comparing results from various bodies of epidemiologic
literature (e.g., British versus American) on particulate matter health effects
arise from differences in specific physical and chemical properties of particulate
matter pollution indexed by different measurement techniques employed in such
studies. In particular, the following points are to be noted:
(1) The British smoke (BS) sampling technique (widely used in Britain and
Europe) mainly collects fine mode particles (<3-5 pm) and, using reflected
light, specifically measures degree of reduction of reflectance by the collected
particles. BS computed by the degree of reduction in reflectance, although
sometimes highly correlated with TSP (r=0.8-0.9) and lead (r=0.8-0.9) as reported
by Ball and Hume (1977) and affected by certain other materials (Pedace and
Sansone, 1972), is most highly correlated with the amount of graphitic carbon*
present (r=.96; Baily and Clayton, 1980). Levels of carbon or other materials
affecting reflectance readings can, however, vary independently of the total
mass of the collected (mainly fine mode) particles. Thus, estimation of. collected
particle mass indexed by BS readings is dependent upon calibration of BS readings
against standard mass readings (weighing) of collected particle samples typifying
* The term "graphitic carbon" is not meant to imply the three-dimensional
structure of graphite, but only to indicate a structure similar to that of
carbon black contained in soot.
-------
a given location, allowing for calculation of corresponding mass concentration
levels (in ug/m ) by taking into account sampling periods and air flow rates.
Relationships between BS reflectance readings and mass concentrations (ug/m )
for a given location are most accurately determined empirically by reflectance
to mass calibrations derived on a site-by-site and, time-specific basis, given
the fact that relative mixes of va n'ous pollutants sampled could vary on an hour
to hour, day to day, or longer-term basis. However, if the relative mix of
particulate matter sampled and the percentage of the total collected mass attribu-
table to graphitic carbon and other materials affecting light reflectance remains,
similar from time to time or site to site, as empirically demonstrated by repre-
sentative calibration determinations, then a common standard calibration curve
relating mass concentration (ug/m ) of particulates collected to reflectance
readings may be generally applicable for BS data obtained from thusly calibrated
sites shown to fit the standard curve well (Wallin, 1965).
(2) Particulate matter measurements by certain American sampling techniques
(e.g., the AISI tape sampler method) which uses light transmittance as the
physical property measured (Katz and Sanderson, 1958), are also most strongly
affected by graphitic carbon levels present among particulate matter mass collected
(mainly fine mode, <5ym size-range). Thus, similar considerations and limitations
as described above for the BS method apply to the AISI tape sampler light trans-
mittance method in regard to converting transmittance readings (coefficient of
haze units or COHS for the AISI method) to estimates of particulate mass collected.
That is, very precise estimates of particulate mass collected or air concentration
(in ug/m ) require that the transmittance readings (in COHS) be calibrated
against representative corresponding sample weights on a site-and time-specific
basis.
-2-
-------
However, practical application of the AISI light transmittance method, as in the
case of BS measurements, usually precludes other than occasional representative
calibrations for a given sampling site due to personnel and other resource
limitations. Rather, to the extent that similar relationships of mass to trans-
mittance readings are consistently obtained at a given site through repeated
calibrations over time, then use of a standard calibration curve for converting
transmittance readings (COHS units) to corresponding particle mass estimates for
that site appears to be reasonably well justified.
Analogously, to the extent that a relatively similar mix of pollutants
(e.g. graphitic carbon) most strongly affecting light transmittance is represented
among the total mass of particulate matter collected at various other sites, it
appears reasonable to employ a common standard calibration curve for conversion
of transmittance values to estimates of particulate mass collected at those
sites by the AISI method, especially if representative calibration data are fit
well by the standard curve. Conversely, substantial variations in the pollutant
mix, e.g. in the percentage of graphitic carbon represented in the overall
particulate mass collected either for the same site at different times or between
different sites, should be reflected by notable deviations from the more usually
applicable calibration curve and would require generation of another one on a
site-and time-specific basis. The likely generalizability of any particular
calibration curve for estimation of collected particulate mass concentrations
based on light transmittance readings appears, then, to be amenable to empirical
testing in terms of: (a) assessment of goodness of fit of data for specific
sites to the particular model defining the given calibration curve, and (b)
assessment of similarities of chemical composition (including percentage of
graphitic carbon and other substances affecting light transmittance) of atmospheric
particulate matter sampled at different times at the same site or between different
sites.
-3-
-------
(3) Since both light reflectance (BS) and transmittance (AISI tape
sampler) measurement methods are both most strongly affected by levels of
graphitic carbon among the particulate matter mass collected, it might
be assumed that results obtained by each method should be readily translatable
into equivalent measurement units employed by the other method (i.e., reflectance
or darkness index BS readings versus COHS units for the AISI tape sampler), using
emperically-derived calibration curves. Similarly, it might be assumed that, if
particles of the same size-range are collected and valid curves exist for estimation
of particulate matter mass levels based on either light absorption or transmittance
readings for a given site, equivalent particulate matter mass or air concentration
estimates should be derivable from reflectance or transmittance readings for a
given set of BS or COHS data points on a site-and time-specific basis. Those mass
concentration estimates, furthermore, would then presumably provide a reasonable basis foi
comparisons of particulate matter levels in the same size-range from one site or
time to another, where atmospheric aerosols of similar chemical composition
are sampled. This might be the case, for example, in terms of samples of
certain urban aerosols containing similar pollutant mixes derived from relatively
similar emission sources, but not obtained from collection sites markedly dominated
by different single emission sources. The interconversion (comparability), or
lack thereof, of particulate matter mass estimates derived from light reflectance
versus light transmittance measurement techniques, therefore, should be empirically
testable in terms of goodness of fit of data from a given site or time in relation
to curves or equations modeling relationships between results from the two
techniques.
-4-
-------
The particulate matter measurements (in terms of mass concentrations
o
in yg/m ) based on standard light reflectance (BS) or transmittance
(AISI tape sampler) methods most often used in community health epidemiology
studies might also be expected to approximate fine mode particulate
matter mass estimates determined by gravimetric methods, in view of
procedures for the former two methods being reported to result mainly in
collection of fine mode fraction particles (<3 urn). However, factors
influencing estimation of mass from BS reflectance or COH units could be
such to result in quantitative mass estimates different from those
obtained for fine particulate mass as determined by gravimetric methods.
(4) Conversion of particulate matter measurement results from
either light reflectance (e.g., BS) or transmittance (e.g., AISI tape
sampler) methods versus high-volume sampler results (expressed in yg/m
TSP) involve additional considerations beyond those outlined above.
First, the high volume sampler has an inlet which is 50% efficient in
collecting 25-30 ym sized particles and therefore collects coarse mode
particles not sampled by the other two methods as standardly employed.
Also, graphitic carbon levels and other materials most strongly affecting
light reflectance and transmittance readings and fine mode particulate
mass levels (indirectly indexed by the same methods) can vary independently
of coarse mode particle levels. They may, therefore, increase or decrease
in directions opposite to changes in total suspended particulate (TSP)
matter mass. It is thusly not surprising that quite different relationships
between particulate matter mass estimates based on BS or COHS readings
and TSP levels can exist from site to site or from sampling
time to sampling time at the same site.
-5-
-------
Reliable determination or estimation of likely TSP levels present at particular
sampling points (times/places) based on corresponding BS or COHS data, then, may
only be possible or justifiable within very circumscribed limits and highly
dependent upon development of applicable empirically-derived intercomparison
models in a fashion analogous to obtaining calibration or interconversion curves
for BS and COHS readings discussed above.
On pg. 3-102, immediately before the last paragraph and heading for Section
3.5.5, new text is to be added which notes that three recently reviewed papers
(by Ledbetter and Cerepaka, 1980; Swinford and Kolaz, 1980; and Heindryckx,
1975) report on relationships between COM and TSP as determined by comparisons
of results obtained from collocated high volume and AISI tape samplers. In each
case, for data obtained from sampling site locations as diverse as areas in
Texas, Illinois, and Belgium, considerable scatter was found for individual
paired observations, suggesting great uncertainty in predicting 24 hour TSP
levels from short-term (1 hour) COH readings. It may be possible to improve on
the relationships by using seasonal calibrations at each site and nonlinear
models calibrated in a manner similar to the ASTM method or the British Standard
1747 Part 2 procedure for smokeshade in order to convert COH readings to units
of mass for comparison with TSP data. However, these studies, and other literature
cited above in this chapter, all appear to indicate that COH measurements are
generally not directly relatable to TSP levels.
On pg. 3-103, immediately before the last paragraph, new text is to be
inserted which notes that a typical finding among the various BS/TSP comparison
studies cited in the preceeding paragraph is that considerable scatter or variability
-6-
-------
exists for individual paired observations for 24 hour BS and corresponding TSP
measurements analyzed in the various studies. However, greater consistency
appeared to exist for BS-TSP relationships expressed as long-term (monthly or
longer) averages of the paired 24 hour observations; and efforts were made in
the different studies to derive equations or models (mainly linear regression)
that best fit the observed data obtained from different individual sites, cities
and time periods (e.g., winter heating seasons versus summer nonheating seasons).
On pg. 3-116, immediately before the last paragraph, new text is to be
inserted noting that neither the Mage (1980)* bounded nonlinear model derived
primarily from the annual average (mean) BS-TSP comparison data of Commins and
Waller (1967) as summarized by Holland et al. (1979) nor any other presently
available model provides a generally reliable basis for interconversion of
corresponding individual 24 hour BS-TSP data points, except perhaps at high
levels of BS (>_ 500 yg/m ). The Mage (1980) and other models such as those
discussed in Lee et al. (1972) or Pashel and Egner (1980), however, may provide
better ore more reliable fits for BS-TSP monthly or annual average data.
On pg. 3-121, there is to be deleted the last three paragraphs discussing
the impact of the use of a sampling flow rate of 0.72 liters per minute by
Pashel and Egner (1980). Also to be deleted are Figure 3-13 on pg. 3-122, all
text but the last paragraph on pg. 3-123, and Appendix C of Chapter 3. These
deletions are based on a personal communication from Pashel and Egner indicating
that a 0.72 liter per minute flow rate was erroneously reported in their draft
1980 Atmospheric Environment article; rather, according to their personal communi-
cation, a 1.5 1/min. flow rate more typically employed in generating BS data
used by British epidemiologists was also used by them in producing the data
reported in their 1980 paper.
* Manuscript now in preparation.
-7-
-------
Detailed discussion of possible methodological errors in the Pashel and
Egner study, starting on pg. 3-118, are to be moved to appendices and only brief
summary statements regarding such included in the main text of Chapter 3. Also,
on pg. 3-128, the paragraph immediately before the heading for Section 3.5.5.3.
is to be entirely deleted and replaced with the following new text: Another
possible explanation for the particular pattern of results obtained by Pashel
and Egner is suggested by the fact that all of their annual mean data for rural,
residential, and commercial sites studied fall on or very near the BNLM (Mage,
1980) model curve in Figure 3-15 on pg. 3-127, whereas data for seven of eleven
industrial sites fall rather far to the right of the BNLM curve. The possibilty
exists that high levels of either noncarbonaceous fine-mode particles, coarse
mode particles, or both in fugitive dust or stack emissions from nearby industrial
sources in the absence of much carbonaceous material from fossil fuel combustion,
o
result in the relatively higher TSP readings (mostly >100 yg/m ) obtained at
those sites. If such were the case, the data would illustrate a likely general
limiting factor in making meaningful comparisons or interconversions between BS
and TSP data, even on a long term annual mean basis. That is, whereas the BNLM
or other analogous models might fit well BS-TSP comparison readings obtained
from sites sampling aerosols dominated by graphitic carbon and other particles
from fossil fuel combustion sources, such models would not likely apply in
markedly different circumstances, eg. in dry rural areas of the American Southwest
or sites strongely affected by fugitive dust from industrial facilities. Especially
highly variable BS-TSP relationships can be expected in such situations where
large amounts of coarse mode crustal particulate matter or mechanically produced
coarse mode particles from anthropogenic activities exists in the air in the
-8-
-------
presence of little fine-mode carbon or other materials from fossil fuel combustion
sources.
On pg. 3-141, additional sentences are to be added to the last paragraph,
as follows: On the other hand, another possible explanation for the marked
divergence of the data set from other published results for BS-TSP comparisons
(mainly sampling urban aerosol mixes dominsted by fossil fuel combustion emission
products) is that significant amounts of particles from fugitive dust or other
emissions from nearby industrial facilities may have contributed to relatively
high TSP readings (ca. 100 - 200 yg/m ) in the presence of low BS readings (<_ 25
o
yg/m ). Such results would be illustrative of one type of circumstance severely
limiting determination of reliable or meaningful relationships between corresponding
BS-TSP data points.
-9-
-------
Chapter 3 - PM/SO
Errata >
Page Par/Line
Delete
Insert
3-1
3-1
2/12
2/13
3-25
3-29
3-42
3-44
3-52
3-59
3-62
3-62
3-63
3-67
3-77
2/2
1/11
2/6
3/2
3/5
1/3
2/7
3/4
1/3
3-82 2/12
3-85 I/last
3-89
3-89 -/last
3-91 2/7
3-93 1/3
Higgins and Ferris, 1978
Speizer and Ferris, 1978
dates
Adams et al. , 1971
are
air (column 4, line 17)
reflectance
MckEllison (1964)
Page 3-52
"period" after: December 15
(column 4, line 8) Moulds,
1961
(Column 5, line 14) 10 cm
Mck
(Entire table heading)
length /mass
can
National Academy of Sciences
1978a
National Academy of Sciences,
19785
was after: which
as
error
National Survey after: for
darkness
Ellison (1968) after: by
Ellison (1968)
Page 3-57
, since
Moulds, 1962
10 mm
uncorrected after: usage of
Relationship of coefficient
of Haze to Particle Counts
(Np), 0.3 - 2.0um, Measured
in New York City
Source: Ingram (1969)
-2
length -mass
could
-------
CHAPTER 3. CRITICAL ASSESSMENT OF PRACTICAL APPLICATIONS OF SULFUR
OXIDES AND PARTICULATE MATTER MEASUREMENT TECHNIQUES
3.1 INTRODUCTION
In the preceding chapter (Chapter 2) of this document, physical and
chemical properties of sulfur oxides (SO ) and particulate matter (PM) were
discussed and numerous methods that have been developed for measuring air
concentrations of those substances were described. Also, information was
provided, in a general form, regarding possible sources of error associated
with various SO and PM measurement approaches, and comments were made
regarding the general specifity, sensitivity, accuracy, and precision of
measurement that can be attained with the different approaches.
In the present chapter (Chapter 3), ways in which some of the methods
described in Chapter 2 have been put to practical use in monitoring ambient
air quality will be discussed, accompanied by: (1) critical assessment of
factors found to affect the quality of measurements obtained with actual
applications of certain of the approaches, and (2) comparison of results
obtained with key applications of those approaches. Particular emphasis will
be placed in the ensuing discussion on those measurement approaches that have
attained a fairly high degree of widespread usage or which have played important
roles in the collection of air quality data discussed in later chapters of
this document. For other discussions of SO and PM measurement applications
consult: National Research Council/National Academy of Sciences (NRC/NAS)
%aJLMfjdc*£»!* rf^^^
reviews on Airborne Particles (H-iggins and ^orrio-, 197&? and Sulfur Oxides
^iuTiJthzMr**! rl^Um^
(Spoizor arva fcrrfs, 1978$; the WHO document, Environmental Health Criteria
(8): Sulfur Oxides and Particulate Matter (1979), ahd the recent American
Journal of Epidemiology report by Holland et al. (1979).
3-1
-------
3.2 HISTORICAL PERSPECTIVE
It is important that consideration of problems associated with the
practical application of SO and PM air quality measurement techniques be
undertaken against the backdrop of a historical recounting of factors that
stimulated and influenced the development and practical application of these
methods. By providing such a historical perspective, it is hoped that a
better appreciation can be achieved concerning the indispensible utility of
air quality measurements, despite their various imperfections, in helping to
quantify air pollutant-effects relationships and to reduce or eliminate such
effects as major public health problems.
There is little question that severe air pollution generated by anthro-
pogenic activities has long exerted significant, even lethal, adverse effects
on the health and welfare of many industrial societies. It has only been
within the past 50 years or so, however, that sufficient recognition of such
effects as a major public health problem has stimulated both (1) extensive
research to better understand the problem and (2) strong governmental and
private sector actions to control or eliminate it.
It is not surprising that the air pollution incidents most often cited as
signal events precipitating great concern about the problem and strong actions
in response to it occurred among the most heavily industrialized societies
then extant, i.e., in Western Europe, Great Britain and the United States. It
is also easy to understand the urgency associated with taking strong actions
to control the problem in light of the magnitude and seriousness of the effects
experienced during those incidents.
For example, in describing the effects of a thick fog that covered the
industrial Meuse Valley in Belgium during early December 1930, Firket (1931)
3-2
-------
noted that several hundred people were afflicted by suddenly appearing acute
respiratory symptoms, complicated in many instances by serious cardiovascular
failure. Firket (1931) further noted: "More than sixty died on the 4th and
5th of December after only a few hours of sickness. A sizeable number of
livestock had to be slaughtered." Also, taking into account that mortality
rates were more than 10 times normal, Firket projected that over 3,000 deaths
would occur if a similar fog were to occur in a city the size of London.
Twenty-two years later, such an event did occur in London and more than
4,000 deaths appeared to be attributable to the four-day London Fog of December
1952, according to Logan (1953). Logan further noted: "The incident was a
catastrophe of the first magnitude in which, for a few days, death rates
attained a level that has been exceeded only rarely during the past hundred
years - for example, at the height of the cholera epidemic of 1854 and of the
influenza epidemic of 1918-19." Indeed, the death rate rivaled or exceeded
that on many of the worst days of other catastrophic events afflicting London
in its more recent past, e.g., the Battle of Britain during World War II.
Similar catastrophic air pollution incidents also occurred around the
same time in the United States. Almost half of the residents of Donora,
Pennsylvania, for example, were afflicted with respiratory symptoms as the
result of a "smog" covering the coke- and steel-producing Monongahela River
Valley during October, 1948 (Schrenk et al., 1949). Twenty people in the
small town of about 10,000 population died during the final week of the "Donora
Smog Episode," in comparison to the 2 or 3 deaths normally expected for the
same period. Over the next few years, catastrophic air pollution incidents
also affected other United States communities. One such case was a dramatic
3-3
-------
increase in infant mortality in Detroit, Mich., attributed to a "pollution
incident" in September 1952 (Int. Joint Commission, 1960). This was followed
by marked increases in respiratory distress cases and fatalities attributable
(Greenburg et al., 1962) to air pollution occuring on an even larger scale
during a "Thanksgiving Day" pollution episode in New York City in November,
1953--an experience later to be repeated several more times in New York City
during the early 1960s.
It was clear from the above incidents and others occurring elsewhere in
the industrialized world that air pollution, especially under certain weather
conditions leading to stagnant masses of pollutant-laden "fog" or "smog", was
capable of causing incidents rivaling natural disease epidemics or man-made
wartime disasters that constituted national emergencies for the affected
societies. Equally clear was the urgent need to take immediate action to
avert or reduce in severity future air pollution disasters. Among the first
to act were two of the most severely affected countries, Great Britain and
the United States.
Many parallel historical threads can be discerned regarding the paths
taken by the two nations in trying to cope with the problem, including a
number of similar mistakes as well as successes. In each country, for example,
there occurred extensive expansion of epidemiological and toxicological research
aimed at identifying components of the killer smogs or fogs responsible for
the observed lethal effects and increased morbidity. Also, in each country,
although occurring within different specific time frames and at different
specific paces or rates, there ensued the upgrading and expansion of air
quality monitoring networks. This included the expansion of monitoring networks
capable of indexing high levels of various industrial pollutants implicated by
3-4
-------
the historical data as being associated with increased incidences of both
morbidity and mortality. The latter invariably included oxides of sulfur and
particulate matter present in the killer fogs or smogs of the 1940s and 1950s,
although the relative contributions of each could not be precisely linked to
observed health effects.
It is important to note in regard to the air quality data that very
precise accuracy and specificity were not then demanded; rather, "benchmark"
ranges of estimates of air pollutant concentrations were acceptable, especially
for indexing the rather high levels of pollution associated with severe morbidity
or mortality effects. It mattered little if the exact amounts of S02, for
3
example, were 143 or 781 or 1408 ug/m plus or minus 5 or 10 percent, when one
was concerned that people would begin to experience severe morbidity or even
o
die when S0? or particulate matter exceeded several hundred ug/m . Being
able to establish that increased mortality or severe morbidity occurred
at, say, a range of 550 to 600 or even 500 to 1000 |jg/m of S0? or particulate
matter in comparison to the incidence of such effects observed at, say, 70 to
100 or even 50 to 200 ug/m of the same pollutants was, understandably, sufficient
scientific evidence to support political and social measures needed to avert
the worst air pollution episodes.
Nor was it particularly important whether or not SO-, specifically, or
particulate matter of whatever specific size-range or chemical composition
could be precisely implicated as the "culprit" causing one or another very
specific health effect. Rather, it was sufficient to recognize that those
substances might not be any more than representative indicators of the total
mix or some other potentially lethal subfraction of pollutants typically
present during the dangerous past pollution episodes. Basically, the main
objective, regardless of specific fine details associated with various pollution
3-5
-------
situations, was to obtain sufficient information: (1) to allow for reasonable
conclusions to be drawn regarding ranges of air levels of various pollutants
(or indices) empirically linked to the occurrence of severe health effects and
(2) to help serve as a guide in directing pollution control efforts toward
sources emitting such pollutants (or mixes containing them).
In addition to the above developments, certain regulatory actions were
taken in response to the severe air pollution situations in Britain and the
United States in the 1950s. The British acted strongly on a national scale by
passing the Clean Air Act of 1956 which forced implementation of very stringent
controls on emissions from coal-fired combustion sources, essentially eliminating
the use of coal for home-heating and many industrial purposes in heavily
industrialized and congested urban areas. The effectiveness of those measures
was reflected in the resulting declines seen by the early 1960's in both
atmospheric sulfur oxides and particulate matter and associated mortality
effects (WSL, 1967).
As for the United States, action initially tended to be taken mostly on
more more restricted geographic bases and consisted mainly of local or state
air pollution control ordinances being passed, often with the cooperation of
local industrial leaders, to reduce air pollution. Among the more notable
examples were early actions taken to control air pollution in one of the most
heavily industrialized regions of the country, the Pittsburgh area. There,
and in other areas of Pennsylvania, extensive steel-making and coking operations
and the burning of locally produced high-sulfur coal contributed to widespread
elevations in both particuate matter and sulfur dioxide air concentrations.
Notable improvements in air quality in various American regions were attained
as the result of such initial actions. However, simultaneous deterioration in
air quality in many other communities or states lacking effective pollution
3-6
-------
control laws, and the growing recognition of air pollution as a multi-state,
regional or national problem eventually led to the passage in the United
States of the Clean Air Act of 1970 and consequent promulgation of National
Ambient Air Quality Standards for sulfur oxides, particulate matter, and other
air pollutants of widespread concern.
Subsequent to many of the above actions taken in the 1950s and early
1960s, questions began to be raised regarding what less severe, but important,
effects on human health and welfare might be caused by lower air levels of
oxides of sulfur, particulate matter, and other air pollutants, especially
under conditions of prolonged periods (months or years) of exposure. Again,
fairly similar paths were followed in both Great Britain and the United States
in trying to deal with such questions.
In both countries, epidemiologists began to design studies to either
retrospectively or prospectively define in more precise terms both qualitative
and quantitative relationships between elevations of various pollutants in the
ambient air and specific sublethal health effects, such as acute or chronic
bronchitis, respiratory infections, temporary decrements in pulmonary function,
and asthma attacks. Air quality monitoring networks set up or expanded earlier
to provide at least representative, but not necessarily thorough, coverage of
geographic areas having varying pollution levels were often looked to by the
epidemiologists to provide requisite, albeit less than perfect, quantitative
estimates of air quality to help define quantitative air pollution-effect
relationships. Only in relatively rare circumstances were sufficient funds or
other resources available to allow for more thorough monitoring coverage to be
arranged specifically for collection of community air quality data to be
coupled with health endpoint measurements.
3-7
-------
At the same time as the above expansion of research efforts, more precise
estimates of air pollutant levels were being demanded or expected from existing
or expanding air monitoring networks to help index progress in reductions of
air pollutants. Bearing on this point, the following was later noted in Her
Majesty's Report on the Investigation of Atmospheric Pollution 1958-1966
(Warren Spring Laboratory, 1967):
The industrial provisions of the Clean Air Act 1956 had come
into force by June 1958, and the first smoke control areas were
declared under the domestic provisions in April 1958. This timetable
added urgency to the view that the existing survey of air pollution
throughout the United Kingdom was too blunt an instrument either to
assess the benefits accruing from the Act or to guide its future
application, and that a scientifically planned National Survey was
necessary. Such a survey was designed and the co-operation of the
local authorities concerned was obtained where measurements were
required in addition to those already being made. Observations were
started in the winter of 1962-63. Not only has the whole pattern of
co-operative observations been transformed in this way, but so also
has the basis of the co-operation: the local authorities and other
organizations are now making measurements as required to conform to
an overall statistical plan.
A growing need was also felt for establishing or maintaining monitoring systems
using sufficiently uniform and reliable measurement approaches to allow for
comparability of air quality data from disparate geographic sites. This need
became perhaps most acutely felt in the United States in the late 1960s in
light of the growing prospect of having to attain national ambient air quality
standards at much lower air concentrations than were earlier envisaged.
Pressure to meet such needs and demonstrate the benefits of air pollution
control in the United States intensified greatly with the passage of the Clean
Air Act of 1970.
Both the above needs and the anticipated use of air quality data in
future epidemiology studies increased the necessity for better air quality
data in terms of their specificity, sensitivity, accuracy, precision, and
3-8
-------
reliability. Thus, it became increasingly more important to be able to
distinguish with confidence between much closer absolute levels of air pollutants
even at relatively low ambient air concentrations (e.g., at levels around or
lower than 100 to 200 ug/m of either S02 or particulate matter). In other
words, the emerging demands required practical applications of available or
newly developed measurement techniques to meet previously unheard-of levels of
accuracy. At times expectations may have exceeded what could realistically
be achieved with field applications of available technology, especially in
comparison to theoretical limits of what could be achieved with particular
measurement techniques under ideal laboratory conditions.
In responding to the above demands, several similar historical parallels
can again be discerned between the British and American experiences, as well
as some quite significant differences. In Britain, steps were taken to assure
that a high degree of uniformity in measurements of pollutants was maintained
or further enhanced; this included the establishment of the National Air
Pollution Survey alluded to above. As part of this effort, the British Smoke
(BS) filter method, widely used in Britain since the early 1900s for the
measurement of black suspended particulate matter, was officially adopted to
monitor air quality across the United Kingdom. The daily smoke filter has
since been the standard air particulate matter measurement instrument used in
the United Kingdom, and responsibility for quality assurance for the use of
the instrument throughout the entire United Kingdom was assigned to the Warren
Spring Laboratory (WSL) of the Department of Industry, where central coordination
and evaluation of uniformity, accuracy, and reliability of all National Survey
air quality measurement have been carried out since the late 1950s. Based
largely on WSL evaluations, the daily smoke filter was later adopted as a
standard also by the British Standards Institute and became one
3-9
-------
of OECD's recommended procedures for measuring suspended particulate matter.
The WSL was also instrumental in selecting SCL measurement methods for
the British National Survey air monitoring program. Under WSL guidance, the
hydrogen peroxide method was adopted in the early 1960s as the standard approach
for measuring SO-. This replaced the lead dioxide gauge still in use at many
sites in Britain (in 1962 there were about 1200 gauges in operation). The
decision to adopt the hydrogen peroxide method was based on recommendations of
a WSL-organized "Work Party" that considered both the deliberations of an OECD
Working Party and detailed comparisons of the two methods under field conditions
by WSL.
The above interplay between the WSL, the British Standards Institute, and
the OECD is illustrative of close working relations that broadly evolved
between the WSL, other British government groups, agencies of various foreign
governments, international agencies and other groups having interests or
responsibilities in research on or application of air pollutant measurement
methods. Such close working relationships further reflect the wide international
respect engendered by the WSL in the field of air pollution monitoring and
broad international acceptance of rigorous standards developed by the WSL for
practical applications of air pollution measurement techniques as used in the
United Kingdom National Air Pollution Survey.
Turning to concomitant developments in the United States during the past
twenty years, one can discern a lag in the development of a standardized nation-
wide approach to air pollution monitoring and control. During the 1950s and
early 1960s numerous air monitoring systems established by different governmental
units sprang up around the country, often to meet needs associated with
enforcement of newly enacted air pollution control -ordinances. It was not
until the mid-to-late 1960s that effective procedures were implemented to feed
3-10
-------
data obtained from the multiplicity of air monitoring sites operated by city,
county, state, and federal agencies into a central data bank as part of a
National Air Survey Network (NASN). Even then, those data were still often
derived from different measurement methods used by various agencies for monitoring
of a given pollutant. Generally, the Hi-Vol TSP measurement method developed
in the United States in the 1950s was used for assessing airborne particulate
levels; but, at times, other methods such as coefficient of haze (CoH) measurements
were also employed to assess particulate levels. Similarly, several approaches,
e.g., the West-Gaeke or sulfation methods, were used to measure SO,,.
Considerably greater uniformity in monitoring approaches has, however,
been achieved over the past decade or so through the publication of "Federal
Reference Methods." Additional efforts were undertaken in the late 1960s and
early 1970s by EPA and its predecessor Federal agencies to establish a new
nationwide air monitoring network using uniform measurement methods. That
network was established as part of what became known as EPA's Community Health
and Environment Surveillance System (CHESS) Program. The "CHESS" monitoring
network, set up in addition to other Federal air sampling stations used for
monitoring compliance with air regulations, included monitoring sites dispersed
in widespread urban and semi-rural areas of the United States to provide air
quality data representative of pollutant exposures experienced by surrounding
population groups. Various health endpoints were also evaluated for those
population groups as part of CHESS Program epidemiology studies. Thus, the
CHESS monitoring network, including sampling sites often situated near or
along side local or state monitoring sites, was designed to provide air quality
data from a nationwide network using uniform measuring methods that supplemented
other data entered into the NASN data bank. The hi-volume TSP sampling procedure
3-11
-------
was usually employed in the CHESS Program to monitor atmospheric particulate
matter levels and the West-Gaeke method was generally employed for S02 measure-
ments, along with additional procedures for estimating suspended sulfates
discussed below. The series of major air pollution/health effects studies
carried out between 1969 and 1975 as part of the CHESS Program, and coupling
such aerometric measurements with community health surveys, have been considered
by many experts as the most comprehensive of their kind.
It should be noted, however, that a number of methodological problems
were engendered by attempts to rapidly deploy air monitoring stations at
widespread sites across the United States and to bring them up to full
operational status in time to collect air quality data to be coupled with
health surveys as part of the CHESS Program. Of particular concern for the
present discussion are problems which were detected regarding certain errors
in air quality data generated from CHESS network sampling sites—several types
of errors which were either not detected at all during the CHESS health endpoint
data collection period (1969-75) or were only detected and corrected through
improved quality control procedures implemented and applied in the last few
years of the Program (i.e. 1972 or 1973 onward).*
*As discussed in more detail later in this chapter, the matter of errors in
air quality data collected as part of the CHESS Program (together with other
concerns regarding the collection of health endpoint data in CHESS studies)
contributed to considerable controversy regarding the validity and accuracy of
results of early CHESS studies, as interpreted and reported in a 1974 EPA
monograph entitled "Health Consequences of Sulfur Oxides: A Report from
CHESS" 1970-71, U.S. EPA Document No. EPA-650/1-74-004 (May 1974). The
controversy eventually led to the 1974 "CHESS Monograph" becoming the subject
of U.S. Congressional oversight hearings in 1976. Subcommittees of the U.S.
House of Representatives Committee on Science and Technology produced a report
on the Monograph, other aspects of the CHESS Program, and EPA's air pollution
research programs generally—a report entitled "The Environmental Protection
Agency's Research Program with Primary Emphasis on -the Community Health and
Surveillance System (CHESS): An Investigative Report." Of primary importance
for the present discussion, that report, widely referred to either as the
3-12
-------
Improvements in the air quality data obtained during the last few years of the
CHESS Program and, also, in other EPA air monitoring efforts, were accomplished
via a substantially expanded in-house EPA quality assurance program. That
program, conducted by EPA's Environmental Monitoring Systems Laboratory in
Research Triangle Park, N.C., has since provided quality assurance backup for
all of EPA's research and nationwide enforcement air monitoring activities.
From the foregoing, it can be seen that similar needs and demands as
those felt by the British in conjunction with the passage of their Clean Air
Act of 1956 were later experienced by the Americans with the passage of their
Clean Air Act of 1970. Remarkably similar paths were also followed in both
countries in responding to those needs; that is, in both cases intensive
efforts were carried out to rapidly expand and improve air monitoring capabilities,
including introduction of more uniformity in measurement approaches across
geographic regions along with increased quality control efforts to help assure
^(continued)
"Brown Committee Report" or the "Investigative Report" (IR), contained various
comments regarding sources of error in CHESS Program air quality data and
quality control problems associated with the data collection and analysis.
The I.R. also contained various recommendations to be implemented by the
Administrator of EPA pursuant to Section 10 of the Environmental Research,
Development, and Demonstration Authorization Act of 1978 ("ERDDAA," P.L.
95-155, 91 Stat. 1257, November 8, 1977). ERDDAA also requires that EPA and
the Agency's Science Advisory Board report to Congress on the implementation
of the IR recommendations.
One recommendation of the IR was that an addendum to the sulfur oxides
monograph be published, to be used in part to qualify the usefulness of the
CHESS studies, and to apprise the public of the controversy surrounding
CHESS. An addendum has been published, and is available from EPA, as
announced in the Federal Register of April 2, 1980, 45 F.R. 21702. The addendum
is incorporated by reference in this document in partial qualification of the
CHESS studies cited herein, and is part of the public file (or docket) established
for revision of this criteria document. The addendum contains the full text of
the IR, reports to Congress by EPA on its implementation of the IR recommendations,
and a report to Congress by EPA's Science Advisory Board on the same subject.
See also Appendix A of of Chapter 14 of the present draft criteria document.
3-13
-------
the validity of the aerometry data collected. Also, in each case it was
contemplated that such improved and uniformly obtained aerometry data from
throughout the two nations could serve as useful data pools to be coupled with
community health epidemiology studies aimed at both (1) improving knowledge of
quantitative air pollution-health effects relationships and (2) demonstrating
the benefits to be accrued from implementation of the respective Clean Air
Acts. In addition, as will become apparent below, remarkably similar problems
were encountered and responded to with roughly comparable degrees of success
in the course of practical applications of air pollution measurement techniques
undertaken to achieve the above objectives.
In addition to the above developments in the United Kingdom and the
United States, many analogous steps have been taken by numerous other industrialized
countries over the past 30-40 years to cope with air pollution problems. In
that regard, again many parallels (and dissimilarities) in the historical
evolution of their air monitoring programs, epidemiologic research efforts,
and political/legal regulatory control activities could be noted in comparison
to developments in Britain and the United States.
However, a historical review of the evolution of such other activities or
detailed analysis below of results obtained with practical applications of
other sulfur oxides or particulate matter measurement approaches employed
outside Britain and the United States is beyond the scope of present purposes.
Rather, major emphasis is placed below on critical assessment of measurement
techniques, mainly British and American, employed in collection of quantitative
air quality data employed in epidemiologic or other evaluations of the health
and welfare effects of sulfur oxides and particulate matter discussed in later
chapters of the present document.
3-14
-------
3.3 CRITICAL ASSESSMENT OF SULFUR OXIDES MEASUREMENT APPLICATIONS
Practical applications of measurement approaches employed in Great Britain
and the United States for determinations of air concentrations of oxides of
sulfur are critically evaluated in what follows. In both countries, lead
oxide "sulfation rate" techniques were often employed prior to 1960 for the
measurement of sulfur oxides. Later, other approaches used in some locations
in the two countries during the 1950s were adopted in an effort to establish
more uniform and improved SOp measurements for comparisons of data from widely
disparate geographic sites.
The following discussion on sulfur oxides measurement applications in
part alludes to published information on the relative specificity, sensitivity,
accuracy, precision, and reliability of the methods discussed when used under
optimum conditions in the hands of technically-expert analysts. Results
actually obtained, in the course of practical applications of the measurement
methods, often by less technically-skilled personnel, are also evaluated,
drawing mainly upon published commentary on quality control assessments for
the different applications.
3.3.1 British Approaches
As noted earlier, the lead dioxide gauge was used extensively in Britain
during the years prior to 1960. However, use of the hydrogen peroxide method
was gradually interspersed with the lead dioxide gauge during the course of
the 1950s, often being coupled in tandem, as it were, with the apparatus for
smoke measurements. Much of the early (1950s) British epidemiology data
discussed later in Chapter 14 of this document has been related to SO-
measurements obtained by the hydrogen peroxide method, especially where 24-hr SO- values
are used. Nevertheless, it is useful to compare the results obtained with the
3-15
-------
two methods, since other British air quality data of historical interest are
derived from the lead dioxide method.
3.3.1.1 Comparison Between the Lead Dioxide Gauge and Hydrogen Peroxide Method-The
National Survey-In 1962, as part of the establishment of the National
Survey, a working party was set up to compare the lead dioxide gauge with the
hydrogen peroxide method which was chosen as the standard method for use in
the survey. As quoted in Atmospheric Pollution, 1958-1966 (WSL, 1967):
The hydrogen peroxide method is subject to the limitation that
its reaction is not confined to sulphur compounds; the lead dioxide
method has the limitation that the extent of the reaction can be
substantially influenced by weather conditions. Despite limitations,
both methods estimate pollution by sulphur compounds; the hydrogen
peroxide method is somewhat more complicated, but has the outstanding
advantage that it can measure concentrations of pollution over short
periods; the lead dioxide method is simple in operation, but it is
incapable of measuring concentrations over short periods.
Even so, it was considered desirable to compare the results
from the two types of instrument under controlled conditions. A
statistical analysis was made by Warren Spring Laboratory of results
from a group of 20 sites at which both lead dioxide and hydrogen
peroxide instruments had been operated over a period of 48 months.
The 20 sites selected were those with a reasonably complete set of
results from March 1957 to February 1961 at which the two instruments
were not more than 100 feet apart.
The correlation between 829 pairs of results from the 20 sites
over a period of four years was highly significant, showing that
both instruments were predominantly affected by the same pollutant,
sulphur dioxide.
The WSL (1967) report presents a plot of these data shown below as Figure 3.1,
with the lead dioxide data reported as mg S03/100cm day.
The WSL fits these data with a linear regression line. Examination of
2
these data suggests that a line y = bx - ex might improve the fit. In this
case the origin is not an absolute, since the presence of ammonia can cause a
reduction in the SO,, reading by the peroxide method, but ammonia would not
3-16
-------
CO
111
cc
(N
X
O
Q
O
V)
O)
REGRESSION LINE
95% CONFIDENCE LIMITS
o - e
0 100 200 300 400 500 600 700
CONCENTRATION OF SO2 BY THE HYDROGEN PEROXIDE METHOD Aig/m3
Figure 3.1 A comparison of lead dioxide and hydrogen peroxide methods for
sulfur dioxide showing wide variations between simultaneous
measurements. The solid line is the regression line, and the
dotted lines are the 95 percent confidence limits. From WSL
(1967).
3-17
-------
influence the lead dioxide results. A correction for ammonia might shift
these data to the right enough to justify a line through the origin.
The ± 2o confidence limits shown in Figure 3.1 correspond to ± 1.8 mg
2 3
SCL/100 cm /day for a given hydrogen peroxide reading and ± 0.18 mg SOp/m for
a given lead dioxide reading. The WSL (1967) report concludes:
"The analyses carried out indicated that there is no generally
applicable calibration for relating lead dioxide and hydrogen peroxide
results. The conversion from lead dioxide to hydrogen peroxide
reading is not recommended except to give a rough indication of the
levels of concentration concerned, the degrees of approximation
being as indicated by the preceding paragraphs."
In other words, estimates of SOp levels derived from lead dioxide sulfation
rate measurements, especially 24-hr estimates, can only be roughly compared
with SOp estimates obtained by the hydrogen peroxide method at other geographic
sites or at later times at the same location(s). Also, it appears from the
data in Figure 3.1, that comparisons between sulfation rate readings are only
likely meaningful when such readings differ by the equivalent of about
180 ug/m of SOp, as measured by the hydrogen peroxide method.
3.3.1.2 Daily Sulfur Dioxide Measurements of the United Kingdom National Survey-
As noted above, when the National Survey began in 1961, it was recognized that
the lead dioxide method could not provide the 24-hour SOp measurements necessary
for correlation with mortality and morbidity effects investigated by epidemiology
studies. The hydrogen peroxide method for SO,, was, therefore, adopted as
being more valid than the old lead dioxide gauge sulfation method. Because
many of the staff making the measurements would be the same people who had
been servicing particle deposit gauges and the lead candles without detailed
technical knowledge of the analyses, however, an Instruction Manual (IM)
issued by WSL in 1966 had to be quite detailed and clearly readable by people
3-18
-------
with no training in analytical techniques. This section reviews the IM (1966)
and some published studies discussing factors affecting accuracy, precision,
and reliability of the British National Survey S0? measurements.
Reagents and Glassware-- In discussing reagents and glassware needed for
proper use of the hydrogen peroxide method, the IM (1966) states:
Distilled water. This should be obtained from a pharmacist or a
supplier of chemicals. It need not be carbon-dioxide free; 'purified'
water from an ion-exchange resin column may also be used...note that
it is not necessary to use Grade A glassware. Grade B quality is
often considerably cheaper and has sufficient accuracy for present
purposes.
The IM (1966) also states:
If it is considered that the cost of the extra Drechsel bottles is
prohibitive other containers can be used provided they have been
conditioned before use...The exposed solution should be treated as
soon as possible if containers other than Dreschel bottles have been
used."
The IM (1966) discusses the problem of alkali being given up by borosilicate
glass and the need to condition the bottles. Polyethylene containers were
permitted and a tolerable range of effect on the pH was given.
Martin and Barber (1971), however, reported that problems were found with
unstable reagents in the course of practical application of the hydrogen
peroxide method in the National Survey. They stated:
In October 1966, zero daily concentrations appeared at some of the
sites and in November alkaline samples were observed with no ammonia
present. The cause of this situation was eventually traced to a
change in the water used for making up the neutral hydrogen peroxide
reagent. A large container of water had been emptied and freshly
prepared demineralised water was apparently then being used. The
freshly prepared water was rich in carbon dioxide. As required, the
reagent was initially neutralised to pH 4.5, but as a result of a
day's bubbling and a week's standing, sufficient carbon dioxide was
evolved to make it alkaline. Demineralised water that stood for a
week before neutralising was reasonably stable in use, but glass-
distilled water was even better, and could be-used as soon as it was
cool. Tests with a spare sample held as long as the bubbled sample
3-19
-------
confirmed this stability. A comparison of results with those from
nearby local authority sites indicated that the use of freshly
prepared demineralised water could,,give monthly average results
around 40 (jg/m low out of 50 ug/m .
The extent to which WSL addressed this problem with carbon dioxide is
unclear, because a later report written by WSL (1977b) discussing quality
control tests states:
The tests also showed an unexpected, small, but significant,
loss in acid from acidified 1-volume peroxide solutions prepared at
WSL for the tests, within a few days of preparation; no explanation
for these losses has been found.
It appears, then, that at least some British National Survey SO- data
obtained in late 1966, and perhaps for a time afterwards, would tend to be
biased toward underestimation of actual SOp air levels due to the carbon
dioxide problem. The size of the error apparently could have been substantial
for late 1966 and, perhaps, early 1967 data reported before the error was
detected. Since then (at least up to 1977), errors due to this problem
presumably were not very large, in light of the steps WSL could have taken
to assure their correction.
Ammonia-- Ammonia is a negative interference since it neutralizes some of the
sulfuric acid. According to the WSL (1975, Volume 5) section on accuracy of
data:
Depending on the site and the season, ammonia may be present in con-
centrations which cause significant under-estimation of,sulphur dioxide,
of up to 80 ug/m on individual days, and up to 40 ug/m in the monthly
mean at country sites in summer (Martin and Barber, 1971). The under-
estimation seldom reaches 25 ug/m in urban areas. Sites have occasionally
been found at which there has been interference by other gases, e.g.
hydrochloric acid, but these are likely to be exceptional cases.
Where gross interference with sulphur dioxide determination is suspected
at National Survey sites, whether by ammonia or by some other gas,
specific measurements of sulphate or of the interfering agent may be made
as a check.
3-20
-------
The 1966 IM states: "If there is any reason to believe that the presence
of ammonia may be interfering with the estimation of sulphur dioxide, it is
not difficult to ascertain whether this is so." However, if the operator is
not adequately trained technically, the presence of alkalinity may not be
suspected until it is so great that the pH is greater than 4.5 without addition
of alkali. Martin and Barber (1971) also state: "The incidence of ammonia was
erratic. It could appear abruptly at many sites on one day and disappear from
them all on another." Thus, if a check is made a week later, ammonia might
not be present on the day the test is made, and this might lead the non-technical
observer to believe that no ammonia was present on the day the pH was over 4.5
and so not to repeat the test.
BS 1747: Part 3 (1969), Note 1 states that:
3
The correction seldom reaches + 25 ug/m SOp except in farming
areas etc. This value has been exceeded in about 10 percent of
samples over the last three summers, including some from town sites.
For Central Electricity Generating Board (CEGB) purposes, the
SOp results are corrected for ammonia, to allow comparison with
conductimetric recorder results (which are virtually unaffected by
these ammonia concentrations), and also because free SOp and ammonia
may co-exist in air without reaction for some time at tnese low
concentrations.
Warren Spring Laboratory, on the other hand, publish uncorrected
S0? values in their Monthly Summary of National Survey results, so
that all their values are on the same basis, i.e., as measured by
simple titration.
The difference between corrected and uncorrected monthly averages
at a site may be negligible or as great as 40 ug/m , depending on
the ammonia found. Forty microgrammes per cubic metre may be the
entire monthly average SOp at a country site.
From the foregoing, it can be concluded that, in general, National Survey
Sn2 data for United Kingdom urban areas probably represent some under-
estimations of actual levels due to ammonia as a source of error. Also,
in rare instances where they might be so affected, -the result would probably
3-21
-------
be no more than a 25 |jg/m underestimation of the actual S0? level. However,
for rural or small town sites with significant ammonia-generating sources
(e.g., livestock barns), the ammonia impact could be quite great—causing even
up to a 90 or 100 percent underestimation of actual SOp levels. Unfortunately,
from the specific information available to the present authors, 1t could not
be clearly determined as to which National Survey data from what areas or time
periods are likely so affected.
Sample Line Conditioning—Martin and Barber (1971) discuss the need for lacquering
to prevent wall loss of SO- to anodized aluminum and araldite in the manifolds.
Analysis of the data showed that "pre-laquering monthly averages were probably
3 3
low by about 25 ug/m out of 50 ug/m ." They also give an example where a
leak was discovered inside a metal conduit supporting the clean PVC tubing
3 3
which led to a drop of about 20 ug/m out of 50 ug/m . Sample lines also
deteriorate with time. Martin and Barber (1971) report that some PVC
deteriorated when exposed to sunlight, becoming dark and sticky and absorbing
SOp. "To a much lesser extent the same phenomena was [sic] found to occur
over 2 or 3 years on the routine surveys, inside windowless huts and even
inside metal conduit...Several instances were found of spiders' webs or insect
carcases, heavily sooted, inside the plastic tubing." Another potentially
more serious problem can be caused by "sucking back" of the bubbler solution
into the system if the air line is blocked. Martin and Barber (1971) stated:
From time to time the air line became blocked, e.g. by kinking,
with ice from freezing fog, or with insects. When this happened the
pump suction sometimes caused the contents of one or more Drechsel
bottles to be sucked back into the manifold, the filter assembly and
the connecting lines. This meant the loss of some samples and
careful cleaning and drying of the system was necessary. In some
cases the air line was freed and in others it was replaced.
3-22
-------
However, there was an increasing suspicion during Summer 1968
that further low results were being obtained. This was partly due
to the air line deterioration discussed in the previous section.
However, tests with a spare instrument at two sites showed that the
lacquering was also faulty and that relacquering would restore the
sensitivity of the instruments. It was not clear, though, whether
the original lacquering had been poor, or whether the lacquer had
deteriorated with age or perhaps by the action of reagent sucked
back.
The original IM (1966) does not discuss this problem. Thus, if it occurred,
the operators might neither have performed the scrupulous cleaning of the
entire system which Martin and Barber performed nor would they have necessarily
checked for lacquer deterioration caused by the reagent action. This could
presumably have led to underestimation of some SOp levels by 40 to 50 percent
based on the above examples cited by Martin and Barber (1971). It is not
presently possible to determine which National Survey or other British air
quality data may have been so affected other than to suspect that some 1968
data may be underestimates. However, an additional source of bias likely exists
from the reported adsorption and conversion of S0? to sulfate on the BS smoke
filter in the line before the collection bottle. This could lead to a con-
sistent underreporting of SOp.
Evaporation of Reagent—In regard to evaporation of reagent the British Standards
Institution (BSI) (1963) noted the following concerning the National Survey
method for determination of sulphur dioxide:
If the apparatus is left to run continously for more than
24 hours or in very hot weather a volume of 75 ml is preferable.
Evaporation should be kept to a minimum and any loss should be
made up with distilled water.
As pointed out by Fry (1970), if the bubblers are maintained inside a heated build-
ing the evaporation loss can be appreciable even in the winter. The pump also
runs hot, which can lead to heating of the bubbler if they are enclosed together.
3-23
-------
As stated in the Fry (1970) report;
It is shown that a significant error can occur in the estimation
of sulphur dioxide for localities of low pollution, due to loss of
water by evaporation during sampling. This error leads to the
reporting of high results and can be overcome by making up the
adsorbing peroxide solution to its original volume, prior to estimating
the net change in acidity.
Fry (1970) cites the following conclusion and recommendation to WSL.
CONCLUSION
It is concluded from the results presented here that evaporation
of the absorbing liquid can lead to significantly higher pollution
results but that this error can be eliminated by making up the
collected solution to its initial volume in accordance with B.S.
1747, prior to determination of net acidity.
RECOMMENDATION
That the Warren Spring Laboratory pollution survey procedure be
amended to include making up the collecting peroxide solution to its
initial volume prior to estimating the net change in acidity.
However, indications are this recommendation was not accepted by WSL, as
indicated by the following later statement by WSL (1975):
Evaporation of the absorbing reagent introduces an error. The
dilute hydrogen peroxide solution used to collect the sulphur dioxide
is neutralized to pH 4.5. If evaporation occurs during exposure the
pH is altered, and some of the alkali added to the exposed solution
in subsequent titration is used to bring the pH back to 4.5. Sulphur
dioxide concentrations are therefore over-estimated when evaporation
occurs. This was realized from the start of the Survey; but it was
decided that any attempt to correct this by ruling that samples be
made up to 50 ml before titration might well introduce larger errors;
recommendations are made regarding avoidance of overheating of the
instruments to minimize evaporation. The effect is likely to be more
prevalent in summer than in winter and can lead to over-estimation
of sulphur dioxide by up to about 15 ug/m .
Fry (1970), on the other hand, reported that a direct determination of
the correction could be made by simply subtracting 0.008 (50-V,) ml of 0.004 N
sodium borate, where Vf is the final volume in ml of solution after evaporation.
3-24
-------
There is no apparent explanation for why WSL did not provide for the recording
of the final volume on form WSL 216 AP (NS) and use of the simple correction
proposed by Fry (1970).
Given the evident failure of WSL to correct for this reagent evaporation
MrtJ
error, which states known to exist from the start of the National Survey, it
might be reasonable to conclude that all National Survey SCL data would be
biased somewhat high as a result. All past and present British National
Survey S02 data would then have to be taken as likely representing overestima-
tions of ambient SOp levels due to this error. The absolute magnitude of
overestimation of specific 50^ data points would vary with specific monitoring
sites (depending on care taken by different operators to avoid overheating of
pump motors) and likely be at a maximum during the summer months. Assuming a
3
maximum 15 ug/m error mentioned above (WSL, 1975), this could constitute
3
overestimations of 15 to 100 percent for obtained SO- values under 100 ug/m ;
3
7.5 to 15 percent for readings of 100 to 200 ug/m ; 3.25 to 7.5 percent for
3 3
readings of 200 to 400 ug/m ; and <3.25 percent for readings over 400 ug/m .
Titration Errors—As stated by WSL (1975), "Uncertain!"ty in the end-point of
the colorimetric titration, and rounding-off of the volume of alkali added in
titration to the nearest 0.1 ml, each introduce errors of up to about ± 5
pg/m3."
3
At low concentrations on the order of 50 ug/m , where only ~0.8 ml of
sodium borate solution is added, the total uncertainty from titration and
3
round off errors is about 7 ug/m or 14%. That is, 68% of the analyses would
be expected to be within ±14%, 95% of the analyses within ±28%, and 5% of the
analyses to be in error of more than 28%.
3-25
-------
Temperature and Pressure Corrections—The BSI (1963) method for sulfur dioxide
discusses temperature and pressure corrections as follows:
The volume of air sampled is measured at the temperature and
pressure in the gas meter; if it is desired to relate the result to
external air volumes, further corrections for the temperature and
pressure outside the laboratory are required . . . The concentration
of atmospheric sulphur dioxide is preferably expressed in terms of
the weight rather than the volume of sulphur dioxide per unit volume
of air.
Clause 5 expresses the concentration in microgrammes per cubic
metre of air sampled, without adjusting the measured volume of air
to standard temperature and pressure and without allowance for the
small drop in pressure at the meter. This procedure is considered
to be adequate for the general run of routine observations. Some
discretion may need to be exercised if the conditions are such that
clogging of the smoke filter causes an unduly large pressure-drop at
the meter.
Where it is considered desirable to make adjustments for
temperature and pressure, the apparatus shown in Fig. 2 should be
modified to provide for measurement of the pressure-drop at the
meter and suitable provision should be made for recording the
ambient temperature and pressure during the period of the test at
such intervals as will give fair average values.
However, clogging of the smoke filter occurs most often on the most
polluted days. Consequently, the volumetric flow at the meter is enhanced by
the drop in pressure, and the meter will report too much flow (1 meter at 1
atmosphere occupies approximately 2 meters at 1/2 atmosphere). Consequently,
the effect will be to underestimate the concentration of both SOp and the BS
on the filter. For example, a 3-inch drop in pressure across the filter
results in approximately a 10 percent underestimate of the S0? and BS.
3.3.1.3 Summary—The assumptions and errors discussed above may be conservative
because in many cases "staff making the measurements had no training in analytical
techniques" (IM, 1966). Of course, if such staff followed each and every
instruction in the manual to the letter then the literature estimates cited
above might be valid. However, if one allows the possibility that techniques
in the field varied measurably from those set forth in the manual or likely
3-26
-------
to be followed under laboratory conditions, then these error estimates may
need to be enlarged. Taking into account all of the uncertainties due to
titration, ammonia, carbon dioxide, variations in quality of distilled or
demineralized water, Grade B glassware, polyethylene bottles, evaporation of
reagents, anodized aluminum, insects and leaks in sample lines, sample line
deterioration, improper laquering of parts in contact with S02, and other
miscellaneous errors such as those due to uncertain waiting time before titra-
tion of samples not transported in Drechsel bottles, one might expect that
some specific sets of British National Survey S0» data could be in error by as
much as 50 to 90 percent.
Ascertaining which specific data sets might be thusly affected to a
greater or lesser extent would, of course, require extensive and extremely
careful reexamination of a!1 National Survey SO,, data points and associated
records documenting pertinent information on operator training, individual
differences in carrying out lab procedures, and day-to-day condition of the
sampling and laboratory equipment, reagents, etc., on a site- and time-period
specific basis.
Alternatively, one could assume that most all of the errors cited, with
some operating to bias the reported data upwards and others downward, would
tend to cancel out, especially when data are averaged over periods of a few
months or a year or for groups of sites, as has been suggested by WSL officials
in Her Majesty's Report: "National Survey of Air Pollution, 1961-71: Vol. 5"
(WSL 1975). It is likely, however, based on the above WSL reports that at
least some errors are generic to all National Survey measurements of a parti-
cular type and would not likely be regularly cancelled out by others. Thus,
for example, the "reagent loss" error discussed above would likely cause
3-27
-------
reported National Survey S0? values to consistently be overestimations of
actual SCL levels in areas of low levels of particulates and ammonia.
3.3.2 American Approaches
As noted earlier, under "Historical Perspective" above, different types
of measurement methods for a given pollutant were at times adopted by various
local, state, and federal agencies in establishing or expanding air quality
monitoring systems that proliferated across the United States during the 1950s
and 1960s. It is beyond the scope or purpose of the present document to
attempt a critical appraisal here of the procedures employed in, and results
derived from, the practical application of SO and particulate matter measurement
A
techniques under the varying specific circumstances associated with the operation
of all of the different American air monitoring systems. Rather, main emphasis
will be placed here on the discussion of only certain key American applications
of measurement methods for SO that are of crucial importance for later discussions
in the present document of quantitative relationships between health and
welfare effects and atmospheric levels of sulfur oxides and particulate matter.
These include applications of S02 measurement methods as employed in the EPA
"CHESS Program" as the single largest attempt to define quantitative relationships
between air pollution and health effects. Also, as appropriate, certain
pertinent information will be discussed regarding air quality measurements
associated with the operation of some local (city or county) and state air
monitoring programs.
In regard to sulfur oxides measurement approaches used in the United States,
lead dioxide or other "sulfation rate" measurement methods were, as in Britain,
widely employed prior to the early 1960s for assessing SOp air levels. However,
probably to a somewhat greater extent than in Britain, sulfation rate measurement
techniques continued to be used later into the mid or late 1960s by some
3-28
-------
monitoring programs in the United States or in connection with certain
community health epidemiology studies, as discussed later in Chapter 14 of
this document. As shortcomings of the "sulfation" methods of the type discussed
in Chapter 2 became more widely recognized, their use was generally abandoned
and more specific methods for the measurement of SC> or other sulfur oxide
compounds were adopted, as was done in Britain. The hydrogen peroxide acidimetric
method (see OECD, 1965) selected for use in the British National Air Pollution
Survey, however, was not very widely adopted in the United States for S02
measurements. Rather, versions of West-Gaeke (1956) colorimetric procedures
were much more widely used in the USA. Conductivity measurements for SO,,
s-ytr-al. , 1971), based on an acidimetric method adaptation often used in
automatic instruments and most suitable for measuring periods of around 24
hours, later began to be applied in the operation of some American air monitor-
ing networks in the 1970s.
The West-Gaeke method was the method mainly employed in the EPA "CHESS
Program" for determining SO- air levels for inclusion in analyses of community
health end point data in "CHESS" epidemiology studies. The application of
that method in the CHESS Program is accordingly most thoroughly discussed
below. Applications of other measurement approaches for sulfur oxide compounds
other than SO- (such as suspended sulfates) as employed in the CHESS Program
are also discussed below. Much of the information provided is derived from a
1976 Congressional Investigative Report (IR) which contained a thorough evaluation
of EPA CHESS Program air quality measurements and other aspects of the Program.
3.3.2.1 Sulfur Oxides Measurements of the U.S. EPA CHESS Program— As indicated
in Section IV A of the IR (1976), the attainment of precise, reliable, reproducible,
and real-time air quality measurements in the field (e.g., SO- and particulates)
was a critical purpose of the EPA CHESS Program. The IR (1976) further notes
3-29
-------
that the methods selected for use in CHESS to meet these objectives, especially
in 1970-71, were probably as good as any available at the time; and attempts
were made to gradually introduce better quality control procedures into the
CHESS program. Thus, the use of the particular methods selected (the best
available), and the goal of implementing quality control procedures in the
Program cannot be criticized. However, the lack of sufficient quality control
efforts to ensure the best possible application of those methods in the field
and the resulting errors in measurement, especially in the early (1970-71)
years of the Program, were and are matters of considerable concern that led to
criticisms of the type voiced in the 1976 IR. Some of the more salient points
raised by the IR regarding CHESS approaches to measurement or estimation of
S0? or suspended sulfate levels are discussed below.
3.3.2.1.1 The West-Gaeke Method for Measurement of Ambient SO,,. As briefly
described in Section IVB of the IR (1976), the West-Gaeke colorimetric procedure
for S02 measurements used in CHESS is:
The designated Reference Method (Federal Register, 36, No. 84,
6168, April 30, 1971). Atmospheric SOp is collected by bubbling air
through a solution of potassium tetracnloromercurate (TCM). The
product of the reaction between S0? and TCM is the nonvolatile
dichlorosulfitomercurate that is tnen determined quantitatively by
reaction with formaldehyde and pararosaniline hydrochloride, followed
by photometric measurement of the resulting intensely colored para-
rosaniline methyl sulfonic acid.
The IR further noted that the field apparatus and sample collection
procedures employed were as follows:
Outside,air is drawn through a sample line at the rate of 200
ml min , then through a 6-inch long glass bubbler stem (tip diameter
of 0.025 in.) immersed in 35 ml (50 ml after January, 1974) of 0.1 M
TCM solution contained in a 32 mm diameter by 164 mm long polypro-
pylene sample container. The exhaust air passed through a glass
wool moisture trap, then through a hypodermic needle used as a
critical orifice to control the flow, through another moisture trap,
3-30
-------
and finally through a vacuum pump. A sample consisted of a 24-hour
collection. Collected samples were stoppered, and mailed to EPA/RTP
for analysis."
IR comments concerning the validity of the method as a laboratory
procedure included:
A collaborative study by McKee et al. (H. C. McKee, R. E.
Childers, and 0. Saenz, Southwest Research Institute, SWRI Project
21-2811, EPA contract CPA 70-40) indicates that "the method cannot
detect a difference smaller than 10 percent between,.two observations
by the same analyst in the range of 0 to 1000 ug m ._ A difference
of 20 percent or less may be detected above 300 ug m , and a _3
difference of less than 50 percent may be detected above 100 ug m ."
For analyses conducted by different laboratories on the same sample,
"the method cannot detect a difference of less than 20 percent betweem
single-replicate observations of two laboratories in the range of 0
to 1000 ug m . At a level of 100 ug m , a difference of less than
100 percent is not detectable." The National Primary Ambient 3
Air Quality Standard for S02 is: For 24 hour average, 365 ug/m . For
annual average, 80 ug/m . Thus3if the standard is met, most values
willobe around or below 80 ug/m , no more than one will be above 365
ug/m .
Regarding the lower linnt of detection, the authors cited above
propose a value of 25 ug m as a practical figure. "A single
determination less than this value is not significantly different
from zero" (Instrumentation for Environmental Monitoring, Air-S02,
Instrumentation, Lawrence Berkeley Laboratories, March 1972).
It is therefore evident that a single analysis is of little
use, considering that the expected concentrations of S023will usually
be less than the ambient air quality standard of 80 ug m . Results
should be regarded as valid only in terms of the mean of multiple
determinations, and only when the analytic method has been followed
rigorously by experienced analysts.
The IR (1976) further noted the following regarding sources of errors in
CHESS S02 measurements and their likely impact:
In an attempt to standardize the methodology and to eliminate
problems associated with interlaboratory errors, a CHESS policy was
instituted whereby all air sampling equipment was assembled and
tested at the central EPA research laboratory and then shipped to
the contractors for field use. Also, bubbler tubes were prefilled
with the appropriate absorber solution, shipped to the contractor
for their daily monitoring use, and shipped back to the central
laboratory for chemical analysis. It was this long distance shipment
of the chemical solutions that led to the first of a series of
field-use problems with the procedure. These problem areas will be
3-31
-------
summarized below with an attempt to evaluate their net effect on the
resultant CHESS S02 data. Following this summary of individual
problem areas, an assessment of the overall S02 data quality will be
given.
a. Spillage of Reagent During Shipment
The first field data were obtained in New york City and
the Salt Lake area (Utah) in November, 1970. By mid-1971, field
personnel at the Utah site reported to their CHESS field engineers
that severe spillage was occurring during shipment. Many
bubbler tubes were arriving partially filled with reagent and
some were completely empty. At the Salt Lake area an attempt
was made to refill with solution from extra tubes those tubes
that were low. However, due to insufficient reagent, this was
only partially successful. This problem was not officially
recognized until October, 1972, at which time an internal
EPA/CHESS memo was written outlining the problem and suggesting
corrective action. The magnitude of the problem can be best
assessed by quoting from the memo. "The present reagent tubes
for S02 and N02 leak during shipment .... The S02 leakage
rate (was found to be) 18% of the total volume, 50% of the
time. ... It follows therefore, that the resultant pollution
data are unreliable." Recommendations were made in this memo
as to possible corrective measures. These recommendations were
not instituted until March, 1973.
During the subsequent years, many attempts were made to
correct this leakage problem. However, none were wholly successful
and as late as January 1975, another EPA memo described losses
of solution in S02 bubblers during shipment and suggesting
appropriate corrective action.
The effects of the reagent spillage problem on the S02
data can be only grossly estimated. Certainly, many samples
were totally lost. These lost samples were not the major
problem. Of more significance was the undetermined amount of
daily S02 data that were in error due to the loss of sample by
spillage and yet included in the network system.
If the reagent was partially lost during shipment to the
sampling site and used as received, an increased concentration
of TCM~S02 complex would occur relative to normal sampling.
This potential positive bias would be corrected by for the
analytical procedure used (Page A-6 CHESS monograph—Analysis
Procedure). "At the laboratory, the sample is brought back to
its original volume by the addition of distilled water to
compensate for water loss during sampling." If however, the
reagent spillage occurred after sampling, the required addition
of water would result in data that were biased low in proportion
to the amount spilled relative to the total volume of solution.
According to the EPA Memo of October, 1972, one half of all S02
data taken between November, 1970 and March, 1973 are likely to
have been biased low by an average of 17%. This problem was
corrected after April, 1973.
3-32
-------
b. Time Delay of the Reagent--S0? Complex
The Reference Method as originally described in the Federal
Register, was to be conducted at 20° C. There was a known error in
the method associated with time delay between sampling and analysis
which was dependent on temperatures. This error was derived from
the spontaneous decomposition over time of the TCM-S02 complex as a
function of temperature. The magnitude of the error and its exact
dependence on temperature was not known but a brief study was conducted
to determine its magnitude by scientists of the CHESS monitoring
group in November, 1971. As a result of this study, a correction
factor of +1.5% per day was arithmetically applied to all CHESS S02
data to compensate for the time delay between sampling and analysis.
A more recent and comprehensive study has been carried out
within the Quality Control Branch, Environmental Monitoring Laboratory
at EPA on the effect of temperature on "The Stability of S02 Samples
Collected by the Federal Reference Method." This study indicated a
much more severe problem than was estimated by the original CHESS
study. The evaluation was carried out over the range of 35 to 278
ug/m3 S02 concentration. The following findings were presented in
the report:
Over a normal range of temperature, the rate of decay
of the TMC-S02 complex increases five-fold for every 10°C
increase in temperature, respectively.
The rate of decay is independent of S02 concentration.
At 20, 30, 40, and 50°C the following S02 losses were
observed: 0.9, 5, 25, and 74% loss per day, respectively.
This study makes abundantly clear a second and even more
severe error associated with S02 measurements conducted by CHESS.
During the summer months, when the S02 absorber solutions were
subjected to high and unknown temperatures between field sampling
and laboratory analysis, significant degradation of the samples did
occur. Estimates of time delay between sampling and analysis range
from 7 to 14 days. Estimates of summer temperature exposures range
from 25 to 40°C being most severe for the Utah CHESS sites. Thus,
CHESS S02 data can be estimated to be negatively biased, mainly
during the summer months. It would normally be difficult or impossible
to estimate the magnitude of the bias except to say that it is
probably large. However, simultaneous S02 measurements were taken
by the New York City Department of Air Resources and by the Utah
State Division of Health. These results were obtained by an independent
method not susceptable to the temperature related error. A consistent
pattern emerged when side by side data are compared. From May to
October, the CHESS S02 data were low with the largest error occurring
in the middle three summer months. The magnitude of the error
varied from month to month and year to year, but the CHESS data were
consistently low and represented only a portio'n of the true ambient
S02 concentration.
3-33
-------
c. Concentration Dependence of Sampling Method
The S02 reference method was subjected to a collaborative
study program in 1973. Four participating laboratories tested the
24-hour version of the Federal Reference Method. A previously
unknown source of error was documented that applies to the CHESS SQ2
data. It was found that the 24-hour sampling method does have a
concentration dependent bias which becomes significant at the high
concentration levels (200 ug/m3). Observed values tend to be lower
than the expected (known) S02 concentration levels. This error
source will yield a negative bias on the daily CHESS S02 data when
they exceed 200 ug/m3 and on all monthly and yearly average data.
d. Low flow correction
The determination of atmospheric S02 concentration was
dependent on, among other factors, the accurate measurement of air
that passed through the TCM solution. This flow was controlled by a
critical flow orifice in the form of a standard hypodermic needle.
In practice, the air flow through the sampling system was measured
at the start and end of each 24-hour sampling period. This was done
to detect low flow due to needle blockage. The Federal Register
Method (Reference Method) calls for an air low of 200 ± 20 ml/min.
In field operation, the CHESS procedure substantially broadened
these tolerances. Replacement needles were installed if the initial
air flow was greater than 220 ml/min which is consistent with the
Reference Method; however, needles were not replaced nor were samples
voided until the measured flow dropped below 100 ml/min. Integrated
flows were calculated by assuming a linear decrease in flow between
the start and end of the 24-hour sampling period. If, however, the
needle was partially blocked near either the beginning or the end of
the sampling period, the linear flow correction would be in error.
Using the Reference Method flow tolerance, only small errors would
be introduced by this correction (less than 10%). Using the CHESS
procedure, however, errors as large as 50% could be introduced and
not detected. These errors would be random (either positive or
negative) depending on when during the sampling period the needle
blockage occurred. Thus a large random error component was added to
the S02 daily data but this component was somewhat damped statistically
in the monthly or yearly averages.
The modification of flow tolerance by the CHESS aerometric
group is a procedure that would not have withstood the critical
review of a competent quality assurance program.
e. Bubbler train leakage
The West-Gaeke method, as described in the Federal Register,
employs a vacuum bubbler train. That is, the sampled air is
drawn through the bubbler train by a vacuum pump rather than
being pushed through by a positive pressure pump. There are
many advantages to the vacuum procedure, most important is that
3-34
-------
the air does not come in contact with any internal pump mechanism.
However, there is a modest pressure differential between the atmosphere
and the internal bubbler; thus all fittings and joints must be gas
tight. The bubbler train used in the CHESS program had two points
where frequent air leak problems were encountered. One was around
the rubber stoppers for the bubbler tube and moisture trap and the
other was the rubber tubing used to hold the glass assembly pieces
together. Field operators reported consistent problems with leakage
in the routine field use of the bubbler train. In a severe leak
situation, the samples were voided due to out of tolerance (Low)
flow rates. There were many cases however, where small leaks occurred
but the final flow was within specifications so the sample was included
as valid. In cases where the leaks formed around the rubber stoppers,
no significant error would be introduced except due to the linear
flow correction as applied to instantaneously developing leaks.
This error is similar in nature to that discussed in the flow section.
In the case of leaks upstream of the bubbler train, room air instead
of outside air is drawn through reagent. In normal situations, it
has been observed that room air is significantly less polluted than
outside air. (See page 6-6, CHESS Monograph—comparison of school
air to outside air). This effect may not be large for the small
buildings used to house CHESS stations, but a somewhat decreased
pollutant level would undoubtedly be sampled. The absolute magnitude
of this error cannot be adequately assessed but it can be stated
that the error would be in a negative direction, that is, again to
underestimate S02 levels.
An overall assessment of CHESS S02 data from Section IV C 2 of the IR (1976), is
as follows:
The S02 data, accumulated at "official" CHESS sites, followed
a remarkably uniform trend as the program progressed. The method
used was the EPA Reference Method which is specific for the chemical
species, S02. Thus, regional changes in pollutant mix, i.e., the
proportion of other pollutant species relative to S02, had minimal
effect on the S02 data. However, the sum effect of the errors
detailed in this section did have a profound effect on both the
accuracy and the precision of the data.
Under normal circumstances, a retrospective evaluation of a
monitoring effort that occurred a number of years in the past and
which had been terminated, could yield only the broadest of estimates
of data quality. Fortunately for this review, two geographically
different locations with six different monitoring sites were involved
in the collection of simultaneous S02 data. Further, the groups
responsible for the two data sets were managed independently and the
methodology used was also independent. This fortunate circumstance
enabled the reviewers to acquire a quantitative understanding of
absolute differences among data sets as well as correlations with
respect to time.
3-35
-------
The locations where side by side data existed were the New York
City sites at Bronx and Queens and the Salt Lake Basin sites at
Ogden, Salt Lake City, Kearns, and Magna. In these locations, the
local environmental monitoring agencies had sites located within 50
meters of the CHESS sites and at similar elevations. At these
sites, the local agencies collected daily S02 and TSP data for the
entire life of the CHESS program. The S02 methodologies used by
both State agencies were variations of the peroxide bubbler method
in which twenty-four 1-hour samples were integrated to form a single
24-hour S02 measurement. In New York the samples were measured
acidimetricly and in Salt Lake City they were quantified conduct-
iometrically. Neither method is as specific for S02 as is the
Reference Method, that is, pollutants that are in a significant
concentration, relative to S02 and that also oxidize to form an
acidic compound will be interpreted as S02. For this reason, when
the NYC Department of Air Resources initially brought to the attention
of the CHESS Aerometric team the large discrepancy between their
respective data, the discrepancy was dismissed as method bias on the
part of the New York method. An EPA memo dated November 3, 1971
described a limited study into the Reference Method. The conclusion
reached was "On the basis of (this study). . .1 feel there is no
sound basis for discrediting the EES (Environmental Exposure System)
methodology.
No further attempt was made to uncover the cause of the discrepancy
in S02 data. Had the CHESS EES team obtained and compared the Salt
Lake Basin data, especially that from Magna site, a disturbing
similarity would have been immediately apparent. This data confirmed
in detail the discrepancies observed in New York. It is important
that the Magna site data were confirmatory since it was in a region
of single source pollution, that from the nearby copper smelter. In
this site very low levels of other pollutants existed relative to
S02, thus the peroxide method was capable of giving reasonable
reliable estimates of S02 concentration. Of equal importance the
general pollutant mix was very different between this rural smelter
site and the urban area of New York City. Despite these diffences
the comparison of side by side Federal-State data indicate the same
discrepancies in both trends and absolute concentrations. The
following conclusions as to S02 data validity can thus be reasonably
drawn from the review of methodological errors and the comparison of
existing side by side data.
The following conclusions regarding CHESS S0? data validity were drawn
in Section IV C 2 of the IR (1976), based on the above review of methodological
errors and the comparison of side by side data:
3-36
-------
From November 1970 until December 1971 the S02 data generated
from CHESS sites using the modified Reference method were biased low
by 50 to 100 percent in the High Exposure sites when compared with
existing State S02 data. Thus, the 1971 annual average S02 exposure
estimates of 60 ug/m3 as reported for Magna in the CHESS monograph
(page 2-24) are more likely in the vicinity of 100 ug/m3. Also, the
same phenomenon occurred in New York and the reported values are
also in similar error.
A confirming fact is that during cool months after 1971 S02
data correlated well both in trends and absolute concentrations
between State and Federal analyses. It thus seems likely that the
State data were reasonably accurate throughout that time period.
However, one consideration must be applied here: namely, that due
to the difference between the independent methods an error bar of at
least one hundred percent must be applied to the data and explicity
correct data cannot be drawn from these observations. In other
words, where two or more independent observations are in disagreement
by a significant amount it cannot be said by inference alone that
one data set is more correct than the other. It is reasonable to
assume, however, from our review of all State and Federal data in
the time period of 1970 through 1971, that the Federal S02 data as
collected in the CHESS program were substantially low and went
through an abrupt upward transition in concentration in December
1971 at all CHESS sites and Federal data taken before that time may
reasonably be expected to have a large unknown negative bias.
In November 1971, the CHESS monthly mean S02 data underwent an
abrupt change in the positive direction. The cause of this change
is not apparent. However, the result was profound. From that time
until the conclusion of the CHESS program in July of 1975, the
fall-winter data were in very good agreement with other exisiting
data and very likely gave reliable estimates of S02 exposures.
Throughout the entire program, the CHESS S02 data had an associated
negative bias during the summer months, becoming most severe during
the hottest periods of July and August. This error usually reached
a maximum of 60 to 80 percent underestimation of exposures and was
variable. As a result, even though wintertime monthly S02 averages
appear valid from 1972-1975, annual averages of the same data are
biased low due to the inclusion of the summer errors. The best
estimate of error in the annual average data 1972-1975 is approximately
minus 15-20 percent relative.
The individual daily S02 levels, when compared to city or State
data or to replicate CHESS measurements taken after 1973 had so
large a random error component that they are not useful to assess
daily S02 exposure (as attempted in the asthma panels). The random
errors associated with the daily values were much larger than the
differences observed over time.
3-37
-------
Due to inherent methodological errors, the following may be
considered as minimum differences between High and Low S02 exposures
which may be considered "real." These are based on EPA's collaborative
study of the reference method and used a 95 percent confidence
interval.
Below 100 ug/m3 S02, a difference of at least 50
ug/m3 is necessary to be statistically significant.
Between 100 and 300 ug/m3 S02, a difference of at
least 60 ug/m3 is necessary to be significant.
Below 25 ug/m3, a single determination is not signifi-
cantly different from zero.
3.3.2.1.2 CHESS program suspended sulfate measurements. Determination of
suspended sulfate (SS) concentrations in the air as part of CHESS aerometry
efforts was carried out as methodological extension of the Hi-Vol TSP method
employed in the CHESS Program. That is, subsamples were cut from the exposed
Hi-Vol filters and analyzed for total water soluble sulfate as a class (but
not according to specific chemical species). The manual turbidimetric method
was used from November, 1970, to September, 1971; and the methyl thymol blue
(MBT) method was employed from September, 1971, to July, 1975. To the extent
that the SS measurements were derived as an extension of the TSP sampling, all
errors associated with the TSP monitoring, as discussed later below, also
apply to CHESS SS measurements. In addition, some more specific considerations
pertaining to the SS measurements can be cited.
Section IV C 4 of the IR (1976), for example, noted the following:
The turbidimetric method is subject to interferences, many of
them being other common pollutants. In areas like the Salt Lake
Basin where the pollutants are dominated by a single source, the
procedure may be adequate. However, in urban areas like Cincinnati
or New York City, where the pollutant mix is derived from many
independent sources and is variable even within the city, the method
is capable of only the crudest estimates of sulfate levels. It
should not be thought of as an accurate measurement of atmospheric
sulfate. Especially, small differences between High and Low exposure
communities, such as were reported in the Cincinnati Study in the
3-38
-------
CHESS Monograph (page 6-5) cannot be identified as real differences.
When a realistic error estimate is applied to the reported sulfate
concentrations, the difference becomes statistically insignificant.
Any correlation of CHESS health effects with sulfate levels where
the sulfate data were obtained using the turbidimetric method must
be carefully qualified.
The MTB method is basically a better measurement method because
most of the aerometric interferences have been eliminated by its
revised methodology. The two remaining interferents, phosphate and
barium, are not normally found in atmospheric concentrations high
enough to cause inordinate problems. However, problems associated
with the sampling aspect of the method have been documented and do
impact on the general CHESS sulfate data quality.
First, problems associated with sulfate blanks (the level of
sulfate on the filter pad as manufactured) were reported to be high
and variable. In the 1971-1973 time period, problems of variable
blanks within the EPA NASN program were documented. The general
blank level was equivalent t6 ari atmospheric sulfate concentration
of 1-2 ug/m3. However, the major problem was variability of the
blank among manufactured lots of the filters. The blank level often
varied by more than 100 percent among lots so that routine and
continuous blank assessment should have been mandatory.
No evidence of routine sulfate blank determination was found in
the CHESS monitoring program until 1974. From that time period on,
adequate blank assessment and correction were applied to the data.
From 1971 until 1974 however, the blank contribuiton to the CHESS
sulfate data was not adequately assessed and consequently a positive
and highly variable bias of unknown magnitude was included in the
data.
Second, adsorption of atmospheric S02 onto the fiberglass
filter material followed by spontaneous oxidation of the S02 to
sulfate had been well documented. A 1966 publication by R. E. Lee
and J. Wagman provided results of their investigation of the problem.
The conversion was clearly documented with severe effects demonstrated
on four-hour samples. The conversion did appear to be an active-site
catalytic conversion that decreased in magnitude .after an initial
saturation of sites. Thus, 24-hour samples were much less affected
by this problem than were those taken for shorter time intervals.
Even so, the paper by Lee and Wagman presented data in which routinely
0.5 to 1 (jg/m3 of the measured sulfate was derived from S02 conversion
products. The maximum conversion presented was 2.1 |jg/m3 derived
from S02; this constituted a 10 percent positive bias of the sulfate
data. A more realistic average bias is likely in the 5 percent
range. However, there is clear evidence that in regions of high
levels of S02, relative to sulfate, the positive measurement bias
becomes much more severe. This is probably the case in the Salt
Lake Basin area.
3-39
-------
The third and most devastating problem associated with the
CHESS sulfate data occurred when the laboratory analysis of sulfates
was contracted to an outside firm. During this time period (October
1972-June 1974) the reported sulfate data underwent a sudden and
sustained decrease in apparent atmospheric sulfate level. Upon
investigation it was determined that the laboratory analysis of all
sulfate data from all CHESS sites were biased low by approximately
50 percent. The reason for this negative bias was and still 1s not
completely clear, but the continued dissemination of poor data was
clearly due to inadequate quality controls. An interim EPA report
on a retrospective quality assurance evaluation of CHESS Sulfate
Data states:
A quality control protocol was designed for CHESS
chemical analysis but has not been implemented as per the
contract The quality control protocol should be implemented
immediately.
In a series of following studies the magnitude of the affected
data and of the error were documented and an attempt was made to
correct and therefore recover the data. This type of procedure is
difficult at best and impossible in most cases. The validity of
this data correction was again assessed by the EPA Quality Assurance
Branch. Their finding was:
The basic question is—How does one make bad data
good? Whatever is tried will be attacked for a multitude
of (justifiable) reasons. Using the existing data set for
relative pollution level assessment will be acceptable,
but statements concerning absolute levels will not be. It
would not be wise to submit these data to the NADB, but
rather answer all requests for these data internally.
Their statement gives a reasonable assessment of the CHESS
sulfate data between 1972 and 1974. The assessment of other year
CHESS sulfate data is more difficult. No comparative sulfate data
exists from the local agencies as it did for S02 and TSP. Based on
the intrinsic capabilities of the methods, and the error assessment
of the field use procedures, it can generally be stated that:
1. From 1970 to September 1971 the sulfate data were
obtained using the turbidimetric method. It should be used
only as a sulfate level indicator. Due to interferences, there
will be severe problems if an attempt is made to correlate
sulfate levels in one part of the country with sulfate levels
in another.
2. From October 1971 until October 1972, the data are
subject to the following considerations:
a. The data are likely biased 'in the positive direction
from 1-2 ug/m3. This bias may be more severe in areas of
high S02 concentration relative to sulfate.
3-40
-------
b. The random error component of the measurement is
probably in the order of ±25% at an atmospheric concentra-
tion of 10 ug/m3.
3. From October 1972 until June 1974, all CHESS sulfate
data were biased negatively by approximately 50% on an annual
average basis due to improper laboratory analysis by the
contractor. These data should be used only on an adjusted
annual average basis to establish local trends within site
locations. The unknown cause of the bias prohibits use of
the data in shorter time structure (i.e., day, week, month)
increments.
4. From July 1974 until July 1975, CHESS sulfate data
underwent a marked improvement and was somewhat better than
that collected in the 1971-1972 era. The positive bias of the
data is probably similar to that of the earlier period but the
random error component was improved due to improved sulfate
blanks on the TSP filters.
3.3.2.2 Summary—The above discussion of the findings of the 1976 Congressional
Investigative Report represents a fairly complete cataloguing of the types of
quality control problems and resulting errors generic to EPA CHESS Program
sulfur oxides measurements, and it likely applies generally to all such data
collected as part of the Program during particular time periods. Certain
other errors, of varying relative magnitude, were detected and discussed in
the IR in connection with specific CHESS studies reported in the 1974 "CHESS
Monograph." Some of these are alluded to in discussions of such studies contained
in Chapter 14 of this document. The reader is also directed to the 1976 IR
for more information pertaining to errors in S02 measurements for specific
CHESS studies.
In general, it can be seen from the above that errors in measurement for
S02 under the CHESS program likely included some that would lead to a consistent
underestimation of actual atmospheric S02 levels by 30 to 40% in some instances
and up to around 100% under other circumstances. Fortunately, other local or
state aerometry measurements, often obtained from monitoring sites side by
3-41
-------
side with CHESS sites and judged to be more accurate and valid than the CHESS
measurements, were at times available and provide very useful "benchmarks" by
which to define the most likely outer bounds of CHESS measurement errors.
3.3.3 Summary of Assessment of Sulfur Oxides Measurement Applications
Tables 3-1 and 3-2 respectively summarize the above information pertaining
to the sources, magnitudes and directional biases of errors associated with
British and American SOp measurements over the past 30 to 40 years. Versions
of the lead dioxide "sulfation rate" technique were used in both countries
before the 1960s, with the lead dioxide gauge (candle) being employed widely
(&
in both countries and sulfation plate methods -ape well, especially in the
United States. As shown in both Tables 3-1 and 3-2, however, several problems
(e.g., temperature and humidity effects) with applications of the lead dioxide
method result in their being essentially useless for 24 hr measurements and in
their having rather large ± error bands associated with readings taken over
longer periods of time, i.e., up to 30 days. Thus the main value of the
sulfation rate data, in its time, was to provide at least a rough index of
sulfur dioxide (or SOp) by which to gauge large increases in sulfur compound
air pollution that came to be recognized as being associated with mortality or
severe morbidity effects.
In the early 1960s, the British adopted the hydrogen peroxide method as
the standard method to be employed for sulfur dioxide measurements as part of
the United Kingdom National Air Pollution Survey. That method was selected
because its sensitivity, reliability and precision were demonstrated to be
much better than that obtained in comparison to the lead dioxide method. More
specifically, the British Standard for sulfur dioxide determination by the
hydrogen peroxide method states that replicate determinations can be expected
3-42
-------
TABLE 3-1. SUMMARY OF EVALUATION OF SOURCES. MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH BRITISH SO, NEASUREHENTS
TIM
period
Measurement
method
Reported source
of error
Direction and magnitude of
reported error
Likely general Impact on
Pre-1961 Lead Oioxldm Humidity (RH)
Temperature (T)
Wind ipeed (WS)
1961-1980 Hydrogen
(British National Peroxide
Air Pol. Survey)
CO
(Overall errors)
Siting of Sample Line
Intake:
a. too near boiler chimneys
b. too near vegetation
Sample line Adsorption:
a. Good care & cleaning
b. Average care
c. Poor care (insects, dirt)
Flow Meter Problems:
a. Dally normal conditions
b. 8-port unit with only
one weekly flow reading
Allowable Filter Clamp
Leakage
Poor Clamp Care I Technique
Grade B Glassware Usage
Improper Alkalinity Buffering
CO, in Demlnerallzed in
H*0
Atmospheric Ammonia
Reaction rate Increases with RH.
Reaction rate Increases 2%
per 5° rise.
Reaction rate Increases
with WS.
SO - 100 ug/aT overestlmatlon.
50-70 percent underestimation.
10 ug/m underestimation.
20-25 ug/a low fro. SO ug/m .
Probable greater underestiaatlon.
t 3 percent variation.
± 5 percent variation.
1-2 percent underestimation.
5-10 percent underestimation.
2-5 ug/m .underestimation.
5-10 ug/m underestimation.
40 ug/m low from SO ug/m
monthly mean.
25 ug/m underestimation on 10X of
summer samples In urban areas.
2S ug/m neg. bias on 10% of summer
samples In urban areas.
Occasional neg. bias In country armat-
up to 80 vg/* daily data & up to 100%
monthly mean In summer.
Presumed t S MO/* precision of data.
Actual t 10 ug/m3 precision level.c
Added t S ug/m precision error.c
IS-100% DOS. bias for SO, data <100 MO/m3.
7.5-15% pos. hies for SO, of 10O-200 tig/m . -
3.25-7.5% pos. bias for $0. of 200-400 yg/m4.
<3.25% pos. bias for S0? dStt >400 |ig/» .
General 5% neg. bias In SO. data.
Occasional - 110% negative bias in 50? data.
"Data from 1%5-1968 most clearly Impacted.
bOata from 1966-1967 most clearly Impacted.
°At <50 ug/m3 uncertainty due to these two errors 1s - 7 ug/m3 or 14%. That is, 68% 6f" the data are within 14% and 5% are >28% in error.
-------
TABLE 3-2. SUMMARY OF EVALUATION OF SOURCES, MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH AMERICAN S02 MEASUREMENTS
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general impact on American S0_ data
1944-1968
Lead dioxide.
1969-1975
(EPA CHESS
PROGRAM)
West-Gaeke
Pararosanaline.
co
i
Humidity (RH).
Temperature (T).
Windspeed (WS).
Saturation of Reagent
(sulfation plate mainly).
(Overall Errors).
Spillage of reagent
during shipment.
Time delay for reagent-
SO- complex.
Concentration dependence
of sampling method.
Low flow correction.
Bubbler train leakage.
(Overall errors).
Reaction rate increases with RH.
Reaction rate increases 2% per 5
rise.
Reaction rate increases with WS.
Variable underestimation beyond
pt. where 15% of Pb03 on plate
reacted.
18% of total volume 50% of time;
occasional total loss
SO, losses of 1.0, 5, 25, and
75% at 20, 30, 40, and 50°C,
respectively.
Underestimation of unspecified
magnitude at daily SO, >200
ug/m3.
±10% to 50% variable error.
Small underestimation a*v of
unspecified magnitude.
Variable positive bias, especially in sunsner.
Variable positive bias, especially in summer.
Variable positive bias, especially in summer.
Possible large negative bias, especially for 30-
day samples for summer saonthly readings.
Generally wide ± error band associated with data.
Possible negative bias up to ^100%, mainly in
summer, with 30-day reading.
Half of S0_ data likely negatively biased by
mean of 17%; some up to 100%.
Usually SHall (<5%) negative bias, but consistent
negative, summer bias up to 25% at 40°C temp,
extreme.
Probable general negative bias, in daily,
monthly, and yearly S0_ data.
Usually error of < ±10%; occasionally up to
± 50% in daily, but dampened statistically in
annual mean.
Slight negative bias suspected.
From Nov., 1970, to Dec., 1971, data biased
low by 50-100%. Fro« Nov 1971, to
conclusion of CHESS Prograa in 1975, fall-
winter data appear valid but suwner data biased
low by maxiMURi of 60-80%. Fro* 1972 to 1975
annual average data approximately 15-20% low.
Daily data highly randotn, not useful.
November, 1970, to April, 1973, CHESS Program data impacted before error corrected.
bApplies to CHESS Program SO- data from all years 1970-1975.
cAs sumnarized by Congressional Investigative Report (IR, 1976).
-------
3 3
to be within ±20 |jg/m for concentrations up to 500 mg/m and within ±4 percent
3
for concentrations above 500 ug/m ; and an OECD Working Parting stated the
accuracy of the method to be ±10% at levels >100 |jg/m . However, as summarized
in Table 3-1, numerous sources of errors have been encountered in the practical
application of the method in collecting data for the British National Survey
over the past 15-20 years.
Certain of the sources of error listed in Table 3-1, it can be seen,
resulted in relatively small errors, whereas others produced errors ranging up
to 50-100% in magnitude. Also, some errors appear to have been restricted to
affecting data from only limited locations (usually unspecified as to specific
names of localities) or during only limited time periods. Many of these types
of errors appear to have been detected fairly quickly and steps taken to
successfully correct or minimize them. Still other sources of errors exist
(e.g., those from reagent evaporation), which have likely affected essentially
all British National Survey S0? data. Some of these appear to remain uncorrected
to this date, in some cases more than 10 or 15 years after they were first
detected and brought to the attention of Warren Spring Laboratory officials
responsible for overseeing quality control for the entire National Air Pollution
Survey.
Taking the above information into account for present purposes, it would
be extremely difficult to determine precisely which errors affected particular
National Survey data sets employed in British epidemiology and other studies
discussed later in this document. That would likely require a thorough examina-
tion, on a time- and site-specific basis, of records detailing information on
how each pertinent data set was collected and the full WSL quality control
assessment reports for each data set. Alternatively, in later evaluations of
3-45
-------
British epidemiology studies one could accept the following overall evaluation
and set on conclusions by the WSL (1975) regarding British National Survey air
pollution data (emphases added):
The actual degree of accuracy attained in the Survey is not known.
Input data are scrutinized by WSL staff, and subjected to computer
checks, and any reflectances, titres, or air flows which are abnor-
mally high or low or show unusually abrupt changes from one day to
the next are queried and data known to be invalid are excluded from
the annual summary tables. Such checks can however eliminate only
some of the gross errors. More information will become available on
accuracy when current (1974) plans to institute additional quality
control, e.g., on reagent solutions, are put into operation.
However, although the accuracy of the Survey data cannot at present
be quantified, many of the errors discussed in the previous para-
graphs will cancel out when data are averaged over periods of a few
months or a year, or for groups of sites. The remainder tend to
show up as anomalies when data are compared with past or subsequent
data at the same site or with data from other sites; anomalies of
this kind have been commented upon throughout the Reports. Members
of Warren Spring Laboratory staff have devoted a large effort over
the years to site visiting and checking on procedures. It is their
experience that the vast majority of the instruments are maintained
and operated with reasonable care and accuracy. The Laboratory is
therefore confident that the accuracy is sufficient for the type of
data analyses carried out in the present series of reports.
Presumably, it is the opinion of the WSL and British epidemiologists that the
accuracy of the survey data is also sufficient to meet the original objectives
of the Survey, ie. to assess the benefits accruing from the Clean Air Act of
1956, which requires use of the survey air quality data along with community
health endpoint evaluations in order to define quanititative air pollution/health
effects relationships. This presumption is supported by the long history of
reliance on these data by British epidemiologists, such as in the making of
statements regarding such quantitative relationships in innumerable journal
articles and reviews appearing during the past twenty years, up to and
including the very recent review by Holland et al. (1979).
3-46
-------
Turning to American applications of S02 measurements since the widespread
abandonment of sulfur dioxide sulfation rate methods in the mid to late 1960s,
one finds summarized in Table 3-2 the several different types of errors identified
as being associated with EPA CHESS Program SCL measurements used as part of
the single largest and most comprehensive community health survey of air
pollution effects conducted in the United States. The types of errors listed
in Table 3-2 were those detected via a thorough evaluation of the CHESS Program
by a Congressional Investigative Committee, as reported in the IR (1976). As
can be seen, the magnitudes of some errors in CHESS SCL measurements spanned
about the same range as those seen for British National Survey S0? measurements
and, at times, derived from analogous sources of error, e.g., evaporation or
other loss or reagents. In the case of the American CHESS Program data,
however, the specific overall impact of the various detected errors on particular
CHESS data sets appears to have been much better defined due to the work of
the IR (1976) which constitutes in one publication a rather comprehensive
review of measurement errors. More specifically, it appears that the CHESS
data generally tended to be somewhat negatively biased in comparison to other
local or state SCL data from monitoring sites proximal to the CHESS sites,
with the local and state data judged by the IR (1976) to be reasonably accurate
and reliable. The specific magnitude of the negative bias for particular
years of CHESS data is summarized in Table 3-2, and more information on the
point can be found in Appendix A of Chapter 14 of this document.
3.4 CRITICAL ASSESSMENT OF PARTICULATE MATTER MEASUREMENT APPLICATIONS
This section presents a critical assessment of principal measurement
approaches used for determination of air concentration of particulate matter
in Britain and the United States. These measurement methods involve: weighing
of filters; measurement of air flow through filters; measurement of the time
3-47
-------
of collection; and measurements of temperature and pressure. In addition, the
British method requires a measurement of reflectance that is converted to a
o
surface loading on the filter (ug/cm ), which in turn is converted into a mass
concentration (ug/m ).
In preparing this critical assessment, the "official" instructions for
respective measurements are discussed as a starting point. In addition,
whenever a reference was cited to support a computation in the official
procedure, the report referenced was obtained, examined, and critiqued. A
literature search was then performed for comparison studies and methods
evaluations which discussed accuracy, precision, and comparability between
methods. This assessment, therefore, represents a synthesis of many studies
over many years. With the accuracy of hindsight it presents an overall picture
of two methods being developed along parallel lines to accomplish similar
purposes. It is not surprising that many of the same problems arose along
each path and common misconceptions may have occurred.
The in-depth analysis contained in this section is provided to address,
and answer where possible, questions raised about the accuracy and precision
of particulate matter measurements employed as aerometry data in the key
quantitative community health studies performed in the United States and Great
Britian, as discussed in Chapter 14 of this document.
3.4.1 British Approaches
3.4.1.1 Daily Smoke Measurements of the United Kingdom National Survey—The
general technique for the British Smoke shade (BS) measurement was described
previously in Chapter 2. This section provides a critical review of the
measurement procedure so that one may obtain an appreciation for the precision,
accuracy, and reliability of the measurements. The details of the BS measurements
are provided by the Instruction Manual (Warren Spring Laboratory, 1966),
3-48
-------
hereafter referred to as IM. At the start of the National Survey in 1961
(WSL, 1961) it was recognized "The daily instrument, while comparatively
simple in design and operation gives reliable results i_n good hands* and
seemed the best choice for the National Survey." WSL circulated the specifica-
tions of the apparatus and methods to all the cooperating organizations as
careful, uniform work was essential if the results from the different sites
throughout the country were to be comparable. However, WSL found that detailed
instructions were necessary as most of the Local Authority staff making the
measurements had no training in analytical techniques. These methods were
reviewed by an O.E.C.D. Working Party and a report "Methods of Measuring Air
Pollution" (OECD, 1964) was prepared, which was accepted into the British
Standards Specification 1747, Parts 2 and 3. The Manual of Instruction (WSL,
1966) incorporated the improvements in techniques, "but apparatus and procedures
are specified in much greater detail to assist operation by observers with ru)
technical knowledge."*
3.4.1.1.1 Sources of imprecision in measurement: physical layout. The question
of the effect of sampling tube length was addressed by Moulds (1962). The
results were reported for a long length (35 feet) and a small length (7 feet)
3
of polyvinyl tubing. The mean concentration for the long tube was 32 ug/m
3
and the mean for the short length was 36 ug/m . Moulds (1962) states "these
differences are not significant", but does not mention if this conclusion was
3
from a statistical test or not. Because the levels sampled (36 ug/m ) are
relatively low, the question of whether the loss is 5 ug/m or 10% is not
addressed, nor is it clear that the loss could not be greater at higher BS
concentrations when the flowrate falls due to mass collection on the filter.
^Underline added for present emphasis.
3-49
-------
It is now recognized (McFarland, 1979), however, that significant gravita-
tional settling of participate matter can occur in a long run of tubing when
laminar flow is occurring. The larger particles, which may be suspended in a
turbulent ambient atmosphere by vertical motions, will be affected by gravity
and even thermal gradients in the inlet tube, if the top of the tubing is
differentially heated by solar radiation or differentially cooled by radiation
to the sky (Katz, 1958). Since the recommended PVC tubing has an internal
diameter of approximately 6.5 mm, and the recommended flowrate is 1.5 liter/min
(25 cc/sec), the linear velocity in the tube is approximately 0.75 m/sec.
Because the IM (1966) originally allowed lengths of tubing up to 40 ft (12.2
m), the time to reach the filter in some instances may have been as high as
16.3 seconds. Consequently, with a 12.2 m inlet line, no particle with a
terminal velocity greater than 0.65 cm/16.3 sec = 0.04 cm/sec could reach the
filter. In addition, particle losses can be caused by sharp bends in the
tubing.
As stated in the IM (1966):
The distance from the funnel to the filter clamps should be
kept as short as possible but there is no critical length under
about 40 ft (12 m). It is important that the apparatus should be
protected from sunlight and from undue heating by radiators or
boilers; sunlight and warmth accelerate the decomposition of hydrogen
peroxide, and excessive heating leads to evaporation of the liquid
in the bubbler.
The run of the connecting tubing should be kept as straight as
possible without kinks or dependent loops. Where the connecting
tubing has to change direction it is essential to keep the radius of
the bend at least 1 in (2.5 cm) and preferably 2 in (5 cm) or more.
A sharp bend causes marked turbulence in the air stream and consequent
deposition of some smoke particles on the walls of the tube.
Because the air is still in turbulent flow for approximately 40 tube
diameters (26 cm) before laminar flow is fully developed, there probably will
3-50
-------
be secondary flow patterns at the 90° bend over the funnel inlet. The particles
may then be preferentially deposited on the glass surface of the bend. An
electrostatic charge on the PVC tubing is another possible source of uncertainty
since this could also cause particles to settle out faster than computed by
Stokes' Law (B.S. 1747, Part 2, 1964). Because there is no uniformity
required by the IM (1966) for either tube lengths or number and size of bends,
uncertainty between measurements must be considered a distinct possiblity.
Leakage—The smoke filter unit consists of two heavy brass cylinders called
filter clamps, drilled out to provide a fixed area for the exposed filter
paper. The seal of the BS filter by the filter clamps is a major source of
possible error. According to WSL (1961): "At one time weights were used to
hold the clamps together, but so many failures occurred that a screw-down
clamp holder is now regarded essential."
As stated by the IM (1966):
By the nature of their construction, filter clamps and 8-port
valves are prone to leakage which cannot be detected without a special
test. The resistance of the filter paper and bubbler to the airflow
means that the side of the clamp assembly or valve connected to the
pump is under a slight vacuum, and a badly seating clamp or dirty
valve will allow air to enter. Leakage can also occur edgewise through
the filter paper if inadequate pressure is applied to the clamps.
If such an inward leak occurs, the amount of air registered by
the gas meter includes air from the room where the apparatus is kept
and which has not passed through the filter paper in the normal way.
This should not be permitted to exceed the limits quoted below.
Those limits for leakage are 1 to 2 percent, depending on clamp size.
This leakage problem was also observed with the early American Iron and
Steel Institute (AISI) CoH instrument when the clamps were out of alignment by
as little as one-thousandth of an inch, according to a 1962 WSL report (WSL
AP/70, 1962). In the early development of the WSL automatic sequential
3-51
-------
samplers, it was discovered that a gross irreproducibility of measurements
with models similar to the AISI instruments was almost certainly caused by
variable leakages at the filter clamps (WSL AP/70, 1962). Thus, first
generation (1959-1961) WSL automatic sampling models were declared unsuitable
for an early survey in Sheffield until the reproducibility could be improved.
For example, one pair of readings of co-located samplers during a test gave
282 and 1065 pg/m3.
In order to prevent a significant error due to leakage, the IM (1966)
provided a section entitled Measurement of Leakage and instructions as to how
to determine that the leakage is in the acceptable ranges of <1 percent for a
1/2-inch or 1-inch filter clamp or <2 percent for a 2-inch or 4-inch filter
clamp. The leakage rate data, however, are apparently not used to make a
correction to the final recorded value of BS, which would be raised by 1 to 2
percent.
Flow measurements—The Whatman No. 1 filter paper used in collecting samples
for/) smoke measurements is quite porous to submicron particles when it is clean
(Liu et al., 1978). As the pores begin to fill up, however, the "window"
closes and the efficiency of the filter increases. The IM (1966) warns against
using a 1/2-inch clamp except in clean areas.
The use of 1/2-in clamps should be confined to extremely clean
areas in open country where 1-in clamps regularly give reflectometer
readings of 90 and over. The main disadvantage of using 1/2-inch
clamps is that the smoke deposit builds up very quickly, soon causing
the reflectometer reading to fall below 40. At the same time the
resistance to the flow of air rises and the flow rate at the end of
the sampling period may be only one-third of that at the beginning,
when the filter paper was clean.
3-52
-------
Another, potentially more serious error is described in WSL (1975, Volume 5)
In the case of 8-port instruments, for which gas meters are read
only once a week, concentrations are calculated on the assumption
that the volume of air sampled was the same every day. When the
throughput of air is less than average on days of high particulate
pollution, because of back pressure at the filter, concentrations are
underestimated, and they are overestimated on the other days of the
week. (For the same reason diurnal variations in back pressure may
lead to under-estimation of pollution when some hours of high pollution
follow hours of low pollution.) This effect is kept to a minimum if
the clamp is sufficiently large, having regard to season and type of
area, and it is unlikely to be significant now that smoke concentrations
are much less than they used to be, except occasionally at sites with
1/2-inch clamps.
This was recognized in the IM (1966) where it is clearly stated:
In principle, the 8-port valve demands a constant air flowrate
through the system, since meter readings are normally taken at intervals
of several days. Obviously if one of a number of stains obtained
between two meter readings is dark, the airflow for that day will be
below the calculated average, and the calculated concentration, based
on the average daily flow over the whole period, will be too low.
To correct for this problem, the IM (1966) states:
Charles Austen (Pumps) Ltd have developed a constant flow pump
capable of being set to work at a constant rate of 1.5 1/min (approximately
3 ft /h) over a wide range of pressure variations in the system. It
is preferable to use this pump in place of the Dymax pump and air
pollution meter.
This recommendation was not made a requirement by WSL, and it is quite
possible that local authority personnel with no training in analytical techni-
ques or sufficient technical knowledge ignored it. Re-examination of pertinent
records would be required to ascertain which National Survey data may have
been affected by flow measurement errors.
3.4.1.1.2 Reflectance measurements. The WSL quality control report (1977b)
states:
3-53
-------
Tests have shown that two WSL operators using different reflectometers
to measure the darkness of a set of smoke stains differ on average
by less than 1 unit of reflectance, the standard deviation being
about 1 unit. Differences of 3 or 4 reflectance units in readings
of individual stains occur occasionally. Co-operating bodies are
therefore contacted regarding their smoke stain assessment if their
reading of an individual stain differs by more than 4 reflectance
units (about 20% of the concentration) from the WSL reading, or if
the mean difference for 4 readings exceeds 3 reflectance units
(about 15% of the concentration). Over 15 percent of the co-operating
bodies involved in the first series of tests, and 15 percent of
those in the second series, were contacted on the basis of these
criteria (22 bodies altogether).
The absolute error in (jg/m associated with this reading difference increases
in magnitude with reduction of the reflectance value. We can derive the error
rate by differentiating the equation for smoke concentration with respect to
reflectance. The results show that the error associated with all unit
reflectance reading is: ± 10% at R = 90; ±7% at R = 80; falling to ±4% at
R = 40. The smoke concentration formulae used for converting reflectance
readings are reported in the IM (1966) as:
Csm = V (91,679.22 - 3,332.0460 R + 49.618884 R2 - 0.35329778 R3
+ 0.0009863435 R4)
where ,
C is the concentration of smoke (ug/m ),
V is the volume of air (ft ),
R is the reflectometer reading, and
F is a factor which depends on the size of clamp used
(1/2-in clamp, F=0.288; 1-in clamp, F=1.00; 2-in clamp,
F=3.68; 4-in clamp, F=12.80)
For darker stains with reflectometer readings below 40, an approximation
of the smoke concentration can be derived from the following formula:
Csm = y (214,245.1 - 15,130.512 R + 508.181 R2 - 8.831144 R3
+ 0.0628057 R4)
3-54
-------
The extensive usage of significant figures in these formulas may lead a
casual reader to believe that they are highly accurate and represent a scienti-
fically valid formulation. In fact, these 5 factors were chosen to fit 5 obser-
vations exactly, which is an unusual procedure. The polynomial fit can lead to
strange behavior, such as the inflection point in the first formula at R = 43.
The factors {F} above deserve particular scrutiny. They arise from the
OECD (1965) report referenced in the IM (1966). The pertinent OECD (1965)
report section states verbatim:
The surface concentration of the smoke in terms of "international
standard" smoke is first obtained from the reflection factor by
reference to the calibration curve appropriate to the filter paper
and reflectometer used. Curves for Whatman No. 1, Schneider CA 32
and fibreglass GF/A filter papers, and Eel and Photovolt reflectometers
are given in Figures 2 to 6.
A discussion of the basis, derivation, and use of these curves is given in
Part 3 of the OECD (1965) report, as stated below:
Let S be the deriued surface concentration of the smoke on the
filter paper, in ug/cm , ?
A be the area of the appropriate filter stain, in cm ,
V be the volume of air sampled, in m , 3
C be the concentration of the smoke in the air, in ug/m ,
then C = SA/V
The following formulae should be used for 2-in and 4-in diameter
stains on Whatman No. 1 filter papers to allow for the lower linear
velocity of the air as it meets the filter paper in these filters
causing less penetration of the smoke particles into the pores of
the paper. This leaves more of the smoke on the surface so that a
given amount produces a darker stain.
C = 0.92 SA/V for a 2-in filter
C = 0.80 SA/V for a 4-in filter.
The factor "F" referred to in the IM (1966) appears to have been derived
from the above OECD (1965) formulae and the IM specifically lists the following
values for F as a function of filter clamp size:
3-55
-------
OECD correction values
0.80
0.92
1.00
?
Filter clamp size
4-inch
2-inch
1-inch
1/2- inch
IM "F" values
12.8
3.68
1.00
0.288
The unreported correction factor from OECD for the 1/2-inch filter clamp
is apparently interpreted by WSL to be 1.15, but its derivation is unknown and
is unreferenced in the IM (1966). This appears to be identical to 0.92/0.80 =
1.15, as might be derived from the OECD (1965) report, but it is quite different
from 1.00/0.92 = 1.09, which could also be reasonably interpreted from the
OECD (1965) figures.
Thus, the 1/2-inch clamp factor of 1.15 is apparently required by WSL in
their 1966 IM to be used to account for particle deposition deeper into the
filter at higher velocities. They therefore assume that the 1/2-inch filter clamp
collects "less" on the surface than the 1-inch filter clamp, and the other
factors are less than 1.0 because they collect more on the surface than the
1-inch BS filter. This is not easily reconciled with the results of Clayton (WSL,
1977), and Liu et al. (1978) who both report a decrease in collection efficiency
at lower face velocities followed by increasing collection efficiency with
increasing face velocities for the 0.32 to 1.0 pm particles at all velocities
tested. That would imply that the particles penetrate through the filter more
readily at low velocities than at the higher velocities such that the particles
would not tend to be trapped on the surface.
3-56
-------
Moulds (1962) compared the penetration of the "smooth" and "rough" sides
of the BS Whatman No.l filter paper. She reported:
A composite filter was used to compare stains on alternative
sides of the paper so that one half of the collecting surface was
the "Rough" side, and the other half the "Smooth" side, with its
closer more regular surface. The stains were uniform within each
half and that on the "Rough" side was never darker than the on the
"Smooth" side, which may allow less penetration or present less
surface area for the deposit.
This statement appears to imply that the smooth side gave less penetration
and the stain was darker. The OECD (1965) report implied that in the case
where less penetration was caused by a higher velocity the stain would be
lighter.
In a further WSL report by Bailey and Nicholson LR 89 (AP) dated October
1968, a factor "F" is developed for a 10 mm filter clamp. The summary of the
report states:
The calibration curve of standard smoke quoted in B.S. 1747:
Part 2: 1964 is derived for a 2.54-cm filter holder. If the diameter
of the stain is reduced whilst the air flow is held substantially
constant the velocity of impact of the smoke particles on the filter
paper is increased causing deeper penetration; hence an increased
concentration per unit area is required to obtain a particular
darkness index. This effect has been examined for a stain of 1-cm
diameter and it was found that, over a range of smoke concentrations
from 5 to 20 ug/m , the 2.54-cm calibration could be used with a
multiplying factor of 1.30.
These authors ran some British High Volume Samples (B.S. 1747: Part 2) simultane-
ously with the standard smoke samples and reported that "The ratio of the concen-
tration by direct weighing to the concentration in terms of equivalent smoke varied
from 0.59 to 0.90 with an average value of 0.76." These values are shown in
Table 3-3 below.
3-57
-------
TABLE 3-3. CONCENTRATIONS BY DIRECT WEIGHING COMPARED WITH
CONCENTRATIONS OF STANDARD SMOKE - 10 mm CLAMP
-
Test Volume,
No. m3
4 19.74
5 19.20
6 19.18
7 19.74
8 20.45
9 8.77
10 9.03
Weight
mg
4.8
6.0
2.5
4.1
5.4
2.0
3.5
Concentration^
pg/m
by weight Standard Smoke
(a) (b)
243
312
130
208
264
228
387
285
410
220
319
312
253
621
Ratio
a/b
0.85
0.76
0.59
0.65
0.85
0.90
0.62
Thus, the standard smoke, which includes the 1.30 factor, is 1/0.76 or 30
percent higher than the concentration collected and weighed. This inconsistency
also appears in a discussion by Appling et al. (1977) of a difference between
a BS reading of 1000 pg/m3 at the MRC site and 850 jjg/m3 at the WSL site in
Endell street, Holburn, on December 15, 1975. Appling et al. state:
The fact that the hourly concentrations were higher at the MRC
site than at the Endell Street site could be wholly or partly due to
a slight difference in the method of measurement at the MRC site,
the instrument there having twice the air throughput rate of the
standard smoke instrument.
However, the reasoning given in the OECD (1964) report for the WSL
correction factors for various filter sizes is that the higher face velocity
cause the particles to penetrate deeper into the filter, thereby producing a
relatively lighter stain. It is not reported what factor, if any, was used at
the MRC (Medical Research Council) to correct for the higher velocity.
3.-58
-------
In light of the above information, an ambiguity presently appears to
exist in regard to the basis for the "F" correction factor in the IM (1966)
and the correction values in the OECD (1965) Manual for the 1/2, 2, and 4-inch
filter clamps. One possible explanation for the apparent inconsistencies may
be that the correction factors {F} are related to a certain computer error.
Figure 3-2 from WSL (1967) shows four different curves. As described in WSL,
(1967):
The National Survey data for 1961 to 1964 were, as described,
processed by computer and after some little time it was discovered that
the computer was not following the curve it was intended to, and that, in
fact, the smoke concentrations were being calculated from the lowest of
the calibration curves shown in Figure 3-2. This set of data has now been
re-calculated to the standard curve and re-issued.
Examination of Figure 3-3 from Moulds (1962), shows that the deviation of
the ratio of 1-inch filter to 2-inch filter concentrations diverges from unity
at a reflectance of 50, where the 1961 DSIR curve labeled C in Figure 3-2
(i.e., Figure 44 from WSL, 1967) diverges from the British Standard Curve (A).
j fl
l]LsuJt>t£**^y/
For example, at reflectance-order 90, the curve A gives about twice as much
mass as curve C, which we presume Moulds used. Consequently, the ratio would
be about 1.0 also. As quoted above, WSL went back and re-computed and re-issued
the 1961 to 1964 National Survey data. However, it appears that WSL may not
have gone back and re-computed all the 1961 to 1964 data, including their data
sets which were used to derive the factors {F} of 1.15, 0.92, and 0.80 (which
were also computed during this period). Of course, if the computer made these
calculations by curve D, the error would be even greater than the 50 percent
difference that Moulds reported.
In summary, the information presently available does not seem to support
the usage of the factors 1.30, 1.15, 0.92, and 0.80, which are apparently
3-59
-------
5QO
4OO
JOO
UJ
o
z
o
o
UJ
<
LL
K.
200
too
I
A - BRITISH STANDARD CURVE
8 - DSIR INTERIM CURVE
C - DSIR -CLARK - OWENS CURVE
D - 1961 TO 1964 NATIONAL SURVEY CURVE
O
/O 2O JO
5O 6O 70
DARKNESS INDEX
Figure 3-2. Comparison of smoke calibration curves for Eel reflectometer,
Whatman No. I paper and a 1-in. diameter filter. From WSL
(1967).
3-60
-------
RATIO OF CONCENTRATION
o r* -i
01 O (,
1 1 1 1 1 | 1 1
1-inch FILTER
2-inch FILTER
Qj"» _ " _^ xO
CO0' ° ^o" " °~
o °0e o° o «to
o
1 I I I I I I I
10 20 30 40 50 60 70 80 9(
REFLECTANCE, percent (DARKER STAIN)
Figure 3-3. Comparison of simultaneous BS measurements using 1-in. and 2-in.
filter clamps, showing that the reflectance ratio (1-in./2-in.)
decreases below approximately 45 percent reflectance of the 1-inch
filter clamp. From Moulds (1962).
3-61
-------
still in use up to this present day. This could have a substantial impact on
BS measurements if the use of the stated corrections cannot be supported in
$&M+silw)
view of the above physical observations by/|WSL (1968), WSL (1977), and Liu et
al. (1978).
3.4.1.1.3 Form of the Reflectance Curve. The International OECD expert curve
may also be improperly based. The reflectance R is the dependent variable,
and the concentration of particulate matter on the filter C must be treated
as the independent variable. Consequently, the regression relation might
better be R = f (C ) not C = f (R). As can be seen from inspection of the
two calibration equations, they can lead to an upper limit at R = 0 of [C =
| (214245.1)] or about 3000 ug/m3. However, tokEllison (196^5 reports that the
reflectance must approach a constant finite value equal to the reflectance of
the particle mixture. That is, C -> », R -> R^ i- 0. Along with the condition
C = 0, R = 100, there are also alternate forms for the equations. If variables
sm ' ' M
are changed to D = 100 - R, where D is the percent decrease in reflectance,
then D = f (C ). A simple equation that places D = 0 @ C =0 and D = D @
v sm' K -i v sm a,
C ->• oo i s:
sm
D = DM [1 - exp (acsm + p Cj)]
Tests for values of D , a, and p can be made to satisfy the curve and the
boundary conditions.
3.4.1.1.4 Averaging error. The IM (1966) states:
On the larger stains, it may sometimes be advisable to take
meter readings on different parts of the stain. If the readings are
found to vary slightly the mean value should be taken as representa-
tive.
The British Standards manual (1747, Part 2, 1964) also states: "Take a
reading at three different positions of the reflectometer head on the stain
and determine the mean."
3-62
-------
However, this procedure will lead to incorrect results. The readings
2
must be converted to (jg/cm and the average filter loading must be computed.
Because the reflectance curve is nonlinear the mean reflectance will lead to
an underestimate of the true filter loading.
3.4.1.1.5 Filter handling. One of the most important aspects of the BS smoke
measurement relates to the handling of the filter in the measurement system.
The Whatman #1 filter is not isotropic. There is a smooth side to the filter
which should face the direction of the airflow and a coarse side which faces
downstream. The IM (1966) procedure calls for the operator to identify the
smooth side thrice, in setting the filter into the clamp properly (smooth side
upstream), again when a clean filter reflectance is measured in order to
establish the 100 percent reflectance point and also when the stain is read.
However, the British Standard instructions (B.S. 1747, Part 2, 1964) only say:
"adjust the reading on a single sheet of clean filter paper" without reference
to smooth or coarse side. A comparison of reflectance of coarse side versus
smooth side shows the coarse side of Whatman #1 paper to be slightly darker
than the smooth side by less than one unit of R (personal communication, D. T.
Mage, 1980).
Failure to distinguish the smooth side from the coarse side in either or
both of these measurements results in an error, since the standard curve being
compared to is smooth filter-smooth filter, both ways. Moulds, (1962) demon-
strated this effect in the experiment described on Page 3-5/Pof this section
as: "Values of the quantity Concentration 'Smooth1/Concentration 'Rough'
expressed as a percentage varied between 100 and 122, with a mean of 106.8 and
a standard deviation of 6.15. An even larger error could exist if the filter
was read in the reflectometer incorrectly, i.e., not on the impact stain side
3-63
-------
but on the opposite side where a lowgrade stain appears from the particles
pulled into the filter paper.
A WSL report (WSL, Vol. 5, January 1975) states on page 113: "Stains have
occasionally been found to have been read the wrong way up, i.e., with the
exposed face of the paper downwards, and the darkness has then been seriously
underestimated." The IM (1966), upon close reading, does not actually say
place the filter sta:in side up. It says
"Replace the clean filter paper by a stained paper. Position
the stain centrally over the white tile and place the masking unit
over the stain, with the hole as near the centre as possible. Insert
the measuring head into the unit while keeping it in position and
record the reading directly on to Form WSL 216 against the date on
which the filter paper was removed from the apparatus. Repeat this
procedure for each stain."
Apparently WSL did not consider reading the stain backwards to be
a serious problem or did not take adequate steps to avert its occurrence,
because a later 1977 report (WSL, 1977) still included the following references
to such errors: "Since the reading of smoke stains on the wrong side of the
paper is the source of the largest errors in smoke assessment, more publicity
should be given to this hazard and how to avoid it." This type of error is a
clear matter of concern if accurate smoke readings are to be obtained, but it
is not possible to determine precisely what National Survey data may have been
affected by such error without a case-by-case examination of the original
records for various monitoring sites during specific time periods (For example,
one monitoring site operator is reported to have read the unstained side of BS
filters for three years).
3.4.1.1.6 Filter weighing procedure. Another question can be raised regarding
the procedure used for providing a weight of smoke in the calibration procedure
3-64
-------
outlined by BS 1474, Part 2, 1964. A control on the weighing is provided by
handling a second filter in an identical manner as the stained filter paper.
The manual states:
The increase in weight of the exposed paper is a measure of the weight of
the dry deposited material or smoke and any change in weight arising from the
handling procedure. A correction for the error due to handling is made by
adding or subtracting the difference between the initial and final weights of
the control paper.
This procedure is quite correct if the handling errors are positively
correlated. If the handling errors are random and independent of each other
then the variance of the error is increased. If, however, the errors are
negatively correlated, i.e., if it is more likely to lose particles from a
stained filter than make it heavier and if it is more likely to soil a clean
filter than make it cleaner, then this procedure could lead to a reduction
in the recorded weights.
3.3.1.1.7 Filter shipment losses. Another possible uncertainty may occur
when filters are shipped to WSL to be read for quality assurance. If the
filters lose particles in the mail, the filters will have a higher reflectance
when they are read at WSL. In the WSL report, "The National Survey of Smoke
and Sulphur Dioxide-Quality Control Tests on Analysis of Samples, October 1975
to February 1977, results of a study of filter readings by the field operators
and WSL are described. The operators sent their filters to WSL to be reread
and WSL reports "More operators under-estimated the reflectance than over-
estimated it." No statistical analysis, however, was reported to test the
hypothesis that no weight was lost in shipment.
3.4.1.1.8 Choice of clamp size. The WSL Instruction Manual (1966) leaves it
to the operators' judgement as to which clamp size to use in anticipation of
3-65
-------
varying pollution levels. For example: "When the stain produced with 1-inch
clamps is close to the 'danger level' for a 24-hour period, it is obvious that
a darker stain will result from a 2- or 3-day exposure over the weekend. In
such cases, a change should be made to 2-inch clamps for the weekend." However,
the IM states: "The daily servicing of the instrument is a simple routine
task; it is often found that a caretaker or other person on duty can undertake
this servicing, leaving the exposed filter paper and hydrogen peroxide to be
collected at a convenient time." However, if the caretaker is not a technically
trained person, then the "danger level" may not be noticed. Thus, the retro-
spective analysis of filters may not provide the feedback necessary to anticipate
sudden increases in pollution.
This situation may have actually arisen during a December, 1975, episode
in London when an excess mortality of 100-200 persons occurred (Holland, et
al., 1979). As described in WSL, LR 263(AP) the period December 14-17, 1975,
produced the highest smoke pollution levels seen since 1967. Because a surface
inversion formed on Sunday the 14th, the network operators had no notice that
an episode was occurring until they came to work on Monday the 15th.
In greater London, during December 1975, 111 monitoring sites were avail-
able for measurement of smoke stain (BS). On Friday, December 12, 1975, only
46 sites were equipped with operational 8-port instruments capable of giving
24-hour samples automatically and the other 65 sites had instruments which
either required daily servicing or they had inoperational 8-port instruments.
The Instruction Manual (1966) allows for manual weekend operations by
providing direction to the operators to use 4-inch diameter clamps for a
72-hour period to prevent "too dark a stain" from being produced by a 2-inch
clamp.
3-66
-------
However, the stations without the 8-port instruments apparently were not
run over the 72-hour period between 11 a.m. Friday, December 12, through 11
a.m. Monday, December 15./^wSL, LR 263 AP makes no mention that these 72-hour
data were available. Thus, 3 days of data were likely lost because of the
weekend at each of the manual sampling sites. Unfortunately, this complicates
the data analysis. For example, during the period 11:00 Sunday, December 14,
to 11:00 Monday, December 15, 1975, the maximum observation out of 46 stations
reporting was BS 696 (jg/m . On the following day, 11:00 Monday, December 15
to 11:00 Tuesday, December 16, 1975, the maximum observation out of 104
3
stations reporting was BS 811 M9/m • If all 111 stations had operated Sunday
to Monday, instead of only 46 stations, then the missing 65 observations might
3
have included an actual maximum reading greater than BS 811 pg/m .
The episode report also makes no mention as to whether any of the 46
8-port instruments which operated over the weekend were not equipped with a
constant flow pump. As described in the IM, lack of such an instrument would
probably lead to underestimated values, because the flow on the episode Sunday
would not be one-seventh of the total weekly flow.
In addition, the WSL report LR 263 (AP) makes no mention whether all the
operators made the necessary changes from 2-inch clamps to 4-inch clamps on
Monday December 15, 1975; thus, it seems possible that the same 2-inch clamps
were used through the weekend as during the immediately preceding and follow-
ing weekdays. If this is so, then standard procedures for changing clamp
sizes would have been violated. More specifically, the WSL IM (1966) states:
In an extremely severe episode of 'smog', when accurate knowledge
of peak levels is desirable, the filter paper should be changed and
the meter read at suitable intervals during the day.
3-67
-------
Further indications that these instructions were not followed at all the
stations during the episode can be derived from the Apling et al. (WSL, 1977)
report. That is , in that report, hourly peak values are only reported for a
few stations for the 24-hour period from Monday December 15 to Tuesday December
16, 1975. Examination of the report (WSL, 1977) further reveals that most of
the stations in the borough of Kensington-and-Chelsea were reporting signifi-
cantly lower BS values than the neighboring borough stations, some less than 1
km away. Since the wind was reported as ~1.5 m/s, the air-mass flight time
between the Kensington-and-Chelsea stations and the neighboring stations was
only 10 minutes, which would lead one to expect similar values between these
upwind and downwind stations. Apling et al. (WSL, 1977) state that this
difference may be produced by differences in distance to the street, but they
make no special analysis of the "cold spot" in Kensington-and-Chelsea.
3.4.1.2 Summary—Similarly to what was indicated earlier for the SCL measure-
ments, the assumptions and errors discussed here for National Survey BS
measurements may be conservative because in many cases "staff making the
measurements had no training in analytical techniques" (IM, 1966) and may not
have followed closely the instructions in the IM (1966). Even if they did,
however, given the nature of certain ambiguities concerning the origin and
derivation of the recommended filter clamp correction factors, substantial
errors may still have been made for that portion of the National Survey BS
data base not obtained by the use of 1-inch filter clamps. Other errors
mentioned above, such as, the "computer error of 1961-64" can be more definitely
identified as likely affecting certain British BS data sets obtained during
specific periods of time. Although this "computer erpor" was reported (WSL,
1967) as being corrected in the National Survey data base at an unspecified
3-68
-------
time, it is neither clear nor likely that all publications in the peer review
literature which employed these incorrect data have been corrected.
3.4.2 American Approaches
As discussed earlier, the hi-volume TSP sampler, since its development in
the early 1950s, has been the instrument most commonly used in United States
for measurement of atmospheric particulate matter; and hi-vol TSP readings
have most typically been used in epidemiologic evaluation of associated air
pollution effects relationships. In contrast, other particulate matter measure-
ment approaches (e.g., the coefficient of haze method) saw only relatively
limited application during the 1950s and early 1960s in certain American
locations and were infrequently used in estimating quantitative relationships
between airborne particulate matter and health or welfare effects. Accordingly,
major emphasis is placed below on the critical appraisal of certain key applica-
tions of hi-volume TSP measurements in the United States; less extensive
attention is accorded to evaluation of other particulate measurement appli-
cations in the United States.
As before, in discussing American applications for measurement of oxides
of sulfur, the present discussion will focus most heavily on evaluation of
applications of hi-volume TSP measurement methods that occurred as part of the
EPA "CHESS Program" as the single most extensive and comprehensive use of such
methods as part of American community health epidemiology studies. Some
consideration of local or state monitoring applications of pertinent hi-volume
TSP measurements is also included as part of the discussion of CHESS TSP
monitoring. Much of the information provided is derived from the 1976
Congressional Investigative Report (IR) which included thorough analysis of
EPA CHESS Program TSP measurements and comments regarding certain local or
state TSP measurements.
3-69
-------
3.4.2.1 Hi-Volume TSP Measurements of the U.S. EPA CHESS Program—Restating
earlier background information, the 1976 Congressionsl.Investigative Report
(IR) noted that the attainment of precise, reliable, reproducible, and real
time air quality measurements for particulate matter and other pollutants was
a key objective of the CHESS Program. Also, as stated in the IR (1976), the
air measurement methods selected for use in the CHESS program to meet that
objective were probably as good as any available at the time; and attempts
were made to introduce and implement better quality control procedures for air
monitoring into the CHESS Program as it progressed. However, quality control
problems with air quality measurements, especially in the earlier years of
the Program, did occur and led to controversy and critisms of the type dis-
cussed in the IR (1976). Most of the more salient points raised by the IR
concerning CHESS TSP measurements are discussed below.
As briefly described in Section IV B 2 of the 1976 IR for the CHESS
Program:
Total suspended particulates (TSP) were measured using the EPA
Reference Method as specified in the Federal Register (36 (84): 8191-8194,
April 30, 1971T).
Total suspended particulates (TSP) were measured by drawing air
through a preweighed 8 x 10 inch glass fiber filter for a period of 24
hours. The apparatus used for this procedure was the standard High
Volume Sampler. At the end of the 24 hour time period, the filter was
reweighed, and the TSP computed on the,basis of total air flow. The air
flow rate was approximately 60 ft min at the start, and must be not
less than 40 ft min at the end for the measurement to be acceptable.
The average air flow rate was computed on the basis of a straightline
interpolation between beginning and ending flow rates.
Comments concerning the validity of the method as a laboratory procedure
included the following from Section IV C 3 of the IR (1976):
3-70
-------
The Reference Method for the determination of total suspended
particulate matter (TSP) is probably the simplest and most reliable
method used by CHESS. It has been well studied and most error
sources are known. However, it is a method that measures an arbitrary
and poorly defined portion of the total atmospheric particulate
burden and the portion measured has unknown relevance to the human
respirable portion. The size fraction measured is somewhat dependent
on the design of the shelter used for Hi-Volume sampler. The design
and dimensions of the Reference Method shelter are specified in the
Federal Register, thus the portion of TSP that is collected by the
method is generally uniform. Best estimates of particle size range
included in the Reference Method are from 0.05 to 60 urn diameter.
Above 60 urn diameter, the particle fall velocity is too great to
navigate the bend around the roof of the shelter. Below 0.05 urn the
collection efficiency of the glass fiber used in the method diminishes.
A collaborative study was conducted on the Reference Method
using 12 different groups sampling ambient air at a common location.
The results of this study indicate the method is capable of reproducible
measurements with less than 5 percent error at the 95 percent confidence
level. Also, the minimum detectable amount of TSP is approximately
2 ug/m3 for a 24-hour sampling period. This sensitivity is more
than sufficient for most 24-hour TSP measurements.
Section IV C 3 of the 1976 Investigative Report further noted the following
regarding sources of error in CHESS TSP measurements and their likely impact:
The TSP measurement method, as used in CHESS, had one notable
difference from the laboratory procedure which was collaboratively
studied. The weighing procedure to determine TSP was performed at
EPA/RTP laboratory not by the CHESS contractors on site. This
necessitated the shipment of individual filter samples through the
mail and the subsequent storages of the samples at EPA. During
laboratory reorganizations at RTP, periods as long as 6 months
elapsed between actual field sampling and laboratory analysis.
The following is a summary of individual errors and an assessment
of overall TSP data quality.
Loss of particulate matter before weighing
V
In the TSP methodology there were field-related procedures that
resulted in partial loss of particulate matter from the Hi-Volume
filter samples. Due to the exposed location of the Hi-Vol TSP samplers,
wind and cold sometimes made it very difficult to remove the filter
paper from the apparatus without losing part of the sample. No
estimate has been made of loss due to this problem; it would, of
course bias the reported results only in the direction of lower-than-actual
atmospheric loadings. This was not a constant problem among CHESS
sites. It was noted by field operators as being a particularly
severe problem in the Salt Lake City area during the winter months.
3-71
-------
Two other error sources have been identified in the determination
of TSP, both of which would also produce a low-side bias: (1) the
shaking-off of particles from the filter during transit from the
field site to EPA/RTP, and (2) the evaporation of organic substances.
In an attempt to quantify the mass loss during traasit, David Hinton,
EPA/RTP, made a comparison of filters collected in Utah, before and
after mailing from Salt Lake City to RTP (22). He found that there
was an average 4 percent loss. Carl Broadhead, of the Utah Division
of Health, conducted a similar comparison; however, he noted an
apparent loss of approximately 25%. This difference may, in part,
be due to the time of year the studies were conducted. During the
dry summer months in the Salt Lake City area, much of the TSP loading
is due to windborn crustal material (sand). This material is much
more easily lost in sample handling that is the finer anthropogenic
particulate material.
A final error source, one more difficult to assess, derives
from wind velocity versus collection efficiency. On days with
relatively high wind (>15 mph), the HiVol sampler is more susceptible
to the inclusion of large diameter particulate material. To compound
this problem, the design of the shelter makes the magnitude of the
error dependent on the wind direction relative to the orientation of
the shelter. The main result of this problem is that two side by
side HiVol samplers, oriented 90 degrees relative to each other,
will produce dissimilar measurements with the discrepancy increasing
as the daily wind velocity increases.
The overall effect of the summed errors with the HiVol TSP
measurement is a slight negative bias. This bias may be as small as
10% or may be as large as 30%. Side by side data from New York and
Salt Lake indicate that this assessment is reasonable. These data
also indicate that the TSP data were by far the best quality data
taken in the CHESS monitoring program. Differences measured between
High and Low sites are probably reasonable estimates of the differences
of TSP exposures as received by populations within these areas.
Some local source variations undoubtedly did occur, but average
annual exposures were reasonable.
In any overall assessment of the CHESS TSP data it should be
noted that all of the sources of errors mentioned previously related
almost exclusively to the loss of large particulate matter and most
likely that matter is associated with crustal weathering. This
material is outside of the normal human respirable size fraction and
by composition, it would be unlikely to be associated with aggravated
health. Thus, loss of that portion of the total material may not
have diminished the quality of data for health effects studies. It
may in fact have rendered that data a closer estimate of the respirable
TSP exposure to which the CHESS population groups were subjected.
3-72
-------
3.4.2.2. Other CHESS Program Particulate Matter Measurements—The 1976 IR
noted that four other methods for estimating particulate concentrations were
used at various times to yield aerometry data employed as part of CHESS Program
evaluations of quantitative air pollution/health effects relationships.
Comments relating to each of these other methods (i.e. the dustfall
bucket, the tape sampler, the cascade impactor, and the cyclone sampler) were
reported in Section IV B 4 of the IR (1976) as follows:
(a) The name 'dustfall bucket' is adequately descriptive. It
is basically an open-topped cylinder, with some protection against
wind and rain loss, that is left out in the open, close to the
ground or on a rooftop, for a month. At the end of that time the
dry matter collected is weighed, and sometimes analyzed for trace
metals. The dustfall bucket method is very crude and misses almost
completely the very significant part of the aerosol, including the
respirable aerosol, that does not settle rapidly. It must be considered
here, however, because dustfall measurements were extrapolated to
obtain estimates of suspended sulfates and sulfur dioxide in New
York City during the period 1949-58 ((Table 5.2.1, CHESS Monograph),
and intermittently in Chicago (Table 4.1.A.3), CHESS Monograph).
Dustfall measurements were used as the basis for these extrapola-
tions because there was no other basis for such estimates, but it
must be remembered that the relationship between suspended sulfates
and dustfall is unknown, and that between sulfur dioxide and dustfall
is another step removed from reality.
(b) Coefficient of Haze (CoH) is determined by the automatically
operating tape sampler. It is determined by measuring the optical
density of an aerosol deposited on a filter tape. The aerosol
deposit is obtained by drawing air at a given flow rate through
white filter paper tape for a known period of time. If one could
assume that the composition and physical characteristics of the
aerosol in a given location did not change with time—that only
atmospheric loadings would change—then the COH would give a fairly
good approximation of the variations of particulate loading and
visibility.
However, this assumption is seldom justified, and even at a
given location the CoH only roughly approximates the true particulate
loading. The CoH method is worthless, or nearly so, for comparisons
between areas with dissimilar aerosols. For example, the aerosols
collected at the Utah sites are primarily the light-colored alumino-
silicate dust, whereas the aerosol collected within the inner core
of large cities has a predominantly sooty character. For a given
3-73
-------
participate loading the Utah aerosol will often have as little as
one-tenth the optical density of the urban aerosol.
(c) The cascade impactor operates on the principle that particles
in an air stream will tend to follow a straight line, when the air
stream is deflected, and thus can be impacted on a surface in their
path. The cascade impactor consists of a series of parallel plates
separated by precisely determined spaces. Alternate plates contain a
certain number of holes of a size that is decreased as one goes
through the series of plates from entrance to exit. Alternating
plates contain a certain number of holes of a size that is decreased
as one goes through the series of plates from entrance to exit.
Alternating with the plates containing the calibrated holes are
plates without holes. These may be coated with a medium for the
trapping of impinged particles. Air is drawn through the apparatus
at a known rate, and the particles are collected in decreasing size
fractions related to the decreasing size of the holes in the plates.
(d) The cyclone sampler is a device for the collection of the
respirable size fraction of an atmospheric particulate loading. It
operates on the principle that the inertia of individual particles
will tend to keep the particles moving in .a straight line when the
air stream in which they are carried is deflected. By this means
the larger size particles are removed by impaction and settling,
while the respirable particles are carried along with the air stream
and are subsequently collected on a filter.
The 1976 IR contained very little information beyond the above remarks
concerning the use of these different particulate matter measurement methods
in CHESS. The few pertinent additional comments made, which also relate to
certain community health studies discussed later in this document, concerned
the fact that dustfall measurements from a Manhattan monitoring site were used
as city-wide values for New York City. Also, note was made of the fact that
such values, together with measured values for suspended sulfates for 1956-1970,
were used to estimate suspended sulfate levels in Queens and New York as part
of CHESS community health studies in New York City. Thus, some values on
which important conclusions from CHESS were based, e.g., sulfates may be
harmful to health, were actually estimated values partly derived in the above
manner.
3-74
-------
3.4.2.3 Additional Sources of Uncertainty in Hi-Volume TSP Measurements—In
addition to the above sources of error noted by the IR (1976) regarding CHESS
Program TSP measurements (which the IR concluded resulted in a "slight negative
bias" of CHESS data), a certain other nuance of procedures included in the
Federal Reference Method (40 CFR 50, Appendix B) may have resulted in the
introduction of an additional slight negative bias in TSP data obtained by EPA
CHESS Program investigators and other American researchers. This, more specifi-
cally, pertains to the manner in which flow rate calculations are made upon
which final TSP concentration determinations are based.
The procedure calls for the averaging of the initial and final recorded
airflow rates. However, as described in Appendix 3-A of this chapter, the
uncontrolled flow rate drops more rapidly at the start of the run than at the
end of the run. Therefore, a linear approximation leads to an overestimate of
the flow rate, which will reduce the measured value. Consequently, all TSP
data computed in this manner have an additional negative bias which is likely
usually of the order of 5 percent; on occassion, however, under circumstances
3
where the flow rate may have fallen below 40 ft /min, larger errors (up to
approximately 15 percent) may have been introduced. Assuming that monitoring
site operators in the United States adhere to the recommended Federal Reference
Method procedures, then this type of bias is likely inherent in all American
TSP data collected without flow rate control or recording.
Also, a recent note by Patterson (1980) reports that hi-volume sampler
exhaust appears to be able to recirculate to the hi-volume air intake as shown
by large copper concentrations found on hi-volume filters. This effect may be
important during periods of calm or low wind speed, but its impact on reported
TSP measurement data cannot yet be fully evaluated.
3-75
-------
3.4.3 Summary for American Hi-Volume TSP Sampling
As indicated above in the discussion of the IR (1976) evaluation of CHESS
Program TSP measurements, certain sources of error (e.g., mailing of hi-volume
sampler filters) led to a slight negative bias in CHESS TSP data. That bias,
according to the IR (1976) was typically about 10 percent but, on occassion,
could have ranged up to 20 or 30 percent. It is thusly necessary, when consider-
ing reported CHESS Program TSP data, to take into account the possibility that
it may be slightly biased low by as much as 10 to 30 percent.
In addition, such CHESS data may be biased low by 5 percent or so due to
the flow rate averaging procedure discussed above. Similarly, other TSP data
sets reported in the American literature may also be biased low by about 5
percent where automatic flow rate control or recording were absent. On rare
occassions, probably only occurring at very high particulate levels when
saturation of the sampling filter might result in a flow rate drop below 40
ft /min, the TSP data may be negatively biased by up to about 15 percent.
3.4.4 Summary of Assessment of Particulate Matter Measurement Applications
Tables 3-4 and 3-5, respectively, summarize the sources, magnitudes and
directional biases of errors associated with major British and American applica-
tions of particulate matter measurement techniques during the past 30 to 40
years. Examination of Table 3-4 reveals that numerous sources of error have
been encountered in the course of British application of methods for measurement
of black smoke (BS). For example, prior to 1961, the use of weights for
sealing purposes led to highly variable errors in BS measurements due to
leakage at filter clamps, and steps were taken to require screw-down clamps as
standard procedure as part of the later British National Survey work implemented
after 1961. It is not clear to what extent any specific British BS data sets
3-76
-------
TABLE 3-4. SUMMARY OF EVALUATION OF SOURCES, MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH BRITISH SMOKE (PARTICULATE) MEASUREMENTS
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general Impact
on published BS data
1944-1950s
Pre-1961
Smoke filter
1961-1964
1964-1980
GO
i
Leakage at clamp.
Weights used to make the
seal.
Highly variable under-
estimation of BS levels.
Depending upon observer
and value of R.
Comparing reflectance to
photographs of painted
standard stains.
Reflectance (R) below 25%, 50-100% underestimation.
stain too dark with use
of Clark-Owens DSIR curve.
Computer not following <80% underestimation at low
proper calibration curve. R if not corrected by WSL
(See Moulds,196^) and
discussion of clamp size
correction factor.
Clamp correction factor
for other than 1-inch
clamp.
Flow rate - normal 1 day.
Flow by 8-port with 1
reading per week.
Variability of reading
reflectance.
Averaging reflectance
instead.of averaging
mass/cm .
Use of coarse side of
filter facing upstream.
Uncertain; derivation
cannot be verified.
Possible +20%.
+3% variation.
-10% underestimation.
+10% overestimation.
+2 units of R
Highly variable under-
estimation due to non-
linearity of R.
6-15% underestimation.
Probable widespread highly
variable negative bias.
Probable widespread relatively
small negative bias.
Occasional 50-100% negative
bias in some data sets.
Negligible for BS <~100 ug/ra .
Increasing negative bias up to 80%
as BS values increase over 100
pg/m .
Possible underestimate for 2-inch
and 4-inch clamps
Possible overestimate for l/2-1nch
and 10 cm clamps.
Presumed ± 3% precision level.
10% negative bias on high BS days.
10% positive bias on low BS days.
Error increases with BS level from3±10%
at 50 pg/m up to ±20% at 400 ug/m .
Probable small negative bias at low
BS levels, could be large at high BS.
Occasional negative bias of 6-15%.
-------
TABLE 3-4 (continued).
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely genera! Impact
on published BS data
Reading of wrong side of
stained filter.
Leakage at filter clamp
a. Normal, with good care
b. With inadequate care.
c. Careless loading where
uneven stains are
produced.
Use of wrong clamp size
a. Stain too light R>90%.
b. Stain too dark R<25%.
50-75% underestimation.
1-2% underestimation.
2-8* underestimation.
10-20% underestimation.
Highly variable over-
estimation.
Highly variable under-
estimation.
Occasional negative bias
of 50-75%.
General 1-2% negative bias.
Occasional 2-8% negative bias.
Occasional 10-20% negative bias.
Data usage not recommended.
Data usage not recommended.
co
i
00
-------
TABLE 3-5. SUMMARY OF EVALUATION OF SOURCES. MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH AMERICAN TOTAL SUSPENDED PARTICULATE (TSP) MEASUREMENTS
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general impact
on published TSP data
1954-1980
Staplex Hi Vol TSP
OJ
^J
<£>
Time Off (Due to power
failure).
Weighing error.
Flow measurement (with
control).
Flow measurement (without
control).
a. Constant TSP—Average
of flows.
1. Low TSP level.
2. High TSP level.
b. Rising TSP-Average of
flows.
c. Falling TSP-Average of
flows.
Aerosol evaporation on
standing.
Condensation of water vapor.
Foreign bodies on filter
(Insects).
Windblown dust into filter
during off-mode.
Wind speed effect on pene-
tration of dust into the
Hi-Vol shelter.
Wind direction effect due to
Hi-Vol Asymmetry.
Artifact formation, NO-
SO. . J
Variable underestimation.
±2% random variation.
±2% random variation.
2% underestimation.
5-10% underestimation.
10-20% underestimation.
10-20% overestimation.
1-2% underestimation.
5% overestimation.
Generally small over-
estimation.
Generally small over-
estimation.
Less penetration at high
windspeed.
Higher penetration when
normal to sides.
5-10 ug/m overestimation.
Negligible impact, rare negative bias.
Negligible impact.
Negligible impact.
Negligible impact.
Possible 5-10% negative bias.
Possible 10-20% negative bias.
Possible 10-20% possible bias.
Probable negligible impact.
Possible 5% positive bias.
Possible 5% positive bias.
Occasional (rare) positive bias.
Occasional (rare) negative bias.
Probable increase in random (±) error.
Occasional positive bias.
-------
TABLE 3-5 (continued).
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general impact
on published TSP data
1969-1975
(EPA CHESS
Program).
Fed. Reference
Method Standard
Hi-Vol Sampler
Loss of sampling material
in field.
Loss of sampling material
in mailing.
Evaporation of organic sub-
stances.
Windflow velocity and
asymmetry.
(Overall errors).
No specific estimate of
magnitude of error; but
would be underestimation.
Reported 4-25X-apparent
loss; max. likely due to
crustal (sand, etc.)
fall-off from selected
Utah sampling sites.
No specific estimate of error
magnitude, but not likely to
exceed 5% underestimation.
No specific estimate of error
magnitude; but most likely to
increase random variation of
small underestimation.
Probable slight negative bias
in Utah winter data. No known impact
on other CHESS TSP data.
Probable general small <10% negative bias;
occasional 25% negative bias.
Probable slight negative bias
of <5% for TSP data from urban/
industrial areas.
Negligible impact or slight
negative bias.
Generally <10% negative bias;
occasional 10 to 30% negative bias.
-------
from the 1950s may have been affected by the clamp leakage problem, but one
must assume that such errors could not have often been very large or serious
and that the WSL took appropriate steps to eliminate or invalidate any data in
gross error as they were detected via their quality control efforts in the
late 1950s. Analogously, there is evidence that WSL did take steps to inform
users of pre-1961 BS data of errors arising from (1) comparing reflectance on
filters to photographs of painted stains and (2) use of reflectance readings
below 25 percent, where the stain was too dark to use the Clark-Owens DSIR
curve. However, it also appears that only a few investigators (e.g., Commins
and Waller, 1970) took steps to go back and correct published reports based on
the affected pre-1961 data and to publish revised analyses taking into account
corrections for the pre-1961 data errors.
Probably of much greater concern than the pre-1961 BS measurement errors
are those encountered after the establishment and initial implementation of
the British National Survey in 1961. These include certain errors, e.g., the
"computer error of 1961-1964," which were eventually detected by WSL and
resulted in steps being taken to correct affected BS data in National Survey
data banks. However, although users of the affected data may have been informed
of such errors by WSL, virtually none of them have taken steps to (1) alert
recipients of publications containing analyses based on the affected data of
the likely inaccuracies or ranges of error involved; (2) to reanalyze the
study results based on the affected data sets; or (3) to reissue or publish
anew any revised analyses. Moreover, even some Warren Spring Laboratory
quality control literature prepared and published during the 1960s or 1970s
and still in use may contain incorrect information and recommended standard
procedures for BS measurements based on analyses "contaminated" by computer
3-81
-------
errors or other problems summarized in Table 3-4 and discussed in the
accompanying text.
In regard to determining which British BS data sets and related epide-
miology studies are affected by different post-1961 National Survey errors, it
is again presently very difficult, as was the case with British SO- measurements,
to specify with any confidence the nature and magnitude of specific errors
impacting particular studies. This would probably require thorough examination
of records and WSL quality control reports concerning each of the pertinent
data sets. On the other hand one can project that certain data sets and
British epidemiology studies were almost certainly affected by some subset of
BS measurement errors and these can be taken into account in evaluating such
studies later in Chapter 14. For example, published reports of the "Ministry
of Pensions" (1965) and Douglas and Waller (1966) studies contain specific
ML/^A^K/
reference to usage of^National Survey data from the 1961-64 period and, there-
fore, the results of those studies should be reevaluated in light of measurement
errors reported by the WSL for that period.
As for errors associated with American particulate matter measurements,
most of those reported to affect TSP data gathered prior to, during and after
implementation of the EPA CHESS Program are summarized in Table 3-5. As seen
in that table, a number of the different types of errors listed are potentially
applicable to essentially all measurements using the Staplex Hi-Volume TSP
sampler widely used in the United States. The general range of uncertainty
derived from these various errors is estimated to be less than 20 percent, in
either a positive or negative direction or on a random (±) basis.
3-82
-------
Other errors, in addition to such general errors for TSP measurements,
which were found by the 1976 Congressional Committee Investigative Report
(IR, 1976) to affect CHESS Program TSP measurements during 1969-1975
are broken out and listed seperately in Table 3-5. Some of those errors
(e.g., loss of sample materials in filter removal from the field monitoring
apparatus) were reported by the IR (1976) as likely affecting only very
restricted CHESS data sets. Others, e.g., errors due to loss of sample
in mailing, appear to have been more widespread and presumably impacted
on many CHESS data sets. The IR (1976) concluded, however, that the net
effect of all of the errors was to introduce, in general, a slight
negative bias of 10 to 30 percent into CHESS TSP data. Such a range
of potential error is not far beyond the range of different types of
errors (e.g., linear flow corrections) more generally associated with
American applications of TSP measurements. Section IV C 3 of the
IR (1976) further concluded that:".'. .the TSP data were by far the best
quality data taken in the CHESS monitoring program. Differences
measured between High and Low sites are probably reasonable estimates
of the differences of TSP exposures as received by populations in these
areas." Accordingly, we concur with the IR (1976) evaluation that CHESS
TSP measurements are reasonable estimates of TSP exposures of CHESS
Program community health study populations, taking into account that
such data may by biased low by no more than 10 or, at most, 30 percent.
3-83
-------
3.5 COMPARISONS OF PARTICULATE MATTER MEASUREMENTS
From the preceding discussions, it can be seen that numerous measurement
approaches have been developed in order to quantify concentrations of airborne
particulate matter. Also, as will become apparent upon reading later chapters
of this document, the practical application of the different techniques for
measuring airborne particulate matter in studies on health or welfare effects
has often made it difficult to compare results from different studies. This
in turn has, at times, contributed to the generation of considerable controversy
and disagreement over the interpretation of literature concerning the adverse
effects of atmospheric particulate matter pollution. Critical evaluation of
the existing literature and emerging new studies on the comparison of different
airborne particulate matter techniques and resolution, if possible, of different
among resulting conclusions are crucial in order to interpret and understand
information discussed on effects of particulate matter later in this document.
The air pollution literature is replete with comparisons between different
measurement methods for airborne particulate matter. Examination of that
literature reveals that the tendency has been to use a linear regression, y =
a + bx, to the exclusion of all other models in making comparisons, even if
the physical theory or boundary conditions call for a different type of model.
Many analysts use a linear model because it is simple and it allows easy
visual comprehension of graphical depictions of comparisons. However, if the
data are not linearly related, by forcing a straight line through the data
erroneous conclusions are apt to be drawn. Significant errors can also be
made by extrapolating comparison data beyond the range of the initial study
observations. In the following sections, various comparison studies are
critically evaluated and alternative nonlinear models are proposed by which to
reassess earlier findings.
3-84
-------
3.5.1 Theoretical Considerations
3.5.1.1 Boundary Conditions and Dependent Variable Selection—In order to
compare measurements such as TSP, CoH, BS, and B ., it must be recognized
SCo T>
that each of these measurements have a common point. That is, when TSP is
zero and there are no particles in the air, CoH and BS will be zero and the
residual B . will be from the scattering of clean air, B . Therefore, any
S Cat- 0
comparison model based upon physical theory must go to 0 in the limit TSP -> 0
and B , -> B . It must also be decided what is to be the dependent
SC3T> 0
variable and what is to be the independent variable in making a comparison.
In the comparison of TSP mass versus CoH readings, for example, CoH should be
the dependent variable because the collected mass does not depend on its
optical properties but the transmitted light depends upon the mass in its
path. The simplest nonlinear equation to use which satisifies the boundary
2
condition is CoH = a TSP + b TSP . Dividing through by TSP we obtain CoH/TSP
= a + b TSP. By plotting (CoH/TSP) versus TSP we obtain a straight line if
the model is a good approximation of the data set. However, if we had chosen
TSP as the dependent variable we would have,
TSP = a CoH + b CoH2 (1)
and CoH = -a + Ja2 + 4bTSP (2)
2b
Note this is quite different from the equation
CoH = a TSP + b TSP2 (3)
Thus with only two adjustable constants we can fit y versus x data which has
substantial curvature. Other considerations may be made at the high con-
centration end as well. took Ellison (1968) observed:
3-85
-------
The calibration curve for transmittance measurements may be
substantially different from that for reflectance measurements even
if allowance is made for the fact that however thick a deposit is
collected, the reflectance will never fall below a limiting value
whereas the transmittance will approach zero asymptotically as
soiling increases.
Therefore, CoH versus TSP is a monotonically increasing function which
eventually saturates because at some particular mass loading the transmitted
light is below the minimum detectable level of the photocell.
In the calibration of an AISI sampler, as another example, percent
transmittances have been reported for different materials as a function of
mass loading (see Figures 3-4 and 3-5). Figure 3-4 is an example of
inappropriate data analysis because (as plotted) mass is the dependent variable
(y), and transmittance is the independent variable (x), and the origin (0,0) is no
used in the relation. Each observation has an experimental error associated with
it; and the intercept of the linear regression is a measure of the lack of fit of
the model and the pure error in the data. An alternative should be used to
fit the data in line with the boundary conditions described above.
The CoH measurement should have a relation to Beer's Law which predicts
T /T = exp (k,m), where k, is a constant and T is the initial value of the
transmitted light when the mass on the filter m = 0. According to the theory,
CoH is proportional to the Optical Density; therefore, In (T /T), should be
proportional to m. By inspection we might expect CoH to be a power series in
2
m and CoH = k,m + k?m +... .
Table 3-6 is a listing of % T and Np (number of particles) collected by
Ingram (1969) in a study later reported by Ingram and Golden (1972).
Figure 3-6, a semilog plot of T versus Np shows curvature downward
indicating that light is being blocked out faster than predicted by Beer's Law
3-86
-------
20
15
o
ui
o
I
ill
o
<
u.
oc
CO
10
100
95
90
85
PERCENT TRANSMITTANCE
Figure 3-4. Surface concentration of particulate matter on filter vs
percent transmittance showing lack of fit at the origin.
3-87
-------
400
COAL-LIME STONE
MIXTURE
100
80 70 60 50
PERCENT TRANSMITTANCE
40
30
Figure 3-5. Surface concentration of particulate matter on filter vs percent
transmittance showing a family of related curves.
3-t
-------
TABLE 3-6. PARTICLE COUNTS (Np)
0.3 - 2.0
Date
3/10/69
3/12/69
3/14/69
3/17/69
3/17/69
3/18/69
3/18/69
3/19/69
3/20/69
3/20/69
3/21/69
3/22/69
Time
1405-1605
0925-1125
0925-1125
0915-1115
1335-1535
0600-0800
0815-1015
1206-1406
1135-1335
1900-2100
1720-1920
1520-1720
Run
3
4
5
6
7
8
9
10
11
12*
13
14
T%
94
92
95
78
88
55
56
84
82
71
95
94
CoH
0.62
0.83
0.51
2.48
1.28
5.98
5.80
1.74
1.98
3.42
0.51
0.62
Np x
ID'3
85
94
58
244
177
448
464
187
224
123
78
87
CoH/Np
x 10"4
7.28
8.87
8.84
10.18
7.22
13.34
12.50
9.32
8.86
27.84
6.58
7.11
Relative
humidity (avg)
40%
47%
47%
31%
23%
45%
32%
40%
68%
86% *
42%
43%
*Run not valid because of humidity effect.
/
3-89
-------
>
to
to
60 -
50
Np x 10 3
Figure 3-6. CoH vs particle count showing deviation from Beers Law.
3-90
-------
and that a power series may be used to model it. By way of example we plot
CoH/Np versus Np in Figure 3-7. The line shown is the quadratic CoH = k,Np +
k2 Np2 where ^ = 6.241 x 10"6, k? = 13.73 x 10"12, and r2 = 0.81. If the
accuracy of the fit is not sufficient for the intended purpose we can "bootstrap
up" by plotting (CoH-k,Np)/Np versus Np and get new values for k~ and k_.
This procedure can be performed indefinitely until the needed accuracy is
attained.
3.5.1.2 Dimensional Analysis—In Figure 3-5 we have a family of curves with
upward curvature. Since they are all representative of the same phenomenon,
the same equation should be used to fit all the curves. That is, the use of y
2 3
= a x + b x for one curve and y = a x + b x for another curve based solely
on which equation gives a better fit is not necessarily appropriate. Not
taken into account is the fact that the terms a and b may not be constants
2
because they have dimensions. For example, 'a1 has dimensions of length -/mass
so it does not represent a pure number. The coefficient 'a1 may be a function
of products of variables which have these dimensions. We could hypothesize
that the transmittance depended only on the average particle diameter (D) and
the average particle density (p) (Horvath and Charlson, 1968) such that a =
kDp where k is a dimensionless constant. If we compare measured values of
(a/Dp) we have a measure of k and can begin to predict the value of 'a1 for
another arbitrary dust from knowledge of its particle size and density. The
literature contains several different studies on the value of the parameters
of TSP = a CoH +b in various cities that do not consider boundary conditions
and dimensions. Also, many journals have published articles which compare the
3-91
-------
NP x 10/ft; 0.3 - 2.0
500
Figure 3-7. CoH/Np vs Np showing evidence of a quadratic relation.
-------
value of 'a' in city X to published values of 'a' in city Y, and statements
are made about the differences in air pollution between them. If the hypothesized
<&*4££
model is valid, however, the unity of all cities ea^be demonstrated by plotting
a/D versus p which will result in a line with slope k through the origin.
Therefore, this chapter on particulate matter measurement comparisons will not
simply compare regression coefficients from published models but it will also
attempt to develop more appropriate models where necessary.
3.5.2 Comparison Between TSP and Other Gravimetric Measurements
Recently, the American hi-volume sampler has been compared to a European
version of the high volume sampler, the LIB Hi-Vol Filtergerat described in
Herpertz (1969). It is significantly different from the U.S. type Staplex
Hi-Vol. The LIB has the filter facing downward, exposed directly to the air.
Consequently, the flow pattern is upwards, which will prevent capture of
unusually large particles which could enter the Hi-vol sampler.
Laskus (1977) reported a detailed study in seven German cities of the
Staplex TSP Hi-vol, a multiple stage Anderson non-viable impactor, the LIB
Hi-vol Filtergerat, and several other devices. The relations of importance to
the United States comparisons are given in Table 3-7.
The most important observation is that the U.S. type Staplex TSP Hi-vol
recovers only 82 percent of the total mass collected by the modified Anderson
non-viable impactor. Laskus (1976) modified the Andersen first two stages to
prevent internal losses and increased the mass collection by about 25 percent.
(A similar modification has also been reported by Wedding, McFarland, and
Cermak, 1977). Laskus (1977) states: "The Staplex filter method (high volume
air sampler) proved to be particularly reliable when variations of wind speed
were considered.
3-93
-------
TABLE 3-7. COMPARISON OF TSP RESULTS OBTAINED BY DIFFERENT HI-VOLUME SAMPLERS
TESTED IN GERMAN CITIES BY LASKUS (1977)
Number of Mean
Instrument comparisons ug/m3
LIB (x) 87
Staplex - TSP (y)
A-n.v.-imp (x) 13
Staplex - TSP (y)
Staplex without
gable roof (x) 54
Staplex - TSP (y)
Modified A-n.v.-imp (x) 80
Staplex - TSP (y)
87.5
87.5
115
118
90
88
107
88
Std
48
49
41
40
47
47
46
50
r Regressions
0.93 y = 2.1 + 0.98 x
0.83 y = 3.4 + x
0.88 y = 0.7 + 0.98 x
0.86 y = 9.5 + 0.91 x
However, this apparatus, like the other instruments, did not record the total
particulate content as separation effects occur at the suction aperature due
to external movement of air." This may not be of great importance to human
health if the particles the Staplex TSP does not capture have aerodynamic
diameters which cannot be inhaled by the mouth or nose breathing patterns
(Stober, 1979). ,
M7f
Eickelpasch and Hotz (19680 also compared the Staplex TSP with the LIB
Hi-Vol Filtergerat. They found that the ratio TSP/LIB varied from 1.0 at wind
speed 2 m/s, up to a ratio of 2.0 and greater at wind speeds over 5 m/s. The
same qualitative effect was found by Laskus (1977) as described in Table 3-8,
but above 5 m/s the LIB dropped only 10 percent below the Staplex TSP in the
range of TSP <200 ug/m . A possible explanation for the difference in the
factor of 2 reported by Eickelpasch (1968) and 1.1 reported by Laskus (1977)
3-94
-------
TABLE 3-8. EFFECT OF WIND SPEED ON COLLECTION EFFICIENCY. From LASKUS (1977),
Modif. Anderson
Staplex
LIB-Gerat
Staplex
Staplex without
gable roof
Staplex
Modif. Anderson
Staplex
LIB-Gerat
Staplex
Staplex without
gable roof
Staplex
Modif. Anderson
Staplex
LIB-Gerat
Staplex
Staplex without
gable roof
Staplex
Number
of
Pairs, n
n.v.-Imp. (x) „
(y) 22
Si 24
$ 18
n.v.-Imp. (x) 40
(y)
(x) 44
(y)
(x)2?
(y)
n.v.-Imp (x) 18
(y)
(x) 19
(y)
(x) 9
(y)
Average
ug/m
Windspeed:
103
85
87
86
90
84
Windspeed:
113
93
97
94
94
93
Windspeed:
98
79
66
73
75
83
s Correlation Regression
Coeff, r
0-2 m/s
41 0.71 y = - 23.3+1.06 x
43
43 0.93 y = - 0.4+0.99 x
43
44 n M y = - 4.7+0.98 x
47 0-82
3-4 m/s
46 0.90 y = - 8.7+0.90 x
50
47 0.83 y = - 0.6+0.98 x
47
46 0.89 y = - 6.6+0.92 x
43
>5 m/s
50 0.90 y = - 15.6+0.96 x
60
57 0.94 y = 2.1+1.08 x
55
61 0.95 y = 3.8+1.15 x
64
3-95
-------
may be the difference in locations of the comparative sampling. Laskus sampled
in 7 cities at various places ranging from the runway at Frankfurt Airport to
heavy industrial, light industrial, and residential areas. If Eickelpasch and
Hotz sampled at a location where windblown dust could be an appreciable factor,
there could be more large particles which would not be able to enter vertically
upward into the LIB. The Hi-vol operates with flow slightly upward through
louvered panels, and there is only a slight rise before turning down to the
filter. Consequently, one would expect the LIB to collect less of the larger
particles and subsequently correlate better with the CoH and BS than the
hi-volume sampler at high loadings of relatively large diameter particles.
3.5.3 Comparison of Gravimetric Methods (TSP) with Light Scattering
Nephelometry (^scat^""^ne integrating nephelometer described previously
in Chapter 2 measures the light scattering properties of the particulate
matter in a fixed light path. Because the wave length of visible light is in
the region 0.3 - 0.8 (jm, particles of this size "physical diameter" will
preferentially scatter the light and the manufacturer calibration is based on
a Freon aerosol (IB = 38 pg/m ) of that size. Because the B . is
SC3T. SCctu
measuring "physical diameter" and the BS, TSP, and CoH are measuring in a
manner related to "aerodynamic diameter" these three "mass" measures appear to
have more relation to each other than they have to B .. Of particular
5 CuT*
importance is the interference from water vapor condensation which requires
the use of heated inlets to reduce relative humidity to at least 50 percent.
Ruppersberg (1971) performed theoretical calculations of the changes of the
scattering coefficient of maritime haze with humidity. In Table 3-6 a data
set on AISI CoH versus particle count from Ingram (1969) shows such an outlier
at 86 percent RH caused by condensation of water vapor.
3-96
-------
Horvath and Charlson (1969) performed a detailed study of the nephelometer
and found the factors that influence the B . measure to be as follows:
1. The density of the aerosol particles,
2. The size frequency distribution of the aerosol,
3. The refractive index of the particles,
4. The shape of the particles,
5. The wave length of the light,
6. The relative concentrations of different types of aerosols.
Figure 3-8 describes the variability of measured mass concentration and B ..
Variations of more than ±50 percent are common, indicating that this method
which measures "visibility" can only be expected to approximate atmospheric
particle loading.
/^
When comparing the nephelometer to BS and TSP, Muy 11 eX1978) applied
linear regression and failed to force the fit through the origin as required
by theory (TSP -> 0, B . -» B of clean air). The data were also inverted and
SCo L 0
TSP was treated as being dependent on B . rather than B .a function of
scat scat
particle mass in the atmosphere. Although Muylle reports that the relation 1
unit of B . = 38 |jg/m gives accurate results in the mean concentration
range, it underpredicts at low concentrations and overpredicts at high concentrations
From Figure 3-8 it is likely that a curve is necessary to reach the origin and
we should try B . = A TSP + B TSP2 and plot B ./TSP versus TSP. Kretzschmar
SCal» SC3L
(1975) reports similar results from a study where a hi-volume sampler and
nephelometer were 400 m apart, and also forces a line through nonlinear data.
3-97
-------
Figure 3-8. Measured mass concentration and light-scattering
coefficient for atmospheric aerosols at different locations.
Showing frequent variations of more than 50%.
3-98
-------
3.5.4 Comparison of Hi-Vol TSP Method With ASTM (CoH) Method
The ASTM coefficient of haze (CoH) method was described previously. Park
et al. (1960) found that flow rates of the AISI samplers varied linearly with
optical density and that a correction could be made to the flowrate quite
easily which would make a significant improvement in accuracy of the measured
CoH values. They also found that ten AISI samplers gave widely different
initial flowrates, varying from 0.30 to 0.40 cfm, while sampling the same
filter paper spot. The flow rates also varied among spots on the same filter
paper roll (0.31 to 0.37 cfm). It seems these findings were ignored in later
papers since the correction for flowrate variation was never adopted in the
references cited herein. Hale and Waggoner (1962) compared CoH versus CoH by
running 12 AISI samplers simultaneously. They found the most variability to
be caused by the inhomogeneity of the paper and the bias of the individual
recording the readings. The variabilities (a) between instrument-reader
combinations of a single CoH observation were in the range 0.20 to 0.25 CoH
unit. The variabilities (a) of 12 CoH units averaged to produce a 24 hr
average CoH were close to the expected value (a/V~12 ) CoH for relating the
variability in CoH to variability in mass. Ingram and Golden (1973) report
that this averaging process leads to an error when data are not linear, so
these error estimates may be conservative.
The comparisons of CoH vs hi-volume sampler available in the literature
all report somewhat different relationships. The references most often cited
are Lee et al. (1972), Ingram and Golden (1973), Pedace and Sansome (1972),
and Dalager (1975). Another problem that occurs in data analyses is the fact
that it is theoretically incorrect to compare a regression of CoH on TSP in
3-99
-------
City 1 with a regression of TSP on CoH in City 2. This comes from the fact
that a least squares fit of y vs x gives a different result than when x vs y
is used. For example, if we assume
y = ax (4)
or
x = by (5)
then in general
a = Ixy/Ix2 and b = Zxy/Iy2. (6)
Then a ^ 1/b unless Zx2Iy2 = (Zxy)2
Therefore one could have the same set of CoH and TSP observations in 2 different
cities and have one city report y = ax compared with y = x/b in the other
city. The differences can be very large if there are outliers in the data
set, which is not an unusual occurrence (Lee et al., 1972).
Lee et al. (1972) reported that there were significant differences between
the regression coefficients for the heating season and non-heating season when
CoH was fit against TSP, and the relation also varied from place to place.
Regan et al. (1979) reported a comparison of fine particle fractions collected
near a coal burning power plant, location unspecified. They used a TSP hi-volume
sampler equipped as an Andersen 2000 5-stage particle sizing fractionator.
Although they report correlation coefficients of CoH vs particulate >0.95
there are some apparent problems with the reported results.
(1) The authors fit mass vs CoH instead of CoH versus mass. As described
previously for a least squares analysis, y = f(x) is not the same as
x = g(y);
(2) The inclusion of an intercept is theoretically incorrect. When TSP
= 0, CoH = 0;
3-100
-------
(3) In order to compare twelve (12) daily 2-hour CoH samples with one
24-hour mass sample, the average of twelve CoH values was computed.
Unfortunately a warning by Ingram and Golden (1973) to "Let mass be
a function of CoH." was not considered. That is, if we have 12
values of x then the correct value of x is found by the following
relationship:
- -1 - _ -| 12 fj\
x = f 1(y), where y = ^ I f (x) U)
Li 1 i
Because Regan et al. assumed that the fraction of mass <15 urn was 92
percent of the total mass then their reported regression of
Mass <15 urn = 18.4 + 424 CoH could be transformed to TSP = 20 + 460 CoH.
These results are 3 times the averages reported by both Lee et al. (1972)
and Ingram and Golden (1973) of approximately 150 pg/m per CoH and these
results even fall outside the -2o limit for the non-heating season in England
reported by Lee (1972). This extreme high TSP/CoH may be caused by the com-
bination of factors cited previously or it may be caused by a site specific
property of the local source.
Certain studies by Simon (1976) and Lisjak (1977) cited in Holland et
al. (1979) should also be evaluated here. Simon (1976) searched the literature
and brought forth eight internal reports and memoranda of the Allegheny County
Health Department, Bureau of Air Pollution Control. Simon said: "The earliest
report was written by Ed Tredway to Ron Chleboski on March 30, 1972 covering
the months October, 1971 to February 1972." This is likely to be the first
such study produced because, as Mr. Tredway states: "tape sampler data prior
to October 1971, is considered unreliable." The Simon (1976) report does not
cite the Ingram and Golden (1973) paper, which demonstrates that the
3-101
-------
procedure of averaging CoH values was incorrect. Papers by Sullivan (1962)
and Katz (1958) also not cited by Simon (1976), demonstrate that CoH
was independent of the mass of large particles on the filter and that
CoH vs TSP relations could be fit by a power curve.
These references demonstrated linear regression was not appropriate,
especially if it did not go to zero CoH at zero TSP. The Lisjak paper (1977)
is even more concise, since it moves directly to the point by only citing the
Simon (1976) report. Consideration of Simon (1976) and Lisjak (1977) alone
may have led Holland et al. (1979) to conclude: "One could not expect any
general relationship between CoH readings (transmittance), smoke filter measure-
ments (reflectance) and high volume measurements (gravimetric)." The CoH
transmittance and British Smoke Reflectance are highly related and Saucier and
Sansone (1972) report a detailed comparison of the two (BS versus CoH).
However, they relied on linear regression without the requirement of fitting
at the origin. (Note that the figures in Saucier and Sansone (1972) are
mislabeled as R vs T, but the text and figure description cites T vs R.)
3.5.5 Comparison of American Hi-volume (TSP) and British Smoke (BS) Measurements
3.5.5.1 Introduction--Hi-vo1 TSP and British Smoke (BS) measurement approaches,
frequently used in community health studies of the type discussed in Chapter
14 were developed to measure two distinctly different properties of atmospheric
particulate matter. The TSP measure was intended to represent the total mass
of particulate matter suspended in the atmosphere, whereas the BS measure was
designed to more specifically estimate the relative concentrations of coal
smoke, that is, black staining particles in the air thought to be more closely
tied to health effects occurring during times of elevated particulate pollution
in England. These basic differences in measurement approaches have contributed
3-102
-------
to considerable confusion and controversy regarding (1) how airborne particulate
matter data obtained with the two methods relate to each other; and (2) possible
interrelationships between associated quantitative dose-effect or dose-response
information from studies using TSP measurements vs those using BS data. It is
important, therefore, to evaluate available studies comparing TSP and BS
measurements, key conclusions that have been or can be derived from the results
of those studies, and supporting information on physical phenomena associated
with the TSP and BS measurement approaches that may help to explain observed
comparison relationships.
Numerous studies have been published during the past decade or so, which
report on comparisons of American Hi-vol TSP and BS measurements from colocated
monitoring units at widely dispersed locations in Britain, other European
countries, and the United States. Such studies include those by Commins and
Waller (1967), Lee et al. (1972), Ferris et al. (1973), Kretzschmar (1975),
Dalager (1975), and Ball and Hume (1977). In addition, other recently conducted
studies, by Pashel and Egner (1980) and Bourbon (1980) have yielded extensive
data on comparisons of side-by-side TSP and BS measurements obtained at widely
disparate geographic locations in the United States and France.
Commins and Waller (1967) obtained side-by-side Hi-vol TSP and BS measurements
at a site in central London over the 8 yr period 1955-1963. When 24-hr TSP
readings usually taken once per week, or more frequently during high pollution
episodes, were compared with corresponding BS readings, it was found that the
2
TSP readings tended to average about 100 pg/m higher than the BS measurements.
The differences were most notable in the summer (non-heating) season and under
windy conditions, but there was a closer correlation between the readings
during winter (heating) season high pollution periods. The latter periods
3-103
-------
were usually associated with low wind speed weather conditions during times of
high particulate emissions from coal fires used for heating of homes. Initially,
no attempt was made by Commins and Waller (1967) to mathematically describe
the complex variation of the TSP/BS relationship in conjunction with season
and wind speeds, but a later unpublished regression analysis (by Waller) of
the overall data set (334 pairs of observations) was reported by Holland
et al. (1979) to have yielded results that can be simply expressed as follows:
TSP (ug/m3) = 119 + 0.824 x BS (ug/m3)
In a later study reported by Lee et al. (1972), side-by-side Hi-vol TSP
and BS readings were obtained during 1970 from 11 sites in the United Kingdom,
including some in London and its suburbs as well as several other industrial
cities. Comparison of daily readings over a 5-month study period covering
both non-heating and heating seasons, revealed a generally close relationship
between the readings in areas not under smoke control regulation but there was
a tendency for the Hi-vol TSP readings to be distinctly higher than BS measure-
ments elsewhere. Regression analysis for measurements taken at the London
sites sampled, where smoke ranged from 20 to 140 ug/m BS, yielded the equation:
TSP (ug/m3) = 44 + 1.61 x BS (ug/m3)
Another regression analysis for all 11 United Kingdom sites combined, including
London and elsewhere, yielded the following equation:
TSP (ug/m3) = 5.09 + 0.618 BS (ug/m3) and
BS (ug/m3) = 36.25 + 0.608 TSP (ug/m3)
Ball and Hume (1977) have since reported findings obtained with side-by-side
Hi-vol TSP and BS sampling at one site in central London during the 1-yr
period May 1975, to April 1976, with 24-hr samples being taken 4 days per wk.
Again, TSP readings were generally higher than the BS readings, with greater
3-104
-------
divergence being seen during the summer than in the winter when airborne
particulate matter concentrations tend to be much higher. The following
equations were derived from regression analyses by Ball and Hume (1977) to
describe these relationships mathematically:
Summer TSP (ug/m3) = 22 + 1.64 x BS (pg/m3)
Winter TSP (ng/m3) = 37 + 1.23 x BS (|jg/m3)
Ball and Hume (1977) further noted a close correlation was found between the
amount of smoke (BS) and that of lead in the air during their 1975-76 sampling,
which they interpreted as indicating that much of the smoke was derived from
traffic sources, e.g., diesel engine smoke and gasoline engine lead emissions.
This would be consistent with what would be expected in terms of highly
probable changes in the particulate matter content of London air following the
enactment of the 1956 Clean Air Act and marked reductions in coal fire emissions
already seen in the early 1960s. That is, as coal fire emissions decreased
tremendously during the 1960s in response to smoke control orders and oil-based
home heating increased along with an ever-growing traffic volume, the particulate
matter content of the air over London and other British cities must have
shifted distinctly toward that typically seen in many large American cities.
Thus, by the 1970s, it should be expected that smoke (BS) measurements, and
any TSP readings taken as well in London, should more closely correspond to the
atmospheric particulate matter found in "American air" long characterized
by notable particulate emissions from vehicular traffic and other oil combustion
sources.
In addition to the above comparison studies relating TSP to BS readings
obtained in London and elsewhere in Britain, several other published studies
have reported on analogous comparisons carried out in other European locations
3-105
-------
and the United States. In regard to the latter, for example, Ferris et al.
(1973) reported on TSP-BS comparisons obtained at two sites in Berlin,
New Hampshire, and found discrepancies between the absolute values obtained
that were, in general, similar to those observed in the above studies. That
is, at low arithmetic mean BS levels of 19 ug/m (for March to December, 1966)
and 20 ug/m (for January to December, 1967), the respective side-by-side TSP
readings were 101 and 91, which represent differences of about the same magnitude
as those seen at British locations where BS readings were below 100 ug/m .
As for other published studies, Dalager (1975) reported comparisons of
TSP and BS readings obtained in an urban (Mbenhavn) area and a semi-industrial
o
(Alborg) region of Denmark. Again, discrepancies were found between the two
measurements, with the TSP readings always being considerably higher (around
3 3
60 to 70 ug/m ) than the invariably very low (<5 ug/m ) BS readings obtained
o
in Alborg, with similar but smaller percentage differences percentage being
seen at higher BS readings from Krfbenhavn. Kretzschmar (1975), too, reported
on comparisons of TSP and BS readings, but from sites in Antwerp, Belgium.
Unfortunately, however, the Hi-vol sampler used for obtaining TSP readings was
sited 15 meters above the ground 400 meters away from the OECD apparatus
situated 3 meters above ground level from which the paired BS readings were
obtained.
Holland et al. (1979) recently reviewed the published literature comparing
American Hi-vol (TSP) and British Smoke (BS) measurement results. The review,
however, is limited to discussion of the studies of Commins and Waller (1967),
Lee et al. (1972), Ferris et al. (1973), and Ball and Hume (1977). All of
these studies share a common thread, in that they used linear regression
procedures in attempting to define relationships between the two measurement
3-106
-------
approaches. In discussing the results of the above studies Holland et al.
(1979) concluded:
The several methods in common use for the determination of
suspended particulates cannot be related to one another in any
simple way.
and further that
the British smoke filter (BS), and the high volume sampler (HV)
do not measure the same properties, and the results are clearly not
directly interchangeable. Relationships between them can only be of
an empirical nature, and those found at a particular site and time
cannot be expected to apply to other sets of circumstances.
Commenting further on the studies, Holland et al. (1979) noted that:
Their results showed clearly, however, that where there is
little of the black smoke characteristic of the incomplete combustion
fuel, there are likely to be huge discrepancies between the British
(smoke) and the US (HV) figures.
Carrying the implications of the above conclusions to a logical endpoint, in
discussing the Ferris et al (1973) comparison results for a study in Berlin,
New Hampshire, Holland and colleagues stated:
"Clearly, the characteristics of the suspended particulates
must have been very different in Berlin, New Hampshire, from those
of smoke in British towns."
Holland et al. (1979), therefore, suggest that consistent or meaningful
relationships between Hi-vol (TSP) and BS measurement results are difficult to
discern. Their report, however, concludes:
Nevertheless, some general guidance can be obtained about how
best to 'translate1 British smoke filter results, as cited in epidemiologic
studies carried out over the past 20 years, into equivalent high
volume sampler figures.... For sites in London, where many of the
epidemiologic studies reported in the following sections were carried
out, each of the studies indicates that, on average, smoke (BS)_
values of 100 pg/m correspond with HV values of about 200 ug/m .
The ratio of HV to smoke (BS) result is greater than 2 at still
lower smoke concentrations, and in the other direction it decreases
towards unity for relatively high smoke concentrations, above about
500
3-107
-------
Holland et al. (1979), in describing the Ball and Hume (1977) study,
however, also noted: "simple linear regression equations were calculated ...
even though it seemed likely that there would be divergences from linearity."
However, Holland et al. (1979) did not go on to discuss or develop any nonlinear
models capable of better accounting for the pertinent data relationships
analyzed.
An in-depth analysis of all of the available data (including those from
studies cited by Holland et al., 1979), which takes into account the nonlinear
character of TSP-BS relationships as a sina-qua-non, leads to the proposal of
a "bounded nonlinear (BNLM) model" that appears to fit the available data over
the entire range of existing observations. The proposed model, the rationale
underlying it, evaluation of its ability to predict TSP-BS relationships, and
physical phenomena underlying observed TSP-BS relationships accounted for by
the model are all discussed below in the ensuing text.
3.5.5.2 Development of a Bounded Nonlinear (BNL) Model—The theoretical basis
for proposing a BNL model arises through consideration of boundary conditions
that generally appear to hold for essentially all of the above published
comparison studies. For example, Dalager (1975) earlier noted that a clear
boundary condition exists at the lower limit of atmospheric particulate matter
concentrations. That is, in the limit as TSP goes to zero, BS must also go to
zero because it is a subset of TSP. Also, at the high concentration end, in
the limit, as BS -> »> BS ->• TSP. This arises from consideration of the fact
that for the high BS pollution periods to occur, wind speed must be low and
the air must be relatively stable. Therefore, the larger particles will
settle out and only the small particles, which are usually optically black,
will remain suspended. Consequently, a model defining a linear relation with
an intercept such as BS = a + b TSP is theoretically incorrect. Rather a
3-108
-------
nonlinear model should be fit within the above boundary conditions. The model
should also fit representative emperically-derived data. One could, for
example, choose a model designed to fit, at a minimum, the consensus values
cited by Holland et al. (1979) as summarizing relationships observed at London
3
sites: "at smoke concentrations of the order of 100 ug/m , the corresponding
high volume results (in London) are about double those of the smoke figures,
3 3
while around 250 ug/m smoke (BS) the ratio is about 4:3, and by 500 |jg/m
smoke (BS), it is approaching unity."
A simple nonlinear model that fits both the above boundary conditions for
BS/TSP and the above consensus values can be stated as:
BS = TSP3/(C2 + TSP2) (8)
Equation 8 fits both the 100 and 250 ug/m BS consensus data points with
3 3
almost identical values of C (200 ug/m and 193 ug/m ) for both low and high
BS values; the equation also, by design, meets the boundary constraints. The
o
model is therefore chosen, for simplicity, with C = 200 ug/m . This function,
equation 8, is plotted on Figure 3-9 as the curve labeled "BNLM" for a bounded
nonlinear model (See Appendix B for a dimensional analysis of this relationship)
This type of equation may also be applicable to a data set of total
Dustfall versus Water Insoluble Dustfall in Stalker, Dickerson and Kramer
(1963). In these dustfall data, the water insoluble dust (WID) must go to
zero when total dustfall (TD) goes to zero and these data suggest a linear
relation at high levels of TD. This model would have 2 parameters, a and B,
WID = - (9)
B2 + TD2
where a <1, and B, TD, and WID have units of Tons/mi leVmonth.
3-109
-------
Data used to fit the BNLM model from Holland et al. (1979) fall on the
BNLM curve, as expected by their use in the fit and as shown by the symbol (6)
in Figure 3-9. In addition, data from several other studies reviewed by
Holland et al. (1979) are also plotted on this figure. The data (A) from
Commins and Waller (1967), based on means, are fit well by the model over the
entire range 50 < BS < 500, and the London data (n) from Lee et al. based on
means (BS = 44 + 1.61 TSP) also fit well over the range of their measurements
3
(20 to 140 ug/m BS). The maximum data point recorded by Ball and Hume (1977)
as shown on Figure 3-10 (at 490 |jg/m BS, 600 ug/m TSP), is plotted in Figure
?
3-9 as (-0). A fourth, U.S., data set (©), shown in Table 3-$ from Ferris et
al. (1973) and obtained in Berlin, New Hampshire, is also fit well by the BNLM
model even though it is in a lower concentration range than the English data
sets. Three additional data sets not reviewed by Holland et al. (1979) are
also plotted in Figure 3-9.
TABLE 3-9. ARITHMETIC MEAN CONCENTRATIONS OF SMOKE (BS METHOD)
AND TOTAL SUSPENDED PARTICULATE (HV METHOD)
AT TWO SITES IN BERLIN, NEW HAMPSHIRE*
Site
Fire station
Brown School
10 months
Mar. - Dec.
Smoke
ug/m
19
6
1966
TSP
ug/m
101
-
12
Jan.
Smoke
ug/m
20
9
months
- Dec. 1967
TSP
ug/m
91
-
*Data from Ferris et al. (1973) The authors do not report the Hi-vol TSP
data fcom the Brown School but the BNL^model predicts ~60 ug/m TSP at
6 ug/nT BS and -70 ug/ni TSP at 9 ug/m BS. (The authors note that
they used a linear average in computing Hi-vol flow rates. For these
relatively low mass loadings the correction would appear to be very small.)
3-110
-------
600 ~
460
400 -
360
300
c
CD
250
200
160
100
60
20
0
A
-O
O HOLLAND 11 il (1979) LONDON
A COMMINS «nd WALLER (19671 LONDON
D LEE et at (1972) - ENGLAND
FERRIS tt il (1973) • BERLIN, NH
O-DALAGER (1975) ALBORG, DK
9DALAGERM975) - K0BENHAVN
-OBALL tnd HUME (1977) • LONDON
VKRETZSCHMAR (1975) - ANTWERP
100
200
300
400
600
600
700
AMERICAN Ht-VOL TSP.
Figure 3-9. Measurements of British Smoke vs Hi-vol TSP, showing a consistent
relation between these measures over the entire range of reported observations
Most points shown are annual mean values; see text for discussion.
3-111
-------
700
4
CD
I/)
VI
U
I-
OC
600
500
400
300
200
100
UPPER LINE: GRAVIMETRIC TSP
LOWER LINE: SMOKE SHADE
BS>TSP
2' 3'4l5 '9 llOhlll2'16h7'18M9l23l24l6 \7*8 I 9
DECEMBER 1975
JANUARY 1976
Figure 3-10.
Daily gravimetric and smokeshade data for two representative
periods as measured at the London County Hall. On December 9th,
the two curves cross, and BS is greater than TSP. From Ball and
Hume (1977).
3-11?
-------
As noted earlier, Dalager (1975) reported on BS versus TSP relationships
for two locations in Denmark: (1) Dejligt Kdbenhavn, and (2) Nrirresundby,
Alborg. The TSP was measured by the LIB Hi-vol sampler described previously
in this chapter. Laskus (1977) compared the LIB with the Staplex used in the
United States and British studies and reported that Staplex =2.1+0.98 LIB
with a correlation coefficient of 0.93. In Krfbenhavn, the relation is TSP =
1.27 BS + 43 ug/m and, with a force fit through the origin, TSP = 3.9 BS.
These lines intersect at TSP = 64, BS = 16.4, close to their mean values which
3
are designated as (Q) in Figure 3-9. The value lies 10 ug/m above the BNLM
o
model line. The air in Alborg is much cleaner than Mbenhavn, and Dalager
o
(1975) does not report the Alborg correlation because "the Hi-vol sampler
often gives figures which are 50 times the OECD-smoke concentration (BS)."
This condition is predicted by the BNLM model at TSP = 28 and BS = 0.56.
Ferris (personal communication, 1980) also reported a similar situation of
extremely high TSP/BS ratios at the Brown School in Berlin, New Hampshire, as
< t
shown in Table 3-9. At the Brown School, the BS was often below the minimum
detectable level of the reflectometer while a TSP mass was in the measurable
I17S
range. Figure 3-11 from Dalager (1975"), shows that the weekly average BS in
o
Nrfrresundby, Alborg, is usually below the minimum detectable level; but it
3
increases at times up to a maximum of about 5 ug/m , whereas the LIB Hi-vol
fluctuates between 60 and 70 ug/m over the same time period. This point (5
3 3
ug/m BS, 65 ug/m TSP) is plotted (0-) on Figure 3-9 and falls slightly below
the BLN model curve.
Also, as alluded to earlier, Kretzschmar (1975) compared BS with a modified
Hi-vol, using cellulose filters as reported by Dams and Heindryckx (1973).
However, the Hi-vol was at a height of 15 meters above ground level, 400
3-113
-------
200
150
O)
100
50
LIB HI VOL. SAMPLER ( /jg/m3)
OECD APPARATUS ( /jg/m3) SMOKE SHADE
14
15 16
APRIL
1 1 I
19 20 21 22 23
I ''"I"
17 18
| MAY | JUNE
WEEKLY AVERAGE, ALBORG.DENMARK
1 T 1 1 T~~
24 25 26 27 28 29 WEEK 1973
JULY
Figure 3-11. Comparison of weekly average LIB Hi-vol and smokeshade from
Alborg, Denmark, showing negligible smoke and appreciable TSP.
From Dalager (1975).
-------
meters away from the OECD sampler, which was 3 meters above ground level.
Since particle concentrations tend to decrease with elevation in a gravitational
field, the TSP probably would have been greater if it had also been measured
at 3 meters, the same elevation as the OECD samplers. Because the Hi-vols
were run by Dams (1973), the corrections for flow variations described by him
(Dams, 1973) were utilized so that the flow rate variation effect would not
enter into this comparison. Kretzschmar (1975) reported a linear relationship
of BS = 10 + 0.37 TSP, which places the regression well above the BNLM curve.
At higher BS levels, which are indicative of lower wind speeds, the 95th
percentile points of the TSP and BS cumulative frequency distributions are 225
3 3
pg/m and 110 ug/m , respectively. This point is plotted (V) on Figure 3-9
and it is consistent with both the BNLM Model and the Commins and Waller
3 3
(1967) equations. The arithmetic mean for BS (55 ug/m ) and TSP (125 ug/m )
from Kretzschmar's (1975) study are plotted as a single point (V) on Figure
3-9, where it falls above the BNLM model line.
A recent draft report by Bourbon (1980) discusses the comparison of
(French) Smoke Shade, Hi-volume TSP and a gravimetric low volume collection (1
3
m /hr). Although the simultaneous TSP and smoke shade data are not listed,
the report states that there appears to be no consistent relation between the
gravimetric and reflectometric measurements. However, a plot of the low-volume
sampler versus smoke shade data averaged over an entire city for each year
appears to fall between the BNLM curve and the BS = TSP line. This is not
inconsistent per st; with the British Smoke versus TSP data because at a given
BS, a low-volume sampler will collect less mass than the hi-volume sampler.
Even so, it appears that a curve of the type of equation (8) could be used to
fit the means of the French data. The report by Bourbon (1980) states that
3-115
-------
Schneider CA32 filter papers are used with a photovolt reflectometer 610Y
Tristmulus green filter. However, Bourbon (1980) reports that the reflectance
is fixed at 92.5 percent by a standard off-white enameled tile; that is,
"s = 92.5; Reflectometre cale sur etalon en gres mmaille (sic) blanc"). It
is not clear how this calibration point relates to both the OECD (1964) report
and WHO (1977) report, which define an international curve with the Photovolt
reflectance set to 100 on a pile of clean Schneider CA 32 filter paper.
In summary, we have analyzed comparative BS versus TSP data described
previously from England, the U.S.A., Denmark, Belgium, and France. All these
data, from disparate locations and independent investigations, appear to be
^te-t
well fit by the e-fi-e-parameter empirical BNLM model, which meets applicable
boundary constraints and also fits well the "consensus" observations described
verbally, but not mathematically, by Holland et al. (1979).
In contrast to the consistent fit of the above data achieved by the
BNLM model (Mage, 1980), only the reported values from one recent study
appear to diverge from the curve defined by the model. More specifically, in
a study supported by the American Iron and Steel Institute (AISI) in cooperation
with five AlSI-member steel companies and described in a draft report by
Pashel and Egner (1980) accepted for publication in Atmospheric Environment,
comparisons were made between BS and TSP readings obtained from side-by-side
BS and hi-volume samplers at 16 locations in the United States. The key
conclusions arrived at as a result of the study are graphically depicted in
Figure 3-12, a duplicate of Figure 7 of the Pashel and Egner report. That is,
it could be inferred that the relationship between BS and TSP readings taken in
the United States is as defined by the "AISI (1977)" line in the figure. This
appears to be markedly different from the relationship defined for BS/TSP
3-116
-------
360
300
260
200
160
100
60
MSI (1977)
LEE, *l it (1972)
60 100 150 200 250 300
BSS CONCENTRATIONS, M9/m3
360
400
450
Figure 3-12. Comparison of hi-volume sampler TSP vs British smoke (BS)
relationship defined by data from United States sampling sites (AISI, 1977)
in contrast to analogous comparison relationship defined by data from
England (Lee, et al., 1972). From Pashel and Egner (1980). Note that
the Lee et al. (1977) line, as plotted here from the original Pashel and Egner
(1980) paper, is incorrect.
3-117
-------
readings obtained in England as depicted by the other line in the figure
(which is incorrectly derived from the original relationship reported by Lee et
al., 1972). Based on this apparent discrepancy and the precise slope of the
"AISI (1977)" line, it is implied: (1) that American measurements of airborne
particulate matter concentrations cannot be meaningfully compared with "British
Smoke" measurements, thus negating any possibility of making meaningful comparisons
between American and British epidemiological studies demonstrating quantitative
relationships between various health effects and air particulate matter concentrations!
measured by different (Hi-vol or BS) techniques; and (2) that "no general
relationship between smoke and TSP exists which would be applicable to all
sampling sites."
Interpretation of this AISI study, characterized by the authors as being
"the first large scale, long term attempt to study this relationship in the
United States," requires in-depth critical analyses of its findings and basic
information on the physical underpinnings of BS and TSP measurements, in order
to understand the obvious discrepancy between the reported relationship defined
by the AISI model (see Figure 3-9) the BNLM model curve (Mage, 1980) and
other available data.
Sampling for the AISI study was performed by personnel from several
different steel corporations throughout the United States. The TSP filters
were weighed on-site, whereas the BS filters were mailed to Bethlehem,
Pennsylvania, for reflectance readings. The reflectance readings were also
"duplicated" at Warren Spring Laboratory (WSL) again following mailing of the
filters to England (after they were read in Bethlehem). As stated by Pashel
and Egner (1980):
3-118
-------
Reflectance readings were taken at Warren Springs (sic) to
assure their comparability with data related to the British epidemiological
studies of the 1950's. The Bethlehem reflectance readings were taken
in order to evaluate what impact, if any, transporting the filters
through the mail had on the stain reflectivity. The mean of the
Bethlehem reflectance readings was 91.4% with a standard deviation of
5.36 while the mean WSL reading was 91.9% with a standard deviation
of 5.99. This indicates that mailing the filters had no significant
impact on filter reflectance.
Apparently, Pashel and Egner (1980) did not follow certain IM (1966)
procedures requiring on-site reflectance readings in order to
ascertain that the reflectance is maintained below 90 percent. When reflectance
approaches 90 percent, the IM (1966) calls for changing to a smaller size
filter clamp in order to keep the reflectance below 90 percent. As quoted
from the IM (1966):
If the concentration of smoke is to be obtained with reasonable
accuracy it is essential that the stain should have a reflectometer
reading in the range 40 to 90. During normal conditions, the readings
should be above 70 so that sudden increases in pollution can be
accommodated within the acceptable limits. With dark stains, the
particles form more than one layer so that the concentration obtained
represents only a minimum value, and the flow rate is affected by
the accumulation of material.
The use of 1/2-in clamps should be confined to extremely clean
areas in open country where 1-in clamps regularly give reflectometer
readings of 90 and over. The main disadvantage of using 1/2-in
clamps is that the smoke deposit builds up very quickly, soon causing
the reflectometer reading to fall below 40. At the same time the
resistance to the flow of air rises and the flow rate at the end of
the sampling period may be only one-third of that at the beginning,
when the filter paper was clean.
Because Pashel and Egner (1980) report a mean reflectance greater than 90
percent, their study may have used too large a clamp size and 1/2-inch, or
even 10 mm, clamps should have been used.
Establishing the lack of statistically-significant differences in reflectance
readings due to the mailing of BS filters is essential in order to establish
the reliability (or credibility) of the basic data derived from sampling and
3-119
-------
related analytical procedures employed in the study. These data, in turn are,
of course, crucial determinants in development of the overall AISI "BS/TSP
relationship" model arrived at in the report of Pashel and Egner (1980) and
shown in Figure 7 of their report (Figure 3-12).
Examined next is the crucial assertion that the mailing of samples had
0
"no significant impact on filter reflectance," i.e., their BS readings in ug/m .
In order to apply a statistical test of the null hypothesis of no reflectance
increase for the mailed filters, we can let reflectance measurements be represented
by the general model:
Xij = Mi + eij 1 ' (10)
j = 1, ---- 860
where p, represents the true reflectance before shipment and \i~ the true
reflectance after shipment and e.. represents a random error term with zero
mean and finite variance. The average reflectance before shipment was reported
as x, = 91.4% with a standard deviation of S, = 5.36. After shipment, the
reflectance was x~ = 91.9% with a standard deviation of S? = 5.99. Estimates
are based on n = 860 samples.
To test the null hypothesis Ho: u. = u« versus the alternative hypothesis
H : |J2 > u. a basic one-tailed t-test, outlined by Steel and Torrie (I960),
was calculated:
X2 ~ xl 05
t = * x = -^ = 1.82
S2 2 0.27
n n
This test showed significance at the a = 0.05 significance level. It must,
therefore, be concluded that the reflectance readings after shipment were
significantly higher than reflectance before shipment.
3-120
-------
Reference to Waller (1964) indicates that the correction [8(BS)/3R] AR is
small at R = 91.4% reflectance so the BS correction is not as large as it
would be at R X50%. However, this was the second mailing for most of the
filters since all filters except the Bethlehem filters were apparently mailed
to Bethlehem for analysis. We would expect the most weight loss to occur in
the first mailing since the most loosely bound particles would fall off
quickly, and the weight loss should decrease with each subsequent mailing.
Also, Pashel and Egner (1980) did not use the 1.5 liter/min flow required
by the IM (1966). That 1.5 1/min flow rate was cited by Holland et al. (1979)
as typically being employed in generating BS data used by British epidemiologists
and was used in almost all of the other BS/TSP comparison investigations
discussed here. Instead, Pashel and Egner (1980) used a flowrate of 0.72
liters per minute, which likely causes a significant change in the character
of the particulate matter collected, since it will cause relatively less
material of each size to be collected than the standard BS sampler using the
1.5 liter/minute flow rate.
This assertion is supported by data from the report by Liu et al. (1978)
^UL jJA^_s/
summarized in a table showing^pressure 'drop on penetration for clean Whatman
No. 1 filters (see Table 2-11 in Chapter 2). In that table, there appears to
be an anomalous point in the Dp = 1.0 row at either AP = 1 or 3 cm Hg since
penetration increases as velocity increases from 6.1 cm/sec to 17.4 cm/sec.
Figure 3-13, a plot of pressure drop versus flow velocity shows that for
a 1-inch (2.54 cm) filter spot the pressure drop will be 0.82 cm Hg for the
standard BS method and 0.40 cm Hg for Pashel and Egners low flow conditions.
Table 2-11 also indicates that Whatman No. 1 filter paper has a relatively
high penetration for small (<1 urn) particles at low flowrates. This large
3-121
-------
en
I
u
STANDARD BS FLOW RATE 25 cm3/jec; V = 5.1 cm/sec
PASHELfEGNER FLOW RATE 12 cm3/$ec; V - 2.4 cm/sec
WHATMAN **1 FILTER PAPER
5.1
VELOCITY, cm/sec
Figure 3-13. Pressure drop on Wahtman #1 filter vs flow rate showing effect
of use of lower flow rate than recommended by IM (1966).
3-122
-------
potential penetration at these low diameters means that some BS will escape
capture although one expects the "window" closes as particles begin to fill
the pores and coat the surface.
In a similar manner, the lower flowrate also manifests itself in changing
the penetration of particles into the funnel and through the inlet tubing. In
order to correct these penetration data for the lower flowrate used by Pashel
and Egner (1980), it is necessary to perform a detailed analysis of the BS
flow system to estimate what their results would have been if they had sampled
at 1.5 liter/min. This analysis is given in Appendix C.
Using Figure 3C-1 from Appendix C, the BS has a penetration to the
filter of 98.6% at Dp = 1 pm when u = 79 cm/sec. When the lower flow is used
the penetration at 1 urn is 97 percent. However, if we consider the wind
speed, as shown in Figure 2-15 from McFarland (1979) the inlet line penetration
curve will be shifted further to the left to lower penetrations still. As
stated later in the discussion of Sullivan (1962), only the particles below
1.5 urn have an appreciable effect on the BS. Consequently, the main effect of
the lower flowrate is to reduce the penetration of particles less than 1.5 urn
in an indeterminate manner.
Another test can be made of the hypothesis that the mailing caused a BS
weight loss. Pashel and Egner report BS data from a monitoring site in Bethlehem
identified as Main Office Garage (MOG) - Industrial. Since this site is in
Bethlehem and the BS filters probably were not mailed, we expect these 60
filters to have a relationship consistent with Ferris (1973) and Holland
(1979). Figure 3-14 is a frequency distribution of the ranked observations,
which is in effect a smoothing process and the data do not appear to be
inconsistent with the BNLM model. The 95 percent values appear to be almost
3-123
-------
FREQUENCY DISTRIBUTIONS FOR BSS AND HVS DATA
FOR SITE MOG
•9.9
99
ui 90
3
i M
Ul
3 70
«
ao
BSS
I
8
Ul
o
c
Ul
T
HVS
60
40
30
20
10
L
0.1
10
20
CONCENTRATION,
60
100
200
600
1000
Figure 3-14.
Frequency distributions of British Smoke and Hi-volume sampler
data for site MOG in Bethlehem, PA, reported by Pashel and
Egner (1980).
3-124
-------
identical to the Ferris values of 20 BS and 91 TSP for 1967. The lower values
diverge in a manner to place them above the EPA BNLM, indicating relatively
high BS/TSP in comparison to the other stations.
Yet another test suggests itself from the geographical distribution of
the sampling sites. Eight of the sixteen sites used in deriving the "AISI"
line shown in Figure 3-9 are in Pennsylvania and the other eight are in Ohio
(6), Kansas City, Missouri, and Colorado. The hypothesis to be tested is that
the data for the eight locations in Pennsylvania are closer to the BNLM curve
than the data from the eight more distant locations from which the filters
were mailed. The median value of the distribution was chosen as a robust
parameter for this test. Although Pashel and Egner do not list the median
value they do list the geometric means and standard deviations of the distributions.
One suspects that the geometric mean is a reasonably good estimator of the
median, since the data for the MOG station appear to be fit by a lognormal
distribution well and the data points shown do not lead to the rejection of
the lognormal distribution by the Kolmogorov-Smirnov test (Mage, 1980c). Of
course the lognormal distribution does not fit the physical boundary conditions
for air pollution so it must only be considered as a first approximation. For
example, the Johnson SD model (Mage, 1980a,b), which provides an upper and
D
lower boundary, could be used to prevent the modeled concentration from either
reaching 2 million ppm for a gas or 1 gm/cc for a TSP level which is predicted
by the lines on Figure 3-14 to occur at a finite frequency. Figure 3-15 is a
plot of the geometric means shown in Table 3-10. Of the eight points closest
to the EPA BNLM curve, 6 are Pennsylvania (P) points and the two non-Pennsylvania
points (non-P) are from Ohio. We can apply a Chi-Square test to these results
as shown in Table 3-11.
3-125
-------
TABLE 3-10. SUMMARY OF PASHEL AND EGNER (1980) BS AND TSP DATA FROM
REPRESENTATIVE LOCATIONS IN THE U.S.A.
Location
Kansas City, MO
Middletown, OH
Bethlehem, PA
Pittsburgh, PA
Anneville, PA
Cleveland, OH
Steel ton, PA
Pueblo, CO
Site code
KAN-Industrial
RYS-Industrial
HAJ-Industrial
MOG-Industrial
NOR-Commercial
HAZ-Industrial
LAB- Residential
MEL-Commercial
ISS-Rural
DOR-Industrial
RIT-Industrial
RRC-Commercial
STD- Industrial
WTP-Industrial
MOR-Industrial
HRS-Industrial
Geometric
rep
1 or
110.9
212.3
76.2
57.3
50.4
153.0
64.9
89.8
59.2
167.6
151.4
55.2
69.8
217.3
69.1
79.2
mean
RC
Do
5.6
11.4
8.1
7.8
7.5
22.9
5.4
11.1
4.7
18.3
16.7
5.8
9.1
20.8
8.6
4.0
Geometric standard
deviation
-rep
1.45
1.72
1.63
1.36
1.40
1.62
1.46
1.45
1.51
1.64
1.48
1.72
1.59
1.47
1.55
1.77
BS
1.84
1.94
1.80
1.86
1.97
2.44
2.22
1.80
1.81
1.66
1.55
1.94
2.02
1.61
1.82
1.96
_
3-126
-------
120
140
Figure 3-15.
Pashel and Egner (1980) British Smoke vs Hi-volume sampler
TSP data from sixteen locations in the USA plotted in
relation to BNLM model of Mage (1980) and showing possible
effect of filter mailing. Note that six of eight data points
(MOR, MOG, NOR, HAJ, LAB, ISS) from Pennsylvania sites fall
closer to the BNLM curve than data points indicated as being
mailed over long distances from Ohio, Colorado, and Kansas
City, Missouri.
3-127
-------
TABLE 3-11. CHI-SQUARE TEST SUMMARY TABLE FOR PASHEL AND EGNER
POINTS PLOTTED IN FIGURE 3-15
Grouping Expected Observed
Closest to EPA BNLM Model 8 P 6 P + 2 non-P
Farthest from EPA BNLM Model 8 non-P 6 non-P + 2 P
P
With 1 degree of freedom, the oun/of xf is 2 (8-6) /8 = 1 which is not significant.
Consequently we cannot reject the hypothesis that mailing the filters influenced
the BS values.
It thus appears that the AISI study results reported by Pashel and Egner
(1980) are subject to the same types of criticisms, concerning measurement
errors due to the mailing of particulate matter sampler filters, as were
advanced in the 1976 Congressional Investigative Report (IR, 1976) in regard
to TSP data obtained as part of the U.S. Environmental Protection Agency CHESS
Program community epidemiology research studies. Taking into account the
errors introduced by the mailing of samples in the Pashel and Egner (1980)
AISI study and the other types of errors discussed above, the likely result
would be underestimation of BS readings. Also, accounting for these errors,
one can project that appropriate correction would bring the AISI
results back into closer alignment with the mainstream of other available data
and that the "AISI line" plotted in Figure 3-9 would become more consistent with
the BNLM model curve.
3.5.5.3 Physical Phenomena Underlying TSP-BS Relationships—In the preceding
discussion, it was demonstrated that the bounded nonlinear model proposed
above describes or fits well essentially all presently available TSP-BS
3-128
-------
comparison data in the mean. It is also important, however, to discuss physical
factors or phenomena which likely underly the observed emperical relationships
and to see how they relate to various specific components of the proposed
model. Of particular importance is the understanding of how both BS and TSP
measurements are impacted by variations in size fractions of the air participate
matter sampled and meteorological conditions, such as wind speed, present at
the time of sampling.
For example, several studies have shown that removal of coarse particles,
D » 1 urn, has a negligible effect on the BS measurement of the sample.
Hemeon (1953) reported that removal by settling of 84 percent of TSP from a
stack sample produced no measurable difference in smoke stain optical density
when compared to an ui situ sample. Sullivan (1962) confirmed these results
for ambient air in Australian samples by measuring the BS smoke reflectance
using colocated samplers with and without a prefilter cascade impactor or
cyclone to remove larger particles before they could reach the final BS filter
(Table 3-12). Sullivan (1962) showed that a filter stage with a mass median
diameter (MMD) cut of 1.5 (jm (Lippman, 1959) had slight effect on the sample
reflectance (Table 3-13). However, when another stage prefilter with a MMD
cut of 0.6 urn was inserted the reflectance of the stains was greatly enhanced
(Table 3-14). Sullivan concluded that "only particles of the order of a
micron or less have significant effect on the production of smoke stains on
filter samples." In addition, Waller, Brooks, and Cartwright (1963) reported
electron microscope studies of particulate matter collected by BS devices
during an epidemiological study begun in 1954. During the episode periods
with documented evidence of excess mortality and excess morbidity the MMD of
the particles estimated from the microscopy studies ranged from 0.5 to 1.0 pm.
3-129
-------
TABLE 3-12. COMPARISON BETWEEN FILTER PAPER STAINS WITH AND WITHOUT CYCLONE
PRE-FILTERS USING TWO-INCH DIAMETER HOLDER AND SAMPLING RATE OF
44 LITERS PER MINUTE
No.
of
Test
1
2
3
4
5
6
Reflectance
With
Cyclone
78.5
84.0
80.0
73.5
76.0
81.5
of Stains
Without
Cyclone
78.0
84.0
80.5
71.0
76.0
80.5
Transmission
With
Cyclone
77.5
83.5
78.0
71.5
73.5
79.0
of Stains
Without
Cyclone
78.0
82.0
79.0
72.5
75.0
79.5
TABLE 3-13. COMPARISON BETWEEN FILTER PAPER STAINS WITH AND WITHOUT THREE STAGES OF
CASELLA CASCADE IMPACTOR IN SERIES AS PRE-FILTER USING ONE-INCH DIAMETER
HOLDERS AND SAMPLING RATE OF 17.5 LITERS PER MINUTE
No.
of
Test
1
2
Reflectance
With
Pre-filter
74.5
79.0
of Stains
Without
Pre-filter
74.5
78.5
Transmission
With
Pre-filter
72.5
77.0
of Stains
Without
Pre-filter
73.0
76.0
3-130
-------
Very few particulate aggregates of equivalent diameter greater than 2 |jm were
observed in their samples. On the other hand, other studies suggest that the
size of particles sampled by BS measurement methods depend upon the precise
size of the intake funnel used and may range upwards to 3 urn or more in some
cases.
TABLE 3-14. COMPARISON BETWEEN FILTER PAPER STAINS WITH AND WITHOUT FOUR STAGES OF
CASCADE IMPACTOR IN SERIES AS PRE-FILTER USING ONE-INCH DIAMETER
HOLDERS AND FLOW RATE OF 17.5 LITERS PER MINUTE
No.
of
Test
1
2
Reflectance
With
Pre-filter
92
90
of Stains
Without
Pre-filter
78.5
76.0
Transmission
With
Pre-filter
90
89
of Stains
Without
Pre-filter
77
73.5
Because the BS is a stochastic subset of the TSP there is little reason to
expect BS measurements to be linearly related with TSP measurements. The
ratio of BS to TSP should be bounded, 0 < BS/TSP < 1. The lower limit of zero
corresponds to the British summer conditions with minimal combustion processes
upwind and relatively large amounts of light colored earth crustal material
suspended in the atmosphere. Prior to establishment of smokeless zones in
London the upper limit of unity corresponded to the British winter conditions
when the temperatures fell to 0°C, the air was calm, and strong temperature
inversions existed at the ground surface extending upwards above the chimneys
of the neighboring houses. In the summer case, TSP is much greater than BS.
However, in the extreme winter case, TSP is almost entirely BS since all the
large particles would be removed from the atmosphere by Stokes law settling
and the predominant particles present would be those particles that settled
slowly or were small enough to remain suspended by Brownian motion. This
3-131
-------
relationship was also shown by Lawther, Lord, Brooks and Waller (1973, 1977)
who reported that after the introduction of smokeless zones in London TSP and
BS reading were still similar on days of high pollution and calm or low wind
speed, though widely divergent at most other times. This is an important
factor to be understood because several of the incidences of high pollution
which produced mortality and morbidity data discussed in Chapter 14 arose at
the upper limit of BS/TSP = 1.
In practice, circumstances can also arise which will lead to the paradox
of BS measures being even greater than TSP. This will be briefly discussed
here, because this situation illustrates some deficiencies in the TSP measure-
ment when flow rate is uncontrolled or when TSP concentration is changing
rapidly. At least four studies in the literature report data with British
smoke greater that TSP. Results obtained in one study by Waller and Lawther
(1955), are shown in Table 3-15. As shown in that table, data obtained in
London at St. Bartholomew's Hospital, between 4 p.m. and 7 p.m., indicated
that the BS level was 3980 ng/m3 and the co-located TSP level was 3580 ug/m .
Note that this table is cited incorrectly in Holland et al. (1979) on page
582: "At that time, the concentrations, as expressed in ug/m » were similar
to one another [TSP (HV) 4700 ug/m3 (sic), with smoke (BS) 4640 ug/m3 (sic)
over a 6-hour period]." Between 4 p.m., and 10 p.m. the correct values are BS
= 4700 |jg/m3 and TSP = 4640 (jg/m3.
In a second study done by Lee, Caldwell and Morgan (1972) during the
heating season at Salford, England, the smokeshade data were often significantly
greater than the Hi-vol data as shown in Figure 3-16. In a third study, by
Ball and Hume (1977), at London County Hall they report that, between 1600
hours on Dec. 8th and 1600 hours on Dec. 9th, 1975, the Hi-vol TSP was 178
3 3
ug/m and BS was 182 ug/m , as was shown in Figure 3-10. The fourth study
(WSL, 1968) is discussed elsewhere in this chapter (see page 3-58).
3-132
-------
TABLE 3-15. CONCENTRATION OF SMOKE AND SULPHUR DIOXIDE ON
JANUARY 19/20, 1955. COMPARISON OF METHODS*
Smoke Concentration, Sulphur Dioxide
mg. per cubic meter p. p.m.
Period**
11:30 a.m. - 4 p.m.
4-7 p.m.
7-10 p.m.
10 p.m. - 10:30 a.m.
Standard
D.S.I.R.
Method
1.35
3.98
5.42
1.27
Short-
period
filter
Paper
1.94
3.18
5.08
—
High- Standard
volume D.S.I.R.
Sampler Method
0.44
3.58 0.66
5.70 0.66
0.52
Short-
period
Colori-
metric
—
0.63
0.52
—
*From Waller and Lawther (1955).
**The periods quoted here refer to those for the standard D.S.I.R. method.
Those for the other two methods are similar, but the terminal times differ
slightly in some cases.
3-133
-------
550
500
450
400
350
S 300
LU
O
250
x
O
e/s
200
150
100
50
I
I
I
50 100 150 200 250 300 350
AVERAGE HI-VOL. MB/™3
400
450
Figure 3-16.
Smokeshade Hi-vol correlations for Salford, England showing
occurrences of a significant number of days with smokeshade
greater than total suspended particulates. From Lee et al.
(1972).
3-134
-------
The ratio of 1.022 in the study by Ball and Hume is probably due to the
joint occurrence of a large negative weighing error of the TSP filter and a
large negative reflectance error of the BS filter. This can be clearly seen
by inspection of the regression plots of duplicate reflectometric and duplicate
gravimetric measurements reported by Muylle, Hachez and Verduyn (1978). As
shown in Figures 3-17 and 3-18 below, an infrequent joint occurrence of ±25
pg/m is quite possible.
Several possible explanations occur for the high ratios of the Salford
data reported by Lee et al. (1972). One possibility may be that in Salford,
the particles may have been darker than the particles used in the original
smoke shade calibration so that an equivalent mass reads as a higher BS level.
However, an equally likely explanation may be due to the variation in Hi-vol
flow rates.
When a Hi-volume sampler is run without flow control, the flow rate
decreases nonlinearly with time as particulate matter builds up on the filter,
according to Harrison, Nader, and Fugman (1960) and Dams and Heindryckx
jtt&vya&.'J
(1973). Cohen (1973) reported that the penetration of the hi-volume/fincreased
by 7 percent as flowrate fell from 60 cfm to 30 cfm. This is indicative of
higher particle wall losses and impactions on the asymmetric hi-volume shelter
at the higher flowrates. Consequently, due to the flowrate reduction, the
hi-volume collects more air to sample in the first half of the run than in the
second half. Therefore, if the pollution level is increasing with time during
the run, the hi-volume will likely sample in a biased manner; i.e., comparatively
more of the cleaner portion than the dirtier portion would be sampled. As
described in more detail in Appendix 3A, this may result in an underestimate
of the true average pollution during the whole period and the assumption that
3-135
-------
REGRESSION BETWEEN DUPLICATE REFLECTOMETRIC MEASUREMENTS
300
200
100
100
200
300
Figure 3-17. Correlation between duplicate reflectometric measurements
showing deviations of up to 25 ug/m . From Muylle et al.
(1978).
3-136
-------
REGRESSION BETWEEN DUPLICATE GRAVIMETRIC MEASUREMENTS
300
200
CD
3.
a.
CO
100
100
200
300
TSP,
Figure 3-18. Correlation between duplicate gravimetric measurements showing
j
deviations greater than 25 ug/m . From Muylle et al. (1978).
3-137
-------
the average flow rate during the sampling period is half of the initial flow
rate plus final flow rate also leads to an underestimate. Lee et al. (1972),
Ferris et al. (1973), and most other United States researchers, however,
routinely averaged the initial and final flow rate readings for most of the
Hi-vol sampler data taken without flow control (Lee, 1974). The result will
usually be an underestimate of the TSP, but the BS method sampling at an order
of magnitude lower flow rate, would not likely be affected as greatly. If BS
= TSP and the TSP measurement is biased low, then BS/TSP > I is possible,
creating a paradox of a part being apparently greater than the whole. Such
"paradoxical" observations, however, have in fact been made in a number of
studies, as alluded to above.
If the average flow rate is computed correctly by integrating v = ^ J0 v dt,
then x = M/V T, and the Hi-vol value will be the same value for x computed by
the BS method. Fortunately, the TSP data reported by the British workers Ball
if? 9
and Hume (19^2-) and Waller and Lawther, (1955) were analyzed by integrating
the flowrate versus time so this factor does not influence their reported
results discussed earlier in this chapter.
In summary, as stated by Harrison et al. (1960), the effect cited will
reduce the measured TSP during periods of rising pollution. Similarly, an
opposite effect occurs when the pollution levels are falling. The measured
TSP will be greater than the BS due to enhanced flow in the dirtier beginning
period and lower flow in the cleaner end period. Since Lee et al. (1972) were
using the average of initial and final flow rates for the TSP flows, a correction
to the data from the recorded ratio of initial to final flow would tend to
raise the TSP values. The Salford data, where BS > TSP, could be explained if
the points above the BS = TSP line occurred with increasing pollution levels
and the points below occurred with decreasing pollution.
3-138
-------
The most likely explanation for the case cited by Waller and Lawther
(1955) is that the rising pollution could have biased the Hi-vol data toward
lower readings. The use of 3-hour samples with heavy pollution present was
intended to prevent both the Hi-vol from burning out and the BS from reaching
optical saturation which would stabilize the BS readings (increase in mass on
the BS filter produces no change in reflectivity). In some cases, Waller
reported that a heavy wet fog could saturate the filter paper and restrict
flow. When flow decreased in this manner, it would not be related to the mass
of particles on the filter and the correction procedures outlined by Dams
(1973) could not be followed. Dams and Heindryckx (1973) determined "weight
factors" for initial and final flow rates so the volume sampled could be
obtained without either averaging or integration of variations in flow rate
over the sampling period.
3.5.5.4 Summary and Conclusions for BS-TSP Comparisons—In the above discussion,
information was reviewed on BS-TSP comparisons derived from a number of studies
published over the past twenty years and reporting results obtained from the
sampling of air in many disparate geographic areas (Britain, other European
countries, and the United States) and varying time periods (from the early
1950s to the mid 1970s). It was earlier hinted at and is now clear from a
present perspective, with the advantage of viewing the various BS-TSP comparison
data sets together in relation to each other, that a nonlinear relationship
exists between BS and TSP measurements obtained with colocated samplers.
That is, regardless of where or when such readings were obtained, TSP values
were usually found to be two or more times higher than corresponding BS readings
up to BS levels of about 100 pg/m3. At higher levels, however, the TSP and BS
readings tend to converge toward each other, such that TSP/BS approaches unity
3-139
-------
at BS levels of 500 ^g/rn^ or more. Thus, above 500 ug/m or so, BS and TSP
readings from colocated samplers are essentially identical.
In carrying out analyses of BS-TSP comparison data sets, various investigators
in the past generally employed linear regression analyses in an effort to
define straight lines that best fit data points obtained by them over a limited
range of BS-TSP values. They also often extrapolated the thusly defined
straight lines to BS-TSP values beyond the range of their emperical observations
and found inconsistencies between the BS-TSP relationships implied by their
line(s) and those defined by linear analyses of different BS-TSP comparison
data sets obtained at other times or locations. Such apparent inconsistencies
between BS-TSP relationships, arising from linear regression analyses of
various data sets obtained over limited and often different ranges of BS-TSP
values, have contributed to and reinforced the view that no consistent
relationships exist between BS and TSP measurements obtained at different
locations or even at different times at the same location. Paradoxically, this
view has gained widespread credence despite the concurrent realization
that corresponding BS-TSP readings are nonlinearly related.
Recognition of the acknowledged nonlinearity of BS-TSP relationships and
the necessity to meet certain boundary conditions defined by emperical observations
and certain theoretical considerations as sine-qua-non starting points led to
the formation of a "bounded nonlinear model" (BNLM) proposed in this chapter
as a unifying concept or means by which to interconvert monthly or annual
average BS and TSP values obtained under vastly different circumstances. The
BNLM essentially employs a power function defining a nonlinear relationship
that meets the boundary conditions of (1) BS •* 0, where TSP -* 0 and (2) BS -» TSP,
3-140
-------
as BS •» °°; or, in other words, when there is no particulate matter in the air,
both TSP and BS readings must be zero or tend toward zero and, also, they must
tend to converge toward each other as BS values become very large as observed
in empirical BS-TSP comparison studies. In addition, the particular BNLM
model chosen defines a curve which fits other empirically-derived observations
to the effect that TSP/BS = 2 at 100 ug/m3 BS and TSP/BS = 4/3 at 250 yg/m3
BS. Plotting of corresponding BS-TSP values from numerous published BS-TSP
i
comparison studies reveals that the BNLM model fits well virtually all presently
available BS-TSP comparison data, in the mean.
Only the reported results from one recently obtained data set comparing
BS and TSP values from colocated samplers at 16 locations in the United States
appear, at first sight, to be greatly inconsistent with the BNLM model. Closer
inspection of the study, however, reveals that numerous methodological errors
were made in conducting the study, including the failure to follow published
standard procedures employed in the collection of BS data earlier used in British
epidemiology studies on the health effects of particulate matter and, also, in
most other BS-TSP comparison studies. Nevertheless, when the particular
methodological errors and other deficiencies are evaluated, it becomes
*
apparent that some of the basic observations reported may not be as inconsistent
with other published results and the BNLM model as initially seems to be the
case.
3-141
-------
3.7 REFERENCES
Apling, A. J. , Keddie, A. W. C., M-L P. M. Weatherley, and M. L. Williams,
The High Pollution Episode in London, December, 1975. LR263(AP), Warren
Spring Laboratory, Stevenage, England, 1977a.
Bailey, D. L. R., and H. L. Nicholson. Smoke Filter Calibration Curve: 1-cm
Filter Holder. LR 89 (AP), Warren Spring Laboratory, Stevenage, England,
October 1968.
Ball, D., and R. Hume. The relative importance of vehicular and domestic
emissions of dark smoke in Greater London in the mid-19701s, the
significance of smoke shade measurements, and an explanation of the
relationship of smoke shade to gravimetric measurements of particulate.
Atmos. Environ. 11:1065-1073, 1977.
Barnes, R. Duplicate measurements of low concentrations of smoke and sulphur
dioxide using two "National Survey" samplers with a common inlet. Atmos.
Environ. 7:901-904, 1973.
Bourbgn, P. A Propos de la Mesure des Poussieres et de 1'Anhydride Sulfureux,
Etude Realisee a la Demande de la C.E.E., 1980.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 2: Determination of Concentration of Suspended Matter. B.S. 1747,
British Standards Institution, London, 1963.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 3: Determination of Concentration of Sulphur Dioxide. B.S. 1747,
British Standards Institution, London, 1963.
Cohen, A. L. Dependence of Hi-vol measurements on airflow rate. Environ.
Sci. Technol. 7:60-61, 1973.
Clayton, P The Filtration Efficiency of a Range of Filter Media for Sub-
Micrometre Aerosols. LR280(AP), Warren Spring Laboratory, Stevenage,
England, 1978.
Commins, B. T., and R. E. Waller. Observations from a ten year study of
pollution at a site in the city of London. Atmos. Environ. 1:49-68,
1967.
Dalager, S. Correlations between methods for measuring suspended particulates
in ambient air. Atmos. Environ. 9:687-691, 1975.
Dalager, S. Sammenligning af metoder til maling of svaevestov og svovldioxide.
Environplan A/S, Copenhagen, 1974.
Douglas, J. W. B., R. E. Waller. Air pollution and respiratory infection in
children. Br. J. Prev. Soc. Med. 20:1-8, 1966.
3-142
-------
Additional References Recommended for Consideration in Chapter 3, PM/SO
s\
Bailey, D. L. R., and P. Clayton. The measurement of suspended particulate
and carbon concentration in the atmosphere using standard smoke shade
methods. Report LR 325 (AP), Warren Spring Laboratory, Stevenage, 1980.
Heindryckx, R. Significance of total suspended particulate matter, as determined
by optical density measurements. BECEWA, 1974, (as cited by Kretzschmar,
1975).
Ledbetter, J. 0. , and B. P. Cerepeka. Obscuration versus aerosol concentration.
J. Environ. Sci. Health A15(2):173-181, 1980.
Rosen, H., A. D. A. Hansen, R. L. Dod, T. Novakov. Soot in urban atmospheres:
Determination by an optical absorption technique. Science 208:741-744, 1980.
Swinford, R. L., and D. J. Kolaz. Field correlation of TSP data from a continuous
particulate monitor and high volume air samples. Paper 80-38.3, 73rd
APCA, Montreal, Canada, 1980.
Wallin, S. C. Calibration of the D.S.I.R. Standard Smoke Filter for Diesel
Smoke. Int. J. Air Wat. Poll. 9:351-356, 1965; and discussions by
Lindsey, A. J., M. Corn, S. R. Craxfor, L. R. Reed, and S. C. Wallin,
IBID, 10:73-76, 1966.
-------
3.7 REFERENCES
Apling, A. J. , A. W. C. Keddie, M-L. P. M. Weatherley, and M. L. Williams,
The High Pollution Episode in London, December, 1975. LR263(AP), Warren
Spring Laboratory, Stevenage, England, 1977.
Bailey, D. L. R. , and H. L. Nicholson. Smoke Filter Calibration Curve: 1-cm
Filter Holder. LR89(AP), Warren Spring Laboratory, Stevenage, England,
October 1968.
Ball, D. J., and R. Hume. The relative importance of vehicular and domestic
emissions of dark smoke in Greater London in the mid-1970's, the
significance of smoke shade measurements, and an explanation of the
relationship of smoke shade to gravimetric measurements of particulate.
Atmos. Environ. 11:1065-1073, 1977.
Barnes, R. Duplicate measurements of low concentrations of smoke and sulphur
dioxide using two "National Survey" samplers with a common inlet. Atmos.
Environ. 7:901-904, 1973.
Bourbon, P. A Propos de la Mesure des Poussieres et de I1Anhydride Sulfureux,
Etude Realisee a la Demande de la C.E.E., 1980.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 2: Determination of Concentration of Suspended Matter. B.S. 1747,
British Standards Institution, London, England, 1963.
British Standards Institution. Methods for the Measurement of Air Pollution.
Part 3: Determination of Concentration of Sulphur Dioxide. B.S. 1747,
British Standards Institution, London, England, 1963.
Cohen, A. L. Dependence of Hi-Vol measurements on airflow rate. Environ.
Sci. Techno!. 7:60-61, 1973.
Clayton, P. The Filtration Efficiency of a Range of Filter Media for Sub-
Micrometre Aerosols. LR280(AP), Warren Spring Laboratory, Stevenage,
England, 1978.
Commins, B. T. , and R. E. Waller. Observations from a ten year study of
pollution at a site in the city of London. Atmos. Environ. 1:49-68,
1967.
Dalager, S. Correlations between methods for measuring suspended particulates
in ambient air. Atmos. Environ. 9:687-691, 1975.
Dalager, S. Samenligning af metoder til maling af svaevestov og svovldioxid.
Jji: Luft kvalitets undersogelese Aalborg. 1. Svaevestovmalinger
Norresundby, April-Juli 1973. Environplan A/S, Copenhagen, Denmark,
April 1974. '
Dams, R., and R. Heindryckx. A high-volume air sampling system for use with
cellulose filters. Atmos. Environ. 7:319-322, 1973.
3-142
-------
Douglas, J. W. B., and R. E. Waller. Air pollution and respiratory infection in
children. Br. J. Prev. Soc. Med. 20:1-8, 1966.
Eickelpasch, D., and R. Hotz. Comparison of various outer air measurements.
Stahl und Eisen 98:469-475, 1978.
Ellison, J. McK. The estimation of particulate air pollution from the soiling
of filter paper. Staub Reinhalt. Luft 28:28-36, 1968.
Ferris, B. G. Jr., J. R. Mahoney, R. M. Patterson, and M. W. First. Air quality,
Berlin, New Hampshire, March 1966 to December 1967. Am. Rev. Respir.
Dis. 108:77-84, 1973.
Firket, M. Bull. The cause of the symptoms found in the Meuse Valley during
the fog of December, 1930. Bull. Acad. R. Med. Belg. 683-741, 1931.
Fry, J. D. Determination of Sulphur Dioxide in the Atmosphere by Absorption
in Hydrogen Peroxide Solution—The Effect of Evaporation of the Solution
During Sampling. SSD/SW/M.246 Central Electricity Generating Board, South
Western Region, England, April 1970.
Hagen, L. J., and N. P. Woodruff. Air pollution from duststorms in the Great
Plains. Atmos. Environ. 7:323-332, 1973.
Hale, W. E., and N. E. Waggoner. The overall variability of CoH unit values and
methods for increasing the accuracy of readout. J. Air Pollut. Control
Assoc. 12:322-323, 1962.
Harrison, W. K. Jr. , J. S. Nader, and F. S. Fugman. Constant flow regulators for
the high-volume air sampler. Am. Ind. Hyg. Assoc. J. 21:115-120, 1960.
Hemeon, W. C. L. , G. F. Haines, Jr., and H. M. Ide. Determination of haze and
smoke concentration by filter paper samplers. Air Repair 3:22-28, 1953.
Herpertz, E. A simple long-term measuring procedure for determining the dust
concentration in the near-ground air layer (LIB Method). Staub
Reinhalt. Luft 29:12-18, 1969.
Holland, W. W., A. E. Bennett, I. R. Cameron, C. du V. Florey, S. R. Leeder,
R. S. F. Schilling, A. V. Swan, and R. E. Waller. Health effects of
particulate pollution: reappraising the evidence. Am. J. Epidemiol.
110:527-659, 1979.
Horvath, H., and R. J. Charlson. The direct optical measurement of atmospheric
air pollution. J. Am. Ind. Hyg. Assoc. 30:500-5^9, 1969.
Human Studies Laboratory. Health Consequences of Sulphur Oxides: A Report
from CHESS, .1970-1971. EPA-650/1-74-004, U.S. Environmental Protection
Agency, May 1974.
3-143
-------
Ingram, W. T. Smoke Curve Calibration. APTD-0928, U.S. Department of Health,
Education, and Welfare, Public Health Service, March, 1969.
Ingram, W. T., and J. Golden. Smoke curve calibration. J. Air Pollut. Control
Assoc. 23:110-115, 1973.
Johnson, N. L. Systems of frequency curves generated by methods of translation.
Biometrika 36:149-176, 1949.
Katz, M. Guide to the Selection of Methods for Measuring Air Pollutants.
WHO/AP/67.29, World Health Organization, Geneva, 1967.
Katz, M. , and H. P. Sanderson. Filtration methods for evaluation of aerosol
contaminants. In: Symposium on Instrumentation in Atmospheric Analysis.
Special Technical Publication No. 250, American Society for Testing and
Materials, Philadelphia, PA, 1958. pp. 29-41.
Kretzschmar, J. G. Comparison between three different methods for the estimation
of the total suspended matter in urban air. Atmos. Environ. 9:931-934,
1975.
Laskus, L. Untersuchung der Korngrb'ssenverteilung des atmospharischen
Staubes in Bodennahe. Staub Reinhalt.Luft 37:299-306, 1977.
Laskus, L. , and D. Bake. Erfahrungen bei der Korngrossenanalyse von
Luftstauben mit dem Andersen. Kaskadenimpaktor. Staub Reinhalt. Luft
36:102-106, 1976.
Lawther, P. J. , A. G. F. Brooks, P. W. Lord, and R. E. Waller. Day-to-day
changes in ventilatory function in relation to the environment. Part
I. - Spirometric values. Part II. Peak expiratory flow values. Environ.
Res. 7:27-53, 1974.
Lawther, P. J. , A. G. F. Brooks, P. W. Lord, and R. E. Waller. Day-to-day
changes in ventilatory function in relation to the environment. Part
III. Frequent measurement of peak flow. Environ. Res. 8:119-130, 1974.
Lawther, P. J., P. W. Lord, A. G. F. Brooks, and R. E. Waller. Air pollution and
pulmonary airway resistance: a six year study with three individuals.
Environ. Res. 13:478-492, 1977.
Lee, R. E. Jr., J. S. Caldwell, and G. B. Morgan. The evaluation of methods
for measuring suspended particulates in air. Atmos. Environ. 6:593-622,
1972.
Lee, R. E., Jr. Measuring particulate matter in air. Iji: Instrumentation
for Monitoring Air Quality. Special Technical Publication No. 555,
American Society for Testing and Materials, Philadelphia, PA, 1974.
pp. 143-156.
3-144
-------
Lippman, M. Review of cascade impactors for particle size analysis and a new
calibration for the Casella Cascade Impactor. J. Am. Ind. Hyg. Assoc.
20:406, 1959.
Lisjack, G. J. Comparison of High Volume and Tape Sampler Data, 1976. Allegheny
County Health Department, Bureau of Air Pollution Control, Pittsburgh, PA,
May 1977.
Liu, B. Y. H., P. Y. H. Pui, K. L. Rubow, and G. A. Kuhlmey. Research in Air
Sampling Filter Media. Progress Report EPA Grant 804600, University of
Minnesota, Minneapolis, MN, May 1978.
Logan, W. P. D. Mortality in the London fog incident, 1952. Lancet 1 (6755):
336-338, 1953.
McFarland, A. R. Test Report — Wind Tunnel Evaluation of British Smoke Shade
Sampler. Air Quality Laboratory Report 3565/05/79/ARM, Texas A&M Univer-
sity, College Station, TX, May 1979.
Mage, D. T. An empirical model for the Kolmogorov - Smirnov Statistic. J.
Environ. Sci. Health Part A. 15: 1980c.
Mage, D. T. An explicit solution for SR parameters using four percentile
points. Technometrics 22:247, May 1980a.
Mage, D. T. Frequency distributions of hourly wind speed measurements.
Atmos. Environ. 14:367-374, 1980b.
Martin, A., and F. R. Barber. Some measurements of loss of atmospheric sulphur
dioxide near foliage. Atmos. Environ. E>:345-352, 1971.
McKee, H. C., R. E. Childers, and 0. Saenz, Jr. Collaborative Study of Reference
Method for the Determination of Suspended Particulates in the Atmosphere
(High Volume Method). APTD-0904, U.S. Environmental Protection Agency,
Research Triangle Park, NC, June 1971.
Ministry of Pensions and National Insurance: Report on an Enquiry into the
Incidence of Incapacity for Work. II. Incidence of Incapacity for Work
in Different Areas and Occupations. Her Majesty's Stationery Office,
London, England, 1965.
Moulds, W. Some instrumental variations arising in routine air pollution
measurements. Int. J. Air Water Pollut. 6:201-203, 1962.
Muyulle, E., D. Hachez, and G. Verduyn. An evaluation of measuring methods
for particulate matter. In: Air Pollution Reference Measurement Methods
and Systems. T. Schneider, H. W. deKoning and L. J. Brasser, eds.,
Elsevier Scientific Publishing Co., Amsterdam, Netherlands, 1978.
pp.113-125.
3-145
-------
Natanson, G. Referenced by N. A. Fuchs in The Mechanics of Aerosols, MacMillan,
New York, 1964. p. 112.
National Research Council. Airborne Particles. National Academy of Sciences.
Washington, DC, 1978a, pp. 243-288.
National Research Council. Sulfur oxides. National Academy of Sciences.
Washington, DC, 1978b. pp. 180-209.
Organization for Economic Co-operation and Development. Methods of Measuring
Air Pollution. Paris, France, 1965.
Park, J. C., D. M. Keagy, and W. W. Stalker. Developments in the use of the
AISI automatic smoke sampler. J. Air Pollut. Control Assoc. 10:303-306,
1960. ~~
Pashel G. E., and D. R. Egner. A comparison of ambient suspended particulate
matter concentrations as measured by the British Smoke Sampler and the
High Volume Sampler at 16 sites in the United States. Atmos. Environ.,
in press, 1980.
Patterson, R. K. Aerosol contamination from High-Volume Sampler exhaust. J.
Air Pollut. Control Assoc. 30:169-171, 1980.
Pedace, E. A., and E. B. Sansone. The relationship between "soiling index"
and suspended particulate matter concentrations. J. Air Pollut. Control
Assoc. 22:348-351, 1972.
Ruppersberg, G. Die anderung des maritimen Dunst - Streukoeffizienten mit der
relativen Feuchte. Meteorol. Forschungsergeb. Reihe B, 4:37-60, 1971.
Saucier, J. Y., and E. B. Sansone. The relationship between transmittance and
reflectance measurements of "soiling index". Atmos. Environ. 6:37-43,
1972.
Simon, M. J. Comparison of High Volume and Tape Sampler Data, 1973-75.
Allegheny County Health Department, Bureau of Pollution Control, Pittsburgh,
PA, 1976.
Stalker, W. W., Dickerson, R. C., and G. D. Kramer. Atmospheric sulfur dioxide
and particulate matter: a comparison of methods of measurements. J. Am.
Ind. Hyg. Assoc. 24:68-79, 1963.
Steel, R. G. D., and J. H. Torrie. Principles and Procedures of Statistics,
McGraw Hill, Inc., New York, 1960. pp. 80-83.
Stober, W. Sampling and evaluation of aerosols under biomedical aspects.
Gesellschaft fur Aerosolforschung, 1979.
Sullivan, J. L. The calibration of smoke density. J. Air Pollut. Control
Assoc. 12:474-478, 1962.
3-146
-------
U. S. Congress. House of Representatives. Committee on Science and Technology.
The Environmental Protection Agency's Research Program with Primary
Emphasis on the Community Health and Environmental Surveillance System
(CHESS): An Investigative Report. U. S. Government Printing Office,
Washington, DC, November 1976.
U. S. Environmental Protection Agency. Addendum to "The Health Consequences
of Sulphur Oxides: A Report from CHESS, 1970-1971," May 1974. EPA-
600/1-80-021, U.S. Environmental Protection Agency, Cincinnati, OH, April
1980.
Waller, R. E. Experiments on the calibration of smoke filters. J. Air PoJlut.
Control Assoc. 14:323-325, 1964.
Waller, R. E., A. G. F. Brooks, and J. Cartwright. An electron microscope study of
particles in town air. Int. J. Air Water Pollut. 7:779-786, 1963.
Waller, R. E., and P. J. Lawther. Some observations on London fog. Br. Med. J.
2:1356-1358, 1955.
Warren Spring Laboratory. Measurement of Atmospheric Smoke and Sulphur Dioxide:
Reproducibility of Results. RR/AP/70, Warren Spring Laboratory, Stevenage,
England, August 1962.
Warren Spring Laboratory. National Survey of Smoke and Sulphur Dioxide:
Instruction Manual. Warren Spring Laboratory, Stevenage, England, 1966.
Warren Spring Laboratory. Accuracy and representativeness of the National Survey
data. In: National Survey of Air Pollution, 1961-1971. Volume 5.
Scotland, Northern Ireland, Accuracy of data, Index. Warren Spring Labo-
tory, Stevenage, England, January 1975. pp. 111-119.
Warren Spring Laboratory. The National Survey of Air Pollution. The Use of the
Daily Instrument for Measuring Smoke and Sulphur Dioxide. Warren Spring
Laboratory, Stevenage, England, December 1961.
Warren Spring Laboratory. The Investigation of Atmospheric Pollution 1958-1966.
Thirty-second Report. Her Majesty's Stationary Office, London, England,
1967.
Warren Spring Laboratory. The National Survey of Smoke and Sulphur Dioxide-
Quality Control Tests on Analyses of Samples, October 1975 to February
1977. Warren Spring Laboratory, Stevenage, England, 1977b.
Wedding, J. B. , A. R. McFarland, and J. E. Cermak. Large particle collection
characteristics of ambient aerosol samplers. Environ. Sci. Technol.
11:387-390, 1977.
World Health Organization. Selected Methods of Measuring Air Pollutants.
WHO Offset Publication No. 24, World Health Organization. Geneva,
Switzerland, 1976.
3-147
-------
Dams, R., and R. Heindryckx. A high-volume air sampler for use with cellulose
filters. Atmos. Environ. 7:319-322, 1973.
Eickelpasch, D. , and R. Hotz. Comparison of various outer air measurements.
Stahl Eisen 98:479-485, 1978.
Ferris, B. G. Jr., J. R. Mahoney, R. M. Patterson, and M. First. Air quality,
Berlin, New Hampshire, March 1966 to December 1967. Am. Rev. Respir.
Dis. 108:77-84, 1973.
Firket, M. Bull. The cause of the symptoms found in the Meuse Valley during
the fog of December 1930. Acad. R. Med. Belg. II (ser.5):683-741, 1931.
Fry, J. D. Determination of Sulphur Dioxide in the Atmosphere by Absorption
in Hydrogen Peroxide Solution—the Effect of Evaporation of the Solution
During Sampling. Central Electricity Generating Board, South West Region,
England, April 1970.
Hagen, L. J., and N. P. Woodruff. Air pollution from duststorms in the Great
Plains. Atmos. Environ. 7:323-332, 1973.
Hale, W. , and N. Waggoner. The overall variability of CoH unit values and
methods for increasing the accuracy of readout. J. Air Pollut. Control
Assoc. 12:322-323, 1962.
Harrison, W. , J. Nader, and F. Fugman. Constant flow regulators for high-volume
air sampler. J. Am. Ind. Hyg. Assoc. 11:114-120, 1960.
Health Consequences of Sulfur Oxides: A report from CHESS, 1970-71. EPA-650/1-74-004
U.S. Environmental Protection Agency, Human Studies Laboratory. Research
Triangle Park, NC, May 1974.
Hemeon, W. , G. Haines, Jr., and H. Icle. Determination of haze and smoke
concentration by filter paper samplers. Air Repair 3:22-28, 1953.
Herpertz, E. (1969) Ein einfaches Langzeitmessverfahren zur Bestimmung der
Staubkonzentration in der bodennahen Atmosphare (LIB-Verfahren). Staub
Reinhalt. Luft 29:408-413, 1969.
Holland, W. , A. Bennett, I. Cameron, C. Florey, S. Leeder, R. Schilling, A.
Swan, and R. Waller. Health effects of particulate pollution: reappraising
the evidence. Am. J. Epidemiol. 110:527, 1979.
Horvath, M. , and R. Charlson. The direct optical measurement of atmospheric
air pollution. J. Am. Ind. Hyg. Assoc. 30:500-509, 1969.
Ingram, W. Smoke Curve Calibration. PHS Contract PH-86-68-66, New York
University, New York, NY, 1969.
Ingram, W., and J. Golden. Smoke curve calibration. J. Air Pollut. Control
Assoc. 23:110-115, 1973.
3-143
-------
Johnson, N. L. Systems of frequency curves generated by methods of translation.
Biometrika 36:149-176, 1949.
Katz, M. Guide to the Selection of Methods for Measuring Air Pollutants.
WHO/AP/67.29, World Health Organization, Geneva, 1967.
Katz, M., and H. P. Sanderson. Filtration methods for evaluation of aerosol
contaminants. J_n: Symposium on Instrumentation in Atmospheric Analysis.
Special Technical Publication No. 250, American Society for Testing and
Materials, Philadelphia, PA, 1958.
Kretzschmar, K. Comparison between three different methods for the estimation
of the total suspended matter in urban air. Atmos. Environ. 9:931-935,
1975.
Laskus, Von L. Untersuchung der Korngrb'ssen-verteilung des atmospharischeu
Staubes in Bodeunahe. Staub Reinhalt. Luft 37:299, 1977.
Laskus, Von L. and D. Bake. Erfahrungen bei der Karngrb'ssenanalyse von
Luflstauben mit dem andersne. Kaskadenimpaktor. Staub Reinhalt. Luft
36:102, 1976.
Lawther, P. J., A. G. F. Brooks, P. W. Lord, and R. E. Waller. Day-to-day
changes in ventilatory function in relation to the environment. Part
I. - Spirometric values. Part II. - Peak expiratory flow values. Environ.
Res. 7:27-53, 1974.
Lawther, P. J., A. G. F. Brooks, P. W. Lord, and R. E. Waller. Day-to-day
changes in ventilatory function in relation to the environment. Part
III. Frequent measurement of peak flow. Environ. Res. 8:119-130, 1974.
Lawther, P-, P. Lord, A. Brooks, and R. Waller. Air pollution and pulmonary
airways resistance: a six year study with three individuals. Environ.
Res. 13:478-492, 1977.
Lee, R. E. Jr., J. S. Caldwell, and G. B. Morgan. The evaluation of methods
for measuring suspended particulates in air. Atmos. Environ. 6:593-622,
1972.
Lee, R. E. , Jr. Measuring particulate matter in air. In: Instrumentation
for Monitoring Air Quality. ASTM STP 555, American Society for Testing
and Materials, Philadelphia, PA, 1974. pp. 143-156.
Lippman, M. Review of cascade impactors for particle size analysis and a new
calibration for the Casella Cascade Impactor. J. Am. Ind. Hyg. Assoc.
20:406, 1959.
Lisjack, G. J. Comparison of High Volume and Tape Sampler Data, 1976. Allegheny
County Health Department, Pittsburgh, PA, 1977.
3-144
-------
Liu, B. Y. H. , P. Y. H. Pui, K. L. Rubow, and G. A. Kuhlmey. Research in Air
Sampling Filter Media. Progress Report EPA Grant 804600, University of
Minnesota, Minneapolis, MN, May 1978.
Logan, W. P. D. Mortality in the London Fog Incident. Lancet. (6755):
336-338, 1953.
MacFarland, A. Test report. Wind Turned Evaluation of British Smoke Shade
Sampler. Air Quality Laboratory Report 3565/05/79/ARM, Texas A&M, College
Station, TX, 1979.
Mage, D. T. An empirical model for the Kolmogorov - Smirnov Statistic. J.
Environ. Sci. Health Part A. 15: 1980c.
Mage, D. T. An explicit solution for SR parameters using four percent!le
points. Technometrics 22 May 19805.
Mage, D. T. Frequency distributions of hourly wind speed measurements.
Atmos. Environ. 14:367, 1980b.
Martin, A., and F. R. Barber. Some measurements of loss of atmospheric sulphur
dioxide near foliage. Atmos. Environ. 5:345-352, 1971.
McK. Ellison, J. The estimation of particulate air pollution from the soiling
of filter paper. Staub Reinhalt. Luft 28:28, 1968.
McKee, H. C. , R. E. Childers, and 0. Saenz, Jr. Collaborative Study of Reference
Method for the Determination of Suspended Particulates in the Atmosphere
(High Volume Method). APTD-0904, U.S. Environmental Protection Agency,
Research Triangle Park, NC, June, 1971.
Ministry of Pensions and National Insurance: Report on an enquiry into the
incidence of incapacity for work. II. Incidence of incapacity for work
in different areas and occupations. London, HMSO, 1965.
Moulds, W. Some instrumental variations arising in routine air pollution
measurements. Int. J. Air Water Pollut. 6:201-203, 1962.
Muyulle, E. , D. Hachez, and G. Verduyn. An evaluation of measuring methods
for particulate matter. In: Air Pollution Reference Measurement Methods
and Systems. T. Schneider, H. W. deKoning and L. J. Brassea, eds.,
Elsevier Science Publ. Co., Amsterdam, 1978.
Natanson, G. Referenced by N. A. Fuchs in The Mechanics of Aerosols, p. 112,
MacMillan, New York, 1964.
Organization for Economic Co-operation and Development. Methods of Measuring
Air Pollution. Paris, 1964.
3-145
-------
Park, J. D. , Keagy, D. M., and W. W. Stalker. Developments in the use of the
AISI automatic smoke sampler. J. Air Pollut. Control Assoc. 10:303-306,
1960.
Pashel G. E. , and D. R. Egner. A comparison of ambient suspended particulate
matter concentrations as measured by the British Smoke Sampler and the
High Volume Sampler at 16 sites in the United States. Atmos. Environ.,
in press, 1980.
Patterson, R. K. Aerosol Contamination from High-volume Sampler Exhaust. J.
Air Pollut. Control Assoc. 30:169-171, 1980.
Pedace, E., and E. Sansone. The relationship between "soiling index" and
suspended particulate matter concentrations. J. Air Pollut. Control
Assoc. 22:348-351, 1972.
Ruppersberg, G. Die anderung des maritimen Dunst - Streukoeffizienten mit der
relativen Feuchte. Meteorol. Forschungsergeb. Reihe B, 4:37-60, 1971.
Saucier, J., and E. Sansone. The relationship between transmittance and
reflectance measurements of soiling index. Atmos. Environ. 6_: 37-43,
1972.
Simon, M. J. Comparison of High Volume and Tape Sampler Data, 1973-75.
Allegheny County Health Department, Pittsburgh, PA, 1976.
Stalker, W. W. , Dickerson, R. C., and G. D. Kramer. Atmospheric sulfur dioxide
and particulate matter: A comparison of methods of measurements. J. Am.
Ind. Hyg. Assoc. 24:68-79, 1963.
Steel, R. G. D., and J. H. Torrie. Principles and Procedures of Statistics,
McGraw Hill, Inc., New York, 1960. pp. 81-82.
Stb'ber, W. Sampling and evaluation of aerosols under biomedical aspects.
Gesellschaft fur Aerosolforschung, 1979.
Sullivan, J. The calibration of smoke density. J. Air Pollut. Control Assoc.
12:10-474, 1962.
U. S. Congress. House of Representatives. Committee on Science and Technology.
The Environmental Protection Agency's Research Program with Primary
Emphasis on the Community Health and Environmental Surveillance System
(CHESS): An Investigative Report. U. S. Government Printing Office,
Washington, DC, 1976.
U. S. Environmental Protection Agency. CHESS Addendum, 1980.
Waller, R. E. Experiments on the calibration of smoke filters. J. Air Pollut.
Control Assoc. 14:323-335, 1964.
3-146
-------
Waller, R. , A. Brooks, and J. Cartwright. An electron microscope study of
particles in town air. Int. J. Air Water Pollut. 7:779-784, 1963.
Waller, R., and P. Lawther. Some observations on London fog. Br. Med. J.
2:1356-1358, 1955.
Warren Spring Laboratory. Measurement of Atmospheric Smoke and Sulphur Dioxide,
Reproducibility of Results. RR/AP/70, Warren Spring Laboratory, Stevenage,
England, August, 1962.
Warren Spring Laboratory. National Survey of Smoke and Sulphur Dioxide,
Instruction Manual. Warren Spring Laboratory, Stevenage, England, 1966.
Warren Spring Laboratory. National Survey of Air Pollution, 1961 - 1971.
Volume 5 Scotland - Ireland. Chapter 15: Accuracy of Data and Representativeness
of National Survey Data. Warren Spring Laboratory, Stevenage, England,
January 1975. pp. 111-118.
Warren Spring Laboratory. Stevenage. The National Survey of Air Pollution.
Methods of Measurement, Warren Spring Laboratory, Stevenage, England,
December 1961.
Warren Spring Laboratory. The Investigation of Atmospheric Pollution 1958-1966.
Thirty-second Report. Her Majesty's Stationary Office, London, 1967.
Warren Spring Laboratory. The National Survey of Smoke and Sulphur Dioxide-
Quality Control Tests on Analyses of Samples, October 1975 to February
1977. Warren Spring Laboratory, Stevenage, England, 1977b.
Wedding, J. B. , A. R. McFarland, and J. E. Cermak. Large particle collection
characteristics of ambient aerosol samplers. Environ. Sci. Technol.
11:387-390, 1977.
3-147
-------
APPENDIX 3A
-------
APPENDIX 3A
EFFECT OF FLOW RATE VARIATION WITH CONSTANT AMBIENT CONCENTRATION
Assuming 100 percent collection efficiency for sampling a mixture of only
black smoke particles, we can model the Hi-vol sampler collection rate as
a? = xo v (1)
where
m = mass collected on filter, ug
X = True mass concentration, ug/m , assumed constant
v = Hi-vol flow rate, m /time
The flow rate (v) will be a function of the total collected mass (m). If
we assume a simple relationship similar to Darcy's law for the decrease in
flow rate as mass is collected on the filter, then
Ho4 <«
where v = initial flow rate with a clean filter
M = a filter constant scaled into a pseudo-mass
Substituting into equation (1) and rearranging
(M + m) dm = v M X dt (3)
0 000
Integrating equation (3) we obtain
(MO + M)2 - Mo2 = 2 % MQ T XQ
where M pgm are collected in time T.
Rearranging
M2 + 2 M M
x° =
A-l
-------
The average X TSP computed by averaging the initial and final flow rates is
y = 2 M (6)
ATSP T /V + V M
I 0 0 0
V M + MO
*
Taking the ratios of equation 5 to equation 6,
XI O J- M I
\£ ~l~ n / f-i\
_o = \ o/ (7)
y 4/1 + M
HTSP
4/1 + M
(
For the case where flow rate drops by a factor of 2 during period T, M/Mo = 1
and X /XTSP = 9/8. This implies that the Hi-vol measuring only black smoke
particles as TSP will register (8/9) of the true value measured by the BS
method.
Harrison, Nader, and Fugman (1960) also provided an analysis of
the effect of variable flow rates with time. They modeled the Hi-vol in a
different manner as follows:
dm
V
and
dv _ _ K dm
dt ~ K dt
where "K is a characteristic of the filter." Equation 2 is relatable to
equation (2a) by differentiation which provides
dm
dt
voMo dm (2b)
A-2
-------
Equation (2b) is equivalent to (2a) only at t = o, where m = o and
K = v /M . Substituting (2a) into equation (1) we obtain a first order linear
model which provides that the log of v is linear with time.
v = v e
K X dt
o o
(8a)
v '
The integration of equation (2a) predicts that there is a maximum mass, M
which can be sampled where v will equal zero; M = v /K. However, the
o
iTlclX
integration of equations 1 and 2 predicts that the longer we sample, the more
mass we will collect.
These relations can be tested by examination of the measured variations
in flow rate with time. Table 3A-1 lists typical variations of Hi-vol flow rate
on days with similar pollution levels observed at MRC (Waller, personal
communication) as part of the series of special measurements over the
8:30 a.m. - 10:30 a.m. period reported by Lawther et al . (1974, 1977).
TABLE 3A-1. HI-VOL FLOW -- CFM
Time
0830
0850
0900
0910
0920
0930
0940
0950
1010
1030
v
52
44
--
37
--
--
32
--
28
27
Day 1
Filter # 3112
vA (
0
1.000
0.846
--
0.685
--
--
0.615
--
0.538
0.519
v /v)2
0
1.000
1.397
--
1.974
--
--
2.641
--
3.450
3.709
v
54
46
44
--
37
34
--
32
--
29
Day 2
Filter # - Not
v/v
0
1.000
0.852
0.815
--
0.685
0.630
--
0.593
--
0.537
recorded
(v/v)2
0
1.000
1.378
1.506
-•-
2.130
2.522
--
2.848
--
3.467
A-3
-------
Substituting equation 4 into equation 2, we obtain:
M
-------
O
DATA FROM WALLER
O DAY 1
Q DAY 2
O
I
G
I
20
40
60 BO
TIME, minute!
100
120
140
Figure 3A-1. 1/v variation of Hi-vol flow with time showing flow pattern predicted
by Darcy's Law. Data from Waller by personal communication (1978).
A-5
-------
60 80
TIME. MINUTES
100
120
140
Figure 3A-2. Test for decrease in flow rate proportional to mass collected
showing devnation from the expected exponential relationship.
Figure 3A-2 is a semi logarithmic plot of flow in CFM versus time If the
exponential relation holds a linear relationship will exist between log V
and t As expected the relation is linear at small values of nUn the first
°
hour, but as m builds up the data appear to diverge as (M
m)2 > >
M
A-6
-------
The original equations (1) and (2) will still hold for this case:
dm u
dt = X v
and
v .
If X varies linearly with time as stated above, then
X = XQ (1+ n t/T) (10)
The true value of X which will be measured by the BS method is the average
value of (1 + n/2)X . However, the TSP measure using flow integration will
be,
X = M/jJ v dt (11)
The ratio of TSP to BS defined as 4» will be computed as,
M (12)
4> =
.5n) X VT
where V = i JT v oM o dt (13)
0
The value of m as a function of time must be evaluated by integrating as
follows; substituting equations (2) and (10) in equation (1) we obtain:
$ = Xo (1+ nt/T) ^ MO (14)
Qt M + m
o
A-7
-------
Rearranging equation 14,
JM(t) (M + m) dm = Xo vo Mo J* (T + nt) dt (15)
o o Y o
Performing the integration to the intermediate time t, where o < t < T, we obtain
(M(t) + M )2 - M 2 = VoMo [(T +nt)2 - T2] (16)
0 ° nT
Solving for M (t);
M(t) = Mo 1(1
Substituting equation (17) into equation 13,
M(t) = Mo 1(1 + XOVQ [(T + nt)2 - T2] /nMJ)* -1] (17)
VQ [(T + nt)2 - T2]/nMQ
(18)
Defining a new variable of integration, y = 1 + n t/T, the equation 18 is
simplified to,
h n + 1
J dy (19)
n MQ/X0V0T -
2
We can now define a parameter 6 = n M /X V T, and
(20)
"l y2 + [e2 -I]"2
Integrating equation 20 we obtain
; = J£ 10g(n * i * f iV)». * a* -ifij (n,
Substituting into equation (12) the ratio of the Hi-vol to BS method becomes
> ° (22)
A-8
-------
In the limit as n -> 0 corresponding to the first example of constant
concentration, $ -* 1 (using L'Hopital's rule). For a situation where the
flowrate is reduced by a factor of 2, M = Mo. If we allow the concentration
being sampled to increase by a factor of 2 during the sampling period, n = 1
and equation 22 is reduced to a function of 6 only,
(23)
(1.5) log
(2 + (3 +
-^-TQ
3
To scale 6, let v = 1.4 m /min
T = 1 day or 1440 min
XQ = 100 ug/m3
X v T = 0.2016 gm
oo y
n = 1.0 since X doubles
assume M - X v T
o oo
e =
Scaled to these levels, <)> = -. ,- , „ = 0.962, that is the Hi vol will measure
~ 96 percent of the BS material. If the same type of filter is assumed so
that M = X v T and 6 = V", M(T)/M can be determined from equation 17 as
M(T)/M = VLn + (1 + n)* - l]/n -1 (24)
o
substituting into equation 22
- /(n + 1)' + (n - 1) Vn
(25)
- ~ - 7 - 2
/n + 1 + [(n + 1) + (n
(1 + 0.5n) logl
\ 1 + Vn
The use of equation 25 gives <|> as a function of n corresponding to a linear rise
•3 o
from 100 ug/m to 100 (1 + n) ug/m over the full day. The results are
tabulated below in Table 3A-2.
A-9
-------
TABLE 3A-2. XQ = 100, T = 1 DAY, * = TSP/BS
0
1.000
0.962
0.933
0.909
0.890
0.853
0.825
0.758
n
0^
I
2
3
4
6.5
9
19
X(T)
100
200
30Q
400
500
750
1000
2000
A-10
-------
APPENDIX 3B
B-l
-------
APPENDIX 3B
Dimensional Analysis of BNLM Model.
The parameter C in the BNLM model has dimensions of Mass Length . It
may, therefore, be a function of several variables which can be arranged into
dimensionless groups. Table 3B-1 lists some of the parameters which can have
an influence on the ratio BS/TSP. These parameters can be arranged in
dimensionless groups corresponding to ratios of variables such as a Reynolds
number (Re), a Richardson number (Ri), and a Froude number (Fr), such as:
and Ri = X( + r)/(21-T).
We might try various combinations of these dimensionless groups, one of which
is shown below;
C = 200 k Ri1 Frm Ren p
where k, 1, m, and r are dimensionless constants and p is the air density
(|jg/m ). Since C is proportional to wind speed raised to the n+2m power, as
wind speed goes to zero, C goes to zero. When wind speed becomes large, u»0,
then C becomes large forcing the BS/TSP ratio to become small as characteristi-
cally found in dust storm conditions (Hagen and Woodruff, 1975).
B-2
-------
TABLE 3B-1. PARTIAL LISTING OF PARAMETERS WHICH CAN INFLUENCE
THE MEASURE OF BS AND TSP
Dimensions M mass, L length, 6 temperature, t time
Parameter Dimension
Stability, 3T/8Z 61 L"1
A Temperature, Heating Degrees (21-T°C) 6
Monin-Obukhov Length, A. L
Wind Speed, u L1 t"1
1 ~2
Local Gravitation, g L t
Elevation above ground, Z L
Distance Between BS and TSP L1
Distance to Nearest Street, X L
Wet adiabatic Gradient, T 6, L"1
Relative Humidity, %
Air Viscosity, u M1 L"1 t"1
Air Density, p M1 L~3
Start TSP flow L3 t"1
Stop TSP flow L3 t"1
Start BS flow L3 t"1
Stop BS flow L3 t"1
B-3
-------
APPENDIX 3C
C-l
-------
APPENDIX 3C
In order to make a qualitative estimate of the effect of flow variation
reference is made to a report by McFarland (1979) which describes a wind
tunnel evaluation of the British Smoke Shade Sampler. In that study a
theoretical model of Natanson (1964) was used to estimate the deposition of
particulate matter in horizontal flow tubes by gravitational sedimentation
with a zero windspeed. The penetration, P, is computed as:
P = 1 - 2 (2* V 1 - * 2/3 + sin -1 * 1/3 - * 1/3 7 1 - * 2/3)/n (1)
where * = 2 VtL/4Du
V. = particle terminal velocity, cm/sec
L = tube length, cm
u = Average flow velocity in tube, cm/sec
The terminal velocity of particles obeying a Stokes-Cunningham sedimentation
relation is
CPp Dp2g (2)
Vt = 18J]
where C = 1+2.46 \/D
P
A = 1.7 x 10 T/P, mean free path, cm
T = Temperature °K
P = Gas pressure, torr
o
p = Particle Density, gm/cm
D = Particle diameter
o
g = gravitational acceleration, cm/s
|j = gas viscosity, poise
C-2
-------
McFarland (1979) computes the penetration as a function of the BS parameters
Tube length = 183 cm
Flow rate = 1.5 liter/min
Diameter = 0.635 cm
-4
Viscosity = 1.85 x 10 poise
3
Particle Density = 1 gm/cm
A = 0.0675 x 10"4 cm
u =79 cm/sec
We compute the air Reynolds number, ND to assure laminar flow
K.
n~, n c-jc; „„, 79 cm 29 gm
Du p 0.635 cm -- 24 ?i
N = - = - ^ - = 327
M 1.85 10 H poise
Since NR <2100 we have laminar flow.
Solutions of equations 1 and 2 give penetration as a function of
diameter D as shown in Table 3C-1.
TABLE 3C-1. PENETRATION OF AEROSOL TO THE BS FILTER AT 1.5 LITER/MINUTE
Particle
Diameter
urn
1.4
2.7
6.5
9.7
11.0
Penetration
P, %
98
90
50
10
0
SR Function
Dp/(ll-Dp)
0.146
0.325
1.444
7.462
oo
C-3
-------
Penetration is bounded by a physical upper limit because no particle with
a terminal velocity greater than 79 cm/sec, corresponding to 11 urn, can enter
into the inlet tube. The logical choice to model the penetration curve is by
use of an SB model defined by Johnson (1949). This model requires that the
logarithm of D /(11-D ) is normally distributed. The values of D /(11-D ) are
listed in Table 3C-1 and are plotted on log-probability paper as Figure 3C-1.
The use of the Johnson SR model for aerometric variables is described by Mage
(1980 a,b). The line drawn on Figure 3C-1 was drawn by the technique of minimizini
the Kolmogorov-Smirnov, KS, statistic for the fit. (Mage, paper in preparation).
This KS statistic is related to the goodness of fit of the model (Mage, 1980c).
In this case the KS statistic is about 0.01 with 2 degree of freedom which
indicates the SR model cannot be rejected as providing a good fit to these
data. The equation of the line is Z = - 0.343 + 0.845 ln[D /(ll - D )] where
Z is a standard normal variable.
With a BS sampler running at 0.72 liter per minute, the new penetration
curve can be determined by a rescaling of * in equations 1 and 2 using V.
0 72
= -.'c 79 = 38 cm/sec. Performing
penetration curve shown in Table 3C-2.
0 72
= -.'c 79 = 38 cm/sec. Performing these computation we obtain the new
C-4
-------
10
a
a
I
x
<
O
"a
a
1.0
0.1
_l
T
T
I I I
PENETRATION OF BRITISH
SMOKE SHADE SAMPLER
JOHNSON SB MODEL
I
98
90 70 50 30
PENETRATION, percent
Figure 3C-1. Penetration of British Smoke sampler showing fit to the
Johnson SB distribution.
C-5
-------
TABLE 3C-2. PENETRATION OF AEROSOL TO THE BS FILTER AT 0.72
liter/minute
Particle
Diameter
urn
-0.95
1.85
4.5
6.7
7.6
Penetration
P, %
98
90
50
10
0
Sg Function
Dp/CLl-Dp)
0.143
0.322
1.452
7.44
00
Comparison of Tables 3C-1 and 3C-2 shows the SB penetration function is almost
identical at both flowrates. We, therefore, can use Figure 3C-1 as a model for
penetration of the BS at 0.72 liters per minute. The effect rate is to shift
the penetration curve to a lower recovery of material.
C-6
-------
4. SOURCES AND EMISSIONS
4.1 INTRODUCTION
Sulfur oxides and particulate matter are among the most pervasive pollutants
emitted into the environment. Although particles and sulfur oxides occur natu-
rally, their presence has also been linked with human activities for centuries.
In the early 1600's, legislation was introduced in British Parliament to remove
noxious materials from the air of London. Much later, these pollutants were
identified as sulfur oxides and particulate matter.
The relation between emissions and ambient air concentrations is complex.
Several issues surrounding the nature of this relation remain unresolved. Among
these issues are the chemical composition of pollutants, their secondary forma-
tion, and long-range transport from sources. Some sources and emissions are
evident, and their contribution to the national air pollution problem is apparent.
The purpose of this chapter is to provide background information in order
to discuss, in later chapters, the health and welfare effects of ambient concen-
trations of particles and sulfur oxides. Accordingly, this chapter highlights
the magnitude of emissions from the many diverse, manmade (anthropogenic) sources
of particulate matter and sulfur oxides throughout the United States. Addition-
ally, it places anthropogenic emissions in perspective with natural emissions of
the substances. The chapter demonstrates that emissions from anthropogenic
sources are of considerably more concern than those from natural sources.
Moreover, while emissions from natural sources tend to be distributed uniformly,
those from anthropogenic sources tend to concentrate, thereby greatly elevating
pollutant levels in specific areas.
4-1
-------
4.2 NATURAL SOURCES
Knowledge of natural sources of particulate matter and sulfur oxides is
important for understanding air pollution. Baseline concentrations in continental
and marine air represent natural exposure levels, and they also provide a reference
for comparing concentrations in polluted air. Thus, the concentrations and
biological effects of air pollutants can be compared with those of natural
atmospheric components.
4.2.1 Terrestrial Dust
Terrestrial dust is transferred to the atmosphere by the action of wind on
soils. Theoretical and experimental studies (Gillette, 1974) indicate that sand
grains, produced by the weathering of rocks and soils and moved by wind, cause
the pulverization of soil minerals, as in sandblasting, to produce fine particles.
These particles may become airborne, and the resulting aerosol may be
transported through the atmosphere for considerable distances. Loose soils in
China and Argentina have accumulated from dust carried by wind (Holmes, 1971);
airborne particles make up a significant fraction of some deep ocean sediments
(Broecker, 1974). Dust from the Sahara Desert may be carried by air currents
across the Atlantic Ocean as far as Florida and Barbados (Delaney et al., 1967;
Junge, 1957).
Terrestrial dust in the atmosphere is derived from earth crustal material.
Al, Si, Fe, and other major elements in rock-forming minerals are present in
aerosol samples to nearly the same extent as in average earth crustal material
(Lawson and Winchester, 1979). For crustal matter like calcite, which disaggre-
gates relatively quickly, greater enrichment of its components, such as Ca, will
occur (Johannson et al., 1976).
4-2
-------
Atmospheric concentrations of many trace elements, including transition
metals and semimetals, are 10- to 1000-fold higher than would be expected from
physical dispersion of soil minerals. These anomalous trace element enrichments
have been observed in many parts of the world, including northern Canada (Rahn,
1974), the South Pole (Zoller et al., 1974), and South America (Adams et al.,
1977). Table 4-1 summarizes geometric mean enrichment factors, relative to Al,
for 29 anomalous and 40 other elements according to Rahn's compilation of all
published data up to 1976 (Rahn, 1976).
The atmospheric enrichment sources of these elements are unknown, but
transport from polluting industries (Rahn, 1974), natural rock volatility
(Goldberg, 1976), and biosphere emanations (Barringer, 1977) have all been
suggested. In general, not enough is known about element ratios in the natural
atmosphere to detect a pollution component.
The major elemental constituents of terrestrial dust occur principally in
coarse particles with aerodynamic diameters (i.e., with inertial properties in
aerosol sampling devices equivalent to those of unit density spheres) greater
than approximately 1 urn. Figure 4-1 shows median particle size distributions of
Si and Fe measured at ground level in New Mexico, Colorado, and New Hampshire
during April 1976 (Winchester et al., 1979). The data indicate that more than
90 percent of the mass occurs on the first three impactor stages, >4, 4 to 2,
and 2 to 1 pm aerodynamic diameter. The low relative abundance of submicron Si,
Fe, and other major dust constituents reflects the greater amount of energy
needed in order for fine particles to be generated from soils by the wind-driven
sandblasting mechanism, normally not provided by the atmosphere near the ground.
In urban St. Louis and west suburban sampling sites, Si and Fe occur
predominantly in coarse particles, with higher relative proportions in the >4-pm
4-3
-------
TABLE 4-1. AEROSOL ENRICHMENT FACTORS RELATIVE TO Al
EFa Elements
0.7-7 Li, Na, K, Rb
Be, Mg, Ca, Sr, Ba
Sc, Y, lanthanides
Al, Ga, Tl
Si, Ti, Zr, Hf, Th, U
Mn, Fe, Co, Nb
F, P
7-70b Cr, Cs, V, W, B, Ni , Ge
70-400b H, In, Cu, Mo, Bi , Zn, As
400-4000b I, Hg, S, Cl , Au, Ag, Sn, Sb,
Pb, Br, Cd, Te, Se, C, N
aGeometric means of element ratios to Al , relative to geochemical average earth
crustal material.
EF=
(element/Al)crust
Anomalously enriched elements arranged in order of increasing EF.
Source: Based on Rahn (1976).
4-4
-------
fraction, impactor stage 1 (Figure 4-1; Winchester et al., 1979). This
situation probably reflects physical dust-raising forces in the urban area
qualitatively similar to those in nonurban areas, although with a relatively
stronger tendency in the city for the largest particles to become airborne by
human activities. The concentrations of Si and Fe in the next-smaller
coarse-particle size fractions appear to be less strongly enhanced in St.
Louis than in nonurban locations.
The results in Figure 4-1 for Si and Fe indicate that total suspended
particulate concentrations may be strongly affected by urban dust-generating
activites. The coarsest particle sizes are affected to the greatest degree,
while particles of aerodynamic diameters below about 2 urn are affected to a
lesser extent. If the smaller particles are more significant in terms of
health effects (U.S. Environmental'Protection Agency, 1975), then it is
essential to monitor concentrations in this size range without interference
from coarser particles.
Figure 4-1 also presents median concentrations of Zn as a function of
particle size. Zn is one of the atmospherically enriched elements listed in
Table 4-1, and much of its mass occurs in the smallest particle size ranges.
The fine-particle Zn concentrations are substantially higher in St. Louis than
in the nonurban locations, suggesting that urban and nonur.ban processes lead
to atmospheric release of submicron particles containing Zn. Total suspended
Zn concentrations contain particles in the 1-10 urn particle size range.
The concentrations and elemental composition of terrestrial dust measured
in remote continental areas of the Northern and Southern Hemispheres are
similar. Table 4-2 compares measurements from Bolivia (Adams et al., 1977),
northwestern Canada (Rahn, 1974), Switzerland (Dams and De Jonge, 1976),
4-5
-------
3000
1000 r
300
100 -
II I I I I
STL W
4/09 15/76
I I I I I I
STL C
4/09 15/76
MEDIANS OF 7-10 SAMPLES
0.1
654321 654321 654321 654321 654321
IMPACTOR STAGE
Figure 4-1. Median concentrations of Si, Fe, and Zn are plotted as a
function of particle size. Samples were obtained during April 1976
at five sites: near Los Alamos, NM; Squaw Mountain, CO; west
suburban St. Louis, MO; urban St. Louis, MO; and Hubbard Brook
watershed NH.
Source: Winchester et al. (1979).
4-6
-------
TABLE 4-2. COMPARISON OF AEROSOL COMPOSITION IN REMOTE REGIONS
•- . - -
Determination
Bolivia
(10)
Location and reference
NW. territory,
Canada (8) Switzerland
North Cape,
Norway (17)
South Pole
(9, 18)*
Fe concn, ng/m 181
Weight ratio to Fe
71
36
51
0.84
Na
K
Mg
Ca
Sc
La
Sm
Al
Si
Ti
Th
Mn
Co
Cr
Cs
V
W
Ni
In
Cu
Zn
As
I
S
Cl
Sb
Pb
Br
Cd
Se
0.33
0.44
0.16
0.46
0.00024
0.0009
0.00015
1.4
-
0.086
0.0003
0.014
0.005
0.0033
0.0004
0.002
0.004
0.003
0.00007
0.0075
0.024
0.0089
0.0014
0.095
0.37
0.0048
0.029
0.0058
0.006
0.0010
0.25
0.76
0.23
0.56
0.0006
0.0013
0.0002
0.93
-
0.070
0.0007
0.021
0.0006
0.0083
-
0.003
0.0002
<0.01
0.00002
0.013
0.053
0.0044
0.0028
-
0.13
0.0018
-
0.0076
-
0.0006
0.61
0.56
0.28
1.4
0.00021
0.0011
0.0002
1.4
2.8
0.068
0.0003
0.042
0.0013
0.010
0.0004
0.008
0.0008
-
0.00003
0.024
0.275
0.0064
0.0075
-
0.20
0.0056
0.12
0.036
0.014
0.0012
8.6
0.94
1.4
0.90
0.00014
0.0008
0.0001
0.84
2.8
0.061
0.0002
0.049
-
0.013
0.0004
0.033
-
-
0.00008
0.045
0.17
0.037
0.012
-
5.8
0.0075
0.11
0.092
-
0.005
5.3
1.1
1.2
0.8
0.00026
0.0007
0.00015
1.3
-
0.16
0.0002
0.021
0.0008
<0.06
0.0002
0.0021
0.0024
-
0.00009
0.047
0.053
0.01 (0.05N)
0.13 (1.2N)
80.
4.2
0.0013
0.75
2.1 (4.2N)
<0.02
0.01
Pb values are from reference 9; all others are from reference 18. "N" denotes
value from Nuclepore filter sample when different from Whatman cellulose filter
bvalue.
Elements grouped according to enrichment factors as in Table 4-1.
4-7
-------
northern Norway (Rahn, as quoted in Dams and De Jonge, 1976), and the South
Pole (Zoller et al., 1974; Maenhaut et al., 1979). The concentrations of Fe,
an element respresentative of terrestrial dust in these locations, ranges over
a factor of 5 from South and North America and Europe. Concentrations are
similar to those found in the Western United States (Figure 4-2; Winchester
and Duce, 1977). Because of the paucity of land areas not covered by snow,
concentrations at the South Pole are 100 times lower than the average of the
other four locations. In general, all five locations have similar elemental
iron ratios. Exceptions may be attributed to the relatively strong influence
of sea salt (high Na, Mg, Cl, Br, and I concentrations at the sites in northern
Norway and the South Pole) and certain pollutants (high V, Cu, Zn, and Pb
concentrations at the sites in Switzerland and Norway).
The terrestrial dust component of atmospheric aerosols appears to be
generally uniform worldwide in relative elemental composition, representing soil
dispersion processes and anomalous enrichments of some elements. Certain excess
elemental concentrations may be due to special processes, such as transport of
sea salt and industrial pollutants, the raising of surface dust by human
activity, or generation of fly ash and other dusts of earth crustal composition.
Judicious comparison of both concentrations and relative elemental composition
in the natural atmospheric aerosol and in more polluted air reveals the presence
of these excess elemental concentrations.
4.2.2 Radioactive Aerosols
The fine fraction of aerosols of terrestrial origin contains radioactive
daughter products of radon. Uranium and thorium in soils support radioactive
Rn-222 (half-life, 3.823 days) and Rn-220 (half-life, 55 sec), which can diffuse
into the atmosphere and subsequently decay to a and p radioactive isotopes of
4-8
-------
TOP JET DROP
6 cm EJECTION HEIGHT
60 rm DIAMETER
0 rm
FILM CAP
FILM DROPS
15-20 PRODUCED
1-10 >m DIAMETER
500 vm DIAMETER
BUBBLE AT INTERFACE
RISE VELOCITY
*e -1
35 cm t
Figure 4-2. Bursting of a bubble, 500 pm in diameter, at the sea-air
interface produces atmospheric particles.
Source: Blanchard (1963).
4-9
-------
Po, Bi, and Pb. These daughter products attach themselves to fine aerosol
particles which settle on surfaces near the ground.
4.2.3 Sea Spray
Aerosol droplets are generated at the ocean surface by the action of wind,
principally through a process whereby air bubbles become entrained and rise to
burst at the surface. Figure 4-2 illustrates how the collapsing bubble cavity
ejects a central jet, which breaks into approximately four jet drops, each about
1/10 the diameter of the original bubble (Blanchard, 1963). The jet drops are
composed of seawater and organic and other surface-active materials which may be
concentrated into the 0.05- to 0.5-|jm thickness of bubble surface (Maclntyre,
1974). Bubbles with diameters of 100 to 1000 pm occur abundantly in the sea
surface and produce jet droplets of 10 to 100 pm in diameter, small enough to
become airborne in moderate winds. Additional fine droplets are produced by
rupture of the film cap from bubbles with diameters of >300 pm, and a 2000-um
bubble may produce up to 100 film drops of <10-um diameter from a cap ^.2 urn
thick.
Both film and jet drops are enriched in surface-active material, but because
of differences in the mechanics of droplet formation, there may be differences
in chemical composition of the two kinds of drops (Berg and Winchester, 1978).
For example, jet drops may be formed from thinner bubble layers than are film
drops, and they may contain surface-active particulate matter scavenged from the
water column by the rising bubbles.
The surface-active material enriched in jet and film drops may be of
natural or pollutant origin and may include organic molecular films, organic and
inorganic particulate matter suspended in near-surface waters, and viable
4-10
-------
particles including viruses, bacteria, and larger living forms (Blanchard and
Parker, 1977; Duce and Hoffman, 1976). Such materials, by becoming components
of sea spray aerosol droplets, may be carried through the atmosphere far from
the point of origin. The potential for virus transfer from coastal waters to
the atmosphere and transport by winds inland to inhabited areas has been
demonstrated (Baylor et al., 1977). Once the processes of formation and
transport of natural sea spray droplets are understood, the potential sea-tb-air
transport of viable and nonviable water pollutants may be predicted.
Sea salt droplets are produced by dispersion of a condensed phase and tend
to be most abundant in large particle sizes. Particles tend to be large because
their formation requires less energy than does that of small particles.
Figure 4-3 shows size distributions of Cl and Br particles in the clean
marine atmosphere of Hawaii (Moyers and Duce, 1972). Both elements occur
predominantly in the 1- to 8-um aerodynamic diameter range (impactor stages D,
C, B), with lower concentrations in the 8-um ranges. The droplets
may originate primarily from film cap rupture, with a contribution of jet drops
to the fraction collected by impactor stage A. The size distributions are
similar, and the elements are in proportions approximating the composition of
seawater, indicating that they are mainly constituents of the dispersed water
rather than surface film material. Lower Br/Cl ratios in the smallest particle
size ranges, however, suggest surface chemical fractionation during droplet
formation or atmospheric chemical reactivity and gas-particle interactions,
leading to changes in aerosol composition prior to sampling.
Figure 4-4 compares the particle size distributions of iodine and chlorine
aerosols in the Hawaiian marine atmosphere (Moyers and Duce, 1972). Unlike Br,
4-11
-------
ID1'
D
m
10"
10
-2
10"
10"
I I I I I ! I I I I I I I I
Br/CI
Br
I I I I I I I
10J
102 g
o
OD
10
r
m
10"
F E D C B A TOT F E D C B A TOT
CASCADE IMPACTOR STAGE
Figure 4-3. Average particle-size spectra for chloride concentration,
bromide concentration, and bromide/chloride ratio were obtained
with an Andersen impactor with a minimum stage E cutoff diameter
of 0.5/;m.
Source: Moyers and Duce (1972).
4-12
-------
E E
I
o
c
§ q
O 1£T
10H
i n r
\ \ i i i
i i
I/CI
IL 1 J J J I I I I I I I I
10"
10
,-2 D
10
•3
F E D C B A TOT F E D C B A TOT
CASCADE IMPACTOR STAGE
Figure 4-4. Average particle size spectra of participate chloride
concentration, p'articulate iodine concentration, and participate
iodide/chloride ratio were obtained with an Andersen impactor.
Source: Moyers and Duce (1972).
4-13
-------
I is found mostly in the smallest particle fractions, stages F, E, D; the I/C1
ratio in particles is much higher than in seawater (3 x 10 ). Rainwater shows
a similar degree of I enrichment (Duce et al., 1965). Volatility of iodine-
containing vapors from the sea surface and reaction with fine aerosol particles
may account for the enrichment; primary injection of iodine-rich surface films
into the atmosphere may also contribute. The mechanisms of element fraction-
ation at the surface of the sea are still poorly understood and are the object
of intense research (Working Symposium, 1972; International Symposium, 1974).
4.2.4 Volcanic Emissions
Emissions from volcanic eruptions and fumaroles may contribute to the
global aerosol and atmospheric sulfur background and have climatic significance.
Materials injected into the stratosphere may persist for many months but then be
rapidly removed in the troposphere. The famous eruption of Krakatoa in 1883
injected enough dust into the stratosphere to cause brilliant sunsets thousands
of kilometers away and a global reduction of incoming solar radiation (Wexler,
1951a,b). In most industrialized continental areas, the contribution of vol-
canic emissions is negligible in comparison with regional-scale anthropogenic
emissions.
The average global emission rates of sulfur and several of its compounds
have been estimated by a number of investigators (Koyama et al., 1965; Holser
and Kaplan, 1966; Kellogg et al., 1972; Storber and Jepson, 1973; Friend, 1973;
Cadle, 1975). Table 4-3 (Granat et al., 1976) shows that the median of four
estimates of elemental sulfur emitted to the atmosphere is 3 Tg of sulfur per
year. Anthropogenic emissions to the atmosphere are estimated at 65 Tg of
sulfur per year, of which 98 percent is SOp and some 2 percent is sulfate
(Friend, 1973).
4-14
-------
TABLE 4-3. SULFUR EMITTED BY VOLCANOES AND FUMAROLES
(Tg/year)a
Reference
To air
To river runoff
To air and river runoff
(Koyama et al. , 1965)
(Holser et al. , 1966)
(Kellogg et al. ,
1972)
7
12
0.75
(Stoiber, 1973)
(Friend, 1973)
(Cadle, 1975)
3.5
2
3.75
4
5
7
al Tg = 1012 g.
Source: Granat et al. (1976).
The Agung eruption of 1963 injected into the stratosphere about 0.3 Tg of
sulfur, 0.2 Tg in the Northern Hemisphere and 0.1 Tg in the Southern Hemisphere
(Granat et al. , 1976). This injection led to the formation of submicron
sulfate aerosol with an estimated residence time of 6 months (Junge et al.
1961) to a year or longer (Castleman et al., 1974). This global burden is 30
times greater than the estimated pre-Agung stratospheric sulfur inventory of
0.01 Tg of sulfur (Granat et al., 1976). Explosive volcanic eruptions have
injected substantial, although smaller, amounts of sulfur into the stratosphere
several times since 1963 (Castleman et al., 1974).
Volcanic emissions of heavy metals into the lower atmosphere can be com-
parable to industrial pollution emissions on a regional basis. Metallic
constituents of particulate matter emitted from Mount Etna, Sicily, in June
1976 (Buat-Menard and Arnold, 1978) were compared with estimated pollution
emissions from bordering countries of the Mediterranean basin. Mount Etna was
4-15
-------
found to be the predominant source of Se, a source comparable in strength to
anthropogenic sources of Cd, Hg, Cu, and Zn, but insignificant in comparison
with automotive emissions of Pb. Consequently, in regions with active volcanism,
natural emissions of heavy metals must be considered in estimating total
metals emissions inventories. In othe'r regions, volcanism has little influence
on atmospheric chemical composition.
4.2.5 Biosphere Emanations
A significant natural source of sulfur, organic matter, and elemental
constituents of aerosols is the terrestrial and marine biosphere. Volatile,
reduced sulfur compounds are released to the atmosphere by microbiological
processes and may become oxidized to SCL and sulfate. The compounds released
include hydrogen sulfide (HLS), methyl mercaptan, dimethyl sulfide (DMS),
dimethyl disulfide, and carbon disulfide (Lovelock et al., 1972; Rasmussen,
1974; Lovelock, 1974). The releases of H^S and DMS may have the greatest
impact on the global sulfur cycle (Granat et al., 1976). The greatest H^S
concentrations occur where anaerobic surface conditions prevail, such as tidal
flats and shallow oceanic and freshwater bodies with reduced bottom conditions.
For DMS, natural soil sources may be predominant, with secondary releases from
senescent leaves, marine algae, and fresh leaves (Granat et al., 1976; Hitchcock,
1975). Terrestrial H-S and DMS releases on a global basis are estimated at 5
Tg of sulfur per year, and oceanic releases may be as high as 27 Tg of sulfur
per year (Granat et al., 1976), constituting a major component of the current
total of natural and anthropogenic sulfur releases to the atmosphere (Kellogg
et al., 1972). Low aerosol sulfur concentrations in the Southern Hemisphere
suggest that little sulfate is formed as a result of these emissions (Lawson
and Winchester, 1979a).
4-16
-------
Volatile organic compounds released from plants have been estimated at
200 Tg/year (Went, 1960). Isoprene derivatives such as terpenes, carotenoids,
and other compounds are believed to predominate. They are likely to be partial-
ly oxidized to form aerosol particles rather than completely oxidized to C0?
and HpO, resulting in blue haze and submicron condensation nuclei (Went, 1960;
Went et al., 1967; Rasmussen and Went, 1965; Schnell and Vali, 1972, 1973).
In cities, anthropogenic emissions form the bulk of the organic content of the
atmosphere; in the countryside, natural organic emissions may dominate.
Aerosol particles, including organic material and trace metals and nutrient
elements such as S, K, and P, are released from plants. Trace metals have
long been known to occur in fluids secreted by plants (Curtis, 1944). Radio-
tracer Sr is transferred from plant foliage to the atmosphere, presumably in
wax particles (Moorby and Squire, 1963) which may be affected by electric
fields (Fish, 1972). Transpiration causes the transfer of both cations and
anions to the atmosphere (Nemeryuk, 1970). Twenty-seven trace elements have
been identified in exudates from coniferous trees (Curtin et al., 1974). Zn
and Pb radiotracer experiments show that particles greater than 5 |jm in
diameter contain most of the metals released (Beauford et al. 1975, 1977). S,
K, and P have also been shown to be associated with tropical forests and to
occur in large aerosol particles (Lawson and Winchester, 1979b). The metal
content of plant-derived aerosols is so high that it has been suggested as an
indicator for geochemical prospecting (Barringer, 1977; Curtin et al., 1974).
4.2.6 Biomass Burning
The worldwide burning of biomass, especially in the clearing of tropical
forests for agriculture, contributes substantially to t^e global release of
trace gases and carbonaceous particulate matter. The gaseous flux of carbon
4-17
-------
to the atmosphere has been the focus of recent research because the release of
C02 from burning natural vegetation may be comparable in magnitude to that
from fossil fuels (Woodwell et al., 1978). Other estimates of CCL release
from forest burning are lower (Bolin, 1977; Baes et al., 1976). The range of
estimates is 0.5 Pg of C to more than Pg of C per year (Pg = 10 g) as
compared with 5-6 Pg of C from fossil fuels (Seiler and Crutzen, 1979). The
role of forests and grasslands in the global C0« balance, whether forested
land areas are decreasing through clearing, and whether the terrestrial biomass
is a net source or a net sink of CO^ are unresolved questions (Bolin, 1979).
Biomass burning results in the formation of charcoal, which cannot be
utilized by microorganisms and consequently persists in the environment,
perhaps for longer than 1000 years (Seiler and Crutzen, 1979). Because of its
low density, much of the charcoal is transported by flowing water and also
becomes airborne and transported by winds to the oceans. Carbon particles
account for some 0.02 to 0.1 percent of the dry weight of Pacific and Atlantic
Ocean sediments (Smith et al., 1973; Griffin and Goldberg, 1975), providing
evidence for long-range atmospheric transport, fallout, and settling of these
particles through the water column.
Chemical analysis of particulate matter from temperate forest burning
indicates approximately 50 percent benzene-soluble organic matter, 40 percent
elemental carbon, and only 10 percent mineral matter (Ryan and McMahon, 1976).
The estimated extent of biomass burning worldwide suggests an annual release
of roughly 200 to 450 Tg of particulate matter containing 90 to 200 Tg of
elemental carbon (Seiler and Crutzen, 1979). With a mean residence time of up
to a week for fine particulate matter generated near the earth's surface
4-18
-------
(Junge, 1963; Martell and Moore, 1974) and uniform mixing up to 1000 m in
altitude, carbon particle concentrations would average 4 ug/m . The estimated
average particulate carbon concentration attributed to biomass burning
indicates that soot carbon constitutes a significant fraction of the aerosol
loading in the atmosphere.
The fluxes of gaseous and particulate sulfur to the atmosphere may also
be due in part to biomass burning. Carbonyl sulfide (COS) has been detected
by gas sampling in a forest fire (Crutzen et al., 1979). Moreover, enhanced
regional concentrations of fine particulate S and K have been found in central
Brazil during the agricultural burning season (Lawson and Winchester, 1978).
As a consequence, estimates of pollution sources in such areas should include
an evaluation of biomass burning near the sampling sites.
4.3 MANMADE SOURCES OF SULFUR OXIDES AND PARTICULATE MATTER
4.3.1 Sulfur Oxide Emissions
4.3.1.1 National and Regional Overview—In 1977, stationary fuel combustion
and industrial processes accounted for about 97 percent of the anthropogenic
emission of sulfur oxides in the United States. The remaining 3 percent was
emitted primarily by motor vehicles (U.S. Environmental Protection Agency,
1978a). Table 4-4 and Figure 4-5 (U.S. Environmental Protection Agency,
1978a) indicate that in 1975, combustion of fossil fuels by electric utilities
and industrial facilities produced three-quarters of the total stationary
source emissions of sulfur oxides. Industrial processes such as smelting
accounted for the remaining one-quarter. This section discusses both national
and regional sulfur oxide emission trends.
4.3.1.1.2 National sulfur oxide emissions and fuel use: Historical trends.
Figure 4-6 (U.S. Environmental Protection Agency, 1978a; U.S. Department of
4-19
-------
TABLE 4-4. NATIONAL SULFUR OXIDES EMISSIONS INVENTORY SUMMARY1
Emissions source
Sulfur oxides, 10
metric tons/yr
% of Total
Utility fuel
combustion
Coal
Oil
Other
Total
Industrial fuel
combustion
Coal
Oil
Other
Total
Industrial processes
Primary
Metal
Petroleum and chemical
Other
Total
Commercial, residen-
tial, and miscellaneous
fuel use
Coal
Oil
Other
Total
Miscellaneous
Total
16.59
1.90
0.11
18.60
3.47
1.70
0.80
5.97
0.22
0.61
0.01
0.84
0.06
29.09
57.0
6.5
0.4
63.4
11.9
5.9
2.7
20.5
Transportation emissions are not included.
Source: U.S. Environmental Protection Agency (1978a).
4-20
-------
25%
COMMERCIAL,
RESIDENTIAL, «
AND
MISCELLANEOUS
FUEL USE
SOURCES
• 0.7% COAL
• 2.1% OIL
• 0.1% OTHER
20.5%
INDUSTRIAL
PROCESS SOURCES
• 115% PRIMARY METAL
• 55% PETROCHEMICAL
2.7% OTHER
63.9%
UTILITY FUEL
COMBUSTION
SOURCES
• 57.0% COAL
• 6.5% OIL
• 0.4% OTHER
125%
INDUSTRIAL FUEL
COMBUSTION SOURCES
• 7.0% COAL
• 4.6% OIL
• 0.9%
OTHER
07%
MISCELLANEOUS
SOURCES
Figure 4-5. Percentage of 1975 national sulfur oxide emissions is
shown by source category. Transportation emissions are not
included.
Source. U.S. Environmental Protection Agency (1978a).
4-21
-------
40
35
c« 30
2
O
I* 25
5 <
UJ Ul
f> £l 20
ui GO
2 ?
X
o
15
10
5
0
1940 1950 1960 1970
YEAR
1975 1985
40
35
v>
o
0 6
20 H m
O v>
2 m
- 15
OT
10-1
CO
1990
Figure 4-6. Bar chart represents nationwide estimates of sulfur oxide emissions are shown for 1940-90.
For 1940-75, transportation sources are not included.
Sources: U.S. Environmental Protection Agency (1978a).
U.S. Department of Energy (1979a).
-------
Energy, 1979a) illustrates the growth of SO emissions from 1940 through 1975.
Similarly, Table 4-5 and Figure 4-7 indicate the growth in the use of fossil
fuels. These trends are discussed below with respect to the three primary
types of fuel-using stationary sources: electric utilities, industrial facilities
and residential and commercial establishments.
4.3.1.1.2.1 Utilities. Electric utilities account for almost two-thirds of
the SO emissions in the United States. These emissions result from the use
of coal, oil, or gas to fire steam electric boilers. Table 4-5 shows that the
proportion of low-sulfur fuels (oil and gas) used in electric power plants
increased during the 1960's and then began to decrease; by 1977, coal accounted
for almost 60 percent of the fossil fuels used by electric utilities. The
increase in oil and gas use in the 1960's reflects the increased availability
and relatively low prices of oil and gas, the lower capital costs of oil- and
gas-fired boilers, and the environmental advantages of using oil and gas
rather than coal. However, the oil embargo of 1973-74 coupled with the sudden
fourfold increase in world oil prices reversed this trend.
The substitution of cleaner fuels and the imposition of environmental
controls on coal-fired power plants account, in part, for the slight reduction
in SO emissions in 1970-75 (Figure 4-6).
P\
4.3.1.1.2.2 Industrial facilities. Table 4-4 and Figure 4-5 (U.S.
Environmental Protection Agency, 1978a) show that industrial facilities account
for about one-third of total SO emissions. Most of these emissions result
A
from specific industrial processes, such as ore smelting.
Over one-third of the industrial SO emissions (and one-eighth of the
total national SO emissions) result from the combustion of fossil fuels to
A
produce process heat and steam for industrial facilities. Table 4-5 shows
4-23
-------
TABLE 4-5. U.S.
FOSSIL FUEL CONSUMPTION FOR STATIONARY SOURCES
BY YEAR BY CONSUMING SECTOR
(quadrillion btu)
Year
1960
Oil
Gas
Coal
Total
1970
Oil
Gas
Coal
Total
1975
Oil
Gas
Coal
Total
1977
Oil
Gas
Coal
Total
1990
Oil
Gas
Coal
Total
Electric
utility
0.56
1.78
4.22
6.56
2.11
4.02
7.45
13.58
3.23
3.25
9.28
15.76
4.03
3.29
10.27
17.59
3.17
0.51
22.05
25.73
Industry
2.32
4.48
4.68
11.48
2.33
9.58
5.05
16.96
2.47
8.64
3.65
14.76
2.71
8.24
3.82
14.77
3.73
9.98
6.64
20.35
Residential
and commercial
5.14
4.25
1.05
10.44
6.71
7.47
0.49
14.67
5.95
7.67
0.17
13.79
5.19
7.56
0.13
12.88
3.78
7.79
0.05
11.62
Total stationary
sources
8.02
10.51
9.95
28.48
11.15
21.07
12.99
45.21
11.65
19.56
13.10
44.31
11.93
19.09
14.22
45.24
10.68
18.28
28.74
57.70
Historical fuel consumption data (1960-77) taken from Federal Energy Data
System (FEDS) (DOE/EIA-0192), July 1979. The 1990 fuel use is based on the
Series C projections of the Department of Energy (it assumes moderate supply
and demand). There are minor discrepancies in the data based used to compile
the 1960, 1970, and 1975 data and the 1977 and 1990 data due to differences
in the definitions of fuel categories (such as inclusion or exclusion of natural
gas liquids in the use of natural gas).
Source: U.S. Department of Energy, (1979a). (The model utilized in these
emission projections is the Regional Emission Projection Systems.
air
REPS
is documented in "An Air Emissions Analysis of Energy Projections for
the Annual Report to Congress," a memorandum by Edward Pechan dated
September 1978).
U.S. Department of Energy (1979b).
4-24
-------
1960 1975 1990
ELECTRIC
UTILITIES
1960 1975 1990 1960 1975 1990
INDUSTRIAL
RESIDENTIAL
AND COMMERCIAL
yi
b
m
z
o
cr
O
<
o
z"
o
E
I
60
55
50
45
40
35
30
25
20
15
10
5
0
OIL AND GAS
1960 1975 1990
ALL STATIONARY SOURCES
Figure 4-7. Bar chart represents stationary-source fossil fuel con-
sumption by consuming sector for 1960, 1975, and 1990 illustrates
data presented in Table 4-5.
4-25
-------
that use of fossil fuels by industry grew less rapidly than utility fossil
fuel use in the 1960's and declined between 1970 and 1975. This decline
appears to be the result of two factors: massive conservation efforts by
industry in response to sharply higher oil and gas prices, and the effects of
the 1974 recession on large fuel-using industries.
Table 4-5 and Figure 4-7 show that although in 1975 total fuel con-
sumption by industry and utilities was about the same, industry burned much
less coal. This difference accounts for the relatively smaller share of total
SO emissions from industrial fuel use.
/\
4.3.1.1.2.3 Residential and commercial establishments. The residential
and commercial fuel-using sector accounts for only about 3 percent of total
SO emissions. But unlike utility and industrial SO emissions, residential
x\ /\
and commercial emissions are released near the ground and thus can signifi-
cantly affect the air quality in the immediate area. Also, most of the SO^
emissions from residential and commercial space heating occur in areas of high
population density. Finally, these emissions are highest during the winter
heating season.
On the other hand, low-sulfur oil and natural gas are burned for space
heating by the commercial sector. The SOp emissions are usually much lower
from this sector than from the other S0? sources. Assuming trends continue,
Table 4-5 shows that coal use in this sector may be almost eliminated by 1990.
Government policies affecting the pricing and allocation of heating oil and
natural gas have promoted the use of these low-sulfur fuels.
4.3.1.1.3 Future trends. Based on data from the National Environmental Data
System (NEDS) and projections of fuel use by the Energy Information Admini-
stration (EIA), national total SO emissions are expected to decline through
1985 and increase thereafter. This forecast is illustrated in Figure 4-6.
4-26
-------
The EIA fuel use projections generally assume reduced overall energy
growth rates, greater use of coal, reduced dependence on oil imports, contin-
ued increases in energy prices, and more energy use by electric utilities.
Rising oil costs will tend to increase coal use and, hence, SO emissions.
The predicted decline in S0x emissions until 1985 assumes stricter New Source
Performance Standards (NSPS) for new power plants and some industrial facilities
and retirement of older facilities. According to DOE projections, this trend
may be counteracted by the use of fossil fuel, especially coal, which is
expected to increase total SO emissions after 1985 (Figure 4-6).
4.3.1.1.3.1 Utilities. The projected increased coal use by electric
utilities is expected to be the result of reduced availability of oil and gas
and sharply higher prices after 1974. In late 1978, Congress enacted the
Powerplant and Industrial Fuel Use Act, which prohibits new oil- and gas-fired
boilers and requires existing utilities to phase out gas use by 1990. This
legislation is expected to further restrict oil and gas use by utilities.
Thus, coal, which accounted for 60 percent of all utility fossil fuel use in
1977, is forecasted by the Department of Energy to increase over 85 percent by
1990, as shown in Table 4-5.
In 1960, electric utilities accounted for only about 23 percent of the
fossil fuel used by stationary sources; by 1990 this share is estimated to
increase to about 49 percent.
4.3.1.1.3.2 Industrial facilities. Because of environmental constraints
and increased availability of natural gas, industrial coal use is projected by
DOE to decline during the 1960's. By 1990, a resurgence of coal use for
industrial boilers and process heat is expected. The Fuel Use Act prohibits
large new industrial boilers from using oil or gas and provides the Department
4-27
-------
of Energy with the authority to extend this prohibition to existing coal-capable
units of future nonboiler combustors. Nevertheless, coal is not anticipated to
account for more than one-third of the fossil fuel use in industrial facilities
by 1990.
4.3.1.1.3.3 Residential and commercial establishments. Conservation,
use of solar power, and electrification are expected to cause decrease in
fossil fuel use in this sector. Coal use by residences and businesses is
expected to diminish by 1990. Therefore, SO emissions from this sector are
not expected to increase over the 1980's.
4.3.1.1.4 Regional SO emissions.
4.3.1.1.4.1 Current emissions distribution. SO emissions vary greatly
throughout the country. As is evident in Figure 4-8, (U.S. Environmental
Protection Agency, 1978a), two of the nine EPA regions—the South Atlantic and
Midwest regions—accounted for more than half of all SO emissions nationwide
in 1975. Figure 4-9 (U.S. Environmental Protection Agency, 1979) indicates
the density of SO emissions on a county basis. This variation is the result
of three factors.
(1) The variation in energy consumption, which is primarily a function
of population density and industrial energy use.
(2) The type of fuel used. Areas with a large portion of hydroelectric,
nuclear, or low-sulfur fossil fuel use have lower emissions than
areas more dependent on coal or high-sulfur oil.
(3) The presence of large SO -emitting industrial facilities, such as
y\
primary metal processors.
4-28
-------
-F*
I
ro
VIRGIN
ISLANDS
ALASKA
HAWAII
II III
IV
VI VII VIII IX
SOX emissions
(106 metric tons) 0.59
Percent of U.S. total 2.0
1.40 4.04 5.92 9.64 2.28 1.59 0.72 2.63 0.31
4.8 13.9 20.3 33.1 7.8 5.5 2.5 9.0 1.1
Figure 4-8. Distribution of 1975 sulfur oxide emissions and percentages of U.S. total sulfur oxide
emissions are shown for the EPA regions.
-------
GO
o
PREPARED BY:
MONITORING AND REPORTS BRANCH
MONITORING AND DATA ANALYSIS DIVISION
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK. N.C. 27711
BASED ON DATA FROM NATIONAL EMISSION
BASE YEAR 1975.
SULFUR OXIDE.
TONS PER SQUARE MILE
Figure 4-9. Sulfur oxide emission density by county.
-------
Electric utility fuel combustion accounted for about 64 percent of total
SO emissions nationwide in 1974. On a regional basis, SO emissions from
X ^
power plants vary from 5.2 percent in Region X (Northwest) to more than 80
percent in Regions IV and V. These variations, illustrated in Table 4-6, are
also affected by the presence of industrial process sources.
4.3.1.1.4.2 Industrial process sources. In 1975, industrial process
sources account for about one-fifth of the national SO emissions from station-
J\
ary sources. Specifically, the primary metals industry contributes nearly 12
percent, and the petroleum and chemical industries together contribute about 6
percent of the national SO emissions. Most SO emissions in EPA Regions VI,
IX, and X (61.5, 86.6, and 56.1 percent, respectively) are from industrial
process sources. Industrial process sources account for less than 7 percent
of the SO emissions in Regions I, IV, and V.
/\
As seen in Figure 4-10 (U.S. Environmental Protection Agency. 1974), most
of the primary copper, lead, and zinc smelters are in EPA Regions VI, IX, and
X. Smelters are large emitters of SO .
/\
El Paso County in Texas, southeastern Arizona counties, and counties in
western Montana have high SO emissions from smelters.
f\
Petroleum-refining capacity is greatest in EPA Regions V, VI, and IX,
especially in Texas, Louisiana, California, and Illinois. Elemental sulfur
production capacity is concentrated in Texas and Louisiana. Sulfuric acid
plants are dispersed widely throughout the Nation, but the majority of larger
plants are located in Florida. The literature does not address the location
of other SO^ sources of the chemical industry.
4-31
-------
CH4BI3 3-20-80
TABLE 4-6.
PERCENTAGE OF EPA REGIONAL SO EMISSIONS BY SOURCE CATEGORY
X
CO
ro
EPA region
Emissions source
Utility fuel
combustion
Coal
Oil
Other
Total
Industrial fuel
combustion
Coal
Oil
Other
Total
Industrial processes
Primary metal
Petroleum and chemical
Other
Total
Commercial, residential
and miscellaneous fuel
use
Coal
Oil
Other
Total
Miscellaneous
I
7.6
37.5
<0.1
45.1
0.9
35.0
0.2
36.1
0.0
0.5
1.2
1.7
0.2
16.2
<0.1
16.4
0.6
II
24.
32.
0.
58.
7.
12.
0.
20.
0.
4.
7.
11.
0.
8.
<0.
9.
0.
9
8
3
1
9
7
2
8
3
3
0
6
9
0
1
0
5
III
61.7
8.1
0.6
70.5
9.7
4.6
0.5
14.8
4.4
4.6
2.3
11.3
1.0
2.4
<0.1
3.4
0.1
IV
75.0
7.0
0.3
82.3
4.3
5.0
0.4
9.7
0.5
3.5
2.5
6.5
0.4
1.0
<0.1
1.4
0.1
V
79.6
1.0
<0.1
80.6
10.0
1.7
0.8
12.5
1.5
1.7
1.5
4.7
1.1
1.0
<0.1
2.1
0.1
VI
11.6
6.2
2.2
20.0
5.4
6.9
3.5
15.8
20.7
34.7
6.1
61.5
<0.1
2.1
<0.1
2.2
0.5
VII
69.0
1.6
0.4
71.0
8.2
0.9
0.3
9.4
9.0
2.9
6.0
17.9
0.6
0.7
<0.1
1.4
0.2
VIII
28.0
0.7
0.2
28.9
5.8
6.5
1.5
13.7
37.1
13.2
3.2
53.5
1.0
2.8
0.1
3.8
0.1
IX
2.1
6.5
<0.1
8.6
0.1
2.5
0.7
3.2
80.5
5.0
1.1
86.6
<0.1
1.2
0.1
1.3
0.2
X
4.2
0.8
0.2
5.2
6.0
15.2
3.7
24.9
39.1
8.4
8.6
56.1
0.2
13.3
0.1
13.6
0.2
transportation sources are not included.
Source: U.S. Environmental Protection Agency (1978a).
-------
-pi
I
co
Figure 4-10. Locations are shown for existing primary metal smelters (midyear 1973).
Source: U.S. Environmental Protection Agency (1974).
-------
4.3.1.1.4.3 Future regional emission patterns. In the 1980's, S0x
emissions are expected to decline in Regions III, IV, and V and increase in
other regions, particularly Region VI (see Table 4-7). Regional SQy emissions
are expected to vary in the future because of differences in regional energy and
economic growth rates. Furthermore, areas now relying on coal are expected to
have lower SO emissions as power plants and industrial facilities that emit
/\
large amounts of SO are retired or fitted with high-efficiency emission controls.
/\
Finally, areas with little previous coal use are expected to convert from oil
and gas to coal and therefore increase SO emissions.
J^.
4.3.1.2 SO Source Emissions Characteristics—Stationary sources may emit SO
in many forms. By volume, over 90 percent of total national SO emissions are
in the form of SQy. The balance consists of primary sulfate, made up of gaseous
sulfur trioxide (SO.,) and sulfuric acid (H^SO.), HLSO. mist, and particulate
sulfates, which include metallic sulfates (MSO., where M is a metal or ammonium
ion) and H2SO. that has been adsorbed onto particles.
Several factors affect the quantity and composition of SO emissions from a
particular source. Among these factors are the source type, operating conditions,
fuel characteristics, and type of emissions control equipment.
Emissions of S0? from major SO sources are documented in Compilation
£-• f\
of Air Pollutant Emission Factors (U.S. Environmental Protection Agency, 1977).
There is comparatively little information on primary sulfate emissions from
major SO sources. Research on the composition of primary sulfate emissions
from utility and industrial boilers has only begun within the last few years.
There is little information on the composition of SO emissions from residential
and commercial fuel combustion sources and industrial process sources.
4-34
-------
TABLE 4-7. CURRENT AND PROJECTED SULFUR OXIDES
EMISSIONS BY EPA REGION3
(10 metric tons)
I
II
III
IV
V
VI
VII
VIII
IX
X
EPA region
New England
New York/New Jersey
Mid-Atlantic
South Atlantic
Midwest
Southwest
Central
North Central
West
Northwest
1975
0.59
1.40
4.04
5.92
9.65
2.28
1.59
0.72
2.63
0.31
1985
0.69
1.66
3.68
5.04
7.53
3.74
1.78
0.69
2.97*
0.33
1990
0.85
2.23
3.92
5.80
8.59
5.47
1.96
0.83
3.02*
0.39
Series C is based on moderate supply and demand assumptions. (The above
estimates ignore potential decreases of emissions from smelters resulting
from controls, e.g., NSPS, SIPs, etc.)
*The DOE model may not account for reductions in sulfur dioxide emissions
expected within the next few years.
Sources: U.S. Environmental Protection Agency (1978a); U.S. Department of
Energy (1979a).
4-35
-------
4.3.1.2.1 Utility and industrial fuel combustion sources.
4.3.1.2.1.1 Coal. The amount of uncontrolled SOp emissions from a
coal-fired utility or industrial boiler depends primarily on the amount of
fuel burned, its sulfur content, and other characteristics of the fuel. Coal
is a slow-burning fuel with a high ash content (Homolya and Cheney, 1978).
Its low-flame temperatures lessen the production of SO, from SO- and atomic
oxygen. Since coal ash is basic, some of the H^SO. formed during combustion
is neutralized by the ash (typically 5 to 15 percent by weight). Although the
sulfur content of coal (ranging from 0.2 to 7 percent by weight) is high, some
of the coal sulfur is trapped in the bottom ash and is not emitted to the
atmosphere.
In the combustion of most coals (most commonly, bituminous), about 95
percent of the coal sulfur is emitted as gaseous SOp and SO-; over 90 percent
of the coal sulfur is converted to S02 (U.S. Environmental Protection Agency,
1977). The balance of the coal sulfur may be emitted as particulate sulfates
or may combine with the slag or ash in the furnace and be disposed of as a
solid waste. Most of the gaseous SO- is hydrated to gaseous or aerosol H-SO.
before exiting the boiler stack (Homolya and Cheney, 1979).
A study by Homolya and Cheney (1978) showed that 0.9 to 3.5 (average 2.1)
percent by weight of SO emissions from coal-fired boilers occurred as primary
J\
sulfate. The coals used in this study were 1.7 to 3.6 percent sulfur by
weight.
Lignite is being used increasingly where it is plentiful at relatively
low cost. The alkali content (especially sodium) of lignite ash has a major
effect on the amount of coal sulfur retained in boiler ash deposits. A
high-sodium lignite may retain over 50 percent of the available sulfur in the
4-36
-------
boiler ash, while a low-sodium lignite may retain less than 10 percent of the
available sulfur in the boiler ash (U.S. Environmental Protection Agency,
1977).
Other studies show that excess boiler oxygen enhances the emission of
primary sulfate from coal-fired boilers (Homolya and Cheney, 1978; Bennett and
Knapp, 1978). The excess oxygen increases oxidation of SO,, to SO., and then
H?SO.. Industrial boilers generally use higher excess oxygen levels than do
utility boilers.
Dirty equipment may increase primary sulfate emissions from coal-fired
boilers because boiler deposits catalyze the oxidation of SO,, to sulfates.
The frequency of soot-blowing operations can affect the emissions of primary
sulfate from boilers with catalytically active deposits. During such operations,
sulfate boiler deposits may be emitted to the atmosphere.
Emission control equipment may have a significant impact on the compo-
sition and concentration of SO from coal-fired boilers. Electrostatic precipi-
tators and flue gas desulfurization systems are used to abate particulate and
SO emissions.
A
Electrostatic precipitators provide a high degree of particulate control.
They usually have little overall effect on SO,, emissions but may reduce total
sulfate emissions by 50 percent or more (Homolya and Cheney, 1978). Some
tests have shown variability in SO,, emissions from boilers equipped with "hot
side" electrostatic precipitators. Conditions in these units may promote
sulfate formation. Arcing can cause localized "hot spots" in which the high
temperatures enhance sulfate formation, and corona discharges can produce
ozone (03), which can oxidize S02 to S03 (McCurley and DeAngelis, 1978).
4-37
-------
In the absence of particulate control devices, the mass median diameter
of coal fly ashes is typically 8 to 30 urn (Natusch, 1978). Fly ashes emitted
from electrostatic precipitators have a mass median diameter of 0.5 urn to 2 urn
(Natusch, 1978). The collection efficiency for particulate sulfates is less
than the collection efficiency for total particulate matter. Accordingly, the
ratio of sulfate particles to total particles is higher in the electrostatic
precipitator outlet stream than in the inlet stream (Homolya and Cheney, 1978;
Bennett and Knapp, 1978; McCurley and DeAngelis, 1978). This phenomenon is
referred to as sulfate enrichment of fine particulate matter.
Flue gas desulfurization systems may be designed to remove SOp with
greater than 90 percent efficiency. Primary sulfate emissions removal from
these systems appears to be highly dependent on design practices. Tests on a
wet scrubber system showed a primary sulfate emissions removal efficiency of
nearly 30 percent (Homolya and Cheney, 1979). However, several individual
test results in the same study showed an increase of primary sulfate emissions
across the wet scrubber. Two causes of these primary sulfate emissions
increases may be penetration of the scrubber demisters by gaseous HpSO,,
forming H-SO. aerosol, and by scrubber liquor reentrainment. Like electro-
static precipitators, flue gas desulfurization systems are not generally
efficient at collecting fine particulate matter.
Fuel modification can reduce SO emissions. Coal may be modified by
washing to reduce sulfur and ash content. Liquefaction and gasification of
coal into new fuels can result in significant reduction of sulfur fuel content
and therefore SO emissions.
4.3.1.2.1.2 Oil. After coal, oil is the largest contributor to S0x
emissions from fuel combustion sources. The amount of oil burned and its
4-38
-------
characteristics are the primary factors affecting uncontrolled SO emissions
from oil-fired boilers.
In contrast to coal, oil is a fast-burning, low-ash fuel (typically 0.01
to 0.20 percent). Its high flame temperatures exacerbate the formation of
SO., H?SO., and particulate sulfates from SOp (Homolya an.d Cheney, 1978). Oil
may contain metals, especially vanadium, which catalyze the formation of SO,,
in the combustion process. Studies show that increasing the vanadium content
of oil increases the formation and emission of primary sulfate (Homolya and
Cheney, 1978; Bennett and Knapp, 1978; Dietz et al., 1978). The mass mean
diameter of uncontrolled particulate matter from oil is smaller than that from
coal, ranging from 0.3 to 10 urn (Scinto, 1979).
Oil-fired boilers generally convert over 90 percent of available fuel
sulfur to SO,,. Tests show that 2 to 12 (average 6.5) percent by weight of SO
£ A
emissions from oil combustion occur as primary sulfate (Homolya and Cheney,
1978). The tests were conducted with a variety of boilers and fuels with
different amounts of sulfur and vanadium.
Excess oxygen enhances primary sulfate emissions from oil-fired boilers
(Homolya and Cheney, 1978; Bennett and Knapp, 1978; Dietz et al., 1978). Fuel
additives (usually containing magnesium oxide) are effective in reducing
primary sulfate emissions from oil-fired boilers. Ideally, the fuel additives
minimize the formation of SO, and can react with or adsorb SO., and HpSO. and
retain them in the bottom ash or combine them with particles to be collected
by an electrostatic precipitator.
Oil-fired utility and industrial boilers can employ the same types of
emissions control equipment as coal-fired boilers, although in many cases they
may have only a mechanical particulate collector. Oil-fired SO emissions are
4-39
-------
affected by electrostatic precipitators and flue gas desulfurization systems
in much the same way as coal-fired SO emissions. Electrostatic precipitators
J\
have a lower particulate sulfate collection efficiency in oil-fired boilers
than they do on coal-fired boilers (Leavitt et al., 1978). Particles from
oil-fired boilers are smaller than those from coal-fired boilers and therefore
collected less effectively. Factors affecting flue gas desulfurization
systems on oil-fired boilers are similar to those affecting coal-fired boilers.
Tests of uncontrolled SOp emissions from coal- and oil-fired utility and
industrial boilers compare favorably with U.S. EPA (1977) emission factors
(McCurley and DeAngelis, 1978; Homolya et al., 1976). Primary sulfur emissions
from oil-fired boilers may be 3 to 10 times greater than those from coal-fired
boilers, given the same sulfur content (Homolya and Cheney. 1978). A recent
study reported that power plants converted from coal to oil emitted more
particulate sulfate and sulfur acid after conversion, even though the coal
had a higher sulfur content and the same heating value as the oil (Homolya and
Cheney, 1979).
4.3.1.2.2 Residential and commercial space heating. Residential and commercial
space heaters mainly use natural gas and oil. Coal and wood are fuels used to
a lesser degree for space heating. The Portland Aerosol Characterization
Study (PACS) found that home heating was a major contributor to total suspended
particle levels in the Portland Airshed (Cooper and Watson, 1979). SO
/\
emissions from home heating may have been accordingly high.
Equipment design, operating parameters, and fuel characteristics affect
the characteristics of SO emissions from space heating. Proper burner adjust-
ment, combustion modifications, and fuel additives may all minimize SO
emissions. Space heaters do not usually have any emissions controls.
4-40
-------
SOp emissions from gas-fired and oil-fired residential sources correlate
well with SOp emission factors in AP-42 (U.S. Environmental Protection Agency,
1977). Some tests have indicated S03 emission rates from an oil-fired source
of nearly three times the expected rate derived with AP-42 factors (Surprenant
et a!., 1979). Further study is needed to determine whether these values are
typical S03 emissions rates from oil-fired space heaters. The existing data
base appears to characterize SO^ emissions from gas, oil, and coal residential
combustion sources, but more information is needed to characterize primary
sulfate emissions from oil-fired and coal-fired space heaters.
4.3.1.2.3 Industrial process sources. The primary metals industries and
the petroleum and chemical industries are major industrial process sources of
SO emissions. Mineral products industries account for a smaller portion of
these emissions.
4.3.1.2.3.1 Primary metals industries. In the primary metals industries,
copper, lead, and zinc smelters are potentially the largest SO emitters.
/\
Most of the sulfur in unprocessed ores is converted to S0? in the smelting
processes (U.S. Environmental Protection Agency, 1977). A relatively small
amount of the sulfur is emitted as particulate sulfate and sulfuric acid.
Factors affecting SO,, emissions from smelters are the amount of raw ore
processed, the sulfur content of the ore, the process configuration, the
degree of sulfur removal for each process step, and the degree of SO^ emission
control used. The bulk of S02 is formed in the roasting, smelting, sintering,
and converting processes of the smelter industry (U.S. Environmental
Protection Agency. 1974).
Sources of fugitive S0? emissions include roaster transfer points, furnace
feed and discharge areas, and converter rollout. In copper smelters, fugitive
4-41
-------
SOp emissions of up to 12 percent of the total generated SOp have been reported
(U.S. Environmental Protection Agency, 1974).
Weak SOp off-gases are usually vented to the atmosphere. Single-contact
sulfuric acid plants are the primary method of treatment for strong SO,, off-
gases. They may recover up to 97 percent of the off-gas SOp as sulfuric acid.
Other SO- control methods with potentially higher removal efficiencies than
single-contact sulfuric acid plants are double-contact sulfuric acid plants
and the dimethyl aniline (DMA) absorption process.
In the past, strong SOp gas streams were controlled because of the market-
ability of the recovered sulfur products. To ensure compliance with local and
national ambient air quality standards, new SOp control equipment has been
installed or is being planned for many smelters (especially copper smelters)
in the United States. New and planned SOp control strategies include the
establishment of sulfuric acid and liquid S0? plants and the installation of
ducting and hoods to minimize fugitive SOp emissions from transfer points
(U.S. Environmental Protection Agency, 1974).
4.3.1.2.3.2 Petroleum industry. The major sources of SO emissions in
/\
the petroleum industry are catalytic cracking and sulfur-recovery processes
and off-gas flares (U.S. Environmental Protection Agency. 1977; Dickerman et
al., 1977). SOp emissions from the catalytic cracking process occur during
the catalyst-regeneration step.
Major sour-gas streams are usually treated in a sulfur plant. Most
sulfur plants utilize a modified Claus process for multistage oxidation of
hydrogen sulfide to elemental sulfur. The sulfur recovery efficiency of these
plants ranges from 92 to 97 percent, depending on the number of catalytic
4-42
-------
stages. Tail gas is usually incinerated, and most of the remaining sulfur
species are oxidized to S02- Some plants have installed tail gas cleanup
systems to further reduce SOp emissions. These units can recover sulfur with
up to 99.8 percent efficiency.
Minor off-gas streams and recovered vapors are often burned in flares.
Most of the sulfur species present in these vapors are oxidized to S0?.
4.3.1.2.3.3 Chemical production industries. Chemical process industries
which emit significant amounts of SO are explosives manufacturing, sulfuric
/\
acid plants, and elemental sulfur plants.
The manufacture of TNT and nitrocellulose explosives produces emissions
of S02 and sulfuric acid mist. A major raw material in the? production of
these explosives is sulfuric acid. Sulfuric acid concentrators, sellite
exhaust, and incinerators are the major SO sources in these processes. SO
A /N
emissions may vary considerably with the efficiency of the process and the
operating conditions (U.S. Environmental Protection Agency, 1977).
Sulfuric acid is manufactured primarily by the contact process. The
three types of raw materials used by sulfuric acid plants are elemental sulfur,
spent acid and hydrogen sulfide, and sulfide ores and smelter gases.
The amount of SO,, emissions in acid-plant gases is an inverse function of
the sulfur conversion efficiency of the process (U.S. Environmental Protection
Agency, 1977). Sulfuric acid mist is generated by SO,, absorbers. The quantity
and size distribution of the acid mist are dependent on the type of sulfur
feedstock used, the strength of the acid produced, and the conditions in the
absorber. Electrostatic precipitators and fiber mist eliminators can reduce
sulfuric acid mist emissions by up to 99 percent.
4-43
-------
4.3.2 Particulate Matter Emissions
4.3.2.1 Major Manmade Sources of Particulate Matter—The 1975 National Emissions
Data System (NEDS) data base (Table 4-8) shows a nationwide participate
emissions total of about 12.5 x 10 metric tons. As shown in Table 4-9, the
major source categories are stationary fuel combustion, mineral products,
primary metals, and land vehicles, which together account for just over 80
percent of the national total. Table 4-10 summarizes data on the characteristics
of particles from certain major sources.
4.3.2.1.1 Geographic distribution. The following discussion, tables, and
figures are directed at characterizing the nature and extent of particulate
emissions on the basis of region, State, major source category, and population
exposure. Table 4-11 summarizes total emission data geographically and includes
additional parameters such as population, land area, and emission densities.
Population and emission densities are also illustrated in Figures 4-11 and
4-12, respectively. Several important points are worth noting with respect to
Table 4-11 and Figures 4-11 and 4-12:
(1) EPA Regions IV and V account for nearly 50 percent of total U.S.
particulate emissions; Regions III, IV, V, and VI total about 72
percent.
(2) Region V is the single greatest contributor (28.1 percent), and Ohio
accounts for just under 40 percent of the Region's total and about
11 percent of the U.S. total.
(3) Except in Regions I, II, and IX, emissions appear to correlate with
population.
4-44
-------
TABLE 4-8. COMPARISON OF CALCULATED AND INVENTORIED
EMISSIONS FOR THE UNITED STATES, 1975
(103 metric tons/year)
Source category
Fuel combustion (point and area)
Residential
Electric generation
Industrial
Commercial and institutional
Other
Industrial processes (point)
Chemical manufacturing
Food and agriculture
Primary metals
Secondary metals
Mineral products
Petroleum industry
Wood products
Evaporation
Other processes
Solid waste disposal (point and area)
Municipal incineration
Residential
Commercial and institutional
Industrial
Other
Transportation (area)
Gasoline vehicles
Diesel vehicles
Aircraft
Vessels
Miscellaneous (area)
Forest fires
Structural fires
Slash burning
Grand total (point and area)
Calculated
by EPA
5,262
91
3,901
1,089
181
0
6,985
181
1,361
1,270
91
3,720
91
272
0
0
454
91
181
91
91
0
1,179
774
272
91
45
454
272
91
91
14,334
1975 NEDS
5,033
182
2,784
1,861
193
13
5,647
207
593
989
126
3,055
119
271
17
271
446
54
227
49
81
34
1,058
677
262
99
20
299
159
1
140
12,489
Section 4.3.2.2 contains more recent data on emissions from transportation
sources.
Source: U.S. Environmental Protection Agency (1978a).
4-45
-------
TABLE 4-9a. MAJOR NATIONAL SOURCES OF PARTICULATE MATTER
(SOURCE: 1975 NEDS INVENTORY)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Source category
Stationary fuel combustion
Mineral products
Primary metals
Land vehicles
Food and agriculture
Solid waste disposal
Wood products
Chemical manufacturing
Forest fires
Slash burning
Emissions
(103 metric
tons/yr)
5033
3055
989
939
593
446
271
207
159
140
Percent
of U.S.
total
40.3
25
7.9
7.5
4.7
3.6
2.2
1.7
1.3
1.1
TABLE 4-9b. MAJOR NATIONAL SOURCES OF PARTICULATE
(SOURCE: 1977 SURE INVENTORY)
Cumulative
percent
of U.S. total
40
65
72.9
80.4
85.1
88.7
90.9
92.6
93.9
95
MATTER
1.
2.
3.
4.
5.
6.
7.
8.
Source category
Stationary fuel combustion
Mineral products
Food and agriculture
Primary metals
Chemical manufacturing
Wood products
Solid waste disposal
Secondary metals
Emissions
(103 metric
tons/yr)
3007
2211
608
585
203
180
128
108
Percent
of U.S.
total
39
29
7.9
7.6
2.6
2.3
1.7
1.4
Cumulative
percent
of U.S. total
39
68
75.9
83.5
86.1
88.4
90.1
91.5
4-46
-------
TABLE 4-10. SUMMARY OF MAJOR SOURCE CATEGORIES AND PARTICLE CHARACTERIZATION DATA
Source category
Industrial Processes
- Mineral Products
Cement
Stone
Asphalt concrete
- Primary Metals
Iron
Aluminum
Steel
Coke
Copper
- Food/Agriculture
Fuel Combustion
- Electric Generation
- Industrial
- Commercial/Institutional
- Residential
Particulate emissions a.
(103 metric tons) Particle size data* Trace element components
1975 NEDS
5,647
3,055
-900
-642
-150
989
-247
-198
-188
-76
-37
593
5,033
2,784
1,861
193
182
1977 SURE
4,347
2,211
312
837
157
585
191
54
59
60
608
3,007
1,670
1,274
63
-
25% I, 75% C (before control)!
40% F (after control) Similar to parent material, i.e.,
— . limestone, clay, shale, gypsum,
100% C (before control)! sand, etc.
30-60% F (after control)
100% I, 75% F{ Fe, Si, Al , Ca, Mg, Zn, Mn, Pb, P
100% I. 60% Ft Fe, Si, Al , Ca, Mg, Mn, Zn, Pb, Ni , Cr, P
100% It . F, Cl, Al
100% I, 75% Ft As, Be, Cd, Cr, Co, Pb, Ni , Se, POM
100% I, 50% Ft As, Pb, Cd, Se, Ag
25% I, 75% c| Grain dust
50% I, 50% C
50% F, 50% C Cl, Fe, Ti, F, V, Ni, B, Mn, As
100% F
100% F
*F = Fine (<2
I = Inhalable (<15
C = Coarse (>15 urn)
See detailed breakdown for fuel
combustion in Section 3
(•Data obtained from FPEIS
-------
TABLE 4-11. STATE-BY-STATE LISTING OF TOTAL PARTICULATE EMISSIONS,
POPULATION, AND DENSITY FACTORS
Region and State
Population
1000's
Total area,
mr
Total emissions,
103 metric tons
Population
density,
people/mi2
Emission
density,
tons/mi2
State
emissions,
percent of U.S.
00
Region 1
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Total
Percent of U.S.
3,117
1,070
5,809
822
927
476
12,221
5.6
5,009
33,215
8,257
9,304
1,214
9,609
66,608
1.8
32.7
50.9
87.1
13.8
8.7
6.3
199.5
1.6
622
32
704
88
764
50
183
6.5
1.5
10.5
1.5
7.2
0.66
3.0
0.26
0.41
0.70
0.11
0.07
0.05
Region II
New Jersey
New York
Puerto Rico
Virgin Islands
Total
Percent of U.S.
7,336
18,084
2,712
62
28,194
12.9
7,836
49,576
3,435
133
60,980
1.7
118.3
366.
106.
4.1
594.8
4.8
936
365
790
466
462
15.1
7.4
30.9
30.8
9.75
0.95
2.93
0.85
0.03
Region III
Del aware
582
District of Columbia 702
Maryland 4,144
2,057
67
10,577
27.2
6.4
119. 2
283
10,478
392
13.2
95.5
11.3
0.22
0.05
O. 95
-------
TABLE 4-11 (continued)
Region and State
Pennsylvania
Virginia
West Virginia
Population
1000 's
11,862
5,032
1,821
Total area,
mi2
45,333
40,817
24,181
Total emissions,
103 metric tons
655.3
395.7
325.5
Population
density,
people/mi2
262
123
75
Emission
density,
tons/mi2
14.5
9.7
13.5
State
emissions ,
percent of U.S.
5.25
3.17
2.61
Total
Percent of U.S.
Region IV
Alabama
Florida
Georgia
Kentucky
Mississippi
North Carolina
South Carolina
Tennessee
Total
Percent of U.S.
Region V
Illinois
Indiana
Minnesota
Michigan
Ohio
24,143
11.1
35,369
16.3
11,229
5,302
3,965
9,104
10,690
123,032
3.4
3,665
8,421
4,970
3,428
2,354
5,469
2,848
4,214
51,609
58,560
58,876
40,395
47,716
52,586
31,055
42,244
383,041
10.6
56,400
36.291
84,068
58,216
41.222
1,529.3
12.2
482.2
252.2
172.7
655.4
167.7
241.0
177.5
347.8
2,442.5
19.5
791.6
530.7
264.0
283.3
1,367.3
196
71
144
84
85
49
104
92
100
92
199
146
47
156
259
12.4
8.3
4.3
2.9
16.2
3.5
4.6
5.7
8.2
6.4
14.0
14.6
3.1
4.9
33.2
3.
2.
1.
5.
1.
1.
1.
6.
4.
2.
2.
43
02
38
25
34
93
42
2.78
34
25
11
27
10.95
-------
TABLE 4-11 (continued)
Ul
o
Region and State
Wisconsin
Total
Percent of U.S.
Region VI
Arkansas
Louisiana
New Mexico
Oklahoma
Texas
Total
Percent of U.S.
Region VII
Iowa
Kansas
Missouri
Nebraska
Population
1000 's
4,609
44,899
20.6
2,109
3,841
1,168
2,766
12,487
22,371
10.3
2,870
2,310
4.778
1,553
Total area,
mi2
56,154
332,351
9.2
-f
53,104
48,523
121,666
69,919
267,338
560,550
15.5
v
56,290
82,264
69.686
77,227
Total emissions,
103 metric tons
275.1
3,512
28.1
132.5
376.9
201.6
127.6
686.2
1,534.8
12.3
262.0
197.0
346.3
215.2
Population
density,
people/mi2
82
135
40
79
10
40
47
40
51
28
69
20
Emission
density,
tons/mi2
4.9
10.6
2.5
7.8
1.7
1.8
2.6
2.7
4.7
2.4
5.0
2.8
State
emissions ,
percent of U. S.
2.20
1.06
3.02
1.61
1.02
5.57
2.10
1.58
2.77
1.72
Total
Percent of U.S.
11,511
5.3
285,467
7.4
1,020.5
8.2
40
3.6
-------
TABLE 4-11 (continued)
Region and State
Population
1000's
Total area,
mv
Total emissions,
103 metric tons
Population
density,
people/mi2
Emission
density,
tons/mi2
State
emissions,
percent of U.S.
i
en
Region VIII
Colorado
Montana
North Dakota
South Dakota
Utah
Wyoming
Total
Percent of U.S.
Region IX
Arizona
Cal i form'a
Hawaii
Nevada
Guam
American Samoa
Total
Percent of U.S.
2,583
753
643
686
1,228
390
6,283
2.9
104,247
147,138
70,665
77,047
84,916
97,914
581,927
16.1
2,270
21,520
887
610
85
27
25,399
11.7
113,909
158,693
6,450
110,540
212
76
389,880
10.8
52.6
83.
79.
53.4
94.5
133.4
496.5
113.4
506.
59.
91.
2.
0.02
772.6
6.2
25
5
9
9
14
4
11
20
136
138
6
401
355
65
0.5
0.6
1.1
0.7
1.1
1.4
1.0
1.0
3.2
9.2
0.8
11.8
0.3
2.0
0.42
0.67
0.63
0.43
0.76
1.07
0.91
4.05
0.48
0.73
0.02
-------
TABLE 4-11 (continued)
in
ro
Region and State
Region X
Alaska
Idaho
Oregon
Washington
Total
Percent of U.S.
U.S. Totals
Population
1000 !s
382
831
2,329
3,612
7,154
3.3
217,546
Total area,
mi2
586,412
83,557
96,981
68,192
835,142
23.0
3.62 x 106
Total emissions,
103 metric tons
51.7
73.5
107.1
154.6
386.9
3.1
12,492
Population
density,
people/mi2
1
10
24
53
9
60
Emission
density,
tons/mi2
0.09
0.9
1.1
2.3
0.5
3.45
State
emissions,
percent of U. S.
0.41
0.59
0.86
1.24
-------
en
to
KEY
POPULATION DENSITY, PEOPLE/mi'
200-400
10-50
> 400
*.*j 50-200
TX.JJ VIRGIN
:X:::i ISLANDS
ALASKA
HAWAII
[x-vi PUERTO
Ift-ftl Rlco
m
AMERICAN
SAMOA
•:|ftj| GUAM
Figure 4-11. Characterization of U.S. population density is shown by State.
Source: U.S. Department of Commerce (1977).
-------
I
en
EMISSION DENSITY. METRIC TONS/mi
20
I-H HAWAII I I AMERICAN
I I SAMOA
5-10
GUAM
lx*:| D.C.
Figure 4-12. Characterization of U.S. particulate emission density is shown by State.
Source: U.S. Environmental Protection Agency (1978a).
-------
4.3.2.1.2 Source identification and attribution. The geographic distribution
of stationary source emissions usually coincides with ambient air quality
concentrations, but more sophisticated techniques are generally used to link
specific sources to levels measured by ambient samplers. Filters from parti-
culate high-volume samplers can be subjected to detailed analyses to determine
the types of materials being collected. Two studies conducted in recent years
serve as examples of this types of approach.
In Worcester, MA, high total suspended particulate levels were consis-
tently recorded during the winter months. A study was performed to determine
what portion of the collected particles was re-entrained road dust, specifi-
cally road salt. The results of this study showed that road salt contributed
40 to 60 percent of the material collected on the filter.
The chemical mass balance method was used to assess the impact of major
particulate sources in Portland and Eugene, OR. The results of this study
showed that on high-impact days, carbon from vegetative burn sources, such as
field-burning, slash burning, and space heating with wood contributed between
26 to 51 percent of the fine particulate mass and up ttf 35 percent of the
total suspended particulate mass. These and other methods used for source
attribution are discussed extensively elsewhere in this document.
4.3.2.1.3 Particulate matter emission trends. Nationwide emissions of parti-
culate matter.have been documented since 1940. After a peak in 1950, there
has since been a downward trend in total emissions (U.S. Environmental
Protection Agency, 1978b, c). Table 4-12 summarizes emissions from the
major source categories for these years, and Figure 4-13 depicts the same
information in a bar graph. Transportation and miscellaneous categories are
included to indicate their estimated magnitude.
4-55
-------
TABLE 4-12. TRENDS IN NATIONAL PARTICULATE EMISSIONS 1940-1977 (106 METRIC TONS/YEAR)
tn
Source category
Transportation
Highway vehicles
Nonhighway Vehicles
Stationary fuel combustion
Electric utilities
Industrial
Other
Industrial processes
Chemicals
Petroleum refining
Metals
Mineral products
Other processes
Solid waste
Mi seel laneous
Forest wildfires and
Managed burning
Agricultural burning
Coal refuse burning
Structural fires
Total
1940
0.5
0.2
0.3
8.7
1.8
5.6
1.3
9.9
0.4
0
3.3
4.0
2.2
0.5
5.2
3.4
1.4
0.4
0
24.8
1950
1.1
0.4
0.7
8.1
2.7
4.1
1.3
12.6
0.6
0.1
3.8
5.5
2.6
0.7
3.7
1.7
1.6
0.4
0
26.2
1960
0.6
0.4
0.2
6.7
4.0
2.2
0.5
14.1
0.4
0.1
2.6
8.0
3.0
0.9
3.3
1.0
1.9
0.4
0
25.6
1970
1.2
0.7
0.5
7.1
4.1
2.6
0.4
11.9
0.3
0.1
2.1
7.8
1.6
1.1
0.9
0.5
0.3
0.1
0
22.2
1971
1.1
0.7
0.4
6.6
4.0
2.2
0.4
11.3
0.2
0.1
1.9
7.4
1.7
0.8
1.1
0.7
0.2
0.1
0.1
20.9
1972
1.2
0.8
0.4
6.4
4.1
2.2
0.3
10.6
0.2
0.1
1.9
6.9
1.5
0.7
0.7
0.5
0.1
0
0.1
19.6
1973
1.2
0.8
0.4
6.5
4.4
2.0
0.3
10.3
0.2
0.1
2.1
6.4
1.5
0.6
0.6
0.4
0.1
0
0.1
19.2
1974
1.2
0.8
0.4
5.6
3.8
1.8
0.3
8.9
0.2
0.1
1.9
5.5
1.2
0.6
0.7
0.5
0.1
0
0.1
17.0
1975
1.1
0.8
0.3
5.0
3.7
1.5
0.2
6.5
0.2
0.1
1.4
3.7
1.1
0.5
0.6
0.4
0.1
0
0.1
13.7
1976
1.1
0.8
0.3
4.6
3.3
1.1
0.2
6.2
0.2
0.1
1.5
3.2
1.2
0.5
0.8
0.6
0.1
0
0.1
13.2
1977
1.1
0.8
0.3
4.8
3.4
1.2
0.2
5.4
0.2
0.1
1.3
2.7
1.1
0.4
0.7
0.5
0.1
0
0.1
12.4
Note: A zero indicates <50,000 metric tons per year. Section 4.3.2.2 updates transportation sources.
Source: U.S. Environmental Protection Agency (1978b, c).
-------
-p»
en
28
-------
It may be noted that although stationary fuel combustion emissions have
declined since 1940, industrial process emissions increased until 1960 and
solid waste emissions increased until 1970 before starting to decline. From
1970 to 1977, major sources of emissions declined by the following percent-
ages: stationary fuel combustion, 32 percent; industrial processes, 55 percent;
solid waste, 64 percent; and total emissions, 44 percent.
The downward trend in emissions particularly since 1970 may be related to
increased regulatory activity at both the State and national levels. The
Clean Air Act Amendments of 1970 encouraged State governments to focus on
their suplementation plans. In addition, EPA initiated Federal regulation of
new and significantly modified air pollution sources in 1971 through the
promulgation of New Source Performance Standards (NSPS). Between 1970 and
1977, these standards were developed for 27 source categories; those which
were regulated for particulate matter are listed, below along with their
effective dates (Federal Register, 1979):
Fossil fuel-fired steam generators - 1971
Incinerators - 1971
Portland cement plants - 1971
Asphalt concrete plants - 1973
Petroleum refineries - 1973
Secondary lead smelting - 1973
Secondary brass and bronze plants - 1973
Iron and steel plants - 1973
Sewage treatment plants - 1973
Primary copper smelters - 1974
Primary zinc smelters - 1974
Coal preparation plants - 1974
Ferroalloy production facilities - 1974
Iron and steel plants: electric arc furnaces - 1974
Kraft pulp mills - 1976
Grain elevators - 1978
Lime manufacturing plants - 1977
There is now a program to promulgate NSPS for 59 additional industries by
August 7, 1982. Those for which particulate standards are likely to be adopted
are listed below in order of priority (Federal Register, 1979):
4-58
-------
Petroleum refineries: Fugitive sources
Acrylic resins
Mineral wool
Stationary internal combustion engines
Industrial boilers
Nonmunicipal incineration
Nonmetallic mineral processing
Metallic mineral processing
Secondary copper
Phosphate rock preparation
Steel and gray iron foundries
Polyethylene resins
Charcoal production
Synthetic rubber
Byproduct coke ovens
Synthetic fibers
Plywood manufacture
Potash
Clay and fly ash sintering
Glass
Gypsum
Sodium carbonate
Secondary zinc
Phenolic resins
Urea - melamine resins
Polystyrene resins
ABS-SAN resins
Fiberglass
Polypropylene resins
Asphalt roofing plants
Brick and related clay products
Ceramic clay manufacturing
Ammonium nitrate fertilizer
Castable refractories
Borax and boric acid
Polyester resins
Ammonium sulfate
Starch
Perlite
Detergent
Another factor in future particulate emission levels is the program
dealing with National Emission Standards for Hazardous Air Pollutants (NESHAPS)
Emissions of asbestos, beryllium, and mercury, currently regulated under
NESHAPS, are expected to decrease in future years (Council on Environmental
Quality. 1978).
4-59
-------
Other regulatory mechanisms which will likely result in declines in
particulate emission levels are the Prevention of Significant Deterioration
(PSD) regulation for attainment areas and the Emission Offset Policy for
nonattainment areas.
4.3.2.1.4 Discussion of major source categories
4.3.2.1.4.1 Stationary fuel combustion. Total emissions of particulate
matter from stationary fuel combustion sources were estimated at just over 5 x
10 metric tons in 1975, about 40 percent of the total U.S. figure for all
categories. The breakdown (in 10 metric tons) within this source category
was as follows:
Electric generation - 2.8 (55 percent)
Industrial - 1.9 (37 percent)
Commercial/institutional - 0.19 (3.8 percent)
Residential - 0.18 (3.6 percent)
Internal combustion/miscellaneous 0.01 (0.2 percent)
As noted, electric generation and industrial fuel consumption accounted for
over 90 percent of total category emissions. The geographic distribution of
these emissions in primarily centered in the "power plant" corridor which runs
from Missouri, Tennessee, and Kentucky, through Illinois, Indiana, Ohio, West
Virginia, and Pennsylvania, into New York. (Figure 4-14). All States contri-
buting >1 percent of U.S. total combustion emissions are indicated by shaded
areas. These areas also correspond to the most populated portions of the
country. Table 4-13 shows the quantities emitted by electric generation and
industrial, commercial and institutional, and residential fuel use. Table
4-14 provides a ranking of States contributing >1.0 percent of the total U.S.
combustion-related emissions. As noted in this table, Ohio emits more than
twice that of any other State and accounts for just over 16 percent of the
4-60
-------
I
CTi
KEY
1-2%
2-5%
Figure 4-14. Percent fractionation of stationary fuel combustion particulate emissions is shown for
States emitting ^ l.Q percent of U.S. mineral products total.
Source: U.S. Environmental Protection Agency (1978c).
-------
TABLE 4-13. STATIONARY FUEL COMBUSTION PARTICULATE EMISSIONS
(10 metric tons/year)
i
en
ro
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
Total
fuel
combustion
272
27
11
21
53
17
14
5
4
64
94
15
10
332
366
69
23
109
63
18
40
42
144
63
24
124
15
27
5
%
of U.S.
5.4
0.5
0.2
0.4
1.0
0.3
0.3
0.1
0.1
1.3
1.9
0.3
0.2
6.6
7.3
1.4
0.5
2.2
1.2
0.4
0.8
0.8
2.9
1.2
0.5
2.5
0.3
0.5
0.1
Electric
generation
184
6
2
1
16
5
2
2
1
44
48
2
0
213
236
40
10
83
21
2
30
8
110
18
5
32
7
14
4
Industrial
82
19
7
15
26
9
6
3
1
17
42
13
8
93
118
24
11
20
39
13
4
16
25
35
15
83
6
11
0.5
Commercial/
institutional
2
1
1
1
5
2
2
0
2
2
1
0
1
17
7
2
1
2
2
1
2
11
3
5
1
4
1
1
—
Residential
4
1
1
4
6
1
4
0
0
1
3
0
1
9
5
3
1
4
1
2
4
7
6
5
3
5
1
1
0.5
-------
TABLE 4-13.
STATIONARY FUEL COMBUSTION PARTICULATE EMISSIONS
(10 metric tons/year)
i
o*
(A)
State
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Guam
U. S. Virgin Islands
Total
fuel
combustion
5
39
68
156
85
46
816
0.5
34
336
26
4
44
9
229
267
33
2
161
73
286
188
50
--
1
--
%
of U.S.
0.1
0.8
1.4
3.1
1.7
0.9
16.2
0.01
0.7
6.7
0.5
0.07
0.9
0.2
4.5
5.3
0.7
0.04
3.2
1.5.
5.7
3.7
1.0
--
0.02
--
Electric
generation
1
12
65
84
48
6
643
--
—
208
1
--
13
4
165
42
7
--
87
--
96
126
32
--
1
--
Industrial
2
15
1
25
24
33
147
--
29
93
24
2
24
3
59
219
24
1
62
67
184
45
16
--
--
--
Commercial/
institutional
1
6
1
33
3
6
18
--
1
16
--
1
3
1
1
5
1
--
4
2
3
10
1
--
--
--
Residential
1
6
1
14
10
1
8
0.5
,4
19
--
1
4
1
4
1
1
1
8
4
3
7
1
--
--
--
-------
TABLE 4-14. TOTAL FUEL COMBUSTION PARTICULATE
EMISSIONS: RANK BY STATE AND CUMULATIVE
PERCENT OF U.S. TOTAL (FOR STATES
EMITTING >1.0 PERCENT OF U.S. TOTAL)
State
Ohio
Indiana
Pennsylvania
Illinois
West Virginia
Alabama
Texas
Tennessee
Wisconsin
Virginia
New York
Michigan
Missouri
Kentucky
Georgia
North Carolina
Washington
Iowa
New Mexico
Florida
Louisiana
Minnesota
Cal ifornia
Wyoming
Percent
of U.S.
16.2
7,3
6.7
6.6
5.7
5.4
5.3
4.5
3.7
3.2
3.1
2.9
2.5
2.2
1.9
1.7
1.5
1.4
1.4
1.3
1.2
1.2
1.0
1.0
Cumulative
percent
16.2
23.5
30.2
36.8
42.5
47.9
53.2
57.7
61.4
64.6
67.7
70.6
73.1
75.3
77.2
78.9
80.4
81.8
83.2
84.5
85.7
86.9
87.9
88.9
4-64
-------
U.S. total. The highest ranked six states account for almost 50 percent of
U.S. emissions; 14 States, 75 percent; and all 24 States contributing >1.0
percent, about 90 percent of the total in the U.S. Combustion of coal is the
primary reason for high emissions in Ohio, Indiana, Pennsylvania, Illinois,
West Virginia, Alabama, Texas, and Tennessee, which are among the leading
coal-producing States.
A review of the 1976 NEDS file (unpublished data, U.S. Environmental
Protection Agency, 1976) shows total stationary fuel combustion emissions
equal to 3.2 x 10 metric tons, but this figure does not include area estimates
for residential fuel or portions of commercial and institutional and industrial
fuel use. The total breakdown for 1976 was as follows (10 metric tons):
Electric generation - 2.2 (68 percent)
Industrial - 1.0 (29 percent)
Commercial/institutional - 0.09 (2.5 percent)
Internal combustion - 0.01 (0.5 percent)
The detailed information available from the 1976 NEDS inventory also gives
an idea of the emissions breakdown by fuel type within each category as shown:
Electric generation
o bituminous coal - 91 percent
o residual oil - 5 percent
Industrial -
o bituminous coal - 57 percent
o wood/bark waste - 19 percent
o residual oil - 10 percent
o bagasse - 5 percent
4-65
-------
Commercial/institutional -
o bituminous coal - 67 percent
o anthracite coal - 17 percent
o residual oil - 10 percent
The size of particles from combustion-related emissions depends primarily
on the fuel and firing method. For example, cyclone and pulverized-coal furnaces,
typically used in electric generation, discharge predominantly fine material;
stoker-fired boilers, used mainly in industrial boilers, emit coarser material.
Table 4-15 shows that oil- and gas-fired units emit mostly fine-sized particles
(Surprenant et al., 1976). Air emissions of various trace elements, based on
1974 emissions data, are shown in Table 4-16 and Table 4-17. As noted, electric
generation is the prime contributor for all of these chemical constituents.
Forecasts concerning anticipated fuel consumption in 1985 are available in
a published report by the Department of Energy (1977). Coal usage is expected
to increase from 590 x 10 metric tons (1976) to about 934 x 10 metric tons
(1985), a 58 percent increase. Total energy consumption of petroleum liquids is
expected to show a modest increase from about 37 x 10 bbls/day (1977) to about
44 x 10 bbl/day in 1985, (a 19 percent increase). Consumption of natural gas
is expected to decline by about 7 percent from 20 trillion ft /year (1977) to
18.7 trillion ft /year in 1985. As noted earlier, increased coal burning will
result in three noteworthy effects:
(1) an unchanged magnitude of overall particulate emissions, since
well-controlled coal-burning plants emit about the same amounts as
uncontrolled oil-fired plants
(2) a redistribution of these overall emissions because of construction of
new plants, and
(3) an increase in trace element emissions.
4-66
-------
TABLE 4-15. FUEL COMBUSTION: PERCENT
OF EMISSIONS LESS THAN 3 urn
FOR FOUR SOURCE CATE-
GORIES AND BY FUEL TYPE3
Source category/fuel % < 3 urn
Electric generation
Coal 34
Residual Oil 91
Distillate Oil 88
Natural Gas 92
Industrial
Coal 3
Residual Oil 90
Distillate Oil 90
Natural Gas 92
Commercial/Institutional
Coal 3
Residual Oil 90
Distillate Oil 90
Natural Gas 90
Residential
Coal 2
Residual Oil
Distillate Oil 90
Natural Gas 90
Based on total mass emission data from 1974, typical fuel firing methods
used within each category, and the degree and extent of emission reduction
normally achieved.
Source: Surprenant et al. (1976).
4-67
-------
TABLE 4-16. TRACE ELEMENT AIR EMISSIONS: PERCENT OF TOTAL FROM CONVENTIONAL STATIONARY
COMBUSTION SYSTEM CATEGORIES
en
oo
Source
Electric generation
Industrial
Commercial /institutional
Residential
Total, metric tons/year
Electric generation
Industrial
Commerci al /i nsti tuti onal
Residential
Total, metric tons/year
Electric generation
Industrial
Commercial /institutional
Residential
Total, metric tons/year
As
90
8
2
0.
2,994
Cr
84
11
5
0.1
1,633
Ni
60
21
19
0.
7,348
Ba Be B Br
88 89 90 84
9 9 9 13
3212
06 0.1 0.06 0.05 1
2,767 236 4,990 6,078
Co Cu F Fe
63 72 83 77
23 16 13 20
14 12 23
0.3 0.07 2 0.2
463 2,540 33,566 154,224
Se Ti U V
85 89 86 63
13 9 10 20
2 2 4 17
02 1 0.06 0.05 0.
789 56,246 1,542 9,979
Cd
61
21
18
0.007
299 644
Pb
92
7
1
0.05
1,179 4
Zn
89
10
1
02 0.1
Cl
83
13
2
2
,112
Mn
89
10
1
0.5
,627
Zr
78
20
2
0.2
Hg
81
14
3
2
54
2,087 2,087
Source: Surprenant et a"l . (1976).
-------
TABLE 4-17. TRACE ELEMENTS IN SOLID WASTE (ASH): PERCENT OF TOTAL GENERATED BY CONVENTIONAL
STATIONARY COMBUSTION SYSTEM CATEGORIES
en
Source
Electric generation
Industrial
Commercial /institutional
Residential
Total, metric tons/year
Electric generation
Industrial
Commercial /institutional
Residential
Total, metric tons/year
Electric generation
Industrial
Commercial /institutional
Residential
Total, metric tons/year
As
88.9
8.9
0.9
1.6
12,247
Co
69.3
8.9
8.2
13.7
1,923
Se
76.0
22.1
1.0
1.0
370
Ba
83.2
13.2
1.4
2.2
15,967
Cu
77.6
12.1
4.1
6.2
4,282
Ti
83.1
12.8
1.7
2.4
178,718
Be
83.1
12.5
1.8
2.6
737
F
0
0
0
0
0
B Br
84.6 0
12.8 0
1.1 0
1.6 0
16,239 0
Fe
86.9
8.8
1.8
2.5
1,369,872 2
U V
84.5 83.6
12.9 12.1
1.1 1.3
1.6 2.0
4,509 8,455
Cd
82.6
14.1
1.4
1.9
110
Pb
80.9
14.6
1.8
2.6
,531 12,
Zn
82.6
12.8
1.8
2.8
6,895
Cl
0
0
0
0
0
Mn
97.7
1.2
0.6
0.4
519
Zr
85.5
11.1
1.4
2.1
17,237
Cr
75.1
11.7
5.4
7.8
5,035
Hg Ni
77.9 80.8
20.2 12.0
1.8 2.9
0.1 4.3
6 4,699
Source: Surprenant et al. (1976).
-------
In 1978, EPA estimated that of 623 large coal- and oil-fired power plants, only
62 percent were in compliance with participate emission standards. (Council on
Environmental Quality, 1978).
4.3.2.1.4.2 Mineral products industries. Particulate emissions from the
mineral industries were 3.1 x 10 metric tons in 1975, which represented 25
percent of the U.S. total. The 1970 total was 7.8 x 106 metric tons, which
represented 35 percent of the U.S. total at that time (U.S. Environmental
Protection Agency, 1978c). States accounting for the largest percentage of
emissions are Kentucky (15 percent), Illinois (10 percent), Pennsylvania (6
percent), Virginia (6 percent), Texas (6 percent), Calfornia (6 percent), and
Ohio (5 percent) as shown in Figure 4-15 and Figure 4-16 (U.S. Environmental
Protection Agency, 1978c).
Mineral industries incorporated in the above emission estimated include
the production of asphalt concrete, brick, castable refractory, Portand cement,
fiberglass, glass, gypsum, lime, mineral wool, cleaned coal, phosphate rock,
potash, crushed and broken stone, sand and gravel, and other similar products.
Production data for some of the larger categories are presented in Table 4-18.
Production facilities for crushed and broken stone, sand and gravel, and
asphalt concrete are located in virtually every States. In contrast, phosphate
rock is produced primarily in Florida, while potash is produced mostly in New
Mexico.
Emissions sources vary with the type of mineral industry but usually occur
in several specific categories. Typical point source emissions occur from
crushing, screening, grinding, conveyor transfer points, loading, and drying
or calcining. Drying and calcining are potentially the largest emission source.
Fugitive dust sources such as blasting, hauling, haul roads, stockpiles, plant
4-70
-------
KEY
L'.l 1-2X
2-5X
6X
• •
• •
• •
ALASKA
HAWAII
1.2
GUAM
D
• •
• •
• •
• »
1 AMERICAN
SAMOA
PUERTO
RICO 1.8
VIRGIN
ISLANDS
Figure 4-15. Percentage fractionation of mineral products industry paniculate
State emitting 2» 1.0 percent of U.S. mineral orariurtc tntai
mineral products total.
emissions is shown for
-------
-p»
>^l
ro
KEY
h.'.l 1-2%
2-5%
i:::::::l > s% OF us. TOTAL
Figure 4-16. Map shows geographic distribution (as percent of total) of paniculate emissions from
primary metal production by State.
Source: Based on unpublished data from 1976 NEDS file.
-------
TABLE 4-18. PRODUCTION SUMMARY FOR MAJOR MINERAL PRODUCTS INDUSTRIES
Mineral
1975 production
quantity
(103 metric tons)
Production Number of
value active
($106) operations
Major producing states
CO
Crushed and broken stone
Sand and gravel
Asphalt concrete
Cement
Phosphate rock
Gypsum
Lime
Potash
820
720
300
68
45
17
17
4
,000
,000
,000
,000
,000
,000
,000
,200
2
1
5
2
1
,100
,300
,300
,100
,100
230
530
220
4,800
5,500
4,300
160
<50
130
170
11
PA,
AK,
CA,
CA,
FL,
CA,
OH,
NM
IL,
CA,
IL,
PA,
NC,
MI,
PA,
TX,
MI,
OH,
TX,
TN
IA,
TX,
FL,
IL,
PA,
MI,
TX
MO,
OH
TX, OH
NY
MO
MI
Source: U.S. Department of the Interior (1977) and Khan et al. (1977).
-------
yards, and uncovered conveyors are significant but have not been quantified.
Fugitive emissions and some point-source emissions (conveyor transfer points
and loading) tend to produce particles larger than 15 urn. Significant fractions
of the fugitive emissions may even settle out on the plant property. Drying
and calcining tend to produce relatively finer particles. Composition of the
emissions is similar to that of the mineral being processed.
(1) Portland cement production. The production of Portland cement
encompasses most operations that are typical of the mineral industry and is
one of the largest sources of emissions. Estimated 1975 emissions are 900,000
metric tons, or one-quarter of the mineral industry total. U.S. production of
Portland cement is expected to increase by 60 percent in the next 25 years
(Hall and Ela, 1978).
Portland cement manufacture accounts for about 98 percent of the cement
production in the United States (U.S. Environmental Protection Agency, 1979).
The more than 30 raw materials used to make cement may be divided into four
basic components: lime (calcareous), silica (siliceous), alumina (argillaceous),
and iron (ferriferous). The raw materials undergo separate crushing after the
quarrying operation, and, when needed for processing, are proportioned, ground,
and blended using either the wet or dry process.
In the dry process, the moisture content of the raw materials is reduced
to less than 1 percent either before or during the grinding operation. The
dried materials are then pulverized into a powder and fed directly into a
rotary kiln. Drying, decarbonating, and calcining are accomplished as the
material travels through the heated kiln, finally burning to incipient
fusion and forming the clinker. The clinker is cooled, mixed with about 5
percent gypsum by weight, and ground to the final product fineness. The
cement is then stored for later packaging and shipment.
4-74
-------
With the wet process, a slurry is made by adding water to the initial
grinding operation. After the materials are mixed, the excess water is removed,
and final adjustments are made to obtain a desired composition. This final
homogeneous mixture is fed to the kilns as a slurry of 30 to 40 percent
moisture or as a wet filtrate of about 20 percent moisture. The burning,
cooling, addition of gypsum, and storage are carried out as in the dry process.
Operation of the kiln and associated clinker cooler are potentially the
largest source of emissions, as indicated by the emission factor of 71-76 kg
per metric ton of feed (U.S. Environmental Protection Agency, 1977). Emissions
after control vary widely from site to site, depending on the applicable
regulation (Federal NSPS, State or local); enforcement practices; and the
type, design and reliability of the control device. The Federal NSPS (promul-
gated in 1971) limits emissions to 0.15 kg/metric ton of feed, an emission
reduction of 99.8 percent.
Very limited data on the particle size and chemical composition of cement
kiln emissions are available. About 10 percent of the emissions before control
are less than 3 urn and 40 percent are less than 3 jam after control as an industry
average (Weart et a!., 1974). When an electrostatic precipitator is used to
meet NSPS, the mean size of the emitted particles is 1 |jm (Barrett, 1979).
The particulate emissions result from raw material attrition, volatilization of
components in the feed, and the fuel (when coal or oil is used). Chemical
composition of the emission is probably similar to the collected dust, as
shown in Table 4-19.
(2) Crushed and broken stone. In 1975, U.S. output of crushed stone was
820 x 10 metric tons, consisting of limestone (74 percent), granite (11
percent), trap rock (8 percent), sandstone (8 percent), and other (4 percent).
Production is expected to double by the year 2000. Crushed stone was produced
4-75
-------
TABLE 4-19. TYPICAL COMPOSITION OF CEMENT KILN DUST
Component
Weight %
Clay (HC1 insoluble,
Organic substance
fired at 800°C)
Cations
Lithium
Sodium
Potassium
Rubidium
Cesium
Magnesium
Calcium
Strontium
Anions
Fluoride
Chloride
Bromide
Iodide
Carbonate
Sulfate
Sulfide
Borate
Phosphate
Li
*b<
Cs
MgH
CaH
Sr
F"
Cl
Br
I
CO
BO
PO
3—
4.61
2.06
0.0064
12.25
24.50
0.475
0.0074
Trace
9.26
0.015
0.46
1.43
0.040
0.0552
29.59
9.06
Trace
0.152
Not detectable
Heavy metals (weight %)
Heavy metal oxides (weight %)
Chromium
Manganese
Iron
Zinc
Lead
Cr
Mn
Fe
Zn
Pb
0.011
0.013
0.84
1.62
0.562
Sum of all determinations
Oxygen (from CaO not bound in carbonate)
Sum of all constituents
Cr^O.
MnO,
Fe2°3
Zn6 *
PbO
0.016
0.021
19
,02
1.
2.
0.607
97.825
2.98
100.805
Source: Barrett (1979).
4-76
-------
in every State except Delaware and North Dakota. The leading producers,
accounting for 30 percent of production, were Pennsylvania, Illinois, Texas,
Missouri, and Ohio. The most important end uses of crushed stone are road
base (55 percent), concrete (13 percent), cement (11 percent), and many other
smaller uses totaling 21 percent (Reed, 1978). Emission sources include all
those typical of the mineral industry except kiln and dryer. EPA emission
factors for plant process facilities (i.e., excluding fugitive dust sources)
indicate potential emissions of 5.5 kg/metric ton, of which 3.1 kg/metric ton
will settle out on the plant property and 2.4 kg/metric ton will remain suspended
(U.S. Environmental Protection Agency, 1977). Potential suspended particulate
emissions based on 1975 production of 820 x 10 metric tons are 1 million
metric tons per year. Suspended particulate emissions after control are
200,000 tO 400,000 metric tons per year.
Chemical composition of emissions is similar to that of the material
processed. Table 4-20 presents chemical composition data for limestone, which
is the primary product of the crushed and broken stone industry.
(3) Asphalt concrete production. Asphalt concrete is a mixture of
aggregate asphalt cement, and occasionally mineral filiter. Production in
1975 was 300 x 10 metric tons, and is expected to increase to 375 x 10
metric tons by 1985 (U.S. Department of the Interior, 1975). Asphalt concrete
is used for highways and streets (67 percent), commercial uses (28 percent),
and airports and private uses (5 percent). Most plants are located in urban
areas where there is a constant market for the product. California, Illinois,
Ohio, Pennsylvania, and New York each contain over 5 percent of the U.S.
asphalt concrete plants. An asphalt concrete plant may include: dryers;
systems for screening, handling, storing, and weighing hot aggregate; systems
4-77
-------
TABLE 4-20. TYPICAL VALUES OF TRACE ELEMENTS IN PARTICIPATE
EMISSIONS FROM CRUSHED AND BROKEN STONE PROCESSES3 (ppm)
Element Limestone
As
Ba
Be
Br
Cd
Cu
Cr
Fe
K
Mn
Ni
Pb
Sb
Se
Sr
Zn
U
V
<10
30-300
<1
<0.3
<0.3
0.5-4
<10
200-2000
100-1000
6-60
<6
<3
<0.3
<3
100-1000
<30
0.1-1
1-2
aData are typical analyses of limestone which are assumed to adequately
represent the crushed and broken stone industry.
4-78
-------
for loading and unloading; and transfer and storage systems associated with
emission control systems. The rotary dryer typically used to dry and heat the
aggregate is potentially the largest emission source, as shown by the emission
factor of 22.5 kg/metric ton (U.S. Environmental Protection Agency, 1977).
However, essentially all plants have at least primary control equipment capable
of reducing emissions to below 7.5 kg/metric tons. The best controlled plants
achieve emissions rates of 0.01 kg/metric ton. Total emissions are 100,000 to
150,000 metric tons per year, which is equivalent to 0.33 to 0.50 kg/metric
tons. Fine particles probably represent 30 to 60 percent of the controlled
emissions.
4.3.2.1.4.3 Primary metals. Particulate emissions from the primary
metal industries were 1.0 x 10 metric tons in 1975, or 11 percent of the U.S.
total. In contrast, emissions in 1970 were 2.1 x 10 metric tons, or 10
percent of the U.S. total (U.S. Department of Interior, 1977). Eight States
(Ohio, Minnesota, Pennsylvania, Indiana, Missouri, Texas, New Mexico, and
Louisiana) account for 73 percent of the total, with Ohio alone accounting for
25 percent and Louisiana 15 percent as shown in Figure 4-16 (unpublished data,
U.S. Environmental Protection Agency, 1976).
The most important primary metal industries included in the above emission
estimates are aluminum, integrated iron and steel (including coke, pig iron,
and steel production), ferroalloys, lead, copper, and zinc. Table 4-21 shows
that the aluminum and integrated iron and steel industries account for 72
percent of the particulate emissions from the primary metal industries.
The iron and steel industry is the largest primary metal industry in
terms of material produced and value of production, as shown in Table 4-22.
Iron, steel, and coke production had a value of $50 billion in 1975, compared
with $5-6 billion for aluminum, ferroalloys, lead, copper, and zinc combined.
Pennsylvania, Indiana, and Ohio are the largest producers of iron and steel.
4-79
-------
TABLE 4-21. DISTRIBUTION OF EMISSIONS AMONG PRIMARY
METAL INDUSTRIES3
Industry or operation
Percent of total emissions
Iron production
Steel production
Byproduct coke
Aluminum
FerroalToy
Copper
Lead
Zinc
Miscellaneous
Other
Each
25
19
7.7
20
11
3.8
1.8
0.6
11
less than 0.03
Source: Unpublished data, U.S. Environmental Protection Agency (1976).
TABLE 4-22. PRODUCTION OF PRIMARY METALS, 1975
Product
Steel
Pig iron
Coke
Aluminum
Ferroalloys
Copper
Lead
Zinc
Production,
106 metric tons
106
73
52
3.5
1.7
1.3
0.5
0.4
Value,
$106
30,700
13,700
4,950
2,976
988
1,814
274
366
Principal
producing states
PA, OH, IN
PA, OH, IN
PA, OH, IN, AL
Northwestern &
South Central U.S.
Eastern U.S.
AZ, NM, UT, MT
MO, ID, CO
PA, TX, ID
Source: Desy (1978).
4-80
-------
The types of particles emitted include heavy metals found on impurities
in the ores, particles of the material being processed, fluorides and fluxes,
and polynuclear aromatic compounds from coke ovens. Most of the emissions
discussed in this report result from high-temperature processing equipment
such as coke ovens, calcining of aluminum hydroxide in reverberatory furnaces,
open and semicovered electric arc furnaces, blast furnaces, basic osygen
furnaces, and sintering operations. Stack and fugitive process emissions from
the above high-temperature operations are predominantly fine particles.
Fugitive dust sources such as coal and ore handling, truck traffic on unpaved
roads, and storage piles may produce larger particles.
(a) Byproduct coke ovens. In 1975, 52 x 10 metric tons of coke were
produced at 62 plants located in 19 states. The largest producers were
Pennsylvania (27 percent), Indiana (16 percent), Ohio (15 percent), and
Alabama (8 percent). Production of coke will probably keep pace with the
increase in steel production, which is expected to be 1.6 percent per year to
the year 2000 (Desy.1978).
Coking is the process of heating coal in an atmosphere of low oxygen to
remove volatile components. In the byproduct process, which represents over
98 percent of the industry, these volatile organic compounds are recovered.
Particulate emissions result from charging of coal to the hot ovens, door and
topside leaks, underfiring, pushing (removal of hot coke), and quenching.
Estimated U.S. particulate emissions from coke ovens were 74 x 10 metric tons
in 1976.
Door and topside leaks and underfiring produce primarily fine particles,
including particulate organic compounds. Charging, pushing, and quenching
produce large particles of coal or coke particles and fine particles that
consist, at least partly, of condensed organic components.
4-81
-------
(b) Iron production. Pig iron was produced at about 200 blast furnaces
in 1975. Total production was 73 x 10 metric tons. The largest producing
states were Pennsylvania (22 percent), Indiana (20 percent), and "Ohio
(18 percent). Pig iron production will keep pace with the expected increase
in steel production of 1.6 percent per year to the year 2000.
A blast furnace consists of a refractory-lined steel shaft in which the
charge is continuously added to the top, and preheated air is blown in through
the bottom and emitted as combustible gas. The charge consists of iron ore,
sinter or pellets, coke, limestone or dolomite, and possibly small amounts of
iron or steel scrap. The combustible gases are cleaned and burned in the
steel plant. Particulate emission sources include the combustion gases,
tapping operations, and blast furnace slips (operations that require bypassing
the control device). Estimated 1976 emissions were 82 x 10 metric tons.
The emitted particles are probably all fine particles that either escape
the control device or result from tapping. Composition of the blast furnace
flue dust is iron (36 to 50 percent), silicon dioxide (9 to 13 percent),
aluminum oxide (2 to 5 percent), calcium oxide (4 percent), carbon (4 to 14
percent), and other components. Data on heavy metals are scarce; one source
reports 0.0014 percent cadmium, 0.083 percent lead, and 0.93 percent zinc
(Katari and Gerstle, 1977).
(c) Steel production. During 1975, the United States produced 106 x 10
metric tons of steel. The largest producers were Pennsylvania (22 percent),
Ohio (17 percent), and Indiana (17 percent). Steel production is expected to
increase 1.6 percent per year to the year 2000.
Steel is produced by a refining process that lowers the carbon and silicon
content of the charge and removes impurities, mainly phosphorous and sulfur.
Excess oxygen in the molten steel is removed by deoxidizing elements such as
4-82
-------
manganese, silicon, or aluminum. Pig iron represents about 55 percent of the
iron feed, and scrap represents the majority of the remainder. Steel is
currently produced by three types of furnaces, as shown in Figure 4-17 (Desy,
1978).
The basic oxygen furnace produces steel from a furnace charge composed of
about 70 percent molten pig iron and 30 percent scrap. The tremendous agitation
produced by the oxygen lancing produces high dust loadings (11 to 18 g/m ) of
iron and small amounts of fluorides. Most of the particles are less than 5
urn. One source reports that the particulate material contained 80 ppm cadmium,
4600 ppm lead, and 45,000 ppm zinc. Estimated 1976 emissions were 91 x 10
metric tons.
From 1960 to 1975, steel production in open hearth furnaces delcined from
90 percent of the U.S. total to 20 percent. Open hearth furnaces are being
replaced by basic oxygen furnaces. The composition of particulate emissions
is similar to that of the basic oxygen furnace. About 50 to 60 percent of the
emissions before control are less than 5 urn, and probably 90 percent are fine
particles after control. Total open hearth particulate emissions were about
30 x 10 metric tons in 1976.
Two types of electric furnaces, the arc furnace and the induction furnace,
are used to produce steel. The arc furnace is used to produce high-alloy
steels, as well as a considerable amount of mild steel. The induction furnace
produces primarily specialty and high alloy steels with no real emission
problems and therefore is omitted in the remainder of this report. Particulate
emissions from arc furnaces consist primarily of oxides of iron, manganese,
aluminum, calcium, magnesium, and silicon. Particulate fluorides are also
emitted. Emissions are primarily fine particles.
4-83
-------
(d) Aluminum production. Almost 4 million metric tons of aluminum were
produced at 31 U.S. plants in 1975. The production of aluminum consumes large
amounts of electric energy, and plants are therefore often located where
energy is inexpensive. The Northwest (Washington and Oregon) and South Central
states (Tennessee, Arkansas, Louisiana,) are large producers. U.S. production,
despite increased imports, is expected to increase 11 percent per year, reaching
18 million tons by the year 2000.
Aluminum is produced from bauxite, a hydrated oxide of aluminum associated
with silicon, titanium, and iron. Most bauxite ore is purified by the Bayer
process, in which the ore is dried, ground in ball mills, and mixed with
sodium hydroxide. Iron oxide, silica, and other impurities are removed by
settling, dilution, and filtration. The aluminum hydroxide is precipitated
from this diluted, cooled solution and calcined to produce pure alumina (AlpO-).
Aluminum metal is manufactured by the Hall-Heroult process, which involves
the electrolytic reduction of alumina dissolved in a molten salt bath of
cryolite (a complex of NaF-AlF.,) and various salt additives.
The largest potential sources of emission are bauxite calcining with an
emission factor of 100 kg/metric tons, and the reduction process, which has an
emission factor of 40 to 50 kg/metric tons. Reported U.S. emissions after
control are 87,000 metric tons for the calcining operation and 105 x 10
metric for the elctrolytic reduction process.
Particulate emissions are primarily alumina, with about 25 percent
particulate fluorides (U.S. Environmental Protection Agency, 1977). Some 35
to 44 percent of the particles are less than 1 pm before control.
4.3.2.1.4.4 Industrial process fugitive particulate emissions (IPFPE).
It appears that three broad categories account for nearly all of the potential
4-84
-------
100
p 80
o
o
o
oc
o.
in
V)
I
oc
u.
O
I-
IU
u
-------
industrial process fugitive particulate emissions in the United States. These
three categores are mineral products (48.1 percent), food and agriculture
(37.2 percent), and primary metals (14.3 percent) (See Table 4-23).
In the mineral products industries, particulate emissions tend to reflect
the composition of their parent materials. While some of these emissions are
in the fine and inhalable range, most of the emissions are not well charac-
terized. For mineral extraction and beneficiation (crushed and broken stone,
surface coal mining, and sand and gravel operations), which by itself accounts
for greater than 20 percent of nationwide emissions, no reliable particle size
data are available.
Grain elevator operations account for the particulate emissions in the
food and agriculture category. These emissions consist almost entirely of
grain dust generated by abrasion of the individual kernels of grain. Few of
these particles are in the inhalable category.
Primary metal production encompasses six separate industies. Fugitive
particulate emissions in this category result both from the handling and
transporting of raw materials and from the smelting and refining of these raw
materials into their finished metal products. While emissions of the first
type are not well characterized, emissions of the second type often consist of
fine and inhalable metal fumes, including toxic trace metals. Some of these
toxic components of particulate emissions are listed in Table 4-24. This
characterization should apply not only to fugitive emissions but to the stack
emissions from these industries as well.
The remaining two categories, secondary metals and wood products, account
for less than 1 percent of the national total of industrial process fugitive
particulate emissions. Metal fume fugitive particles from secondary metal
melting operations also include toxic components, which are presented in Table
4-24.
4-86
-------
TABLE 4-23. INDUSTRIAL PROCESS FUGITIVE PARTICULATE EMISSIONS
i
CD
Category
percent
of U.S.
total
Mineral
products
(48. IX)
Food and
agriculture
(37.2%)
Industry
Mineral extraction
and beneficiation
Portland cement
Asphaltic concrete
Lime manufacturing
Crecrete batching
Total
Grain elevators
Annual
uncontrolled
-emission
10 metric tons
730a
698
95
47
31
1,601
1,239
Estimated X
of total
plant uncon-
trolled emissions
100.0
5.7
1.4
1.4
100.0
100.0
Fraction
emissions Size characteristics
represented
L No data
M 10 to 15% <10 umc
I 50 to 70% <4 umb
H 45 to 70X <5 umb
H 10 to 20* <5 M"ib
H 10 to 100 vmd
Major components
Same as parent material (impor-
tant for toxic minerals such .
as asbestos, beryllium, silica
Limestone, clay, shale, gypsum,
iron-bearing and siliceous
materials
Sand, crushed stone, limestone,
hydrated lime
Limestone, lime
Cement dust (see Portland cemen
above)
Grain dust
-------
TABLE 4-23 (continued)
£•
I
CD
CO
Category
percent
of U.S. Industry
total
Primary Coke/iron/steel
metals
(14.3%)
Foundries
Aluminum
Copper
Zinc
Total
Secondary Lead
metals
(0.2%) Aluminum
Annual Estimated %
uncontrolled of total
-emission plant uncon-
10 metric tons trolled emissions
234 10.6
122 53.5
57 26.0
43 37.7
5 5.7
475
4 6.2
2 5.6
Fraction
emissions Size characteristics
represented
H 50% inhalable6
H 100% inhalable,
26% fine
M 80% inhalable,
55% fine0
H 67% inhalable,6
at least 25%
fine"'*
L 100% inhaJable;b>k
50% fine
H 100% inhalabje,
67% fine°'R
H 99% inhalable,
94% fine8'"
Major components
Polycyclic organic matter, coal
dust, coke dust, iron oxide dust,
kish, (graphitic material),, metal
fume (primarily iron oxide) '
Metal oxide fume (primarily
oxides of silicon.aod iron) fine
carbonaceous fume '
Particulate fluorides, alumina
(Al_0,), carbon dust, condensed
hydrocarbons, tars
Cu, Fe, S. SiO_ from ore con-
centrate; metai fume consisting
of.oxides of As, Pb, Zn, Cu,
Cd°'J
Ore concentrate dust; metal fume
consisting of oxides of Zn, Pb, Cd
Metal fume consisting of oxides
of Zn, Pb, Sn
Al 0 Aid,, fluorides. NaCl ,
oxtdes of alkali metals
-------
TABLE 4-23 (continued)
i
00
Category
percent
of U.S. Industry
total
Copper (brass/bronze)
Zinc
Total
Wood Lumber and furniture
products
(0.2%)
All
categories
(100.0%)
Annual Estimated %
uncontrolled of total Fraction
-emission plant uncon- emissions Size characteristics Major components
10 metric tons trolled emissions represented
1 10.9 M 100% <5 unih Metal fume consisting .
primarily of oxides of Zn, Pb
1 6.9 M 100% <5 Mm,^'h Metal fumes consisting of oxides
702 < 1 umn of Zn, Al, Cu; NH.C1, . l
ZnCl_, carbonaceous materials '
8
9 52.9 L No Data Sawdust
3,331
*L = Low, <40 percent.
M = Moderate, 40 to 70 percent.
H = High >70 to 100 percent.
.Total includes surface coal mining and crushed stone extraction only
Jutz et al. (1977).
^Sussman (1977).
°Thimsen and Aften (1968).
fIndustry total estimated from data presented in Drehmel et al. (1979)
Steiner (1977).
9Bohn et al. (1978).
^Drehmel et al. (1979).
•Oaugherty and Coy (1979).
jJNelson et al. (1977).
,Zoller et al. (1978).
"Vandergrift et al. (1971).
-------
TABLE 4-24. TOXIC COMPONENTS OF FUGITIVE (AND STACK) PARTICIPATE EMISSIONS
i
vo
o
Industry
Primary metals
1. Coke
2. Iron
3. Steel
4. Foundries
5. Aluminum
6. Copper
7. Lead
8. Zinc
Secondary Metals
1. Lead
2. Aluminum
3. Copper (brass/bronze)
4. Zinc
POMa As Be Pb Cr Cd Hg Se Ag Co Ni Fluorides References
+++ + + + + + * + 1
2,3
+ + + 2,3,4
+ + + + + 5
+ ++5,6
++ ++ + + 5,6
+ ++ + + + 3,5,6,7
++ + + 5
++ 3
++ 3
++ + 3
+ + + + 3,5
Polycyclic organic matter. The symbol (*) is used to indicate the presence of trace or minute quantities of a toxic component in
the particulate emissions from an industry.
.If large quantities (>S percent by weight) of the toxic component are present in the emission, the symbol (++) is used.
1, U.S. Environmental Protection Agency (1978).
2, Steiner (1977).
3, Vandegrift et al. (1971).
4, Bahn et al. (1978).
5, Jutze et al. (1977).
6, Nelson et al. (1977).
7, Daugherty and Coy (1979).
-------
Fugitive participate emissions, by definition, are difficult to collect
and control. For a given industry, there are generally a large number of
fugitive emission sources; e.g., 20 separate sources have been identified for
foundaries (Jutze et al., 1977). In terms of total emissions, one or two of
these sources may predominate. Fugitive particulate emissions on the plant
site may result from open sources, such as wind erosion of storage piles and
vehicular traffic over plant roads; from loading, unloading, transfer, and
handling operations; from incompletely controlled point sources, such as
furnace charging and tapping; and from poorly maintained equipment, such as
leaking furnaces or coke oven doors.
Even though fugitive particulate emission totals may appear small when
compared with large point sources, they may take on added importance because
of the concentration of control efforts on point source emissions. In
situations where point sources are well controlled or emissions are released
from high stacks, fugitive particulate emissions will exert the primary effect
on local air quality. Extremely high particulate levels have been measured in
areas with a predominant industrial-process fugitive particulate influence
(Lynn et al., 1976; Lebowitz, 1975).
Measurement and characterization of IPFPE. Many of the emission factors
currently used for IPFPE are based on engineering judgment or extrapolation
from similar processes. There is often little test data available to support
these estimates. This is partially because of the inherent difficulties in
the measurement of fugitive particulate emissions. If a pollutant is not
easily collected and controlled, it follows that it will not be easy to
<••)•
measure.
4-91
-------
The measurement techniques that have been used for fugitive particulate
emissions are of three general types. The quasi-stack method is considered to
be the most accurate. This technique involves hooding a source Of emissions
and sampling from the captured air stream. This method is not feasible for
many sources. For sources that are enclosed in a building with a limited
number of openings to the ambient atmosphere, the roof monitor method can be
used. Pollutant concentrations in the air flowing to the ambient atmosphere
are measured, along with the airflow rate through the opening. For sources
which are not enclosed, an upwind-downwind sampling technique can be used.
From pollutant concentrations measured upwind and downwind from the source,
windspeed, and diffusion equations, the source strength can be estimated.
With the exception of the quasi-stack method, the measurement techniques are
indirect, and the accuracy of the results can vary considerably (Kolnsberg,
1976).
Characterization of fugitive particulate emissions according to size and
composition presents some problems. When information on particle size is
required, all of the difficulties described previously for emission rate
sampling would seem to apply. One indirect appraoch which has been used is
the application of stack emission size measurements to fugitive particulates
from the same source. While this approach may be valid in some cases, it must
be carefully applied. Often stack emission size measurements are made at the
inlet to an air pollution control device. In some cases, there may have been
a preferential removal of large particles prior to this point by gravitational
settling or passage through a mechanical collector.
4-92
-------
Composition of fugitive participate emissions is easier to characterize.
In the absence of direct measurements, indirect methods may be sufficient.
For a process which is a source of both ducted and fugitive emissions, it is
not unreasonable to assume that both types will be similar in composition
(although the relative proportions of the components may be different). For
open sources and for handling, conveying, transfer, and unloading operations,
emissions should be of the same composition as the parent material. For
sources which do not fit into either of these categories, direct measurement
is necessary.
Derivation of estimated uncontrolled IPFPE inventory. The most complete
reference on industrial process fugitive particulate emissions is a detailed
literature review compiled by Jutze et al. (1977). Nineteen industries which
are major emitters of fugitive particulates were identified. Within each
industry, the processes and operations with fugitive emission potential were
evaluated. For each process, a fugitive emission factor was developed. A
later report applied some of these emission factors to nationwide production
data for each industry in an attempt to estimate national totals for
uncontrolled fugitive particulate emissions. This later report included only
process emissions, omitting such plant-site open sources as storage pile
erosion, vehicular traffic, unloading, and some handling and transfer opera-
tions.
The estimates in Table 4-23 attempted to include these omitted sources.
For each source which had been neglected, the appropriate emission factor was
applied to nationwide production data (Jutze et al. 1977). The uncontrolled
emissions estimates derived in this matter were added to the nationwide totals
described above (Zoller et al., 1978).
4-93
-------
The only exception to this procedure was in the integrated iron and
steel industry. Fugitive particulate emissions from this industry have been
extensively measured (Cusino, 1979), and more up-to-date estimates of annual
uncontrolled emissions were readily available (Spawn, 1979). Since these
estimates were based on emission factors derived from extensive testing, they
were considered more reliable and were used in place of the earlier totals
(Zoller et al., 1978). Since the integrated iron and steel industry includes
coke, iron, and steel production, all of which had previously been reported as
separate industries (Jutze et al, 1977), the number of industrial categories
was reduced from 19 to 17.
In addition to estimates of annual uncontrolled fugitive particulate
emissions from each industrial category, Table 4-23 contains some information
on size and composition characteristics of the emissions. As previously
noted, there are often numerous fugitive particulate emission sources within
an industry. Because size distribution data are lacking for many of these
sources, it is not possible to characterize all emissions within an industry.
In particular, industries which have a large open source contribution tend to
be poorly characterized.
4.3.2.1.4.5 Nonindustrial fugitive particulate emissions. This section
addresses the fugitive particulate emissions related to traffic entrainment
of dust from public paved and unpaved roads, agriculture operations, construction
activities, surface mining operations, and fires. With the exception of
fires, all of these sources may be classified as open-dust sources; i.e., they
entail dust entrainment by the interaction of machinery with aggregate materials
and by the forces of wind on exposed materials.
4-94
-------
Table 4-25 gives estimates of nationwide annual emissions of suspended
participate matter from these sources (Cooper et al., 1979). Off-road motor
vehicles and inactive tailings piles are also potentially significant open
dust sources (TRW, Inc., 1977). It is estimated that fugitive dust emissions
exceed particulate emissions from traditional point sources in 90 percent of
the Air Quality Control Regions that are not meeting the standards for total
suspended particles (Carpenter and Weant, 1978).
Emissions from the interaction of machinery with aggregate materials vary
strongly from one case to the next. The parameters which cause these variations
may be grouped into threee categories: (1) measures of source activity or
energy expended (for example, the speed and weight of a vehicle traveling on
an unpaed road); (2) properties of the material being disturbed (for example,
the content of silt in the surface material on an unpaved road); (3) Climatic
parameters (for example, number of precipitation-free days per year on which
emissions tend to be at a maximum.
Cowherd et al. (1974) have advanced the concept that open-dust source
emission factors be considered in mathematical equations with multiplicative
correction terms containing the above parameters as appropraite, so that the
factors are applicable to a wide range of source conditions. For example,
measured emissions factors for unpaved roads, which span two orders of
magnitude, are predicted by an emission factor equation with a precision
factor of 1.46; 'the 95 percent confidence interval for a predicted emission
value, P, was found to extend from P/1.46 to 1.46 P (Cowherd et al., 1979).
Table 4-26 groups nonindustrial fugitive emission source categories by
dust generation mechanism. Also shown are the emission-related properties of
4-95
-------
TABLE 4-25. ESTIMATED ANNUAL SUSPENDED PARTICULATE EMISSION RATES
FROM OPEN-DUST SOURCES IN THE UNITED STATES
Source category
Estimated emission rate,
106tons/year
Unpaved roads
Wind erosion of cropland
construction activities
Paved roads
Wild fires
Agricultural tilling
Minerals
Suface mining - other
Tailings
Surface mining - coal
Prescribed fires
Total
Total, nonroad
320.0
44
27
7.9
3.4
3.2
2.8
0.8
0.4
0.43
409.0
81.0
Source: Cooper et al. (1979).
4-96
-------
TABLE 4-26. PARAMETERS AFFECTING EMISSIONS FACTORS FOR OPEN DUST SOURCES
Generic source category
Material
Equipment
Climate
10
Vehicular traffic on unpaved
surfaces
Vehicular traffic on paved
surfaces
Batch-drop operations
Continuous-drop operations
Wind erosion of storage
piles and exposed areas
Blasting
Surface silt content
Surface silt content
Surface loading
Silt content
Moisture content
Silt content
Moisture content
Surface credibility
Surface silt content
Surface moisture content
Silt content
Moisture content
Hardness
Vehicle speed
Vehicle weight
Vehicle weight
Loader capacity
Quantity and dis-
tribution of
blasting agent
Wind speed
Wind speed
Wind speed
Source: Cowherd et al. (1979).
-------
the materials being disturbed, the equipment which effects the disturbance
(except in the case of wind erosion), and the prevailing climate where the
source operation occurs. The use of the silt content as a measure of the
dust-generation potential of a material acted on by the forces of wind or
machinery was an important step in extending the applicability of the emission
factor equations to the wide variety of aggregate materials of industrial
importance. The upper size limit of silt particles (75 urn in diameter) is the
smallest for which size analysis by dry seiving is practical, and this size is
also a reasonable upper limit for particles which can become airborne. The
wind erosion parameters given in Table 4-26 were taken largely from the Wind
Erosion Equation developed by Woodruff and Siddoway (1965).
Construction activities and surface mining operations may entail emis-
sions from all of the generic source categories given in Table 4-26. Emissions
during building and road construction are generated by land clearing, blasting,
ground excavation, cut and fill operations, and wind erosion of disturbed
land. Dust-producing operations in surface mining include: topsoil and
overburden removal; drilling and blasting; ore loading, transport, and dumping;
and land reclamation and wind erosion of disturbed areas.
Open-dust sources produce emissions which span a wide range of particle
sizes. In developing emission factors which relate to total suspended particles,
it is necessary to consider the particle capture efficiency of the standard
high-volume sampler. In the past, the 30-jjm Stokes diameter, based on a
o
typical dust particle density of 2 to 2.5 g/cm , has been used to approximate
the effective cutoff diameter for the high-volume sampler (Cowherd et al.,
1974 and Chapter 3 of this document). Recent wind tunnel experiments have
provided more definitive data on particle capture efficient (McFarland et al.,
1979).
4-98
-------
Table 4-27 presents data on fractions of inhalable particles (IP) in
measured particle emissions from open-dust sources. These data were obtained
with a high-volume parallel-slot cascade impactor with a cyclone preseparator
and directional sampling intake. The reliability of these data is weakened by
the need to correct for residual particle bounce and to extrapolate the measured
size distributions from the cyclone cutpoint of about 10 pm (Cowherd, et al.,
1979).
4.3.2.2. Mobile Source Particulate Matter—Mobile sources, excluding off-road
vehicles, account for about 2.5 percent of the nationwide total particulate
emissions from all sources. Mobile-source emissions may be divided into two
general categories: engine-related particles emitted in vehicle exhaust and
vehicle-related particles including tire wear debris and asbestos from clutch
and brake lining wear (Bradow et al., 1979; U.S. Environmental Protection
Agency, 1978c; Dennis, 1974; Jacko and DuCharme, 1973). Less than 10 percent
of the annual vehicle-related particulate material (~22 thousand metric
tons/year in 1977) is airborne (Pierson and Brachaczek, 1974). Engine-related
particles were emitted at a rate of about 325 thousand metric tons/year in
1977. Because of their size, 0.01-10 urn (Khatri and Johnson, 1978), they may
be deposited in the lung. Engine-related particles are mostly under 1 pro in
diameter and have relatively low settling velocities in the atmosphere.
Therefore, they tend to be transported in the ambient air in a manner similar
to gaseous pollutants (Whitby and Cantrell, 1976). Such particles become
widely dispersed and can penetrate and deposit in the human respiratory system
(Lee et al., 1976).
Engine-related particulate emissions from mobile sources may be divided
into three types of aerosols: those dominated by lead halides, by sulfate,
4-99
-------
TABLE 4-27 SUMMARY OF INHALABLE PARTICULATE (IP) FRACTIONS FOR OPEN-DUST SOURCES
o
o
Source/control method
Unpaved roads/surface material
Uncontrol led
Dirt
Dirt and slag
Crushed slag
Gravel
Sand and gravel
Controlled
Dirt and slag treated with
Coherex
Crush rock treated with
Trex
Paved roads (uncontrolled)
Aggregate storage pile stacking
Iron ore pellets
Coal
No. of
tests
9
2
17
5
6
2
3
11
3
2
Avg. IP fraction
in TSP emissions
0.45
0.72
0.54
0.38
0.35
0.88
0.57
0.77
0.49
0.55
Standard
deviation
0.17
0.01
0.14
0.13
0.08
0.12
0.22
0.16
0.26
0.31
Source: Relder and Cowherd (1979).
-------
and by carbonaceous matter with associated absorbed organics. Emission rates of
these particles are classified according to engine type—either gasoline
(spark ignition) or diesel (compression ignition).
Spark ignition vehicles may be subdivided further by fuel type:
o Leaded gasoline--pre-1975 domestic passenger
cars, some imported passenger cars, and some heavy-duty trucks.
o Unleaded gasoline—almost all 1975-to-present domestic passenger
cars and all 1980-to-present domestic and imported passenger cars.
Compression ignition vehicles may be subdivided into three classes by engine
type:
o Four-stroke light-duty engines—domestic and imported diesel-powered
passenger cars.
o Four-stroke heavy-duty engines—domestic and imported trucks and
buses.
o Two-stroke heavy-duty engines—domestic trucks and buses.
Table 4-28 lists total particulate emissions and major aerosol components
by the various vehicle categories. All of the particulate emission rates are
derived from results of laboratory tests which simulate actual urban driving
patterns. The lowest particulate emitter is an unleaded gasoline vehicle,
catalyst-equipped without an air pump; and the highest emitter is a two-stroke
heavy-duty diesel truck. The particulate emission rate of an average two-stroke
heavy-duty diesel truck is over 2000 times that of a gasoline vehicle with
catalyst and no air pumps.
Particulate matter from vehicles using leaded gasoline consists mostly of
the inorganic compounds of lead, bromine, and chlorine and substantial quantities
of carbonaceous material (Springer, 1978). The dominant lead compound present
4-101
-------
TABLE 4-28. MOTOR VEHICLE ENGINE PARTICIPATE EMISSIONS
(simulated urban driving conditions)
Total particulate
Engine Vehicle emission rate,
type type (g/mi)
Fuel specific
Exhaust particle
emission rate, g/kg fuel Leadc
Major component emission rates, g/km (g/mi)
Carbon
Organic Sulfate
o
no
Gasoline Leaded gasoline 0.18(0.29)
pre-1970
1970-later 0.08(0.13)
Unleaded gasoline
Catalyst, no airO.009(0.01)
Catalyst, air 0.031(0.05)
Light-duty dieselO.31 (0.49
Heavy-duty diesel1.0(1.6)
2-stroke truck 1.8(2.8)
1.6
0.69"
0.078
0.26 c
4.7
2.6
4.3
0.067 (0.108) 0.059(.095) 0.061(0.098) 0.002(003)
0.048 (0.077) 0.026(.042) 0.008(.013) 0.01 (.002)
0.003(.005) 0.001(.002) 0.003(.005)
0.006(.010) 0.001(.002) 0.010(.016)
0.001(0.002) 0.23(.37) 0.047(.076) 0.007(.011)
0.004(0.006) 0.75(1.21) 0.12(0.19)
0.001(0.002) 1.33(2.14) 0.90(1.45)
0.059(.093)
0.09K.130)
As PbBrCl for gasoline; as Pb for diesel.
Including both elemental carbon and organic carbon.
CBased on 0.03 with sulfur gasoline and 0.23 with sulfur diesel fuel
Assumes 15 miles per gallon.
eNo. 1 fuel, low-speed driving conditions.
Source: Emissions summarized from Bradow et al. (1979).
-------
is PbBrCl and is due to the presence of the scavengers ethylene dichloride and
ethylene dibromide in leaded fuel. Table 4-28 shows that pre-1970 vehicles
(0.18 g/km) emit about twice the total particulate mass of 1970 and later
vehicles (0.08 g/km). Data on major particulate components for 1970 and later
are limited. When compared to date from the pre-1970 grouping, there are
indications of increased percentages (but decreased masses) of lead compounds
and a decreased percentage of carbon and organic compounds. This shift is
probably a result of the leaner engine combustion of the later model years.
Lead emissions are directly related to fuel lead content, and all the leaded
particulate data presented in Table 4-28 have been normalized to a lead content
of 1.8 g/gal. As a result of gasoline lead phaseout (Federal Register, 1973),
the increased relative abundance of catalyst-equipped cars (1975-present), and
the certification of all 1980 domestic and imported vehicles on unleaded fuel
only, annual mobile source lead particulate emissions are decreasing (U.S.
Department of the Interior, 1977). The available data on particle size indicate
a mass median diameter less than 0.54 urn, probably about 0.25 urn (Moran, et
al., 1971). The introduction of catalyst-equipped vehicles requiring unleaded
gasoline has substantially decreased engine particulate emissions since 1975.
The particulate matter from catalyst vehicles using unleaded gasoline is
dominated by sulfate and carbonaceous material. Sulfate emissions vary directly
with fuel sulfur level and are greater from air-pump-equipped catalyst cars
(0.010 g/km) than from catalyst cars without air (0.003 g/km) (Kawecki, 1978).
The particles emitted from catalyst vehicles burning unleaded gasoline have
aerodynamic diameters smaller than those from vehicles burning leaded gasoline
or diesel fuel these smaller particles have a mass median diameter of about
4-103
-------
0.05 pm (Groblicki, 1976). The smaller percentages of trace metals and soluble
organics from catalyst vehicles are due to the predominance of sulfate and the
efficiency of the catalyst systems for the reduction of organic emissions.
Diesel particulate matter is appropriately described by examining three
engine groups: light duty, four-stroke heavy duty, and two-stroke heavy duty
(includes both trucks arid buses). All of these engine groups emit carbon
particles together with an oily mixture of organic compounds, some minor
amounts of sulfates, and traces of metallic constituents such as iron, copper,
calcium, and zinc (Lee and Duffield, 1979). Sulfate emissions from diesel
passenger cars can be substantial (about 7mg/km) and are typically comparable
with many in-use catalyst cars with air pumps (lOmg/Km).
Carbon-containing species dominate the composition of diesel particles.
Elemental carbon is usually the major component. The amount of associated
organic material varies substantially among engine types and is greatest with
two-stroke engines (Hare and Bradow, 1979). Particulate emission rates for
all diesel groups are sensitive to many parameters, including vehicle size and
operating conditions (speed, load) and fuel (aromatics and sulfur content).
Table 4-28 presents estimates of particulate emission rates and characteristics
for the various engine groups under urban driving conditions. Most of these
data were developed with dynamometer simulations of normal driving with a
typical No. 2 diesel fuel. The total particulate emission rate (0.31 g/km)
reported for light duty diesels is a population-weighted value for vehicles on
the road in 1977. The emission rates of the two-stroke bus (1.1 g/km) are
lower than those of the two-stroke truck (1.8 g/km). The reduced emission
rates of the bus probably are caused by the use of No. 1 fuel and by lower
driving speeds.
4-104
-------
The two-stroke diesel has the highest emission rate of participate
organic matter, typically about 50 percent of the total participate mass (Hare
and Bradow, 1979). A major portion of this material is similar to lubricating
oil (Black and High, 1978). A smaller fraction of organic matter is associated
with four-stroke diesel particles.
This oily matter contains a great number of organic chemicals, all with
rather low volatility. For example, the substance with the lowest molecular
weight conventionally measured in these mixtures contains 15 carbon atoms, and
the most abundant group of chemicals contains 21 to 25 carbon atoms (Black and
High, 1978). These substances have rather low vapor pressures and, partially
because of their association with the carbon in diesel soot, they exist primarily
as particles rather than gases under ambient conditions (Black and High,
1978).
Because of the large number of compounds, their molecular complexity,
identification, and measurement of individual species in the oily component of
diesel soot have been a challenging analytical problem. Recently, attempts
have been made to limit the problem by studying the biological activity of
various chemical fractions of the mixture (Huisingh et al. , 1978). Ames
bioassay results have suggested that the major mutagenic activity in such
mixtures is located in oxygenated subfractions (Huisingh et al., 1978). The
size distribution of diesel particulate matter suggests a mass median diameter
of about 0.2 urn (Dolan and Kittelson, 1979). These particles have low settling
velocities and thus remain suspended in the ambient air for long periods of
time. Their size permits respiratory system penetration and deposition (Schreck,
1978).
Figure 4-18 illustrates the relative emissions of specific particulate
components by vehicle type. Approximately 90 average unleaded gasoline passenger
4-105
-------
100
580
o
CTi
v>
z
o
V)
tn
5
UJ
"X
UJ
80
60
40
20
HDD2 -HEAVY DUTY DIESEL. 2-STROKE
HDD4 - HEAVY DUTY DIESEL. 4-STROKE
UJD -LIGHT DUTY DIESEL
LDGL - LIGHT DUTY GASOLINE. LEADED
LOGUL - LIGHT DUTY GASOLINE.
UNLEADED
TOTAL PARTICULATE
MATTER
LEAD
ORGANIC MATTER
SULFATE
Figure 4-18. Relative paniculate emissions are shown for vehicle categories (total and major components).
-------
cars would be required to emit total particulate mass per mile comparable to a
single two-stroke heavy-duty diesel truck under urban driving conditions.
Similarly, about 580 unleaded gasoline passenger cars would be required to
emit particulate organic mass per mile comparable to that of a two-stroke
heavy-duty diesel truck. About 10 catalyst-equipped passenger cars would be
required for sulfate emissions comparable to a single heavy-duty diesel.
Light-duty diesels and catalyst gasoline cars emit similar amounts of sulfate.
The heavy-duty and light-duty relative emission rates are important for esti-
mation of ambient particulate burden, but the work functions of the various
vehicle categories differ dramaticaly. On a per-passenger mile basis or
per-ton mile basis, the emission rates are not nearly so different in magni-
tude. Examination of the fuel-specific emissions (i.e. mass of total particu-
late matter emitted per unit of fuel combusted) indicates somewhat lower
particle emissions from heavy-duty diesels than from light-duty diesels.
Estimation of the atmospheric particulate loading from highway vehicles
requires specific factors detailing the emission rate per vehicle (in grams/
mile), the vehicle population, and the distance traveled. Based on these
factors, a scenario for the estimation of the national vehicle particulate
emission in thousands of metric tons per year has been developed (see Figure
4-19). This scenario assumes that light-duty diesel-powered passenger vehicles
will reach a 25 percent penetration of total U.S. sales by 1995 with no stricter
governmental regulation of particulate emissions. Figure 4-19 indicates that
for 1977, the total vehicular suspended particulate level is 345,000 metric
tons/year (engine and vehicle related) and is composed of 190,300 metric
tons/year of elemental carbon; 67,900 metric tons/year of soluble organic
compounds; 77,022 metric tons/ year of lead salts; and 9,600 metric tons/year
4-107
-------
800
>ELEMENTAL CARBON
......*.• SOLUBLE ORGANICS .*
••*••«»•••
•*••••••••
1977
1985
SULFATE^:
1990 1995
2000
YEAR
Figure 4-19. Total suspended particulate composition is projected, based
on best diesel estimate without particulate regulation.
4-108
-------
of sulfate. In this scenario, the total suspended participate level declines
from 1977 to 1985 and after reaching a mimimum in 1985, increased markedly as
the proportion of diesels in the vehicle population increases. By the year
2000, diesel particulate emissions (i.e., elemental carbon and soluble
organics) dominate the total suspended particulate composition, representing
greater than 90 percent of the total emitted particulate matter. Scenarios
utilizing different assumptions (e.g., estimates of diesel populations, future
governmental particulate regulations, relaxing of emissions standard) would
indicate different projected total suspended particulate levels. However,
the overall trend of diesel-powered vehicles dominating total suspended parti-
culate composition would remain unchanged.
4.4 SUMMARY
Sulfur oxides and particulate matter are emitted into the atmosphere
from a variety of sources, both natural and man-made. Natural souces of
these pollutants include terrestrial dust (windblown soil and rock
particles), radioactive aerosols, sea spray, volcanic emissions, biosphere
emanations (products of biological processes), and biomass burning.
Although natural contributions of sulfur oxides and particulate matter
to the atmosphere greatly exceed the total man-made contributions, natural
emissions are globally distributed whereas man-made emissions tend to be
much more concentrated.
Over 90 percent of the national sulfur oxides emsisions are in the form
of sulfur dioxide; the balance consists of sulfates in various forms. The
quantity and composition of these emissions vary from source to source and
depend upon several factors, including fuel characteristics, operating con-
ditions, and types of emissions-control equipment in use.
4-109
-------
Total man-made emissions of sulfur oxides in the United States are
currently calculated to be approximately 29 million metric tons per year from
stationary sources. Transportation sources contribute less than 1 million
metric tons of sulfur oxides, or not quite 3 percent of the total emissions.
Of the total stationary sources, utility plant combustion of coal and oil is
the primary contributor (62 percent), while other industrial processes account
for about 32 percent, and residential and commercial use of coal and oil
amounts to less than 3 percent of the total sulfur oxide emissions. These
patterns vary considerably from one region to another, with over 75 percent of
the total national emissions coming from the eastern half of the United States.
Historically, overall emissions of sulfur oxides have increased from about 23
million metric tons per year in 1940 to as high as 35 million metric tons per
year in 1973. Although the past few years have seen a slight decline in
emissions, an increase to nearly 40 million metric tons is expected by 1990.
As with sulfur oxides, the characteristics of particulate matter vary
according to type of source, emissions-control equipment, and other factors.
Overall man-made emissions of particulate matter amounted to about 12.5 million
metric tons per year in 1975 in the United States. The distribution of sources
varies geographically but, again, the eastern half of the United States domi-
nates by accounting for 66 percent of the national total. The nature of
sources for particulate matter is also varied. Forty percent of these
emissions come from stationary fuel combustion sources, principally electric
generation and industrial fuel use, and 25 percent from mineral processing
industries. The remainder is distributed among primary metal production, land
vehicles, food and agricultural processes, solid waste disposal, and other
sources. In addition, there are also considerable emissions of fugitive
particulate matter from both industrial (3.3 million metric tons/yr) and
non-industrial (409 million tons/yr) sources.
4-110
-------
4.4 REFERENCES
Adams, F. , R. Dams, L. Guzman, and J. W. Winchester, Background aerosol
composition on Chacaltaya Mountain, Bolivia. Atmos. Environ.
11:629-634, 1977.
Baes, C. F. , Jr., H. E. Goeller, J. S. Olson, and R. M. Rotty. The global
carbon dioxide problem. Report ORNL-5194, Oak Ridge National
Laboratory, Oak Ridge, TN, 1976.
Barrett, K. W. A Review of Standards of Performance for New Stationary
Sources-Portland Cement Industry. Prepared by MITRE Corporation for
the U.S. Environmental Protection Agency. EPA-450/3-79-012. March
1979.
Barringer, A. R. , U.S. Geological Survey Professional Paper 1015, 1977.
pp. 231-251. Airtrace--an airborne geochemical exploration technique.
Baylor, E. R., M. B. Baylor, D. C. Blanchard, L. D. Syzdek, and C. Appel,
Virus transfer from surf to wind, Science 198:575-580, 1977.
Beauford, W., J. Barber, and A. R. Barringer. Heavy metal release from
plants into the atmosphere. Nature 256:35-37, 1975.
Beauford, W. , J. Barber, and A. R. Barringer, Release of particles containing
metals from vegetation into the atmosphere. Science 195:571-573, 1977.
Benkoritz, Karmen M. Compiling a Multistate Emissions Inventory. In: Pro-
ceedings of a Specialty Conference on Emission Factors and Inventories.
West Coast Section, APCA, Anaheim, CA, November 1978. pp. 122-139.
Bennett, L. and K. T. Knapp. Sulfur and trace metal particulate emissions
from combustion sources. In: Workshop Proceedings on Primary Sulfate
Emissions from Combustion Sources. Volume 2, Characterization.
EPA-600/9-78-020b. U.S. Environmental Protection Agency, ORD, Research
Triangle Park, NC, August 1978. pp. 165-183.
Berg, W. W. Jr., and J. W. Winchester. Aerosol chemistry of the marine
atmosphere. In: Chemical oceanography, 2nd ed., vol. 7. J. P. Riley
and R. Chester, eds., Academic Press, London, 1978.
Black, F., and L. High. Diesel hydrocarbon emissions, particulate and
gas phase. Presented at Symposium on Diesel Particulate
Emissions Measurement and Characterization, Ann Arbor, MI,
May 1978.
4-111
-------
Blanchard, D. C., The electrification of the atmosphere by particles from
bubbles in the sea. Prog. Oceanogr. 1:71-202, 1963.
Blanchard, D. C., and B. C. Parker. In: The Aquatic Microbial Community,
vol. 15. J. Cairns, ed. , Garland, New York, 1977. pp. 625-653.
Bonn, R., T. Cuscino, and C. Cowherd. Fugitive Emissions from Integrated
Iron and Steel Plants. EPA-600/2-78-050. U.S. Environmental
Protection Agency, March 1978.
Bolin, B. Changes of land biota and their importance for the carbon cycle.
Science 195:613-615, 1977.
Bolin, B. On the role of the atmosphere in biogeochemical cycles. Q. J.
Meteorol. Soc. 105:25-42, 1979.
Bradow, R. L., F. M. Black, J. N. Braddock, C. T. Hare, M. Ingalls, and
J. M. Kawecki. Study of Particulate Emissions from Motor Vehicles.
A Report to Congress. Draft report prepared by Southwest Research
Institute, San Antonio, TX, under Contract No. 68-02-2951 and by
Biospherics Inc., Rockville, MD, under Contract No. 68-02-2926 to
the U.S. Environmental Protection Agency, Research Triangle Park,
N.C., 1979.
Broecker, W. S. Chemical Oceanography. Harcourt Brace Jovanovitch,
New York, 1974.
Buat-Menard, P., and M. Arnold. The heavy metal chemistry of atmospheric
particulate matter emitted by Mount Etna volcano. Geophys. Res.
Lett. 5:245-248, 1978.
Cadle, R. D.-Volcanic emissions of halides and sulfur compounds to the
troposphere and stratosphere. J. Geophys. Res. 80:1650-1652, 1975.
Carpenter, B. H. , and G. E. Weant III. Particulate Control for Fugitive
Dust. Report No. EPA-600/7-78-071. U.S. Environmental Protection
Agency, Research Triangle Park, NC, April 1978.
Casteleman, A. W. Jr., H. R. Munklewitz, and B. Manowitz. Istopic
studies of the sulfur component of the stratospheric aerosol layer.
Tellus 26:222-234, 1974.
Curtis, L. C., The influence of guttation fluid on pesticides. Phyto-
pathology 34:196-205, 1944.
Cooper, D. W., J. S. Sullivan, M. Quinn, R. C. Antonelli, and M. Schneider.
Setting Priorities for Control of Fugitive Particulate Emissions from
Open Sources. EPA-600/7-79-186. U.S. Environmental Protection Agency,
Washington, DC, August 1979.
4-112
-------
Cooper, J. A., and J. G. Watson. Portland Aerosol Characterization Study
(PACS). Application of Chemical Mass Balance Methods to the
Identification of Major Aerosol Sources in the Portland Airshed.
Oregon Graduate Center, Beaverton, OR, April 1979.
Cowherd, C. , Jr., et al., 1974
Cowherd, C., Jr., R. Bohn, and T. Cuscino, Jr. Iron and Steel Plant
Open Source Fugitive Emission Evaluation. EPA-600/2-79-103.
U.S. Environmental Protection Agency, Washington, DC, May 1979.
Council on Environmental Quality. Environmental Quality--9th Annual
Report of the Council on Environmental Quality. December 1978.
Crutzen, P. J. , L. E. Heidt, J. P. Krasnec, W. H. Pollack, and W. Seiler.
Biomass burning as a source of the atmospheric gases CO, H9, N,0,
NO, CH3C1, and COS. Nature in press, 1979. * L
Curtin, G. C., H. D. King, and E. L. Mosier. Movement of elements into
the atmosphere from coniferous trees in subalpine forests of
Colorado and Idaho, J. Geochem. Explor. 3:245-263, 1974.
Cuscino, T. A. Particulate Emission Factors Applicable to the Iron
and Steel Industry. Midwest Research Institute Project No.
4468-L(23). EPA Contract No. 68-02-2814, Task 23. Final Report.
August 7, 1979.
Dams, R., and J. De Jonge. Chemical composition of Swiss aerosols from
the Jungfraujoch. Atmos. Environ. 10:1079-1084, 1976.
Dannis, M. L. Rubber dust from the normal wear of tires. Rubber Chem.
Techno. 47:1011-1037, 1974.
Daugherty, D. P. , and D. W. Coy. Assessment of the Use of Fugitive
Emission Control Devices. EPA-600/7-79-045. U.S. Environmental
Protection Agency, February 1979.
Delany, A. C. , D. W. Parkin, J. J. Griffin, E. D. Goldberg, and B. E.
F. Reinman. Airborne dust collected at Barbados. Geochim.
Cosmochim. Acta 31:885-909, 1967.
Desy, D. H. Iron and Steel, Mineral Commodity Profile. Report No.
MCP-15, U.S. Department of the Interior, Washington, D.C. July
1978.
Dickerman, J.D., et al. Industrial Process Profiles for Environmental
Use Chapter 3, Petroleum Refinining Industry. EPA-600/2-77-023c.
U.S. Environmental Protection Agency, IERL, Research Triangle
Park, NC, January 1977.
4-113
-------
Dietz, R. N., R. F. Wieser, and L. Newman. Operating Parameters Affecting
Sulfate Emissions from an Oil-Fired Power Unit. In: Workshop
Proceedings on Primary Sulfate Emissions from Combustion Sources.
Vol. 2, Characterization. EPA-600/9-78-020b. U.S. Environmental
Protection Agency, ORD, Research Triangle Park, NC, August 1978.
pp. 239-270.
Dolan, D. F., and D. B. Kittelson. Roadway measurements of diesel exhaust
aerosols. SAE Paper No. 790492, Detroit, Mi, March 1979.
Drehmel, D. C., D. P. Daugherty, and G. H. Gooding. State of control
technology for industrial fugitive process particulate emissions.
In: Symposium on the Transfer and Utilization of Particulate
Control Technology. Vol. 4, Fugitive Dusts and Sampling, Analysis
and Characterization of Aerosols. EPA-600/7-79-044d. U.S.
Environmental Protection Agency, February 1979.
Duce, R. A., and E. J. Hoffman. Chemical tractionation at the air/sea
interface. Annu. Rev. Earth Planet. Sci. 4:187-228, 1976.
Duce, R. A., J. W. Winchester, and I. W. Van Nahl. Iodine, bromine,
and chlorine in the Hawaiian marine atmosphere. J. Geophys. Res.
70:1775-1799, 1965.
Eimutis, E. C., and R. P. Quill. Source Assessment: Overview Matrix for
National Criteria Pollutant Emissions. EPA-600/2-77-107c. U.S.
Environmental Protection Agency, Research Triangle Park, NC, 1977.
Federal Register, 44:49223, 49225-49226, 1979.
Federal Register, Regulation of fuels and fuel additives. 38:33, 734-33,
741, 1973.
Fish, B. R. Electrical generation of naturalaerosols from vegetation.
Science 175:1239-1240, 1972.
Friend, J. P. The global sulfur cycle. Ir\: Chemistry of the Lower Atmosphere,
S. I. Rasool, ed., Plenum Press, New York, 1973. pp. 177-201.
Gillette, D. A. On the production of soil wind erosion aerosols having the
potential for long range transport. J. Rech. Atmos. 8:735-744,
1974.
Goldberg, E. D. Rock volatility and aerosol composition. Nature 260:128-
129. 1976.
Granat, L., H. Rodhe, and R. 0. Hallberg. The global sulfur cycle. Ecol.
Bull. (Stockholm) 22:89-134, 1976.
4-114
-------
Griffin, J. J. , and E. D. Goldberg. The fluexes of elemental carbon in
coastal marine sediments. Limnol. Oceanogr. 20:456-463, 1975.
Groblicki, P. E. General Motors Sulfate Dispersion Experiment: Aerosol
Sizing Measurements. Publication No. GMR-2127, EV-28. Prepared by
General Motors Research Laboratories, Warren, MI to the Environmental
Protection Agency Symposium on General Motors Sulfate Dispersion
Experiment, Research Triangle Park, NC. April 1976.
Hall, W. B., and R. E. Ela. Cement and Mineral Commodity Profiles. Report
No. MCP-26, U.S. Department of the Interior, Bureau of Mines, November
1978.
Hare, C. T., and R. L. Bradow. Characterization of Heavy-Duty Diesel Gaseous
and Particulate Emissions, and Effects of Fuel Composition. See Paper
No. 790490. Detroit, MI, February 26-March 2, 1979.
Hidy, G. M. et al. Design of the Sulfate Regional Experiment (Sure). Vol I:
Supporting Data and Analysis. Electric Power Research Institute,
Palo Alto, CA, February 1976.
Hitchcock, 0. R. Biogenic contributions to atmospheric sulfate levels.
Presented at the Second Annual Conference on Water Reuse, Chicago,
1975.
Holmes, 1971.
Holser, W. I. and I. R. Kaplan. Isotope geochemistry of sedimentary
sulphates. Chem. Geol. 1:93-135, 1966.
Holser, et al., 1966.
Homolya, J. B., H. M. Barnes, and C. R. Fortune. A characterization of the
gaseous sulfur emissions from coal and oil-fired boilers. In: Proceedings
of the Fourth National Conference on Energy and the Environment. AICHE,
New York, 1976. pp. 490-494.
Homolya, J. B. , and J. L. Cheney. An assessment of sulfuric acid and sulfate
emissions from the combustion of fossil fuels. I_n: Workshop Proceedings
on Primary Sulfate Emissions from Combustion Sources. Volume 2,
Characterization. EPA-600/9-78-020b, U.S. Environmental Protection
Agency, ORD, Research Triangle Park, NC, August 1978. pp. 3-11.
Homolya, J.'B., and J. L. Cheney. A study of primary sulfate emissions
from a coal-fired boiler with FGD. J. Air. Pollut. Control Assoc.,
vol. 29, September 1979.
Homolya, J. B. , and J. L. Cheney. Primary Sulfate Emissions from Oil
Combustion. U.S. Environmental Protection Agency, Research Triangle
Park, NC 1978.
4-115
-------
Husingh, J. et al. Application of bioassay to the characterization of
diesel particle emissions. lr\: Symposium on Application of Short-
Term Bioassays in the Fractionation and Analysis of Complex Environ-
mental Mixtures. EPA 600/9-78-027. U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1978.
International Symposium on the Chemistry of Sea/Air Particulate Exchange
Processes (collected papers). J. Rech. Atmos. 8:501-1007, 1974.
Jacko, M. G. , and R. T. DuCharme. Brake emissions—emission measurements
from brake and clutch linings from selected mobile sources. Report
68-04-0020, prepared by Bendix Research Laboratories, Southfield,
MI. U.S. Environmental Protection Agency, Ann Arbor, MI. March 1973.
Johansson, I. B., R. E. Van Grieken, and J. W. Winchester. Elemental
abundance variation with particle size in north Florida aerosols.
J. Geophys. Res. 81:1039-1046, 1976.
Junge, C. E. Air Chemistry and Radioactivity, Academic Press, New York,
1963.
Junge, C. E. Chemical analysis of aerosol particles and of gas traces
on the island of Hawaii. Tellus 9:528-537, 1957.
Junge, C. E., C. W. Chagnon, and J. E. Manson. Stratospheric aerosols.
J. Meteorol. 18:81-108, 1961.
Jutze, G. A., et al. Technical Guidance for Control of Industrial Process
Fugitive Particulate Emissions. EPA-450/3-77-0110. U.S. Environ-
mental Protection Agency, March 1977.
Katari, V. S., and R. W. Gerstle. Industrial Process Profile for Environ-
mental Use. Chapter 24, The Iron and Steel Industry. Prepared by
Radian Corp. for the U.S. Environmental Protection Agency, EPA-600/
2-77-023x, February 1977.
Kawecki, J. M. Emission of Sulfur-Bearing Compounds from Motor Vehicle
and Aircraft Engines. A Report to Congress. Prepared by Biospherics,
Inc., Rockville, MD, under Contract No. 68-02-2926 to the U.S.
Environmental Protection Agency, Washington, DC. Publication No.
EPA-600/9-78-028. August 1978.
Kellogg, W. W., R. D. Cadle, E. R. Allen, A. L. Lazrus, and E. A. Martell.
The sulfur cycle. Science 175:587-596, 1972.
Khan, Z. S., et al. Source Assessment: Asphalt Hot Mix. Prepared by
Monsanto Research Corp. for the U.S. Environmental Protection Agency.
EPA-600/2-77-107n (PB-276-731). December 1977.
4-116
-------
Khatri, N. J. , and J. H. Johnson. Physical size distribution character-
ization of diesel particulate matter and the study of the coagulation
process. SAE Paper No. 780788. September 1978.
Kolnsberg, H. J. A Guideline for the measurment of air-borne fugitive
emissions from industrial sources. In: Symposium on Fugitive
Emissions Measurement and Control. "ETA-600/2-76-246. U.S. Environ-
mental Protection Agency, September 1976.
Koyama, T. E. , N. Nakai, and E. Kamata. Possible discharge rate of hydrogen
sulfide from polluted coastal belts in Japan. J. Earth Sci. 13:1-11,
1965.
Lawson, D. R., and J. W. Winchester. A standard crustal aerosol as a
reference for elemental enrichment factors. Atmos. Environ. 13:925-
930, 1979.
Lawson, D. R. , and J. W. Winchester. Atmospheric sulfur aerosol concen-
trations and characteristics from the Sourth American continent.
Science 205:1267-1269, 1979a.
Lawson, D. R. , and J. W. Winchester. Sulfur and trace element relationships
in aerosols from the South American continent. Geophys. Res. Lett.
5:195-198, 1978.
Lawson, D. R., and J. W. Winchester. Sulfur, potassium, and phosphorus
associations in aerosols from South American tropical rain forests.
J. Geophys. Res. 84:3723-3727, 1979b.
Leavitt, C. et al. Environmental Assessment of Coal and Oil-Firing in a
Controlled Boiler, Vol. II, Comparative Assessment. Prepared by TRW,
Inc., for the U.S. Environmental Protection Agency. EPA-600/7-78-164b.
U.S. Environmental Protection Agency, Research Triangle Park, NC
August 1978.
Lee, Robert E. , Jr., and F. V. Duffield. Sources of Environmentally Important
Metals in the Atmosphere. _In: Ultratrace Metal Analysis in Biological
Sciences and Environment, A Symposium, 174th meeting, American Chemical
Society, Chicago, Illinois, August 29-30, 1977. Advances in Chemistry
Series, American Chemical Society, Washington, DC, 1979. pp. 146-171.
Lee, S. D., M. Malanchuk, and V. N. Finelli. Biologic effects of auto
emissions. 1. Exhaust from engine with and without catalytic
converter, J. Toxicol. Environ. Health 1:705-712, 1976.
Lebowitz, M. F. Short-term testing for fugitive dust effect. Presented at
the 68th Air Pollution Control Association Meeting, Boston, MA. Paper
No. 75-25.4. June 1975.
4-117
-------
Littman, Fred E. Regional Air Pollution Study. Emission Inventory Summari-
zation. EPA-600/4-79-004. U.S. Environmental Protection Agency, ORD,
Research Triangle Park, NC, January 1979.
Lovelock, J. E. CS7 and the natural sulfur cycle. Nature 248:625-626.
1974. i
Lovelock, J. E., R. J. Maggs, and R. A. Rasmussen. Atmospheric dimethyl
sulfide and the natural sulfur cycle. Nature 237:452-453, 1972.
Lynn, D. A., G. L. Deane, R. C. Galkiewicz, and R. M. Bradway. National
Assessment of the Urban Particulate Problem. Volume 1, Summary of
National Assessment. Report No. EPA-450/3-76-024. U.S. Environ-
mental Protection Agency, Research Triangle Park, NC June 1976.
Maclntyre, F. Chemical fractionation and sea-surface microlayer processes.
In: The Seas, vol. 5. E. D. Goldberg, ed., Wiley, New York, 1974.
Maenhaut, W., W. H. Zoller, R. A. Duce, and G. L. Hoffman. Concentration
and size distribution of particulate trace elements in the South
Polar atmosphere, J. Geophys. Res. 84:2421-2431, 1979.
MarteH , E. A., and H. E. Moore. Tr'opospheric aerosol residence times:
a critical review. J. Rech. Atmos. 8:903-910, 1974.
McCurley, W. R., and Daryl G. DeAngelis. Measurement of sulfur oxides
from coal-fired utility and industrial boilers. In: Workshop
Proceedings on Primary Sulfate Emissions from Combustion Sources.
Vol. 2, Characterization. EPA-600/9-78-020b. U.S. Environmental
Protection Agency, ORD, Research Triangle Park, NC, August 1978.
pp. 67-85.
McFarland, A. R., C. A. Ortiz, and C. E. Rodes. Characteristics of aerosol
samplers used in ambient air monitoring. Presented at the 86th
National Meeting of the American Institute of Chemical Engineers,
Houston, TX April 1979.
Moorby, J., and H. M. Squire. The loss of radioactive isotopes from the
leaves of plants in dry conditions. Radiat. Bot. 3:163-167, 1963.
Moran, J. B., 0. Manary, R. Fay, and M. Baldwin. Development of Particulate
Emission Control Techniques for Spark-Ignition Engines. Final Report
prepared by Organic Chemicals Department, The Dow Chemical Co., Midland,
MI, under Contract EHS70-101. U.S. Environmental Protection Agency,
Ann Arbor, MI, July 1971.
Moyers, J. L., and R. A. Duce. Gaseous and particulate bromine in the marine
atmosphere. J. Geophys. Res. 77:5330-5338, 1972a.
Moyers, J. L. , and R. A. Duce. Gaseous and particulate iodine in the marine
atmosphere. J. Geophys. Res. 77:5229-5238, 1972.
4-118
-------
Natusch, D. F. S. Characterization of fly ash from coal combustion. In:
Workshop Proceedings on Primary Sulfate Emissions from Combustion
Sources. Vol. 2, Characterization. EPA-600/9-78-020b. U.S.
Environmental Protection Agency, ORD, Research Triangle Park, NC
August 1978. pp. 149-163.
Nemeryuk, G. E. Migration of salts into the atmosphere during transpiration.
Sov. Plant Physiol. 17:560-566, 1970.
Nelson, K. W. , M. 0. Varner, and T. J. Smith. Nonferrous metallurgical
operations. In: Air Pollution, (3rd ed.) vol. IV, A. C. Stern,
ed., Academic~Press, New York, 1977.
Pierson, W. R., and W. W. Brachaczek. Airborne particulate debris from
rubber tires. Rubber Chem. Technol. 27:1275-1299, 1974.
Rahn, K. A. The chemical composition of the marine aerosol. Technical
Report, University of Rhode Island, Kingston, July 1976.
Rahn, K. A. Sources of trace elements in aerosols—an approach to clean
air. Ph.D. Thesis, University of Michigan, Ann Arbor, 1974.
Rasmussen, R. A. Emission of biogenetic hydrogen sulfide. Tellus 26:254-
260, 1974.
Rasmussen, R. A., and F. W. Went. Volatile organic material of plant origin
in the atmosphere. Proc. Natl. Acad. Sci. U.S.A. 53:215-220, 1965.
Reed, A. H. Stone, Mineral Commodity Profiles. U.S. Department of the
Interior, Bureau of Mines. Report No. MCP-17, July 1978.
Reider, J. P., and C. Cowherd, Jr. Inhalable Particulate Emission Factor
Assessment and Development. Final Report, EPA Contract No. 68-02-2609,
Assignment No. 11. U.S. Environmental Protection Agency, Research
Triangle Park, NC April 1979.
Ryan, P. W., and McMahon. Some chemical and physical characteristics of
emissions from forest fires. Paper 76-2.3, presented at the 69th
Annual Meeting of the Air Pollution Control Association, Portland,
OR, June 22-July 1, 1976.
Shcreck, R. M. Health Effects of Diesel Exhaust. Biomedical Sciences
Dept., General Motors Research Laboratories, Warren, MI, 1978.
Schnell, R. C., and G. Vali. World-wide source of leaf-derived freezing
nuclei. Nature 246:212-213, 1973.
Schnell, R. C., and G. Vali. Atmospheric ice nuclei from decomposing
vegetation. Nature 236:163-165, 1972.
4-119
-------
Scinto, L. L. Primary Sulfate Emissions from Coal and Oil Combustion-
Rough Draft.. Prepared by TRW, Inc, for the U.S. Environmental
Protection Agency, IERL, Research Triangle Park, NC, August 1979.
Seiler, W. , and P. J. Crutzen. Estimates of gross and net fluexes of
carbon between the biosphere and the atmosphere from biomass burning.
Climatic Change, in press, 1979.
Smith, D. M., J. J. Griffin, and E. D. Goldberg. Elemental in marine
sediments: a baseline for burning. Nature 241:268-270, 1973.
Spawn, P. Letter report submitted under Contract No. 68-02-2687, Technical
Directive No. 007. September 20, 1979.
Springer, J. J. Current and future trends of exhaust emissions from HD
highway vehicles. Presentation to Federally Coordinated Program
of Highway Research and Development, Federal Highway Administration,
College Park, MD. October 1978.
Steiner, B. A. Ferrous metallurgical operations. In: Air Pollution,
(3rd ed.), vol. IV. A. C. Stern, ed. , AcademicTress, New York,
1977.
StOiber, 1973.
Stoiber, R. E., and A. Jepson. Sulfur dioxide contributions to the
atmosphere by volcanoes. Science 182:2577-578, 1973.
Suprenant, N. F. et al. Emissions Assessment of Conventional Stationary
Combustion Systems. Volume 1, Gas and Oil-Fired Residential Heating
Sources Prepared by TRW, Inc., for the U.S. Environmental Protection
Agency. EPA-600/7-79-020b, U.S. Environmental Protection Agency, IERL,
Research Triangle Park, NC, May 1979.
Surprenant, N. et al. Preliminary Emissions Assessment of Conventional
Stationary Combustion Systems. Vol. II, Final Report. Prepared by
TRW, Inc., for the U.S. Environmental Protection Agency. EPA-600/2-76-
046b. U.S. Environmental Protection Agencies, Research Triangle
Park, NC, March 1976.
Sussman, V. H. Mineral product industries. In: Air Pollution, 3rd. ed.,
vol. IV. A. C. Stern, ed., Academic Press, New York 1977.
Timsen, D. J., and P. W. Aften. A Proposed Design for Grain Elevator Dust
Collection.. S. Air Pollut. Control Admin. 18:738, 1968.
TRW, Inc. Guideline for Development of Control Strategies in Areas with
Fugitive Dust Problems. OAOPS Report No. 1.2-71. U.S. Environmental
Protection Agency, Research Triangle Park, NC, October 1977.
U.S. Department of Commerce, Bureau of the Census. Statistical Abstract of
the United States - 1977. U.S. Department of Commerce, Washington, DC.
4-120
-------
U.S. Department of Energy. Annual Report to Congress, Vol. II. DOE/EIA-
0036/2, U.S. Department of Energy, Washington, DC, 1977.
U.S. Department of Energy. Annual Report to Congress, 1978, Volume III.
DOE/EIA-0173/3. U.S. Department of Energy, Energy Information Adminis-
tration, Washington, D.C. 1979a. (The model utilized in these air
emissions projections is the Regional Emission Projection Systems.
REPS is documentated in "An Air Emissions Analysis of Energy Projections
for the Annual Report to Congress," a memorandum by Edward Pechan
dated September 1978).
U.S. Department of Energy, 1979b.
U.S. Department of Energy. 1985 Air Pollution Emissions. DOE/PE-0001.
U.S. Department of Energy, Washington, DC, December 1977.
U.S. Department of the Interior, 1975.
U.S. Department of the Interior, Bureau of Mines, Division of Fuels Data.
Carbon black in 1976. Iji Mineral Industry Surveys, 1977.
U.S. Department of the Interior. Minerals Yearbook 1975. Vol. I, Metals,
Minerals, and Fuels. U.S. Department of the Interior, Washington,
DC., 1977.
U.S. Environmental Protection Agency. An Assessment of the Health Effects
of Coke Oven Emissions. External Review Draft. U.S. Environmental
Protection Agency, Research Triangle Park, NC, April 1978.
U.S. Environmental Protection Agency. Background Information for New Source
Performance Standards: Primary Copper, Zinc, and Lead Smelters. Vol. 1,
Proposed Standards. EPA-450/2-74-002a. U.S. Environmental Protection
Agency, OAQPS, Research Triangle Park, NC, October 1974.
U.S. Environmental Protection Agency. Unpublished data, 1976.
U.S. Environmental Protection Agency. Compilation of Air Pollutant Emission
Factors, 3rd ed. Publication AP-42. U.S. Environmental Protection Agency,
OAQPS, Research Triangle Park, NC, August 1977.
U.S. Environmental Protection Agency. National Air Quality and Emissions
Trends Report 1977. Publication No. 450/2-78-052. U.S. Environmental
Protection Agency, Research Triangle Park, NC, December 1978a.
U.S. Environmental Protection Agency. National Air Data Branch, Monitoring
and Data Analysis Division. 1975 Narional Emissions Report. EPA-450/2-
78-020, U.S. Environmental Protection Agency, Research Triagnel Park, NC,
May 1978b. pp. vii and 1.
U.S. Environmental Protection Agency. National Air Pollutant Emission Estimates,
1940-1976. EPA-450/1-78-003. U.S. Environmental Protection Agency,
July 1978c.
4-121
-------
U.S. Environmental Protection Agency, 1979.
U.S. Environmental Protection Agency. Position Paper on regulation of
atmospheric sulfates. Report EPA-450/2-75-007. U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1975.
Vandegrift, A. E. et al. Participate Pollutant System Study. Volume III,
Handbook of Emission Properties. NTIS Report No. PB-203-522. May 1971.
Weart, T. E. et al. Fine Particulate Emission Inventory and Control Survey.
Prepared by Midwest Research Institute for the U.S. Environmental
Protection Agency. EPA-450/3-74-040. U.S. Environmental Protection
Agency, Research Triangle Park, NC, January 1974.
Went, F W. Organic matter in the atmosphere, and its possible relation to
petroleum formation. Proc. Natl. Acad. Sci. U.S.A. 46:212-221, 1960.
Went, F. W. , D. E. Slemmons, and H. N. Mozingo. The organic nature of
atmospheric condensation nuclei, Proc. Natl. Acad. Sci. U.S. 58:69-74,
1967. ~~
Wexler, H. On the effects of volcanic dust on insolation and weather.
Bull. Am. Meteorol. Soc. 32:10-15, 1951.
Wexler, H. Spread of the Krakotoa volcanic dust cloud as related to the
high-level circulation. BU11. Am. Meteorol. Soc. 32:48-51, 1951.
Whitby, K. T. , and B. Cantrell. Atmospheric aerosols—characteristics and
measurements. Session 29, Fine Particles. In: International Conference
on Environmental Sensing and Assessment, Las Vegas, NV, September
14-17, 1975. Institute of Electrical and Electronics Engineers,
New York, 1976.
Winchester and Duce, 1977.
Winchester, J. W., R. J. Ferek, D. R. Lawson, J. 0. pilotte, M. H. Thiemens,
and L. E. Wangen. Comparison of aerosol sulfur and crustal element
concentrations in particle size fractions from continental U.S. locations,
Water Air Soil Pollut., in press, 1979.
Woodruff, N. P., and F- H. Siddoway. A wind erosion equation. Soil Sci.
Soc. Am. Proc. 29:602-608, 1965.
Woodwell, G. M., R. H. Whittaker, W. A. Reiners, G. E. Likens, C. C. Delwiche,
and D. B. Botkin. The biota and the world carbon budget. Science 199:
141-146, 1978.
Working Symposium on Sea-to-Air Chemistry (collected papers). J. Geophys.
Res. 77:5059-5349, 1972.
4-122
-------
Zoller, J. , T. Bertke, and T. Janzen. Assessment of Fugitive Participate
Emission Factors for Industrial Processes. EPA-450/3-78-107. U.S.
Environmental Protection Agency. Research Triangle Park, NC, September
1978.
Zoller, W. H. E. S. Gladney, and R. A. Duce. Atmospheric concentrations and
sources of trace metals at the South Pole. Science 183:198-200, 1974.
4-123
-------
5. ENVIRONMENTAL CONCENTRATION AND EXPOSURE
5.1 DIMENSION OF EXPOSURES TO PARTICLES AND SULFUR DIOXIDE
The impact of air pollution on health and welfare is a function of dose
delivered to the receptor and the ability of the receptor to cope with the
resultant stress. It is almost impossible to measure directly the air pollution
dose to a population or even an individual except in the laboratory. Most often
the dose is estimated. This is particularly true for sulfur dioxide and airborne
particles. The stress experienced by a critical organ or receptor tissue from
particle inhalation depends on particle size, composition, morphology, acidity
or basicity, and other physical-chemical properties of the aerosol. The delivered
dose is also a function of the anatomical features of the receptor as well as
manner of breathing, breathing rate, and integrity of bodily defense systems.
Throughout their lifetime people inhale a complex mixture of gases and
particles. Other segments of the environment such as vegetation, materials, ect.
are also exposed to the same complex mixtures. Its composition varies with time
at any given location because of changing atmospheric conditions. The mixture
of particles and gases comes from a variety of sources (either directly or after
partial reaction), and the composition is therefore quite different at different
.locations.
As an alternative to direct measurement of dose, exposure could be used as
an approximation for studies on air pollution risk and effects. The exposure-
response relationship for air pollution is most important for establishing standards
Unfortunately, to extrapolate from measurements of ambient levels at a few
locations to an individual or population exposure is a very difficult task at
present. The outdoor air's contribution to indoor concentrations is still being
5-1
-------
investigated. The additional exposures to gases and particles from nonoccupa-
tional indoor sources are not adequately quantified.
Indoor air quality and activity patterns complicate air pollution exposure
estimates and are discussed later in this chapter. First, the ambient outdoor
concentrations of particulate matter and sulfur dioxide are examined. The
status of national air quality is presented for TSP (total suspended particulate
matter) and SOp. This first section discusses mass concentration measurements
of TSP, including information by size fraction. Both particle size and com-
position are important characteristics of atmospheric aerosols. The relation-
ships between ambient exposures to pollutants are presented in the concluding
section after a discussion of the relationship between indoor and outdoor
concentrations.
5.2 AMBIENT MEASUREMENTS
5.2.1 Ambient TSP Concentrations
The accuracy and precision of TSP monitoring are limited by three general
considerations: (1) sampling methods, including instrumentation, analytical
methods, and quality assurance; (2) sampling frequency; and (3) location of
monitors.
Chapter 3 discusses the first of these considerations; the second and third
are discussed in this chapter. Sampling frequency affects the confidence limits
on mean TSP concentrations and annual or seasonal trends. It is appropriate to
discuss this limitation at the beginning of this section before the 1978 national
TSP data base is presented. The siting of TSP monitors influences significantly
the levels measured, and hence the interpretation of data. These considerations
are presented with examples in several sections of this chapter.
In 1978 there were 4105 TSP monitoring sites in the United States and its
territories that reported data to the National Aerometric Data Bank'(NADB) of
5-2
-------
the U.S. Environmental Protection Agency. Of these, only 2882 had enough
observations per quarter and per year for the data to be considered valid
for estimating annual averages.
Table 5-1 lists a number of such sites for each State and territory. The number
of sites reporting valid data range from zero in Delaware and American Samoa to
318 in Ohio. The most populous state, California, had 60 and New York had 236.
The U.S. Environmental Protection Agency has established a uniform sampling
schedule to be followed by all State and local agencies. It requires a 24-hr
sample (midnight to midnight) every 6th day. Hence, in 1 year there are 60 or
61 possible sampling days from which to derive the mean value and distribution,
and to determine attainment of standards. In 1978 14 percent of all
reporting sites had 60 or more observations.
Sampling days are missed and samples must be voided for a variety of
reasons. Therefore, a minimum requirement has been established for considering
the data from any site as valid in determining an annual average: there must be
at least five observations during each quarter of a calender year and at least
26
& observations for the year. Of the Federal, State, and local TSP sites
reporting data to NADB, 70 percent met this requirement.
The distribution of observations for the 2882 valid monitoring sites in
1978 is shown in Figure 5-1. Ten (10) percent of these sites had less than 47
observations; 50 percent had more than 56 observations. However, 80 percent of
the monitoring sites collected fewer than 60 samples. Three percent of the
monitors sampled at an equivalent frequency of 1 day in 3, and fewer than 2
Percent collected samples at a frequency of 1 day in 2.
The NAAQS for TSP is expressed as an annual geometric mean and as a once-per-year
daily value. Frequency of monitoring is a fundamental parameter of the air
5-3
-------
TABLE 5-1. VALID MONITORING SITES OPERATING IN 1978
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvana
Puerto Rico
Rhode Island
South Carolina
South Dakota
Number
of
Valid Sites
76
15
9
37
60
61
36
NA
8
105
37
9
29
131
47
50
22
82
27
22
52
37
124
54
15
27
21
37
20
23
79
56
236
83
18
318
26
37
112
10
6
46
10
Percent
of total
3
1
<1
1
2
2
1
NA
<1
4
1
<1
1
5
2
2
1
3
1
1
2
1
4
2
1
1
1
1
1
1
3
2
8
3
1
11
1
1
4
<1
<1
2
<1
Percent
of U.S. population
1.70
0.19
1.05
0.99
10.03
1.20
1.45
0.27
0.33
3.90
2.32
0.41
0.39
5.21
2.47
1.34
1.07
1.60
1.81
0.50
1.92
2.70
4.25
1.84
1.10
2.23
0.35
0.72
0.29
0.38
3.42
0.55
8.41
2.54
0.30
4.98
1.29
1.08
5.50
NA
0.44
1.32
0.32
5-4
-------
TABLE 5-1. (continued)
State
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Guam
Virgin Islands
Number
of
Valid Sites
90
148
20
4
115
47
13
107
22
NA
3
3
Percent
of total
3
5
1
1
4
2
1
4
1
NA
<1
<1
Percent
of U.S. population
1.97
5.87
0.57
0.22
2.35
1.68
0.85
2.15
0.18
NA
NA
NA
aNA, Not applicable.
5-5
-------
1000
900 -'
800 -
7001—
600
500
400
300
z
<
z
z
<
&
o
u.
V)
K
LU
tfl
03
O
E
Si
u.
O
ffi
ui
CO
2S
200
100
90
80
70
60
50
40
30
20
I I
I I
I I
mi I I I I i i I
I I
till
I
I
I
I I
0.01 0.1 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 99.99
% WITH NUMBER OF OBSERVATIONS LESS THAN
Figure 5-1. Distribution of the number of observations in 1978 per valid site is for a total of 2882 sites.
5-6
-------
quality data used for comparison with the standards. The period of determining
an annual average for comparison is a calendar year. If the number of 24-hr
observations is less than 365, then the true mean concentration for the year can
only be expressed as residing within a range of values. By assuming that the
actual distribution of values is log normal, then confidence intervals can be
calculated from the geometric mean and the standard deviation. Figure 5-2 shows
the effect of sample size on the 95 percent confidence intervals for a hypo-
thetical site whose true annual geometric mean is equivalent to the standard.
From this example we can only conclude that the annual geometric mean is 75 +_ 13
3
yg/m , if the mean for that year was determined from a sample size of 61.
Increasing the sampling frequency to 1 day in 2 reduces the level of uncertainty
o
so that the annual mean can be expressed as 75 +_ 3 yg/m . Hunt (1972) demonstrated
that the confidence intervals around the geometric annual mean were broadened as
the sampling size decreased and as the standard geometric deviation increased.
A critical factor in evaluating compliance with once-a-year standards is
the effect of sampling frequency. Figure 5-1 shows that in 1978 the majority of
valid sites (80 percent) had fewer than 60 sampling days. The sites with more
frequent sampling had a greater chance of sampling the higher concentrations.
Table 5-2 illustrates this point. Assuming that there are a number of days on
which the observations are above the standards, the probability of selecting 2
or more days on which standards are exceeded is a function of sampling frequency.
If there are 10 days above the standards, there is only a slightly better than
50 percent chance of actually monitoring on 2 of those days given a sampling
frequency of 61 out of 365 days. When the sampling frequency is doubled to 122
sampling days, the probability of capturing 2 days out of 10 that exceed the
standards increases to 80 percent. This analysis serves to show the relative
5-7
-------
95
90 -
* 85
O
a 80
2
UJ
O
I 75
a.
40
70
65
D
ABOVE THE STANDARD
AREA OF.
UNCERTAINTY
BELOW THE STANDARD
61 91 122 183 365
NUMBER OF SAMPLING DAYS PER YEAR
Figure 5-2. The 95 percent confidence intervals about the annual
primary standard for TSP is shown for various sampling frequencies
(assume the standard geometric deviation equals 1.6).
5-8
-------
TABLE 5-2. PROBABILITY OF SELECTING TWO OR
MORE DAYS WHEN SITE EXCEEDS STANDARD
Actual number Sampling frequency, days/year
of excursions
61
2 0.03
4 0.13
6 0.26
8 0.40
10 0.52
12 0.62
14 0.71
16 0.78
18 0.83
20 0.87
22 0.91
24 0.93
26 0.95
122
0.11
0.41
0.65
0.81
0.90
0.95
0.97
0.98
0.99
0.99
0.99
0.99
0.99
183
0.25
0.69
0.89
0.96
0.99
0.99
0.99
0.99
0.99
0.99
0.99
0.99
0.99
365
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
5-9
-------
changes in the magnitude of the problem relating to sampling frequencies. In
actuality, samples are not taken randomly; they are taken systematically, usually
at a rate of once every 6 days. The problem is further complicated in that the
log normal distribution of TSP concentrations does not apply uniformly to all
sites.
The 1978 TSP data were examined to determine whether there was a relation-
ship between the number of observations and the maximum concentration observed.
Figure 5-3 is a plot of the maximum concentration versus the number of obser-
vations for 1448 sites. No relationship is apparent. For the 37 sites having
maximum 24-hr values exceeding 450 ug/m , 21 are reported from sites with fewer
than 60 observations. Where there were fewer than 35 observations, no high
values were found. This comparison is not totally valid because each monitor is
sampling from a different set of values. It does indicate that analyses using
the aggregated data sets may not be adversely affected by differences in obser-
vations. However, comparing sites on the basis of their violating the once-per-
year standard for TSP without regard to sampling frequency, is questionable.
An additional complication occurs when the meteorological regimes affecting
the TSP concentrations are considered. Attainment of standards may depend on
the number of "clean" sampling days versus the number of "dirty" sampling days.
Watson (1979) exemplifies this problem with Portland, OR, TSP data. The annual
geometric mean TSP data show a decreasing trend from 1973 through 1975, with a
significant increase in 1976. If these data are reexamined and weighted by the
meteorological regimes actually sampled in each year, the conclusions are changed.
The stratified mean TSP values show a large drop in concentrations occurring
between 1974 and 1975, with the levels being constant for the year before
and after. Since these means are determined from a sample set varying from 49
5-10
-------
40.00 74.00 108.00 142.00 176.00 210.00 244.00 278.00 312.00 346.00
1447.00
1305.00
1163.00
1021.00
•)
a.
§ 879.00
H
OC
5 737.00
o
8
1- 595.00
5
|
| 453.00
311.00
169.00
127.00
_ 1 1 1 1 1 ,1
••
-
— • *
t
— • «
2
• 2
• «2 * • •
«23« • • •
« • 22 • •
• • 344 • • •
• 22*4 * * •• •
— »»»2 131 2 • »2 2
*23iBS *• * 3 •* • *•
•227?»»» 2«*» • •••
23W» 2* 3 3«*3*
* «*H?m?*22 » 2* 2 »2 »3
-••• 43»m?J« 2 3»«22»» 3»» «
2 2JWW«»» • 42« 4 42* •
••2 237W7?* 24* 2« »
••«324TW» 3 • «
•2324334 3
~ 1 *l 1* 1 1 1
1 1 1 1 1 1 I
*
*
*
1 1 1 1 1 1 _
—
—
—
—
—
—
* *
—
» »
( —
•
« * • •** » •
•*
•
2 * * • • •
*•
1 1 1 1 1 1 1
* * •
* • *
* *
1 1 1 1 1 1 ~
23.00 57.00 91.00
125.00 159.00 193.00 227.00 261.00 295.00 329.00 363.00
NUMBER OF OBSERVATIONS FOR 1978
Figure 5-3. In this scattergram of maximum TSP concentration against number of observations in 1978,
the total number of sites is 1448. The number 9 indicates nine or more; hence, the numbers do not sum
to the total.
5-11
-------
samples in 1974 to 79 samples in 1976, a statistical test is required to deter-
mine whether the means of any year are significantly different from those of any
other year. The 95 percent confidence intervals for all stratified means do
overlap. Watson concludes, then, that there is a reasonable chance that the
true means do not really vary from year to year.
The choice of sampling location can obviously affect the concentrations
measured. Remote sites are chosen specifically to gain understanding of back-
ground levels of either natural or transported suspended particulate matter. At
a specific location, the height of the monitor above the ground influences
concentrations. If the monitor is elevated above surface sources, lower concen-
trations might be expected. The elevations of the 1978 valid sites ranged from
ground to 748 feet above surface, with a mean height of 23 feet. Only 5 percent
of all monitors were higher than 40 feet above grade. Table 5-3 gives the
height distribution of monitors.
A scattergram (Figure 5-4) of the site median value against the site
elevation suggests that larger values might be associated with lower monitoring
sites. However, careful inspection of the cross-tabulation of the median value
by elevation (Table 5-4) and the cross-tabulation (Table 5-5) of the 90th
percentiles by elevation fails to reveal any significant relationship. The cut
points for the median and 90th percentiles were preselected. The intent was to
group all the national TSP monitoring sites by, in percentages, first 5, next
20, next 25, next 25, next 20, and final 5. This approximate breakdown can be
seen in the column on the right labeled "ROW TOTAL." To determine whether
elevation might be affecting concentrations, inspect two rows, 70 to 97 yg/m
2
and >97 yg/m . If the monitors closer to the ground were actually systematically
"seeing" a greater mass of suspended particles, then for both rows the column
percent (third number in grid) would be higher in the first and second columns
5-12
-------
9.00
25.00 41.00 57.00 73.00 89.00 105.00 121.00 137.00 153.00
187.00
169.00
151.00
133.00
n
a
a.
h
0 115.00
5
C
5 97.00
8
0.
M
•- 79.00
z
5
5 61.00
43.00
25.00
7.00
_ I I I I 1 1
'
**
* *
— *
* * *
• • *
* •
— * *
• *»
•» 2
• »* 2
- » 3 *
•* « 2 • *
• 3 2 *
» • 33
• 2 4*** • * •
2* * 2 » «
2 .24 • « • » 2 • »
3*2 3 «2 *
«* 2228 •• • 3* • • • •
* • 3224 f»24 • 3 3 *2 •
-•• *2*35 2* 2 3* * 2*
23 3424 •» 332*8** • • • *
**4* *44f 8 23 22 3 • 2
373 *43*27**«8 3*2 4 3 •• * * •'
-•23* 3728* 32 8 8* 2 3 2*
4434 8W«2»2 822f *3 » «4* 4 '
3 * »W2232*4 4 *4 *3 • 2 2
3 23**84*«3f 3?«*8 33 2 • 3* '
7 34 ?44» 5» 472*3 »32*» 3 •
-2*2* •T7»233 *72 4 32 3» 2
2423 3*48«*7**4* 3* •» 3 • 4 3
2233*4334*23 *« 3 • • • 2
•2* * 34*t* • 22 • •
4332 24 4 • 22 2 • •»
-• 22*4 • 2 •• 4 •
2 32 4
2* I 3
» 3*** •
• 2 2
r .' i i i i i
! 1 1 1 1 1 I
»
*
*
2
2 * 2
« * «
3 * •
•
• • *
• • • > »
• 3 • *
2 •
2 • •
• «
2
• •
•
1 1 1 1 1 1 1
II II 1 1 _
w
—
—
-
*
• -
•
*
* _
—
II II 1 1 -
1.00 17.00 33.00 49.00 65.00 81.00 97.00 113.00 129.00 145.00 161.00
ELEVATION OF MONITOR ABOVE GRADE, ft
Figure 5-4. In this scattergram of median TSP concentration against elevation of monitor above grade,
total number of sites is 1448. The number 9 indicates nine or more; hence, the numbers do not
to the total.
5-13
-------
TABLE 5-3. HEIGHT ABOVE GRADE FOR 1978 TSP MONITORS
Height, ft
Number
Percent
of total
Cumulative
number
Cumulative
percent
0-5
5-10
10-16
16-20
20-25
25-35
35-40
>40
396
549
645
474
275
160
241
142
13.7
19
22.4
16.4
9.6
5.6
8.4
4.9
396
945
1590
2064
2339
2499
2740
2882
13.7
32.7
55.1
71.5
81.1
86.7
95.1
100.0
5-14
-------
TABLE 5-4. CROSS-TABULATION OF MEDIAN CONCENTRATION
OF TSP BY HEIGHT OF MONITOR ABOVE GRADE
Median TSP
concentration,
v9/m
<27
Number of sites
Percent of row
Percent of column
Percent of total
27-44
Number of sites
Percent of row
Percent of column
Percent of total
44-56
Number of sites
Percent of row
Percent of column
Percent of total
56-70
Number of sites
Percent of row
Percent of column
Percent of total
70-97
Number of sites
Percent of row
Percent of column
Percent of total
>97
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Height above grade
<4
6
5.4
6.9
0.2
22
4.3
25.3
0.8
27
4.0
31.0
1.0
20
2.8
23.0
0.7
11
2.0
12.6
0.4
1
0.6
1.1
0.0
87
3.2
4-10
39
34.8
8.6
1.4
106
20.6
23.5
3.9
118
17.3
26.1
4.3
97
13.6
21.5
3.6
68
12.4
15.0
2.5
24
14.8
5.3
0.9
452
16.5
10-16
29
25.9
4.0
1.1
139
27.0
19.1
5.1
184
26.9
25.3
6.7
178
25.0
24.5
6.5
144
26.3
19.8
5.3
54
33.3
7.4
2.0
728
26.6
16-25
16
14.3
2.5
0.6
87
16.9
13.4
3.2
157
23.0
24.3
5.7
186
26.1
28.7
6.8
148
27.0
22.9
5.4
53
32.7
8.2
1.9
647
23.7
, ft
25-40
10
8.9
2.3
0.4
85
16.5
19.5
3.1
104
15.2
23.9
3.8
121
17.0
27.8
4.4
96
17.5
22.1
3.5
19
11.7
4.4
0.7
435
15.9
>40
12
10.7
3.1
0.4
76
14.8
19.8
2.8
93
13.6
24.3
3.4
110
15.4
28.7
4.0
81
14.8
21.1
3.0
11
6.8
2.9
0.4
383
14.0
Row
total
112
4.1
515
18.9
683
25.0
712
26.1
548
20.1
162
5.9
2732
100.0
5-15
-------
TABLE 5-5. CROSS-TABULATION OF 90TH PERCENTILE
CONCENTRATION OF TSP BY HEIGHT OF MONITOR
ABOVE GRADE
90th percent! le
TSP concentration,
pg/m
54
Number of sites
Percent of row
Percent of column
Percent of total
54-79
Number of sites
Percent of row
Percent of column
Percent of total
79-99
Number of sites
Percent of row
Percent of column
Percent of total
99-124
Number of sites
Percent of row
Percent of column
Percent of total
124-188
Number of sites
Percent of row
Percent of column
Percent of total
188
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Height above qrade,
>4
7
6.1
8.0
0.3
20
3.7
23.0
0.7
23
3.4
26.4
0.8
16
2.4
18.4
0.6
19
3.3
21.8
0.7
2
1.3
2.3
0.1
87
3.2
4-10
41
36.0
9.1
1.5
119
22.1
26.3
4.4
122
17.9
27.0
4.5
95
14.1
21.0
3.5
54
9.5
11.9
2.0
21
13.6
4.6
0.8
452
16.5
10-16
29
25.4
4.0
1.1
131
24.3
18.0
4.8
182
26.8
25.0
6.7
167
24.7
22.9
6.1
168
29.4
23.1
6.1
51
33.1
7.0
1.9
728
26.6
16-25
18
15.8
2.8
0.7
88
16.4
13.6
3.2
160
23.5
24.7
5.9
178
26.4
27.5
6.5
157
27.5
24.3
5.7
46
29.9
7.1
1.7
647
23.7
ft
25-40
7
6.1
1.6
0.3
99
18.4
22.8
3.6
103
15.1
23.7
3.8
113
16.7
26.0
4.1
95
16.6
21.8
3.5
18
11.7
4.1
0.7
435
15.9
>40
12
10.5
3.1
0.4
81
15.1
21.1
3.0
90
13.2
23.5
3.3
106
15.7
27.7
3.9
78
13.7
20.4
2.9
16
10.4
4.2
0.6
383
14.0
Row
total
114
4.2
538
19.7
680
24.9
673
24.7
571
20.9
154
5.6
2732
100.0
5-16
-------
than the others in that row. This is not the case. In fact, on a percentage
basis there are more monitors at 16 to 25 feet above the ground experiencing
higher levels.
Monitor elevation does not appear to be systematically biasing the national
TSP data. The influence appears to be unimportant in the aggregate. However,
there may be a correlation between moritor height and location with respect to
pollution sources. Some studies in 5.2.1.24 indicate a relation of monitor
elevation, distance from roadway, and TSP concentrations at specific locations.
Other siting considerations are discussed later.
The inferences drawn about air quality levels, trends, and population
exposures from the TSP data presented in this chapter are made in full knowledge
of the following limitations of TSP monitoring: (1) sampling sites are not
standardized; (2) frequency of sampling is quite varied; (3) the vast majority
of sites reporting have fewer than 60 sampling days per year; (4) the frequency
of sampling is not randomized with respect to meteorological conditions. (5) no
spatial averaging is used in analyzing or reporting data; and (6) though the
monitor is stationary, the population it is intended to represent is highly
mobile, spending a portion of its time indoors.
5.2.1.1 Composition and Range of Concentrations—The particles that make up
atmospheric aerosols, are solid or liquid, ranging in size from a few tenths of
an Angstrom to several hundred micrometers. They may be described by their size
range, physical or chemical composition, origin, or effects. The actual classi-
fication of atmospheric aerosols is determined by the measurement technique.
Suspended particulate matter is a subset of this larger class of aerosols, and
is ubiquitous in man's environment. It includes primary particulate matter
(pollen, dust, soot, fibers) emitted from natural or anthropogenic sources and
suspended in ambient air, and secondary particulate matter (sulfates, nitrates,
5-17
-------
condensed metal, and organic vapors) which is formed in the atmosphere by
gaseous reactions involving sulfur dioxide, nitrogen dioxide, metal, and organic
vapors. Aerosols can be formed by condensation of gases or vapors, by grinding
and abrasion, or residues of evaporation of cloud droplets, ocean spray or
biological fluids. These particulates, once in the atmosphere, can be trans-
formed by coagulation and condensation at the same time that they are being
transported and diluted by air movement. The mix of particulates to which a
population is exposed contains both inert substances and biologically active
substances. In addition, these aerosols may be transformed either on mixing
with indoor air, on inhalation, or on deposition. Health effects from aerosol
exposures may be induced or exacerbated by simultaneous exposures (or in some
cases, prior exposures) to other air contaminants such as sulfur dioxide,
nitrogen dioxide, and ozone. It is therefore important to know not only the
properties of the particles, but also to what gases the population would be
simultaneously exposed.
The standard method of measuring the particulate content of outdoor air has
been to weigh the total amount of particulate matter retained on a filter paper
through which ambient air has been drawn by a high-volume air pump for 24 hr.
This measurement technique defines only the mass per unit volume of particles
present in the air over the sampling period.
Figure 5-5 is drawn from the distribution of 1978 annual arithmetic means
(Table 5-6). Also displayed in Figure 5-5 is the cumulative frequency of sites
for which the 90th percentile value is equal to or less than given value. Half
3
of all the Nation's sites had annual arithmetic mean values less than 60 yg/m .
The overall mean value for the country is 64 yg/m . Annual mean values range
o
from 9 to 288 yg/m . Only 14 valid sites had annual mean concentrations equal
o
to or less than 16 yg/m . At the other end of the distribution, 25 percent of
5-18
-------
I I I I I I I I
Q Q 90TH PERCENTILE
D D MEAN
I I I I I I I
1
0.01 0.1 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 99.99
% OF SITES REPORTING ANNUAL MEAN CONCENTRATION LESS THAN
Figure 5-5. Distribution of mean and 90th percentile TSP concentrations is shown for valid 1978 sites.
5-19
-------
TABLE 5-6. DISTRIBUTION OF VALID SITES BY
1978 ANNUAL AVERAGE TSP CONCENTRATIONS3
2 3 4
9
11
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
1
1
2
4
3
3
2
, 3
3
6
4
10
2
7
6
6
9
13
11
12
18
17
23
18
20
22
24
31
32
27
41
40
30
45
43
47
44
55
50
55
64
0
0
0
0
0
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
2
1
2
2
2
2
2
2
0
0
0
0
0 -
0
1
1
1
1
1
1
2
2
2
2
2
3
3
4
4
5
6
6
7
8
9
10
11
12
13
15
16
17
19
20
22
24
26
27
30
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
67
51
65
38
55
51
82
65
61
65
52
65
57
51
45
45
49
48
53
39
36
45
33
30
47
35
31
42
30
33
26
30
24
31
22
19
24
11
19
16
13
2
2
2
1
2
2
3
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
0
32
34
36
37
39
41
44
46
48
50
52
55
56
58
60
61
63
65
67
68
69
71
72
73
75
76
77
78
79
80
81
82
83
84
85
86
87
87
88
88
89
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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
13
11
13
12
12
14
11
7
12
5
9
11
7
7
8
7
9
3
6
5
6
12
3
6
7
2
2
1
5
2
5
4
3
4
4
5
3
4
2
2
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
89
89
90
90
91
91
92
92
92
92
93
93
93
94
94
94
94
95
95
95
95
96
96
96
96
96
96
96
96
97
97
97
97
97
97
97
98
98
98
98
90
134
135
136
137
138
139
140
141
142
143
144
145
146
147
149
150
151
153
154
156
158
159
162
163
165
167
169
170
173
175
176
177
184
187
188
190
194
195
196
197
201
256
2
3
1
3
2
1
3
2
1
4
1
1
1
1
1
1
1
1
1
1
1
1
1
3
1
1
2
1
1
1
3
2
1
1
1
1
1
1
1
1
1
1
0 98
0 98
0 98
0 98
0 98
0 98
0 98
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99.
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 99
0 100
0 100
0 100
0 100
0 100
0 100
0 100
0 100
0 100
0 100
0 100
0 100
aKey to columns: 1, TSP concentration, -m|/m ; 2, number of sites; 3, adjusted
percentage of sites; 4, cumulativevof sites.
/K/UMfeu^-
5^-20
-------
sites had annual means greater than 76 pg/m and 10 percent were greater than 96
pg/m . The histogram of sites against concentrations (Figure 5-6) shows that
over a third of all monitoring sites had annual mean values between 40 and 60
pg/m . Slightly less than another third had annual averages between 60 and 80
3
pg/m .
The distribution of 90th percentile values is also plotted in Figure 5-5.
To interpret this distribution, consider that half of all valid monitors had a
90th percentile value in excess of 97 pg/m . For 10 percent of the monitors, 10
percent of the observations exceeded 160 pg/m . For one monitor, 10 percent of
3
the observations exceeded 600 pg/m .
The annual mean TSP concentrations range from approximately 10 to over 250
3
pg/m . The lower values are associated with remote monitoring sites. The
background sites in the National Air Sampling Network (NASN) like Glacier
National Park and Acadia National Park had annual averages of 11 and 21 pg/m
in 1977.
Higher annual concentrations are found in many populated and industrialized
areas. Table 5-7 is a listing of the 30 highest sites by annual averages.
Topping the list are four central-city sites in commercial, residential, or
industrial settings. The Phoenix, AZ, site (0136) had the highest annual mean
of 256 pg/m , followed by a site in Calexico, CA at 201 pg/m and an industrial
site in Granite City, IL, at 197 pg/m3.
It is apparent that these extremely high annual TSP concentrations are
associated with commercial and industrial locations. Of the top 30 sites, 15
are industrial. Many of the higher concentrations (19 of 30) were found at
central-city locations. Only four were classified as rural sites, most of which
are also residential areas.
It is also likely that arid climates and dusty conditions in the vicinity
of some monitoring sites might lead to suspension of surface material. However,
5-21
-------
I—I—\—I
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280
TSP CONCENTRATION, pg/m3
Figure 5-6. Histogram of number of sites against concentration shows that over one-third of the.sites had
annual mean concentrations between 40 and 60
5-22
-------
TABLE 5-7. THIRTY SITES FOR HIGHEST TSP ANNUAL MEANS IN 1978
Location
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Phoenix, AZ
Calexico, CA
Granite City, IL
Libby, MT
Granite City, IL
Valencia Co., NM
Middletown, OH
Routt Co. , CO
Denver, CO
Brawl ey, CA
Kansas
Pittsburgh, PA
Caribou Co. , ID
Dearborn, MI
Houston, TX
Lancaster Co. , NE
Azusa, CA
El Paso, TX
Granite City, IL
Cleveland, OH
Braddock, PA
Riverside, CA
Site
code
013G
0011
009F
010F
014F
001F
002H
003F
002 F
0011
015F
017G
015J
002G
035H
002G
0021
022G
010F
013H
001G
003F
TSP concen- -
tration, ug/m
256
201
197
196
195
194
190
188
187
184
177
177
176
176
176
175
173
170
169
169
167
165
Site type
Center city
Commerical
Center city
Residential
Center city
Industrial
Center city
Residential
Rural
Industrial
Rural
Industrial
Suburban
Industrial
Rural
Commercial
Center city
Conmercial
Center city
Commercial
Center city
Commercial
Center city
Industrial
Suburban
Industrial
Suburban
Industrial
Center city
Commercial
Suburban
Industrial
Suburban
Industrial
Center city
Industrial
Center city
Residential
Center city
Industrial
Center city
Residential
5-23
-------
TABLE 5-7 (continued)
23.
24.
25.
26.
27.
28.
29.
30.
Location
Bakersfield, CA
Lyon Co., NV
Cleveland, OH
Des Moines, IA
Stuttgart, AR
San Bern' to, TX
East Chicago, IL
Detroit, MI
Site
code
003 F
002F
008H
046G
005F
001F
004H
023G
TSP concen- 3
tration, yg/m
163
163
163
162
159
158
156
154
Site type
Center city
Commercial
Rural
Industrial
Center city
Industrial
Center city
Industrial
Center city
Residential
Center city
Commercial
Center city
Industrial
Suburban
Industrial
5-24
-------
it is impossible to ascertain the contribution of fugitive or resuspended dust
to the concentrations measured at these 30 sites without more detailed analysis.
Daily, or 24-hr, TSP concentrations have a wide range. In remote areas
such as the Pacific islands, daily values may be as low as a few micrograms per
cubic meter. Over the continental United States, concentrations from 5 to 20
o
pg/m are routinely reported. In other locations, daily TSP values can exceed
10 times the levels found in remote areas, on occasion exceeding 3000 yg/m .
Values exceeding 1000 yg/m are observed in remote arid regions as well as in
populated urban areas. Daily TSP levels approaching these higher values, 500
3
to 1500 yg/m , are frequently associated with adverse meteorological conditions:
low-level inversion, stagnation, or high winds resuspending surface material.
Table 5-8 lists the valid TSP monitoring sites with the highest 30 24-hr
values. Only some of these sites are listed in the top 30 annual average list
(Table 5-7). In cities like Topeka, KS, and Libby, MT, which are not densely
populated or industrialized, these high concentrations may result from chance
occurrences, such as fire or dust storm. In other cities like El Paso, TX, and
Granite City, IL, which are industrialized, the maximum concentrations are more
likely to be related to persistent sources of pollution.
5.2.1.2 TSP Concentrations by Location—Ambient TSP can be simply considered as
arising from the particulate emission of three broad source categories: tradi-
tional sources (large and clearly recognized sources), nontraditional sources such
as resuspension of surface dust, and natural sources. The relative contributions
of these three sources are affected by meteorological conditions and location.
Hence it is not surprising to see a wide variation in daily measurements at a
single site, variation over an urban area, or variation among regions of the
country. This section will, by illustration, examine these differences in TSP
concentrations by location.
5-25
-------
TABLE 5-8. THIRTY MONITORING SITES WITH THE HIGHEST 24-HR
CONCENTRATIONS IN 1978
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Location
Soda Springs, ID
Lawrence Co. , OH
Shoshone Co. , ID
Libby, MN
Harris Co., TX
Ohio
Phoenix, AZ
San Bern' to, TX
Tulsa, OK
Caribou Co. , ID
Granite City, IL
Carey, OH
Menominee, MI
Weld Co., CO
Storey, Co., NV
Kennewick, WA
Charleston Co. , SC
Pasco, WA
Topeka, KS
El Paso, TX
Clarkston, WA
Rochester, MN
Site
code
004 F
003H
021F
010F
018H
004J
013G
001F
112F
015J
014F
002J
001F
005F
001F
0011
038F
001F
002F
019G
001F
015G
TSP concen- -
tration, vg/m
1441
1433
1408
1384
1359
1259
1241
1172
1061
918
837
835
791
770
744
728
719
691
672
671
663
658
Site type
Suburban
Residential
Rural
Commercial
Suburban
Industrial
Center city
Residential
Suburban
Industrial
Center city
Commercial
Center city
Commercial
Center city
Commercial
Center city
Industrial
Rural
Industrial
Suburban
Residential
Center city
Commercial
Rural
Commercial
Center city
Residential
Center city
Commercial
Center city
Residential
Suburban
Residential
Center city
Commercial
Center city
Residential
5-26
-------
TABLE 5-8 (continued)
Location
Site
code
TSP concen-
tration,
Site type
23. El Paso, TX
24. .Nampa, ID
25. Cook Co., IL
26. Denver, CO
27. Cedar Rapids, IA
28. Valencia Co., NM
29. Big Spring, TX
30. Turlock, CA
022G
001F
002H
002F
024G
001F
002F
0011
647
643
630
629
624
619
614
606
Suburban
Industrial
Center city
Commercial
Suburban
Residential
Center city
Commercial
Suburban
Industrial
Rural
Industrial
Suburban
Commercial
Suburban
Residential
5-27
-------
5.2.1.2.1 Industrial-commerical-residential. A study of TSP samples collected
in 14 U.S. cities (EPA, 1976) has concluded that traditional sources, principally
3 3
industrial, account for 15 yg/m in residential areas and over 60 yg/m at sites
nearer industrial areas. Auto exhaust nontraditional emissions of road dust,
fugitive dust, unducted emissions and a variety of widely dispersed sources can
"3
contribute 25 to 35 yg/m to the citywide TSP levels. There is a background
windborne particle level, both natural and manmade. Its contribution to annual
TSP levels is estimated at 5 to 10 yg/m and at times may be greater than 40
3
yg/m . This type of particle is found in higher concentrations in the East.
The general differences in annual TSP concentration among locations is seen
in Figure 5-7. These differences reflect the character of the neighborhoods
where the monitors are located. This figure summarizes the mean concentrations
from 154 sites in 14 cities. Residential neighborhoods in and near cities have
TSP levels between 50 and 70 yg/m . Commercial sites have a wider range of
o
concentrations (60 to 110 yg/m ). Industrial locations generally range between
80 to 150 yg/m3.
The 1978 data base has been analyzed on the basis of a different set of
descriptors. One description scheme classifies monitoring sites by their
purpose: population exposures, source receptors, or background sites. Another
scheme identifies sites by the amount of development: central city, suburban,
rural, or remote. These classifications are not mutually exclusive.
When sites are distributed by descriptors, a distinct weighting becomes
apparent. Almost 80 percent of the sites are population oriented; approximately
15 percent are source related, and less than 6 percent are background monitors.
The distribution by development also reflects its population emphasis. Of the
total monitors, 83 percent are at either central-city or suburban sites, 15
5-28
-------
IDU
n
.E
3.
z 100
o
H
oc
z
Z 50
8
t
ft
^™
^
*M
-_
RESIDENTIAL
COMMERCIAL
INDUSTRIAL
Figure 5-7. Histogram of mean TSP levels by neighborhood shows
lowest levels in residential areas, higher levels in commercial areas,
and highest levels in industrial areas.
Source: U.S. Environmental Protection Agency (1976).
5-29
-------
percent are at rural sites, and 2 percent are at remote sites. More detail on
how site location affects observed concentration is given in Tables 5-9 through
5-12, where sites are grouped by median values and the 90th percentile values in
a cross-tabulation by site descriptors. It is informative to compare the common
percentages in each grid and the row total. For instance, 38 percent of the
background sites had median values less than or equal to 27 pg/m , whereas only
4.4 percent of all sites had these low values. Only 30 percent of the back-
ground sites had median values above 44 yg/m , and none had values above 97
3 3
yg/m , whereas 75.5 percent of source sites had median values above 44 yg/m .
The pattern is consistent for the distribution of the 90th percentiles cross-
tabulated by site purpose. Cross-tabulations of site median values and site
90th percentile values with the development-related site descriptors is further
confirmation of the influences of siting on measured TSP concentrations. Rural
and remote sites have lower median values and lower 90th percentile values. The
suburban sites reflect the overall national distribution. The central-city
sites have proportionately more sites in the higher concentration ranges.
5.2.1.2.2 Intracity comparisons. Because of the strong neighborhood influence
on TSP concentrations, it is not unusual to find considerable variation in
peak and mean concentrations across a community. It is instructive to examine
intracity differences because it illustrates the difficulty in estimating
population exposures to TSP.
Data on the nine cities having the highest annual TSP concentrations in
1977 are given in Table 5-13. Only sites having enough observations per quarter
to report an annual mean are used. Although TSP concentrations in these cities
were generally high, the less developed or less industrialized areas in each
city had 1977 concentrations below the primary NAAQS, with the exception of
Granite City, IL. The annual mean concentration for the highest sites can be
two to four times higher than that for the cleanest sites within the same city.
5-30
-------
TABLE 5-9. CROSS-TABULATION OF TSP SITES BY MEDIAN
CONCENTRATION OF SITE DESCRIPTOR
Median TSP
concentration,
yg/m
<27
Number of sites
Percent of row
Percent of column
Percent of total
27-44
Number of sites
Percent of row
Percent of column
Percent of total
44-56
Number of sites
Percent of row
Percent of column
Percent of total
56-70
Number of sites
Percent of row
Percent of column
Percent of total
70-97
Number of sites
Percent of row
Percent of column
Percent of total
>97
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Purpose of site
Population
50
40.0
2.2
1.8
410
74.5
18.1
14.4
567
80.9
25.1
20.0
635
85.2
28.1
22.3
469
83.9
20.8
16.5
129
79.6
5.7
4.5
2260
79.5
Source
13
10.4
3.1
0.5
91
16.5
21.7
3.2
100
14.3
23.9
3.5
95
12.8
22.7
3.3
87
15.6
20.8
3.1
33
20.4
7.9
1.2
419
14.7
Background
62
49.6
38.0
2.2
49
8.9
30.1
1.7
34
4.9
20.9
1.2
15
2.0
9.2
0.5
3
0.5
1.8
0.1
0
0.0
0.0
0.0
163
5.7
Row
total
125
4.4
550
19.4
701
34.7
745
26.2
559
19.7
162
5.7
2842
100.0
5-31
-------
TABLE 5-10. CROSS-TABULATION OF TSP SITES BY 90TH PERCENTILE
VALUES AND SITE DESCRIPTOR
90th percentile
TSP
concentration,
yg/m
<54
Number of sites
Percent of row
Percent of column
Percent of total
54-79
Number of sites
Percent of row
Percent of column
Percent of total
79-99
Number of sites
Percent of row
Percent of column
Percent of total
99-124
Number of sites
Percent of row
Percent of column
Percent of total
124-188
Number of sites
Percent of row
Percent of column
Percent of total
>188
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Purpose of site
Population
55
44.7
2.4
1.9
441
76.8
19.5
15.5
576
81.6
25.5
20.3
595
84.3
26.3
20.9
484
83.2
21.4
17.0
109
72.2
4.8
3.8
2260
79.5
Source
14
11.4
3.3
0.5
88
15.3
21.0
3.1
94
13.3
22.4
3.3
95
13.5
22.7
3.3
87
14.9
20.8
3.1
41
27.2
9.8
1.4
419
14.7
Background
54
43.9
33.1
1.9
45
7.8
27.6
1.6
36
5.1
22.1
1.3
16
2.3
9.8
0.6
11
1.9
6.7
0.4
1
0.7
0.6
0.0
163
5.7
Row
total
123
4.3
574
20.2
706
24.8
706
24.8
582
1
20.5
4.
151
5.3
2842
100.0
5-32
-------
TABLE 5-11. CROSS-TABULATION OF TSP MEDIAN
VALUE BY SITE DESCRIPTOR
Median TSP<
concentration,
vg/m
<27
Number of sites
Percent of row
Percent of column
Percent of total
27-44
Number of sites
Percent of row
Percent of column
Percent of total
44-56
Number of sites
Percent of row
Percent of column
Percent of total
56-70
Number of sites
Percent of row
Percent of column
Percent of total
70-97
Number of sites
Percent of row
Percent of column
Percent of total
>97
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Central city
8
6.5
0.7
0.3
149
28.0
12.7
5.4
252
37.1
21.5
9.1
346
48.6
29.5
12.5
315
57.2
26.9
11.4
103
62.4
8.8
3.7
1173
42.5
Location of monitor
Suburban
26
21.1
2.3
0.9
204
38.3
18.2
7.4
320
47.1
28.6
11.6
310
43.5
27.7
11.2
210
38.1
18.8
7.6
49
29.7
4.4
1.8
1119
40.5
Rural
58
47.2
14.0
2.1
159
29.9
38.3
5.8
106
15.6
25.5
3.8
55
7.7
13.3
2.0
24
4.4
5.8
.9
13
7.9
3.1
0.5
415
15.0
Remote
31
25.2
55.4
1.1
20
3.8
35.7
0.7
2
0.3
3.6
0.1
1
0.1
1.8
0.0
2
0.4
3.6
0.1
0
0.0
0.0
0.0
56
2.0
Row
total
123
4.5
532
19.3
680
24.6
712
25.8
551
19.9
165
6.0
2763
100.0
5-33
-------
TABLE 5-12. CROSS-TABULATION OF TSP 90TH PERCENTILE
VALUES BY SITE DESCRIPTOR
90th percent! le
TSP concentration,
yg/m
<54
Number of sites
Percent of row
Percent of column
Percent of total
54-79
Number of sites
Percent of row
Percent of column
Percent of total
79-99
Number of sites
Percent of row
Percent of column
Percent of total
99-124
Number of sites
Percent of row
Percent of column
Percent of total
124-188
Number of sites
Percent of row
Percent of column
Percent of total
>188
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Location of monitor
Central city
8
6.6
0.7
0.3
168
30.2
14.3
6.1
253
36.9
21.6
9.2
328
48.4
28.0
11.9
328
58.1
28.0
11.9
88
56.1
7.5
3.2
1173
42.5
Suburban
31
25.4
2.8
1.1
229
41.2
20.5
8.3
315
46.0
28.2
11.4
296
43.7
26.5
10.7
197
34.9
17.6
7.1
51
32.5
4.6
1.8
1119
40.5
Rural
56
45.4
13.5
2.0
141
25.4
34.0
5.1
109
15.9
26.3
3.9
53
7.8
12.8
.1.9
38
6.7
9.2
1.4
18
11.5
4.3
0.7
415
15.0
Remote
27
22.1
48.2
1.0
18
3.2
32.1
0.7
8
1.2
14.3
0.3
1
0.1
1.8
0.0
2
0.4
3.6
0.1
0.0
0.0
0.0
0.0
56
2.0
Row
total
122
4.4
556
20.1
685
24.8
678
24.5
565
20.4
157
5.7
2763
100.0
5-34
-------
TABLE 5-13. RANGE OF ANNUAL GEOMETRIC MEAN CONCENTRATIONS IN
AREAS WITH HIGH TSP CONCENTRATIONS IN 1977
Number of sites Annual
Number with annual range range,
City of sites >75 yg/ml vg/ml
Range of
maximum 24-hr
value, vg/ml
Tucson, AZ 7
Pocatello, ID 4
Chicago, IL 25
Granite City, IL 8
Taos County, NM 1
Middletown, OH 3
Cleveland, OH 23
Youngstown, OH 5
El Paso, TX 14
3
3
12
8
2
13
4
10
156-67
218-65
170-50
185-85
168
192-64
152-48
172-66
158-60
591-178
1371-344
1106-152
485-227
577
707-157
705-128
602-163
691-205
5-35
-------
5.2.1.2.3 Regional differences in background concentrations. It has been
demonstrated that TSP concentrations can vary across an urban area and among
cities with different sources and meteorology. Superimposed on this intercity
difference may be regional differences in the natural or transported fraction of
TSP concentrations. Figure 5-8 shows the contribution of these sources to
nonurban levels. It was assumed that the global and local contributions in the
average would be similar. The greatest difference among regions is the contri-
bution from "continental" and transported emissions. These two cate-
gories of particles contribute in such a way that nonurban sites in the West are
300
typically 15 yg/m , in the Midwest, 25 yg/m , and in the East, 35 yg/m . Except
for the Acadia National Park site (15 yg/m ) and Millinocket (23 yg/m3), all
sites in Maine had 1977 annual averages above 30 yg/m . Nonurban sites in
o
Wisconsin typically had TSP levels less than 25 yg/m . Nonurban sites in
Montana had levels less than 20 yg/m in 1977. The individual values were: Big
33 3
Horn County, 17 yg/m ; Custer County, 15 yg/m ; Powder River County, 14 yg/m .
5.2.1.2.4 Special local problems with TSP. Neighborhood and regional influences
on TSP measurements have already been highlighted. However, the influence of
roadside sampling deserves special mention. Several studies have demonstrated
that TSP measurements are very sensitive to vehicular traffic, road conditions,
and distance from and above a nearby roadway. Measurements at sampling locations
are dominated by open-area sources and mechanically resuspended dust from
vehicular traffic within a 1-km radius around the site. It has been demonstrated
in a few studies that the distance from the roadway and the level of traffic
dramatically influence the concentrations measured. Figures 5-9 and 5-10 are
monthly-average hourly particle data for a site in Fall River, MA. The concen-
tration of the fine fraction (< 2.5 ym) varies only slightly over the day, whereas
the coarse fraction, and hence TSP concentrations, vary with the traffic volume
5-36
-------
50
40
5 30
<
cc
20
10
LOCAL
TRANSPORTED PRIMARY
TRANSPORTED SECONDARY
| | CONTINENTAL
K'.l GLOBAL
WEST
MIDWEST
EAST
Figure 5-8. Average estimated contributions to nonurban levels in the
East, Midwest, and West are most variable for transported secondary
and continental sources.
Source: GCA.
5-37
-------
120
I ' ' ' I '
I I I I I I
I , , , I , , , I
00
04
08 12 16
TIME OF DAY, start hour
24
Figure 5-9. Average diurnal variation of coarse particles and Plymouth Avenue traffic in Fall River during
April shows large contributions from suspended particles.
Source: Record et al. (1978).
5-38
-------
1800
1600 —
1400
V)
LU
_l
U
I 1200
111
u.
O 1000
ce
LU
I
Z
800
600
400
200
TTI I I I I I I I I I I I I I I | | | I I I
COARSE
PARTICLES
Pf-rT I I I I I I I I I I I I I I I I I i I
90
80
70
60
50
40
30
20
10
00
04
08 12 16
TIME OF DAY. start hour
20
24
Z
cc
t-
ui
u
o
u
IU
_J
u
OC
Figure 5-10. Average diurnal variation of ARM concentrations at Fall River trailer during April is less
for coarse particles than near road. The respirable particles were those recorded with a GCA ambient
particle monitor.
Source: Record etal. (1978).
5-39
-------
on a nearby road. Thus the population exposure to fine particles is poorly
approximated by TSP measurements when the latter are dominated by local sources.
Variations in TSP concentrations have been considered in some detail. When
reviewing the national status for TSP standards, it will be difficult to ascertain
the severity, extent, or primary cause of higher concentrations without detailed
site specification. In some instances, peak concentrations are the result of
emissions from local industries; in other cases, some specific source is responsible.
However, higher concentrations are also noted in areas where resuspended dust
from vehicular traffic or wind erosion is the most likely source.
There are cases where high concentrations would be expected but are not
recorded as such because the monitor location or frequency of sampling biases
the sample.
5.2.1.3 Diurnal and Seasonal Variation in TSP Concentrations
5.2.1.3.1 Diurnal. TSP concentrations vary with local emissions strength,
meteorological conditions, and the changes in the contributions from background
particles. The TSP mass loadings to the atmosphere in general increase during
the day and decrease at night (Figure 5-10). The atmosphere undergoes greater
vertical mixing during the day, and wind speeds near the surface increase as a
result. Greater vertical mixing coupled with increased source emissions causes
particles mass loadings to increase. At night, decreased mixing and the resultant
decreased surface winds permit settling of larger particles. With increased
atmospheric stability, local elevated sources are not as likely to mix to the
ground. There may be exceptions to this. Low-level sources such as automobiles,
unducted industrial emissions, and residential furnaces and fireplaces emitting
into an intense ground-level inversion at night can result in extremely high
night concentrations. This situation does occur in many valley communities.
Unfortunately, diurnal cycles are not well established because the standard
sampling procedure for TSP measurements is a 24-hr sample, midnight to midnight.
5-40
-------
5.2.1.3.2 Weekly patterns. Since human activity follows distinct weekly
cycles, it is likely that anthropogenic sources of particles will also have
weekly patterns. The most distinct weekly patterns are weekdays versus weekends.
Trijonis et al. (1979) have examined the RAMS TSP and dichotomous data base for
weekend-weekday differences in particle loadings. They concluded that there was
only a slight (-9 percent) difference between weekend TSP values and weekday
values for the average of five urban sites in St. Louis. For three suburban
sites the difference was -5 percent and for two rural sites the difference was
-12 percent. The urban difference was dominated by readings from one monitor in
a heavily industrial and commercial area.
Greater differences were found for the dichotomous samples and for their
elemental constituents. The TSP measurements are dominated by the larger particles,
which are more affected by meteorological factors. Hence the weekly cycles in
human activities would be masked more in the TSP measurements than in the
weekly TSP dichotomous samples.
5.2.1.3.3 Seasonal. Analyzing temporal patterns can frequently provide insight
into the nature and source of particulate matter. Meteorological parameters
affect the generation and dispersion of particles. These parameters include,
among others, degree-days, mixing height, ventilation factors, frequency of
calms and stagnations, and precipitation. These are also seasonal patterns to
sources. Many industries are seasonal in nature. Vacation periods in urban
areas are accompanied by reduced automobile traffic.
Because meteorological parameters are so important, it is likely that
seasonal patterns in one area cannot be generalized to other areas. Trijonis
(1979) found a modest annual pattern of higher TSP concentrations in the summer
months in St. Louis. Figure 5-11 supports this observation. It is a comparison
between the TSP monthly mean values and the data from dichotomous sampling.
5-41
-------
TSP
—•URBAN
SUBURBAN
-••RURAL
•URBAN
SUBURBAN
.-.-• RURAL
URBAN
SUBURBAN
RURAL
JL
I
I
I
J_
I
URBAN
SUBURBAN
CRURAL
J
JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC
MONTH
Figure 5-11. Seasonal variations in urban, suburban, and rural areas are shown for four size ranges of
particles.
5-42
-------
Particle Type Size, ym
Inhalable particles (IP) <15
Coarse particles 2.5 - 15
Fine particles <2.5
The really distinct seasonal pattern is in the fine aerosol fraction. Summer
fine-particle concentrations are twice as great as winter values. As discussed
later, sulfates aerosol makes up most of the fine-fraction particles and shows a
distinct seasonal pattern.
To illustrate the geographic specificity of these seasonal cycles, 3 years
of monthly averaged TSP data are presented in Figure 5-12. The data are from
Steubenville, OH, an industrialized site in the upper Ohio River Valley. Each
monthly mean is derived from 20 or more sampling days. The TSP concentrations
are considerably higher than the St. Louis values. The months with the highest
TSP in Steubenville were March, April, and May in 1977, July, August, September,
and November in 1978, and February and June in 1979. No clear seasonal pattern
emerges from this 3-year period.
5.2.1.4 Trends in TSP concentrations—In 1957, a National Air Sampling Network
(NASN) began to operate routinely on a national basis. The U.S. Public Health
Service, with cooperation from State health departments, operated 231 urban and
37 nonurban stations. Some of these stations operated every other year, so in a
given year there were 143 urban and 37 nonurban TSP high-volume monitoring sites
in operation. These sites collected one 24-hr sample every other week for a
total of 26 samples per year. In 1977, over 4000 stations reported TSP values
to the National Aerometric Data Bank of the U.S. EPA. Not only has the number
of sites greatly increased, but the sampling frequency has been 1 day in 6 since
1971. For some cities there are now data for more than 20 years of TSP
Monitoring. Although the sites may not be in exactly the same locations for
5-43
-------
Z
g
cc
200
150
u
110°
a so
S
_i
Z
I °
I I I | I I I I I I I I M TT I I I I I I I I J I I I I II I I | I
I I I I I I I I I I I I I I 1 I I I I I I I I I 1 I I I I I I I I I I I
JFMAMJ JASONDJ F M A M J J ASONDJ FMAMJ JASOND
1977 1978 1979
YEAR AND MONTH
Figure 5-12. Monthly mean TSP concentrations are shown for the Northern Ohio Valley Air Monitoring
Headquarters, Steubenville, OH. No clear seasonal pattern is apparent.
5-44
-------
every city, general trends in TSP concentrations can be obtained. Figure 5-13
plots the annual geometric mean TSP concentrations for three groups of cities.
In 1958, these five cities, classified as industrial, had annual mean TSP con-
centrations between 140 and 170 pg/m . By 1974 the annual mean concentrations
had dropped to between 80 and 110 yg/m . Similarly, three of four cities
classified as moderately industrialized showed substantial decreases. Only
Denver increased. The four cities classified as lightly industrialized showed
less overall change.
Examining the expanded data set for TSP from high-volume samplers shows
that for 2707 sites, the composite median concentration has remained about 60
yg/m between 1972 and 1977. The geometric mean over this period has decreased
by approximately 8 percent. The decrease in the 90th percentile of the annual
average concentrations is most pronounced over this period (see Figure 5-14).
Lowering the TSP concentrations in locations with very high levels has been a
target of State air pollution control strategies. In addition, displacing
sources to rural regions, building new sources with taller stacks, converting to
cleaner fuels, and restricting open burning have decreased the number of loca-
tions experiencing annual concentrations of over 100 yg/m .
For the period 1970-77 EPA reported an almost 50 percent reduction in TSP
emissions. Most of this reduction occurred in the early 1970's as State air
pollution control programs started many major emitters on compliance schedules.
The rather modest composite overall reduction of 8 percent in annual TSP levels
is difficult to interpret. The later part of this record follows an earlier
period of substantial reduction in emissions. In many areas during this period,
emphasis shifted from compliance to air quality maintenance. One partial
explanation that must be considered is the fact that direct emissions from
5-45
-------
230
200
ISO
160
140
120
100
90
60
40
20
0
I I I I I 1 I I I I I I I I I I
I
O BALTIMORE
NA BIRMINGHAM
Q CINCINNATI
A CLIVf LAND
• PHILADELPHIA
• ST. LOUIS
HEAVILY INDUSTRIALIZED CITIES
I I I I I I I |
I I
I I I 1
J I
1957
1960
1966
YEAR
1970
1974
I I I I I I ! I I I I I I I I I I I I
I I I I I I
O CHATTANOOGA
A DENVER
O PROVIDENCE
• SEATTLE
1967
1960
1965
1970
1974
YEAR
160
140
120
10°
80
60
40
20
0
T
T I
O MIAMI
Q OKLAHOMA CITY
A SAN FRANCISCO
• WASHINGTON. O.C.
_ LIGHTLY INDUSTRIALIZED CITIES
J L
19S7
.1960
1966
YEAR
1970
1974
Figure 5-13. Annual geometric mean TSP trends are shown for NASN sites.
5-46
-------
stationary sources contribute only a fraction to the TSP loadings of the atmos-
phere. As mentioned before, transported particles and particles from nontradi-
tional sources also contribute to the mix of TSP.
Another perspective on regional differences is gained from observations of
the 1978 data. Table 5-14 provides a summary statistic on the 50th and the 90th
percentiles for valid monitors. Region IX ranks first for the mean 50th and
90th percentiles, followed by Region VII, Region VI and Region V. Regions I and
II had consistently lower values.
The column presenting the standard deviations of the mean values for the
50th and 90th percentiles is also of interest. Smaller standard deviations
suggest that there is more uniformity in reported concentrations among monitoring
sites. Regions I, II, and IV have less variance among sites than other regions.
This could be interpreted as either more uniform distribution of pollution
levels or more uniformity and consistency in placing monitoring sites. The
larger standard deviations in other regions, particularly in the West, probably
mean that there is greater variation in pollution levels.
There are distinct regional differences in the trends of TSP concentrations.
The distribution of site means and the actual rate of change in TSP levels
differ among regions of the country. These trends are shown in Figures 5-15 and
5-16; it should be realized that the differences between years and even over the
entire period have not been tested for significance. Therefore, intraregion and
interregion comparisons are presented qualitatively.
In the Eastern United States, in EPA Regions I and II, the composite
3 3
average across sites has decreased from 60 yg/m to approximately 55 yg/m .
The distribution of concentrations is much narrower in Regions I and II than it
^ in the more industrialized Regions III, IV, and V.
5-47
-------
160
140
i
i
\ 120
| 100
I *
S 60
8 4°
& 20
0
2 I*
1972
1973
.1974 1975
YEAR
1976
1977
1
O
A
^—
I
90THPERCENTILE
75TH PERCENTILE
COMPOSITE AVERAGE
MEDIAN
25TH PERCENTILE
10TH PERCENTILE
Figure 5-14. (Top) Nationwide trends in annual mean total suspended
paniculate concentrations from 1972 to 1977 are shown for 2707
sampling sites. (Bottom) Conventions for box plots.
£-48
-------
1M
140
120
n 100
i "
a «
40
20
0
REGION 1 -
n I r
REGION ; -
i i i i i i_
REGION! -
1972 197JI 1974 I97S
YEAR
1977 197J 1»74 1875 I97g l>77
YEAR
1972 1973 1974 197S 197« 1977
YEAR
1*0
140
.120
100
K
60
•0
20
0
1*0
140
120
100
•0
W
40
20
0
i i I i r
I I I I I
\ \ 1 I I
REGION 6
1M
140
120
100
go
•o
40
20
1974> »
YIAR
1973 1974 197S 1971 1977
YEAR
Figure 5-15. U.S. Environmental Protection Agency air quality control regions,
and regional trends of annual mean TSP concentrations, are shown for Eastern States,
197277.
5-49
-------
1«0
140
120
"i 100
* 80
ft 60
10
20
0
REGION ( -
1977 1973 1974 197» 1971 1977
YEAR
1972 1973 1974 1971 1979 1977
YEAR
1»7J 1973 1974 1971 1971 1977
YIAR
ISO1
140
120
REGION 9
I I I I I
REGION 10
140
120
100
go
w
40
20
1971 1973 1974 1971 1979 1977
YEAR
1971 1974 1971 H7i 1977
YEAR
Figure 5-16. U.S. Environmental Protection Agency air quality control regions,
and regional trends of annual mean TSP concentrations, are shown for Western States,
197277.
5-50
-------
TABLE 5-14. REGIONAL SUMMARIES OF TSP VALUES FROM VALID MONITORS
en
r
en
Number
of sites Region
128
315
300
534
781
294
136
152
89
113
15
9
10
2882
I
II
III
IV
V
VI
VII
VIII
IX
X
Alaska
Hawaii
Puerto Rico
Total0
Median
Minimum
14.0
10.0
27.0
22.0
13.0
12.0
34.0
7.0
16.0
11.0
11.0
25.0
32.0
7.0
Mean
49.2
43.2
59.8
55.3
64.1
65.0
69.7
54.8
76.5
60.3
48.1
39.7
54.3
58.9
Maximum
100.0
114.0
171.0
137.0
189.0
166.0
154.0
164.0
226.0
129.0
94.0
70.0
85.0
226.0
SDa
14.6
15.3
20.3
17.2
22.0
20.9
20.2
32.8
38.0
24.5
22.7
13.7
13.9
22.8
Minimum
32.0
29.0
52.0
41.0
26.0
37.0
58.0
18.0
37.0
23.0
35.0
40.0
66.0
18.0
90th j>ercent11e
Mean
87.3
85.0
105.7
93.5
122.4
110.4
123.6
107.8
133.4
123.6
137.1
63.6
90.7
107.9
Maximum
181.0
286.0
296.0
256.0
383.0
436.0
359.0
412.0
381.1
361.1
250.0
99.0
134.0
436.0
SDU
27.7
30.6
42.1
30.9
42.8
45.8
44.2
64.0
66.0
52.5
68.7
18.9
18.9
44.9
. SD, Standard deviation of the mean.
Including American Samoa and Guam.
-------
3
In Region III the composite average decreased from 78 to 60 yg/m , with the
90th percentile in the distribution of annual mean concentrations decreasing
from slightly over 100 to about 95 yg/m . In Region IV the composite average
decreased only slightly, from 65 to 60 yg/m , but has remained relatively stable
or has even increased slightly since 1975. In Region V the composite average
decreased from 80 to 70 yg/m , and the 90th percentile has decreased from 100 to
85 ug/m , reflecting the effectiveness of point source control.
The Western States make up Regions VI through X. In Region VI the compo-
3
site average has remained at approximately 75 yg/m and the 90th percentile has
increased slightly since 1973, to about 100 yg/m . Industrial, utility, and
related growth in this area as well as in Region IV is probably responsible for
keeping TSP concentrations from decreasing. In region VII the composite average
has been almost constant, varying only slightly between 80 and 75 yg/m . The
90th percentile has varied between 110 and 100 ymg/ . Region VIII shows wide
distribution in the concentrations. The 10th percentile, at about 20 yg/m , is
the lowest among all regions. The 90th percentile, approximately 100 yg/m , is
equal to about the highest concentrations in any region. The composite average
has varied over the 6-year record, but is essentially the same, approximately 80
yg/m , in 1977 as it was in 1972. The background air quality in the upper
States of this region (Montana, North and South Dakota, Wyoming) is among the
o
highest in the country. Thus, some of the low levels (20 yg/m and below)
represent some of the lowest background concentrations measured in the United
States. The high composite and high 90th percentile levels reflect the impact
of locating monitors near industrial sources such as smelters and the fugitive
dust emissions from windblown soils. Region IX has a composite average of 100
3 3
yg/m , which is up from 90 yg/m in the early 1970's. The 90th percentile is
also high, at 120 yg/m . Thus, Region IX has some of the highest levels in the
5-52
-------
3
country. Region X has a composite average of approximately 70 yg/m , which is
3
up slightly from a low of 60 yg/m in 1975. The 90th percentile varies between
3
90 and 100 yg/m .
The overall trend in improvement from 1972 through 1975 was followed by a
reversal in some regions in 1976. Despite this short-term reversal in 1976, 60
percent of the sites showed long-term improvement from 1972 to 1977. For those
sites at which TSP concentrations violated annual standards, 77 percent showed
long-term improvements. Approximately 25 percent of these sites reported their
lowest annual values in 1977- Possibly, the short-term reversal in 1976 was
due to unusually dry weather, resulting in windblown dust that may have contributed
to elevated TSP levels throughout the Central Plains, Far West, Southwest, and
Southeast.
5.2.1.5 National Status of NAAQS for TSP--The present primary NAAQS for TSP are
3 3
75 yg/m annual geometric mean concentration and 260 yg/m for the second-
3
highest 24-hr concentration. There is a secondary standard of 60 yg/m for an
3
annual geometric mean and 150 yg/m for the second-highest 24-hr concentration.
Of the 4008 stations reporting TSP values in 1977, only 2699 had enough
samples per quarter to qualify for a valid annual average (EPA, 1978). From
these stations 40 percent, or 1070, reported a violation of the secondary TSP
annual standard (60 yg/m ); 17 percent, or 456 stations, reported violations of
the national primary annual standard (75 yg/m ). Of all stations reporting at
least minimum data, 36 percent, or 1424, showed violations of the 24-hr secondary
•3
TSP standards (second-highest value not to exceed 150 yg/m ); 8 percent, or 314
monitoring stations, reported violations of the 24-hr primary standards for TSP
(second-highest value not to exceed 260 yg/m ). Table 5-15 shows the number of
stations that recorded violations of the national primary or secondary standards
5-53
-------
TABLE 5-15. NATIONAL SUMMARY OF TOTAL STATIONS REPORTING
DATA AND NUMBER REPORTING VIOLATIONS OF AIR QUALITY
STANDARDS FOR TSP, 1977 AND 1978
Percent of sites
Data record and Number of sites exceeding NAAQS
standard exceeded 1977 1978 1977 1978
Valid annual data3 2699 2486
Annual secondary
(guide only)
Annual primary
At least minimal data
24-hr secondary
24-hr primary
1070
456
4008
1424
314
1025
439
3767
1191
215
40
17
36
8
41
17
32
6
Must contain at least five of the scheduled 24-hr samples per quarter for EPA-
recommended intermittent sampling (once every 6 days) or 75 percent of all
.possible values in a year for continuous intruments.
At least three 24-hr samples for intermittent sampling monitors.
Source: 1977 data from National Air Quality, Monitoring and Emission Trends
Report, 1977, U.S. Environmental Protection Agency, EPA 450/2-78-052,
December 1978.
1978 data from direct communication with OAQPS, U.S. Environmental
Protection Agency.
5-54
-------
for TSP in 1977 and 1978. Forty-two States plus the District of Columbia showed
violations of the primary ambient air standard for TSP in at least one monitoring
location.
Preliminary summaries of 1978 air quality TSP data indicate that 1025 of
the 2486 sites with valid data for annual averages exceeded the secondary standard
and 434 exceeded the primary standard. This represents essentially no change in
the percentage of sites not complying between 1977 and 1978.
There were only 3767 stations reporting data in 1978. Of those with at
least partial data, 32 percent, or 1191, violated the 24-hr secondary standard
and 6 percent, or 215, violated the primary standard. This indicates a slight
improvement over 1977-
Figure 5-17, based on 1977 NADB TSP data, shows the AQCR attainment status
for the annual primary NAAQS. The AQCR's classified as nonattainment were
determined by violations at at least one sampling site. Nonattainment deter-
mined on the scale of an entire AQCR could be misleading. Local sources may
dominate the levels recorded at a particular site in an otherwise less polluted
area. EPA has recognized this and has allowed States to subdivide AQCR's into
attainment and nonattainment regions. For this refined analysis, the reader is
referred to the State implementation plans revised as of January 1979 for each
State.
It appears that no region is exempt from attainment problems for TSP. The
North, Midwest, Southeast, Southwest, and Northwest all had monitors showing
violations of annual standards.
Finer spatial resolution of TSP concentrations is shown in Figure 5-18,
prepared by EPA from 1974-76 data. Again, it should be remembered that the
concentrations attributed to each county were derived from the monitor in that
5-55
-------
2MI PUERTO RICO
S3 GUAM
24SI AMERICAN SAMOA
03 VIRGIN ISLANDS
EM* AQCR WITH AT LEAST ONE MONITOR
EXCEEDING 75Mg/m3
I I AQCR WITH NO MONITORS EXCEEDING
75Mg/m3
Rv-:J INADEQUATE DATA
Figure 5-17. Map of the United States shows the AQCR's violating the primary annual TSP standard
(>75pg/m3), 1977.
5-56
-------
TOTAL SUSPENDED PARTICULATE MAXIMUM ANNUAL AVERAGE BY COUNTY,1974-197S
Figure 5-18. Maximum annual average of TSP concentrations for 1974-76 are shown by county. Note
that the annual mean primary national ambient air quality standard is 75 pg/m^, which is not to be
exceeded.
5-57
-------
county having the highest annual average value. It is not possible to determine
whether concentrations are more or less uniform across the county or whether
they are localized. However, several general impressions are obtained about
national TSP conditions. High concentrations can be found 1n almost every
State. Many populated counties have high concentrations (for example, New
Jersey-New York City, Pittsburgh, Harrisburg, Chicago, and Los Angeles).
Several sparsely populated counties also have high concentrations. Arid regions
as well as industrialized counties have high levels.
The AQCR attainment status for the daily NAAQS is shown in Figure 5-19,
which is based on the same 1977 NADB TSP data. The same comment made above
applies to the 24-hr measurements. A violation of NAAQS for TSP at one location
does not necessarily imply a higher health risk for the entire population of
that area. The health implications even for those living near a site in vio-
lation are not clear. Populations living in attainment areas but exposed to TSP
high in trace metals, for example, might have a high health risk.
A closer look at the site descriptions for stations that recorded violations
suggests that the reasons for violation are quite variable. As discussed earlier,
it seems clear that industrial sources contribute significantly to TSP levels at
many sites. This is not so obvious at other sites, however. Some extremely
high concentrations experienced at monitors in Arizona, New Mexico, and elsewhere
are most likely associated with surface dust suspended by the wind. Without a
careful site inventory or perhaps detailed analysis of TSP chemical and elemental
composition, the specific reasons for TSP violations are unknown.
5.2.1.6 Severity of Peak TSP Concentrations—The geographic displays of attain-
ment status are only one way of conveying the extent of the TSP pollution problem.
To indicate the severity of TSP ambient exposures, the 90th percentile concen-
5-58
-------
Putrto Rico I I
Amtriun Samoa I 1
Guam I }
Virgin lilandi I Zl
Figure 5-19. Map of the United States shows the AQCR's violating the primary 24-hr TSP standard
(second highest, > 260 M9/m3), 1977.
5-59
-------
tration of the 24-hr measurements was examined for all 4008 sites in the 1977
NADB. The concentrations of TSP and other air pollutants have been widely
reported to be log normally distributed (Larsen, 1971). This statistical
relationship, however, appears inappropriate at the high and low ends of the
distribution (Mage and Ott, 1978). Because the extreme values at the high end
are subject to wide scatter, the 95th or 99th percentile was found to be less
representative of the severity of high TSP levels. The 90th percentile was
therefore chosen as being a more stable indicator. It represents the TSP level
that is exceeded on approximately 36 days of the year.
Table 5-16 shows, for each AQCR, the number of TSP monitoring sites whose
90th percentile concentrations were <100, 100-200, 200-260, and >260 yg/m3. In
Figure 5-20 the AQCR's having at least one monitoring station whose 90th per-
centile exceeds 260 yg/m are displayed. AQCR's in Montana, Arizona, and New
Mexico have a large number of monitoring sites for a relatively sparse popu-
lation (approximately twice EPA's minimum requirement). A number of these sites
are near smelters. Hence, the high levels do not necessarily imply high popu-
lation exposure to TSP. In addition, windblown soil contributes to the higher
levels in these States. In the Northeast and East, the elevated TSP concen-
trations reflect the higher density of industrial and urban emissions. In these
cases, the high levels (in Pennsylvania, Ohio, New Jersey, New York, Connecticut,
and Massachusetts) indicate a larger population exposed to peak TSP concentrations.
The high 90th percentile levels in North Dakota, Nebraska, Iowa, and Colorado
perhaps reflect an influence of fugitive emissions from agriculture.
Figure 5-21 shows the number of AQCR's whose monitors have their 90th
percentile TSP concentration within the various categories. Of the country's
254 AQCR's, only 20 had air quality to the extent that none of their 90th percentiles
1
exceeded 100 yg/m . One hundred and fifty-four AQCR's had 90th percentile values in at
5-60
-------
Figure 5-20. Shaded areas indicate AQCR's in which at least one monitoring site in 1977 had a 90th
percentile TSP concentration exceeded 260 pig/m3.
5-61
-------
180
170
160
150
140
130
120
0
< 100
u.
0 go
cc
ffl 80
D
Z 70
60
50
40
30
20
10
0
20
184
35 44
<100
<200
<260
>260
90TH PERCENTILE TSP CONCENTRATIONS. M9/m3
Figure 5-21. Severity of TSP peak exposures is shown on the basis of
the 90th percentile concentration. Four AQCR's did not report.
5-62
-------
TABLE 5-16. 1977 NATIONWIDE COMPLIANCE STATUS FOR TOTAL SUSPENDED
PARTICULATE MATTER AND SEVERITY OF TSP EXPOSURES, BASED ON THE
90TH PERCENTILE CONCENTRATION OF 24-HR VALUES
•
Number of sites reporting
Number 90th percentile concentrations
State
AL
AK
AZ
AR
CA
AQCR
1
2(I)b
3
4
5(1)
6
7(1)
8
9
10
11
12(1)
13(1)
14(1)
15
16
17(1)
18(1)
19(1)
20
21
22
24
25
27
28
29
30
31
32
33
14(1)
of sites
5
8
9
20
49
4
41
11
8
1
7
13
22
60
38
16
8
16
6
11
2
15
36
3
2
7
7
20
13
1
8
<100
3
7
4
6
37
2
23
3
4
1
4
4
9
35
9
4
4
3
4
1
9
0
3
2
5
2
16
1
1
1
100-200
2
1
5
10
12
2
17
8
4
3
6
11
21
18
12
4
13
2
9
1
6
19
2
5
4
7
4
200-260
2
1
2
3
3
1
12
4
2
(ug/m ) of:
>260
2
3
1
3
1
5
1
1
5-63
-------
TABLE 5-16 (continued)
Number of sites reporting
Number 90th percentile concentrations
State
CO
CT
DE
DC
FL
GA
HI
ID
IL
AQCR
34
35
36
37
38
39
40
41 .
42(I)b
43(1)
44
45(1)
46
47(1)
5(0
48
49(1)
50
51
52
2(1)
49(1)
53(1)
54
55(1)
56
57
58(1)
59
60
61
62(1)
63
64
65
66
67(1)
68(1)
69(1)
70(1)
71
72(1)
73(1)"
74
75
of sites
4
14
23
13
9
5
5
4
35
135
2
59
3
68
13
38
61
12
48
7
8
16
22
1
10
3
10
14
34
5
4
11
2
128
8
23
49
6
27
9
4
8
<100
1
1
2
3
1
3
2
4
24
105
1
40
3
47
13
35
54
12
41
6
6
10
15
1
6
2
8
1
5
1
32
5
10
10
3
10
1
1
100-200
3
11
19
8
8
2
1
10
30
1
19
18
3
7
7
1
2
6
7
3
1
2
9
25
2
2
10
2
90
3
12
29
2
15
7
4
7
200-260
2
1
2
3
1
1
2
1
2
5
7
1
2
1
(ug/m33) of
>260
1
2
3
2
1
1
1
1
3
5-64,
-------
TABLE 5-16 (continued)
Number of sites reporting 3
Number 90th percentile concentrations (pg/m ) of:
State
IN
IA
KS
KY
LA
ME
AQCR
67(I)b
76
77(1)
78(1)
80
81
82(1)
83
84
65(1)
68(1)
69(1)
85(1)
86(1)
87(1)
88
89
90
91
92
93
94(1)
95
96
97
98
99
100
72(1)
77(1)
78(1)
79(1)
101
102
103(1)
104
105
19(1)
106(1)
107(1)
108
109
110
of sites
9
27
20
18
3
17
4
13
18
5
5
13
6
3
5
17
1
35
14
6
6
5
14
5
64
9
12
43
14
6
29
30
2
21
7
<100
7
9
4
5
2
12
2
3
5
1
3
2
2
2
1
3
7
10
4
3
2
9
1
20
1
7
11
11
4
23
22
1
14
7
100-200
2
17
14
13
1
5
2
9
13
4
2
11
3
1
4
12
1
26
4
1
3
3
5
4
42
4
5
28
3
2
6
6
1
7
200-260 >260
1
1
1 1
2
1
1 1
4
4
2
5-65
-------
TABLE 5-16 (continued)
State
MD
MA
MI
MN
MS
MO
MT
NE
AQCR
47(I)b
112
113(1)
114
115
116
117
118
119
120(1)
121(1)
82(1)
122
123
124
125
126
127
128(1)
129(1)
130
131
132
133
5(1)
18(1)
134
135
70(1)
94(1)
137
138
139
140
141
142
143
144
85(1)
86(1)
145
146
Number of sites reporting
Number 90th percentile concentrations
of sites
9
16
3
42
8
5
10
18
22
26
47
44
23
15
25
7
17
47
4
21
4
1
4
7
7
6
18
7
5
11
15
15
15
21
<100
6
9
2
23
7
2
8
14
17
22
30
10
7
9
18
7
11
35
2
11
1
1
3
2
2
11
7
2
5
13
9
8
3
100-200 200-260
3
7
1
16 4
3
1
4
5
4
17
31 3
16
6
7
6
11 1
2
10
3
1
3
4
5
4
7
3
5
2
5 1
6 1
17
(pg/m ) of
>260
1
1
1?
1
5-66
-------
TABLE 5-16 (continued)
Number of sites reporting 3
Number 90th percentile concentrations (yg/m ) of:
State
NV
NH
NJ
NM
NY
NC
ND
OH
AQCR
13(Db
147
148
107(1)
149
43(1)
45(1)
150
151(1)
12(1)
14(1)
152
153(1)
154
155
156
157
43(1)
158
159(1)
160
161
162
163
164
136(1)
165
166
167
168
169
170
171
130(1)
172
79(1)
103(1)
124(1)
173
174
175
176
of sites
8
21
2
8
33
16
28
3
7
7
8
41
20
38
48
50
13
23
24
12
10
31
5
12
8
14
36
24
107
20
18
<100
8
2
8
20
6
2
1
1
3
35
19
37
45
28
13
20
11
8
8
19
3
9
7
8
29
3
25
3
3
100-200
2
11
12
9
18
3
6
4
4
6
1
1
3
19
3
13
4
2
12
2
3
1
6
5
20
72
16
200-260 >260
6
2
1
1
5 3
1 1?
1
3
1 1
1 1
8 2
1
5-67
-------
TABLE 5-16 (continued)
Number of sites reporting
Number 90th percentile concentrations
State
OH
OK
OR
PA
PR
RI
SC
.SD
AQCR
176
177 .
178(I)b
179(1)
180
181(1)
182
183
17(1)
22(1)
184
185
186
187
188
189
190
191
192
193(1)
194
45(1)
151(1)
178(1)
195
196
197
244
120(1)
53(1)
58(1)
167(1)
198
199
200
201
202
203
204
87(1)
205
206
of sites
18
21
35
6
19
35
8
13
21
3
19
4
6
5
3
3
2
34
6
10
20
39
11
4
10
9
3
19
2
9
11
8
<100
4
6
5
2
4
1
2
4
10
2
8
3
1
1
2
17
1
1
8
5
5
3
5
8
16
2
4
8
5
100-200
14
14
22
4
14
28
6
9
10
1
10
4
3
3
2
1
16
5
6
11
19
6
1
5
1
3
3
5
3
3
200-260
5
6
1
1
2
1
2
1
5
(yg/m3) of:
>260
1
3
1
1
1
10
5-68
-------
TABLE 5-16 (continued)
Number of sites reporting 3
Number 90th percentile concentrations (yg/m ) of:
State
TN
TX
UT
VT
VA
WA
WV
AQCR
7(Db
18(1)
55(1)
207(1)
208
209
22(1)
106(1)
153(1)
210
211
212
213
214
215
216
217
218
14(1)
219
220
159(1)
221
47(1)
207(1)
222
223
224
225
226
62(1)
193(1)
227
228
229
230
103(1)
113(1)
179(1)
181(1)
232
233
234
235
236
of sites
54
39
10
4
4
11
11
18
43
61
11
5
4
23
8
34
20
9
23
32
5
8
21
8
1
1
12
6
3
<100
16
9
4
1
1
7
1
9
22
15
4
2
4
8
24
11
9
14
22
6
12
1
5
3
1
100-200
36
27
6
3
2
4
9
9
21
41
4
5
2
17
10
9
9
10
5
2
9
7
1
1
7
3
2
200-260 >260
1 1
2 1
1
1
2 3
3
2
5-69
-------
TABLE 5-16 (continued)
Number of sites reporting .
Number 90th percentile concentrations (yg/m ) of:
State
WI
WY
GU
VI
AQCR
68(I)b
73(1)
128(1)
129(1)
237
238
239
240
241
242
243
246
247
of sites
22
17
47
17
6
8
40
3
4
<100
13
12
21
10
5
8
37
1
2
100-200
9
5
25
7
1
3
2
200-260 >260
1
1 1
Source: Based on Air Quality Data, 1977 Annual Statistics. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, NC, EPA-450/2-78-040, September 1978.
5-70
-------
least one site between 100 to 200 yg/m . These data and those in Table 5-16,
indicate that most of the U.S. population might experience ambient TSP concen-
trations exceeding 100 yg/m for at least 36 days of the year.
5.2.2 Composition of Ambient Participate Matter
Atmospheric particles can be characterized in a variety of ways, including
the number of particles per volume of air, the mass of particles per volume, the
concentration of chemical species, the elemental composition of particles, or
the acidity of the aerosol. Characterizing particles in different ways may be
important in analyzing health significance and welfare effects. In addition,
the further characterization of atmospheric particles beyond the traditional
mass-per-volume measurements may provide useful information on source contributions
and thus be of use in the design of effective control strategies. The following
section will summarize information on the concentrations of TSP by various
fractional components.
Particles can be classified into five general categories: Inorganic salts
such as sulfates and nitrates; minerals, such as silica and silicates; fibrous
materials (mineral, natural organic, or synthetic organic); organic compounds;
biological material. The concentrations of TSP sulfates, nitrates, various
organic constituents, and certain metals will be described. The availability of
data on these constituents varies widely. Sulfate has been measured for some
time. Recent studies have added much to our understanding of its spatial and
temporal variation. However, data on the organic fractions of TSP are sparse.
Finally, this section presents results of studies that have identified the
source contributions to urban and rural TSP's.
5-2.2.1 Ambient sulfates—The term "atmospheric sulfates" describes a variety
°f sulfur compounds, including ammonium sulfate, ammonium bisulfate, sulfuric
5-71
-------
acid, calcium sulfate, and a variety of metal salts. Most of the historic data
on atmospheric concentrations of sulfates is based on the water-soluble extract
of TSP filters and measurements of the sulfate ion. These measurements were
subject to artifact formation on the glass fiber filters used in the early NASN
measurements. In attempts to overcome this problem, Teflon and special neutral
glass fiber filters have been introduced. They have not been trouble-free.
Hydrophobic surfaces do not wet sufficiently to ensure that all water-soluble
sulfates are removed. In general, it is now accepted that pre-1974 or -1975 TSP
sulfate measurements using the traditional glass fiber filters may have over-
estimated sulfates by as much as 2 yg/m in areas where ambient S02 concen-
trations were high.
The range of annual average TSP sulfate concentrations is from less than 1 pg/rrT
3
in some states to almost 20 yg/m in urban industrial areas of the Northeast.
For 24-hr average concentrations, sulfate concentrations have ranged from near
3
zero to more than 80 yg/m .
5.2.2.1.1 Spatial and temporal distribution. The spatial distribution of
sulfate concentration for 1974 is displayed in Figure 5-22. Figure 5-22A
presents the annual average concentrations. An area having an annual average of
more than 15 yg/m extends from the lower Ohio Valley through the upper Ohio
Valley, including major portions of Kentucky, West Virginia, Ohio, and western
o
Pennsylvania. The areas with annual averages exceeding 10 yg/m include almost
all of the United States east of the Mississippi, except for the South Atlantic
States and the upper New England area.
o
Through the central Midwest area, values of 4 to 9 yg/m are reported. The
high values seen in the Rocky Mountain States may originate from local smelters
and coal-fired power plants. The Far Western States and the Pacific Northwest
o
experience annual sulfate levels below 2 to 3 yg/m , except for the Los Angeles
5-72
-------
Figure 5-22. Contour maps of sulfate concentrations for 1974 are
shown for: (a) annual average; (b) winter average; (c) summer
average.
Source: National Research Council, Sulfur Oxides, National
Academy Washington, DC, 1977.
5-73
-------
area. The Los Angeles levels are not shown in this figure, but a 1975 National
Academy of Sciences report on air quality and stationary source emission con-
trols indicates that they are between 7 and 13 yg/m (National Academy of
Sciences, 1975).
Seasonal variations in sulfate concentrations are shown in Figures 5-22B
and 5-22C tor the winter months and the summer months. The area of elevated
sulfate greatly expands during the summer months. As demonstrated by several
regional studies on atmospheric sulfate transport, sulfate concentrations can be
elevated over large geographical regions under certain meteorological conditions
(Elisassen, 1978; Lyons and Dooley, 1978; Perhak, 1978; Whelpedale, 1978). This
is support for the transport and conversion beyond the source regions of sulfur
dioxide emissions. It is clear from these contour maps of high sulfate levels
that a large portion of the U.S. population is exposed to annual sulfate con-
centrations in the ambient air of more than 10 yg/m . In view of the increasing
sulfur dioxide emissions from increased use of coal throughout the United States,
particularly in the South Central States, the area of maximum sulfate levels
might expand and shift to the lower Ohio Valley area and the Southeast.
In a large-scale study of atmospheric sulfate in eastern Canada, Whelpedale
(1978) reports mean levels of 10 yg/m over southern Ontario. The mean levels
of sulfates dropped to less than 2.5 yg/m3 above the 49th parallel. Figure 5-23
displays these values for the period of study. During episodic conditions that
affect primarily the lower Great Lakes region, the 24-hr concentrations have
3
been reported as high as 40 to 50 yg/m . These episodic conditions are as-
sociated with the position of a high-pressure cell over eastern Canada with
southwest flow occurring on the back side of the high pressure. This synoptic
situation favors transport of sulfur dioxide and sulfates from the high sulfur
dioxide source regions of the industrialized Northeastern United States.
5-74
-------
FORTCHIMO i
ARMSTRONG •
.6 SABLE ISLAND
HALIFAX
QUEBEC CITY
MANIWAKI" 8-°
* 3.1/LAWRENCE
ATLANTIC OCEAN
5 f [ MOUNT FOREST (" • £oNTAR (Q
MICHIGAN ^^
UNITED STATES / ^JT^V> ^^"^^^ ^*"» 10
SCALE
I I I I i
0 100 200 Ml
ii in
200 KM
Figure 5-23. Intensive Sulfate Study area in Eastern Canada shows the geometric mean of the
concentration of particulate soluble sulfate during the study period. Units are micrograms of
sulfate per cubic meter.
Source: Whelpedale (1978).
5-75
-------
Recently new information on the interrelationship of sulfur dioxide,
nitrogen dioxide, ozone, ISP, sulfates, and nitrates has become available from
large-scale regional study. The Electric Power Research Institute (EPRI)
Sulfate Regional Experiment (SURE) involves intensive monitoring from some 54
rural stations and an aircraft sampling program. The area being studied is 2400
x 1840 km; it extends from Kansas to the Atlantic coast and from mid-Alabama to
southeastern Canada (see Figure 5-24) (Hidy et al., 1979).
Mueller et al (1979) reported on the earlier SURE data collected in 1974
and 1975 and presented the preliminary results of an intensive field study made
during July 1977 through February 1978. Using the limited historic data base,
they indicate that the rural stations experience a frequency of occurrence of
sulfates similar to that observed around large metropolitan areas such as New
York City. As seen in Figure 5-25, 24-hr values greater than 10 yg/m occurred
in approximately half the data, and the occurrence of 24-hr sulfate levels
exceeding 20 yg/m was about 10 to 12 percent.
Based on concentrations of 10 to 20 yg/m as an indicator of elevated
exposure, the average concentrations over the SURE data were estimated by a
linear interpolation procedure with a resolution of 80 x 80 km grids. The
occurrence and spatial extent of elevated sulfates is displayed in Figure 5-26.
Episodes of elevated sulfates are extensive; during an episode in early August
1977, the area where sulfate levels exceeded 20 yg/m expanded to more than
2
500,000 km. The month of August shows elevated sulfates over a large area as
compared with the month of October. In spite of the lower mean values during
the winter months, large regions of elevated sulfur can occur. Two regional
episodes occurred in January and early February 1977. In August, 39 percent of
the sulfate values exceeded 10 yg/m; in January the figure was 30 percent.
Five percent of the values exceeded 20 yg/m3. In October, 20 percent of the
5-76
-------
Figure 5-24. Map of SURE regions shows locations and numbers of
ground measurement stations.
Source: Hidy et al. (1979).
5-77
-------
100
50
m
1
t-
oc
8
UJ
20
10
O RIVERHEAD, NY (chimpt)
A BRONX, NY (champs)
D ROCKPORT, Uur.l)
• SCRANTON (iur« I)
n—i—r
RANGE OF OCCURRENCE
FOR SURE I AND NYC
CHAMP STATIONS
0.01
Figure 5-25. Cumulative plots show the frequency of sulfate concentrations in the SURE region on
the basis of the 1974-75 historical data.
Source: Mueller et al. (1979).
5-;
-------
300
200 -
V)
_i
LLJ
O
g
£
(9
§
X
§
u.
O
oc
UJ
O
30
300
200
100
I I I
JANUARY-FEBRUARY 1978
>20M9/m
10
T
15
I
20
1
25
DAY
I
30
I
5
10
Figure 5-26. Plot shows the number of grid cells in which sulfate
concentration equaled or exceeded 10 and 20 MQ/m^ for 24-hr
mean in the SURE region for August 1977, October 1977, and
January-February 1978. Total number of grid cells was 690.
Source: Mueller et al. (1979).
5-79
-------
values exceeded 10 pg/m , and less than 1 percent of the values exceeded 20
o o
pg/m . Figure 5-27 shows the estimated number of days exceeding 10 pg/m , for
August 1977 and January-February 1978. In August, almost the entire Northeast
had at least 10 days with sulfate concentrations greater than 10 pg/m . The
q
area having 20 or more days with more than 10 pg/m involved Ohio, West Virginia,
Maryland, Pennsylvania, and New York. By contrast, in the winter months the
area of prolonged elevated sulfate concentrations shifts toward the West and
Southeast. The upper Ohio Valley remains high, and an increase in the number of
days with more than 10 pg/m also occurs over Tennessee, Alabama, and Georgia.
Studies of seasonal variations have reported elevated concentrations in the
sunnier months (Hitchcock, 1976; Hidy, 1976). The summer monthly mean concen-
trations of sulfate in some regions can be twice those for the winter months.
These distinct variations (Figures 5-28 and 5-29) for New York City and Los
Angeles are less distinct in the Southeast and Midwest. Although biological
production of hydrogen sulfide has been offered as an explanation for high
summer values, more credence is now given to increased photochemical and chemical
reactions.
Figure 5-30 is based an daily observations of sulfur oxides and sulfates
obtained in 1974 and 1975 from stations in Indiana and Illinois eastward to
Albany, NY. The data show a summer maximum in 24-hr monthly average sulfate
concentrations over a distance of 1600 km (Hidy et a!., 1978). All of the
stations showed increased monthly mean sulfate levels during the summer months,
the highest value occurring in July and centered near Wheeling, WV. For July
1975, daily values exceeding 30 pg/m are quite frequent. Measurements of the
24-hr values for sulfates as high as 80 pg/m have been recorded in the Wheeling,
WV, area.
5-30
-------
250
Figure 5-27. Map shows the spatial distribution of number of days
per month that the sulfate concentration equaled or exceeded
10 pg/m^. Station data were extrapolated according to r"2.
(A) January-February 1978 (31 days)? (B) August 1977 (31 days).
Source: Mueller et al. (1979).
5-81
-------
12
Figure 5-28. 1977 seasonal patterns of SO2 emissions and 24-hr
average 862 and SO^ ambient levels in the New York area are
normalized to the annual average values.
Source: Lynn et al. (1975).
5-82
-------
a.
cc
ui
300
250
0
200
150
>
<
(M
100
50
I I I I I
30
Z
20 *
10
m
m
'S
m
V)
JAN MAR MAY JUL SEP
NOV
Figure 5-29. Monthly variation in monthly mean of 24-hr average
sulfate concentration at downtown Los Angeles is compared with
monthly mean 1973 Los Angeles County power plant SO2
emissions.
Source: Mirabella (1977).
5-83
-------
MONTHLY AVERAGE SULFATE CONCENTRATION, (
WEST EAST
in
r.
o»
s
o»
Figure 5-30. Regional and seasonal distributions of sulfate concentra-
tions are shown for the Northeast quadrant of the United States.
Source: Hidy et al. (1976).
5-84
-------
The seasonal differences are elucidated further in the daily time-distance
plots for July 1974 and January 1975 in Figures 5-31 and 5-32. In winter, the
high sulfate concentrations are localized, and generally the maximums are lower
than those in summer. The winter episodes appear to last only 1 or 2 days, but
the summer episodes have much greater geographical extent and seem to have
durations of 1 to 5 days. Hidy et al. (1979) point out that the intense epi-
sodes may occur every 5 to 10 days on a cycle roughly equivalent to the passage
of synoptic-scale high-pressure systems.
Lavery et al. (1979) postulate the existence of two meteorological conditions
that result in regional accumulation of particulate sulfate concentrations above
20 vjg/m in the Northeastern United States:
The first regime consists of cases where widespread stagnation occurs with
a large high pressure area slowly moving eastward over the midwestern and
eastern United States. Zones of polluted air collect over areas within
100-300 kilometers of high sulfur dioxide emissions sources. These zones
maintain themselves over periods of one to four days in warm, moist air,
with light winds, around the southern and western parts of the high pressure
area. The second regime appears to be conducive to long-range (greater
than 500 km) sulfate transport and involves a channeling of air flow
between the west side of the Appalachian Mountains and weak cold fronts
approximately oriented west-southwest to east-northeast and traveling
south-eastward. The channeling appears to be combined with capped vertical
mixing associated with subsidents around the frontal system. These epi-
sodes can last up to four days.
5.2.2.1.2 Urban variations. The preceding discussion of spatial and temporal
variations of sulfate was derived for the most part from widely spaced rural
monitoring stations. It is of interest to note spatial variations on the much
smaller scale of a metropolitan area. The sulfate measured on this scale may
consist of a natural background component, a long-distance transported com-
ponent, a component formed locally in the atmosphere, and/or an artifact formed
on the filter. Hidy et al. (1978) compare urban sulfate distributions from the
previously reported works of Lynn et al. (1975) for the New York City area, and
5-85
-------
DAILY AVERAGE SO4,
WEST
31
29
27
25
23
21
19
17
15
13
11
9
7
5
3
1
EAST
20
i no.
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
w ZZt-
ZO<£C
j^s
8
Q
Ul
UCN
o
D
X
Ul
LU
i
cc
o
Figure 5-31. Spatial and temporal distributions of daily mean sulfate concentrations over the Northeast
quadrant of the United States are shown for July 1974.
Source: Hidy et al. (1976).
5-86
-------
DAILY AVERAGE SULFATE CONCENTRATION,
WEST
EAST
>
I
31
29
27
25
23
21
17
15
13
11
9
7
5
3
1
5 s
• •
z o
a
CO Z
s
o
Z
o
a
Ul
°32
w llj rf
< o
S z "
o
Z
a
oc
o
<
00
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
Figure 5-32. Spatial and temporal distributions of daily mean sulfate concentrations over the Northeast
quadrant of the United States are shown for January 1975.
Source: Hidy et al. (1976).
5-87
-------
Kurosaka (1976) did the same of the Los Angeles area. These areas differ in
meteorology and climate, but the population and total sulfur dioxide emissions
are similar.
The population density and sulfur dioxide emission densities are greater
for the New York City area, however (see Table 5-17). As seen in Figure 5-33,
•*
there is a significant difference in sulfate concentrations across the New York
urban area, with the highest values observed in a strip from Staten Island
northeast into Brooklyn. The highest concentrations of sulfur dioxide emissions
are in eastern New Jersey, Staten Island, Brooklyn, and the high-density areas
of Manhattan. Within a distance of 10 to 50 km from the sources of highest
concentration, the sulfate concentrations have decreased by 30 to 40 percent
from their maximum values.
As shown in Figure 5-34, the mean annual average concentrations derived
from 24-hr values in Los Angeles show a relatively uniform distribution across
the Los Angeles basin area. A weak maximum is seen in the area near Burbank,
and another maximum may occur in the San Bernardino area. The areas of major
sulfur dioxide emissions are El Segundo and Long Beach areas and Fontana. A
pattern similar to New York is found in the Los Angeles area; at distances
exceeding 50 km from the areas of highest concentration, the sulfate levels drop
off significantly.
Spengler et al. (1979)(have measured respirable sulfates from a network of
10 to 12 sites in each of six cities for periods of up to 2 years. Analysis of
variance shows no significant variations among sites within the cities of
Topeka, KS; Portage, WI; Kingston, TN; and Watertown, MA. Some slight variations
occur among the sites in St. Louis, and significant variations occur among the
sites in Steubenville, OH. Only the Corondolet area of southeast St. Louis was
monitored, not the entire city. There is a coke plant and a lead pigment plant
5-f
-------
NEW JERSEY
JERSEY
CITY
LONG
ISLAND
1
STATEN
ISLAND
ATLANTIC OCEAN
I I I I I I
0.5
km
Figure 5-33. Map shows annual mean 24-hr average sulfate levels in
micrograms per cubic meter in the New York area, based on 1972
data from Lynn et al. (1975). Triangles are locations of three
CHAMP site stations. The fourth station is at the tip of Long Island
about 160 km from Manhattan.
Source: Hidyetal. (1978).
5-89
-------
VENTURE CO.
LOS ANGELES CO.
RIVERSIDE CO.
CIRCLED NUMBERS: STATION DATA
CHAMP STATIONS :
a : UNCERTAIN BECAUSE OF DISCREPANCIES
BETWEEN AGENCY ANALYTICAL METHODS
Figure 5-34. Distribution of annual average sulfate concentration in micrograms per cubic meter in the
greater Los Angeles area is based on 1972-74 data.
Source: Kurosaka (1976).
b-90
-------
TABLE 5-17. SOME CHARACTERISTICS OF POLLUTION IN THE
NEW YORK AND LOS ANGELES AREAS
Parameter
Los Angeles
New York
a 2
Surface area considered, km
21.000
Population estimate (1970)2 9.000JOOO
Population density, no. /km
S02 emissions, tons/yr «
S0« emission density, kg/km /yr
Maximum temperature, °C^
Minimum temperature, C
Relative humidity, %e
Normal precipitation, cm
Mean wind speed, m/sec
Mixing height, «m9
Ventilation^ nr/sec
S09, yg/m 9 ., .
Water-soluble sulfate (SO/~), yg/nT n
N02, yg/m3 n 4 3
Water-soluble nitrate (NO, ), yg/m
03, yg/m3 " 3
Total particulate mass concentration (ISP)
less sulfate and nitrate, TSPM, yg/m
430
238,000
10,300 H
22.8 (5.5)a
10.8 (4.6)
50.2 (17.0)
36
3.3 (1.4)
849 472)
2690 2160)
12.5 19.9)
10.1 (7.9)
83.9 (44.3)
9.1 7.7)
52 (34)
64.5 (27.4)
17,000
12,000,000
710
266,000
14,200
15.0 (7.4)
9.3 (8.4)
59.6 (16.5)
106
5.8 (2.3)
1290 906)
7460 (6200)
42.9 (45.0)
8.9 (5.7)
67.6 (36.0)
2.6 (2.1)
20 (22)
40.4 (19.9)
Greater metropolitan areas; Los Angeles, South Coast Air Basin; New York, tri
uState metropolitan area.
cBased on EPA Air Quality Control Regions.
^Annual mean of daily maximum or minimum hourly temperature.
Cumbers in parentheses are standard deviations.
^Annual mean of daily minimum humidity.
-Annual mean of noon wind speed at surface.
^Defined by annual mean of daily midday radiosonde sounding.
Annual mean of 24-hr averaged values, 1974-75; Los Angeles, seven stations,
New York, four stations (see Hidy et al., 1977b for details).
Source: Hidy et al. (1978).
5-91
-------
nearby, which causes large S02 gradients and perhaps also sulfate gradients. In
Steubenville, the TSP and SOp values near the river are approximately twice the
concentrations 5 km to the west of the river. For respirable sulfates the
pattern is similar and the gradient is not as pronounced, but the differences
among sites are significant.
An attempt has been made to explain the variability in sulfate data for
both the Los Angeles area and the New York City area by means of stepwise linear
regression. Table 5-18 displays the three principal independent variables and
the r values associated with them in explaining the variance of the daily
sulfate concentrations. The results are very consistent in both areas except
for Vista, CA, a community about 100 miles southeast of Los Angeles.
The results indicate that the most important variables are the 24-hr ozone
level, the midday relative humidity, and the total mass concentration, minus the
sulfate and nitrate fraction. Hidy et al. (1978) also suggest that these three
factors are important in determining the daily variations of sulfate concen-
trations. The ozone or oxidant levels are an indication of photochemical
oxidation, the relative humidity is an indication of water content of the air
mass, and TSP is an indication of reactions involving particulate matter.
The local sulfur dioxide concentrations did not enter into the correlation
sequence as one of the three principal variables. The findings of Spengler et
al. (1979) appear quite consistent with these findings since the only city with
a significant spatial variation among sites for sulfates also had a variation
among sites for the respirable particles and TSP.
5.2.2.2. Ambient Nitrate Aerosols--Nitrate aerosols make up a varying amount of
the total suspended particulate matter. Although widely reported to be signifi-
5-92
-------
TABLE 5-18. PRIMARY RANKING OF VARIABLES FOR CORRELATING AIRBORNE
SULFATE IN TWO CITIES BASED ON A STEPWISE LINEAR REGRESSION OF
15 VARIABLES FROM CHAMP AND RELATED MONITORING STATIONS
A. Los Angeles area
Variable
1
2
3
Correlation
coefficient (R)
B. New York
Variable
1
2
3
Correlation
coefficent (R)
Anaheim
°3
TSPM
RH
0.71
Brooklyn
TSPM
RH
0,
3
0.60
Garden
Grove
°3
TSPM
RH
0.77
Queens
TSPM
RH
°3
0.63
West
Covina
°3
TSPM
RH
0.79
Bronx
TSPM
0.
Dn
r\n
0.54
Santa Thousand
Glendora Monica Oaks
TSPM 0, TSPM
RHD RR RH
0, TSPM 0
3 x
0.79 0.79 0.72
Riverhead, L.I.
TSPM
RH
°3
0.62
Vista3
T .
Rfi
0.56
?Located 50 km north of San Diego and 16 km inland from the coast,
RH, Relative humidity.
OX, 1-hr daily maximum ozone value.
Source: Hidy et al. (1978).
5-93
-------
cantly less than the sulfate fraction, nitrates nevertheless represent an
important constituent. Most nitrates in the atmosphere are formed in gas-to-
aerosol reactions, principally involving nitrogen dioxide and nitric oxide.
These reactions may yield nitric acid (gas or aerosols), ammonium nitrate,
sodium nitrate, and lesser amounts of other compounds. A minor fraction of the
nitrate aerosols measured in the atmosphere can be attributed to wind erosion of
soil and resuspension of fertilizers (National Research Council, 1978). These
sources may be more important locally near fertilizer plants or transfer facil-
ities, as well as munition factories. A study in Chattanooga dramatizes the
influence of local sources (Helms et al., 1970; National Research Council,
1977). The average nitrate concentration from a site close to the Volunteer
3
Army Ammunition Plant was 48.9 yg/m. This is more than three times the NASN
o
maximum station average for 1965 (13.5 yg/m ). This station average was 15 to
20 times higher than that of the four other Chattanooga sites presumably not
influenced directly by the munitions plant. Their averages ranged from 2.4 to
3.8 yg/m3.
Accurate measurement of nitrates has been difficult (EPRI, 1979). Much of
the historic data is of uncertain quality because of the formation of "par-
ticulate" nitrate on glass fiber filters. Spicer (1976) found that the con-
centration of nitrate "aerosols" was 3 to 10 times higher on glass fiber high-
volume filters than on quartz fiber filters. Quartz filters are neutral and are
thought to collect only particulate nitrates, but glass fiber filters contain
basic sites and impurities that apparently extract at least some nitric acid
vapor, increasing "particulate" nitrate readings. For several years the NASN
stations operated by EPA have been analyzing the nitrate ion from high-volume
samples. Again, because these samplers used glass fiber filters for collection,
some of the historic data may overestimate the nitrate concentrations because
5-94
-------
of absorption of nitric acid. The reader is referred to Table 5-14 of the
National Research Council's study of nitrogen oxides (1977). This table presents
the annual mean and maximum concentrations of nitrate 1on for 92 NASN sites from
1958 through 1970.
The annual mean nitrate aerosol concentrations from 55 urban and 17 non-
urban NASN sites is summarized in Figure 5-35. Concentrations in urban air are
substantially higher than those in nonurban air. Although year-to-year variations
are substantial, both urban and nonurban averages show upward trends. These
trends are consistent with the increase in emissions of nitrogen dioxide.
These measurements are limited in that the analytical methods for aerosol
nitrate analysis do not differentiate among particles containing neutral ammonium
nitrate, sodium nitrate, and nitric acid. It is therefore impossible to estimate
the relative quantities of neutral and acid nitrates in the aerosol.
The spatial distributions of nitrate ions are shown in Figure 5-36 for
mostly urban sites and in Figure 5-37 for nonurban sites. The annual average
concentrations are in micrograms per cubic meter, as measured from high-volume
3
samples. A zone of high urban concentrations exceeding 4 vg/m extends eastward
from Chicago through the industrialized Northeast through Pennsylvania to the
Philadelphia area. Other zones of high nitrates are found in southern Louisiana,
around Birmingham, AL, and near Little Rock, AR. In general, a zone of high
urban nitrates 3 yg/m and larger extends up from southeastern Texas through the
Midwest and across through the Northeast.
It is obvious from this display that the data base is quite incomplete for
the west coast. No data are reported for the Los Angeles area, nor for the
large metropolitan areas of San Francisco, Seattle, and Portland. The areas of
high nitrate are similar to the areas of high nitrogen dioxide emissions (Figure
5-38). Areas of localized nitrogen dioxide emission from automobiles, electric
5-95
-------
a.
O
<
oc
UJ
o
o
(J
-I
O
O
c
UJ 1.5
UJ
OC
z
D
Z
Z
UJ
3.5 —
3.0 —
2.5
2.0
URBAN (55 sites)
1.0
0.5
NONURBAN (17titn)
1965 1966 1967
1968 1969 1970 1971
YEAR
1972 1973 1974
Figure 5-35. Mean annual nitrate aerosol concentrations in urban and nonurban air were obtained by
the National Air Surveillance Network.
Source: Data from up to 1970 are from U.S. Environmental Protection Agency (1974). Data for
1971-74 are from Akland (1977).
5-96
-------
Figure 5-36. Map shows U.S. mean annual ambient nitrate levels in micrograms per cubic meter.
Source: Akland (1977).
5-97
-------
1.7
Figure 5-37. Mean nitrate concentrations in micrograms per cubic meter were measured at nonurban
sites by the U.S. Environmental Protection Agency (unpublished data).
Source: National Research Council (1979).
5-98
-------
Figure 5-38. Map shows U.S. ambient emission densitiewin grams per square meter per year.
Source: Akland (1977).
5-99
-------
utilities, and petroleum refineries might be expected to have elevated nitrate
concentrations.
A few studies have sought information on nitrate concentrations by com-
position and particle size. Orell and Seinfeld (1977) compared the formation,
sizes, and concentrations of ambient sulfate and nitrate particles. Unlike
sulfuric acid, the nitric acid that is formed tends to remain 1n the gaseous
phase, although it may be an important component of acid precipitation. The
EPRI SURE Project (1979) reports nitrate ion concentrations one-tenth the
concentration of sulfate ions. The monthly mean values for August and October
3
1977 are less than 0.6 yg/m ammonium nitrate at three locations across the
Northeast (see Figure 5-39). On a few occasions the daily levels exceded 1.5
yg/m .
Data from the sulfate and nitrate data base of the California Aerosol
Characterization Experiment are reported by Appel et al. (1978). Summer mea-
surements for 1972 and 1973 from five fixed and one mobile site indicate that
the mass mean diameter for nitrates is between 0.3 and 1.6 ym.
Twenty-four-hour averaged concentrations of nitrate ion varied across the
Los Angeles basin, increasing to 15 to 30 yg/m in the eastern communities of
Rubidoux and Riverside. Table 5-19 shows the averaged values from high-volume
sampling for different times during 1972-73. The values do not reflect the
annual concentrations; in fact, the highest concentrations are now in doubt
because of an artifact involving the filters. The authors do report 1968 NASN
data in the South Coast Basin as corroborating evidence of high nitrate ion
concentrations and spatial variations (see Table 5-20). In contrast to sulfate,
the diurnal pattern for nitrate often has a maximum during the morning close to
the maximum for gas-phase nitrogen oxides. They concluded that the ratios of
5-100
-------
in
o
TABLE 5-19. AEROSOL CHARACTERIZATION EXPERIMENT DATA FOR NITRATE
AND SULFATE IN THE CALIFORNIA SOUTH COAST BASIN 1972-73
(24-h average values)
Sampling
location
Dominguez Hills
Harbor Freeway
Pasadena
West Covina
Pomona (1972)
Pomona (1973)
Rubidoux
Riverside
No.
episodes
2
2
6
5
5
2
3
9
NO "
j
pg/m
4.1
6.3
6.5
8.2
19.8
7.9
31.5
15.2
% of
mass
3.1
5.6
9.0
4.5
14.6
4.8
12.5
13.4
SO-
ug/m
17.8
5.7
7.0
21.8
9.6
11.5
10.7
7.7
2-
% of
mass
13.6
5.1
9.7
12.0
7.1
7.0
4.2
6.8
aAnalysis of high-volume samples collecter on Whatman 41 filters during July-October periods
Source: Appel et al. (1978).
-------
TABLE 5-20. NASN NETWORK DATA FOR NITRATE AND SULFATE IN THE CALIFORNIA
SOUTH COAST BASIN FOR 1968
(annual 24-h average values)
Sampling
location
Long Beach
Los Angeles
Ontario
Riverside
ug/m
5.7
7.7
9.1
10.2
No3
% of
mass
5.0
6.0
7.9
8.8
SO,2"
ug/m
12.7
10.2
9.0
8.3
% of
11.1
7.9
7.8
7.2
aData based on geometric mean of 26 24-h sampling periods.
L Source: U.S. Environmental Protection Agency (1972).
o
-------
100
80--
0
\ 60
ff
tu
0
§ 40
8*
*'.+
__EZ52
?,
(184)
- -2.0
4-1-6 I
CJ
4- 8
o
1.2 5
0.8
- -0.4
O
2
MONTAGUE PHILO ROCKPORT
AUGUST 1977
MONTAGUE PHILO ROCKPORT
OCTOBER 1977
Figure 5-39. Histogram shows monthly mean and extremes of 24-hr particle concentrations for August
and October 1977 at three sites in the SURE region.
Source: Hidy et al. (1979).
5-103
-------
ionic constituents and ambient ammonia levels suggested that ammonium salts were
i
the principal form of sulfate and nitrate.
Until recently such high nitrate levels were not suspected in other regions
of the United States. However, the Environmental Protection Agency has just
completed the first analysis from a dichotomous particle sampling program in
Denver, CO. The 24-hr nitrate levels, primarily in the fine fraction, often
exceeded 10 yg/m (Stevens, 1979).
Japanese workers have been investigating atmospheric nitrates for some
time. Kadawaki (1977) has found a bimodal distribution of nitrates in the Ngoya
area of Japan. The submicron particles (0.4 to 0.6 ym in diameter) are ammonium
nitrate, and the course particles (3 to 5 ym in diameter) are sodium nitrate.
Background nonurban levels on the outer islands of Japan have been reported as
o
low as 0.8 to 0.9 yg/m (Kito, 1977). Much higher concentrations were reported
in the city of Kawasaki, as high as maximum average concentrations of almost 7
yg/m3 (Terabe, 1977).
In summary, our knowledge of nitrates in the atmosphere is rather limited.
No comprehensive data set exists. The NASN measures nitrate ion every 12th day
at relatively few sites; spatial and short-term temporal variations cannot be
discerned. In fact, many cities have no measured values of nitrates reported.
Furthermore, historic data before 1977 are in doubt because of the artifact
formation on the filters. It is now believed that most of the aerosol nitrate
is in the form of ammonium nitrate. There is some evidence that ammonium
nitrate is in the fine fraction, while the artifact is predominantly in the
coarse fraction. There are spatial patterns in nitrate concentrations. Cities
tend to have higher levels of nitrates than do rural regions. Both the Chattanooga
study and the Los Angeles study indicate that localized areas may have sub-
stantially higher nitrate levels. High average levels (>10 yg/m ) of nitrates
5-104
-------
found in Denver raise the concern that available data on nitrate concentrations
may underestimate the actual population exposure. In the near future, new
sampling and analysis techniques should substantially expand our knowledge of
nitrate aerosols, nitric acid, and other nitrogen compounds.
5.2.2.3 Airborne Organic Particulate Matter—This section will be Included in
later draft.
5-105
-------
5.2.3 Ambient Sulfur Dixoide Concentrations
Ambient concentrations of sulfur dioxide are determined by the following
factors: (1) the density of emissions sources; (2) the source characteristics
such as stack height, exit velocity, and source strength (3) the local meteorological
conditions (4) the local topography and surrounding buildings (5) the reaction
rate of sulfur dioxide in the plume, and (6) the removal rates by precipitation,
deposition at the surface, and other reactions. These factors interact in
such a way that in urban and industrialized areas with high densities of S0?
emissions, the SO,, emissions, the SOp concentrations are much higher than in
surrounding rural areas. It is quite common to find gradients in S02 concen-
tration within these industrialized and areas, with a central core area report-
ing the highest SO- concentrations. This pattern is shown diagrammatically in
Figure 5-40.
Where SOp emissions are dominated by a single or a few point sources, the
pattern of SO,, concentrations could be different from the pattern displayed in
Figure 5-40. Depending on topography, meteorology, and source characteristics,
the concentration patterns may be asymmetrical and the temporal distribution
may be skewed to low mean values with a few intermittent high peaks. These
differences in concentration patterns may be important to the effects experienced
in exposed human populations.
It will be shown in this section that significant progress has been made
toward meeting the primary NAAQS for S0?. By 1978 there remained only a few
AQCRS with daily SO* levels above 365 ug/m , or annual averages above 80
o
ug/m . Most urban areas experienced dramatic improvements in air quality as a
result of: (1) restrictions on sulfur in fuel (2) better controls on existing
sources; (3) displacement of sources and the building of new sources in less
populated regions; and (4) the building of taller stacks.
5-106
-------
0
h
K
UJ
o
o
u
N
s
UJ
c
MICRO AND MIDDLE
REGIONAL
NEIGHBORHOOD
I
PEAKS ASSOCIATED
WITH MAJOR POINT SOURCES
RURAL AREAS
SUBURBS |
I
URBANCORE
CITY LIMITS
I SUBURBS
RURAL AREAS
Figure 5-40. Relative locations for sites measuring concentrations represent several spatial scales of
measurement in an urban complex, with respect to annual averaging times.
5-107
-------
This section presents SOp concentration data for specific locations and
areas where levels are currently high. The national status of S02 concentra-
tions is reviewed, along with trends data. A comparison is made between S09
levels in six cities in the early 1960's and the concentration in the late
1970's. Insights on factors that are important determinants of population
exposures are presented in the discussion of diurnal and seasonal S0? concen-
tration patterns. Since SCL can be measured by a variety of methods (see
chapter 3), a brief discussion of SOp monitoring and instruments precedes the
substantive sections on concentrations.
5.2.3.1 Sulfur Dioxide Mom'toring--The Environmental Protection Agency is now
in the process of revising Federal, State, and local air monitoring networks.
By 1981, States will be operating a selected number of sites in the National
Air Monitoring Station (NAMS) Network. These sites are to be located in the
areas of highest pollution concentrations and areas of high population density.
They are designed to serve in assessing the trends and progress in meeting
standards. By 1983, the State and local agencies are to be operating the
State and Local Air Monitoring Station (SLAMS) Network. This network is
designed to be part of each State's implementation plan. It is expected that
this will mean fewer sites than are currently in operation; however, the
effort of Federal coordination of air monitoring should provide the much-needed
quality control. The trend to reduction in the number of stations is already
apparent in the 1977 sulfur dioxide data. There were almost 120 fewer monitor-
ing sites reporting data in 1977 than in 1976 (2365 vs. 2482). For sulfur
dioxide this trend is becoming more pronounced. Many States terminated all or
most of their 24-hr West-Gaeke bubbler sampling in 1978. It is well documented
that sulfur dioxide measurements by this method are subject to degradation at
5-108
-------
high temperature. Rather than modify the West-Gaeke bubbler method and maintain
the dense network of bubblers and continuous monitors, State and local agencies,
where they can, are relying continuous monitoring equipment. (See chapter 3).
Table 5-21 give the number and type of SCL monitors operated by Federal,
State, and local agencies during 1976, 1977, and 1978. It should be noted
that in addition to those included on the tabulation there are many sulfur
dioxide monitors operated by electric utilities, paper companies, smelting
companies, and others. The data collected by these industries are not,
however, part of the National Aerometric Data Bank. Table 5-21 indicates the
wide variety of sampling methods currently used to measure S02 in the atmosphere
There are some serious, unresolved questions about the direct comparability of
these methods. By 1978 it was recognized that S02 could bias the calibrations
of the flame photometric instruments. It was also recognized that hydrocarbons
were an important interference for the pulse fluorescent instruments.
Nationally, SC^ monitoring is not as extensive as TSP monitoring. In
1978 there were 947 sites with continuous monitoring equipment and 1298 bubbler
sites. Every State conducted S0? monitoring. Table 5-22 lists the number of
S(L monitors operating in each State in 1978. All operating sites are eligible
for comparison with 3-hr and 24-hr NAAQS. However, only those sites with a
sufficient number of hourly or daily observations are considered valid for
comparison with the annual NAAQS. To be considered valid for this purpose, a
continuous monitor must have been operational for at least 6570 hours during
the year. A bubbler must have operated at least 5 days in each of the four
calendar quarters. It is with respect to the number of vaild SCL sites that
the national coverage appears inadequate. Only 99 of the 1298 bubbler sites
or 7.6 percent are considered valid in 1978. Only 385 of the 947 (or 40.7
5-109
-------
TABLE 5-21. TOTAL MONITORS REPORTING S02, BY METHOD. 1976-1977
Type
Continuous
Colorimetric (West-Gaeke)
Conductometric
Coulometn'c
Flame Photometric
Hydrogen Peroxide (Titrimetry)
Catalyst-flame photometric
Pulsed fluorescent
Second Derivative
spectrophotometry
Sequential- conduct i metric
24-Hour Bubbler
Paraarosaniline
Total
1976
82
20
339
121
11
1
17
2
10
603
1879
2482
1977
95
25
365
105
76
1
—
--
9
676
1689
2365
1978
__a
--
--
--
--
—
947
1298
2245
Data not available.
5-110
-------
TABLE 5.22. NUMBER OF BUBBLER AND CONTINUOUS SO. MONITORING SITES IN
OPERATION IN 1978 AND VALID FOR 1978 ANNUAL MEANS
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.
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of
Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyomi ng
American Samoa
Guam
Virgin Islands
TOTAL
Bubblers
Operational Valid
10
7
0
21
0
0
0
3
0
106 25
23
9
0
36
64 25
35
0
153
15
38 5
43
49
0
31 20
0
8
1
20
0
16 11
12 2
25
0
52
15
156 10
12
7
4
22
5
24 1
9
0
123
0
0
64
8
26
20
15
0
4
7
T298 99
Continuous
Operational
30
4
15
1
55
8
24
15
4
41
14
0
14
57
42
16
5
45
0
11
19
19
30
22
4
16
21
1
1
10
23
9
64
2
2
46
6
4
43
0
4
8
0
59
16
15
4
23
11
4
57
3
0
0
0
947
Valid
12
1
19
5
5
2
18
2
13
22
15
4
2
26
1
1
12
28
10
7
2
2
19
4
32
21
2
1
14
1
42
6
5
1
7
4
2
15
"385
5-111
-------
percent) continuous sites are considered valid. There are 7 States with no
valid S02 monitoring for 1978. EPA is currently taking steps to improve the
quality of SO- data and increase the number of representative sites reporting
valid data.
For valid bubbler sites, the average number of 24-hr observations in 1978
was 60. The number of observations per site ranged from 28 to 322. For the
valid continuous sites the mean number of observations was 7806 hourly measure-
ments. This ranged from a minimum of 6578 hr to a maximum of 8755 hr.
Overall, the SOp monitoring should improve as EPA requires quality control
procedures of State and local agencies. However, the historic data are of
unknown accuracy. Presumably the West-Gaeke bubbler data were subject to
comparable exposures, collection, and analysis procedures by the collecting
agencies. To the extent that errors are consistent in time S02 bubbler data
may still be useful for trend analysis, health studies and welfare effect
studies, using the same methods.
5.2.3.2 Distribution of Sulfur Dioxide Concentrations—Although there are
natural sources of SOp such as volcanoes (see chapter 4), they are of minor
importance. Sulfur dioxide has a rather short half-life in the troposphere
(see chapter 6). Background levels are often measured as zero. Therefore it
is not surprising that the annual mean S02 concentration is 3 ug/m in some
locations.
Monitoring in urbanized areas near industrial sources that use sulfur-
bearing fuels shows rather high concentrations of SO^. In 1978 the annual
mean concentrations obtained by SO^ bubblers ranged from 3 to 79 ug/m . The
valid continuous monitors registered 1978 annual mean concentrations ranging
from 3 to 152 ug/m .
5-112
-------
The concentration of S02 like that of TSP is affected by meteorological
variables influencing transport, dispersion, and removal, as well as by
topography and the configuration of sources. Spatial and temporal variations
in these parameters are reflected in the range of maximum concentration and of
the 90th percentiles reported across the Nation. For bubbler sites, the
lowest 24-hr maximum value reported by a site was 3 ug/m and the highest was
3
907 ug/m . For the valid continuous sites the spread of 24-hr maximum values
3 3
was greater, ranging from 10 ug/m at one site to 2512 ug/m another site.
Among all continuous sites reporting in 1978, regardless of validity, the
extreme 24-hour value was 3931 ug/m . These "daily" maxima are derived from
moving 24-hr means.
Table 5-23 provides summaries of statistics on SOp from valid 1978 sites.
Note the differences between bubbler and continuous data for the annual mean
values. There appears to be a substantial difference in the overall mean
concentrations as well as the maximum annual mean concentrations. The
statistics on the 90th percentile and maximum concentrations also vary greatly
between methods. Explanations of these differences are offered in a later
section.
5.2.3.3 Sulfur Dioxide Concentration By Location
5.2.3.3.1 1978 highest annual average concentrations. The reasons for SO,,
variability have been mentioned in earlier sections. This section examines
the locations with the highest annual averages and the highest maximum concen-
trations, and analyzes the distribution of SO^ concentration nationally by
site descriptors. However, because of the differences in SO,, concentrations
between bubbler and continuous monitoring, the distributions by site descriptor
will be treated separately for each method.
5-113
-------
TABLE 5-23. STATISTICS ON THE VALID S02 MONITORING SITES FOR 1978
Statistic
Bubblers (N=99)
Mean
Standard Deviation
Minimum
Maximum
Continuous (N=385)
Mean
Standard Deviation
Minimum
Maximum
Observations
60.0
38.0
28.0
322.0
7806
654
6578
8755
Annual
Mean,
|jg/m
19.5
15.4
3.0
79.0
31.5
19.3
3.0
152.0
90th Percent! le
3
|jg/m
43.6
37.5
2.0
176.0
65.2
42.1
3.0
488.0
Maximum
Concentration
pg/m
112.2
126.5
3.0
907.0
221.8
189.7
10.0
2512.0
5-114
-------
Table 5-24 lists the annual mean and the maximum 24-hr concentration for
the 11 valid continuous monitoring sites with the highest annual means in
1978. The highest annual mean concentration was 152 ug/m . This site in
Montana was situated 2 miles northeast of a smelter. It also had the highest
24-hr concentration (2512 ug/m ) of any valid continuous monitor. The maximum
3 3
hour was 7205 ug/m , and the second highest hour was 6026 ug/m at this site.
Of the highest 11 sites, 5 are associated with smelters, 5 are associated with
industrialized areas or towns, and one, New York City, is a densely populated
city. In New York City. S02 emissions from space heating, power plants, and a
variety of industrial sources resulted in a high annual mean concentration.
5.2.3.3.2 1978-highest daily average concentrations. Listed in Table 5-25
are the 45 continuous sites with the highest single-day 24-hour SO,, concentra-
tions. All but one of these sites have a second highest daily value exceeding
o
265 ug/m . Several of these sites are in the same general areas. Most likely
the high levels in these areas are the result of a single source. For example,
the sites in Deer Lodge County, Montana are experiencing high maximum 24-hr
concentrations as a result of S0? emissions from a local smelter.
Many of these sites having high maximum daily values are located near
specific industrial sources such as smelters, steel manufacturers, and paper
mills. A few of the sites like Toledo, Ohio, Hammond, Indiana, Pittsburgh,
Pennsylvania, and East Lake, Ohio, are in urbanized industrial areas. Most of
the other sites are in less populated rural regions.
High 24-hr SO^'concentrations occur in all regions of the country but not
in all States. Ten of the high sites are in Montana, six in Wisconsin and six
in Minnesota. In Table 5-25, listing the 45 sites with the highest maximum
readings, only 17 states are included.
5-115
-------
TABLE 5-24. SO, MONITORING SITES WITH THE HIGHEST 11 ANNUAL MEAN
CONCENTRATIONS IN 1978 (VALID CONTINUOUS SITES ONLY)
Location
Annual Means
ug/m
Maximum 24-hr,
ug/m
Description
Helena, Deer Lodge 152
Co., Montana
Pittsburgh, Pennsylvania 140
Helena, Deer Lodge 95
County, Montana
Magna, Salt Lake Co., 93
Utah
2512 Rural-mine smelter
602 Center city industrial
1450 Rural-industrial
1.6 miles east of
smelter
811 Suburban-industrial
Toledp, Ohio
Pittsburgh, Pennsylvania
Buffalo, New York
Kellogg, Shoshone Co.,
Idaho
Shoshone Co. , Idaho
New York City, New York
Mingo Junction, Ohio
84
79
78
78
77
77
76
915
376
267
294
493
296
329
Center city industrial
Suburban-industrial
Suburban-industrial
Suburban- residential
Suburban-industrial
Center city residential
Center city industrial
5-116
-------
TABLE 5-25. THE 45 MONITORING SITES WITH THE HIGHEST 24 HOUR SO,
CONCENTRATIONS IN 1978 (Continuous methods.) '
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Location
Eau Claire County
Wisconsin
Deer Lodge County
Montana
Deer Lodge County
Montana
Unknown
Montana
Dakota County
Minnesota
Deer Lodge County
Montana
Dougherty County
Georgia
Dakota County
Minnesota
Berlin
New Hampshire
Marathon County
Wisconsin
Toledo
Ohio
Millinocket
Maine
Tarpon Springs
Florida
Magna, Utah
Maximum 24 hr,
pg/m
3931
2512
1450
1084
1030
1015
945
933
925
918
915
894
816
811
Second
Highest 24-hr,
pg/m
3266
1880
1185
700
850
862
440
653
721
791
613
792
653
706
Description
•>
Rural agricultural
2.75 mine of smelter
Rural -industry
1.6 mi. E of smelter
?
Rural-industry
Rural-industry
Rural-industry
Rural-industry
Center city-
residential
Rural-industry
Center city-
industry
Suburban-
industrial
Rural-commercial
Suburban- industry
5-117
-------
TABLE 5-25, (continued)
Maximum 24 hr,
Location pg/m
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Laurel County
Montana
Milwaukee
Wisconsin
St. Charles County
Missouri
Gil a County
Miami, Arizona
Unknown
Minnesota
Deer Lodge County
Montana
Unknown
Montana
St. Paul Park
Minnesota
Oregon Lucas County
Ohio
Deer Lodge County
Montana
Unknown
Minnesota
Laurel
Montana
Deer Lodge County
Montana
Green! ee County
808
799
782
779
753
720
711
703
697
689
683
670
668
668
Second
Highest 24-hr,
Mg/m Description
736
482
751
677
689
530
415
663
414
682
603
279
659
561
Suburban- residential
3/8 mi. NE of re-
finery
Center city-
industry
Rural-agricultural
Rural-agricultural
7
Center city-
commercial
?
Center city-
commercial
Suburban- residential
Rural -industry
?
Suburban-residential
3/8 mi. NE of re-
finery
Rural-industry
Rural-industry
Arizona
5-118
-------
TABLE 5-25. (continued)
Maximum 24 hr,
Location ug/m
29.
30.
i
31.
n
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
St. Paul Park
Minnesota
Millinocket
Maine
Pittsburgh
Pennsylvania
Millinocket
Maine
Wood River
Illinois
Philadelphia
Pennsylvania
Philadelphia
Pennsylvania
Iron County
Cedar City, Utah
Pi ma County
Ajo, Arizona
Hammon Lake County
Indiana
Milwaukee
Wisconsin
Greenlee County
Arizona
Shoshone County
Idaho
Unknown
Alabama
655
647
602
569
553
548
534
526
522
521
518
510
493
486
Second
Highest 24-hr
ug/m
429
612
556
556
354
393
405
439
481
458
401
347
398
401
»
Description
Center city-
commercial
Rural-industry
Center city-
industry
Rural -industry
Suburban- industry
Suburban- industry
Center city
commercial
Suburb- residential
Suburb-industry
Center city-
industry-Gary area
Center city-
residential
Rural industry
1.3 mi. NW of smelter
Suburban- i ndustry
?
5-119
-------
TABLE 5-25. (continued)
Location
Maximum 24 hr.
Second
Highest 24-hr,
Description
43. Bannock County
Idaho
44. Greenbay County
Wisconsin
45. Greenbay County
Wisconsin
486
482
470
497
405
374
Rural-industry
Privately operated
5-120
-------
5.2.3.3.3 Sulfur dioxide concentrations by site. In this section the
distributions of annual mean and 90th percentile concentrations by site
descriptors are presented for bubbler and for continuous methods. As with
TSP, a two-descriptor code has been associated with each site. Distributions
for every combination of Type 1 (population, source, background) and for Type
2 (central city, suburban, rural, remote) will not be presented. In some
cases the designations are contradictory, such as "population-remote" or
"background-central city." The purpose of presenting these distributions is
to permit comparison of these two categories of sampling methods and to examine
the SCL concentrations as a function of location.
In Table 5-26 the cross-tabulation of mean concentrations by method is
presented for "population-central city" and "population-source" sites. The
third and fourth numbers in each cell are the column percentage and total
percentage of sites having concentrations within the designated range.
Examination of each of these numbers reveals that bubblers are, on the average,
reporting lower concentrations than the continuous instruments. Three explana-
tions are plausible:
1. Continuous monitors are systematically overestimating SO^ concentra-
tions.
2. Bubbler monitors are systematically underestimating S02 concentrations
3. Continuous monitors are biased upwards by placement in areas with
higher levels.
In light of the reported temperature decay problems with SOp bubblers
(chapter 3), the most likely explanation is that bubblers are underestimating
St^. In Table 5-26, 14.4 percent of the continuous monitors report mean
concentrations above 62 ug/m while only 1.6 percent of the bubblers report
5-121
-------
TABLE 5-26. CROSSTABULATION OF ANNUAL MEAN S0? CONCENTRATION BY METHOD (BUBBtER OR CONTINUOUS)
FOR POPULATION-ORIENTED AND FOR SOURCE-ORIENTED CENTER CITY SITES
in
i
no
Annual mean
S0? concentration,
Wml
2-7
Number of sites
Percent of row
Percent of column
Percent of total
7-18
Number of sites
Percent of row
Percent of column
Percent of total
18-33
Number of sites
Percent of row
Percent of column
Percent of total
33-62
Number of sites
Percent of row
Percent of column
Percent of total
>62
Number of sites
Percent of row
Percent of column
Percent of total
Bubbler
139
90.3
27.6
17.7
159
83.2
31.6
20.2
106
55.2
21.1
13.5
91
45.3
18.1
11.6
8
16.3
1.6
1.0
Population
Continuous
15
9.7
5.3
1.9
32
16.8
11.3
4.1
86
44.8
30.3
10.9
110
54.7
38.7
14.0
41
83.7
14.4
5.2
Purpose
Row Total
154
19.6
191
24.3
192
24.2
201
25.5
49
6.2
of Site
Bubbler
9
75.0
16.1
8.7
18
64.3
32.1
17.3
12
46.2
21.4
11.5
13
46.4
23.2
12.5
4
40.0
7.1
3.8
Source
Continuous
3
25.0
6.3
2.9
10
35.7
20.8
9.6
14
53.8
29.2
13.5
15
53.6
31.3
14.4
6
60.0
12.5
5.8
Row Total
12
11.5
28
26.9
26
25.0
28
26.9
10
9.6
-------
TABLE 5-26. (continued)
Annual mean
SQy concentration,
M9/mr
Column total
Number of sites
Percent of total
Bubbler
503
63.9
Population
Continuous
284
36.1
Purpose
Row Total
787
100.0
of Site
Bubbler
56
53.8
Source
Continuous
48
46.2
Row Total
104
100.0
ro
co
-------
these higher concentrations. Of the "source-center city" sites a higher
o
percentage (7.1 percent) of the bubblers are above 62 ug/m , but this is still
less than the 12.5 percent of the continuous monitors in this category.
The pattern is consistent with the display of the "population-suburban"
and "source-suburban" distributions in Table 5-27. Where monitors were located
in rural regions near industrial sources, only 2 percent of the bubblers
reported means greater than 33 ug/m , while 17 percent of the continuous sites
in similar locations were above this level.
Similar distributions were tabulated for the 90th percentiles. Table
5-28 displays these distributions for all "source"-oriented monitors in central
cities, suburban, rural and remote areas. The concentrations reported by
bubbler sites tends to be lower than those reported by continuous sites. This
can be seen by comparing the row and column percentages between methods in
Table 5-28. In part this may be due to the low sensitivity of the bubbler
method. EPA reports that the method is not very reliable below 24 ug/m .
A systematic difference between methods is more apparent in the cross-
tabulation of the "population" sites in Table 5-29. Disregarding for a moment
the difference between methods, only 4.6 percent of the central-city population-
o
oriented monitors reported 90th percentiles over 127 ug/m . Less than 4
percent of the suburban population-oriented monitors reported 90th percentile
o
levels over 127 ug/m . Again, this does not necessarily imply that less than
5 percent of the population is exposed less than 10 percent of the time to
o
concentrations above 127 ug/m . Referring to the previous table of source
oriented monitors, the percentage of monitors in central cities and suburban
locations that exceed 127 ug/m for their 90th percentile concentration is
somewhat greater (5 to 7 percent) than for the population-oriented monitors.
5-124
-------
TABLE 5-27.
CROSSTABULATION OF ANNUAL MEAN SO, CONCENTRATION BY METHOD (BUBBLER OR CONTINUOUS)
FOR POPULATION-ORIENTED AND FOR SOURCE-ORIENTED SUBURBAN SITES
ro
in
Annual mean
S0? concentration,
pg/ml
2-7
Number of sites
Percent of row
Percent of column
Percent of total
7-18
Number of sites
Percent of row
Percent of column
Percent of total
18-33
Number of sites
Percent of row
Percent of column
Percent of total
33-62
Number of sites
Percent of row
Percent of column
Percent of total
>62
Number of sites
Percent of row
Percent of column
Percent of total
Bubbler
139
89.7
36.5
22.3
101
71.6
26.5
16.2
88
51.8
23.1
14.1
50
37.6
13.1
8.0
3
12.5
0.8
0.5
Purpose
Population
Continuous Row Total
16 155
10.3
6.6
2.6 24.9
40 141
28.4
16.5
6.4 22.6
82 170
48.2
33.9
13.2 27.3
83 133
62.4
34.3
13.3 21.3
21 24
87.5
8.7
3.4 3.9
of Site
Bubbler
21
91.3
23.9
13.0
35
72.9
39.8
21.6
16
32.0
18.2
9.9
13
43.3
14.8
8.0
3
27.3
3.4
1.9
Source
Continuous
2
8.7
2.7
1.2
13
27.1
17.6
8.0
34
68.0
45.9
21.0
17
56.7
23.0
10.5
8
72.7
10.8
4.9
Row Total
23
14.2
48
29.6
50
30.9
30
18.5
11
6.8
-------
TABLE 5-27. (continued)
Annual mean
SCL concentration,
ug/ml
Column total
Number of sites
Percent of total
Bubbler
381
61.2
Population
Continuous
242
38.8
Purpose of
Row Total
623
100.0
Site
Bubbler
88
54.3
Source
Continuous
74
45.7
Row Total
162
100.0
en
ro
en
-------
TABLE 5-28. CROSSTABULATION OF 90TH PERCENTILE OF DAILY AVERAGE CONCENTRATIONS OF S0? BY METHOD
(BUBBLER OR CONTINUOUS) FOR SOURCE-ORIENTED SITES IN CENTER CITY, SUBURBAN, RURAL,
AND REMOTE LOCATIONS
ro
90th Percentile of
daily average S0« ^
concentration, (jg/m
2-14
Number of sites
Percent of row
Percent of column
Percent of total
14-39
Number of sites
Percent of row
Percent of column
Percent of total
39-68
Number of sites
Percent of row
Percent of column
Percent of total
68-127
Number of sites
Percent of row
Percent of column
Percent of total
>127
Number of sites
Percent of row
Percent of column
Percent of total
Location
Center City
Bubbler
13
76.5
23.2
12.5
14
66.7
25.0
13.5
10
35.7
17.9
9.6
13
48.1
23.2
12.5
6
54.5
10.7
5.8
Contin-
uous
4
23.5
8.3
3.8
7
33.3
14.6
6.7
18
64.3
37.5
17.3
14
51.9
29.2
13.5
5
45.5
10.4
4.8
Row
Total
17
16.3
21
20.2
28
26.9
27
26.0
11
10.6
Suburban
Bubbler
24
88.9
27.3
14.8
28
65.1
31.8
17.3
18
38.3
20.5
11.1
14
46.7
15.9
8.6
4
26.7
4.5
2.5
Contin-
uous
3
11.1
4.1
1.9
15
34.9
20.3
9.3
29
61.7
39.2
17.9
16
53.3
21.6
9.9
11
73.3
14.9
6.8
Row
Total
27
16.7
43
26.5
47
29.0
30
18.5
15
9.3
Rural
Bubbler
36
65.5
43.4
13.2
30
37.0
36.1
11.0
12
16.7
14.5
4.4
3
6.7
3.6
1.1
2
10.5
2.4
0.7
Contin-
uous
19
34.5
10.1
7.0
51
63.0
27.0
18.8
60
83.3
31.7
22.1
42
93.3
22.2
15.4
17
89.5
9.0
6.3
Row
Total
55
20.2
81
29.8
72
26.5
45
16.5
19
7.0
Remote
Bubbler
1
50.0
100.0
12.5
0
0.0
0.0
0.0
0
0.0
0.0
0.0
Contin-
uous
1
50.0
14.3
12.5
5
100.0
71.4
62.5
1
100.0
14.3
12.5
Row
Total
2
25.0
5
62.5
1
12.5
-------
TABLE 5-28. (continued)
90th Percent! 1e of
daily average SOp 3
concentration, pg/m
Column total
Number of sites
Percent of total
Location
Center City
Bubbler
56
53.8
Contin-
uous
48
46.2
Row
Total
104
TOO.O
Suburban
Bubbler
88
54.3
Contin-
uous
74
45.7
Row
Total
162
100.0
Rural
Bubbler
83
30.5
Contin-
uous
189
69.5
Row
Total
272
100.0
Remote
Bubbler
1
12.5
Contin-
uous
7
87.5
Row
Tota'
8
100.0
en
i
f\J
00
-------
TABLE 5-29. CROSSTABULATION OF 90TH PERCENTILE OF DAILY AVERAGE CONCENTRATIONS OF
S0? BY METHOD (BUBBLER OR CONTINUOUS) FOR POPULATION-ORIENTED MONITORS
CENTER CITY, SUBURBAN, AND RURAL LOCATIONS.
ro
10
90th Percent! le of
daily average S0? -
concentration, \jg/m
2-14
Number of
Percent of
Percent of
Percent of
14-39
Number of
Percent of
Percent of
Percent of
39-68
Number of
Percent of
Percent of
Percent of
68-127
Number of
Percent of
Percent of
Percent of
>127
Number of
Percent of
Percent of
Percent of
sites
row
column
total
sites
row
column
total
sites
row
column
total
sites
row
column
total
sites
row
column
total
Location
Center City
Bubb
1
90
33
21
ler
67
.8
.2
.2
136
77
27
17
.7
.0
.3
104
52
20
13
44
16
10
30
2
1
.0
.7
.2
85
.3
.9
.8
11
.6
.2
.4
Contin-
uous
17
9.2
6.0
2.2
39
22.3
13.7
5.0
96
48.0
33.8
12.2
107
55.7
37.7
13.6
25
69.4
8.8
3.2
Row
Total
184
23.4
175
22.2
200
25.4
192
24.4
36
4.6
Suburban
Bubbler
146
89.0
38.3
23.4
95
66.0
24.9
15.2
90
54.2
23.6
14.4
48
37.5
12.6
7.7
2
9.5
0.5
0.3
Contin-
uous
18
11.0
7.4
2.9
49
34.0
20.2
7.9
76
45.8
31.4
12.2
80
62.5
33.1
12.8
19
90.5
7.9
3.0
Row
Total
164
26.3
144
23.1
166
26.6
128
20.5
21
3.4
Bubbler
25
86.2
43.1
29.4
23
85.2
39.7
27.1
9
45.0
15.5
10.6
1
12.5
1.7
1.2
0
0.0
0.0
0.0
Rural
Contin-
uous
4
13.8
14.8
4.7
4
14.8
14.8
4.7
11
55.0
40.7
12.9
7
87.5
25.9
8.2
1
100.0
3.7
1.2
Row
Total
29
34.1
27
31.8
20
23.5
8
9.4
1
1.2
-------
TABLE 5-29. (continued)
90th Percent! le of
daily average SCL ,
concentration, ug/m
Column total
Number of sites
Percent of total
Location
Center City
Bubbler
503
63.9
Contin-
uous
284
36.1
Row
Total
787
100.0
Suburban
Bubbler
381
61.2
Contin-
uous
242
38.8
Row
Total
623
100.0
Bubbler
58
68.2
Rural
Contin-
uous
27
31.8
Row
Total
85
100.0
in
i
U)
o
-------
These comparisons are quite general. In order to make more accurate
approximations of the population exposures to high ambient SO,, concentrations,
two steps must be taken. First the discrepancy in SOp sampling methods must
be resolved so that measurements can be compared. Second, a careful analysis
around each high monitor must be performed. Perhaps this would included
dispersion modeling, but as a minimum the population actually exposed should
be tabulated. The last point deserves emphasis. The current data bases
contain no information on the population each monitoring site is intended to
represent.
5.2.3.4 Sulfur Dioxide Cycles
5.2.3.4.1 Diurnal patterns. In some locations SCL concentrations have distinct
temporal patterns. These patterns depend on the variability of meteorological
factors and on the variability of source emissions.
Diurnal variations in SOp concentrations reflect the changing dispersion
characteristics of the lower atmosphere and variations in mixing height. If
emissions are predominantly from low-level sources such as residential and and
institutional space heating, the highest hourly concentrations will frequently
occur at night and in the early-morning hours. At these times, low mixing
height and decreased wind speeds lead to higher concentrations. During the
day more vertical mixing usually occurs and wind speeds increase; this results
in the dilution of low-level emissions. Figure 5-41 give the monthly mean
hourly concentrations for S02 in December 1978 in Watertown, MA. The pattern
just described is apparent.
In locations where SCL emissions from taller stacks are the major S02
source, a different diurnal pattern can occur. In these situations, typical
of power plants and smelters, the highest concentrations usually occur in the
5-131
-------
0.040
0.030
<
oc
u
o
u
CM
8
0.020
0.010
1 1 1 1 1 T
1 1 r
J I I I J
0 2 4 6 8 10 12 14 16 18 20 22 24
HOUR
Figure 5-41. Monthly means of hourly sulfur dioxide concentrations are shown for Watertown, MA, for
December 1978.
5-132
-------
morning hours just after sunrise. Levels can be almost zero at night if the
source is emitting into a stratified region above a lower level inversion.
Upon inversion breakup, when heating at the surface causes vertical mixing, an
elevated plume can be mixed to the ground. Fumigation conditions lasting from
several minutes to several hours can occur. The monthly mean hourly plot of
S0? concentrations at a site in Kingston, TN, illustrates this point (See
Figure 5-42). In 1975, a nearby 1500-MW power plant was emitting through nine
stacks less than 320 ft high. In 1978 the emission had been switched over to
two 1000-ft stacks. Analysis by Montgomery and Coleman (1975) of the effects
of tall stacks on the peak-to-mean ratios for different averaging times discusses
the influence of inversion breakup. In essence, even with tall stacks, inversion
breakup that catches the plume and brings it to the surface can occur. So the
peak-to-mean ratio is almost independent of stack height. The frequency of
occurrences on the other hand would most likely be less with taller stacks.
In Figure 5-42 the plot for the monthly mean of hourly values reveals a
different pattern from that found in Watertown, MA. The maximum hours occur
in the late morning. By 1978, the emissions from the nearby power plant were
discharged through two 1000-ft towers. The monthly mean of the hourly observa-
tions in January 1978 shows a less pronounced diurnal cycle. Some similarity
can be found in comparing the monthly mean hourly averages for Watertown,
Massachusetts (Figure 5-41), and St. Louis, Missouri (Figure 5-43). In
February, 1977, a major local source of SO^ was still in operation in St.
Louis. The midmorning and late night maxima are again associated with diurnal
variations of meteorological factors. By February, 1978, the source had shut
down and the S0? levels at the monitoring stations reflect this fact. The
absence of low-level stacks emitting into a stable layer of air near the
5-133
-------
0.050
0.040
a
a
•>
§ 0.030
Ul
0
S 0.020
CM
0.010
JAN 1975
10
.12 14
HOUR
16
18
20
22
24
Figure 5-42. Monthly means of hourly sulfur dioxide concentrations are shown for Kingston (TVA site
44-1714-003, "Laddie Village") for January 1975 and 1978.
5-134
-------
0.200
I I I I I I I I I I I I I I I I I
i I I I
FEBRUARY 1977
FEBRUARY 1978
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0.010
1 2 3
Figure 5-43. Monthly means of hourly sulfur dioxide concentrations are shown for St. Louis (city site
no. 26-4280-007, "Broadway & Hurck") for February 1977 and 1978.
5-135
-------
surface at night is noticeable. Concentrations do not build up at night as in
the previous year.
The SOp monthly mean hourly values for the industrialized river valley
town of Steubenville, OH, are shown in Figure 5-44. In June of 1976, a distinc-
tive maximum in the diurnal pattern appears. In July of 1977 there is no
apparent variation across the hours.
In concluding this section it can be said that the variations in hourly
concentrations are influenced by source configuration and meteorological
dispersion. Therefore, it is difficult to generalize about the pattern of the
hourly concentrations. It has been shown that, although there may be some
similarities, the daily patterns in SOp concentrations are different for
different locations and can change in time for a given location.
5.2.3.4.2 Seasonal Patterns. Concentrations of SOp display seasonal variability.
The variability is most pronounced in areas in which there is strong seasonal
variation in the emission-source strength or in meteorological conditions.
Obviously, in urban areas where space heating is the major source of SOp, the
levels will be much higher during the heating season. Figure 5-45 illustrates
just such situations in Watertown, MA and Steubenville, OH. The highest
monthly mean concentrations occur in the winter months.
Figure 5-45 also shows the data for St. Louis, Missouri, where the seasonal
pattern is different. Here a local industrial source dominates SOp concentration
patterns around the monitor. The higher monthly mean concentrations occur in
the months with the higher frequency of south winds. The source is to the
south of the monitoring station. Any increase in SOp concentrations as a
result of the winter heating season is not apparent.
5-136
-------
0.080
0.070
0.060
0 0.050
c
0.040
ui
0
S 0.030
8N
0.020
0.010
I I I I
10
12
14
16
18
20
HOURS
22
24
Figure 5-44. Monthly means of hourly sulfur dioxide concentrations are shown for Steubenville, OH
(NOVAA site 36-6420-012) for June 1976 and 1977.
5-137
-------
0.040
0.030
a
a.
o
H
K
Z
UJ
u
8
IN
0.020
0.010
CITY SITE
STEUBENVILLE 26 6420 012
ST. LOUIS 26 4280 007
WATERTOWN 22 2380 003
I
I
I
I
I
I
I
I
JAN
FEB MAR APR MAY
JUNE JULY AUG SEPT OCT
MONTH
NOV DEC
Figure 5-45. Seasonal variations in sulfur dioxide levels are shown for Steubenville, St. Louis, and
Watertown.
5-1?.?.
-------
5.2.3.5 National Status of Sulfur Dioxide Concentrations
5.2.3.5.1 National View. Of the 2365 monitoring sites reporting data in the
year 1977, only 1355 had sufficient data in each quarter and for the entire
year to support a valid annual mean. Of these sites, nineteen reported values
above the primary annual mean standard for sulfur dioxide of 80 ug/m (0.03
ppm). This represents approximately 1 percent of all the stations reporting.
Of 2365 monitoring stations, fifty-eight reported levels above the 24-hr
primary standard. This is based on the second-highest 24-hour average exceeding
o
365 ug/m . These 58 stations represent only 2 percent of all the reporting
stations. Thirty sites reported levels above the 3-hr secondary standard of
3
1300 ug/m or 0.5 ppm. Again this represents approximately 1 percent of all
the monitoring sites. (See Table 5-30).
In 1978, five of the 484 valid sites reported values above the annual
standard. This is only 1 percent of valid sites. Forty-four continuous-analyzer
».
and two bubbler sites report a second-highest daily concentration above 365
pg/m . Again, this is approximately the same proportion of nonattainment
sites as in 1977.
Areas of the country where S0? levels exceed NAAQS are usually associated
with large point sources or a complex of industrial sources. However, many
nonindustrialized or lightly industrialized urban areas in the Northeast have
elevated S0? because of widespread use of high-sulfur fuels for space heating
and in power plants. In the South, emissions from industrial sources are
important sources of SOp. In the West, power plants and smelters are the
dominant sources of SO^. Over the entire Nation, SO^ exposures vary greatly
in complexity. A single source in flat or complex terrain can cause SOp
concentrations with a large standard deviation, depending on wind direction
5-139
-------
TABLE 5-30. NATIONAL SUMMARY FOR SULFUR DIOXIDE TOTAL STATIONS REPORTING DATA, NUMBER AND PERCENTAGE
OF SITES REPORTING CONCENTRATIONS ABOVE THE NATIONAL AMBIENT AIR QUALITY STANDARDS 1977-1978
Number of Sites Percent of sites
Data record and Reporting Violating exceeding NAAQS
Standard exceeded 1977 1978 1977 1978 1977. 1978
Valid annual data3
annual standard
Minimal data
1355
2365
484
2245
19
58
5
44
1
2
1
2
24-hr standard
3-hr Standard 30 18 1 <1
y, _ - --
-L Valid annual data record must contain at least five of the scheduled 24-hour samples per quarter for EPA- recommended
o intermittent sampling (once every 6 days) or 75 percent of all possible values in a year for continuous instruments.
Minimal data consist of at least three 24-hour samples for intermittent sampling monitors or 400 hourly values for
continuous instruments.
-------
and mixing conditions. In contrast those living in nonindustrial-urban
communities could experience seasonal variability with high concentrations in
the winter. Those living in an industrialized river valleys could experience
persistently high SO,, levels.
The regions experiencing elevated S02 concentrations are displayed in a
series of national maps. Figure 5-46 shows the AQCR's where valid monitors in
1977 reported annual concentrations greater than 80 ug/m . The eastern AQCR's
were associated with industrial urban areas for the most part, many of them in
river valley settings. The AQCR's in the west have strong point sources in
complex terrain settings. Unfortunately, 59 AQCR's had insufficient data.
Another 179 AQCR's had monitoring, with all sites reporting annual levels
3
below 80 ug/m .
The AQCR's where at least one monitor reported its second highest daily
3
value above 365 ug/m (0.14 ppm) is 1977 are shown in Figure 5-47. More
AQCR's appear to have a problem with daily concentrations. Twenty-nine AQCR's
3
had a monitor whose second highest concentration was greater than 365 ug/m .
Eighteen AQCR's had insufficient or no data, and in the remaining 200 all
sites reported second highest daily values below 365 ug/m .
Although the data base used in compiling Figure 5-48 was collected between
1974 and 1976, it does offer finer spatial resolution of national SOp concentra-
tions. The second highest 24-hr average S02 concentration by county is displayed
Some areas in the west with extremely high concentrations are still problem
areas in the late 1970's. Comparing this figure with Table 5-25, which shows
the 45 sites with the highest concentrations, confirms that several counties
and cities are still reporting high concentrations. As with the previous
figures showing S0? levels by AQCR's, one should not infer spatial averaging.
5-141
-------
Mm AQCR WHERE AT LEAST ONE MONITOR EXCEEDS 80 jig/m
CD AQCR WHERE NO MONITORS EXCEED 80 g/m3
NO DATA OR INADEQUATE RECORD
Figure 5-46. National status is shown for average annual (1977) sulfur dioxide concentrations; less
than 100, 100 to 200, 200 to 365, and greater than 365 M9/m3.
5-142
-------
£•£} 200 pg/mj £ 90TH %<365
* i 100 pg/rr
I I BOTH % <100
<200 pg/m
Figure 5-47. Statistical characterization of 1977 national sulfur dioxide status is shown by 90th percen-
tile concentrations.
5-143
-------
Figure 5-48. Characterization of 1974-76 national SO2 status is shown by second highest 24-hr average
concentration. The 24-hr primary national ambient air quality standard is 365 pg/m3, which is not to
be exceeded more than once per year.
Source: Monitoring and Reports Branch, Monitoring and Data Analysis Division, Off ice of Air Quality
Planning and Standards, U.S. Environmental Protection Agency.
5-144
-------
High readings may exist at one or more monitoring sites (for example, Deer
Lodge County, MT), but it is likely that there are substantial gradients
across the county, and almost certainly across an AQCR.
As indicated earlier, the range of valid annual averages from continuous
monitors in 1978 extends from 3 to 158 ug/m . The range for bubblers is from
3 to 79 ug/m . Hence no bubbler site exceeds the NAAQS, whereas five continuous-
o
monitor sites, or roughly 1 percent did exceed 100 ug/m in 1978. Plotting
the distribution of the percentage of sites less than a specific concentration
offers further evidence of the discrepancy in sampling methods. Figure 5-49
plots the cumulative frequency distribution for valid 1978 bubbler and continuous
monitor sites. Figures 5-50 and 5-51 are histograms of the annual average S02
concentrations (ug/m ) for the valid bubbler and valid continuous-monitor
sites in 1978.
According to the bubbler method, 50 percent of the monitors are averaging
3 3
less than 12 ug/m . The 50th percentile for continuous monitors is 27 ug/m .
A few interesting observations can be made from these distributions. If the
five continuous-monitor sites represent special situations, such as smelter
monitoring, and were set aside, then the distributions for both methods would
be very similar. Second, the appearance of a "light tail" to the bubbler
distributions indicates that the concentrations are not lognormally distributed.
Presumably if the bubbler method is understimating S02 ambient concentrations
H is biasing the higher values to a greater extent.
Figure 5-52 shows the cumulative frequency distributions of the 90th
percentiles for valid bubbler and continuous-monitor sites. The distributions
indicate 5 percent of the sites had approximately 10 percent of their reported
24-hr averaged concentrations exceeding 120 ug/m . One percent of the sites
5-145
-------
200
100
90
.80
70
60
50
40
C8
X
z
tu
u
z
O
CJ
IN
O
V)
Z
30
20
10
9
8
7
111 111
I I I
I I
CONTINUOUS
III III
|
J I J J I I I I I
I
I
MAX SITE
MAX SITE .^
—
I I
I
0.01 0.1 0.2 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.9 99.99
% OF VALID MONITORS WITH CONCENTRATION LESS THAN STATED VALUE
Figure 5-49. Distribution of annual average sulfur dioxide concentration in 1978 is shown for con-
tinuous and bubbler methods.
5-146
-------
30
30
V)
Ul
55 25
cc
LU
CD 20
CO
o
NUMBER OF
o ui o w
—
—
—
1 1
—
—
, ,
10 20 30 40 50 60 70
ANNUAL AVERAGE CONCENTRATION
80
Figure 5-50. Histogram shows annual average sulfur dioxide
concentrations for valid bubbler sites, 1978.
5-147
-------
120
110
100
90
V)
Ul
5 80
V)
O
Z
O
O
u.
O
cc
UJ
CO
70
60
50
40
30
20
10
0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
ANNUAL AVERAGE CONCENTRATION
Figure 5-51. Histogram shows annual average sulfur dioxide concentrations for valid continuous sites,
1978.
5-148
-------
I I I I I I I
0.01 0.1 0.2 0.5 1 2 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 99.99
PERCENT OF SITES WITH 90TH PERCENTILE CONCENTRATIONS LESS THAN VALUE INDICATED
Figure 5-52. Cumulative frequency distribution of the 90th percentile is shown for valid bubblers and
continuous monitors.
5-149
-------
had daily observations exceeding 150 |jmg/m 10 percent of the time (equivalent
to 36 days per year at each of these sites).
The convergence of the distributions for bubbler and continuous-monitor
measurements at higher concentrations is intriguing. At first, it appears to
contradict the distributions of the annual averages. However, if bubbler
results are primarily affected at high temperatures then there is an explana-
tion. For bubbler sites located in areas where SOp emissions and higher S0?
concentrations occur primarily in the winter, the temperature effect would not
substantially reduce the higher concentrations. However, high measurements in
warmer weather would be affected. In addition, if the laboratories using
bubblers to determine SOp were to report all values below the lower detectable
limit as zero, the distribution would be affected in a similar manner.
The upper end of the distributions in Figures 5-49 and 5-52 represent
very few sites. It is therefore not surprising to find that the maximum or
second-highest site is not consistent with the full distribution.
5.2.3.5.2 Regional distribution of SOp
Regional differences in SOp concentrations are not striking. In part,
this is the result of the locational dependency of the monitor. In previous
sections it was shown that high SOp levels are found around smelters in
otherwise clean areas. In the eastern and northern states, most continuous
SOp monitoring is in urbanized areas. Mean concentrations across all continuous
monitors in Regions I, II, III, IV, and V range from 23 ug/m to 51 ug/m .
(See Table 5-32). The maximum annual mean for a single site in these regions
ranges from 59 ug/m in Region I to 140 ug/m in Region III among the valid
sites. In the less industrialized or populated regions (VI through X), the
3
mean annual concentration across all sites in each region ranges from 8 ug/m
o
to 40 ug/m .
5-150
-------
TABLE 5-31. BUBBLER S02 MONITOR RESULTS BY REGION
en
i
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
Type
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Number of
sites
16
108
2
12
0
140
26
368
55
307
0
196
0
63
0
40
0
0
0
15
Min.
40
2
28
5
—
1
30
1
41
I
_
1
_
5
-
2
-
—
-
6
Number of
observations
per site
Mean Max. s.d.
92
34
30
13
_
27
52
23
56
33
_
18
_
24
-
21
-
-
-
12
322
322
31
31
—
181
61
111
61
181
_
44
_
47
_
38
_
-
-
15
90
50
2
9
_
20
8
15
5
18
_
9
_
11
_
10
_
-
-
4
Arithmetic Means
Min. Mean Max. s.d.
15
3
5
3
_
3
3
3
3
3
_
3
_
3
_
3
_
-
-
11
22
27
7
21
_
22
9
13
24
27
_
5
_•
9
_
3
_
-
-
27
54
111
8
57
— f
70
57
68
79
104
_
32
_
48
_
8
_
-
-
63
10
18
2
18
_
13
13
12
15
16
_
4
_
7
_
1
_
-
-
21
90th
Min.
13
2
13
2
_
2
2
2
2
2
_
2
_
2
_
2
_
-
-
13
Percentile
Mean Max.
47
51
15
46
_
49
18
28
56
58
_
9
_
18
_
3
_
-
-
34
146
245
16
96
_
209
127
173
176
210
_
102
_
149
-
22
_
-
-
76
s.d.
31
35
2
33
_
33
29
27
37
35
_
14
_
22
-
4
-
—
-
25
-------
TABLE 5-31. (continued)
Region
Alaska
Hawaii
Nation
Type
Valid
All
Valid
All
Valid
All
Number of
observations
Number of per site Arithmetic Means 90th Percentile
sites Min. Mean Max. s.d. Min. Mean Max. s.d. Min. Mean Max. s.d.
0 ..__ _.__-.--
7 1 21 29 11 5 9 11 2 10 18 31 9
0 _.-_ ._.__-_-
9 10 13 15 2 3 10 36 11 2 18 73 24
99
1265
en
i
en
ro
-------
TABLE 5-32. CONTINUOUS S02 MONITOR RESULTS BY REGION.
en
co
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
Type
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Number of
sites
22
72
51
87
26
108
100
203
111
254
13
32
13
38
12
49
19
52
18
29
Min.
6665
185
6597
140
6578
94
6678
421
6580
129
6631
1669
6769
334
6741
373
6857
105
6651
625
Number of
observations
per site
Mean Max. s.d.
7519
4582
7540
5815
7562
4381
8305
5754
7640
5512
7443
5461
7540
4676
7739
4694
7952
4973
7854
6158
8416
8416
8697
8697
8638
8638
8755
8755
8715
8715
8452
8452
8325
8325
8624
8624
8638
8638
8677
8677
567
2720
546
2380
670
2534
574
2848
625
2327
607
2072
439
2624
514
2358
507
2525
464
2526
Arithmetic Means
Min. Mean Max.
16
8
15
15
12
7
5
3
7
3
3
3
6
4
3
3
3
3
13
13
33
39
37
41
51
45
23
23
36
37
13
12
31
25
40
34
8
24
34
33
59
138
78
94
140
140
63
77
84
192
31
56
47
82
152
152
29
87
78
78
s.d.
12
23
16
19
23
21
12
13
16
25
7
13
14
20
47
39
6
20
18
17
90th Percentile
Min. Mean Max.
34
14
35
35
34
14
9
3
10
5
3
3
13
5
3
3
3
3
35
29
65
77
72
78
97
86
54
49
70
73
31
29
62
52
100
89
16
49
90
72
147
340
159
173
282
282
135
180
167
501
69
160
94
155
488
488
48
213
150
150
s.d.
27
52
30
33
46
40
27
29
30
50
19
38
25
41
146
113
12
49
38
38
-------
TABLE 5-32. (continued)
Region Type
Alaska Valid
All
Hawaii Valid
All
Nation Valid
All
i
en
Number of
sites
0
4
0
0
385
928
Number of
observations
per site Arithmetic Means 90th Percentile
Min. Mean Max. s.d. Min. Mean Max. s.d. Min. Mean Max. s.d.
1965 2057 2114 65 3 8 17 6 3 13 22 8
---- ________
-------
Even with the summary of the 1978 continuous SCL data as displayed in
Table 5-32, it is difficult to speculate on regional differences in SCL concen-
trations. Concentrations are influenced primarily by local sources. The
locations of monitors have clearly not been randomly chosen in each region,
nor have they been systematically deployed for source population or background
sampling. Therefore, a better indicator of regional differences in S02 concen-
trations and population exposures can be obtained from examining sulfur emission
patterns. (See Chapter 4).
Another approach to examining the differences between regions in SO,,
concentrations is offered in Table 5-33. In this table the number of sites
3
exceeding an annual average of 80 ug/m and the number of sites where the
q
second highest 24-hour average exceeds 365 ug/m are tabulated. The "potential"
for higher S0? concentrations can be inferred from the number of all sites
(valid and invalid) that fall into these two categories. The definition of
"valid site" in this case refers to the number of observations, not the accuracy
of the data.
Table 5-33 indicates that Region V has the greatest number of sites with
a second highest daily value exceeding 365 ug/m . This is followed by Region
VIII with 12 "potentially" high sites.
The general conclusions that might be drawn from this table are that the
regional and national status of SO,-, attainment of standards might be
substantially underestimated. There are two possible reasons. One is that
bubbler SOp monitors do not report the same proportion of violations as do
continuous monitors. The second is that only 41 percent of the continuous
monitors are considered valid for determining annual averages; additional
locations with high S0? concentrations are not noted in national summaries.
5-155
-------
TABLE 5-33. COMPARISON OF BUBBLER AND CONTINUOUS S02 MONITOR RESULTS BY REGION
Total
number of sites
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
Alaska
Hawai i
Nation
Type
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
Valid
All
continuous
22
72
51
87
26
108
100
203
111
254
13
32
13
38
12
49
19
52
18
29
0
4
0
0
385
928
bubbler
16
108
2
12
0
140
26
368
55
307
0
196
0
63
0
40
0
0
0
15
0
7
0
9
99
1265
Number of sites with
arithmetic mean Number with second highes
> 80 ug/m3 value > 365 ug/m3
continuous
0
4
0
4
1
6
0
0
1
11
0
0
0
1
3
6
0
2
0
0
0
0
0
0
5
34
bubbler
0
1
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
continuous
1
4
0
0
2
3
6
7
8
20
0
0
0
2
3
12
0
4
4
4
0
0
0
0
24
56
bubbler
0
1
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
• —
5-156
-------
5.2.3.6 Severity of Ambient Exposures—The 90th percentile of the 1977 sulfur
dioxide data was examined for every AQCR. The number of S02 monitors in each
AQCR and the subsets that reported their 90th percentile level <100 ug/m ,
between 100 and 200, between 200 and 365, and >365 ug/m3 are tabulated in
Table 5-34. The 90th percentile serves as an indicator of the severity of the
SOp ambient concentrations. It represents the concentrations that the popula-
I
tion might be repeatedly exposed to as frequently as once each ten days. In
other words, it represents an ambient SO,, concentration that might be experienced
some 870 hr/yr, or for 36 days of the year.
The severity of S02 exposure, according to the 90th percentile level is
displayed for each AQCR in Figure 5-53. On the basis of the 1977 aerometric
data, most of the urban areas of the East and Northeast frequently have S0?
levels exceeding 100 ug/m . The urban areas around Washington, D.C., western
Pennsylvania, eastern Ohio, southern Indiana, northern Kentucky, St. Louis,
and Indianapolis have levels that repeatedly exceed 200 mg/m and some are
even greater than 365 mg/m . Close examination of the aerometric data indicates
that the peak concentrations are often two to five times the 90th percentile
value. Again, the large point sources in the West are responsible for the
high 90th percentile levels in Montana, Nevada, Idaho, Arizona, and New Mexico.
Throughout the Midwest and high Plains States, the 90th percentile levels are
less than 100 mg/m , indicating very low means in these areas.
5.2.3.7 Sulfur Dioxide Concentrations And Historic Trends
5.2.3.7.1 National trends. The SOp levels in most urban areas in the United
States have improved steadily since the mid-1960's. The trend of decreasing
SO^ concentrations can be resolved into three distinct periods. From 1964-69
the improvement was gradual. In the middle period, between 1969 and 1972, the
5-157
-------
TABLE 5-34. NINTIETH-PERCENTILE SULFUR DIOXIDE LEVELS, 1977.
(NADB 1977, S02 24-HR BUBBLERS AND 1-HR CONTINUOUS)
State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
AQCR
1
2a
3
4
5a
6
7a
8
9
10
11
12a
13a
14a
15
16a
17a
18a
19a
20
21
22a
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Number
of sites
1
1
1
5
69
2
17
1
1
4
13
3
16
18
7
3
9
2
1
1
6
34
2
5
16
4
3
1
8
Number of Sites Reporting
90th percentile concentrations (uq/m )
< 100 100-200 200-365 > 365
1
1
1
5
68 1
2
16 1
1
1
4
10 1 2
3
15 1
15 1 2
7
3
9
2
1
1
6
29 5
1 1
5
16
4
3
1
8
Interstate AQCR, reported once
5-158
-------
TABLE 5-34 (continued)
State
Connecticut
Delaware
D.C.
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
AQCR
41
42a
43a
44
45a
46
47
48a
49a
50
51
52
53a
54
55a
56
57a
58a
59
60
61
62a
63
64
65a
66
67
68a
69a
70
71
72a
73
74
75
76
77a
78a
79a
80
81.
82a
83
84
Number
of sites
3
34
100
1
47
3
75
11
48
28
11
53
8
7
15
15
1
13
3
60
4
14
1
2
14
3
114
4
7
29
4
29
3
5
10
6
28
31
36
22
2
17
5
11
Number of Sites Reporting 3
90th percentile concentrations (p3/m )
< 100
2
31
93
1
13
3
56
11
48
28
11
50
8
7
12
14
1 (1 obs)
12
3
10
2
8
1
2
11
3
96
4
7
15
4
27
3
3
8
6
28
18
30
13
2
12
4
10
100-200 200-365 > 365
1
3
7
23 1
17 2
2 1
3
1
3
9 1
2
132
3
18
11 2 1
' 2
2
2
10 3
6
8 1
5
1
1
5-159
-------
TABLE 5-34. (continued)
State
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massa-
chusetts
AQCR
69a
85a
86a
87a
88
89
90
91
92
93
94a
95
96
97
98
99
100
72a
77a
78a
79a
101
102
103a
104
105
19a
22a
106a
107a
108
109
110
47a
112
113a
114
115
116
42a
117
118
119
120a
121a
Number
of sites
9
2
3
4
1
1
2
13
1
20
14
2
3
2
14
2
7
13
25
8
6
24
25
2
20
8
7
20
6
46
3
6
13
26
24
28
Number of Sites Reporting
90th percentile concentrations (uq/m )
< 100 100-200 200-365 > 365
9
2
3
3 1
1
1
2
13
1
19 1
14
2
3
2
14
2
7
13
25
8
6
24
23 2
2
17 1 2
6 2
7
17 3
6
44 2
3
6
13
25 1
22 2
25 3
5-160
-------
TABLE 5-34. (continued)
State
Michigan
Minnesota
Missis-
sippi
Missouri
Montana
Nebraska
Nevada
New
Hampshire
New Jersey
AQCR
82a
122
123
124a
125
126
127a
128a
129a
130a
131
132
133
58
18a
134
135
70a
94a
137
138
139
140
141
142
143
144
85a
86a
145
146
13a
147
148
107a
121a
149
43a
45a
150
151
Number
of sites
15
21
14
6
1
3
9
25
1
15
2
1
2
1
3
1
6
4
1
9
5
3
4
1
1
1
2
1
Number of Sites Reporting 3
90th percent! le concentrations ipg/m )
< 100 100-200 200-365 > 365
15
19 2
12 2
5 1
1
3
9
25
1
13 1 1
2
1
2
1
3
1
6
1111
1
5 3 1
5
3
4
1
1
1
2
1
5-161
-------
TABLE 5-34. (continued)
State
New Mexico
New York
North
Carolina
North
Dakota
Ohio
Oklahoma
AQCR
12a
14a
152
153a
154
155
156
157
43a
158
159a
160
161
162
163
164
136a
165
166
167a
168
169
170
171
130a
172
79a
103a
124a
173
174
175
176
177
178
179a
180
181a
182
183
17a
184
185
186
187
188
189
Number
of sites
2
15
1
4
1
1
15
7
17
17
35
4
10
11
6
6
25
3
3
6
11
16
13
91
9
14
6
13
5
2
27
5
7
8
1
12
1
2
3
Number of Sites Reporting
90th percentile concentrations (uq/m )
< 100 100-200 200-365 > 365
2
14 1
1
4
1
1
14 1
7
12 5
13 4
26 8 1
4
7 3
11
6
6
25
3
3
6
10 1
16
13
55 31 5
9
14
5 1
9 4
4 1
2
5 21 1
5
5 2
8
1
12
1
2
3
5-162
-------
TABLE 5-34. (continued)
State
Oregon
Penn-
sylvania
Puerto
Rico
Rhode
Island
South
Carolina
South
Dakota
Tennessee
Texas
Utah
Vermont
AQCR
190
191
192a
193a
194
45a
195
196
197
244
120a
53a
58a
167a
198
199
200
201
202
203
204
87a
205
206
7a
18a
55a
207
208
209
22a
106a
153a
210
211
212
213
214
215
216
217
218
14a
219
220
159a
221
Number
of sites
1
1
2
10
1
7
9
4
9
7
2
15
3
6
3
2
46
29
1
4
3
12
2
14
20
55
8
6
4
14
2
Number of Sites Reporting 3
90th percentile concentrations (ug/m )
< 100 100-200 200-365 > 365
1
1
2
8 2
1
3 3 1
9
4
9
7
2
15
3
6
3
2
44 5
27 2
1
4
3
12
2
14
20
55
8
6
4
73 4
2
5-163
-------
TABLE 5-34. (continued)
State
Virginia
Washington
West
Virginia
Wisconsin
Wyoming
Guam
Virgin
Islands
AQCR
47:
207a
222
223
224
225
226
62
193a
227
228
229
230
103a
113a
179a
181a
232
233
234
235
236
128a
129a
237
238
239
240
241
242
243
246
247
Number
of sites
10
14
1
17
9
1
1
3
10
1
1
1
6
4
1
14
13
11
15
4
3
9
4
4
Number of Sites Reporting
90th percentile concentrations (uq/m)
< 100 100-200 200-365 > 365
10
11 3
1
15 2
8 1
1
1
2 1
9 1
1
1
1
6
4
1
13 1
11 2
10 1
15
4
3
9
4
4
5-164
-------
AQCR WITH AT LEAST ONE MONITOR HAVING SECOND
HIGHEST 24-HR OBSERVATION EXCEEDING 365 fig/m3
1 I AOCR WITH NO MONITORS HAVING SECOND HIGHEST
24-HR OBSERVATION EXCEEDING 365 pg/m3
ED NO DATA
Figure 5-53. Characterization of 1977 national sulfur dioxide status is shown by 24-hr average concen-
trations.
5-165
-------
improvement in most urban areas was more pronounced. Since 1973 the improvement
has again become slower. The 1977 EPA trends report states: "In most urban
areas, this is consistent with the switch in emphasis from attainment of
standards to maintenance of air quality; that is, the initial effort was to
reduce pollution to acceptable levels followed by efforts to maintain air
quality at these lower levels." From 1972 through 1977 annual averaged SCL
levels dropped by 17 percent, or an annual improvement rate of about 4 percent
per year. Figure 5-54 summarizes the annual averaged SC^ concentrations for
32 urban NASN stations for the years 1964-71. In this figure the first two
periods are apparent. In Figure 5-55 the national trends in annual average
S02 concentrations from 1972-1977 at 1333 sampling sites are displayed. The
symbols in Figure 5-55 are the same as those used in earlier figures on TSP
trends. The diamond symbolizes the composite annual average concentration;
the triangle is the median value.
Over the period of 1970-77 SOp emissions are reported to have decreased
only slightly (EPA, 1978). In 1970 the estimated annual manmade S02 emissions
were 29.8 million metric tons. By 1977 this was reduced only to 27.4 million
metric tons. The improvement in the ambient air quality levels for SOp reflects
displacement of sources from urban areas to rural areas, the restriction of
sulfur content of fuels used in low-level area sources, and the building of
newer sources with taller stacks.
5.2.3.7.2 Urban trends. The first air quality criteria document for sulfur
oxides, published in 1969, presented the frequency distributions for sulfur
dioxide levels in selected American cities for 1962-1967 (U.S. Department of
Health, Education, and Welfare, 1969). Improvements in sulfur dioxide levels
in each of the six cities can be demonstrated by comparing the available
5-166
-------
z
til
o
o
o
1972
Figure 5-54. Average sulfur dioxide concentrations are shown for 32
urban NASN stations.
Source: National Academy of Sciences (1975).
5-167
-------
56
48
*)
" 40
2*
0
K
CC
£ 32
LU
U
0
u
111
0 24
x
0
o
tr
D
"j 16
00
8
n
1
i
O
__
—
O
^MH
O
A
' T
0
MM
0
A
r
•
O
^
O
A
T
Ml
c
^M
<0
A
T
•
-
j)
^
O
A
M
\
O
—
J_ _
O
—
A
f "
1972
1973
1974 1975
YEAR
1976
1977
Figure 5-55. Nationwide trends in annual average sulfur dioxide concentrations from 1972 to 1977 are
shown for 1233 sampling sites.
Source: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards (1978).
5-168
-------
recent SOp data. The 1960's data came from the Continuous Air Monitoring
Project (CAMP), which operated continuous monitors in a few of the largest
U.S. cities. The cities are: Chicago, Philadelphia, St. Louis, Cincinnati,
Los Angeles, and San Francisco. In each city there is more than one continuous
monitor now operating. The station reporting the highest levels was used in
order not to overemphasize any improvement. The comparison is only an approxi-
mation because the locations of the monitors and the instrumental methods used
were not the same in the 1969 document (see Table 5-35).
In each city the peak concentration has decreased. In most cities the
peak is less than one-half of the earlier values. The only exception is St.
Louis, where the earlier peak was 0.72 ppm and the 1977 peak was 0.67 ppm. The
result is not unexpected in that the earlier summary is a composite frequency
distribution of 5 years of monitoring.
A better indicator of improved air quality would be the 50th, 70th, and
90th percentile concentration. In Chicago, the 50th percentile concentration
dropped from 0.08 ppm to 0.022 ppm; the primary annual standard is met but
there are still violations of the 24-hr S02 NAAQS. In Philadelphia the levels
have improved substantially; the 50th, 70th, and 90th percentile concentrations
are less than one-half the earlier values. Modest improvement is shown for
St. Louis and Cincinnati; their 50th percentile concentrations are lower now
than they were in the mid-19601s. The highest concentrations occur as frequently
in St. Louis as they did before, but in Cincinnati they occur less frequently.
Review of the St. Louis S02 data shows improved air quality for most of the
city. The high concentrations now reported are typical of only a smaller
section of the city. Los Angeles shows improvement in reducing the high
concentrations, but the 50th percentile concentration is actually slightly
5-169
-------
TABLE 5-35. COMPARISON OF FREQUENCY DISTRIBUTION OF.SO- CONCENTRATION (PPM)
DURING 1962-67a AND DURING 1977°. *
Frequency Distribution
City
Chicago
Philadelphia
St. Louis
Cincinnati
Los Angeles
San Francisco
Year
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
30
0.03
0.01
0.03
0.01
0.02
0.005
0.01
0.013
0.01
0.01
50
0.08
0.022
0.05
0.02
0.03
0.01
0.018
0.02
0.015
0.02
0.001
70
0.17
0.032
0.09
0.04
0.05
0.03
0.03
0.025
0.02
0.02
0.01
0.01
90
0.32
0.06
0.21
0.08
0.11
0.1
0.07
0.04
0.04
0.03
0.03
0.01
of S00 (ppm)
99
0.65
0.12
0.45
0.23
0.26
0.37
0.18
0.085
0.08
0.05
0.07
0.03
Maximum
0.95
0.25
0.85
0.44
0.72
0.67
0.53
0.29
0.25
0.09,
0.17
0.03
Concentrations from CAMP stations as reported in Air Quality Criteria for
Sulfur Oxides document, U.S. Dept. of Health, Education, and Welfare, NAPCA,
1969.
^Concentrations from NADB, U.S. Environmental Protection Agency, 1.977.
5-170
-------
higher than it was the previous decade. Similar, San Francisco, has trimmed
the peaks, but is has a very low median value.
5.2.3.8 Summary—The frequency of peak levels has been reduced in most urban
areas. The steady improvement of S02 ambient air quality has been slowed
somewhat in recent years. Currently, only 1 percent of the sulfur dioxide
monitoring sites show levels above the annual NAAQS. In 1974 the annual mean
sulfur dioxide standard was exceeded in 3 percent of the monitoring stations
(31 of 1030), compared with 16 percent in 1970. In 1977 and 1978, 2 percent
of the sites reported violations of the 24-hr standard. In 1974 this standard
was exceeded in 4.4 percent of the reporting stations (99 of 2241), compared
with 11 percent in 1970. Many of these sites reporting violations of the
24-hr standard are in remote areas near large point sources. A large percentage
of the U.S. population is exposed to ambient concentrations of sulfur dioxide
exceeding 100 mg/m some 10 percent of the time.
5-171
-------
5.3 REFERENCES
Ahmed, A. Karim. PCBs in the Environment. Environment, p. 6, March 1976.
Air Quality Criteria for Participate Matter, U.S. DHEW, Public Health Service,
Washington, DC, Pub. No. AP-49, 1969; and National Ambient Air Quality
Standards: 40 CRF §550.6.
Akland, G. G. Air quality for non-metallic inorganic loss 1971-1974 from the
National Air Surveillance Networks, Report No. EPA-600/4-77-003. RTP, NC:
U.S. EPA, 1977.
Altshuller, A. P. Characteristics of the chemical composition of the five
particulate fraction in the atmosphere. May 1976 (unpublished paper).
Alzona, J. , B. L. Cohen, H. Rudolph, H. N. Jow and J. 0. Frohliger. Indoor-
outdoor relationships for airbourne particulate matter of outdoor origin.
Atmospheric Environment 13: 55-60, 1979.
Amdur, M. 0. The physiological response of guinea pigs to atmospheric pollu-
tants. Institute of Air Pollution 1: 170, 1959.
Anderson, I. Relationships between outdoor and indoor air pollution. Atmos-
pheric Environment 6: 275-8, 172.
Appel, Bruce R., E. L. Kothny, Emmanuel M. Hoffer, Geroge M. Hidy, Jerome J.
Wesolowski. Sulfate and nitrate data from California Aerosol Characteri-
zation Experiment (ACHEX) ES&T 12:418-25, 1978.
Atmospheric Nitrosamine Assessment Report. Office of Air and Waste Management,
Draft, March 12, 1976.
Benson, F. B., J. J. Henderson, and D. E. Caldwell. Indoor-Outdoor Air Pollution
Relationships: A Literature Review. Publication No. AP-112. RTP, NC:
U.S. EPA, 1972.
Berdyev, Kh. B., N. V. Pavlovich, and A.A. Tuzhilina. Effect of motor vehicle
exhaust gases on atmospheric pollution in dwelling and in a main street.
Hygiene and Sanitation (Moscow) 32:424-426.
Biersteker, K., H. de Graaf, and C.A. G. Nass. Indoor air pollution in Rotterdam
homes. Intl. Journal of Air and Water Pollution 91232-50, 1965.
Binder, R. E., C.A. Mitchell, H. R. Hosein, and A. Bouhuip. Importance of the
indoor environment in air pollution exposure. Arch. Environmental Health
31:277, 1976.
f
Blifford, I. H. and G. 0. Meeker. A factor analysis model of large scale pollu-
tion. Atmospheric Environment 1:147-157, 1967.
Bradow, Ron. Emissions measurement and charterization division. Environmental
Sciences Research Laboratory, RTP, EPA, 1976.
Bretschreider, K. and J. Matz. Nitrosamines in the atmospheric air at places of
employment. Arch. Gesh. 13:1974 and Exc. Medica Sec. 46., Environmental
Health & Pollution Control. 6, 1974.
5-218
-------
Bridbord, K. et. al. Human exposure to polynuclear aromatic hydrocarbons. In:
Carcinogenesis: Vol. I, Polynuclear aromatic hydrocarbons: chemistry,
metabolism and carcinogenesis, ed. R. I. Freudenthal and P. W. Jones. New
York: Raven Press, 1976.
Brosette, C. (1976) Airborne particulates: Black and white episodes. Ambio 5:
157-163, 1976.
Cahill, T. Personal remarks. Committee on Particulate Control Technology,
National Research Council, 2nd Meeting, March 29-30, 1979, Los Angeles,
Cal i form'a.
Calvert, Jack G. Hydrocarbon involvement in photochemical smog formation in Los
Angeles atmosphere. ES&T 10:256, 1976.
"Cancer and the Environment," N. J. Dept. of Env. Prot. , May 25, 1976.
Carey, G. C. R., J. J. Phair, R. J. Shephard, and M. L. Thomson: The effects of
air pollution on human health. Amer. Ind. Hyg. Assn. J. 19:363-370, 1958.
Cavanugh, G. et. al. Potentially hazardous emissions from the extraction and
processing of coal and oil. Columbus, Ohio: Batelle Columbus Laboratories,
prepared for the U.S. EPA, 1975.
Charlson, R. J. , A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C. Ahlquist.
Sulfuric acid - ammonium sulfate aerosol: Optical detection in the St.
Louis region. Science 184: 156-158, 1974.
Charlson, Robert J., David S. Covert, Timothy V. Larson, and Alan P. Waggoner,
"Chemical properties of tropospheric sulfur aerosols," Atm. Env. 12:39-53,
1978.
Coburn, W. G. , R. B. Huzar, and J. D. Huzar. Continuous _i_n situ monitoring of
ambient particulate sulfur using flame photometry and thermal analysis.
Atm. Env. 12: 89-98, 1978.
Colome, S. D. and J. D. Spengler. Instrumental neutron activation analysis of
indoor/outdoor respirable aerosols. APCA, Cincinnati, Ohio, 72nd Annual
Meeting, June 1979, pp. 24-29.
Colucci, J. M. and C. R. Begeman. The automotive contribution to airborne
polynuclear aromatic hydrocarbons in air. Natl. Cancer Institute Monogr.
9:225, 1962.
Corn, M. Aerosols and the primary air pollution - nonviable particles: their
occurrence, properties, and effects, Ch. 3 in Stern, A. C. (ed.), Air
Pollution, v. I (3d ed.). New York: Academic Press, 1976.
Corn, M. Nonviable particles in the air. In: A. C. Stern, ed., Air Pollution.
v. I: Air pollution and its effects. (2nd ed.) New York: Academic Press,
1968, pp. 47-94.
Cote, W. A., W. A. Wade III, and J. E. Yocom: Final report on a study of indoor
air quality. EPA Contract 68-02-0745, TRC Project No. 32247. Wethersfield,
Conn.: Quality Assurance and Environmental Monit. Lab and Human Studies
Lab, U.S. EPA, 1974.
5-219
-------
Cox, R. A. and S. A. Perikett. Effect of relative humidity on the disappearance
of ozone and sulphur dioxide in contained systems. Atmospheric Environment
6:365-68, 1972.
Cunningham, P. T. and S. A. Johnson (1976). Spectroscopic observation of acid
sulfate in atmospheric particulate samples. Science 191:77-79, 1976.
De Maio, L. and M. Corn. Polynuclear aromatic hydrocarbons associated with
particulates in Pittsburgh air. JAPCA 16:67, 1966.
Dockery, Douglas W. Personal exposure to respirable particulates and sulfates:
measurement and prediction. Doctoral Thesis, Harvard University, School of
Public Health, May 1975.
Dzubay, T. G. Chemical-elemental balance method applied to dichotomous samples
data. Symposium, Annals of the N.Y. Acad. of Sci., Jan. 9-12, 1979.
Dzubay, T. G. and R. K. Stevens. Ambient air analysis with dichotomous sampler
and x-ray fluorescence spectrometer. Envir. Sci. Tech. 9:663-668, 1975.
Dzubay, T. G., R. K. Stevens, and C. M. Peterson. Application of the dichotomous
sampler to the characterization of ambient aerosols. In T. G. Dzubay, Ed.
X-Ray Fluorescence Analysis of Environmental Samples. Ann Arbor: Ann
Arbor Science Publishers, Inc., 1977, pp. 95-106.
Economic Commission for Europe. Working Party on Air Pollution Problems. Task
Force on Fine Particulate Pollution. Report. Geneva, WHO, December 1977.
Gartrell, G., Jr. and S. K. Friedlander. Relating particulate pollution to
sources: the 1972 California Aerosol Characterization Study. Atmospheric
Environment 9:279-299, 1975.
Gate, D. F- Relative contributions of different sources of urban aerosols:
aplications of a new estimation method to multiple sources in Chicago.
Atmospheric Environment 9:1-18, 1975.
Goldwater, L. J., A. Manoharan, and M. B. Jacobs: Suspended particulate matter,
dust in "domestic" atmospheres. Arch. Envir. Health 2: 511-515, 1961.
Gordon, Robert J. Distribution of Airborne Polycyclic Aromatic Hydrocarbons
throughout Los Angeles. ES&T 10:370, 1976.
Haagen-Smib, A. J. et. al. The State of California plan for achieving and
maintaining the National Ambient Air Quality Standards, January 30, 1972.
Hardy, K. A., R. A. Kselsson, J. W. Nelson, J. W. Winchester. Elemental con-
stituents of Miami aerosol as function of particle size. ES&T 10:176-182,
1976.
Hauser, Thomas R. and John N. Pattison. Analysis of aliphatic fraction of air
particulate matter. ES&T 6:549, 1972.
Heidelberger, Charles. Chemical carcinogenesis. Ann. Rev. Biochem. 44:79, 1975.
5-220
-------
Helms, G. T. , J. H. Southerland, K. R. Woodward, J. V. Hinda?, D. H. Coventry,
and C.D. Robson. Chattanooga, Tennessee -- Rossville, Georgia. Interstate
Air Quality Study, 1967-68. National Air Pollution Control Administration
Publication No. APTD-0583. Washington, DC:GPO, 1970.
Hetling, L. , E. Horn and J. Tofflemire. Summary of Hudson River PCG Study
Results. NY State Department of Environmental Conservation, Bureau of
Water Research, April 1978.
Hidy, G. M., B. Appel, R. J. Charlson, W. E. Clark, D. Covert, D. Dockweiler, S.
K. Friedlander, P. Friedman, R. Giauque, S. Green, S. Heisler, W. W. Ho, J.
J. Hunt-Ficker, R. B. Husar, K. Kubler, G. Lauer, P. K. Mueller, T. Novakov,
R. Ragaini, L. W. Richards, T. B. Smith, E. R. Stephens, G. Sverdrup, S.
Twiss, A. Waggoner, S. Wall, H. H. Wang, J. J. Wesolowski, K. T. Whitby,
and W. White (1974). Characterization of Aerosols in California (ACHEX).
Final Report. Vols. I-IV. Submitted to the Air Resources Board, State of
California, Sept. 30, 1974.
Hidy, G. M., B. R. Appel, R. J. Charlson, W. E. Clark, S. K. Friedlander, D. H.
Hutchison, T. B. Smith, J. Suder, J. J. Wesolowski, and K. T. Whitby (1975).
Summary of the California aerosol characterization experiment. J. APCA
25:1106-1114.
Hidy, G. M. , P. K. Mueller, V. Dio, and K. C. Detor. Study and implementation of
the Sulfate Regional Experiment (SURE). Proc., 4th Symposium on Turbulence
and Diffusion and Air Pollution, AMS, Jan. 1979.
Hidy, G. M., P. K. Mueller, and E. Y. Tong. Design of Sulfate Regional Experi-
ment (SURE), v. 1 & 2, Electric Power Research Institute, Report EC-125,
Palo Alto, California, 1976.
Hidy, G. M., P. K. Mueller & E. Y. Tong. Spatial and temporal distributions of
airborne sulfates in parts of the United States. Atm. Env. 12:735-752,
1978.
Hitchcock, D. Atmospheric sulfates from biological sources. JAPCA 26: 210-
215, 1976.
Hoegg, U. R. Cigarette smoke in closed spaces. Envir. Health Persp. 117-128,
1972.
Hollowell, C. D. , R. J. Beednitz, and G. M. Traynor. Combustion generated
indoor aic^pollution. 4th International Clean Air Congress, Tokyo. May
16-vu, iy//.
Hopke, P. K. et. al. The use of multivariate analysis to identify sources of
selected elements in the Boston urban aersol. Atmospheric Environment
10:1015-1025, 1976.
Husar, R. B., W. E. Wilson, Jr., and G. J. D'Alessia. Sulfate: Alternative
means of reducing their atmospheric concentrations. AICLE. 85th National
Meeting, Philadelphia. June 4-8, 1978.
5-221
-------
Inter-Agency Energy-Environment Research and Development Program. Survey of
Indoor Air Quality Health Criteria and Standards. Publication #EPA-600/7-
78-027. Research Triangle Park: Office of Research and Development, U.S.
Environmental Protection Agency, March 1978.
Ishido, S. Variations in indoor and outdoor dust intensities. Bull. Dept.
Home Econ., Osaka City University. 6:53-59, 1959.
Ishido, S. , K. Kamada, and Nakagawa: Free dust particles and airborne
microflora. Bull. Dept. Home Econ. , Osaka City Univ. 21: 31-37, 1956.
Jacobs, M. B., L. J. Goldwater, and A. Fergany: Comparison of suspended parti-
culate matter of indoor and outdoor air. Intl. J. Air Water Pollution
6:377-380, 1962.
Junge, C. Die Konstitution der atmospharischen aerosole. Ann. Meteorol. 5
(Beiheft) 1-55, 1952.
Kadawaki, Atoshi. Size distributions and chemical composition of atmospheric
particulate nitrate in the Ngoya area. Atm. Env. 11: 671-5, 1977.
Kerteszhe-Saringer, M. et. al. On the size distribution of B(a)P containing
particles in urban air. Atm. Env. 5:429, 1971.
Ketchum, Brian T.. et. al. Final report to the National Highway Traffic Adminis-
tration, Non-passenger Automobile Average Fuel Economy Standards Model
Years 1980-81. (Docket No. FE. 77-05, Notice 1) by Citizens for Clean Air,
In. and Environmental and Resources Technology, February 7, 1978.
Ketchum, Brian T. et. al. Final Report to the National Highway Traffic Adminis-
tration. U.S. Dept. of Transportation, 1981-84. Passenger Automobile
Average Fuel Economy Standards (Docket No, FE-76-1), April 11, 1977.
Kito, T. Survey of air pollution on outlying islands. Kuki Seijo 15(2): 10-
21, 1977.
Kleinman, M. T. The use of long term and seasonal trends of elemental compounds
as an aid to the identification of sources of airborne pollutants.
Conference on Aerosols, N.Y. Acad. of Sciences, January 1979.
Kotin, P. and H. L. Falk. Atmospheric factors in pathogenesis of lung cancer.
Adv. in Cancer Res. 7:475, 1963.
Kowalcyzk, G. S., C. E. Choquette, and G. E. Gordon. Chemical element balances
and identification of air pollution sources in Washington, D.C. College
Park , MD: Dept. of Chemistry, University of Maryland, 1979 (draft).
Krost, Ken. Atmospheric Chemistry and Physics Division, Environmental Sciences
Research Lab. Personal commun. RTF:EPA, 1977.
Kurosaka, D. Sulfate concentrations in the South Coast Air Basin. Report DTS-
76-1, Calif. Air Resources Board, Sacramento, Calif. 1976.
5-222
-------
LaBelle, C. W. , J. A. Long, and E. E. Christofano. Synergistic effects of
aerosols: participates as carriers of toxic vapors. Arch. Ind. Health II;
197, 1955.
Larsen, Ralph I. A mathematical model for relating air quality measurements to
air quality standards. EPA Publ. AP-89. RTP, NC: EPA, 1971.
Lavery, T. F. , G. M. Hidy, R. L. Basket, and J. Thrasher. Occurrence of long-
range transport of sulfur oxides in the northeastern United States. Proc. ,
4th Symposium on Turbulence Diffusion and Air Pollution AMS, Reno, Nevada,
Jan. 15-18, 1979.
Lee, R. E., Jr. The size of suspended particulate matter in air. Science
178:567-575. 1972.
Lee, R. E., Jr., S. S. Joranson, R. E. Enrione, and G. B. Morgan. National Air
Surveillence cascade impactor network. II. Size distribution measurements
of trace metal components. Environ. Sci. Tech. 6:1025-1030 1972.
Lee, R. E. , Jr. , R. K. Patterson, and J. Wagman. Particle-size distribution of
metal components in urban air. Environ. Sci. Tech. 2:288-290 1968.
Lefcoe, N. M. and I. I. Inculet: Particulates in domestic premises: I. Ambient
levels and central air filtration. Arch. Envir. Health 22:230-238, 1971.
Lefcoe, N. M. and I. I. Inculet: Particulates in domestic premises: II. Ambient
levels and indoor-outdoor relationships. Arch. Environ. Health 30:575,
1975.
Lijinsky, William and Samuel S. Epstein. Nitrosamines as environmental carcinogens.
Nature 225:21, 1970.
Ludwig, J. H. , G. B. Morgan, and T. B. McMullen. Trends in urban air quality.
In W. H. Matthew, W. W. Kellogg, and G. 0. Robinson, Eds. Man's Impact
on Climate. Cambridge: MIT Press, 1971, pp. 321-338.
Lynn, D. A., G. L. Deane, R. C. Galkiewicz, and R. M. Bradway. National Assessment
of the Urban Particulate Problem. Vol I. RTP, NC: U.S. EPA Office of Air
Quality Planning and Standards. Publ. #EPA-450/3-76-024, July 1976.
Lynn, D. A., B. S. Epstein, and C. K. Wilcox. Analysis of New York City Sulfate
Data. Final report EPA Contr. 68-02-1337, Task 10, GCA Corp., Bedford, MA,
1975.
Lyons, W. A. and J. C. Dooley, Jr. Satellite detection of long range pollution
transport and sulfate aerosol haze. Atmos. Envt. 12:621-631, 1978.
MITRE Technical Report Series, EPA Contract No. 68-02-1495, Project No. 077B,
Dept. W-54, prepared by the MITRE Corp., McLean, VA, 1976.
MSA Research Corporation. Hydrocarbon Pollutant Systems Study: I. Stationary
Sources, Effects and Control. Final Report. MLA Research Corp., 1972.
Mage, D. T. and W. R. Ott. Refinements of the lognormal probability model for
analysis of aerometric data. JAPCA 28:796-7 1978.
5-223
-------
Maunsell, K. Concentration of airborne spores in dwellings under normal con-
ditions and under repair. Intl. Arch. Allergy 5:373-376, 1954.
Megaw, W. S. The penetration of iodine into buildings. Intl. J. Air Pollution
5:121-128, 1962.
Miller, F. J. et. al. Particle size considerations for establishing a standard
for inhaled particles. RTP, NC: U.S. EPA, Office of Air Quality Planning
and Standards, January 1979 (draft).
Moschandreas, D. J., J. W. Winchester, J. W. Nelson, and R. M. Burton. Fine
particle residential indoor air pollution. Atmospheric Environment (in
press).
Mueller, P. K., R. W. Mosley, and L. B. Pierce. Chemical composition of Pasadena
aerosol by particle size and time of day. J. Colloid and Interface
Science 39:235, 1972.
Mueller, P. K. et al. Concentration of fine particulate and lead in car exhaust.
Symposium on Air Pollution Measurement Methods, Philadelphia. American
Society for Testing and Materials. Special technical publication, 1964.
Mueller, P. K., G. M. Hidy, T. F. Lavery, K. Warren, R. L. Bassett. Some early
results from the Sulfate Regional Experiment (SURE). Proc. 4th Symposium
on Turbulence Diffusion and Air Pollution, AMS, Reno, Nevada, Jan. 15-18,
1979, pp. 322-329.
Ragda, Niren, John R. Ward, Benjamin J. Mason and Leon S. Pocinki. Retrospec-
tive modeling of ambient concentrations of Benzo(a)pyrene. Pres., 71st
Annual Meeting. Air Pollution Control Association, Houston, June 25-30,
1978.
National Academy of Sciences. Air Quality and Stationary Source Emissions
Control. Ser. #94-4. March 1975.
National Academy of Sciences. Report by the Commission on Natural Resources.
Air Quality and Stationary Source Emission Controls, March 1975.
National Academy of Sciences. Vapor-Phase Organic Pollutants. Washington, DC;
NAS, 1976.
National Academy of Sciences. Committee on Biologic Effects of Atmospheric
Pollutants. Particulate Polycyclic Organic Matter. Washington, D.C., NAS,
1972.
National Research Council. Airborne Particles, EPA-600/1-77-053. RTP, NC:
U.S. EPA, 1977.
National Research Council. Nitrates: An Environmental Assessment. Washington,
D.C.: National Academy of Sciences, 1978.
National Research Council. Nitrogen Oxides. Washington, D.C.: National
Academy of Sciences, 1977.
5-224
-------
National Research Council, Assembly on Life Sciences, Division of Medical
Sciences, Committee on Medical and Biologic Effects of Environmental
Pollutants, Subcommittee on Airborne Particles. Airborne Particles.
Baltimore: University Park Press, 1979.
Natusch, David F. S. and John R. Wallace. Urban aerosol toxicity: the influence
of particle size. Science 186:695, 1974.
Nitrosamines found in commercial pesticides. Chem & Eng News, p. 33, Sept. 20,
1976.
Oreel, A. and J. Seinfeld. Nitrate formation in atmospheric aerosols. Environ-
mental Sci. & Tech. 11:1000-7, 1977.
Pace, T. G. U.S. EPA, personal communication. April 1979.
Pace, T. G. and E. L. Meyer, Jr. Preliminary characterization of inhalable
particulate in urban areas. Proceedings, 79th Annual Meeting, Air Pollution
Control Assn., Cincinnati, Ohio, June 1979.
Parvis, D. Condensation nuclei in the air of artificially heated environments.
Annali del la Sanita Publica 13:1569-1581, 1952.
Pellizzari, Ed. Development of Analytical Techniques for Measuring Ambient
Carcinogen Vapors. RTP:Research Triangle Institute, EPA, 1976.
Perera, F. P. and A. K. Ahmed. Respirable particulates: impact of airborne
fine particulates on health and the environment. New York: Natural
Resources Defense Council, Inc., 1978.
Perhack, Ralph M. Sulfate regional experiment in northeastern United States:
The "SURE" program. Atm. Env. 12:641-647, 1978.
Preussman, R. Formation of carcinogens from precursors occurring in the environ-
ment: new aspects of nitrosamine induced tumorigenesis. Rec. Results
Cancer Res. 44:9, 1974.
Record, Frank M. , R. M. Brodway, and K. W. Wiltsee, Jr. A study of ambient
particulates in Massachusetts. Proc. Northeast Atlantic Internatl.
Section, APCA, Fall Technical Meeting, Windsor, Conn., Oct. 18, 1978.
Romagnol, G. Studies on the climatic conditions in some elementary classrooms
of Novara. Italian Rev, of Hyg. 21:410-419. 1961.
Organic Vapors Identified in Ambient Air. RTI Quartery Rept. No. 2, Table 9,
1976. And KIN-BUC Report, EPA, 1976.
Sawicki, E. The chemical composition and potential genotoxic aspects of polluted
atmospheres. Workshop for the Investigation on the Carcinogenic Burden by
Air Pollution in Man. Hanover, Germany. Oct. 22-24, 1975.
Sawicki, E. Dimethylnitrosamine in the Baltimore and Belle, West Virginia
Environments. Proceedings, Meeting. Environmental Research Center,
Environmental Protection Agency, RTP, Jan. 6, 1967.
5-225
-------
Sawicki, Eugene. Atmospheric Chem. and Phys. Div. ESRL, RTF, EPA. Personal
communication, April 1976.
Sawicki, E. et. al. Benzo(a)pyrene content of the air of American communities.
Amer. Ind. Hyg. Assn. J. 21:443, 1960.
Sawicki, E. et. al. Polynuclear aromatic hydrocarbons: Composition of the
atmosphere in some large American cities. Am. Ind. Hyg. Assn. J. 23: 137,
1962.
Sawicki, E. et. al. Quantitative composition of the urban atmosphere in terms
of polynuclear aza heterocyclic compounds and aliphatic and polynuclear
aromatic hydrocarbons. Int. J. Air & Water Poll. 9:515, 1965.
Scanlan, Richard. N-Nitrosamines in Foods. CRC Critical Reviews in Food
Technology, April, 1975, p. 357.
Schaefer, V. S. , V. A. Mohnen, and V. R. Veirs: Air quality of American homes.
Science 175:133-175, 1972.
Scott Research Laboratory. 1970 Atmospheric Reaction Studies in the Los Angeles
Basin, Vols. 1, 2, 4. Plumsteadville, Pa.: Scott Research Lab., 1970.
Seisaburo, S., K. Kiyoko, and N. Tatsuko: Free dust particles and airborne
microflora. Bull. Dept. Home Econ. , Osaka City Univ. 4:31-37, 1959.
Selected Environmental Carcinogens. Draft #1. New Jersey Dept. of Env. Prot.,
Oct. 1976.
Shapley, Deborah. Nitrosamines: scientists on the trail of prime subject in
urban cancer. Science 191:235, 1976.
Shepherd, R. J., G. C. R. Carey and J. J. Phair: Critical evaluation of a
filter-strip smoke sampler used in domestic premises. AMA Arch. Ind.
Health 17:236-252, 1958.
Spedding, D. J. Sorption of sulphur dioxide by indoor suraces -- II. Wood.
J. Appl. Chem. 20:226-28, 1970.
Spedding, D. J. and R. P. Rowlands. Sorption of sulphur dioxide by indoor
surfaces -- I. Wallpapers. J. Appl. Chem. 20:143-46, 1970.
Spedding, D. J. , R. P. Rowlands, and J. E. Taylor. Sorption of sulphur dioxide
by indoor surfaces — III. Leather. J. Appl. Chem. 21:68-70, 1971.
Spengler, J. Analysis of preliminary dichotomous sampling data, Harvard School
of Public Health, Unpublished, 1979.
Spengler, John D. Preliminary analysis of dichotomous and TSP measurements in
Topeka, Kansas and Kingston, Tennessee.
Spengler, J. D. and D. W. Dockery. Long term measurements of respirable sulfates
and particles inside and outside homes. Electro 79 Professional Program
Paper 2413, New York City, April 25, 1979.
5-226
-------
Spicer, C. W. The fate of nitrogen oxides in the atmosphere. Ecological
Research Series, Office of Research and Development Report No. EPA-60013-
76-030. RTP, NC: U.S. EPA, 1976.
Sproul, Wayne T., Air Pollution and its Controls. NY: Exposition Press, 1970.
Steigerwald, B. J. Participate control strategy. Office of Air Quality Plan-
ning and Standards, EPA, memo to Roger Strelow, Asst. Administrator for Air
and Waste Management, July 2, 1975; submitted by EPA to the Subcommittee,
House Committee on Interstate and Foreign Commerce.
Sterling, T. D. and D. M. Kobayashi. Exposure to pollutants in enclosed "living
spaces." Envir. Res. 13:1-35, 1977.
Stevens, R. , National Environmental Research Laboratory, U.S. EPA, Personal
communication. May 1979.
Stevens, Robert K. and Thomas G. Dzubay. Sampling and analysis of atmospheric
sulfate and related species. Atmospheric Envt. 12:55-68, 1978.
Summary Report on Nitrosamines (draft), RTP: Health Effects Research Lab., EPA,
1976.
Sverdrup, G. M., K. T. Whitby, and W. E. Clark: Characterization of California
aerosols -- II. Aerosol size distribution measurements in the Mojave
Desert. Atmos. Enviroment 9:483-494, 1975.
Terabe, M. Air pollution in Japan: present state area trends. Kuki Seizo
15(2):1-10, 1977.
Thomas, J. F., M. Mujai, and B. D. Tebbens. Fate of airborne benzo(a)pyrene.
Envir. Sci. Tech. 2:33, 1968.
Thompson, C. R. , E. G. Hensel, and G. K. Kats: Outdoor-Indoor levels of six air
pollutants. JAPCA 23:881-6, 1973.
22 High Volume Industrial Chemicals (HIVOC). Preliminary Assessment for Air
Pollution Potential, U.S. EPA, Office of Air Quality Planning and Standards,
1976.
U.S. Dept. of Health, Education & Welfare Public Health Service. Natl. Air
Pollution Control Administration. Air Quality Criteria for Sulfur Oxides.
Publication No. AP-50. Washington, D.C.: HEW, 1969.
U.S. Environmental Protection Agency. National Air Quality, Monitoring, and
Emissions Trends Report, 1977. EPA-450/21-78-052, December 1978.
U.S. Environmental Protection Agency. National Assessment of the Urban Parti-
culate Problem. Vol. I. National Assessment (EPA-450/3-76-024). July
1976.
U.S. Environmental Protection Agency. National primary and secondary ambient
air quality standards. Federal Register 36(84):8186-8201, 1971.
5-227
-------
U.S. Environmental Protection Agency. Report to Congress on the National
Participate Problem. Feb. 1979. Noted to be cited or quoted.
U.S. Environmental Protection Agency. Environmental Criteria and Assessment
Office. Health Assessment Document for Polycyclic Organic Matter. RTF,
NC:U.S. EPA, 1978. External review draft #1.
U.S. Environmental Protection Agency. Environmental Criteria and Assessment
Office. Health effects of short-term exposure to nitrogen dioxide. EPA-
600/8-78-009. RTP, NC: EPA, 1978.
U.S. Environmental Protection Agency. Office of Air Quality Planning and Standards.
Air Quality Data. 1977 Annual Statistics. EPA 450/2-78-040. RTP, Sept.,
1978.
U.S. Environmental Protection Agency. Office of Air Quality Planning and Standards,
National Air Quality, Monitoring and Emissions Trends Report, 1977. RTP,
NC. Publ. 0EPA-450/2-78-052. Dec. 1978.
U.S. Environmental Protection Agency. Office of Air Quality Planning and Standards.
Optimum Site Exposure Criteria for S0? monitoring. Publ. #EPA-450/3-
77/013. RTP, NC, U.S. EPA, April 1977.
U.S. Environmental Protection Agency. Office of Air Quality Planning and
Standards. Preferred Standards Path Report for Polycyclic Organic Matter.
Durham, NC: EPA, 1974.
Walker et al. Environmental Assessment of Atmospheric Nitrosamines. EPA
Contract No. 68-02-1495, Feb. 1976.
Walsh, M., A. Black, and A. Morgan. Sorption of SO- by typical indoor sources
including wool carpets, wallpaper, and paint. Atmospheric Environment,
received to be published February 24, 1977.
Warren Springs Laboratory: British Standard 1747, Part 2. Stevenage, G. B.,
1964.
Weatherly, M. L. Air pollution inside the home in symposium on Plume Behavior.
Intl. J. Air-Water Poll. 10:404-09, 1966.
Whelpedale, D. M. Large scale atmospheric sulfur studies in Canada. Atm. Eriv.
12:661-670, 1978.
Whitby, K. T. , A. B. Algren, R. C. Jordan, and J. C. Anm's: The ASHAE airborne
dust survey. Heating, Piping and Air-Cond. 29(11): 185-192, 1957.
Wigby, T. K. and B. Cantrell. Atmospheric Aerosols — Characteristics and
Measurements. Intl. Conference on Environmental Sensing and Assessment
(ICESA) (75CH1004-1), 1975.
Wilson, M. L. G. Indoor air pollution. Proc. Royal Soc. , Ser. A. (London)
307:215-221, 1968.
5-228
-------
Winchester, J. W. and G. D. Nifong. Water pollution in Lake Michigan by trace
elements from pollution aerosol fallout. Water, Air and Soil Pollution
1:50-64, 1971.
Yocom, J. R.
ships.
, W. L. Clink, and W. A.
JAPCA 21:251-259, 1971.
Cote: Indoor/Outdoor air quality relation-
5-229
-------
6. TRANSMISSION THROUGH THE ATMOSPHERE
6.1 INTRODUCTION
Relating ambient concentrations of SO and particles to source emission
rates is a central problem for the development of control measures. Processes
occurring in the atmosphere play a dominant part in determining source-receptor
relationships. Until the early 1970's, the atmosphere was regarded primarily as
a medium that dispersed primary emissions, thereby reducing ambient concentrations
downwind of individual sources by dilution. During this period, emission rates
from major point sources were increasing; at the same time, however, stack
heights were increasing, which improved the local ambient air quality because of
increased atmospheric dilution.
Since the early 1970's, it has become increasingly apparent that the role
of the atmosphere in the chemical transformation and removal of SO and particles
A
is also important. Many harmful environmental effects are now attributed to the
reaction products of SO NO , and hydrocarbons that are distributed and deposited
J\ A
by the atmosphere. To quantify these effects and their causes, it is necessary
to understand many processes that occur in the atmosphere. Among these processes
are physical and chemical transformations (section 6.2), dry and wet removal
processes (section 6.3), and dispersion (section 6.4).
The process by which a pollutant is carred past a fixed point or across
a fixed plane is called transport; this process is carried out by horizontal
wind. The process by which a pollutant is mixed in the atmosphere is
called diffusion; this process is caused by horizontal and vertical motions
resulting from gradients of pressure and temperature and from surface roughness
features. In air pollution meterology, it is common to use the term "dispersion"
to describe transport and diffusion together. The term "transmission" is used
the joint effects of transport, diffusion, transformation, and deposition.
6-1
-------
Listed below are the relevant factors that are discussed and substantiated
in this chapter.
(1) Atmospheric particle mass is distributed bimodally with respect to
particle size. Particles smaller than 2 ym in diameter are referred to
as fine particles; particles 2 ym and larger are often referred to as coarse
particles.
(2) Comparable masses of fine and coarse particles are measured within
urban areas. With increasing distance downwind from urban areas, the fine-
particle mass tends to exceed the coarse-particle mass, except in areas in-
fluenced by wind or vehicle generated fugitive dust from soil.
(3) Much of the fine-particle mass is secondary, i.e. formed in the
atmosphere by the transformation of gases, mainly sulfur dioxide, nitrogen
oxides, and organics containing more than six carbon atoms. The remainder of
the fine-particle mass is due to primary emissions and consists mostly of
combustion-derived sulfates and carbonaceous particles, lead aerosols, small
amounts of other metal-containing particles from combustion and industrial
processes, and small amounts of finely divided mineral dusts.
(4) The coarse particles are mostly primary and consist of substances
emitted directly from industrial sources and from combustion processes, along
with substantial amounts of suspended or resuspended dusts of various types.
Some nitrate and sulfate is found in coarse particles.
(5) Because removal by deposition may take several days, fine particles
may be transported for distances of 1000 km or more from their origin or the
origin of precursor gases. Secondary fine particles are formed continuously by
atmospheric reactions and are likely to undergo long-range transport.
6-2
-------
(6) Urban-area sources and large-point sources contribute precursors of
aerosols formed during atmospheric reactions; the contributions of these many
sources are therefore superimposed downwind.
(7) Gas-to-particle conversion processes depend on a variety of environmental
factors, such as solar radiation, concentrations of oxidizing radicals, and
humidity. Precursor-gas emission rates alone are therefore of very limited
usefulness in estimating the mass of secondary fine particles.
Because of these complexities, it is essential that several very different
approaches be used to establish source-receptor relationships (section 6.5).
The annual U.S. emissions of SO ranged from 20 to 23 x 10 tons per year
between 1940 and 1960 (chapter 4). Until 1950, SO emissions from industrial
^
sources were the largest; near-ground emissions from homes, commercial establish-
ments, and locomotives were next; emissions from electric power plants were
least. Between 1950 and 1960, SO emissions from homes, commercial establish-
ments, and locomotives decreased rapidly and those from industrial sources
decreased somewhat, whereas those from electric power plants increased rapidly.
By the 1960's, SO emissions from electric power plants exceeded emissions from
/\
any other class of sources. The increase in SO emissions from electric power
plants continued into the 1970's; emissions from other classes of sources continued
to decline or level off at much lower levels than in earlier years.
Sulfur dioxide concentrations in urban areas declined substantially during
the 1970's. This decline reflects shifts in fuel usage toward lower sulfur fuels
for most emission sources. In several urban areas, power plants also burned
fules with lower sulfur content. The trends in sulfur dioxide concentrations
reflect these changes. During the same period, urban sulfate levels decreased
in the winter months, particularly in the Northeast, reflecting SO emission
^
trends. Sulfates increased during the wanner months in both urban and nonurban
6-3
-------
areas into the 1970's, with some subsequent declines later in the decade in some
regions. The differences among seasonal trends obscure the effect of SO
A
control on sulfate concentrations computed on an annual average basis. The high
sulfate concentrations measured during the summer months in recent years are
best explained by long-range transport and the transformation of SO to sulfates.
A
Historical trends during the winter months can be accounted for by shifts to
lower sulfur fuels.
The historical data base for particulate nitrate is in doubt because of
sampling difficulties. Positive artifacts on basic filter material due to absorp-
tion of nitric acid vapor and negative artifacts due to volitilization of
NH4N03 vapor make the existing data base unreliable until these problems have
been defined quantitatively.
A downward trend in benzene-soluble organics was observed at most urban and
nonurban sites during the 1960's. No trend data on carbon-containing aerosols
are available for the 1970's. Nevertheless, carbon-containing aerosols still
seem to be the second most abundant particulate species in the Eastern United
States. At some western urban sites, carbon-containing aerosols exceed sulfates
in concentration because of low SO emissions.
/\
The preceding discussion indicates the complexity in developing source-
receptor relationships for the SO /aerosol/precursor complex.
X
6.2 TRANSFORMATIONS
6.2.1 Transformation of SO^
The chemical transformation of S02 in the atmosphere generally leads to the
formation of sulfuric acid or other sulfates, although the formation of metal
sulfides and organic sulfite complexes has been reported (Eatough et al., 1978).
The oxidation of S0 to sulfuric acid is of grave concern because it leads to
6-4
-------
the acidification of aerosols, fogs, and rain (acidification of rain is treated
separately in Chapter 8). However, the formation of these sulfates, commonly called
secondary sulfates," is responsible for large quantities of submicron aerosols
in the troposphere.
6,2.1.1 Atmospheric Investigations--Investigations of the plumes of coal-
burning power plants have suggested pseudo-first-order conversion rates in the
range of 0 to 8 percent per hour. However, Wilson (1978) critically reviewed
most of these studies reported before 1978 and concluded that they do not
provide reliable information on the rate of sulfate formation. In general, the
investigators used insensitive instrumentation, improper sampling techniques, or
inappropriate flight patterns. In 1978, more advanced studies were reported at
the International Symposium on Sulfur in the Atmosphere. For example, Husar et
al. (1978) reported data (see Figure 6-1) for the rate of sulfate formation as a
function of solar radiation for the Labadie power plant plume. As can be seen
in Figure 6-1, the conversion rate shows a strong correlation with solar radiation,
which suggests that photochemical pathways are important. However, Gillani et
al. (1978) also noticed that the rate was significantly affected by the background
concentration of ozone and by the degree of mixing. For example, as shown in
Figure 6-2, the Labadie plume received approximately the same solar radiation
intensity on July 9 and 18. However, on July 18 the background 03 concentration
and mixing were greater; on that day the conversion rate was about twice that
on July 9.
It is likely that the inability to replicate experiments as discussed above
and the inability of various research laboratories to agree on experimental
conversion rates are due to complex parameters such as mixing, solar radiation
intensity, and background concentration of reactive species (Wilson, 1979).
it should be expected that other parameters, including temperature,
6-5
-------
<
oc
z
o
tt 3
a.
O
o
a 2
LL
_)
D
C/3
JULY 9
D
a
•
JULY 18
O
o
•
PLUME AGE
2-5 HR
5-8 HR
> 8 HR
SOLAR RADIATION
/ c/ T^. o 7
/ / \ \
« — / » _ »
8 10 12 14 16 18 20 22 24
Figure 6-1. The daytime conversion rate in the MISTT study was
quite variable, ranging from 1 to 4 percent per hour, whereas night-
time values were consistently below 0.5 percent per hour. Either
photochemical conversion or liquid-phase oxidation in daytime
cumulus clouds is consistent with the daytime peak of conversion
rate.
Source: Husar et al. (1978).
6-6
-------
20
16
(A
oc
£ 12
w
D
u
p
cc
a.
oc
u.
JULY ft,
D
a
•
JULY 11
0
a
•
PLUME AGE
2-5 HR
64 HR
> t HR
JULY 9
1.00
2.00
3.00
4.00
6.00
6.00
TOTAL SOLAR RADIATION DOSE DR kW-hr/m3
Figure 6-2. The ratio of participate sulfur, Sp, to
total sulfur concentrating tot, (consisting of partic-
ulate S in sulfates plus gaseous S in SCL, average
for traverse) is shown as a function of total plume
solar radiation dose, DR, (time integral of solar-
radiation, R,., between plume release time and plume
sampling time). The least squares regression lines
shown have the following equations:
July 9: y = 2.05x - 0.89, r = 0.92;
July 18: y = 3.22x + 0.75, r = 0.96.
Source: Gillani et al. (1978).
6-7
-------
relative humidity, and source operational parameters, will influence the rate of
sulfate formation. These recent plume studies have established that atmospheric
conversion rates of sulfate formation are in the range of 0 to 8 percent per
hour (Husar et al., 1978). However, at this time, field observations of plumes
have not been completely described in terms of the fundamental processes,
including the chemical pathways, atmospheric mixing, and removal at the earth's
surface. A model including variable first-order conversion, mixing, and removal
has been presented (Gillani et al., 1978). Only data for the summertime are
available at present.
There have been a limited number of investigations of SOp conversion rates
in urban areas. Measurements in the Los Angeles area during the summer smog
season suggested an SO^ oxidation rate in the range of 1 to 13 percent per hour
for the daylight hours (Roberts and Friedlander, 1976). Measurements using St.
Louis air in a Teflon bag give maximum SOp oxidation rates of 1.6 to 5.5 percent
per hour. These values, and the increase in rate with increasing ratio of non-
methane hydrocarbons (NMHC) to nitrogen oxides, agree with the results of smog
chamber studies at higher pollutant concentrations (Miller, 1978). For the St.
Louis urban plume, White et al. (1976) measured the mass flow rate of ozone and
particulate sulfur. Their measurements at discrete distances suggested that the
overall sulfate formation rates varied from 1 to 15 percent per hour. That data
set has been modeled by Isaksen et al. (1978), who obtained good agreement by
assuming a photochemical pathway involving oxidation of SOp by OH, HOp, and R02
radicals.
The oxidation rate of SOp in the troposphere on a regional scale of dis-
persion has not been measured directly, owing to the complexity of such an
experiment. However, based on atmospheric concentration data from monitoring
networks and an estimate of natural and manmade SOp source strengths, lifetimes
6-8
-------
of S02 and sulfate are estimated to be 25 and 80 hr, respectively, at midlati-
tudes for Europe (Rodhe, 1978). Accounting for the relative importance of
various S02 transformation and removal processes is difficult, but Rodhe (1978)
estimated that the turnover times of SO^ were partitioned between oxidation to
sulfate (80 hr), dry deposition (60 hr), and wet deposition (100 hr), with about
30 percent of the SO^ being oxidized and then removed as sulfate.
Measurement of oxygen isotope ratios has been applied to the study of S0?
oxidation in the atmosphere in an attempt to elucidate the principal pathways:
homogeneous gas-phase oxidation and heterogeneous oxidation (either in liquid
aerosols or on the surface of solid aerosol). Measurements of oxygen isotope
ratios for water vapor, precipitation water, sulfate in precipitation, sulfate
in aerosols, and S02 as a function of season indicate that aerosol sulfate shows
no correlation with the isotope ratio of water vapor or precipitation water, but
sulfate in rain does (Holt et al., 1979). It now appears that the turnover
times of water and sulfate are too dissimilar for oxygen isotope ratios to yield
information on sulfate aerosol pathways. However, the correlation of isotope
ratios for sulfate in precipitation and for precipitation water indicates that
the major pathway for sulfate formation in raindrops is in the condensed phase.
Because of the complexity of atmospheric processes, it appears at this time
that insight into the chemical pathways requires the study of simple systems in
the laboratory in conjunction with smog chamber and field studies.
6.2.1.2 Laboratory Investigation and Modeling
6-2.1.2.1 Smog chamber studies. Smog chamber studies such as those conducted
at Battelle's Columbus laboratories have provided information on the influence
of pollutant concentrations and the NMHC/NO ratio on sulfate formation (Miller,
/\
1978).
Experiments in which NMHC and NO concentrations were held constant while
J\
the S02 concentration was varied indicate that the conversion is first order in
6-9
-------
SCL, as would be expected for radical reactions. As shown in Figures 6-3A and
6-3B (Miller, 1978), the rates do vary somewhat with changes in NMHC and NO
concentrations. There is a greater variation in maximum rates (1.1 to 11 percent
per hour) than in total sulfate formed during a 6-hr radiation period (23 to 65
o
). It is not clear which measure is most appropriate for plumes, since the
smog chamber was operated under constant illumination instead of the diurnal
cycle of solar radiation.
As shown in Figure 6-3C (Miller, 1978), the maximum rate increases as the
NMNC/NO ratio increases. This is relevant to plume conversion because the
J\
plume contains high concentrations of NO and very low concentrations of NMHC.
/\
Thus, as background air with a higher NMHC/NO ratio mixes with or diffuses into
J\
the plume, the conversion rate within the plume increases. This inhibiting
effect of NO on photochemical smog reactions occurs because NO competes with
X X
S02 for OH and peroxy radicals. Since NO reacts rapidly with ozone, no ozone
will be found in the plume until enough ozone has diffused into the plume to
convert all the nitric oxide to nitrogen dioxide.
6.2.1.2.2 Plume simulation with a chemical kinetics model. A very simplified
model of plume behavior has been used to investigate how diurnal variations in
solar intensity, changes in solar intensity, and changes in amount of dilution
affect sulfate formation (Easter et aT. , 1980). The chemical kinetics model was
based on the average values of the pollutant concentrations in the plume and
thus neglected cross-plume variations in concentration and reaction rates, which
are thought to be very important. However, the simulation provides insight into
the interaction between mixing and solar radiation.
The model predictions of SO^ oxidation rates are presented in Figure 6-4
(Wilson, 1979). Each plot shows the S02 oxidation rate in ambient air and in
6-10 '
-------
1.4
1.3 h
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
I I I I I I I I .I I I
VI •
3%(hr)
4.5
A —
1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
INITIAL NMHC CONCENTRATION, ppm C
I
zf
o
c
z
UJ
O
O
u
o*
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
I I I II I T J II II I I
B —
3 -
I I
1 2 3 4 5 6 7 8 9 10 11 12 13 14
INITIAL NMHC CONCENTRATION, ppm C
14
I 8
I O 6
i ill
T 111
4 6 8 10
20
40 60 80100 200
INITIAL NMHC/NOX CONCENTRATION RATIO, ppm C/ppm
Figure 6-3. The initial NMHC and NOX concentrations affect (A) the maximum rates of S0: oxidation
and (B) the total conversion of SO2 to sulfate aerosol in a 6-hr irradiation interval. (C) The maximum
We of SOj oxidation is plotted as a function of initial NMHC/NOX ratio.
Source: Miller (1978).
6-11
-------
3
1 2
S i
^CLEAR SKY - UNSTABLE
BOUNDARY LAYER _
57 9 11 13 15 17 19 21
LOCAL TIME, hr
a 5
5'
i 3
1 2
<
£ 1
x
o
in
CLOUDY SKY - UNSTABLE
BOUNDARY LAYER _
9 11 13 15 17 19 21
LOCAL TIME, hr
CLEAR SKY - NEUTRAL
BOUNDARY LAYER
9 11 13 15 17 19 21
LOCAL TIME, hr
CLOUDY SKY - NEUTRAL
BOUNDARY LAYER —
9 11 13 15 17 19 21
LOCAL TIME, hr
CLEAR SKY - STABLE
BOUNDARY LAYER
9 11 13 15 17 19 21
LOCAL TIME, hr
01
| 3
1 2
<
S i
x
CLOUDY SKY - STABLE
BOUNDARY LAYER_
9 11 13 15 17 19 21
LOCALTIME.hr
Figure 6-4. The rate of oxidation of SO: is simulated by kinetic modeling. The solid curve in each plot
is the SO2 oxidation rate in ambient air. The dashed curves are the net S0: oxidation rates predicted
for power plant plumes emitted at 6 a.m., 11 a.m., and 3 p.m.
Source: Wilson (1979).
6-12
-------
plumes emitted at 6 a.m., 11 a.m., and 3 p.m. for a particular set of solar-
intensity and plume-dispersion conditions. The plume SOp oxidation rates were
calculated using "net-plume" SC^ and sulfate concentrations (average plume
concentrations minus ambient concentrations). For determining dispersion rates,
stability class was held constant for the neutral and stable conditions but was
varied for the unstable condition by following the diurnal pattern of boundary
layer evolution for a clear summer day.
The model predicts that the rate of SOp conversion to sulfate in plumes
will be bounded by the conversion rate in the ambient, background air. This is
not surprising since the plume contains higher levels of NO , which competes
A
with SOp for the OH and peroxy radicals. Plume SOp oxidation rates show a
distinct dependence on solar intensity: the rates are lowest near dawn and
dusk, and they are significantly lower in the simulations of cloudy versus clear
conditions. The 50 percent reduction in solar intensity for the cloudy con-
ditions results in a somewhat greater than 50 percent decrease in the ambient
S02 oxidation rates. According to the modeling estimates, the primary contri-
butions to SOp oxidation come from the reaction of SOp with OH, HOp and R02-
The OH concentration is dependent on solar intensity and on the concentration of
other species which vary during the simulations.
The plume SOp oxidation rates are also predicted to be sensitive to the
plume dispersion rate. Several hours are usually required after emission
begins before the oxidation rate in the plume reaches the ambient rate, and this
induction period is lengthened by slower plume dispersion. Free-radical con-
centrations in the plume are depressed below ambient levels during this initial
period, until plume NO concentrations are reduced by dispersion and by reaction
with ambient ozone. The induction period is also lengthened by lower solar
6-13
-------
radiation because of the resulting lower ambient ozone levels. The induction
period is also observed in field measurements and corresponds to the "early" and
"intermediate" stages in the chemical life history of the plume (Wilson, 1979).
It is clear that measurements of plume conversion rates made during the
periods from 7 to 9 a.m. and from 6 to 8 p.m. will give much lower rates than
similar measurements during midday. The lower rates are due both to lower solar
radiation and to reduced mixing.
6.2.1.3 Homogeneous Gas-Phase Reactions—The reactions involved in the homo-
geneous oxidation of S02 have been critically reviewed by Calvert et al. (1978).
It is now thought that the oxidation is governed by reactions of photochemically
generated OH, HCL, and CH-^CL radicals. For midlatitudes in the Northern Hemis-
phere, the oxidation is mainly due to the OH radical (Figure 6-5). The oxidation
rate is season dependent, with a maximum value of 0.2 percent per hour for July
and a minimum value of 0.09 percent per hour for January (Northern Hemisphere
average). Calculated conversion rates for the clean troposphere as a function
of season and latitude have been computed by Altshuller (1980) as follows:
July, 30°N, 0.43 percent per hour, 50°N, 0.24 percent per hour; January, 30°N,
0.10 percent per hour, 50°N, 0.01 percent per hour.
In polluted air masses, however, Calvert et al. (1978) estimated that the
rates of oxidation by RCHOO, 0, CH302, and H02 radicals greatly exceed the rate
of oxidation by the OH radical (see Figure 6-5 for a specific example of the
relative importance of these species in a polluted air mass). Recent laboratory
measurements (Graham, 1979) suggest that the rate of reaction of H02 with S02
may not be as great as the value used by Calvert et al. (1978). However, smog
chamber studies demonstrate that some species other than OH does convert S02 to
sulfate (Kuhlman et al., 1978; McNelis, et al, 1975). Additional investigations
are needed to .obtain reliable rates for the reaction of the HOp and R02 radicals
with S02.
6-14
-------
4 -
30 60 90
IRRADIATION TIME, min
120
Figure 6-5. The theoretical rate of reaction (percent per hour) of
various free-radical species on SO: is shown for a simulated sunlight-
irradiated (solar zenith angle of 40°) polluted atmosphere. The initial
concentrations (in ppm) were as follows: SOj , 0.05; NO, 0.15; N0:,
0.05; CO, 10;CH4, 1.5; CH2O, 0; CH3CHO, 0. The relative humidity
was 50 percent, and the temperature was 25CC.
Source: Calvert et al. (1978).
6-15
-------
6.2.1.4 Heterogeneous Aqueous Reactions—Whereas photochemical gas-phase
reaction rates appear to be high enough to dominate S02 oxidation during the
summer and in daytime hours, the pathway in the winter and nighttime hours is
not known at present. Although not well established, it appears that SC^
oxidation in liquid aerosols and fogs can reasonably proceed during these
periods through catalysis (by dissolved transition metal ions and carbon) and
through reaction with dissolved oxidants (ozone, hydrogen peroxide, and nitrogen
dioxide).
Studies of the autooxidation of dissolved SOp in pure water have yielded
rates that are too low to be significant for the lower troposphere, although it
has been suggested that this reaction may be important in clouds (Hegg and
Hobbs, 1978). It is highly unlikely that aerosols and fogs in the lower troposphere
consist of pure water, and it has not been demonstrated that the uncatalyzed
autooxidation may be dominant in cloud water.
There have been many investigations of the metal-catalyzed oxidation of
dissolved SCL species in water. The metals of main interest include manganese,
iron, copper, and nickel. Unfortunately, vanadium, which is also of interest,
has not been widely studied.
The kinetics of the manganese-catalyzed reaction have been investigated for
over 75 years. One of the first to criticize such studies was Titoff in 1903,
who remarked: "In Bigelow's (1898) work the reaction occurred between two phases,
and the retardation could be determined by a change in the boundary layer or by
a decrease in the solution rate of the oxygen." Although Titoff recognized the
possibility that the oxidation is dominated by mass transport, many later
investigators up to the present have not treated it adequately. Apart from work
done in two-phase systems, there have been only three studies. It is unusual
6-16
-------
that in none of these three reports were the rate expressions clearly stated;
instead, it was left to the reader to extract the rate constant. Estimates of
the rate expressions are presented in Table 6-1.
The work of Hoather and Goodeve (1934) and that of Coughanowr and Krause
(1965) are in good agreement, each showing a second-order dependence on manganese
concentration. However, Neytzell-de Wilde and Taverner (1958) reported a first-
order dependence on manganese concentration. This difference is currently
unresolved.
The iron-catalyzed oxidation of dissolved SO- species has not been investi-
gated as extensively as that of manganese. The only studies not using two-phase
systems (subject to mass transport limitations) are those of Neytzell-de Wilde
and Taverner (1958) and of Karraker (1963). Hegg and Hobbs (1978) have pointed
out that Karraker (1963) did not investigate the catalyzed oxidation in which
dissolved oxygen is the oxidant, but instead the redox system associated with
3+ - 2+
the couple Fe + e -»• Fe in an oxygen-free system. Thus, Karraker1 s work is
not considered applicable. Neytzell-de Wilde and Taverner (1958) reported that
2
the sulfate formation rate was second order for [S03 ], but Karraker (1963),
who reanalyzed their data, demonstrated that it was in fact first order. As
before, Neytzell-de Wilde and Taverner did not report the rate constant for
their data; however, an estimate is presented in Table 6-1. To date, the
kinetics of copper-catalyzed and vanadium-catalyzed oxidations have not been
satisfactorily investigated.
For the metal ion-catalyzed oxidation of dissolved 50^ in the absence of
mass transport considerations, the reaction kinetics for managanese are in
doubt, there is only one study for iron, and there are no satisfactory investi-
gations for copper and manganese. For the few systems that have been studied,
6-17
-------
TABLE 6-1. RATE EXPRESSIONS FOR THE Mn-, Fe-, Cu-, AND V-
CATALYZED OXIDATION OF DISSOLVED S02 IN WATER
Expression
Investigators
la.
b.
c.
2.
d[S02"]
dt
d[S02"]
~~dt
d[S02"]
~dt
d[S02"]
~~dt
44[Mn2+]1-7[S02-]°
pH 3-4
1.7 x 10'5[Mn2+] [SO2"]
pH ~ 2.2
8[Mn2+]2[S02~]°
pH 3-4
0.04 [Fe3+l [SO2'] [H*]"1
Hoather and Goodeve
(1934)
Neytzell-de Wilde
and Taverner (1958)
Coughanowr and Krause
(1965)
Neytzell-de Wilde
and Taverner (1958)
3. For copper: none
4. For vanadium: none
5. Mixed catalysts (e.g., manganese plus iron): none
Units = mole, liter, second.
6-13
-------
the ranges of pH, temperature, ionic strength, and type and concentration of
inhibitors have been insufficiently investigated. Also, the behavior of mixed-
catalyst (e.g., manganese plus iron) systems needs to be investigated. The
catalysis of the oxidation of dissolved SCL by carbon particles suspended in the
/w.
water has been studied by Chang et al . (H#9-). It was found that the oxidation
rate of dissolved SCL species was
-H20]= k[C] [02]°-69[S02.H20]° exp(-Ea/RT)
dt
With an activation energy of Efl = 11.7 kcal/mol over the pH range of 1.45 to 7.5
for the carbon studied, which was Nuchar-190. (The investigators demonstrated
that Nuchar-190 behaved similarly to soot from acetylene and natural gas flames.)
An average value of k = 1.17 x 10 mol ' x liter ' /g-sec was reported. The
kinetics have been interpreted in terms of the rate-limiting step being the
formation of an activated complex between molecular oxygen and the carbon
surface (Chang et al., 1978; Eatough et al . ,
Chang et al . (1978) have estimated that for 10 yg of fine carbon soot
o
suspended in 0.05 g of liquid water and dispersed in 1 m of air, the atmospheric
sulfate production would be about 1 yg/hr. At this time, it remains to be
demonstrated that the laboratory soots used by Chang et al . (1978) correspond to
those present in the atmosphere or that the suspension of soot at ambient levels
(<10 pg/m ) in aerosols, cloud droplets, or rain is similar to the laboratory
system.
Hydrogen peroxide, ozone, and nitrogen dioxide may be important in the
oxidation of SOp in aqueous aerosols and fogs. Although these compounds do not
demonstrate high reactivity toward S02 in air, their reactivity is enhanced in
the liquid phase. Again, caution is advised in accepting the results of studies
6-19
-------
of two-phase systems in which the investigators have not completely accounted
for the possibility of the mass transport limitation of the oxidation rate.
Therefore, only the recent results for single-phase systems are discussed here.
Penkett et al. (1979) used a stopped-flow reactor to determine the kinetics
of oxidation of dissolved S02 species by hydrogen peroxide. It was found that
the rate of sulfate formation is given by
+
= k[H202] [HSO~] [H+] + ka[H202] [HSO~] [HA]
5 2
where k = 5.2 x 10 liters/mol -sec, with k and k being the third-order rate
a
constants for the catalysis by free protons and proton-donating buffers, respectively
At pH <_ 4, it is found that k/k2 >_ 3200. Therefore, the second term is probably
not important for acid aerosols and fogs. It is of great significance that the
reaction rate increases as the solution becomes more acidic, which is in contrast
to aqueous oxidation by metal ions and by ozone. The activation energy and the
effect of ionic strength on the reaction have been measured by Penkett et al.
(1979).
The oxidation of dissolved S02 by ozone has been investigated with stopped-
flow systems. Penkett et al. (1979) have interpreted their work in terms of a
decomposition of ozone to initiate a free-radical chain reaction involving OH,
HS03, and HS05 radicals, after Backstrom (1934). Penkett et al. (1979) suggested
that the rate expression is
] , ,
= k[HSOg] [03] [H+r1
where k = 71 sec . Erickson et al. (1977) reported the fractional contributions
of the oxidation of the three sulfur oxide species by ozone at various pH
values; their rate expressions are
6-20
-------
d[soh
[0]
dt j^ 3
]
= k2[HS03-] [03]
] 2
= k3[S032 ] [03]
where kj = 590 1 iters/mol -sec, k2 = 3.1 x 10 liters/mol -sec, and k3 = 2.2 x 10
liters/mol -sec. The differences between the results of Penkett et al. (1979)
and Erickson et al . (1977) remain to be resolved.
The oxidation of S02 dissolved in water by NOp and HNOp may also be an
important sulfate formation process. Ross (1978) has reported that nitrous acid
completely oxidizes' S02 dissolved in water at pH = 4 within 15 to 30 sec when
both reactants are present in the 10 to 10 M range. At present, the reactions
involved in the oxidation of S02 by nitrous acid and nitrogen dioxide are
incompletely characterized, and their environmental significance is unknown.
At this time, the significance of reactions occurring in aqueous aerosols
in unknown. That is due principally to the lack of laboratory data for properly
designed two-phase reaction studies on complex chemical systems that simulate
environmental conditions.
6.2.1.5 Heterogeneous Surface Reactions—Recent laboratory studies of the
removal of S02 by airborne particulate matter have been conducted on a wide
variety of pure solids (Chun and Quon, 1973; Judeikis and Siege! , 1973; Haury et
al., 1978). Although reaction kinetics have not been identified, two general
types of processes have been: a capacity-limited reaction for S02 removal and a
catalytic S02 oxidation process. The initial contact of S02 with the solid
Produces a rapid loss of S02 from the gas phase; the reaction rate decreases
6-21
-------
with time. For the capacity-limited reaction, the rate slowly approaches zero;
for the catalytic process, the rate levels off for a time and then approaches
zero. The latter phenomenon is attributed to a pH decrease caused by sulfuric
acid formation.
Urone et al. (1968) and Smith et al. (1969) found a number of solids to be
effective in removing SO-. In Drone's studies, SO- admitted to a flask
containing a particular solid, allowed to react with no mixing, and the
product and remaining S0? were determined. Only the average reaction rates can
be calculated from these experiments; more importantly, with this experimental
procedure the rates will be diffusion limited. The highest rate determined was
for SOp with ferric oxide; the value was >75 percent per minute. Other materials
found to be slightly less reactive than ferric oxide were magnetite, lead
oxide, lead dioxide, calcium oxide, and aluminum oxide. The rate for the ferric
oxide experiment was for 20 mg of ferric oxide in a 2-liter flask; the ferric
oxide concentration would thus be 10 yg/m . Assuming a direct proportionality
between rate and particle concentration, the SOp removal rate in the atmosphere
would be calculated to be 0.04 percent per hour for 100 yg/m of particles with
the same reactivity as ferric oxide.
In Smith's experiments, an exploding-wire technique was used to generate
aerosols. There was a rapid decrease in SO- concentration, the rate depending
on the materials tested; the capacity for removal was in the order PbO > A1203
> Fe304.
Chun and Quon (1973) measured the reactivity of ferric oxide to SO^, using
a flow system involving a filter containing suspended particles. They determined
311
a removal rate constant of 9.4 x 10" ppm min (-din 0/dt), where 0 is the
fraction of surface sites available for reaction. Extrapolating this to an
atmospheric particle concentration of 100 yg/m with an equivalent reactivity
6-22
-------
and an S02 concentration of 0.1 ppm, the data project an atmospheric removal
rate of 0.1 percent per hour.
Stevens et. al. (1978) report total iron concentrations in six U.S.
cities ranging between 0.5 and 1.3 vg/m . Other species such as manganese,
copper, or vanaduim had total concentrations usually below 0.1 pg/m . Thus
actual ambient air concentrations are a factor of approximately 50 times less
3
than that assumed in the above papers. A reactive particle concentration of 2 pg/m
would yield a predicted S02 removal rate of no more than 0.002 percent per hour.
Therefore, surface reactions are probably not important except in sources prior
to or immediately after emission.
The most comprehensive study to date on SO,, removal by pure solids was made
byJudeikis (1974), who used a tubular flow reactor in which solids were supported
on an axial cylinder to measure reactivities; the calculated removal rates for
various solids are listed below.
Removal rate,
Substance percent per hour
MgO 32
Fe203 18
A1203 13
Mn02 10
Pb02 3
Nad 0.1
Charcoal 1
The values were estimated based on a previously derived model (Judeikis and
Siege!, 1973).
Charcoal is not a pure solid, but it is listed for comparison purposes.
Because of the ubiquitous nature of carbonaceous matter in ambient air partic-
samples, various workers have studied the S02 removal rate by carbon.
6-23
-------
Because of the varieties of carbon available for study, such as activated charcoal,
graphite, acetylene flame products, and combustion products of diesel oil and
heating oil, a comparison of the results is rather difficult. We cite here a
few investigations that deal with the gas-solid reaction of SCL with carbon.
Tartarelli et al. (1978) studied the interaction of S02 with carbonaceous
particles collected from the flue ducts of oil-burning power stations. They
concluded that the amount of adsorption' is increased by the presence of oxygen
and water in the gas stream. Reaction rates were not determined in this study.
Liberti et al. (1978) studied the adsorption and oxidation of S0? on various
particles, including soot from an oil furnace and various atmospheric particulate
samples. They concluded that the main interaction between the SOp and particu-
late matter is adsorption, with most catalytic reactions occurring at high
temperatures, near the combustion source. Their experiments with atmospheric
particulate samples led them to the conclusion that any heterogeneous nonphoto-
chemical sulfate formation is strongly dependent on the reactivity of the
aerosol surface, and hence the history (aged, freshly emitted), of the aerosol.
6.2.2 Physical Transformations
6.2.2.1 Aerosol Size Distribution—The size distribution of atmospheric aerosol
is important because it influences the rate of dry and wet deposition, transport
times, visibility impairment, and the site and extent of deposition in the lung.
The existence of a bimodal distribution was suggested in the early 1970's based
both on aerosol volume distributions derived from particle counts (Whitby et al.,
1972) and on mass distributions measured with impactors (Whitby et al., 1974;
reinterpretation of impactor data of Lee, 1972). There is now an overwhelming
amount of evidence that not only are two modes usually observed in the mass
6-24
-------
or volume distribution of well-mixed urban and rural aerosols, but that the
fine and coarse modes are normally quite different in chemical composition.
The physical separation of the fine and coarse modes occurs because con-
densation/coagulation processes produce fine particles, while mechanical pro-
cesses produce mostly coarse particles (Whitby, 1978).
This is shown in an idealized schematic in Figure 6-6. Individual sources
of primary aerosols may produce fine or coarse aerosols; some chemical species
in the coarse mode may have a tail extending into the fine mode. Secondary
aerosols, formed in the atmosphere from primary gaseous emissions, will normally
be fine. Recently it has been observed that fine particles may be further
subdivided into a nuclei mode and an accumulation mode (Wilson et al., 1977;
Whitby and Sverdrup, 1980).
The nuclei mode is observed only near combustion or other high temperature
sources. These particles rapidly coagulate into the accumulation mode. The
dynamics of particle growth normally prevent accumulation mode particles from
growing larger than about 1 ym.
Figure 6-7 shows a typical urban-type size distribution presented in five
different ways.
Typically, atmospheric aerosol number size distributions have been pre-
sented as log dN/d log D vs log D , as shown in Figure 6-7(a). Because a
good portion of such plots could be fitted by a power function of the form dN/d
k
iog D = AD this has been used by many investigators to model the number-size
distribution (Junge, 1953; Clark and Whitby, 1967). However, when such number
distributions are transformed to surface area-size, or volume-size distributions
[Figs. 6-7(d) and (e)], it is seen that the apparently minor deviations form the
power law are actually significant modes.
6-25
-------
Figure 6-6. Schematic of an atmospheric aerosol
size distribution showing the three modes, the main
source of mass for each mode, and the principal pro-
cesses involved in inserting mass into, and removing
mass from, each mode.
6-26
-------
10*
10*
10'
o.
0 10'
£
| ,
IO'1
10-
10'
(o)
/«•»•
•ss
0.001 0.01 O.I 1.0
Dp, pin
10
Figure 6-7. Average urban model aerosol distribution
plotted in five different ways. In (a) a power function
has been fitted to the number distribution over the size
range 0.0 to 32 ym. In (b) a log-normal distribution has
been fitted to the range 0.1 to 32 ym. It is evident that
the modal nature of the aerosol is shown best by the plots
of (c), (d), and (e).
6-27
-------
The linear-log forms shown in Figs. 6-7(c), (d), and (e) are useful
because the apparent area under the distribution curves is proportional to the
integral in the size range. From these plots, it is possible to visually judge
the relative N, S and V in the different modes of the distribution.
The form log dN/d log D, vs log D is dominated by the nuclei mode for most
urban and near-urban situations. For background aerosols, the number in the
accumulation mode may occasionally dominate the distribution. This plot
usually masks the model structure of aerosols, and erroneously suggest that
the whole distribution can be modeled by a continuous function. Although a
power function is often an adequate model for the number distribution in the
0.1-10 pm range, it is a poor model for the surface and volume distributions.
Cumulative plots of N, 5, V, or m vs log D [Fig. 6-7(b)] have widely
used for the presentation of powder size distributions and for the presentation
of impactor results (Wagman et al., 1967).
From such a plot of a unimodal distribution on logarithmic probability
paper, it is possible to determine the geometric mean by inspection. However,
the cumulative plot has several deficiencies. First, the cumulative form masks
the modes in the distributions. Second, the median used to determine the
geometric mean is a function of the upper and lower cutoffs of the measurement
method, since the distribution must be normalized by the total N, S, or V.
Thirdly, the median obtained from a cumulative plot is kind of an average
between the fine and coarse modes, and therefore does not have a clear-cut
physical significance. For the distribution shown in Fig. 6-7(e) the geometric
mean diameters by volume (DGV) of the accumulation and coarse particle
modes are 0.31 and 5.7 ym, respectively. The volume median from the
cumulative plot [Fig. 6-7(b)] is 2.5 ym.
The fundamentally modal nature of atmospheric size distribution was not
seen until about 1970 because the majority of workers in aerosols only plotted
6-28
-------
data as in Figs. 6-7(a) and (b). It is now clear that the N, S and V size
distributions must be all examined and then plotted in several ways if a
proper understanding is to be achieved (Whitby, 1978).
The fine and coarse particle modes in general: originate separately,
are transformed separately, are removed from the atmosphere by different
mechanisms, require different control techniques, have different chemical
composition, and have different optical properties. Therefore, the distinction
between fine and coarse particles is of fundamental importance to any discussion
of the physics, chemistry, measurement, or air quality standards of aerosols.
Of particular importance is the general division of acidic material into the
fine fraction and of basic material into the coarse fraction. Measurements
of acidity (pH) require that the sample be dissolved in water. If the fine and
coarse fractions are collected in the same sample, the acid particles in the fine
fraction will dissolve and be neutralized by coarse basic particles.
The existence of a bimodal distribution with fine and coarse modes has
been clearly demonstrated by cascade impactor studies which yield mass-size
distributions and by number distribution studies which may be converted into
volume distribution. These size distribution studies suggest 1-3 urn as the
most appropriate range for a cut-point for fine and coarse aerosols. However,
practical considerations of reducing plugging of impactor orifices indicate
that 2.5 ym is a more appropriate cut-point, especially for particle size
fractionating devices such as the dichotomous sampler (Miller et al., 1979).
Impactor studies in which chemical composition has been determined as a
Unction of particle size also demonstrate the division into fine and coarse
Nodes and show the difference in chemical composition of the two modes. Except
fora few trace elements, the chemical species are either primarily coarse,
6-29
-------
primarily fine, or bimodal. On the basis of such studies, it is possible to
divide the major chemical species observed in atmospheric aerosols into several
groups shown in Table 6-2 (Miller et al., 1979).
The major components of the fine fraction of the atmospheric aerosol are
sulfate, ammonium, nitrate ions, lead, carbon-containing material Including
soot, and condensed organic matter. Also, several studies have shown that
potentially toxic carcinogenic species, such as polynuclear aromatic compounds,
As, Se, Cd, Zn, which can exist as vapors, are more concentrated in the fine
particle fraction. In urban areas the fine fraction, as a percent of total
suspended particulate matter, varies from 15-25% in Denver to 40-60% in the
Los Angeles area and New York-New Jersey urban areas. The percent of the fine
particle fraction which is secondary varies from 60-80% in these urban areas.
These percentages are based on short-term intensive studies (Miller et al. 1979).
Sulfate is usually considered a secondary aerosol. However, in a few cases
in areas downwind of oil-fired power plants with relatively short stacks using
high vanadium oil primary sulfate can make an appreciable contribution to the
aerosol mass (Wilson et al, 1979).
The coarse fraction consist mainly of soil or curstal material, the major
elements being oxides of Si, Al, Ca, and Fe. The major source in most sampling
areas is soil or crustal material suspended by vehicular traffic from dirt roads,
dirt deposited on paved roads, unpaved shoulders, and rock used as filler in
pavement or from recreational or agricultural activity in rural areas (Harrison,
1976; Graf et al, 1977; Draftz, 1979; Hardy, 1979). Studies in Chicago have
shown that on the average reflotation (soil suspension from terrestrial surfaces)
accounts for approximately 20% of the annual TSP value in the city of Chicago.
On an annual average basis it was found that vehicular traffic suspends about
14 yg/m , while winds averaging above 13 mph suspend 5 yg/m (Newman et al., 1976)
6-30
-------
TABLE 6-2. CLASSIFICATION OF MAJOR CHEMICAL SPECIES
ASSOCIATED WITH ATMOSPHERIC AEROSOLS
_ • •
Normally
Fine
Normally
Coarse
Normally
Bimodal
Variable
So4=, C (soot), Fe, Ca, Ti, Mg N03", Cl" Zn, Cu, Ni, Mn
organic (con- K, P04~, SI, Al Sn, Cd, V, Sb
densed vapors), organic (pollen,
Pb, NH4+, As, spores, plant parts)
Se, H4
Source: Miller et al., 1979.
6-31
-------
In some samples from industrial areas or near specific sources the coarse
particle fraction may be dominated by local industrial sources (Draftz and
Severin, 1980). Other sources include wind erosion products, primary emissions,
sea spray, volcanic eruptions, and biological material such as pollen and
spores.
6.2.2.2 Aerosol Transformation—Aerosol dynamics concerns the evolution of the
particle size distribution. At any given moment, the particle size distribution
reflects the past history of the aerosol. It is determined by such physical
processes as nucleation (formation of new particles), condensation on existing
particles, coagulation (collision and subsequent adhesion), and depositional
loss from the atmosphere. The state of theoretical work in describing these
processes for application in phenomonological models appears to be satisfactory
(Hidy and Brock, 1970; Middleton and Brock, 1976; Middleton et al., 1977; Suck
/?7f
et al., 1*7-7-; Gelbard and Seinfeld, 1979). Factors affecting the rate at which
the particle size distribution changes include the mechanisms of chemical formation,
the prexisting particle size distribution, and the relative humidity. Recent
reviews of aerosol dynamics with special emphasis on the formation of sulfur
compounds include those by Friedlander (1978), Whitby (1978), and Boulaud et al.
(1978).
Aerosol formation in photochemical chambers has been frequently studied by
measuring the particle size distribution. By integrating and weighting the
size distribution, the aerosol number concentration, N, area concentration, A,
and volume concentration, V, can be measured as a function of time [Figure 6-8(b)j
Friedlander, 1978).
For aerosol growth processes involving homogeneous gas-phase reactions as
the source of condensable molecules, the growth can be divided into three time
domains (Husar and Whitby, 1973; Friedlander, 1978). In the first domain,
6-32
-------
HOMOGENEOUS QAt
PHASE REACTION
NEW
PARTICLE
FORMATION
CONDENSATION
AND
COAGULATION
FORMATION
AND
COAGULATION
TIME
Figure 6-8. (A) Condensible1 molecules (monomer) produced by gas-
phase reactions may either produce new particles or condense on
larger preexisting particles. New particles (cluster*) may also be
jcavenged by the preexisting particles. (B) Three time domains for
aerosol formation in an irradiated chamber.
Sources: (A) McMurry and Friedlander (1978).
(B) Friedlander (1978). after Husar and Whitby (1973).
6-33
-------
molecules are generated by gas-phase chemical reactions and collide to form
particles (nucleation). The nucleation of sulfuric acid has received consider-
able attention (e.g., Middleton and Kiang, 1978), and there is a qualitative
agreement between theory and experiment.
In the second domain, collision and adhesion (i.e., coagulation among the
newly formed particles) become important. The aerosol surface area Increases
and then approaches an approximately constant value in the third domain, in
which coagulation and condensation on the existing aerosol are the dominant
processes (Figure 6-8; McMurry and Friedlander, 1978; Friedlander, 1978).
The growth processes are illustrated in Figure 6-8A (McMurry and Friedlander,
1978). If the surface area of the preexisting aerosol is sufficiently high, the
formation of new particles will be suppressed by scavenging onto the existing
particles. The scavenging of newly formed nuclei and condensable molecules
constitutes the key mechanism by which "accumulation mode" aerosols are formed
in the atmosphere (Whitby, 1978).
The gas-to-particle conversion rate can be measured directly by monitoring
the rate of change of the aerosol volume concentration, dV/dt. The theory of
aerosol dynamics for such a system was formulated by McMurray and Friedlander
(1978). Clark et al. (1976) used such a system in Los Angeles, where they added
SOp to ambient air containing hydrocarbons and NO . Their observations showed
t- /\
that aerosol formation is linearly proportional to the concentration of S02
(Figure 6-9A; Clark et al., 1976) and to that of nonmethane hydrocarbons (Figure
6-9B; Clark et al., 1976).
In summary, the phenomenology and theory of aerosol dynamics in photochemical
systems are reasonably well understood. Nucleation, condensation, and coagulation
are the dominant mechanisms, scavenging being responsible for the growth of
6-34
-------
I I I I
CORRELATION COEFFICIENT • 0.99
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
INITIAL S02 CONCENTRATION, ppm
I
I
I
I
0.5 1.0 1.5 2.0
INITIAL NMHC CONCENTRATION, ppm
2.5
Figure o-9. (A) Volumetric conversion rate is shown as a function of
initial SOS concentration for an ultraviolet light intensity of 30.5
W/mJ and an NMHC concentration of 1.57 ppm. (B) Volumetric
conversion rate is shown as a function of initial NMHC concentration
for SOj concentration of 0.017 ppm and an ultraviolet light inten-
sity of 30.5 W/mJ.
Source: Clark etal. (1976).
-------
accumulation-mode aerosols in the size range from 0.5 to 1.0 pm. Unfortunately,
there is no comparable understanding of aerosol dynamics for liquid-phase or
surface-controlled reactions.
6.2.3 Water Uptake and Release
An important form of physical transformation is the uptake or release of
water vapor from atmospheric aerosols. Most atmospheric sulfates and nitrates
are water soluble and hygroscopic. Concurrent with the identification of
sulfates as major constituents of atmospheric fine particles, there has been
considerable interest in the hygroscopic properties of pure and mixed sulfur
compounds (Junge, 1952, 1963; Winkler, 1973; Charlson et al., 1974; Tang, 1976).
The hygroscopic growth of sulfate-containing particles has a profound effect on
their size, shape, refractive index, pH, and reactivity, which in turn influence
their toxicity, removal efficiency, and effects on climate and weather.
The hygroscopic properties can be obtained from observation of aqueous
solution properties such as equilibrium water vapor pressure or relative humidity
(RH) as a function of solute/solvent mole fraction. The dependence of the
sulfate mole fraction on RH for three different sulfate compounds is illustrated
in Figure 6-10 (Charlson et al., 1978). (Raoult's law behavior assuming complete
dissociation is included for comparison.) Sulfuric acid exhibits a monotonic
curve characteristic of simple hygroscopic behavior. By contrast, ammonium
bisulfate and ammonium sulfate exhibit stepwise increases at 39 and 81 percent
RH, respectively, and monotonic curves at higher RH. This latter phenomenon, a
sudden uptake of water when the RH exceeds a certain level, is termed "deliquescence1
and is exhibited by these two compounds and a number of other organic and
inorganic compounds, primarily salts. The existence of and differences in
deliquescence points among compounds provide a possible basis for identification
(e.g., Charlson et al., 1978).
6-36
-------
o
5
<
oc
u.
01
_l
O
O
V)
A AMMONIUM
SULFATE
•—RAOULT'S LAW ASSUMING
COMPLETE DISSOCIATION
I I
0 0.2 0.4 0.6 0.8 1.0
WATER VAPOR PRESSURE AS FRACTION OF SATURATION
Figure 6-10. Solute mole fraction is shown as a function of water
vapor pressure (as a fraction of saturation) for three sulfate com-
pounds: sulfuric acid (Coordinating Research Council, 1968);
ammonium bisujfate (Tang and Munkelwitz, 1977); and ammonium
sulfate (Low, Wtt). Curve showing data according to Raoult's law is
from Charlson et al. (1978).
6-37
-------
Under atmospheric conditions, hygroscopic/deliquescent compounds in the
aerosol phase equilibrate rapidly with their surroundings and exhibit changes in
size and other properties in response to RH changes. Figure 6-11 (Charlson et
al., 1978) shows theoretical growth (increasing RH) curves for some sulfate
compound aerosols based on vapor pressure data. In the aerosol phase, most
compounds seem to remain as solution droplets at concentrations well above
saturation, that is, at an RH well below the deliquescence point. This is
caused by the phenomenon called hysteresis (Orr et al., 1958; Junge, 1963).
Atmospheric aerosols probably exhibit hysteresis (Winkler and Junge, 1972), but
this has been confirmed only on deposited aerosol samples. If so, they may
contain considerably more water in the range 30 to 70 percent RH than would be
expected from their deliquescence curves.
Since vapor pressure data are based on bulk measurements of pure compounds,
corrections must be made for the increase in vapor pressure over curved surfaces
(Kelvin effect) and, in the case of deliquescence, for the small number of
crystal units present. For the complex mixture of compounds generally found in
atmospheric aerosols, hygroscopic properties are rather difficult to model.
A number of techniques have been used to directly investigate the hygro-
scopic properties of aerosols (Orr et al., 1958; Tang and Munkelwitz, 1977;
Kapustin et al., 1974; Sinclair et al., 1974; Winkler, 1973; Covert, 1974;
Charlson et al., 1974). Charlson et al. (1974) have developed a humidograph
that measures the light-scattering coefficient of the particulate matter, b ,
as a function of increasing RH. This technique provides an in situ measurement
of b and is most sensitive to the concentration of particles 0.3 to 1.0 ym in
diameter.
Humidograms of pure laboratory-generated aerosols are shown in Figure 6-12
(Charlson et al., 1978). The humidograms for sulfuric acid and ammonium bisulfate
6-38
-------
O AMMONIUM BISULPATE
D SULFURIC ACID
A AMMONIUM SULFATE
Figure 6-11. Theoretical growth curves based on vapor
pressure data are shown for three sulfate aerosols
exposed to increasing humidity. The dry particles
have initial radius, ro, and they grow to radius, r,
which is a function of relative humidity.
Sgyrpe; Charlson et al. (1978).
§=11
-------
T I I I 1 I I I I
I I I I I I I I I
1 I I [ I I I
I I T l^ I IT
I 1 I I l I l I I
I I I I I I I I I
i I i r
ITT
4 -
I I I I I II IT
l i i i i I
so
100
1 -
Figure 6-12. Humidograph data for pure, laboratory-
generated sulfate compounds show deliquescence steps
for some compounds. Dashed curves are calculated
except in graph I.
Source: Charlson et al. (1978). €-40
-------
are for practical purposes indistinguishable, the expected deliquescence point
for ammonium bisulfate was not observed experimentally. Those for both
(NH4)3H(SO.)2 and ammonium sulfate are distinguishable from the rest because of
the presence of their deliquescence steps. The mixture of these two salts
(Figure 6-12E; Charlson et al., 1978) shows two deliquescence steps at 45 and 80
percent. If (NH^HtSO^ is not an important atmospheric species, these
differences in response to relative humidity, coupled with the fact that the
acid sulfate aerosols can be changed into ammonium sulfate (thus altering their
hygroscopic nature) by in situ reaction with gaseous ammonia, provide a means
for detecting sulfate and for distinguishing between ammonium sulfate and the
more acidic forms (Charlson et al., 1974).
Atmospheric aerosols seldom consist of pure compounds, however. The "mixed
aerosol" hypothesis advanced by Junge (1952) and extended by Winkler (1973)
defines two extreme cases for chemical mixing in aerosols: (1) internal mix-
tures, in which each individual aerosol particle contains a mixture of compounds,
and (2) external mixtures, in which each particle is a pure compound, yet the
population of aerosol particles consists of various compounds.
All intermediate degrees of mixing are possible. External mixtures will
have hygroscopic properties that are simple linear combinations of the individual
compounds. Internal mixtures exhibit nonlinear behavior with respect to hygro-
scopic growth because of ion interactions, which tend to increase solubility.
An example of this effect is illustrated in Figure 6-121 (Charlson et al., 1978)
for external and internal mixtures of ammonium sulfate, sodium chloride, and
potassium nitrate in mass ratios of 2:1:1.
In summary, the hygroscopic and deliquescent behavior of pure sulfate
compounds in aerosol form can be predicted from solution equilibrium calculation.
However, water uptake by mixed salts and atmospheric aerosols can be obtained
only from aerosol measurements as a function of relative humidity.
6-41
-------
6.2.4 Carbon-Containing Aerosols
The carbon in atmospheric aerosols consists of an elemental component
(such as graphite or soot) and a nonvolatile, organic component. There are
significant differences in the optical properties of these two components.
Elemental carbon is formed during the combustion of fossil fuels and Is emitted
as primary particles (^0.1 urn), which strongly absorb light. The nonvolatile,
organic component consists of primary hydrocarbons emitted in combustion exhaust
and of secondary organics formed by photochemical reactions. These primary
hydrocarbons and secondary organic vapors either nucleate or condense on existing
aerosols. They do not strongly absorb light, but usually make important contributions
to light scattering in urban hazes.
The importance of elemental carbon has been emphasized in a conference on
/f/f
"Carbonaceous Particles in the Atmosphere" (Novakov, Wf9). However, there
are only limited data on the mass ratio of elemental/(primary + secondary)
organics for a few cities. Appel et al (-t9f9) found for a four-day period in
»
July 1975 in Pasadena, Pomona, and Riverside that elemental carbon was the most
abundent carbon species; also, the secondary organic mass was usually twice
that of primary hydrocarbons. Of the secondary organics, adipic and glutaric
acids were among the most abundent products; most likely they were oxidation
products of cyclohexene and cyclopentene emitted by motor vehicles. For total
carbon, benzene soluble organics, and hydrogen analyses of fine aerosols collected
in Denver in Nov. 1971, it was estimated that the elemental carbon was 2.3-3.6
vg/m for the episode days observed; from measured Pb concentration, it was suggested
that in Nov. 1973 in Denver the elemental carbon in fine aerosol was 1.7-4.4 pg/m
tiy?
(Durham et al , 1-979") . Also, for Denver in Nov. 1973, Pierson and Russell (1979)
3
estimated from Hi-Vol samples the total elemental carbon to be 2.9-27.6 yg/m .
Although atmospheric measurements of carbon-containing aerosols are less
complete than those of sulfates, available results suggest that carbon-containing
6-42
-------
aerosols in many locations, both urban and nonurban, are the second most abundant
fine-particle species after sulfates. At some western urban locations where S0x
emissions have been small, carbon-containing aerosols have made the largest
contribution to fine-particle mass. The concentration of primary carbonaceous
particles is likely to have been even higher in the past in the Eastern United
States when coal was more widely used as a fuel. With the growing use of wood
combustors for home heating, carbonaceous particle concentrations are likely to
increase.
The National Academy of Sciences (Grosjean, 1977) has extensively reviewed
the methods of primary and secondary organic aerosol identification, and the
physical and chemical aspects of their formation. Primary organics emitted
into the atmosphere by industrial sources, motor vehicles, agriculture activities,
and natural sources include: linear and branched alkanes and alkenes, substituted
benzenes and styrenes, quinones, acridines, quinolines, phenols, cresols,
pthlalates, fatty acids, carbonyl compounds, polyaromatic hydrocarbons, terpenes
and pesticides. Secondary organic aerosols are formed ty the oxidation reactions
of the primary organics, ozone, and nitrogen oxides. Typical products that have
been identified are: aliphatic organic nitrates, dicarboxylic acids, benzoic
and phenylacetic acids, and terpene products such as pinonic acid (Grosjean and
Friedlander, 1975; Miller et al., 1972; Schuetzle et al., 1975). By using
computer-controlled high-resolution mass spectrometry and thermal analysis
Schuetzel et al. (1975) and Cronn et al. (1977) obtained diurnal variations of
primary and secondary organics from two-hour size-resolved samples. The
variations of hexanedioic acid and total amides with ozone are shown in Figure
6-13 for West Covina. The concentrations of these secondary organic aerosols and
the ozone are highly correlated.
6-43
-------
10
\
K> o 4 t e
TIME (PST)
•D »<
04
OT
a«
S 04
01
01
O.I
1™
VEST
TOT* i
•
•
. WtIT (
•l
. . , i .
CO VIM 1
• '11
4MIDC
*
:OVINJ
* i ,
5
. . i . .
-
-
'>
^
^
\
V
—
s— ^
B
too 4 • a*
TIME
K>a»
Figure 6-13. (a) Diurnal variation of hexanedioic
acid, a presumed secondary component.
Figure 6-13. (b) Comparison of the diurnal variation
of total amides (a presumed secondary component) with
ozone.
Source: Cronn et al., (1977).
6-44
-------
In an attempt to understand the atmospheric oxidation pathways that
yield secondary organic aerosols, simple mixtures have been investigated in
laboratory chamber studies. As discussed in more detail in the National Academy
of Sciences report (Grosjean, 1977), the following trends have been observed
by chamber researchers: (a) Most paraffins do not generate aerosols during
irradiation, (b) acetylenics do not form aerosols, (c) all unsaturated compounds
with six or more carbon atoms can form organic aerosols, (d) cyclic olefins and
diolefins form more aerosol than their 1-alkene analogs, (e) conflicting
results have been reported on the aerosol-forming ability of aromatics, (f) carbonyl
compounds do not generate aerosol; and (g) mechanical stirring inhibits aerosol
formation. Cyclic olefins are the most important class of organic aerosol
precursors, due mainly to their high gas-phase reactivity and their ability
to form non-volatile dicarboxylic acids.
The chemical composition of "model" organic aerosols generated from
single hydrocarbons is not well-established for suspected important
aerosol precursors. Functional group analyses for the products of olefins,
benzene and benzene substituted compounds, and terpenes that have reacted with
ozone show that the bulk consists of high oxygenated compounds, which include:
carbonyls, carboxylic acids, and nitrate esters. Only a few studies of species
identification have been reported. Detailed aerosol product identification has
been reported for the ozone-1-butene reaction (Lipeles, 1973); the NO -toluene,
A
NO -cyclohexene, and NO -a-pinene photo-reactions (Schwartz, 1974), and the NO -
Ax x
cyclopetene, NO -cyclohexene, and 1-7-octadiene photo-reactions (Grosjean, 1977).
J\
Good agreement was indicated by Grosjean (1977) with Schwartz (1974) for the
N0x-cyclohexene photo-reaction, except that Grosjean observed adipic acid to
« the major product (not found to be present by Schwartz). It is significant
6-45
-------
that most of the polyfunctional compounds identified (see the cited papers and
the NAS report for details) have also been identified as important constituents
in ambient aerosols.
The secondary aerosols formed from alkenes having seven or more carbons
(cyclic olefins, diolefins, and terpenes) grow into the light-scattering range
and produce appreciable visibility reduction. For example, aerosols formed from
cyclic olefins and diolefins have particle sizes between 0.1 and 0.3 ym. For
such systems, the gas-to-aerosol conversion process consists of the formation of
supersaturation in the gas phase and subsequent condensation on preexisting
particles.
The rates of conversion of precursor organic vapors to organic aerosols in
Los Angeles have been estimated to average 1 to 2 percent per hour. This
moderate rate of conversion is consistent with the observation that organics
account for an important fraction of the fine aerosols under conditions of
intense photochemical activity, while only a small part of the precursor organic
vapors are converted to particulate matter.
6.2.5 Particulate Nitrate
A number of topics related to the transport of nitrogen oxides and their
transformation to gaseous and particulate nitrates are discussed in the Air
Quality Criteria for Oxides of Nitrogen (to be published, 1980). These topics
include visibility, environmental transport and transformation, and acidic
precipitation.
At eastern urban sites such as St. Louis, Mo., Dayton, Oh., and Charleston,
3
W.V., particulate nitrate concentrations did not exceed 1 to 2 yg/m , and these
nitrate concentrations often were below 1 pg/m (Spicer, 1976; Stevens et
a]., 1978). At several eastern nonurban sites, nitrates ranged from 0.1 to 0.6
o
yg/nT (Mueller et a!., HW). In the measurements in Charleston, W.V- (Stevens et
a!., 1978), the concentration of particulate sulfate was 5 to 50 times greater
6-46
-------
than that of participate nitrate. At the nonurban sites, the ratio of participate
sulfate to participate nitrate ranged from 10:1 to 100:1 (Mueller et a!.,
1979). The results should not be extrapolated to western sites.
Atmospheric nitrates most likely result from photochemical reactions
involving the oxidation of NO and NOp to yield HN03 and organic nitrates
(Demerjian et al., 1974; Orel and Seinfeld, 1977). The measurement of ambient
nitrate aerosols has been recognized to be subject to significant biases. Glass
fiber filters may exhibit a positive bias due to the absorption of gaseous HN03
(Spicer and Schumacher, 1979; Appel et al, 1979), and negative biases may result
from the evaporation of NH^NOg in both solid (Stelson et al., 1979) and aqueous
solution (Tang, 1979). Also, a negative bias may result from displacement of
HNO, from particulate nitrate by H9SO. aerosols being collected on the filter
J ftq <- *
(Barker et al., W-9-). The measurement methodology for HN07 and inorganic
r J
particulate nitrate measurement and their biases have been reviewed (Stevens,
1979). Until an understanding is gained for particulate nitrate sampling with
glass fiber, quartz fiber and Teflon filters, nitrate data bases cannot be regarded
as valid.
Because of the substantial increases anticipated in nitrogen oxide emissions
from utility sources, current nitrate levels must be carefully determined and
followed.
6.3 WET AND DRY REMOVAL OF S02 AND AEROSOLS
Deposition processes limit the lifetime of sulfur compounds and other
compounds in the atmosphere, control the distance traveled before removal, and
limit their atmospheric concentrations. In addition to fulfilling the essential
role of cleansing the atmosphere, these processes are also important because
6-47
-------
they deliver sulfur compounds to surface waters, soils, and ecosystems (see
Chapter 8 on acidic precipitation). Because of these features and the growing
concern over adverse physical and biological effects, the overall question of
deposition has received increasing attention in the fields of atmospheric
chemistry and biology.
SOp and aerosols are removed from the atmosphere by wet deposition (incor-
poration in precipitation) and dry deposition (direct transfer in the absence of
precipitation). These two processes are of approximately equal magnitude.
P
According to Rodhe (1978), the estimated overall turnover times of S0?, SO^T
2_
and sulfur (SO- + SO^") in Europe at midlatitudes are roughly 25, 80, and 50 hr,
respectively; the turnover times for wet and dry deposition are approximately
100 and 60 hr, respectively.
6.3.1 Dry Deposition
The dry deposition of pollutant gases and aerosols has been recently
tf?6
reviewed by Chamberlain (1975), Garland (1978), Slinn (-t9?4}, and Slinn et al.
(1978). It is convenient to view the overall mass transfer rate as the product
of a mass transfer coefficient (V.) and the difference between the bulk concentration
and the surface concentration (Garland, 1978). It it is also convenient to
envision the deposition velocity as the inverse of the overall resistance to
mass transfer. The total resistance is the sum of several independent resistances.
The surface resistance incorporates absorption, adsorption for gases, and
adhesion for aerosol particles. It is generally believed that the surface of
vegetation is responsible for most of the SOp uptake by dry deposition. The
surface resistance of dry vegetation is governed primarily by the diffusion rate
of SOp through the stomatal openings (Garland, 1978). The opening and closing
of stomata depends on the type of vegetation, solar radiation» water stress, and
other important variables. It has been proposed, however, that the S0? uptake
6-48
-------
is closely related to the transfer of water vapor. Fowler and Unsworth (1974)
point out that the surface resistance over vegetation wetted by dew is nearly
zero until the deposition is limited by the decrease in the solubility of S0? as
the pH of the dew decreases as a result of the oxidation of S02 to sulfuric
acid.
For fine particles in the submicron size range, the collision efficiency
with surfaces is unity; therefore their surface resistance is zero.
The aerodynamic resistance is determined by the turbulent eddy diffusion
that delivers the pollutants from a height of several meters to the surfaces.
The mean horizontal wind speed, surface roughness, and to some extent atmospheric
stability are the key factors determining the aerodynamic resistance.
Field and laboratory data suggest that over water, calcareous soil, and
short grass, surface resistance to SCL uptake is low and is often exceeded by
atmospheric resistance to transfer. However, over forests, tall crops, acidic
soil, and snow, surface resistance contributes substantially to the total transfer
resistance. In such cases, not only are meteorological parameters important in
determining deposition rates, but also the chemical and biological characteristics
of the surface (e.g., wetness, degree of leaf stomatal opening) (Garland, 1978).
Table 6-3 gives an estimate of deposition velocities for S02 over various
types of vegetation. Table 6-4 gives similar information for calcareous and
acidic soils (ISSA Workshop, 1978). A midday maximum in Vd is frequently observed,
reflecting diurnal maxima in stomatal opening and in turbulent transport. A
seasonal variation, with minimum values less than 0.3 cm/sec when the vegetation
is dry or senescent, is also observed. The presence of water on foliage leads
to larger values (1 cm/sec or greater) providing the pH of the water is greater
than about 4.
6-49
-------
TABLE 6-3. S02 DEPOSITION VELOCITIES OVER VEGETATION
Vegetation
Height
Short
Medium
Tall
Example
Grass
Crops
Forest
Height,
m
0.1
1.0
10.0
Range of V,
cm/sec
0.1-0.8
0.2-1.5
0.2-2.0
Typical V,a
cm/ sec
0.5
0.7
Uncertain
These values were obtained in a humid climate. Much lower values are
likely in arid climates.
Source: ISSA Workshop (1978).
6-50
-------
TABLE 6-4. S02 DEPOSITION VELOCITIES OVER SOIL3
Type of soil
Calcareous
Calcareous
Acid
Acid
PH
>7
>7
-4
~4
State
Dry
Wet
Dry
Wet
Range of V,
cm/sec
0.3 to 1.0
0.3 to 1.0
0.1 to 0.5
0.1 to 0.8
Typical V,
cm/sec
0.8
0.8
0.4
0.6
aAs yet, no information is available to assess v on desert sand or
lateritic soils. 9
Source: ISSA Workshop (1978).
6-51
-------
Several other types of surface should be mentioned. Sparse experimental
evidence suggests that dry snow is a very poor sink for SOp (V, - 0.1 cm/sec).
Additionally, the surface is naturally smooth so that atmospheric mixing is
minimal. Wet snow behaves like water. Natural water surfaces are efficient
sinks for SCL (V,, 0.2 to 1.5 cm/sec, typically 0.7 cm/sec) unless very acidic.
Rates of deposition are determined by atmospheric mixing intensity. There are
very few studies of dry deposition from plumes or of atmospheric budgets. These
few studies provide estimates of V . over large areas of countryside. These
estimates are generally consistent with estimates for homogeneous surfaces (V,,
0.2 to 2.0 cm/sec, typically 0.8 cm/sec). Laboratory measurements show that
many building materials absorb SOp at significant rates, but it is difficult
to use the results to estimate deposition on built-up areas. Observations in
London during a prolonged period of air stagnation suggest a deposition velocity
of 0.7 cm/sec, but this may be atypical.
The dry-deposition velocity for aerosols depends primarily on their size
and density, the intensity of turbulent mixing near the surfaces (the friction
velocity, u*), and the roughness height of the surfaces. Large particles
/72T
deposit mainly by sedimentation (Figure 6-14; SI inn et al., HW-j, but particles
in the range 1.0 to 50 ym are also transported toward the surfaces by turbulence
where sedimentation is supplemented by impaction and interception by roughness
elements. Very fine particles diffuse to surfaces by Brownian motion, which is
greater for smaller particles. Sehmel and Hodgson (1976) found that the deposition
velocity for the accumulation-mode aerosols in the size range from 0.1 to 1.0 vm
is minimal, since none of the above processes is efficient. The dry removal
rate of accumulation-mode particles depends strongly on the surface roughness
111f
(Figure 6-14; SI inn et al., W79). The roughness height (ZQ) in Figure 6-14
6-52
-------
•O
(J
o
_J
iu
z
o
UJ
O
I I I |MI! | II
STABLE ATMOSPHERE WITH
ROUGHNESS HEIGHT
Zo
-------
is related to, but is not identical with, the physical roughness height. The
data of Sehmel and Hodgson (1976) also indicate that the dominant resistance to
particle dry deposition occurs within a few millimeters of a deposition surface.
Consequently, the airborne particle concentrations vary little above the surfaces.
Since deposition velocity depends on particle size, estimation of the rate
of deposition of particles requires a knowledge of aerosol size distribution.
Most sulfate mass lies in the size-range from 0.1 to 1.0 urn (Whitby, 1978;
Stevens et a!., 1978), for which the dry deposition velocity is not expected to
exceed an average of 0.5 cm/sec.
The present understanding of dry deposition is such that reasonable estimates
can be made for deposition rates of S02 for use in models, but not for sulfate
or other aerosols. It should be emphasized that current dry deposition studies
refer to uniform terrain and their appropriateness for use in cases of rough or
complex terrain is not known.
6.3.2 Wet Removal
Wet deposition can be defined as the removal of material from the atmos-
phere by incorporation into precipitation through various scavenging mechanisms.
Mechanisms for the incorporation of pollutants into precipitation depend on
the diffusion coefficient of gases, size distribution of particles, characteristics
of the precipitation, and location of the pollutant within the precipitating
cloud. The main attachment mechanisms include convection, impaction, interception,
molecular and Brownian diffusion, turbulent transport, nucleation, thermophoresis
and diffusiophoresis, and electrical effects.
The predominant attachment mechanisms for S0? are convection and diffusion,
for both in- and below-cloud scavenging. According to a recent review by Hales
(1978), for a known distribution of SO-, the SO- scavenging rate can be calculated
with reasonable confidence.
6-54
-------
Typical sulfate concentrations and wet removal rates for sulfur compounds
are summarized in Table 6-5 (ISSA Workshop, 1978). The ranges are given in
order to show spatial variations in each region.
Wet deposition of sulfate formed by SCL oxidation in falling raindrops is
an important removal mechanism. Modeling calculations (Overton et al., 1979;
Hill and Adamowicz, 1977) suggest that sulfate from the oxidation of S02 in
these droplets may range from 3 to 87 pmol/liter (and pH from 4.5 to 6.5) in the
presence of SO^ (10 ppb), ozone (0 to 50 ppb), ammonia (5 to 20 ppb), carbon
dioxide (320 ppm), and iron (10" ymol/liter).
In summary, dry and wet deposition processes are such complex phenomena
that even after several decades of research our understanding is inadequate.
Dry deposition has been widely investigated, but the representativeness of the
results is not known. No routine monitoring of dry deposition of S02 and
sulfate has been conducted. On the other hand, wet deposition of sulfate has
been monitored in Europe since the late 1950's and precipitation monitoring
networks have been set up in North America. Identifying the rates and mechanisms
of atmospheric deposition on land and water is a key problem of atmospheric
chemistry.
6.4 DISPERSION IN THE ATMOSPHERE
Both horizontal and vertical dispersion mechanisms contribute to the movement
and dilution of pollutants in the atmosphere. Horizontal transport, characterized
by wind speed and wind direction, determines long-range transport and plume
trajectory. Turbulence provides vertical mixing that contributes to dilution
and is a major factor in deposition. The mixing of plume material with potentially
reactive background air is facilitated by both horizontal and vertical dispersion.
6-55
-------
TABLE 6-5. REPRESENTATIVE ANNUAL AVERAGE SULFUR WET AND DRY DEPOSITION RATES
Location
Heavily
industri-
alized
areas
Rural
Remote
North
America
Europe
North
America
Europe
North
Atlantic
Other
oceans
Continents
Excess sulfate in
precipitation,
mg sec/liter
3-?
3-20
0.5-2
0.5-3
0.2-0.6
0.04
0.1
Wet deposition
rate
g sec/(rr/yr)
0.1a-3
2-4
0.1-2
0.2-2
0.1-0.3
0.01a-0.2
0.01a-0.5
Dry deposition
rate,
g sec/m /yr)
?
3-15
0.2-2.6
0.5-5.0
0.04-0.4
<0.1
0.4
Low deposition rates result from low precipitation.
Source: ISSA Workshop (1978).
6-56
-------
In the past there have been a number of severe air pollution episodes, when
meteorological conditions have resulted in minimal atmospheric dilution over
regions for extended periods. Historically, these episodes (e.g., Donora,
London, Los Angeles) can be identified as resulting from stagnant air masses
having light winds and limited vertical mixing heights in which pollutants
accumulate.
By building tall stacks, pollutants can be injected into the faster winds
of air layers at higher elevations. Often these layers are decoupled from the
layers of air near the surface.
6.4.1 Spatial and Temporal Scales
The transmission of air pollutants in the atmosphere is a function of
chemical transformation rates within the atmosphere and by removal rates, which
are in turn affected by meteorological conditions. There are basically two
approaches to estimating the time scales of transformation, transport, and
removal processes. The most common approach is to study the atmospheric processes
themselves by measurements in the atmosphere or in the laboratory, or by theoretical
calculations. The other approach is to use mass conservation or "budget" calcula-
tions to estimate the space and time averages of such processes. Geochemical
budgeting constitutes a systematic comparison of sources, sinks, and concentrations
of a given compound averaged over reasonably large time and space scales.
The lifetime of atmospheric trace constituents emitted directly by man's
activities ranges from a few seconds for large particles of tire debris to many
years for fluorocarbons, methane, and carbon dioxide (Figure 6-15). The typical
gaseous pollutants emitted into the lower troposphere, such as SOp and nitrogen
oxides, reside in the atmosphere for days.
6-57
-------
10
SYNOPTIC -
REGIONAL
FINE PARTICLES
«2 urn)
MESOSCALE
NO,
COARSE PARTICLES
2
-------
The transport distance of a given compound .is determined by its physical
and chemical characteristics and by the mean wind speed between the source and
the receptor. Thus, atmospheric processes and a pollutant's chemical and physical
properties determine the residence time of a chemical compound 1n the atmosphere.
For the description of the spatial-temporal behavior of chemical compounds in
the atmosphere, it is convenient to use established meteorological terminology.
The microscale is defined as a scale in which atmospheric processes occur within
1 km or less. There are only a few primary pollutants that have a mean residence
time within the microscale; examples are particles from tire erosion and heavy
windblown dust.
The spatial scale of one to a few hundred kilometers and the corresponding
3 *>
temporal scale from 15 min to a day (10 to 10 sec) are referred to as the
mesoscale. This scale encompasses major thunderstorms and discernible single or
urban plumes. Atmospheric coarse particles in the size range of 2 to 20 pm tend
to fall out within this scale. Of the gaseous primary emissions, the oxides of
nitrogen, NO and NO^, are typically consumed by chemical reactions or deposition
within the first day after emission; a large fraction oxidize.
At a few hundred kilometers the mesoscale merges with the synoptic scale,
involving migratory high- and low-pressure systems of the lower troposphere with
wavelengths of 1000 to 2500 km. This scale, extending from 1 to 10 days in
time, is also referred to in air pollution literature as the regional scale.
Many of the primary and secondary pollutants, including S02» nitric acid, ozone,
and the constituents of fine particles (sulfate, nitrate, ammonium, organics,
soot) are dispersed within the synoptic or regional scale of 100 to 2500 km and
1 to 10 days.
6-59
-------
The hemisphere scale bridges the synoptic and the global scales. The
hemispheric/global scales of 2,500 to 40,000 km are limited by the circumference
of the earth but unlimited in time. Meteorologically, they include the global
circulation pattern, such as the westerly winds in the Northern Hemisphere, the
equatorial trade winds, and the mean vertical motions of the atmosphere that
rise at the equator and descend at the poles. There are numerous manmade and
natural trace chemical substances whose residence times are greater than the
mean circulation time around the globe of 2 weeks. These substances include
carbon monoxide, carbon dioxide, freons, and methane. If fine particles or
their precursors are dispersed to the upper troposphere (e.g., by thunderstorms),
they may be transported on a global scale.
In the study of the overall effect of a given pollutant, the spatial/
temporal scales of the effect have to be matched to the characteristic scale of
pollutant lifetime in the atmosphere. In the case of the manmade sulfur compounds,
for example, most of the global emissions and depositions are confined to a few
regional "hot spots," such as Central Europe and portions of North America. The
global budgeting of manmade and natural sulfur compounds may indicate marginal
impact when assessed on a global scale. On the other hand, assessing the effect
of S0~ or fine particles only on a scale of tens or hundreds of kilometers may
be incomplete since that scale is small compared with distance transported.
At the International Symposium on Sulfur in the Atmosphere (ISSA Workshop,
1978), it was concluded that the typical atmospheric residence time of SO- is
about 1 day and that of sulfate aerosol is on the order of 3 to 5 days (Figure
6-16A) depending on precipitation patterns. A mean transport of 500 km per day
is common in the Eastern United States (Pack et al., 1978); hence reasonable
scales for S02 and sulfate effects are 500 and 1000 to 2000 km, respectively.
The implication, as seen in Figure 6-16B, is that the ambient sulfate aerosol
6-60
-------
SOJ AEROSOL
r
« S
A
IDRY
I ^7
• <
^\
WET
\ 1 J
•^M
/n
' I ^
^
^ ^
' 1
r
Y/J///////////AS///////////X///////////A////////.
FIRST DAY SECOND DAY THIRD DAY
Fi gure 6-16 . (A) This flow diagram of sulfur transmission through
the atmosphere shows that over half of the S02 is removed or trans-
formed to sulfate within the first day of its atmospheric residence.
Most of the sulfate may remain beyond 4 days of residence in the
atmosphere. (B) Schematic illustration shows the range of transport
of fine particles. The circles have radii of 500, 1000, and 1500 km;
1500 km is considered a reasonable mean transport distance. The
actual transport depends on the synoptic-scale wind field.
-------
concentration at a receptor site can depend on the additive effects of tens or
even hundreds of SOp plumes of different atmospheric ages.
6.4.2 Planetary Boundary Layer
The planetary boundary layer extends from the earth's surface to the
geostrophic wind level; the latter marks the upper limit of frictional influence
of the earth's surface. It is the layer into which heat, moisture, and air
pollutants from man's activities are injected. The planetary boundary layer
averages about 1.5 km in depth and varies from a few tens of meters in stable
conditions to several kilometers in very unstable situations. With decreasing
height, the wind speed tends to fall below the speed of the geostrophic wind
found above the layer, theoretically reaching zero at the surface (Figure 6-17A;
Husar et a!., 1978). The decrease in wind speed with height may not be monotonic,
especially in stable conditions.
The planetary boundary layer is always in a state of change; it is responding
either to a change in wind velocity, temperature, moisture, or pressure field or
to some change in the underlying surface. With respect to air pollution, the
most important question is how the upper limit of the planetary boundary layer
changes, that is, the height of the mixing layer at any given time and location.
Figure 6-18, developed by Smith and Hunt (1978), shows a simple nomogram for
estimating the mixing layer height for daytime conditions based on time of day,
month, cloud cover, and wind speed. The time of day, month, and cloud cover
determine the integrated heat input into the boundary layer. The nomogram also
recognizes wind speed as another source of turbulence. Holzworth (1972) has
documented the variation in mixing heights for National Weather Service rawin-
sonde (balloons from which temperature, wind, and humidity can be derived from
on-board instruments and tracking) sites in the United States.
6-62
-------
0100
0700
1300
1900
10
1400
0400
0 10 0
WIND SPEED, m/wc
1000 1400
10
10
1 I
I
2200
I
290 310 290 310 290 310 290
POTENTIAL TEMPERATURE,°K
310
O
1400
1200
1000
100
600
400
200
0 10 20 30 0 10 20 0 10 20 0 10 20
CHANGE OF WIND DIRECTION WITH HEIGHT, d»jrte»/100 m
1200
E
O 1000
z
i «00
O
S 600
g
* 400
Z
a
S 200
1 l I i I r (_„ «t
MIXING HIIGHT-.
f
»LUME MlICMT //•;' . /•
24 6 8 10 12 14 16 IB 20 22 24
TIME OF DAY
6-17- The wind data obtained from hourly releases by the
RAPS balloon toundmg network are averaged here for July 1976.
(A| Vertical profiles of velocity for 0100. 0700, 1300, and 1900
CDT releases, (Bl vertical profiles of potential temperature (refer-
•need to 1000 mb) and its standard deviation; (C) absolute change of
wind direction with height (degrees per 100 m), (D) schematics of
plume geometry at four different parts of the diurnal cycle. The
horizontal bars are plume height data reported by Forrest and
Newman (1977) for the Labadie plume in the summer.
Source: Hus*r ft al. (1978).
6-63
-------
INTEGRATED HEAT
INPUT FROM DAWN
CLOUD
AMOUNT
(ok tat)
I I I I ' I I I
04 06 08 10 12 14 16 18 20
ANEMOMETER
WIND SPEED ,
AT 10m (m/wc) I
TIME: 14.00 hr
MONTH: JULY
CLOUD: 4/8 THS
WIND: 3m/Mc
Figure 6-18. This nomogram permits estimation of the depth of the
boundary layer in the absence of marked advective effects or basic
changes in weather conditions. The marked example shows how the
nomogram is to be used.
Source: Smith and Hunt (1978).
6-64
-------
The dispersal of pollutants in the vertical dimension is also complex.
Diurnal changes in atmospheric stability lead to the vertical dispersion of
pollutants up to several kilometers during daytime mixing periods and then
isolation of pollutants aloft as low-level nocturnal inversions are formed
(e.g., see Hess and Hicks, 1975). Depending on wind speed, these Isolated
layers can be transported long distances during the night, before daytime mixing
again brings the pollutant into contact with the surface. Figures 6-17A and
6-17D (Husar et al., 1978) illustrate, for example, that within the diurnal
cycle in the St. Louis region, nighttime plume transport is faster than midday
plume transport by a factor of nearly 2 because of variations in the vertical
wind and temperature profiles of the atmosphere. Nocturnal plumes emitted from
tall stacks are affected by 10- to 20-m/sec low-level jets in two ways: the
effective stack height is reduced to approximately the actual stack height and
the distance traveled overnight may be increased by 500 km (Smith et al., 1978).
Such plumes are subject to minimal vertical mixing at night, and it is thus
feasible to follow the plumes with instrumented aircraft for as much as several
hundred kilometers. In the afternoon, the lower wind speed and greater instability
of the atmosphere produce larger effective stack heights and greater vertical
mixing of pollutants emitted into the mixed layer, resulting in greater dilution
than in the nighttime regime. The large effective stack height of afternoon
buoyant plumes has important consequences for sulfur transformations and removal--
that is, the sulfur budget.
Daytime convective activity tends to mix pollutants uniformly within the
planetary boundary layer. For problems involving long-range transport, it is
usually sufficient to assume well-mixed conditions during daytime periods with
Aspect to dispersion, although often not with respect to transformation. When
6-65
-------
the sun sets, the surface cools, convective mixing diminishes and finally
stops, and stable layers of air form. If pollutants are injected into these
stable layers, concentrations can build up. After sunrise, as convection
begins, these pollutants can be mixed to the surface.
This diurnal cycle plays an important role in determining the loss rate of
pollutants from dry deposition at the surface. During the daytime convective
period, near-surface concentrations from ground-level pollutant sources are
reduced because mixing occurs through a large volume. During the night, near-
surface concentrations are increased because the vertical mixing depth is very
restricted. For pollutants emitted from tall stacks, the diurnal patterns of the
surface concentrations are reversed: peak concentrations are observed during
daytime hours, as the result of vigorous downward mixing of the plume material.
At night, plumes from tall stacks are embedded in elevated stable layers that
remain aloft, cut off from the ground by stable inversion layers.
Over water, conditions can differ dramatically, depending on temperature
differences between water and air. During the summer, cold water creates a
stable layer near the surface that prevents dry deposition of pollutants.
Conversely, warm water increases mixing at night so that dry deposition is
increased.
Given the importance of vertical profiles of wind and temperature in the
atmosphere, it is imperative to include these factors in simulation models that
account for variations in vertical mixing height. Profiles of vertical wind
speed, temperature, and moisture are usually determined from rawinsondes.
Typically, rawinsondes are released by the National Weather Service every 12 hr.
Two releases per day are often inadequate to define the changing diurnal pattern
6-66
-------
for a given location. In addition, the vertical resolution of the data near the
surface may not be sufficient for detailed analysis or extrapolations to adjacent
areas of interest.
Wind direction normally changes with height. The effect of vertical wind
shear on air pollutants is to increase the horizontal dispersion about the mean
flow direction; this effect usually is greater under stable nocturnal conditions
than under daytime convective conditions. Under unstable or neutral conditions,
typically daytime, the effect of vertical wind shear in the first 1 to 2 km
above the surface is usually small. During periods of strong convection, however,
pollutants could be taken to heights much greater than 5 km. Once convective
activity begins cloud processes provide mechanisms for lifting polluted air to
great heights and also for chemical transformations in the liquid water of cloud
droplets. Once the cloud evaporates, the remaining sulfur compounds could be
strongly affected by the different wind speeds and directions found at these
great heights, providing greater dispersion relative to near-ground dispersion
processes. The fate of pollutants injected into layers above the planetary
boundary layer must be dealt with in long-range transport studies.
Topographic effects on plume dispersion also require special treatment.
Mountainous terrain increases vertical mixing and causes channel flow through
valleys.
Vertical dispersion is also important in determining the amount of the
pollutant involved in the in-cloud condensation process above the mixing layer
that leads to washout by precipitation. Vertical mixing within air masses,
Particularly along fronts, can permit pollutants to mix to high altitudes,
thereby avoiding precipitation processes. However, such pollutants injected at
high altitudes will ultimately return to the surface at greater distances.
6-67
-------
6.4.3 Horizontal Transport
The main flow of manmade pollutants in the atmosphere is horizontal, whereas
vertical mixing of pollutants is generally confined to the planetary boundary
layer. The measurable extent of horizontal transport of SOp and fine particles
from a given source is about 1000 km; thus, the dispersion of air pollutants
occurs within a shallow layer (a few kilometers) of the earth's atmosphere.
The regional or synoptic-scale transport of air pollutants is most frequently
assessed from computed "air parcel trajectories." The computational methods,
until recently, involved manual determination of the transport wind speed and
direction from synoptic weather charts. The trajectories were of two kinds:
dynamic (from pressure or pressure-height charts) or kinematic (from streamline
and isotach analyses of the observed winds). In the former case, the instantaneous
geostrophic wind was estimated from the pressure charts. With the streamline-
isotach charts, a best estimate of the existing wind was made for the desired
location from the analyses.
With the advent of computers in the meteorological services, programs have
been devised to compute trajectories. Computer methods are still based, however,
on two types of information: observed pressures or observed winds. Some
techniques analyze these fields and interpolate from the irregular observing
network into a regularly spaced grid system from which, by further interpolation,
the transport vectors at any arbitrary point can be evaluated. Other methods
interpolate the observed data directly to the desired trajectory location.
The early work of Durst et al. (1959) is an example of geostrophic com-
putations for the upper atmosphere. A more recent article by Sykes and Hatton
(1976) provides a comprehensive discussion of the two alternative approaches and
6-68
-------
describes an advanced geostrophic computational technique (using recorded
surface pressure observations) applied to the surface boundary layer.
The kinematic approach has been used in the OECD Long Range Transport of
Atmospheric Pollutants (LRTAP) (Ottar, 1976; Eliassen, 1976) employing winds at
the surface and at the 850-mb level (approximately 1.5 km).
Recently, work by Meyers et al. (1977) has extended interpolation as a
means for determining regional flow fields. Their technique uses variational
analysis as a basis for further adjusting the interpolated values in order to
force consistency of the data sets with additional constraints. The constraint
used by Meyers et al. (1977) is conservation of energy and momentum.
Data from which a comparison of calculated trajectories with "real" tra-
jectories can be made are not numerous for the regional scale. The ideal would
be a continuous point source of a unique material with a known emission rate.
Pack et al. (1978) discuss specific data sets available for verification studies
over the short, intermediate, and long ranges.
In a recent experiment in Oklahoma City, Angell et al. (1973) attempted to
determine the accuracy of various approaches to trajectory determination over
the relatively flat Midwestern terrain. Of 55 tetroon flights, return tags were
received from 27, 13 of which traveled a distance considered suitable for
analysis (Hoecker, 1977). Twelve of the recoveries were from more than 500 km
from the launch site, and three were from more than 1000 km.
The Oklahoma City tetroon trajectories, while few, agree qualitatively with
other, more numerous observations and illustrate the systematic differences in
trajectory directions created by different synoptic situations. Taken together,
it is evident that in many cases trajectory estimates perform well. They are at
their best in relating air quality data at isolated locations to pollution
transported from large area sources.
6-69
-------
A second general approach, especially for anticyclone-induced stagnation
episodes, can be inferred from a series of recent reports (e.g., Ferber et al.,
1976; Husar et al., 1976a, 1976b). These reports show the coherence of dew
point, visibility, and ozone respectively, within well-defined regions, generally
when under the influence of high-pressure areas. Vukovich et al. (1977) have
calculated the "residence time" within (admittedly idealized) anticyclones of
different shapes and moving with different speeds. Reconstruction of the paths
of anticyclones is easier than the reconstruction of multiple trajectories from
the various pollutant sources. Further, prediction of the movement of these
systems is one of the more accurate aspects of modern weather forecasting. When
large-area estimates are needed, especially for secondary pollutants, this use
of the trajectory of a large, easily indentifiable entity as a moving, changing,
but calculable air mass should be investigated and the technique tested on
actual weather systems.
Finally, the best approach would be the quantitative study of known amounts
of material placed into the atmosphere specifically to study transport and
diffusion. However, no tracer currently meets the criteria for a long-range
transport tracer. Particles and reactive gases are nonconservative. Many other
candidates, such as sulfur hexafluoride, already are present in the air in
concentrations high enough to impose very high release amounts to overcome the
background and dilution over long travel distances. Radioactive gases present
environmental difficulties.
6.5 SOURCE-RECEPTOR RELATIONSHIPS
The relationship between emission sources and ambient concentrations at
receptor sites is determined by transmission processes in the atmosphere:
horizontal and vertical transport, transformation, and removal. Transformation
6-70
-------
and removal rates for a given pollutant are competitive, and both depend on
horizontal and vertical transport. Hence, to completely understand the source-
receptor relationship, the specific role and interaction of each of the three
processes must be quantitatively evaluated.
In this section, five different approaches to source-impact Identification
are discussed. The first three are receptor-oriented empirical methods. They
use existing monitoring data on aerosol or S0? concentrations at various receptor
sites in conjunction with other relevant data such as the aerosol chemical
composition (section 6.5.1), historical trends of emissions and ambient con-
centrations (section 6.5.2), or the direction from which the pollutants are
coming (section 6.5.3) as clues for their probable origin. The remaining two
methods begin at the source and examine the pollutant transmission processes
through the atmosphere and their impact at a receptor by: (1) field observations
(section 6.5.4) and (2) diagnostic modeling, that is, the use of source data,
ambient concentration data, and a model to clarify what happen between source
and receptor (section 6.5.5). Several of these methods can only provide circum-
stantial evidence for the source-receptor relationships. Other approaches may
provide causal evidence but may impose heavy demands on sparse or nonexistent
data. Hence, it is evident that establishing the source-emission relationships
for aerosols over the various regions of the United States requires prudent use
of all the available source resolution techniques as well as new ones as they
are developed.
6.5.1 Aerosol Chemical Composition
A knowledge of the chemical composition of aerosols is essential to under-
standing their effects. The chemical composition serves also as a tracer of
their probable origin.
6-71
-------
In the chemical mass balance method, the chemical composition of an ambient
aerosol is used to trace its origin (Friedlander, 1973; Miller et al., 1972;
Heisler et al., 1973; Gartrell and Friedlander, 1975). Characteristic tracer
elements (e.g., vanadium, which comes primarily from residual oil) can be used
to estimate the contribution of specific sources.
The first comprehensive study of the chemical composition of various size
fractions of the haze aerosol was conducted in the Los Angeles air basin as part
of the 1969 Pasadena Aerosol Study (Hidy, 1972). This was followed by project
ACHEX (Hidy, 1975). For measured aerosol concentrations at seven locations in
m*
the Los Angeles basin, White and Roberts (W7-5-) constructed a chemical mass
fjfd
balance (Figure 6-19; White and Roberts, i9>5). The key contributing species
for the total aerosol mass concentration were nitrates, sulfates, organics, and
unidentified substances. On the basis of statistical analysis of b.0.. and
scat
chemical composition data, they concluded that sulfates are the most efficient
scatterers among the measured chemical species (see chapter 9). Oil and gasoline
combustion products, and their oxidation products, appear to dominate the Los
Angeles aerosol.
A separate chemical-mass balance for fine and coarse particles collected in
Charleston, WV, was reported by Lewis and Macias (1980) and is shown in Figure
6-20. Of the fine-particle mass concentration (33 pg/m ) about 30 percent was
sulfate, about 13 percent was ammonium ion, 18 percent was carbonaceous material;
the rest consisted of trace constituents and undetermined chemical species. The
coarse-particle mass concentration of 27 yg/m was found to be largely associated
with elements that occur in the earth's crust. Lewis and Macias concluded that
some elements were found mainly in the fine fraction; other elements were found
in the coarse fraction. Carbon was the only major element that was almost
equally distributed between the two size fractions.
6-72
-------
PASADENA
cn
i
CO
COVINA
195 "
OOMINGUEZ
HILLS 118
• NITRATES
O SULFATES
• OHGANICS
0 OTHER
!/•< \
. \
NITRATES
• SULFATES
• ORGANICS
O OTHER
WEST
COVINAX
195
DOMINGUEZ
HILLS
118
Fi gure 6-19. Sulfates evidently contribute only about 25 percent of the total mass but cause about half
of the light scattering, as shown in these maps of mass concentration (A) and light-scattering coefficient
(B) in the Los Angeles basin. The pie diagrams show the relative contributions of nitrates, sulfates,
organics, and other compounds. The contributions of source types (oil, gasoline, and other) to the mass
concentration (C) and light-scattering coefficient (D) are also shown.
Source: White and Roberts
-------
FINE PARTICLES
MASS •= 33.4
COARSE PARTICLES
MASS = 27.1 jjg/m3
VALUES IN PERCENT
Fi gure 6-20. Compositions of fine and coarse particles collected in Charleston. WV, are shown in these
pie diagrams.
Source: Lewis and Macias
-------
For several years an extensive air pollution monitoring program was con-
ducted in St. Louis as part of project RAPS (Regional Air Pollution Study).
Size-segregated aerosol samples (Stevens et al., 1978) were collected auto-
matically with dichotomous samplers and analyzed for elemental composition and
mass concentration of fine and coarse particles.
The results from the 10-station monitoring network (Figure 6-21; Dzubay,
1979), analyzed by Dzubay (1979), show the distribution of the aerosol species
within the St. Louis metropolitan area. Here again, sulfur compounds were the
dominant species in the fine-aerosol fraction, contributing about 60 percent of
the mass. For stations 103 through 112, within the central city, primary motor
vehicle contribution was estimated to be about 10 percent of the fine-particle
mass. For the peripheral stations, 25 km away from the city center (stations
122 and 124), the sulfate concentrations were comparable to those within the
city, but motor vehicles accounted for only a few percent of the fine-particle
mass. It is thus concluded that the automobile contributions were of local
origin while the sulfate is distributed regionally, and the addition of the
sulfate by the St. Louis metropolitan area is only 10 to 20 percent over the
"regional background." In his mass balance analysis, Dzubay (1979) was able to
account for practically all the coarse-particle mass as crustal shale and limestone.
Figures 6-21A and 6-21B (Dzubay, 1979) also contain mass and chemical composition
data for a Smoky Mountain site at Elkmount, Tenn. The fine-particle mass concentration
o
there was about 25 yg/m , comparable to the values in the St. Louis region
outside the city. There again, about 60 percent of the fine-particle mass was
contributed by sulfur compounds. The unknown compounds may have included carbonaceous
compounds, nitrates, and water. The coarse-particle mass concentration at the
Smoky Mountain site was only 6 pg/m , substantially below that in the St. Louis
region.
6-75
-------
103 105 106 108 112 IIS 118
SAMPLING SITE
1JO 122 124 SMOKY
MIS
103 105 106 108 112 115 118 120 122 124 SMOKY
Fl gure 6-21 . Sourci resolution of the St. Louis aerosol shows that less than 10 percent of the fine
particle mass is due to primary automobile emissions; about 60 percent is from sulfur oxide sources
Almost ill of the coarse-panicle mass n accounied for by dust from the eanh's crust.
Source: Dzubay (1979).
6-76
-------
Sulfur compounds constitute the most significant chemical component of the
fine-particle mass over the Eastern United States. Statistical evidence for the
importance of sulfates in the fine-particle mass was found from the analysis of
RAPS data by Loo et al. (1978) (Figure 6-22). The particulate sulfur was con-
sistently about 10 to 15 percent of the fine-particle mass for a sampling period
of 2 months. Similar observations were made by Lewis and Macias (1980) for
Charleston, WV, and by numerous other investigators over the past decade.
There is great concern over the origin of the fine-particle haze in the
lightly populated areas of the Western and Southwestern United States and the
potential impact of future emissions in that region. Reporting the results of
project VISTTA, Macias et al. (1979) have presented size-chemical composition
data for size-segregated aerosol collected in the Four Corners area during
aircraft flights (Figure 6-23; Macias et al., 1979). As anticipated, the
coarse-particle fraction could be accounted for by the crustal element contributions.
In the fine-particle mass balance, about 40 percent of the 5.3 pg/m was sulfate,
another 10 percent was trace constituents, and 22 percent could not be accounted
for. However, Macias et al. (1979) have also reported a 29 percent contribution
of silica to the fine-particle mass. This is unusual because the crustal elements
normally accompanying silicon were not present in the fine-particle samples.
Macias et al. (1979) argued, therefore, that the fine-particle silicon may be
due to direct emissions from combustion sources.
More detailed size distribution data for sulfur, silicon, and other com-
pounds were reported by Winchester et al. (1979) for sites in New Mexico,
Colorado, St. Louis, and New Hampshire. Invariably, the particulate sulfur
concentration was highest for impactor stages between 0.25 and 1.0 ym. The size
distribution of silicon, on the other hand, was consistently the highest for
6-77
-------
I
I
II •*
I ,
, inn1;' i,
1 ' . .1 I
I
I
' 'si' -
is;;,"
• •••
•ViH • I
ll I ll ' '
11111 1 '
l"l
I '»
I
20 40 60 80
MASS (F),
Figure 6-22. In the St. Louis area, monitoring data for about a 2-
month period shows that sulfur was about 10 to 12 percent (calcu-
lated as elemental sulfur) of the fine-particle mass.
Source: Loo et al. (1978).
6-78
-------
80.
s,o2
61%
COARSE PARTICLES
MASS » 4.0 jjg/m3
SO
C*0.
FINE PARTICLES
MASS - 5.3
Fi gure 6-23 . Composition of fine and coarse particles collected dur-
ing flights in the Four Corners region is presented in pie diagrams.
The total aerosol mass was estimated from m situ size distribution
measurements. In this data set. silica accounted for an estimated 29
percent of the fine-particle mass.
Source: Macias et al. (1979).
6-79
-------
impactor stages above 2 ym and lowest for the stages below 1 ym. Hence, in the
Winchester et al. (1979) data set, the fine-particle silicon content was practically
negligible, even at the New Mexico and the Colorado sampling sites. The resolu-
tion of the difference between the data reported by Macias et al. (1979) and
Winchester et al. (1979) requires further data from the western States.
Factor analysis has been applied recently to data on the chemical composi-
tion of aerosols to determine which clusters of elements have high temporal
correlations. The high correlation of a set of elements may indicate a common
source.
In summary, the chemical composition of light-scattering aerosols provides
a valuable, if not the most important, currently available clue to their sources.
Sulfur compounds contribute about half of the fine-particle mass over most of
the continental United States; they occur generally in the most effective light-
scattering size range, 0.4 to 0.7 ym.
6.5.2 Historical Analysis of Trends in Emission Concentrations
The relationship between sources and their effects on ambient concentrations
can also be inferred from historical trend analysis of both. If the 10- to 20-
year trend of concentrations exhibits the same patterns as does the emission of
some pollutant, then a possible cause-effect relationship may be inferred. On
a shorter time scale, such as during the 1967-68 copper smelter strike in the
Southwest, the complete shutdown of a major source can yield valuable clues to
that source's role in the deterioration of air quality.
6.5.3 Pollution Roses and Sector Analysis
The estimation of source-receptor relationships by means of "pollution
roses" has been used successfully for decades in the case of primary pollutants,
such as carbon monoxide. In its simplest form, the method consists of classify-
ing each pollutant measurement according to the corresponding wind direction and
6-80
-------
computing the average pollutant concentration for each wind direction class.
The plot of average concentration versus wind direction is referred to as a
pollutant rose; with careful selection of the wind direction classes, it is
possible to infer the individual effect of local sources. The major assumption
required in such analysis is that the plume arrives at the receptor from the
direction in which the source lies.
Wind rose analysis was applied by Latimer et al. (1978) to historical
visibility data from Farmington, N.M. (Figure 6-24; Latimer et al., 1978). The
percentage of daylight observations for which RH < 60 percent and visual range >
121 km was chosen rather than the mean. The visual range was significantly
improved for the south-southeast to west wind direction classes during the
shutdown of the copper smelter, lying in the same directions at distances of
more than 400 km. Thus it may be inferred that the smelters cause up to 85
percent of the reduction of visual range below 121 km associated with south-
southeast winds.
As noted by Altshuller (1976), long-range transport in the Eastern United
States accounts for at least 50 to 75 percent of the measured sulfate. Since
sulfate appears to cause more than its mass share of light scattering, most of
the observed haziness may be unrelated to local sources. During long-range
transport, the plume may meander and arrive at the receptor from almost any
direction. Thus, in the Eastern United States, the traditional pollution rose
"iay be inadequate for determining the sources of haziness.
The utility of the sulfate rose in determining the source-receptor relation-
ship may be improved by the more sophisticated approach of trajectory sector
a"alysis. Backward air parcel trajectories are calculated to determine the
source region that most strongly contributes to the measured concentration. The
6-81
-------
">
z
2 S
100
<
>
V) LU
CD U
o x
O
2
80
60
40 O-
20
o S
T I I I I
I I I I I I I
— 1957-1966
— — 1967-1976
O—OJULY 1967-MARCH 1968 (COPPER STRIKE)
D
I I
WIND DIRECTIONS ASSOCIATED WITH SIGNIFICANTLY.
IMPROVED VISUAL RANGE (AT THE 95% CONFIDENCE
LEVEL) WHEN COMPARED WITH THE REMAINDER OF
THE DECADE 1967-1976
I J J J J J I I I I I I
£ O N. NNE.NE. ENE. E. ESE. SE. SSE. S. SSW. SW. WSW. W. WNW NW. NNV\
u.
WIND DIRECTION
Fi gure 6-24. This plot of the percentage of daylight observations with
RH < 60 percent for which visual range was > 121 km as a function
of wind direction shows that during the period of the copper strike,
there was a significant improvement of visual range from the direc-
tion of copper smelters, SSW. to W., implicating these SOj sources.
Data are from Farmington, NM.
Source: Latimer et al. (1978).
6-82
-------
direction from the receptor to the source may then be used in place of the local
wind direction in constructing the pollution rose. Samson (1978) used this
approach to establish the importance of the Ohio River Valley sources of sulfate
at nonurban sites in New York State. Galvin et al. (1978) showed that in New
York State the elevated sulfate concentrations were consistently associated with
trajectories that originated in the Ohio River Valley region (Figure 6-25;
Galvin et al., 1978). Chung (1978) used trajectory analysis to implicate the
same region as an important source of sulfate in southwestern Canada. Rodhe et
al. (1972), Brosset et al. (1975), and others established the importance of
continental European sources by this technique. The OECD Program on the Long
Range Transport of Air Pollutants (LRTAP) included trajectory sector analysis of
sites across Europe (Figure 6-26; OECD, 1977). Thus, the relative contribution
of each region to any site is clearly established. For example, it appears that
most of the sulfate (and thus, probably, most of the haze) received in southern
Sweden is from sources in Central Europe.
The more elaborate trajectory analysis techniques are easily adapted to
include simple gas-particle conversion and removal kinetics along the trajectory.
Such models are used to extract the regional-average rate constants from source
emissions and measured concentrations, as in the OECD project. Such empirical
approaches to data analysis are known as diagnostic models and are discussed
further in section 6.5.5.
In summary, observed measurements of aerosol and haziness can be attributed
to sources or source regions when the meteorological transport between source
and receptor is known. For conditions in which long-range transport and unsteady
winds are significant, the utility of pollution roses may be increased by
receptor-back-to-source trajectory computations.
6-83
-------
WHITJFACE 7/W76
Fl
WMIT|F»Ci
..
- '—...•• ..,-'' liv .v
TIME OF ARRIVAL IDT — 1 00 • m 4 00 t m
^~^~ § 00 • m -•-. 10 00 p m
TIME 0' AHKIVAl. (DT
: oo.m -«BO»-
> I 00 • m .... 10 00 e «
_ 6-^D . Air mats traiectories to Whitetace, Schohane. and Holland on JU!Y 10 and 22. 1976, ind'
cate tnat elevated suHate at nonurban New York sites originated in tt\e Ohio River Valley
Source: Galvin et al. (1978).
6-84
-------
0 10 20 30
I I I I
Figure 6-26. Sulfate trajectory roses in Europe reveal that most of
the sulfate in southern Scandinavia is from emissions in other coun-
tries lying to the south, rather than from local contributions.
Source: OECD (1977).
6-85
-------
6.5.4 Direct Measurements of Source Impacts: Plumes and Regional Hazes
In the previous three sections, the source-receptor relationships were
examined from the point of view of the receptor. The approach was to determine
what source types contributed and how much they contributed to the total burden
at a given site. An alternative approach discussed in this section 1s that of
starting at the source and following the transmission of the air pollutants
through the atmosphere until they are ultimately removed. This more fundamental
approach requires a rather detailed understanding of the chemical transformation
processes, the removal processes, and the horizontal and vertical transport of
matter in the atmosphere. In this approach, the specific roles of transport,
transformation, and removal processes are identified. This approach facilitates
consideration of these processes in the appropriate control strategies.
This section focuses almost exclusively on the transmission of sulfur
compounds through the atmosphere. The empirical source-receptor methods of the
preceding sections have clearly implicated sulfates as the principal components
of the fine-particle aerosol population, contributing perhaps half or more of
the fine-particle mass over a large part of the continental United States. The
second reason for focusing on sulfur compounds is that the origin of the nonsulfur
compounds in the aerosol phase, such as nitrates and organics, is not as well
understood.
6.5.4.1 Power Plant Plume Studies—The actual rates and mechanisms of conversion
as well as the creation of light-scattering aerosols must be studied in field
experiments. In recent years, a number of power plant plume conversion studies
were specifically designed to study the SCL oxidation rate and the formation of
sulfate and light-scattering aerosols.
6-86
-------
Since the late 1960's, an increasing fraction of total SCL emission has resulted
from coal combustion in electric power plants equipped with stacks 150 to 200 m
high.
The atmospheric transmission of tall-stack effluents has been studied
extensively during the past decade, most recently in project MISTT (Midwest
Interstate Sulfur Transformation and Transport) (Wilson, 1978) and STATE (Sulfur
Transport and Transformation in the Environment) (Schurmoior. 1980), where the
transport, chemical transformation, removal, and the interaction of these
processes in determining the sulfur budget of large plumes were assessed. One
set of results and conclusions of these studies is given in Figures 6-1 (Husar
et al., 1978) and 6-27.
It is also important to examine the long-distance effects of the Labadie
power plant plume. Figure 6-28 illustrates the plume geometry and the measured
SOp concentrations attributed to the Labadie plume for two long-range sampling
days, July 9 and July 18, 1976 (Gillani et al., 1978). On both days the plumes
were tracked to about 300 km from the source. The properties of the aging plume
are shown in Figure 6-29 (Gillani et al., 1978).
6.5.4.2 Urban Plumes—Urban plumes aggregate from various sources within a
metropolitan area. The best-studied urban plume is that of metropolitan St.
Louis. The individual sources include coal-fired power plants with a combined
capacity of 4600 MWe, oil refineries with a combined capacity of 4.4 x 10
barrels per day, various other industries, and a population of about 2 million
(White et al., 1976). Because St. Louis is far from other major metropolitan
areas, its impact on the surrounding ambient air quality is relatively easy to
identify; air that has been modified by the aggregate emissions of the metropolitan
6-87
-------
10 12 14 16
TIME OF DAY
20 22 24
Fi gure 6-27. The amount of paniculate sulfur formed increases when
the plume is removed from the surface by dilution or by decoupling
from the surface layer. Hence, daytime emissions into deep!-, mixed
layers or elevated stable layers are expected to produce more sulfate
than nighttime emissions.
6-f
-------
I
00
Figure b-28. Horizontal profiles of SO, were made during selected constant altitude traverses on July 9
and July 18, 1976. July 9 traverses are at about 450 m AGL, and July 18 traverses are at about 750 m
AGL. The Labadie plume sections are shaded. Also shown are backward trajectories for the Labadie
plume.
-------
3
<
J
HEIGH1
2400
2000
1600
1200
800
400
I
—
I I
GASEOUS
MIXING HEIGHT
^
^^
I
0 50
2400,
2000
1600
1200
800
400
Q
I
"—
^^
gt*""' %
t
L--'
a.
^
100 150
km
1 1
PARTICULATE
_ MIXING HEIGHT
*
^^
^
\
I
0 50
2400
2000
1600
1200
800
400
0
I
^^~~ »
i i
4 1
1 1
1 1
100 150
1 1
1
SULFUR _
1000
> 1 —
* %
mg/m'
i "^
i STABLE
^1
200 250
1
SULFUR —
r100,-
^^^^^t
mfl/m —
JiAe:
^ 1
200 250
1
— LIGHT-SCATTERING COEFFICIENT —
— MIXING HEIGHT
X
'-"\
V ^
I
: \
2.00 —
1 |
P J*T*ST"ABLE
2400
2000
1600
1200
800
400
o
r- '— r 1 -i
MIXING HEIGHT ' '
— ^ GASEOUS SULFUR _
J^1
—
t
I ,
YN.
\ ,
V
L
•
•
-— •
0 50 100 150
km
1000 __
mg/m2_
— 1 STABL?
200 2!
MIXING HEIGHT
I I
PARTICULATE SULFUR —
100
km
50 100 150 200
DOWNWIND DISTANCE, km
250
14UU
2000
1600
1200
800
400
I 1 I 1
^MIXING HEIGHT1 ucHT-SCATTERING-
^ «*•*
_ "1
-
-
\
|
H
•
1
|
1 I
.^ COEFFICIENT
\
1
1
! K
» 2.00 ~
\ ^ -
^— J-t STABL?
0 50 100 150 200 25
DOWNWIND DISTANCE, km
0
1
1
30 65
1 1
PLUME AGE, hr
•1100- 1530-
1200 1630
1 1
85
1 1
2000-
2100
1 1
0
1
1
20 60
PLUME AGE.hr
1500- 1530-
1600 1630
I I
90
I J
1800-
1930
I J
APPROXIMATE TIME OF SAMPLING
APPROXIMATE TIME OF SAMPLING
Fi gure 6-29 . Vertical profiles show the crosswind integrals of excess plume concentrations of gaseous
sulfur (Sg), particulate sulfur (Sp), and light-scattering coefficient (bscat) for the transport of the
Labadie plume of July 9 and 18, 1976. Broken portions of the vertical profiles represent extrapolations.
Vertical integrals of the profiles of Sg and Sp denote masses Mg and Mp, respectively, of gaseous and
particulate sulfur in vertical plume slabs of unit downwind depth. The region marked "stable" repre-
sents the inversion layer forming near the ground during the evening.
Source: Gillani et al. (1978).
6-90
-------
area forms an "urban plume downwind." The Fate of Atmospheric Pollutants Study
(FAPS) (e.g., Haagenson and Morris, 1974) has shown that this plume can often be
identified as far as 120 km from the city.
As part of project MISTT (Wilson, 1978), the three-dimensional flow of
aerosols and trace gases in the St. Louis urban plume was studied. The plume
was successfully tracked to a distance of 240 km, and it was mapped quantitatively
to 160 km (Figure 6-30; Wilson, 1978). At these distances, the plume was still
defined and about 50 km wide.
An increased concentration of light-scattering aerosol was a conspicuous
characteristic of the St. Louis urban plume. The primary contribution of project
MISTT was to quantify the flow of materials at increasing downwind distances so
as to study the transformations of pollutants in the atmosphere.
The flux of ozone, b.,,.. , and particulate sulfur (S) all increased with
5Cal P
distance downwind of St. Louis on July 18, 1975, reflecting the secondary origin
of ozone and most of the light-scattering aerosols (White et al., 1976). Most
of the increase in the b_ ,. flux was observed downwind of the major increase in
SCa L
ozone flux, which is consistent with the finding of laboratory studies that
aerosol production lags behind ozone production in a photochemical system
(Wilson et al., 1973). The ratio of the flux of b,_a. to the flux of So indicates
SCa t p
that sulfate compounds accounted for most of the newly formed aerosol in the
urban plume. This case study illustrates the fact that emissions in urban
plumes from a metropolitan area such as St. Louis cause reduced visibility and
elevated ozone concentrations long after the primary precursors have been diluted
to low concentrations.
The visibility reduction in the St. Louis urban plume was also studied as
part of project METROMEX. Komp and Auer (1978) have made direct observations of
6-91
-------
CMAMFAlOtt
en
vo
ro
OZONE
AEROSOLS AS bscat
KILOMETERS
Figure 6-30. The three-dimensional flow of aerosols and ozone in the St. Louis plume
is plotted for a distance of 240 km.
Source: Wilson (1978).
-------
visual range from aircraft downwind of St. Louis and presented the data as
contour maps of aging in the urban plume. The visual range was reduced by a
factor of 2.
These direct measurements of the impact of a single major source area on
its surrounding region show that the urban plume of one metropolitan area may
become a "background haziness" for another. It is therefore reasonable to
assume that the pollutant backgrounds in nonurban regions of the Eastern United
States are probably the result of the combined emissions of metropolitan and
industrial areas as well as of power plants far upwind. Regrettably, direct
measurements of plume impacts have not been made beyond a few hundred kilometers
from their sources. Beyond those distances, aircraft mapping of single plumes
and even urban plumes tends to become difficult; furthermore, plumes are likely
to be superimposed. It has been suggested that the observed large-scale haziness
over a large part of the Eastern United States during stagnation is in fact the
result of the superimposition of numerous urban, industrial, and power-plant
plumes, as discussed in the next section.
;6.5.4.3 Regional-Scale Episodes of Haziness—The past several decades have seen
a shift in the spatial scale of particulate matter distribution. The scale has
expanded from areas immediately surrounding each emission source to entire
metropolitan areas, and to multistate regions. One reason for the lessening of
local high concentrations has been the control of primary particulate emissions.
'However, the emission of precursor gases such as SO,, hydrocarbons, and NO has
L. X
Increased over this period.
Early work on the potential for air pollution episodes focused on local
'"version situations rather than stagnation of a regional-scale air mass.
(1961) found that the southeast coast of the United States and the Smoky
6-93
-------
Mountain region had the highest incidence of surface inversions nationally.
Holzworth (1967) evaluated the pollution potential of several U.S. urban areas
on the basis of mean mixing heights and wind speeds. Korshover (1967) constructed
a climatology of stagnating anticyclones in the Eastern United States from 1936
to 1965. He found the highest potential for large-scale stagnation east of the
Rocky Mountains to be in the southern Appalachian and Smoky Mountain regions.
One of the earliest examinations of a regional-scale episode detailed the
evolution and impact of the Thanksgiving 1966 episode over the eastern seaboard,
in which a stagnating high-pressure system caused prolonged elevated regional
concentrations of primary pollutants.
In the past decade, several investigators have presented evidence for long-
range transport of visibility-reducing fine particles. Rodhe et al. (1972),
Brosset et al. (1975), Eliassen and Saltbones (1975), Smith and Jeffrey (1975),
and Szepesi (1978) are among those who have confirmed the existence of long-
range transport of sulfates in Europe. Chung (1978) found evidence of the
transport of sulfate from the northeastern Midwest of the United States into
southeastern Canada.
In the United States, Altshuller (1976) noted the anomaly of decreasing
urban S02 concentrations with increasing rural sulfate trends (Table 6-6) and
proposed long-range transport as the cause of the increase in regional sulfate
levels. Wilson et al. (1976) also proposed long-range transport as the cause of
the increased regional sulfate concentrations on the basis of field studies of
plume transmission. Gage et al. (1977) demonstrated that a high sulfate-episode
potential occurs when multiple S0? sources lie along a given wind direction.
One of the earliest case studies of transport of large-scale hazy air
masses was that of Hall (1973). Since about 1975, the evolution and transport
6-94
-------
en
TABLE 6-6. THREE-YEAR RUNNING AVERAGE OF SULFATE BY GEOGRAPHICAL
REGION FOR NONURBAN SITES IN THE UNITED STATES
Region
East Coast
Southeast
Eastern Midwest
Western Midwest
Southwest
Rocky Mountain
West coast
1965-67
8.1
5.4
7.3
3.1
3.5
1.7
3.0
1966-68
8.2
6.1
7.8
2.7
3.3
1.2
3.0
1967-69
8.4
7.2
8.3
3.2
4.2
1.5
3.2
1968-70
(7.0)a
9.5
5.2
1.8
3.7
1969-71
(7.1)3
8.4
9.8
3.7
5.8
2.1
3.8
1970-72 1971-73
(7.7)a (8.9)a
8.1
10.4 10.9
3.6
5.6 5.0
2.1
Computations based on smaller number of sites.
-------
of regional-scale hazy air masses has received increasing attention from many
research groups: long et al. (1976), Husar et al. (1976a), Lyons and Husar
(1976), Wolff et al. (1977), Samson and Ragland (1977), Vukovich et al. (1977),
Galvin et al. (1978), and Hidy et al. (1978).
A common finding in recent studies is that formation of regional-scale
haziness is usually associated with the presence of slow-moving high-pressure
systems. Since precipitation is relatively infrequent in anticyclonic systems,
the residence time of fine particles may be increased to a week or more.
An example of one such episode over a 2-week period in June-July 1975 is
presented in Figure 6-31 (Husar et al., 1976a). Inspection of the sequence of
contour maps reveals that multistate regions are covered by a hazy layer in
which noon visual range is less than 10 km (b .= 4 x 10 m ), indicated by the
outer contours.
The apparent motion of the haze was confirmed by geostationary satellite
photographs. The region of haziness on July 30 is clearly seen in Figure 6-32
(Lyons and Husar, 1976), which shows the hazy air mass over Arkansas, Missouri,
Kansas, Iowa, and Minnesota.
Two passages of the hazy air mass over St. Louis resulted in sharp increases
of b . over the entire metropolitan region (Figure 6-33A; Husar et al.,
1976a). Sulfate concentrations also increased during the haze episode, from
about 9 to 33 yg/m . The spatial coherence of the haziness is seen in the
correspondence between the extinction coefficient at St. Louis and that at
Springfield (Figure 6-33B; Husar et al., 1976a), again confirming that the
observed haziness was primarily due to inflow into the polluted "background"
material of the hazy air mass rather than local contributions.
Sufficient sulfate data were available from the National Aerometric Data
Bank for comparison with visibility on 2 days during the episode period.
6-96
-------
CTi
Fi gure 6-31. Sequential contour maps of noon visibility for the period June 25 to July 5, 1975, illus-
trate the evolution and transport of a large-scale hazy air mass.
Source: Husar et al.
-------
Fi gure 6-32. On June 30, 1975. the hazy air mass covered large parts of Arkansas, Missouri. Kansss,
Jowa. and Minnesota, as seen in this geostationary satellite photograph.
Source: Lyons and Husar (1976).
6-98
-------
— 40
E
*r
o
6/23 6/21 6/25 6/26 6/27 6,28 629 6/30 7/1 7/2 73 74
Fl
6-33 . Local monitoring data in the St. Louis. MO, area during the June-July 1975 haziness
•pisode (A) Light-scattering coefficient (btcal) recorded at three widely spaced locations in the Si
Louis metropolitan area and daily average sulfate concentrations. (Bi extinction coefficients (u€X.>
obtained from visibility observations at St. Louis. MO. and Springfield. IL. 150 km apart
Source: Husar et at. (1976).
6-99
-------
Figure 6-34 (Husar et al., 1976a) indicates the substantial correspondence
between highest sulfate and lowest visibility.
During this period, the visual air quality was beyond the control of any
local jurisdiction. The Alabama Air Pollution Control Commission reported the
following (Bulletin of the AAPCC, 1975):
During the weekend of July 5, 1975, a heavy haze layer enveloped
the state of Alabama and much of the Southeastern United States.
At that time, the AAPCC technical staff received many comments from
the public concerning the origin and composition of the haze.
The National Weather Service in Birmingham did issue an air
stagnation advisory (ASA) for Alabama for this same time period;
however, the traditional pollutant measurements made by the AAPCC and
local programs did not show excessive levels. In fact, the measured
local levels were lower than had been measured under previous ASA's
making the dramatic decrease in visibility more intriguing.
Husar et al. (1976b) reported that in June-August 1975 there were at least
six episodes similar to the one above. The work of other investigators confirms
that episodes of regional-scale hazy air masses are not rare in the Eastern
United States. At present, however, only the qualitative features of such
episodes are understood: the observed effect on visibility, secondary sulfate
and ozone composition, and the apparent motion of the haze.
Important questions remain to be answered about regional-scale episodes of
haziness, including the following:
(1) Do the hazy air mass and the meteorologically defined anticyclone
completely coincide?
(2) What are the quantitative effects of superimposing multiple SCL
plumes and urban reactive plumes?
(3) What are the effects (and possible feedback mechanisms) of high
pollutant concentrations on dry and wet deposition and cloudiness?
6-100
-------
cn
o
JUNE 23. 1975
JUNE 23. 1975
LJ 6 »8
•i ••
•••-*'V_^,J I \
SULFATE \.*
CONCENTRATION.
/I g/m3
•i • »
U.J TO to 30
cb
-------
(4) What is the exact residence time of the fine particles in the atmosphere
during such episodes? It may be, for example, that lack of precipitation
leads to extremely long sulfate lifetime.
The current Sulfate Regional Experiment (SURE), sponsored by the Electric
Power Research Institute, is yielding valuable information about sulfur transmission
in the Eastern United States. The U.S. Environmental Protection Agency's Sulfur
Transformation and Transport in the Environment (STATE) is directed toward the
complicated source-receptor relationship. The upcoming Prolonged Elevated
Pollution Episode (PEPE) project of STATE is specifically designed to sample
such regional-scale episodes of haziness from their inception throughout their
residence over the Eastern United States.
The Eastern United States has experienced the most severe episodes of
manmade haziness to date because the sources of precursor gases are concentrated
in that area. Empirical evidence indicates that manmade sulfate and nitrate
are also important factors in the visual air quality of the Western and Southwestern
United States. Figure 6-35 (Holzworth, 1972) reveals that the meteorological
potential for air pollution or haziness for the Western United States is as high
or higher than that for the Eastern United States. Since the frequency of
precipitation is low in the western regions, the residence time and the overall
visual impact of the same emissions could be greater there than in the East.
It is not currently known whether regional haze occurs in the Western and
Southwestern United States. Extinction coefficients for remote western areas
-4 -1
may be as low as 0.2 x 10 m ; a haze episode thus may require very little
mass to cause a significant deterioration of visibility. For example, only
about 4 yg/m of fine-particle mass is needed to produce an increase in the
6-102
-------
CTi
I
O
U)
Figure o-3b. Isopleths of total number of forecast days of high meteorological potential for air pollu-
tion in a 5-year period demonstrate that the potential for regional-scale manmade haziness is at least as
high in the Western United States as in the Eastern.
Source: Holzworth (1972).
-------
-4 -1
extinction coefficient of 0.2 x 10 m . In pristine Western areas, this mass
could double the extinction coefficient and reduce the visual range by half,
from about 240 km to 120 km (Bachman, 1979).
6.5.5 Diagnostic Models
Knowledge of the transmission from source to receptor completes quantification
of the source-receptor relationship (see chapter 5). Atmospheric transformations
and removal during transport are the key processes involving SO and particles,
/\
particularly for the secondary fine-particle species (i.e., sulfate, nitrate,
and organics). These processes entirely determine the impact of a source.
The kinetics of atmospheric transmission cannot in general be measured
directly because of the interference caused by removal. However, the available
data on emissions, trajectories, and resulting concentrations can be processed
through a model incorporating the key processes. The unknown rate constants are
then extracted by adjusting the rate constants until a good fit is obtained
between the model and the observed data. The mathematical formulation used for
this purpose is referred to as a diagnostic model.
One of the best-known applications of a diagnostic model is the analysis of
the OECD monitoring data (OECD, 1977). An emission inventory and transmission
conditions for Europe were used in the model. The rates of gas-to-particle
conversion and of removal were then extracted by tuning these parameters until
the best fit between calculated and observed concentrations was achieved. The
results for sulfur transmission from the OECD study are listed in Table 6-7.
The year-round average conversion rate of 1-2 percent per hour and the
overall average dry-removal rate of 3-4 percent per hour were major new results
of the OECD program. Studies being conducted in the United States with similar
scope and objectives include the Sulfate Regional Experiment (SURE) of the
6-104
-------
TABLE 6-7. VALUES APPLIED IN CALCULATIONS WITH
THE LAGRANGIAN DISPERSION MODEL IN THE OECD PROJECT
Characteristic
Value
Fraction of emitted sulfur deposited
locally
Fraction of emitted sulfur transformed
directly to sulfate
Decay rate of sulfur dioxide, sec"
Rain
Dry
2- -1
Transformation rate SO,, to SO. , sec
2- -1
Loss rate of SO. , sec
Mixing height, m
0.15
0.05
4-10"| (14.4% hr"1)
1-10"D (3.6% hr"1)
3.5-10"6 (1.26% hr"1)
4-10"6 (1.44% hr"1)
1000
"Data from OECD, 1977.
6-105
-------
Electric Power Research Institute (Perhac, 1978), the Multistate Atmospheric
Power Production Pollutant Study (MAP3S) of the Department of Energy and EPA
(MacCracken, 1978), and the Sulfur Transport and Transformation in the Environment
(STATE) project of EPA. Similar models have been developed by Eliassen and
Saltbones (1975), Fisher (1978), and Johnson et al. (1978).
The main advantage of the regional approach is that the rate constants
represent averages over all sources and spatial-temporal scales of interest.
The OECD estimates of the dry and wet deposition of sulfur compounds emitted
from the Federal Republic of Germany (Figure 6-36; OECD, 1977), which is the
size of a Midwestern State, indicate that 30 percent is dry deposited and 5
percent is wet deposited within the country and 65 percent is transported.
Nevertheless, the impact on air quality is greater within the state or country
than without.
Isaksen et al. (1978) applied a diagnostic model of free-radical conversion
chemistry to the urban plume of St. Louis» Mo., on July 18, 1975. Inputs to the
model included aircraft measurements of ozone, nitric oxide, nitrogen dioxide,
sulfur dioxide, sulfate, and transport information. By tuning the chemical rate
constants to fit both ozone and sulfate at a series of plume ages, the magnitude
of the conversion rate of sulfur dioxide to sulfate was extracted as a function
of plume age (Figure 6-37; Isaksen et al., 1978). The highest conversion rate
occurred at 2 p.m. at a plume age of about 6 hr. Of the peak 5 percent per hour
rate computed, about 1 percent per hour was attributed to oxidation by the
hydroxyl radical.
Gillani et al. (1978) have applied a model with diurnally varying parameters
of mixing height, mixing intensity (turbulence), conversion rate, and removal
rate to three-dimensional aircraft measurements of a power plant plume. With
6-106
-------
I
o
DRY DEPOSITION
WET DEPOSITION
Figure 6-36. Maps of dry and wet deposition of sulfur compounds from the Federal Republic of
Germany show that about 5 percent of Germany's emissions are wet deposited and 30 percent are dry
deposited within the country.
Source: OECD (1977).
-------
5 -
>e
g
V)
cc
8
V
O
1/5
'CN
O
OH.HO, HO.
10
12 14
TIME
16
18
Figure 6-37. Extracted conversion rates of SOj to sulfate in the St.
Louis urban plume show a peak at about 5 percent per hour at 2
p.m. The rates due to the OH radical alone are also derived.
Source: Isakson et at. (1978).
6-108
-------
aircraft concentration measurements and trajectory data, simultaneous rates of
conversion and removal were estimated at 200 km. Once tuned, the model was
applied to estimate the effect of SO- emission height and time of day of release
on the total sulfate formation.
In summary, the determination of the source-effect relationship of secondary
fine particles requires the filtering of measurable data (emissions, transport
path, and concentrations) through a diagnostic model. The impact of individual
sources can be estimated once the model is properly tuned.
6-109
-------
6.6 SUMMARY
The relationship of emissions of sulfur oxides and particulate matter to
resulting ambient concentrations of these substances and their derivatives
depends on the process of transmission, which includes the transport, diffusion,
transformation, and deposition of atmospheric pollutants. Since each of these
component processes is a function of numerous physiochemical and meteorological
variables, source-receptor relationships are necessarily complex. Despite the
difficulty in analyzing the transmission process, certain findings have been
substantiated.
Atmospheric particulate matter tends to be bimodally distributed in terms
of size; thus, fine particles are less than 2 urn in diameter and coarse particles
are 2 urn or more in diameter. Fine and coarse particles also tend to differ
in their sources and chemical composition, with much of the fine particulate
matter being derived in the atmosphere from the transformation of gases (notably
sulfur dioxide, nitrogen oxides, and organics) vis-a-vis coarse particles
generally being emitted directly from industrial sources or combustion processes,
or simply consisting of various types of suspended dust. Nevertheless, it is
clear that certain chemical reactions do occur, most notably the transfer^
mation of sulfur dioxide in the atmosphere to various sulfates, including
sulfuric acid.
Although comparable masses of find and coarse particles are measured in
urban areas, coarse particles settle out more readily during transport than
fine particles. Thus, with increasing distance downwind from urban areas,
the fine particle fraction exceeds the coarse, except perhaps for fugitive
dust from soil. Fine particles may be trasnported 1000 km or more from their
origin before being deposited. During this long-range transport additional
6-110
-------
fine particles may be formed from gaseous precursors of aerosols. Moreover,
other downwind sources may be superimposed and further contribute to the fine
particulate mass.
6-111
-------
6.7 REFERENCES
Altshuller, A. P. Model Predictions of the Rates of Homogeneous Oxidation of
Sulfur Dioxide to Sulfate in the Troposphere. Atmos. Environ. 13:1653-1662,
1980.
Altshuller, A. P. Regional transport and transformation of sulfur dioxide to
sulfates in the United States. J. Air Pollut. Control Assdc. 26:318-324,
1976.
Angell, J. K., W. H. Hoecker, C. R. Dickinson, and S. H. Pack. Urban influence
on a strong daytime air flow as determined from tetroon flights. J.
Appl. Meteoral. 12:924-936, 1973.
Appel B. R. , S. M. Wall, Y. Tokiwa, and M. Haik. Interference effects in
sampling particulate nitrate in ambient air. Atmos. Environ. 13:319-325,
1979.
Bachman, J. , ed. Protecting Visibility: An EPA Report to Congress. EPA-450/5-
79-008, U.S. Environmental Protection Agency, Research Triangle Park, NC,
1979.
Backstrom, H. L. J. The chain mechanism in the auto-oxidation of sodium
sulfite solutions. Z. Phys. Chem. 625:122-128, 1934.
Bigelow, J. Z., 1898.
Boulaud, D., J. Bricard, and G. Madelaine. Aerosol growth kinetics during SO-
oxidation. Atmos. Environ. 12:171-177, 1978.
Brosset, C., K. Andreasson, and M. Perm. The nature and possible origin of
acid particles observed at the Swedish west coast. Atmos. Environ.
9:631-642, 1975.
Bull. Alabama Air Pollut. Control Comm., 1975.
Calvert, J. G. , F. Su, J. W. Bottenheim, and 0. P. Strausz. Mechanism of the
homogeneous oxidation of sulfur dioxide in the troposphere. Atmos.
Environ. 12:197-226, 1978.
Chamberlain, A. C. Movement of particles in plant communities. Iji: Vegetation
in the Atmosphere. J. L. Monteith, ed. , Academic Press Inc., New York.
1975. pp. 155-201.
Chang, S. G., et al., 1978.
Chang, S. G., R. Brodzinsky, R. Toossi, S. Markowitz, and T. Novakov. Catalytic
oxidation of SO,, on carbon in aqueous solutions. In: Proceedings of
Carbonaceous Particles in the Atmosphere, Lawrence Berkeley Laboratory,
Berkeley, Calif. , 1979.
6-112
-------
6.7 REFERENCES
Altshuller, A. P. Model Predictions of the Rates of Homogeneous Oxidation of
Sulfur Dioxide to Sulfate in the Troposphere. Atmos. Environ. 13:1653-1662,
1980. ~
Altshuller, A. P. Regional transport and transformation of sulfur dioxide to
sulfates in the United States. J. Air Pollut. Control Assoc. 26:318-324,
1976. ~~
Angell, J. K., W. H. Hoecker, C. R. Dickinson, and S. H. Pack. Urban influence
on a strong daytime air flow as determined from tetroem flights. J.
Appl. Meteoral. 12:924-936, 1973.
Appel B. R., S. M. Wall, Y. Tokiwa, and M. Haik. Interference effects in
sampling particulate nitrate in ambient air. Atmos. Environ, 13:319-321%
1979. ~~
Appel, B. R., E. L Kothny, S. Wall, M. Haik, and R. L. Knights. Diurnal and
spatial variations of organic aerosol constituents in the Los Angeles basin.
In: Proceedings of the Conference on Carbonaceous Particles in the Atnw
sphere, Lawrence Berkley Laboratory, Berkley, CA, March 2Q~22J> 1978.
Bachman, J., ed. Protecting Visibility: An EPA Report to Congress, EPA-450/5-
79-008, U.S. Environmental Protection Agency,, Research Triangle Park, NCr
1979.
Backstrom, H. L. J. The chain mechanism in the auto-oxidation of sodium
sulfite solutions. I. Phys. Chem. 625:122-128, 1934.
Bigelow, J. Z., 1898.
Boulaud, D. , J. Bricard, and G. Madelaine. Aerosol growth kinetics during SO,,
oxidation. Atmos. Environ. 12:171-177, 1978.
Brosset, C. , K. Andreasson, and M. Perm. The nature and possible origin of
acid particles observed at the Swedish west coast. Atmos. Environ.
9:631-642, 1975.
Bull. Alabama Air Pollut. Control Comm., 1975.
Calvert, J. G., F. Su, J. W. Bottenheim, and 0. P. Strausz. Mechanism of the
homogeneous oxidation of sulfur dioxide in the troposphere. Atmos.
Environ. 12:197-226, 1978.
Chamberlain, A. C. Movement of particles in plant communities. Li: Vegetation
in the Atmosphere. J. L. Monteith, ed., Academic Press Inc., New York.
1975. pp.' 155-201.
Chang, S. G. , R. Brodzinsky, R. Toossi, S. Markowitz, and T. Novakov. Catalytic
oxidation of SO,, on carbon in aqueous solutions. Im Proceedings of
Carbonaceous Particles in the Atmosphere, Lawrence Berkeley Laboratory,
Berkeley, Calif., 1978.
6-112
-------
Charlson, R. J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. H?SO./(NH.)pSO. background aerosol: optical detection in the
St. Louis regiori. AtmOs. Environ. 8:1257-1267, 1974.
Charlson, R. J., D. S. Covert, T. V. Larson, and A. P. Waggoner. Chemical
properties of tropospheric sulfur aerosols. Atmos. Environ. 12:39-53,
1978.
Chun, K. C., and J. E. Quon. Capacity of ferric oxide particles to oxidize
sulfur dioxide in air. Environ. Sci. Technol. 7:532-538, 1973.
Chung, Y. C. The distribution of atmospheric sulfates in Canada and its
relationship to long range transport of air pollutants. Atmos. Environ.
12:1471-1480, 1978.
Clark, W. E., D. A. Landis, and A. B. Harker. Measurements of the photochemical
production of aerosols in ambient air near a freeway for a range of SOp
concentrations. Atmos. Environ. 10:637-644, 1976.
Clark, W. E. and K. T. Whitby. Concentration and size distribution measurement
of atmospheric aerosols and a test of the theory of self-preserving size
distributions. J. Atmos. Sci. 24:677-687, 1967.
Coordinating Research Council. Handbook of Chemistry and Physics 49:0-180,
E-37, Chemical Rubber Co., Cleveland, OH, 1968.
Coughanowr, D. R., and F. E. Krause. The reaction of SOp and 0? in aqueous
solution of MnSO^. Ind. Eng. Fund. 4:61-66, 1965.
Covert, D. W. A study of the relationship of chemical composition and humidity
to light scattering by aerosols. Ph.D. Thesis, University of Washington,
Seattle, 1974.
Cronn D. R. , R. J. Charlson, R. L. Knights, A. L. Crittenden, and B. R. Appel.
A survey of the molecular nature of primary and secondary components of
particles in urban air by high-resolution mass spectrometry. Atmos.
Environ. 11:929-937, 1977.
Demerjian K., J. A. Kerr, and J. Calvert. Mechanism of photochemical smog
formation. In: Advances in Environmental Science and Technology, Vol.
4. (Edited by J. N. Pitts, and R. J. Metcalf). John Wiley. NY, 1974.
Draftz, R. G., and K. Severin. Microscopical Analysis of Aerosols Collected
in St. Louis, Missouri. EPA-600/3-80-27, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1980.
Draftz, R. G. Aerosol Source Characterization Study in Miami, Florida:
Microscopical Analysis. EPA-600/3-79-097, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1979.
Durham, J. L., R. K. Patterson, and R. G. Draftz. Carbon in Denver's Urban
Plume. Proceedings: Carbonaceous Particles in the Atmosphere,
Lawrence Berkeley Laboratory, Rep. No. LBL-9037, March 20-22, 1978.
pp. 102-106.
6-113
-------
Durst, C. S. , A. F. Crossley, and N. E. Davis. Horizontal diffusion in the
atmosphere as determined by geostrophic trajectories. J. Fluid Mech.
6:401-422, 1959.
Dzubay, T. G. Chemical element balance method applied to dichotomous sampler
data. In: Proceedings of Symposium on Aerosols, Anthropogenic and
Natural Sources and Transport, Annals of the New York Acad. of Sci., New
York, 1979.
Easter, R. C., D. F. Miller, and W. E. Wilson. Kinetic simulation of the
diurnal rates of S0? oxidation in a power plant plume. Atmos. Environ.
vol. 14, 1980. £
Eatough, D. J., T. Major, J. Ryder, M. Hill, N. F. Mangelson, N. L. Eatough,
L. D. Hansen, R. G. Meisenheimer, and J. W. Fischer. The formation and
stability of sulfite species in aerosols. Atmos. Environ. 12:262-271,
1978.
Eatough, D. J., W. P. Green, and L. D. Hansen. Oxidation of sulfite by
activated charcoal. In: Proceedings of the Conference on Carbonaceous
Particles in the Atmosphere, Lawrence Berkley Laboratory, Berkley, CA,
1978a.
Eliassen, A. The Trajectory Model: A Technical Description. Norwegian
Institute for Air Research, Kjeller, Norway, 1976.
Eliassen, A., and J. Saltbones. Decay and transformation rates of SOp as
estimated from emission data, trajectories and measured air concentrations.
Atmos. Environ. 12:489-501, 1975.
Erickson, R. E., L. M. Yates, R. L. Clark, and C. M. McEwen. The reaction of
sulfur dioxide with ozone in water and its possible atmospheric signifi-
cance. Atmos. Environ. 11:813-817, 1977.
Ferber, G. J., K. Telegada, J. L. Heffter, and A. E. Smith. Air concentrations
of krypton-85 in the mid-west United States during January-May, 1974.
Atmos. Environ. 10:379-385, 1976.
Fisher, B. E. A. The calculation of long term sulphur deposition in Europe.
Atmos. Environ. 12:489-501, 1978.
Forrest and Newman, 1977.
Fowler, D. , and M. H. Unsworth. Dry deposition of sulfur dioxide on wheat.
Nature 249:389-290, 1974.
Friedlander, S. K. A review of the dynamics of sulfate containing aerosols.
Atmos. Environ. 12:187-195, 1978.
Friedlander, S. K. Chemical element balances and identification of air pollu-
tion sources. Environ. Sci. Technol. 7:235-240, 1973.
6-114
-------
Gage, S. J., L. F. Smith, P. M. Cukor, and B. L. Nieman. Long-range transport
of SO /MSO. from the U.S. Environmental Protection Agency/Teknekron
integrated technology assessment of electric utility energy systems. In:
Proceedings of the International Symposium on Sulfur in the Atmosphere
(ISSA), Dubrovnik, Yugoslavia, 1977.
Galvin, P. J., P. J. Samson, P. E. Coffey, and D. Ramano. Transport of sulfate
in New York state. Environ. Sci. Techno!. 12:580-584, 1978.
Garland, J. A. Dry and wet removal of sulphur from the atmosphere. Atmos.
Environ. 12:349-362, 1978.
Gartrell, G., Jr., and S. K. Friedlander. Relating particulate pollution to
sources: the 1972 California aerosol characterization study. Atmos.
Environ. 9:279-299, 1975.
Gelbard F- , and J. H. Seinfeld. The general dynamic equation for areosols—
Theory and application to aerosol formation and growth. J. Colloid
Interface Sci. 68:363- , 1979.
Gillani, N. V., R. B. Husar, J. D. Husar, D. E. Patterson, and W. E. Wilson.
Project MISTT: kinetics of particulate sulfur formation in a power plant
plume out to 300 km. Atmos. Environ. 12:589-598, 1978.
Graf, J., R. H. Snow, and R. G. Draftz. Aerosol Sampling and Analysis -
Phoenix, Arizona. EPA-600/3-77-015, U.S. Environmental Protection Agency,
Research Triangle Park, 1977.
Graham, R. Rate constants for the reaction of H02 with H02,S02,CO,N20,
trans-2-butene, and 2,3-dimethyl-2-butene at 300 K. 3. Pnys. Cnem., vol.
83, 1979. *
Grosjean D. Chapters, Aerosols. In: Ozone and Other Photochemical Oxidants.
National Academy of Sciences, Washington, DC, 1977. pp. 45-125.
Grosjean D., and S. K. Friedlander. Gas-particle distribution factors for
organic and other pollutants in the Los Angeles atmosphere. J. Air
Pollut. Control Assoc. 25:1038:1044, 1975.
Haagenson, P. L., and A. L. Morris. Forecasting the behavior of the St.
Louis, Missouri pollutant plume. J. Appl. Meteoro. 13:901-909, 1974.
Hardy, K. A. Aerosol Source Characterization Study in Miami, Florida: Trace
Element Analysis. EPA-600/7-79-197, September, 1979.
Hales, J. M. Wet removal of sulfur compounds from the atmosphere. Atmos.
Environ. 12:389-399, 1978.
Hall, F. P., Jr., C. E. Duchon, L. G. Lee, and R. R. Hagan. Long-range
transport of air pollution, a case study, August 1970. Mon. Weather Rev.
101:404, 1973.
6-115
-------
Marker A. B., L. W. Richards, and W. E. Clark. The effect of atmosperhic SCL
photochemistry upon observed nitrate concentrations in aerosols. Atmos.
Environ. 11:87-91, 1977.
Harrison, P. R., R. G. Draftz, and W. H. Murphy. Identification and Impact of
Chicago's Ambient Suspended Dust. In: Atmosphere-Surface Exchange of
Particulate and Gaseous Pollutants (1974), pp 540-560.
Haury. G. , S. Jordan, and C. Hofmann. Experimental investigation of aerosol-
catalyzed oxidation of S0? under atmospheric conditions. Atmos. Environ.
12:281-287, 1978. ^
Hegg, D. A., and P. V. Hobbs. Oxidation of sulfur dioxide in aqueous systems
with particular reference to the atmosphere. Atmos. Environ. 12:241-253, 1978.
Heisler, N. L. , S. K. Friedlander, and R. B. Husar. Atmos. Environ. 7:633, 1973.
Hess, G. D., and B. B. Hicks. A study of PBL structure: the Sangamon experiment
of 1975. Argonne National Laboratory Report ANL 75-60, Part IV, 1975. pp. 1-4.
Hidy G. M., and J. R. Brock. The Dynamics of Aerocolloidal Systems. Pergamon
Press, NY, 1970.
Hidy, G. M. Summary of the California aerosol characterization experiment.
J. Air Pollut. Control Assoc. 25:1106-1114, 1975.
Hidy, G. M. , ed. Aerosols and Atmospheric Chemistry. Kendall Award Symposium.
Academic Press Inc., New York. 1972.
Hidy, G. M., P. K. Mueller, and E. Y. Tong. Spatial and temporal distributions
of airborne sulfate in parts of the United States. Atmos. Environ. 12:
735-752, 1978.
Hill, F. B., and R. F. Adamowicz. A model for rain composition and the washout
of sulfur dioxide. Atmos. Environ. 11:917-927, 1977.
Hoather, R. C., and C. F. Goodeve. The oxidation of sulphmous acid. III.
Catalysis of mangenous sulphate. Trans. Faraday Soc. 30:1149-1156, 1934.
Hoecker, W. H. Accuracy of various techniques for estimating boundary layer
trajectories. J. Appl. Meteorol. 16:374-383, 1977.
Holt, B. D., P. T. Cunningham, and R. Kumar. Use of oxygen isotopy in the
study of transformations of SO- to sulfates in the atmosphere. Presented
at WMO Symposium, Sofia, Bulgaria, Oct. 1-5, 19/j.
Holzworth, G. C. Mixing depths, wind speeds and air pollution potential for
selected locations in the United States. J. Appl. Meteorol. 6:1039-1044,
1967.
Holzworth, G. C. Mixing Heights, Wind Speeds and Potential for Urban Air Pol-
lution Throughout Contiguous United States. Environmental Protection
Agency, Office of Air Programs Publication No. AP-101. U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1972. p. 35.
6-116
-------
Hosier, C. R. Low level inversion frequency in the contiguous United States.
Mon. Weather Rev. 89:319-339, 1961.
Husar, R. B., and K. T. Whitby. Growth mechanisms and size spectra of photo-
chemical aerosols. Environ. Sci. Technol. 7:214, 1973.
Husar, R. B., D. E. Patterson, C. C. Paley, and N. V. Gillani. Ozone in hazy
air masses. In: Proceedings of the International Conference on Photo-
chemical Oxidants and Its Control, Research Triangle Park, NC, Sept.
12-20, 1976b.
Husar, R. B., D. E. Patterson, J. D. Husar, N. V. Gillani, and W. E. Wilson.
Sulfur budget of a power plant plume. Atmos. Environ. 12:549-568, 1978.
Husar, R. B., N. V. Gillani, J. D. Husar, C. C. Paley, and P. N. Turcu. Long
range transport of pollutants observed through visibility contour maps,
weather maps, and trajectory analysis. Presented at the Third Symposium
on Atmospheric Turbulence, Diffusion, and Air Quality. Raleigh, NC, Oct.
19-22, 1976a.
International Symposium on Sulfur in the Atmosphere (ISSA) Proceedings. Atmos.
Environ. 12:1-769, 1978.
Isaksen, I. S. A. , E. Hesstved, and 0. Hov. A chemical model for urban plumes:
test for ozone and particulate sulfur formation in St. Louis urban plume.
Atmos. Environ. 12:599-604, 1978.
Johnson, W. B., D. E. Wolf, and R. L. Mancuso. Long term regional patterns
and transfrontier exchanges of airborne sulfur pollution in Europe.
Atmos. Environ. 12:511-527, 1978.
Judeikis, H. S. The Role of Solid-Gas Interactions in Air Pollution.
EPA-650/3-74-007, U.S. Environmental Protection Agency, August 1974.
Judeikis, H. S., and S. Siegel. Particle-catalyzed oxidation of atmospheric
pollutants. Atmos. Environ. 7:619-631, 1973.
Junge, C. E. Air Chemistry and Radioactivity. Academic Press Inc., New York,
1963.
Junge, C. E. Die Konstitution der atmospheuschen Aerosols. Ann. Meteorol.
Hamburg 5:1-55, 1952.
Junge, C. E. Die rolle der aerosole und der gasformingen beimengungen der
luft in spurenstoffhaushalt der troposphere. Tellus 5:1-26, 1953.
Kapustin, V. N., Yu. S. Lynbortseva, and G. V. Rozenberg. Variability of
aerosols under the influence of cloud modulation of the radiation field.
Izv. Atmos. Oceanic Phys. 10:1327-1331, 1974.
Karraker, D. G. The kinetics of the reaction between sulphurous acid and
ferric ion. J. Phys. Chem. 67:871-874, 1963.
6-117
-------
Komp, M. J., and A. H. Auer, Jr. Visibility reduction and accompanying aerosol
evolution downwind of St. Louis. J. Appl. Meteorol. 17:1357-1367, 1978.
Korshover, J. Climatology of stagnating anticyclones east of the Rocky Mountains
1936-1965. Public Health Service Publication No. 999-AP-34, Cincinnati,
OH, 1967. 15 pp.
Kuhlman, M. R., D. L. Fox, and H. E. Jeffries. The effect of CO on sulfate
aerosol formation. Atmos. Environ. 12:2415-2423, 1978.
Latimer, D. A., R. W. Bergstrom, S. R. Hayes, M. K. Liw, J. H. Seinfeld, G. Z.
Whitten, M. A. Wojcik, and M. J. Hillyar. The Development of mathematical
models for the prediction of anthropogenic visibility impairment. EPA-450/
3-78-110 a, U.S. Environmental Protection Agency, 1978.
Lee, R. E. Jr. The Size of Suspended Particulate Matter in Air. Science
178:567-575, 1972.
Lewis, C. W., and E. S. Macias. Composition of size fractioned aerosol in
Charleston, West Virginia. Submitted for publication. Atmos. Environ.
14:185-194, 1980.
Liberti, A., D. Brocco, and M. Possanzini. Adsorption and oxidation of sulfur
dioxide on particles. Atmos. Environ. 12:255-261, 1978.
Lipeles M., C. S. Burton, H. H. Wang, E. P. Parry, and G. M. Hidy. Mechanisms
of Formation and Composition of Photochemical Aerosols. EPA-650/73-036.
Thousand Oaks, CA: Rockwell International Science Center, 1973. 98 pp.
Loo, B. W., W. R. French, R. C. Gatti, F. S. Goulding, J. M. Jaklevic, J.
Llacer, and A. C. Thompson. Large-scale measurement of airborne particulate
sulfur. Atmos. Environ. 12:759-771, 1978.
Low, R. D. H. A theoretical study of nineteen condensation nuclei. J. Rech.
Atmos. 4:65-78, 1969.
Lyons, W. A., and R. B. Husar. SMS/GOES visible images detect a synoptic-scale
air pollution episode. Mon. Weather Rev. 104:1623-1626, 1976.
MacCracken, M. C. MAP3S: an investigation of atmospheric energy related
pollutants in the northeastern United States. Atmos. Environ. 12:649-659,
1978.
Macias, E. S., D. L. Blumenthal, J. A. Anderson, and B. K. Cantrell. Character-
ization of Visibility Reducing Aerosols in the Southwestern United States.
Interim report on Project VISTTA. MRI Report 78 1R, 15-85, 1979.
McMurry, P. H., 'and S. K. Friedlander. Aerosol formation in reacting gases:
relation of surface area to rate of gas-to-particle conversion. J.
Colloid Interface Sci. 64:248-257, 1978.
6-118
-------
McNeils, D. N., L. A. Ripperton, W. E. Wilson, P. L. Hanst, and B. W. Gay.
Gas Phase Reactions of Ozone and Olifin in the Presence of Sulfur Dioxide.
In: Removal of Trace Contaminants from Air, B.R. Dietz, ed., A.C.S.
Symposium Series 17:187-200, 1975.
Meyers, R. E., R. T. Cederwall, W. D. Ohmstede, and J. Kampe. Transport and
diffusion using a diagnostic meso-scale model employing mass and total
energy conservation constraints. Jji: Proceedings of the Third Symposium
on Atmospheric Turbulence, Diffusion and Air Quality, American Meteorological
Society, Boston, MA, 1977. pp. 90-97.
Middleton P. B., and J. R. Brock. Simulation of aerosol kinetics. J. Colloid
Interface Sci. 54:249- , 1976.
Middleton P. B, S. H. Suck, and J. R. Brock. On the multimodality of density
functions of pollutant aerosols. Atmos. Environ. 11:251-255, 1977.
Middleton, P. , and C. S. Kiang. Experimental and theoretical examination of
the formation of sulfuric acid particles. Atmos. Environ. 12:179-185,
1978.
Miller, D. F. Precursor effects on S09 oxidation. Atmos. Environ. 12:273-280,
1978. *
Miller, F. J. , D. E. Gradner, J. A. Grahm, R. E. Lu, Jr., W. E. Wilson, and J.
D. Bachman. Size Considerations for Establishing a Standard for Inhalable
Particles. J. Air Pollut. Control Assoc. 29:610-615, 1979.
Miller, M. S., S. K. Friedlander, and G. M. Hidy. A chemical element balance
for the Pasadena aerosol. J. Colloid Interface Sci. 39:165-176, 1972.
Mueller, P. K. , G. M. Hidy, K. Warren, T. F. Lavery, and R. L. Baskett. The
occurrence of atmospheric aerosols in the northeastern United States.
Annals of the New York Academy of Sciences 338:462-482, 1980.
Newman, J. E., M. D. Abel, P. R. Harrison, and K. J. Yost. Wind as Related to
Critical Flushing Speed Versus Reflotation Speed by High-Volume Sampler
Particulate Loading, 466-496. In: Atmospheric-Surface Exchange of
Particulate and Gaseous Pollutants (1974). Englemann, R. J., ed. Energy
Research and Development Administration symposium series, CONF-740921,
1976.
Neytzell-de Wilde, F. G., and L. Taverner. Experiments relating to the possible
production of an oxidizing acid leach liquor by auto-oxidation for the
extraction of uranium. In: Proceedings of the Second U.N. Conference
Peaceful Uses of Atomic Energy, vol. 3, 1958. pp. 303-317.
Novakov, T. Conference on Carbonoceous Particles in the Atmosphere. Lawrence
Berkley Laboratory, Berkley, CA, March 20-22, 1978.
OECD. The OECD Program on Long-Range Transport of Air Pollutants—Measurements
and Findings. Organization for Economic Cooperation and Development,
Paris, 1977.
6-119
-------
Orel A. E., and J. H. Seinfeld. Nitrate formation in atmospheric aerosols.
Environ. Sci. Techno!. 11:1000-1007, 1977.
Orr, C. , F. K. Hurd, and W. J. Corbert. Aerosol size and relative humidity.
J. Colloid Interface Sci. 13:472-482, 1958.
Ottar, B. Monitoring long-range transport of air pollutants: the OECD study.
Ambio 6:203-206, 1976.
Overton, J. H., V. P- Aneja, and J. L. Durham. Production of sulfate in rain
and raindrops in polluted atmospheres. Atmos. Environ. 13:355-367, 1979.
Pack, D. H., G. J. Ferber, J. L. Hefter, K. Telegadas, J. K. Angell, W. H.
Hoecker, and L. Machta. Meteorology of long-range transport. Atmos.
Environ. 12:425-444, 1978.
Penkett, S. A., B. M. R. Jones, K. A. Brice, and A. E. J. Eggleton. The
importance of atmospheric ozone and hydrogen peroxide in oxidizing sulphur
dioxide in cloud and rainwater. Atmos. Environ. 13:123-137, 1979.
Perhac, R. M. Sulfate regional experiment in northeastern United States: the
'SURE' program. Atmos. Environ. 12:641-647, 1978.
Pierson, W. R., and P. A. Russell. Aerosol carbon in the Denver area in
November 1973. Atmos. Environ. 13:1623-1628, 1979.
Roberts, P. T. , and S. K. Friedlander. Photochemical aerosol formation--S02,
1-heptene, and NO in ambient air. Environ. Sci. Technol. 10:573,. 1976.
Rodhe, H. Budgets and turn-over times of atmospheric sulfur compounds.
Atmos. Environ. 12:671-680, 1978.
Rodhe, H., C. Persson, and 0. Akesson. An investigation into regional trans-
port of soot and sulfate aerosols. Atmos. Environ. 6:675-693, 1972.
Ross, D. S. Comments on Are! and Seinfeld. Environ. Sci. Technol. 11:1000,
1977; 12:726, 1978.
Samson, P. J. Ensemble trajectory analysis of summertime sulfate concentrations
in New York. Atmos. Environ. 12:1889-1893, 1978.
Samson, P. J., and K. W. Ragland. Ozone and visibility reduction in the
Midwest: evidence for large scale transport. J. Appl. Meteorol.
16:1101-1106, 1977.
Schuetzle D., D. Cronn, A. L. Crittenden, and R. J. Charlson. Molecular
composition of secondary aerosol and its possible origin. Environ. Sci.
Technol. 9:838-845, 1975.
6-120
-------
Schiermeier, F. A., F. Pooler, N. V. Gillani, J. F. Clarke, J. K. S. Ching,
and W. E. Wilson. Experimental investigation of atmospheric sulfur
transport and transformation in the STATE program. In: Proceedings
of the 10th International Technical Meeting on Air Pollution Modeling
and its Application. NATO/CCMS, Rome, Italy., Oct. 23-26, 1979.
Schwartz W. Chemical Characterization of Model Aerosols. EPA-650/3-74-011.
Columbus, Ohio: Battelle Memorial Institute, 1974. pp. 124
Sehmel, G. A., and W. H. Hodgson. Particle dry deposition velocities. Iji:
Atmosphere-Surface Exchange of Particulate and Gaseous Pollutants. 1974
Proceedings of ERDA Symposium Series 38, Richland, WA, Publication No.
CONF-740921 from NTIS, Springfield, VA, 1976.
Sinclair, D., R. J. Countess, and G. S. Hoopes. Effect of relative humidity
on the size of atmospheric aerosol particles. Atmos. Environ. 8:1111-1118,
1974.
SI inn, W. G. N. Dry Deposition and Resuspension of Aerosol Particles - A New
Look at Lome Old Problems. In: Atmosphere-Surface Exchange of Particulate
and Gaseous Pollutants (1974). Englemann, R. J. ed. Energy Research and
Development Administration symposium series, CONF-740921, 1976. pp. 1-40.
Slinn, W. G. N., L. Hasse, B. B. Hicks, A. W. Hogan, D. Lai, P. S. Liss, K. 0.
Munnich, G. A. Sehmel, and 0. Vittori. Some aspects of the transfer of
atmospheric trace constituents past the air-sea interface. Atmos. Environ.
12:2055-2087, 1978.
Smith, B. M., J. Wagman, and B. R. Fish. Interaction of airborne particles
with gases. Environ. Sci. Technol. 9:558-562, 1969.
Smith, F. B. , and G. H. Jeffrey. Airborne transport of sulfur dioxide from
the United Kingdom. Atmos. Environ. 9:643-659, 1975.
Smith, F. B. , and R. D. Hunt. Meteorological aspects of the transport of
pollution over long distances. Atmos. Environ. 12:461-477, 1978.
Smith, T. B. , D. L. Blumenthal, J. A. Anderson, and A. H. Vanderpol. Transport
of S0? in Power Plant Plumes: Day and Night. Atmos. Environ. 12:605-611,
1978.
Spicer C. W., and P. M. Schumacher. Particulate nitrate: Laboratory and
field studies of major sampling interferences. Atmos. Environ. 13:543-552,
1979.
Spicer, C. W. The Fate of Nitrogen Oxides in the Atmosphere. EPA 600/3-76-030,
U.S. Environmental Protection Agency, Research Triangle Park, NC, 1976.
6-121
-------
Stelson A. W., S. K. Friedlander, and J. H< Seinfeld. A note on the equilibrium
realtionship between ammonia and nitric acid and particulate ammonium
nitrate. Atmos. Environ. 13:369-371, 1979.
Stevens, R. K., T. G. Dzubay, G. Russwurm, and D. Rickel. Sampling and analysis
of atmospheric sulfates and related species. Atmos. Environ. 12:55-68,
1978. ~~
Stevens, R. K., editor. Current Methods to Measure Atmosperhic Nitric Acid
and Nitrate Artifacts. EPA-600/2-79-051, U.S. Environmental Protection
Agency, Research Triangle Park, N.C. 99 pp. (1979).
Suck, S. H., E. C. Upchurch, and J. R. Brock. Dust transport in Maricopa
County, Arizona. Atmos. Environ. 12:2265-2271, 1978.
Sykes, R. I., and L. Hatton. Computation of horizontal trajectory based on
the surface geostrophic wind. Atmos. Environ. 10:925-934, 1976.
Szepesi, D. J. Transmission of sulfur dioxide on local, regional and continental
scale. Atmos. Environ. 12:529-535, 1978.
Tang, I. N. Phase transformation and growth of aerosol particles composed of
mixed salts. J. Aerosol Sci. 7:361-371, 1976.
Tang, I. N., and H. R. Munkelwitz. Aerosol growth studies. III. Ammonium
bisulfate aerosol in a moist atmosphere. J. Aerosol Sci. 8:321-330,
1977.
Tang, I. N. On the Equilibrium Partial Pressures of Nitric Acid and Ammonia
in the Atmosphere. Atmos. Environ, (accepted for publication) 1980.
Tartarelli, R., P. Davine, F- Morelli, and P. Corsi. Interactions between SO^
and carbonaceous particulates. Atmos. Environ. 12:289-293, 1978.
Titoff, A. Beitrage zur Kaentnis der negativen Katalyze im homogenen System.
Z. Phys. Chem. 45:641-683, 1903.
Tong, E. Y., G. M. Hidy, T. F. Lavery, and F. Berlandi. Regional and local
aspects of atmospheric sulfates in the northeast quadrant of the United
States. In: Proceedings Third Symposium of Turbulence, Diffusion and
Air Quality, American Meteorological Society, 1976.
Urone, P., H. Lutsep, C. M. Noyes, and J. F- Parcher. Static studies of
sulfur dioxide reactions in air. Environ. Sci. Technol. 8:611-618, 1968.
Vukovich, F. M., W. D. Bach, Jr., B. W. Cressman, and W. J. King. On the
relationships between high ozone in the rural boundary layer and high
pressure sys.tems. Atmos. Environ. 11:967-983, 1977.
Wagman, J., R. E. Lee, and C. J. Axt. Influence of Some Atmospheric Variables
on the Concentration and Particle Size Distribution of Sulfate in Urban
Air. Atmospheric Environ 1:479-489, 1967.
Whitby, K. T. The physical characteristics of sulfur aerosols. Atmos. Environ.
12:135-159, 1978.
6-122
-------
Whitby, K. T. and G. M. Sverdrup. California Aerosols: Their Physical and
Chemical Characteristics. In: The Characters and Origin of Smog Aerosols.
G. M. Hidy et al., eds. John Wiley, New York, 1980.
Whitby, K. T. , R. B. Hausar, and B. Y. H. Lui. The Aerosol Size Distribution
of Los Angeles Smog. J. Colloid Interfrace Sci. 39:177-204, 1972.
Whitby, K. T., R. E. Charlson, W. E. Wilson, and R. K. Stevens. The Size of
Suspended Particle Matter in Air. Science 183:1098-1199, 1974.
White, W. H. , and P. T. Roberts. On the nature and origins of visibility-
reducing aerosols in the Los Angeles air basin, In: The Character
and Origins of Smog Aerosols. Advances in Environ. Sci. and Technol.
10:715-753, 1980.
White, W. H., J. A. Anderson, D. L. Blumenthal, R. B. Husar, N. V. Gillani, J.
D. Husar, and W. E. Wilson. Formation and transport of secondary air
pollutants: ozone and aerosols in the St. Louis urban plume. Science
194:187-189, 1976.
Wilson, W. E. Sulfates in the atmosphere: a progress report on Project
MISTT. Atmos. Environ. 12:537-548, 1978.
Wilson, W. E. Transformation during transport: a state of the art survey of
the conversion of S0? to sulfate. Presented at WMO Symposium, Sofia,
Bulgaria, Oct. 1-5, 1979.
Wilson, W. E. , D. F. Miller, A. Levy, and R. K. Stone. The effect of Fuel
Composition on Atmospheric Aerosol Due to Auto Exhaust. J. Air Pollut.
Control Assoc. 23:949-956, 1973.
Wilson, W. E. , et al. , 1979.
Wilson, W. E. , L. L. Spiller, T. G. Ellestad, P. L. J. Lamothe, T. G. Dzuby,
R. K. Stevens, E. S. Macias, R. A. Fletcher, J. D. Jusar, R. B. Husar, K.
T. Whitby, D. B. Kittelson, and B. K. Cantrell. General Motors Sulfate
Dispersion Experiment: Summary of EPA Measurements. J. Air Pollut.
Control Assoc. 27:46-51, 1977.
Wilson, W. E. , R. J. Charlson, R. B. Hasar, K. T. Whitby, and D. L. Blumenthal.
Sulfates in the atmosphere. Presented at the 69th Annual Meeting of Air
Pollution Control Assoc., Portland, OR, June 27-July 1, 1976.
Winchester, J. W., R. J. Ferek, D. R. Lawson, J. 0. Pilotte, M. A. Thiemens,
and L. E. Wagon. Comparison of aerosol sulfur and crystal element con-
centrations in particle size fractions from continental United States
locations. Submitted for publication, 1979.
Winkler, P. The growth of atmospheric aerosol particles as a function of the
relative humidity. II. An improved concept of mixed nuclei. Aerosol
Sci. 4:373-387, 1973.
6-123
-------
Winkler, P. , and C. Junge. The growth of atmospheric aerosol particles as a
function of the relative humidity. I. Methods and measurements of
different locations. J. Rech. Atmos. 6:617-638, 1972.
Wolff, G. T. , P. J. Lioy, R. E. Meyers, and R. T. Cederwall. An investigation
of long-range transport of ozone across the Midwestern and Eastern United
States. Atmos. Environ. 11:797-802, 1977.
6-124
-------
Charlson, R. J., A. H. Vanderpol, D. S. Covert, A. P. Waggoner, and N. C.
Ahlquist. H2S04/(NH.)2SO. background aerosol: optical detection in the
St. Louis region". Almos. Environ. 8:1257-1267, 1974.
Charlson, R. J. , D. S. Covert, T. V. Larson, and A. P. Waggoner. Chemical
properties of tropospheric sulfur aerosols. Atmos. Environ. 12:39-53,
1978. ~~
Chun, K. C. , and J. E. Quon. Capacity of ferric oxide particles to oxidize
sulfur dioxide in air. Environ. Sci. Technol. 7:532-538, 1973.
Chung, Y. C. The distribution of atmospheric sulfates in Canada and its
relationship to long range transport of air pollutants. Atmos. Environ.
12:1471-1480, 1978.
Clark, W. E. , D. A. Landis, and A. B. Harker. Measurements of the photochemical
production of aerosols in ambient air near a freeway for a range of S0?
concentrations. Atmos. Environ. H):637-644, 1976.
Clark, W. E. and K. T. Whitby. Concentration and size distribution measurement
of atmospheric aerosols and a test of the theory of self-preserving size
distributions. J. Atmos. Sci. 24:677-687, 1967.
Coordinating Research Council,- 1968.
Coughanowr, D. R. , and F. E. Krause. The reaction of S0? and 0? in aqueous
solution of MnS04. Ind. Eng. Fund. 4:61-66, 1965. * i
Covert, D. W. A study of the relationship of chemical composition and humidity
to light scattering by aerosols. Ph.D. Thesis, University of Washington,
Seattle, 1974.
Cronn D. R., R. J. Charlson, R. L. Knights, A. L. Crittenden, and B. R. Appel.
A survey of the molecular nature of primary and secondary components of
particles in urban air by high-resolution mass spectrometry. Atmos.
Environ. 11:929-937, 1977.
Demerjian K., J. A. Kerr, and J. Calvert. Mechanism of photochemical smog
formation. In: Advances in Environmental Science and Technology, Vol.
4. (Edited by J. N. Pitts, and R. J. Metcalf). John Wiley, NY, 1974.
Draftz, R. G. , and K. Severin. Microscopical Analysis of Aerosols Collected
in St. Louis, Missouri. EPA-600/3-80-27, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1980.
Draftz, R. G. Aerosol Source Characterization Study in Miami, Florida:
Microscopical Analysis. EPA-600/3-79-097, U.S. Environmental Protection
Agency, Research Triangle Park, NC, 1979.
Durham, et al. , 1979.
6-113
-------
Durst, C. S., A. F. Crossley, and N. E. Davis. Horizontal diffusion in the
atmosphere as determined by geostrophic trajectories. J. Fluid Mech.
6:401-422, 1959.
Dzubay, T. G. Chemical element balance method applied to dichotomous sampler
data. lr\: Proceedings of Symposium on Aerosols, Anthropogenic and
Natural Sources and Transport, Annals of the New York Acad. of Sci., New
York, 1979.
Easter, R. C. , D. F. Miller, and W. E. Wilson. Kinetic simulation of the
diurnal rates of SCL oxidation in a power plant plume. Atmos. Environ.
vol. 14, 1980.
Eatough, D. J., et al., 1979.
Eatough, D. J., T. Major, J. Ryder, M. Hill, N. F. Mangelson, N. L Eatough,
L. D. Hansen, R. G. Meisenheimer, and J. W. Fischer. The formation and
stability of sulfite species in aerosols. Atmos. Environ. 12:262-271,
1978.
Eliassen, A. The Trajectory Model: A Technical Description. Norwegian
Institute for Air Research, Kjeller, Norway, 1976.
Eliassen, A., and J. Saltbones. Decay and transformation rates of S0? as
estimated from emission data, trajectories and measured air concentrations.
Atmos. Environ. 12:489-501, 1975.
Erickson, R. E., L. M. Yates, R. L. Clark, and C. M. McEwen. The reaction of
sulfur dioxide with ozone in water and its possible atmospheric signifi-
cance. Atmos. Environ. 11:813-817, 1977.
Ferber, G. J. , K. Telegada, J. L. Heffter, and A. E. Smith. Air concentrations
of krypton-85 in the mid-west United States during January-May, 1974.
Atmos. Environ. 10:379-385, 1976.
Fisher, B. E. A. The calculation of long term sulphur deposition in Europe.
Atmos. Environ. 12:489-501, 1978.
Forrest and Newman, 1977.
Fowler, D., and M. H. Unsworth. Dry deposition of sulfur dioxide on'wheat.
Nature 249:389-290, 1974.
Friedlander, S. K. A review of the dynamics of sulfate containing aerosols.
Atmos. Environ. 12:187-195, 1978.
Friedlander, S. K. Chemical element balances and identification of air pollu-
tion sources. Environ. Sci. Techno!. 7:235-240, 1973.
6-114
-------
Gage, S. J. , L. F. Smith, P. M. Cukor, and B. L. Nieman. Long-range transport
of SO /MSO. from the U.S. Environmental Protection Agency/Teknekron
integrated technology assessment of electric utility energy systems. Jjn:
Proceedings of the International Symposium on Sulfur in the Atmosphere
(ISSA), Dubrovnik, Yugoslavia, 1977.
Galvin, P- J., P. J. Samson, P. E. Coffey. and D. Ramano. Transport of sulfate
in New York state. Environ. Sci. Technol. 12:580-584, 1978.
Garland, J. A. Dry and wet removal of sulphur from the atmosphere. Atmos.
Environ. 12:349-362, 1978.
Gartrell, G., Jr., and S. K. Friedlander. Relating particulate pollution to
sources: the 1972 California aerosol characterization study. Atmos.
Environ. 9:279-299, 1975.
Gelbard F. , and J. H. Seinfeld. The general dynamic equation for areosols--
Theory and application to aerosol formation and growth. J. Colloid
Interface Sci. 68:363- , 1979.
Gillani, N. V., R. B. Husar, J. D. Husar, D. E. Patterson, and W. E. Wilson.
Project MISTT: kinetics of particulate sulfur formation in a power plant
plume out to 300 km. Atmos. Environ. 12:589-598, 1978.
Graf, J., R. H. Snow, and R. G. Draftz. Aerosol Sampling and Analysis -
Phoenix, Arizona. EPA-600/3-77-015, U.S. Environmental Protection Agency,
Research Triangle Park, 1977.
Graham, R. Rate constants for the reaction of H0? with H02,SOp,CO,N?0,
trans-2-butene, and 2,3-dimethyl-2-butene at 300 K. 6. PHys. Cnem., vol.
83, 1979.
Grosjean D. Chapters, Aerosols. Jn: Ozone and Other Photochemical Oxidants.
National Academy of Sciences, Washington, DC, 1977. pp. 45-125.
Grosjean D. , and S. K. Friedlander. Gas-particle distribution factors for
organic and other pollutants in the Los Angeles atmosphere. J. Air
Pollut. Control Assoc. 25:1038:1044, 1975.
Haagenson, P. L. , and A. L. Morris. Forecasting the behavior of the St.
Louis, Missouri pollutant plume. J. Appl. Meteoro. 13:901-909, 1974.
Hales, J. M. Wet removal of sulfur compounds from the atmosphere. Atmos.
Environ. 12:389-399, 1978.
Hall, F. P., Jr., C. E. Duchon, L. G. Lee, and R. R. Hagan. Long-range
transport of air pollution, a case study, August 1970. Mon. Weather Rev.
101:404, 1973.
Harker A. B., L. W. Richards, and W. E. Clark. The effect of atmosperhic S02
photochemistry upon observed nitrate concentrations in aerosols. Atmos.
Environ. 11:87-91, 1977.
6-115
-------
Harker, A. B., et al. , 1979.
Haury, G. , S. Jordan, and C. Hofmann. Experimental investigation of
aerosol-catalyzed oxidation of SO- under atmospheric conditions. Atmos.
Environ. 12:281-287, 1978.
Hegg, D. A., and P. V. Hobbs. Oxidation of sulfur dioxide in aqueous systems
with particular reference to the atmosphere. Atmos. Environ. 12:241-253,
1978. ~~
Heisler, N. L., S. K. Friedlander, and R. B. Husar. Atmos. Environ. 7:633,
1973.
Hess, G. D. , and B. B. Hicks. A study of PBL structure: the Sangamon experi-
ment of 1975. Argonne National Laboratory Report ANL 75-60, Part IV,
1975. pp. 1-4.
Hidy G. M., and J. R. Brock. The Dynamics of Aerocolloidal Systems. Pergamon
Press, NY, 1970.
Hidy, G. M. Summary of the California aerosol characterization experiment.
J. Air Pollut. Control Assoc. 25:1106-1114, 1975.
Hidy, G. M. , ed. Aerosols and Atmospheric Chemistry. Kendall Award Symposium.
Academic Press Inc. ^ New York. 1972.
Hidy, G. M., P. K. Mueller, and E. Y. long. Spatial and temporal distributions
of airborne sulfate in parts of the United States. Atmos. Environ. 12:
735-752, 1978.
Hill, F. B., and R. F. Adamowicz. A model for rain composition and the washout
of sulfur dioxide. Atmos. Environ. 11:917-927, 1977.
Hoather, R. C., and C. F. Goodeve. The oxidation of sulphmous acid. III.
Catalysis of mangenous sulphate. Trans. Faraday Soc. 30:1149-1156, 1934.
Hoecker, W. H. Accuracy of various techniques for estimating boundary layer
trajectories. J. Appl. Meteorol. 16:374-383, 1977.
Holt, B. D., P- T. Cunningham, and R. Kumar. Use of oxygen isotopy in the
study of transformations of S0? to sulfates in the atmosphere. Presented
at WMO Symposium, Sofia, Bulgaria, Oct. 1-5, 1979.
Holzworth, G. C. Mixing depths, wind speeds and air pollution potential for
selected locations in the United States. J. Appl. Meteorol. 6:1039-1044.
1967.
Holzworth, G. C. Mixing Heights, Wind Speeds and Potential for Urban Air Pol-
lution Throughout Contiguous United States. Environmental Protection
Agency, Office of Air Programs Publication No. AP-101. U.S. Environmental
Protection Agency, Research Triangle Park, NC, 1972. p. 35.
6-116
-------
Hosier, C. R. Low level inversion frequency in the contiguous United States.
Hon. Weather Rev. 89:319-339, 1961.
Husar, R. B. , and K. T. Whitby. Growth mechanisms and size spectra of photo-
chemical aerosols. Environ. Sci. Techno!. 7:214, 1973.
Husar, R. B., D. E. Patterson, C. C. Paley, and N. V. Gillani. Ozone in hazy
air masses. I_n: Proceedings of the International Conference on Photo-
chemical Oxidants and Its Control, Research Triangle Park, NC, Sept.
12-20, 1976b.
Husar, R. B. , D. E. Patterson, J. D. Husar, N. V. Gillani, and W. E. Wilson.
Sulfur budget of a power plant plume. Atmos. Environ. 12:549-568, 1978.
Husar, R. B. , et al. , 1976.
Husar, R. B. , N. V. Gillani, J. D. Husar, C. C. Paley, and P. N. Turcu. Long
range transport of pollutants observed through visibility contour maps,
weather maps, and trajectory analysis. Presented at the Third Symposium
on Atmospheric Turbulence, Diffusion, and Air Quality, Raleigh, NC, Oct.
19-22, 1976a.
International Symposium on Sulfur in the Atmosphere (ISSA) Proceedings. Atmos.
Environ. 12:1-769, 1978.
Isaksen, I. S. A., E. Hesstved, and 0. Hov. A chemical model for urban plumes
test for ozone and particulate sulfur formation in St. Louis urban plume.
Atmos. Environ. 12:599-604, 1978.
Johnson, W. B. , D. E. Wolf, and R. L. Mancuso. Long term regional patterns
and transfrentier exchanges of airborne sulfur pollution in Europe.
Atmos. Environ. 12:511-527, 1978.
Judeikis, H. S. The Role of Solid-Gas Interactions in Air Pollution.
EPA-650/3-74-007, U.S. Environmental Protection Agency, August 1974.
Judeikis, H. S. , and S. Siegel. Particle-catalyzed oxidation of atmospheric
pollutants. Atmos. Environ. 7:619-631, 1973.
Junge, C. E. Air Chemistry and Radioactivity. Academic Press Inc., New York,
1963.
Junge, C. E. Die (Constitution der atmospheuschen Aerosols. Ann. Meteorol.
Hamburg 5:1-55, 1952.
Junge, C. E. Die rolle der aerosole und der gasformingen beimengungen der
luft in spurenstoffhaushalt der troposphere. Tellus 5:1-26, 1953.
Kapustin, V. N. , Yu. S. Lynbortseva, and G. V. Rozenberg. Variability of
aerosols under the influence of cloud modulation of the radiation field.
Izv. Atmos. Oceanic Phys. 10:1327-1331, 1974.
6-117
-------
Karraker, D. G. The kinetics of the reaction between sulphurous acid and
ferric ion. J. Phys. Chetn. 67:871-874, 1963.
Komp, M. J., and A. H. Auer, Jr. Visibility reduction and accompanying aerosol
evolution downwind of St. Louis. J. Appl. Meteorol. 17:1357-1367, 1978.
Korshover, J. Climatology of stagnating anticyclones east of the Rocky Mountains
1936-1965. Public Health Service Publication No. 999-AP-34, Cincinnati,
OH, 1967. 15 pp.
Kuhlman, M. R., D. L. Fox, and H. E. Jeffries. The effect of CO on sulfate
aerosol formation. Atmos. Environ. 12:2415-2423, 1978.
Latimer, D. A., R. W. Bergstrom, S. R. Hayes, M. K. Liw, J. H. Seinfeld, G. Z.
Whitten, M. A. Wojcik, and M. J. Hillyar. The Development of mathematical
models for the prediction of anthropogenic visibility impairment. EPA-450/
3-78-110 a, U.S. Environmental Protection Agency, 1978.
Lee, R. E. Jr. The Size of Suspended Particulate Matter in Air. Science
178:567-575, 1972.
Lewis, C..W., and E. S. Macias, 1979.
Lewis, C. W., and E. S. Macias. Composition of size fractioned aerosol in
Charleston, West Virginia. Submitted for publication, Atmos. Environ.,
1980.
Liberti, A., D. Brocco, and M. Possanzini. Adsorption and oxidation of sulfur
dioxide on particles. Atmos. Environ. 12:255-261, 1978.
Lipeles M. , C. S. Burton, H. H. Wang, E. P. Parry, and G. M. Hidy. Mechanisms
of Formation and Composition of Photochemical Aerosols. EPA-650/73-036.
Thousand Oaks, CA: Rockwell International Science Center, 1973. 98 pp.
Loo, B. W., W. R. French, R. C. Gatti, F. S. Goulding, J. M. Jaklevic, J.
Llacer, and A. C. Thompson. Large-scale measurement of airborne particulate
sulfur. Atmos. Environ. 12:759-771, 1978.
Low, R. D. H., 1971.
Lyons, W. A., and R. B. Husar. SMS/GOES visible images detect a synoptic-scale
air pollution episode. Mon. Weather Rev. 104:1623-1626, 1976.
MacCracken, M. C. MAP3S: an investigation of atmospheric energy related
pollutants in the northeastern United States. Atmos. Environ. 12:649-659,
1978.
Macias, E. S. , D. L. Blumenthal, J. A. Anderson, and B. K. Cantrell. Character-
ization of Visibility Reducing Aerosols in the Southwestern United States.
Interim report on Project VISTTA. MRI Report 78 1R, 15-85, 1979.
6-110
-------
McNeil's, D. N. , L. A. Ripperton, W. E. Wilson, P. L. Hanst, and B. W. Gay.
Gas Phase Reactions of Ozone and Olifin in the Presence of Sulfur Dioxide.
^n: Removal of Trace Contaminants from Air, B.R. Dietz, ed., A.C.S.
Symposium Series 17:187-200, 1975.
McMurry, P. H., and S. K. Friedlander. Aerosol formation in reacting gases:
relation of surface area to rate of gas-to-particle conversion. J.
Colloid Interface Sci. 64:248-257, 1978.
Meyers, R. E. , R. T. Cederwall, W. D. Ohmstede, and J. Kampe. Transport and
diffusion using a diagnostic meso-scale model employing mass and total
energy conservation constraints. In: Proceedings of the Third Symposium
on Atmospheric Turbulence, Diffusion and Air Quality, American Meteorological
Society, Boston, MA, 1977. pp. 90-97.
Middleton P. B., and J. R. Brock. Simulation of aerosol kinetics.
Interface Sci. 54:249- , 1976.
J. Colloid
Middleton P. B, S. H. Suck, and J. R. Brock. On the multimodality of density
functions of pollutant aerosols. Atmos. Environ. 11:251-255, 1977.
Middleton, P., and C. S. Kiang. Experimental and theoretical examination of
the formation of sulfuric acid particles. Atmos. Environ. 12:179-185,
1978.
Miller, D.
1978.
F. Precursor effects on
S0? oxidation.
Atmos. Environ. 12:273-280,
Miller, F. J. , D. E. Gradner, J. A. Grahm, R. E. Lu, Jr., W. E. Wilson, and J.
D. Bachman. Size Considerations for Establishing a Standard for Inhalable
Particles. J. Air Pollut. Control Assoc. 29:610-615, 1979.
Miller, M. S. , S. K. Friedlander, and G. M. Hidy. A chemical element balance
for the Pasadena aerosol. J. Colloid Interface Sci. 39:165-176, 1972.
Mueller, P. K. , et al., 1979.
Newman, J. E. , M. D. Abel, P. R. Harrison, and K. J. Yost. Wind as Related to
Critical Flushing Speed Versus Reflotation Speed by High-Volume Sampler
Particulate Loading, 466-496. In: Atmospheric-Surface Exchange of
Particulate and Gaseous Pollutants (1974). Englemann, R. J., ed. Energy
Research and Development Administration symposium series, CONF-740921,
1976.
Neytzell-de Wilde, F. G. , and L. Taverner. Experiments relating to the possible
production of an oxidizing acid leach liquor by auto-oxidation for the
extraction of uranium. lr\: Proceedings of the Second U.N. Conference
Peaceful Uses of Atomic Energy, vol. 3, 1958. pp. 303-317.
Novakov, 1979.
6-119
-------
OECD. The OECD Program on Long-Range Transport of Air Pollutants—Measurements
and Findings. Organization for Economic Cooperation and Development,
Paris, 1977.
Orel A. E. , and J. H. Seinfeld. Nitrate formation in atmospheric aerosols.
Environ. Sci. Technol. 11:1000-1007, 1977.
Orr, C. , F. K. Hurd, and W. J. Corbert. Aerosol size and relative humidity.
J. Colloid Interface Sci. 13:472-482, 1958.
Ottar, B. Monitoring long-range transport of air pollutants: the OECD study.
Ambio 6:203-206, 1976.
Overton, J. H., V. P. Aneja, and J. L. Durham. Production of sulfate in rain
and raindrops in polluted atmospheres. Atmos. Environ. 13:355-367, 1979.
Pack, D. H. , G. J. Ferber, J. L. Hefter, K. Telegadas, J. K. Angel 1, W. H.
Hoecker, and L. Machta. Meteorology of long-range transport. Atmos.
Environ. 12:425-444, 1978.
Penkett, S. A., B. M. R. Jones, K. A. Brice, and A. E. J. Eggleton. The
importance of atmospheric ozone and hydrogen peroxide in oxidizing sulphur
dioxide in cloud and rainwater. Atmos. Environ. 13:123-137, 1979.
Perhac, R. M. Sulfate regional experiment in northeastern United States: the
'SURE' program. Atmos. Environ. 12:641-647, 1978.
Pierson and Russell, 1979.
Roberts, P. T. , and S. K. Friedlander. Photochemical aerosol formation--S02,
1-heptene, and N0x in ambient air. Environ. Sci. Technol. 10:573, 1976.
Rodhe, H. Budgets and turn-over times of atmospheric sulfur compounds.
Atmos. Environ. 12:671-680, 1978.
Rodhe, H., C. Persson, and 0. Akesson. An investigation into regional trans-
port of soot and sulfate aerosols. Atmos. Environ. 6:675-693, 1972.
Ross, D. S. Comments on Arel and Seinfeld. Environ. Sci. Technol. 11:1000,
1977; 12:726, 1978.
Samson, P. J. Ensemble trajectory analysis of summertime sulfate concentrations
in New York. Atmos. Environ. 12:1889-1893, 1978.
Samson, P. J., and K. W. Ragland. Ozone and visibility reduction in the
Midwest: evidence for large scale transport. J. Appl. Meteorol.
16:1101-1106, 1977.
Schuetzle D., D. Cronn, A. L. Crittenden, and R. J. Charlson. Molecular
composition of secondary aerosol and its possible origin. Environ. Sci.
Technol. 9:838-845, 1975.
6-120
-------
Schurmeier, 1980.
Schwartz W. Chemical Characterization of Model Aerosols. EPA-650/3-74-011.
Columbus, Ohio: Battelle Memorial Institute, 1974. pp. 124
Sehmel, G. A., and W. H, Hodgson. Particle dry deposition velocities, I_n:
Atmosphere-Surface Exchange of Particulate and Gaseous Pollutants. 1974
Proceedings of ERDA Symposium Series 38, Richland, WA, Publication No.
CONF-740921 from NTIS. Springfield, VA, 1976.
Sinclair, D. , R. J. Countess, and G. S. Hoopes. Effect of relative humidity
on the size of atmospheric aerosol particles. Atmos. Environ. 8:1111-1118,
1974.
Slinn, W. G. N., 1974.
Slinn, W. G. N. Dry Deposition and Resuspension of Aerosol Particles - A New
Look at Lome Old Problems. In: Atmosphere-Sufrace Exchange of Particulate
and Gaseous Pollutants (1974JT Englemann, R. J. ed. Energy Research and
Development Administration symposium series, CONF-740921, 1976.
Slinn, W. G. N. , L. Hasse, B. B. Hicks, A. W. Hogan, D. Lai, P. S. Liss, K. 0.
Munnich, G. A. Sehmel, and 0. Vittori. Some aspects of the transfer of
atmospheric trace constituents past the air-sea interface. Atmos. Environ.
12:2055-2087, 1978.
Slinn, W. G. , et al., 1979.
Smith, B. M. , J. Wagman, and B. R. Fish. Interaction of airborne particles
with gases. Environ. Sci. Technol. 9:558-562, 1969.
Smith, F. B. , and G. H. Jeffrey. Airborne transport of sulfur dioxide from
the United Kingdom. Atmos. Environ. 9:643-659, 1975.
Smith, F. B. , and R. D. Hunt. Meteorological aspects of the transport of
pollution over long distances. Atmos. Environ. 12:461-477, 1978.
Smith, T. B. , D. L. Blumenthal, J. A. Anderson, and A. H. Vanderpol. Transport
of S0? in Power Plant Plumes: Day and Night. Atmos. Environ. 12:605-611,
1978.
Spicer C. W. , and P. M. Schumacher. Particulate nitrate: Laboratory and
field studies of major sampling interferences. Atmos. Environ. 13:543-552,
1979.
Spicer, C. W. The Fate of Nitrogen Oxides in the Atmosphere. EPA 600/3-76-030,
U.S. Environmental Protection Agency, Research Triangle Park, NC, 1976.
Stelson A. W., S. K. Friedlander, and J. H. Seinfeld. A note on the equilibrium
realtionship between ammonia and nitric acid and particulate ammonium
nitrate. Atmos. Environ. 13:369-371, 1979.
6-121
-------
Stevens, R. K., T. G. Dzubay, G. Russwurm, and D. Rickel. Sampling and analysis
of atmospheric sulfates and related species. Atmos. Environ. 12:55-68,
1978. ~~
Stevens, R. K. , 1979.
Suck, S. H., E. C. Upchurch, and J. R. Brock. Dust transport in Maricopa
County, Arizona. Atmos. Environ. 12:2265- , 1978.
Sykes, R. I., and L. Hatton. Computation of horizontal trajectory based on
the surface geostrophic wind. Atmos. Environ. 10:925-934, 1976.
Szepesi, D. J. Transmission of sulfur dioxide on local, regional and continental
scale. Atmos. Environ. 12:529-535, 1978.
Tang, I. N. Phase transformation and growth of aerosol particles composed of
mixed salts. J. Aerosol Sci. 7:361-371, 1976.
Tang, I. N., and H. R. Munkelwitz. Aerosol growth studies. III. Ammonium
bisulfate aerosol in a moist atmosphere. J. Aerosol Sci. 8:321-330,
1977.
Tang, I. N., 1979.
Tartarelli, R. , P. Davine, F. Morelli, and P. Corsi. Interactions between S0?
and carbonaceous particulates. Atmos. Environ. 12:289-293, 1978.
Titoff, A. Beitrage zur Kaentnis der negativen Katalyze im homogenen System.
Z. Phys. Chem. 45:641-683, 1903.
Tong, E. Y., G. M. Hidy. T. F. Lavery, and F. Berlandi. Regional and local
aspects of atmospheric sulfates in the northeast quadrant of the United
States. J_n: Proceedings Third Symposium of Turbulence, Diffusion and
Air Quality, American Meteorological Society, 1976.
Urone, P., H. Lutsep, C. M. Noyes, and J. F. Parcher. Static studies of
sulfur dioxide reactions in air. Environ. Sci. Technol. 8:611-618, 1968.
Vukovich, F. M. , W. D. Bach, Jr., B. W. Cressman, and W. J. King. On the
relationships between high ozone in the rural boundary layer and high
pressure systems. Atmos. Environ. 11:967-983, 1977.
Wagman, J., R. E. Lee, and C. J. Axt. Influence of Some Atmospheric Variables
on th? Concentration and Particle Size Distribution of Sulfate in Urban
Air. Atmospheric Environ 1:479-489, 1967.
Whitby, K. T. The physical characteristics of sulfur aerosols. Atmos. Environ.
12:135-159, 1978.
Whitby, K. T. and G. M. Sverdrup. California Aerosols: Their Physical and
Chemical Characteristics. In: The Characters and Origin of Smog Aerosols.
G. M. Hidy et al., eds. John Wiley, New York, 1980.
6-122
-------
Whitby, K. T. , R. B. Hausar, and B. Y. H. Lui. The Aerosol Size Distribution
of Los Angeles Smog. J. Colloid Interfrace Sci. 39:177-204, 1972.
Whitby, K. T. , R. E. Charlson, W. E. Wilson, and R. K. Stevens. The Size of
Suspended Particle Matter in Air. Science 183:1098-1199, 1974.
White, W. H. , and P. T. Roberts, 1975.
White, W. H. , J. A. Anderson, D. L. Blumenthal, R. B. Husar, N. V. Gillani, J.
D. Husar, and W. E. Wilson. Formation and transport of secondary air
pollutants: ozone and aerosols in the St. Louis urban plume. Science
194:187-189, 1976.
Wilson, W. E. Sulfates in the atmosphere: a progress report on Project
MISTT. Atmos. Environ. 12:537-548, 1978.
Wilson, W. E. Transformation during transport: a state of the art survey of
the conversion of S0~ to sulfate. Presented at WHO Symposium, Sofia,
Bulgaria, Oct. 1-5, 1979.
Wilson, W. E., D. F. Miller, A. Levy, and R. K. Stone. The effect of Fuel
Composition on Atmospheric Aerosol Due to Auto Exhaust. J. Air Pollut.
Control Assoc. 23:949-956, 1973.
Wilson, W. E. , et al., 1979.
Wilson, W. E. , Jr., D. F. Miller, A. Levy, and R. K. Stone. J. Air Pollut.
Control Assoc. 23:949, 1973.
Wilson, W. E., L. L. Spiller, T. G. Ellestad, P. L. J. Lamothe, T. G. Dzuby,
R. K. Stevens, E. S. Macias, R. A. Fletcher, J. D. Jusar, R. B. Husar, K.
T. Whitby, D. B. Kittelson, and B. K. Cantrell. General Motors Sulfate
Dispersion Experiment: Summary of EPA Measurements. J. Air Pollut.
Control Assoc. 27:46-51, 1977.
Wilson, W. E. , R. J. Charlson, R. B. Hasar, K. T. Whitby, and D. L. Blumenthal
Sulfates in the atmosphere. Presented at the 69th Annual Meeting of Air
Pollution Control Assoc., Portland, OR, June 27-July 1, 1976.
Winchester, J. W., R. J. Ferek, D. R. Lawson, J. 0. Pilotte, M. A. Thiemens,
and L. E. Wagon. Comparison of aerosol sulfur and crystal element con-
centrations in particle size fractions from continental United States
locations. Submitted for publication, 1979.
Winkler, P. The growth of atmospheric aerosol particles as a function of the
relative humidity. II. An improved concept of mixed nuclei. Aerosol
Sci. 4:373-387, 1973.
Winkler, P., and C. Junge. The growth of atmospheric aerosol particles as a
function of the relative humidity. I. Methods and measurements of
different locations. J. Rech. Atmos. 6:617-638, 1972.
6-123
-------
Wolff, G. T., P. J. Lioy, R. E. Meyers, and R. T. Cederwall. An investigation
of long-range transport of ozone across the Midwestern and Eastern United
States. Atmos. Environ. 11:797-802, 1977.
6-124
-------
Note to Reader: This material was received too late
to be incorporated in the body of the chapter.
6.8 STATUS OF SOURCE APPORTIONMENT RESEARCH
Federal, state, and local agencies generally use atmospheric dispersion
models to provide guidelines in the control of particulate loadings in the
atmosphere. These source oriented dispersion models based on emission inven-
tories and meteorological parameters were developed to predict the impact of a
particulate emission source on a receptor site.
An alternative to the predictive dispersion models to assist in air
quality management is the application of receptor oriented models. The
Chemical Element Balance (CEB) Model is one type of receptor model that has
been widely used over the past ten years to measure the impact of particulate
sources on air quality (1-3).
These CEB models assume that the elemental composition of ambient parti-
cles collected at a receptor site is a linear combination of the components of
the particulate matter originating from various sources. In theory, know-
ledge of the elemental composition of the ambient air particles and the
emissions of all important sources permits the solution of a set of simul-
taneous equations which will provide quantitatively the contributions of each
source of the aerosol to a selected receptor.
In practice, the information is never as complete or reliable as desired,
so the assumptions that go into the model must be simplified. Typically,
rather than use all the emission data from each source, a set of marker ele-
ments is used to characterize a few prominent sources. The marker elements
are normally those that are strongly associated with specified sources.
6-A
-------
Examples are lead and bromine with motor vehicles, sodium with sea salt, and
Vanadium or nickel with combustion of fuel oil.
45 6
Kowalczyk, et al. , Watson , and Dzubay have reported results from CEB
analysis on filters collected from networks of samples in the three cities,
Washington, DC, Portland, OR, and St. Louis, MO, respectively. Watson and
Dzubay analyzed aerosol that was fractionated into two size ranges to sep-
arate primary from secondary aerosols. Figure lisa figure from the work of
Dzubay which shows a comparison of the average mass concentration in the
coarse fraction (particles >2.5 pm) and the components deduced in the chemical
element balance. The corresponding fine fraction was mainly composed of
sulfate.
These previous applications of CEB to source apportionment studies suffer
from the following shortcomings (1) all major sources contributing to the
ambient aerosols at the time of the study were not well characterized, (2) the
computational methods used had not been verified nor their sensitivities to
variations in input parameters established, and (3) ambient and source aerosol
characterization was incomplete.
Whereas CEB methods apply knowledge about source characterization to a
relatively small set of filters to derive a source contribution, multi variate
analysis methods, such as factor analysis pattern recognition methods extract
information about a source contribution on the basis of the variability of
elements measured on large numbers of filters. In factor analysis, data on
the concentrations of each element in each sample of the data set are manipu-
lated to find groupings of variables (common factors) that best explain the
variations of elemental composition from their average values.
Factor analysis has been applied by Hopke et al. to a set of 18 elements
in 90 samples from the Boston area. From this data set, 77.5 percent of the
6-B
-------
variance could be accounted for by six common factors. From the composition
of the factors, primary airborne particles could be attributed to several
sources: soil mixed with emission from coal fired power plants, sea salt,
combusion of fuel oil, auto emissions, and incineration of refuse. A sixth
factor, with large loadings for only manganese and selenium could not be
identified with a particular source.
Factor analysis has several advantages over CEB in that no a priori
assumptions need to be made about either the number or composition of the
sources. Thus, secondary particles that become associated with primary par-
ticulates between release and collection can be incorporated in the analysis.
o
For example, Gaarenstroom found that sulfate, ammonium, and nitrate ions in
the ambient aerosol can, be included in factor analysis.
The major weaknesses in factor analysis is that it requires a data set
where there are large variation in the concentration of the elements from
sample to sample. In addition, factor analysis can only provide information
on the number of source contributing to a receptor but not the magnitude of
their contribution.
9
The source 'of aerosols can also be determined by microscopic methods .
These methods have high resolving capabilities for sources with characteristic
morphological features such as wood fiber, pollen,quartz, mica, and tire dust.
To be quantitative, however, one must estimate the number of particles, their
density and volume and must analyze enough particles to be representative of
the total sample.
Microscopic methods can be divided into two categories: optical, which
is generally limited to particles greater than 1 urn, and scanning electron
microscopy (SEM) which is applicable to smaller particles.
6-C
-------
Recently, Lee et al. coupled an SEM with an energy dispersive x-ray
fluorescence system and automated the instrument to scan and analyze a large
number of particles. The system promises to extend the general applicability
of microscopic methods.
Davis has applied x-ray transmission and diffraction techniques to the
analysis of hi-volume filters collected in Rapid City, SD-. These analyses
revealed that a majority of the aerosol pollution in Rapid City was a result
of emissions from quarrying activities near the city. These x-ray diffraction
techniques show promise of being able to quantitatively measure the source of
aerosol pollution in variety of urban locations where reentrained dust is a
major aerosol pollution problem.
6-D
-------
REFERENCES FOR SECTION 6.8
1. Miller, M. S., S. K. Friedlander, and G. M. Hidy. A chemical element
balance for Pasedena aerosol. J. Colloid Interface Sci. 3_9:165, 1972.
2. Friedlander, S. K. Chemical element balances and identification of air
pollution sources. Environ. Sci. Techno!. 7:235-240, 1973.
3. Gatz, D. F. Relative contributions of different sources of urban
aerosols: Application of new estimation method to multiple sites in
Chicago. Atmos. Environ. 9:1-18, 1975.
4. Kowalczyk, G. S., C. E. Choquette, and G. E. Gordon. Chemical element
balances and identification of air pollution sources in Washington, D.C.
Atmos. Environ. 12:1143-1153, 1978.
5. Watson, J. G., Jr. Chemical Element Balance Receptor Methodology for
Assessing the Sources of Fine and Total Suspended Particulate Matter in
Portland, Oregon. Ph.D. Thesis, Department of Chemistry, Oregon Graduate
Center, Beaverton, OR, 1979.
6. Dzubay, T. G. Chemical element balance method applied to dichotomous
sampler data. In: Proceedings of the Conference on Aerosol: Anthro-
pogenic and Natural -- Sources and Transport. New York Academy of
Sciences, in press, 1979.
7. Hopke, P. K. , E. S. Gladney, G. E. Gordon, W. H. Zoller, and A. G. Jones.
The use of multivariate analysis to identify sources of selected elements
in the Boston urban aerosol. Atmos. Environ. 10:1015-1025, 1976.
8. Gaarenstroom, P. D., S. P. Perone, and J. L. Moyers. Application of
pattern recognition and factor analysis for characterization of atmo-
spheric particulate composition in the Southwest Desert atmosphere.
Environ. Sci. Techno!. 11:795-800, 1977.
9. Graf, J., R. H. Snow, and R. G. Draftz. Aerosol Sampling and Analysis -
Phoenix, Arizona. EPA-600/2-77-015, U.S. Environmental Protection Agency,
Washington, DC, 1977.
10. Lee, R. J. , E. J. Fasiska, P. Jonocko, D. McFarland, and S. Penicala.
Electron beam particle analysis. Ind. Res. Dev. 21:105, 1979.
11. Davis, B. L. Additional suggestions for x-ray quantitative analysis of
high volume filter samples. Atmos. Environ. 12:1403-1406, 1978.
6-E
-------
GLOSSARY
AaOOy Alveolar-arterial difference or gradient of the partial pressure of
oxygen. An overall measure of the efficiency of the lung as a gas ex-
changer. In healthy subjects, the gradient is 5 to 15 mm Hg (torr).
A/PR/8 virus: A type of virus capable of causing influenza in laboratory
animals; also, A/PR/8/34.
Abscission: The process whereby leaves, leaflets, fruits, or other plant
parts become detached from the plant.
Absorption coefficient: A quantity which characterizes the attenuation with
distance of a beam of electromagnetic radiation (like light) in a substance
Absorption spectrum: The spectrum that results after any radiation has
passed through an absorbing substance.
Abstraction: Removal of some constituent of a substance or molecule.
Acetaldehyde: CH^CHO; an intermediate in yeast fermentation of car-
bohydrate anfl in alcohol metabolism; also called acetic aldehyde,
ethaldehyde, ethanal.
Acetate rayon: A staple or filament fiber made by extrusion of cellulose
acetate. It is saponified by dilute alkali whereas viscose rayon remains
unchanged.
Acetylchol ine: A naturally-occurring substance in the body which can
cause constriction of the bronchi in the lungs.
Acid: A substance that can donate hydrogen ions.
Acid dyes: A large group of synthetic coal tar-derived dyes which
produce bright shades in a wide color range. Low cost and ease
of application are features which make them the most widely used
dyes for wool. Also used on nylon. The term acid dye is derived
from their precipitation in an acid bath.
Acid mucopolysaccharide: A class of compounds composed of protein
and polysaccharide. Mucopolysaccharides comprise much of the
substance of connective tissue.
Acid phosphatase: An enzyme (EC 3.1.3.2) which catalyzes the disassociation
of phosphate (PO.) from a wide range of monoesters of orthophosphoric
acid. Acid phosphatase is active in an acidic pH range.
Acid rain: Rain having a pH less than 5.6, the minimum expected from
atmospheric CO-.
G-1
-------
Acrolein: CH2=CHCHO; a volatile, flammable, oily liquid, giving off
irritant vapor. Strong irritant of skin and mucuous membranes. Also
called acrylic aldehyde, 2-propenal.
Acrylics (plastics): Plastics which are made from acrylic acid and are
light in weight, have great breakage resistance, and a lack of
odor and taste. Not resistant to scratching, burns, hot water,
alcohol or cleaning fluids. Examples include Lucite and Plexiglass.
Acrylics are thermoplastics and are softened by heat and hardened
into definite shapes by cooling.
Acrylic fiber: The generic name of man-made fibers derived from acrylic
resins (minimum of 85 percent acrylonitrite units).
Actinic: A term applied to wavelengths of light too small to affect
one's sense of sight, such as ultraviolet.
Actinomycetes: Members of the genus Actinomyces; nonmotile, nonspore-
forming, anaerobic bacteria, including both soil-dwelling saprophytes
and disease-producing parasites.
Activation energy: The energy required to bring about a chemical reaction.
Acute respiratory disease: Respiratory infection, usually with rapid
onset and of short duration.
Acute toxicity: Any poisonous effect produced by a single short-term
exposure, that results in severe biological harm or death.
Acyl: Any organic radical or group that remains intact when an organic
acid forms an ester.
Adenoma: An ordinarily benign neoplasm (tumor) of epithelial tissue;
usually well circumscribed, tending to compress adjacent tissue rather
than infiltrating or invading.
Adenosine monophosphate (AMP): A nucleotide found amoung the hydrolysis
products of all nucleic acids; also called adenylic acid.
Adenosine triphosphatase (ATPase): An enzyme (EC 3.6.1.3) in muscle
and elsewhere that catalyzes the release of the high-energy, ter-
minal phosphate group of adenosine triphosphate.
Adrenalectomy: Removal of an adrenal gland. This gland is located near
or upon the kidney and is the site of origin of a number of hormones.
Adsorption: Adhesion of a thin layer of molecules to a liquid or solid sur-
face.
Advection: Horizontal flow of air at the surface or aloft; one of the
means by which heat is transferred from one region of the earth
to another.
6-2
-------
Aerodynamic diameter: Expression of aerodynamic behavior of an irregularly
shaped particle in terms of the diameter of a sphere of unit density
having identical aerodynamic behavior to the particle in question.
Aerosol: Solid particles or liquid droplets which are dispersed or sus-
pended in a gas.
Agglutination: The process by which suspended bacteria, cells or similar
particles adhere and form into clumps.
Airborne pathogen: A disease-causing microorganism which travels in the
air or on particles in the air.
Air pollutant: A substance present in the ambient atmosphere, resulting
from the activity of man or from natural processes, which may cause
damage to human health or welfare, the natural environment, or
materials or objects.
Airway conductance: Inverse of airway resistance.
Airway resistance (R ): The pressure difference between the alveoli
and the mouth required to produce an air flow of 1 liter per second.
Alanine aminotransferase: An enzyme (EC 2.6.1.2) transferring amino
groups from L-alanine to 2-ketoglutarate. Also known as alanine
transaminase.
Albumin: A type of simple, water-soluble protein widely distributed
throughout animal tissues and fluids, particularly serum.
0
n
Aldehyde: An organic compound characterized by the group -C-H.
Aldolase: An enzyme (EC 4.1.2.7) involved in metabolism of fructose
which catalyzes the formation of two 3-carbon intermediates in the
major pathway of carbohydrate metabolism.
Algal bloom: Sudden spurt in growth of algae which can affect water
quality adversely.
Alkali: A salt of sodium or potassium capable of neutralizing acids.
Alkaline phosphatase: A phosphatase (EC 3.1.3.1) with an optimum pH of
8.6, present ubiquitously.
Allergen: A material that, as a result of coming into contact with appro-
priate tissues of an animal body, induces a state of sensitivity result-
ing in various reactions; generally associated with idiosyncratic
hypersensitivities.
Alpha-hydroxybutyrate dehydrogenase: An enzyme (EC 1.1.1.30), present
mainly in mitochondria, which catalyzes the conversion of hydro-
xybutyrate to acetoacetate in intermediate biochemical pathways.
6-3
-------
Alpha rhythm: A rhythmic pulsation obtained in brain waves exhibited
in the sleeping state of an individual.
Alveolar capillary membrane: Finest portion of alveolar capillaries,
where gas transfer to and from blood takes place.
Alveolar macrophages (AM): Large, mononuclear, phagocytic cells found
on the alveolar surface, responsible for the sterility of the lung.
Alveolar oxygen partial pressure (PAOp): Partial pressure of oxygen in the
air contained in the air sacs of the lungs.
Alveolar septa: The tissue between two adjacent pulmonary alveoli, con-
sisting of a close-meshed capillary network covered on both surfaces
by thin alveolar epithelial cells.
Alveolus: An air cell; a terminal, sac-like dilation in the lung. Gas
exchange (02/C02) occurs here.
Ambient: The atmosphere to which the general population may be exposed.
Construed here not to include atmospheric conditions indoors, or in
the workplace.
Amine: A substance that may be derived.from ammonia (NH_) by the re-
placement of one, two or three of the hydrogen (H) atoms by hydro-
carbons or other radicals (primary, secondary or tertiary amines,
respectively).
Amino acids: Molecules consisting of a carboxyl group, a basic amino
group, and a residue group attached to a central carbon atom. Serve
as the building blocks of proteins.
p-Aminohippuric acid (PAH): A compound used to determine renal plasma
flow.
Aminotriazole: A systemic herbicide, Co^A^A' used in areas other than
croplands, that also possesses some antithyroid activity; also called
amitrole.
Ammonification: Decomposition with production of ammonia or ammonium
compounds, esp. by the action of bacteria on nitrogenous organic
matter.
ium: Anion (NH.) or radical (NhL) derived from ammonia by combination
tfith hydrogen. Present in rainwater, soils and many commercial ferti-
Ammoni
with
lizers.
Amnestic: Pertains to immunologic memory: upon receiving a second
dose of antigen, the host "remembers" the first dose and responds
faster to the challenge.
6-4
-------
Anaerobic: Living, active or occurring in the absence of free oxygen.
Anaerobic bacteria: A type of microscopic organism which can live in
an environment not containing free oxygen.
Anaphylactic dyspneic attack: Difficulty in breathing associated with
a systemic allergic response.
Anaphylaxis: A term commonly used to denote the immediate, transient
kind of immunological (allergic) reaction characterized by contraction
of smooth muscle and dilation of capillaries due to release of pharmacologically
active substances.
Angiosperm: A plant having seeds enclosed in an ovary; a flowering plant.
Angina pectoris: Severe constricting pain in the chest which may be
caused by depletion of oxygen delivery to the heart muscle; usually
caused by coronary disease.
0 -8
Angstrom A: A unit (10 cm) used in the measurement of the wavelength
of light.
Anhydride: A compound resulting from removal of water from two molecules
of a carboxylic (-COOH) acid. Also, may refer to those substances
(anhydrous) which do not contain water in chemical combination.
Anion: A negatively charged atom or radical.
Anorexia: Diminished appetite; aversion to food.
Anoxic: Without or deprived of oxygen.
Anthraquinone: A yellow crystalline ketone, C,.Hg02, derived from
anthracene and used in the manufacture of dyes.
Anthropogenic: Of, relating to or influenced by man. An anthropogenic
source of pollution is one caused by man's actions.
Antibody: Any body or substance evoked by the stimulus of an antigen
and which reacts specifically with antigen in some demonstrable way.
Antigen: A material such as a foreign protein that, as a result of
coming in contact with appropriate tissues of an animal, after a latent
period, induces a state of sensitivity and/or the production of antibody.
Antistatic agent: A chemical compound applied to fabrics to reduce or
eliminate accumulation of static electricity.
Arachidonic acid: Long-chain fatty-acid which serves as a precursor
of prostaglandins.
G-5
-------
Area source: In air pollution, any small individual fuel combustion
or other pollutant source; also, all such sources grouped over a
specific area.
Aromatic: Belonging to that series of carbon-hydrogen compounds in
which the carbon atoms form closed rings containing unsaturated
bonds (as in benzene).
Arterial partial pressure of oxygen (PaCL): Portion of total pressure of
dissolved gases in arterial blood as measured directly from arterial
blood.
Arterialized partial pressure of oxygen: The portion of total pressure
of dissolved gases in arterial blood attributed to oxygen, as
measured from non-arterial (e.g., ear-prick) blood.
Arteriosclerosis: Commonly called hardening of the arteries. A condition
that exists when the walls of the blood vessels thicken and become
infiltrated with excessive amounts of minerals and fatty materials.
Artifact: A spurious measurement produced by the sampling or analysis
process.
Ascorbic acid: Vitamin C, a strong reducing agent with antioxidant proper-
ties.
Aspartate transaminase: Also known as aspartate aminotransferase
(EC 2.6.1.1). An enzyme catalyzing the transfer of an amine group
from glutamic acid to oxaloacetic, forming aspartic acid in the
process. Serum level of the enzyme is increased in myocardial in-
farction and in diseases involving destruction of liver cells.
Asphyxia: Impaired exchange of oxygen and carbon dioxide, excess of
carbon dioxide and/or lack of oxygen, usually caused by ventilatory
problems.
Asthma: A term currently used in the context of bronchial asthma in
which there is widespread narrowing of the airways of the lung.
It may be aggravated by inhalation of pollutants and lead to
"wheezing" and shortness of breath.
Asymptomatic: Presenting no subjective evidence of disease.
Atmosphere: The body of air surrounding the earth. Also, a measure of
pressure (atm.) equal to the pressure of air at sea level, 14.7 pounds
per square inch.
Atmospheric deposition: Removal of pollutants from the atmosphere onto
land, vegetation, water bodies or other objects, by absorption,
sedimentation, Brownian diffusion, impaction, or precipitation in rain.
6-6
-------
Atomic absorption spectrometry: A measurement method based on the
absorption of radiant energy by gaseous ground-state atoms. The
amount of absorption depends on the population of the ground state
which is related to the concentration of the sample being analyzed.
Atropine: A poisonous white crystalline alkaloid, C,7H?^NO_, from
belladonna and related plants, used to relieve Spasms and to dilate
the pupil of the eye.
Autocorrelation: Statistical interdependence of variables being analyzed;
produces problems, for example, when observations may be related
to previous measurements or other conditions.
Autoimmune disease: A condition in which antibodies are produced against
the subject's own tissues.
Autologous: A term referring to cellular elements, such as red blood cells
and alveolar macrophage, from the same organism; also, something
natually and normally ocurring in some part of the body.
Autotrophic: A term applied to those microorganisms which are able to
maintain life without an exogenous organic supply of energy, or which
only need carbon dioxide or carbonates and simple inorganic nitrogen.
Autotrophic bacteria: A class of microorganisms which require only
carbon dioxide or carbonates and a simple inorganic nitrogen com-
pound for carrying on life processes.
Auxin: An organic substance that causes lengthening of the stem when
applied in low concentrations to shoots of growing plants.
Awn: One of the slender bristles that terminate the glumes of the
spike!et in some cereals and other grasses.
Azo dye: Dyes in which the azo group is the chromophore and joins
benzene or napthalene rings.
Background measurement: A measurement of pollutants in ambient air due
to natural sources; usually taken in remote areas.
Bactericidal activity: The process of killing bacteria.
Barre: Bars or stripes in a fabric, caused by uneven weaving, irregular
yarn or uneven dye distribution.
Basal cell: One of the innermost cells of the deeper epidermis of the
skin.
Benzenethiol: A compound of benzene and a hydrosulfide group.
G-7
-------
Beta (b)-lipoprotein: A biochemical complex or compound containing both
lipid and protein and characterized by having a large molecular
weight, rich in cholesterol. Found in certain fractions of human
plasma.
Bilateral renal sclerosis: A hardening of both kidneys of chronic
inflammatory origin.
Biomass: That part of a given habitat consisting of living matter.
Biosphere: The part of the earth's crust, waters and atmosphere where
living organisms can subsist.
Biphasic: Having two distinct successive stages.
Bleb: A collection of fluid beneath the skin; usually smaller than
bullae or blisters.
Blood urea: The chief end product of nitrogen metabolism in mammals,
excreted in human urine in the amount of about 32 grams (1 oz.)
a day.
Bloom: A greenish-gray appearance imparted to silk and pile fabrics
either by nature of the weave or by the finish; also, the creamy
white color observed on some good cottons.
Blue-green algae: A group of simple plants which are the only N^-fixing
organisms which photosynthesize as do higher plants.
Brightener: A compound such as a dye, which adheres to fabrics in order
to provide better brightness or whiteness by converting ultraviolet
radiation to visible light. Sometimes called optical bleach or
whitening agent. The dyes used are of the florescent type.
Broad bean: The large flat edible seed of an Old World upright vetch
(Vicia faba), or the plant itself, widely grown for its seeds and
for fodder.
Bronchi: The first subdivisions of the trachea which conduct air to
and from the bronchioles of the lungs.
Bronchiole: One of the finer subdivisions of the bronchial (trachea)
tubes, less than 1 mm in diameter, and having no cartilage in
its wall.
Bronchiolitis: Inflammation of the smallest bronchial tubes.
Bronchiolitis fibrosa obliterans syndrome: Obstruction of the bronchioles
by fibrous granulation arising from an ulcerated mucosa; the condition
may follow inhalation of irritant gases.
G-8
-------
Bronchitis: Inflammation of the mucous membrane of the bronchial tubes.
It may aggravate an existing asthmatic condition.
Bronchoconstrictor: An agent that causes a reduction in the caliber
(diameter) of a bronchial tube.
Bronchodilator: An agent which causes an increase in the caliber (diameter)
of a bronchus or bronchial tube.
Bronchopneumonia: Acute inflammation of the walls of the smaller bronchial
tubes, with irregular area of consolidation due to spread of the in-
flammation into peribronchiolar alveoli and the alveolar ducts.
Brownian diffusion: Diffusion by random movement of particles suspended
in liquid or gas, resulting from the impact of molecules of the
fluid surrounding the particles.
Buffer: A substance in solution capable of neutralizing both acids
and bases and thereby maintaining the original pH of the solution.
Buffering capacity: Ability of a body of water and its watershed to
neutralize introduced acid.
Butanol: A four-carbon, straight-chain alcohol, C.HqOH, also known as
butyl alcohol.
Butylated hydroxytoluene (BHT): A crystalline phenolic antioxidant.
Butylated hydroxyanisol (BHA): An antioxidant.
14
C labeling: Use of a radioactive form of carbon as a tracer, often
in metabolic studies.
14
C-proline: An amino acid which has been labeled with radioactive carbon.
Calcareous: Resembling or consisting of calcium carbonate (lime), or
growing on limestone or lime-containing soils.
Calorie: Amount of heat required to raise temperature of 1 gram of
water at 15°C by 1 degree.
Cannula: A tube that is inserted into a body cavity, or other tube
or vessel, usually to remove fluid.
Capillary: The smallest type of vessel; resembles a hair. Usually
in reference to a blood or lymphatic capillary vessel.
Carbachol: A chemical compound (carbamoylcholine chloride, CgH,cClN_CL) that
produces a constriction of the bronchi; a parasympathetic stimulant
used in veterinary medicine and topically in glaucoma.
6-9
-------
Carbon monoxide: An odorless, colorless, toxic gas with a strong affinity
for hemoglobin and cytochrome; it reduces oxygen absorption capacity,
transport and utilization.
Carboxyhemoglobin: A fairly stable union of carbon monoxide with hemo-
globin which interferes with the normal transfer of carbon dioxide
and oxygen during circulation of blood. Increasing levels of
Carboxyhemoglobin result in various degrees of asphyxiation, in-
cluding death.
Carcinogen: Any agent producing or playing a stimulatory role in the
formation of a malignancy.
Carcinoma: Malignant new growth made up of epithelial cells tending to
infiltrate the surrounding tissues and giving rise to metastases.
Cardiac output: The volume of blood passing through the heart per unit
time.
Cardiovascular: Relating to the heart and the blood vessels or the
circulation.
Carotene: Lipid-soluble yellow-to-orange-red pigments universally
present the photosynthetic tissues of higher plants, algae, and the
photosynthetic bacteria.
Cascade impactor: A device for measuring the size distribution of particulates
and/or aerosols, consisting of a series of plates with orifices of
graduated size which separate the sample into a number of fractions
of decreasing aerodynamic diameter.
Catabolism: Destructive metabolism involving the release of energy and
resulting in breakdown of complex materials in the organism.
Catalase: An enzyme (EC 1.11.1.6) catalyzing the decomposition of hydrogen
peroxide to water and oxygen.
Catalysis: A modification of the rate of a chemical reaction by some
material which is unchanged at the end of the reaction.
Catalytic converter: An air pollution abatement device that removes
organic contaminants by oxidizing them into carbon dioxide and
water.
Catecholamine: A pyrocatechol with-an alkalamine side chain, functioning
as a hormone or neurotransmitter, such as epinephrine, morepinephrine,
or dopamine.
Cathepsins: Enzymes which have the ability to hydrolyze certain proteins
and peptides; occur in cellular structures known as lysosomes.
Cation: A positively charged ion.
G-10
-------
Cellular permeability: Ability of gases to enter and leave cells; a
sensitive indicator of injury to deep-lung cells.
Cellulose: The basic substance which is contained in all vegetable
fibers and in certain man-made fibers. It is a carbohydrate and
constitutes the major substance in plant life. Used to make cellulose
acetate and rayon.
Cellulose acetate: Commonly refers to fibers or fabrics in which the
cellulose is only partially acetylated with acetate groups. An
ester made by reacting cellulose with acetic anhydride with SO.
as a catalyst.
Cellulose rayon: A regenerated cellulose which is chemically the same
as cellulose except for physical differences in molecular weight
and crystal!inity.
Cellulose triacetate: A cellulose fiber which is completely acetylated.
Fabrics of triacetate have higher heat resistance than acetate and
may be safely ironed at higher temperature. Such fabrics have improved
ease-of-care characteristics because after heat treatment during
manufacture, a change in the crystalline structure of the fiber
occurs.
Cellulosics: Cotton, viscose rayon and other fibers made of natural fiber
raw materials.
Celsius scale: The thermometric scale in which freezing point of water
is 0 and boiling point is 100.
Central hepatic necrosis: The pathologic death of one or more cells,
or of a portion of the liver, involving the cells adjacent to the
central veins.
Central nervous system (CNS): The brain and the spinal cord.
Centroacinar area: The center portion of a grape-shaped gland.
Cerebellum: The large posterior brain mass lying above the pons and
medulla and beneath the posterior portion of the cerebrum.
Cerebral cortex: The layer of gray matter covering the entire surface
of the cerebral hemisphere of mammals.
Chain reaction: A reaction that stimulates its own repetition.
Challenge: Exposure of a test organism to a virus, bacteria, or other
stress-causing agent, used in conjunction with exposure to a pollutant
of interest, to explore possible susceptibility brought on by the
pollutant.
G-ll
-------
Chamber study: Research conducted using a closed vessel in which pollutants
are reacted or substances exposed to pollutants.
Chemiluminescence: A measurement technique in which radiation is pro-
duced as a result of chemical reaction.
Chemotactic: Relating to attraction or repulsion of living protoplasm
by chemical stimuli.
Chlorophyll: A group of closely related green photosynthetic pigments
occurring in leaves, bacteria, and organisms.
Chloroplast: A plant cell inclusion body containing chlorophyll.
Chlorosis: Discoloration of normally green plant parts that can be
caused by disease, lack of nutrients, or various air pollutants,
resulting in the failure of chlorophyll to develop.
Cholesterol: A steroid alcohol C^yH.j-OH; the most abundant steroid in
animal cells and body fluids?
Cholinesterase (CHE): One (EC 3.1.1.8) of a family of enzymes capable
of catalyzing the hydrolysis of acylcholines.
Chondrosarcoma: A malignant neoplasm derived from cartilage cells,
occurring most frequently near the ends of long bones.
Chromatid: Each of the two strands formed by longitudinal duplication
of a chromosome that becomes visible during an early stage of cell
division.
Chromophore: A chemical group that produces color in a molecule by absorbing
near ultraviolet or visible radiation when bonded to a nonabsorb-
ing, saturated residue which possesses no unshared, nonbonding valence
electrons.
Chromosome: One of the bodies (46 in man) in the cell nucleus that is the
bearer and carrier of genetic information.
Chronic respiratory disease (CRD): A persistent or long-lasting intermittent
disease of the respiratory tract.
Cilia: Motile, often hairlike extensions of a cell surface.
Ciliary action: Movements of cilia in the upper respiratory tract, which
move mucus and foreign material upward.
Ciliogenesis: The formation of cilia.
G-12
-------
Citric acid (Krebs) cycle: A major biochemical pathway in cells, in-
volving terminal oxidation of fatty acids and carbohydrates. It
yields a major portion of energy needed for essential body functions
and is the major source of CO^. It couples the glycolytic breakdown
of sugar in the cytoplasm witn those reactions producing ATP in the
mitochondria. It also serves to regulate the synthesis of a number
of compounds required by a cell.
Clara cell: A nonciliated mammalian cell.
Closing volume (CV): The lung volume at which the flow from the lower
parts of the lungs becomes severely reduced or stops during expiration
presumably because of airway closure.
Codon: A sequence of three nucleotides which encodes information re-
quired to direct the synthesis of one or more amino acids.
Coefficient of haze (COH): A measurement of visibility interference in the
atmosphere.
Cohort: A group of subjects included in a test or experiment; usually
characterized by age, class or other characteristic.
Collagen: The major protein of the white fibers of connective tissue,
cartilage, and bond. Comprises over half the protein of the mammal.
Collisional deactivation: Reduction in energy of excited molecules
caused by collision with other molecules or other objects such
as the walls of a container.
Colorimetric: A chemical analysis method relying on measurement of the
degree of color produced in a solution by reaction with the pollutant
of interest.
Community exposure: A situation in which people in a sizeable area are
subjected to ambient pollutant concentrations.
Compliance: A measure of the change in volume of an internal organ (e.g.
lung, bladder) produced by a unit of pressure.
Complement: Thermolabile substance present in serum that is destructive
to certain bacteria and other cells which have been sensitized by
specific complement-fixing antibody.
Compound: A substance with its own distinct properties, .formed by the
chemical combination of two or more elements in fixed proportion.
Concanavalin-A: One of two crystalline globulins occurring in the jack
bean; a potent hemagglutinin.
Conifer: A plant, generally evergreen, needle-leafed, bearing naked seeds
singly or in cones.
G-13
-------
Converter: See catalytic converter.
Coordination number: The number of bonds formed by the central atom in
a complex.
Copolymer: The product of the process of polymerization in which two or
more monomeric substances are mixed prior to polymerization. Nylon is
a copolymer.
Coproporphyrin: One of two porphyrin compounds found normally in feces
as a decomposition product of bilirubin (a bile pigment). Porphyrin
is a widely-distributed pigment consisting of four pyrrole nuclei
joined in a ring.
Cordage: A general term which includes banding, cable, cord, rope, string,
and twine made from fibers. Synthetic fibers used in making cordage
include nylon and dacron.
Corrosion: Destruction or deterioration of a material because of reaction
with its environment.
Corticosterone: A steroid obtained from the adrenal cortex. It induces
some deposition of glycogen in the liver, sodium conservation, and
potassium excretion.
Cosmopolitan: In the biological sciences, a term denoting worldwide
distribution.
Coulometric: Chemical analysis performed by determining the amount of a
substance released in electrolysis by measuring the number of
coulombs used.
Coumarin: A toxic white crystalline lactone (CgHgO-) found in plants.
Coupler: A chemical used to combine two others in a reaction, e.g. to
produce the azo dye in the Griess-Saltzman method for NCL.
Crevice corrosion: Localized corrosion occurring within crevices on metal
surfaces exposed to corrosives.
Crosslink: To connect, by an atom or molecule, parallel chains in a complex
chemical molecule, such as a polymer.
Cryogenic trap: A pollutant sampling method in which a gaseous pollutant
is condensed out of sampled air by cooling (e.g. traps in one method
for nitrosamines are maintained below -79 C, using solvents maintained
at their freezing points).
Cuboidal: Resembling a cube in shape.
Cultivar: An organism produced by parents belonging to different species
or to different strains of the same species, originating and persist-
ing under cultivation.
G-14
-------
Cuticle: A thin outer layer, such as the thin continuous fatty film
on the surface of many higher plants.
Cyanosis: A dark bluish or purplish coloration of the skin and mucous
membrane due to deficient oxygenation of the blood.
Cyclic GMP: Guanosine 5'-phosphoric acid.
Cytochrome: A class of hemoprotein whose principal biological function
is electron and/or hydrogen transport.
Cytology: The anatomy, physiology, pathology and chemistry of the cell.
Cytoplasm: The substance of a cell exclusive of the nucleus.
Dacron: The trade name for polyester fibers made by E.I. du Pont de Nemours
and Co., Inc., made from dimethyl terephthalate and ethylene glycol.
Dark adaptation: The process by which the eye adjusts under reduced
illumination and the sensitivity of the eye to light is greatly in-
creased.
Dark respiration: Metabolic activity of plants at night; consuming oxygen
to use stored sugars and releasing carbon dioxide.
Deciduous plants: Plants which drop their leaves at the end of the grow-
ing season.
Degradation (textiles): The decomposition of fabric or its components
or characteristics (color, strength, elasticity) by means of light,
heat, or air pollution.
Denitrification: A bacterial process occurring in soils, or water, in
which nitrate is used as the terminal electron acceptor and is re-
duced primarily to N?. It is essentially an anaerobic process; it
can occur in the presence of low levels of oxygen only if the micro-
organisms are metabolizing in an anoxic microzone.
De novo: Over again.
Deoxyribonucleic acid (DMA): A nucleic acid considered to be the carrier
of genetic information coded in the sequence of purine and pyrimidine
bases (organic bases). It has the form of a double-stranded helix
of a linear polymer.
Depauperate: Falling short of natural development or size.
Derivative spectrophotometer: An instrument with an increased capability
for detecting overlapping spectral lines and bands and also for
suppressing instrumentally scattered light.
G-15
-------
Desorb: To release a substance which has been taken into another substance
or held on its surface; the opposite of absorption or adsorption.
Desquamation: The shedding of the outer layer of any surface.
Detection limit: A level below which an element or chemical compound
cannot be reliably detected by the method or measurement being used for
analysis.
Detritus: Loose material that results directly from disintegration.
DeVarda alloy: An alloy of 50 percent Cu, 45 percent Al, 5 percent Zn.
Diastolic blood pressure: The blood pressure as measured during the period
of filling the cavities of the heart with blood.
Diazonium salt: A+chemical compound (usually colored) of the general
structure ArN»Cl , where Ar refers to an aromatic group.
Diazotizer: A chemical which, when reacted with amines (RNH2, for example),
produces a diazonium salt (usually a colored compound).
Dichotomous sampler: An air-sampling device which separates particulates
into two fractions by particle size.
Differentiation: The process by which a cell, such as a fertilized egg,
divides into specialized cells, such as the embryonic types that
eventually develop into an entire organism.
Diffusion: The process by which molecules or other particles intermingle
as a result of their random thermal motion.
Diffusing capacity: Rate at which gases move to or from the blood.
Dimer: A compound formed by the union of two like radicals or
molecules.
Dimerize: Formation of dimers.
1,6-diphosphofructose aldolase: An enzyme (EC 4.1.1.13) cleaving fructose
1,6-bisphosphate to dihydroxyacetone phosphate and glyceraldehyde-
3-phosphate.
D-2,3-diphosphoglycerate: A salt or ester of 2,3-diphosphoglyceric acid,
a major component of certain mammalian erythrocytes involved in the
release of 0^ from HbCL. Also a postulated intermediate in the bio-
chemical patnway involving the conversion of 3- to 2-phosphoglyceric
acid.
Diplococcus pneumoniae: A species of spherical-shaped bacteria belonging
to the genus Streptococcus. May be a causal agent in pneumonia.
G-16
-------
Direct dye: A dye with an affinity for most fibers; used mainly when
color resistance to washing is not important.
Disperse dyes: Also known as acetate dyes; these dyes were developed
for use on acetate fabrics, and are now also used on synthetic
fibers.
Distal: Far from some reference point such as median line of the body, point
of attachment or origin.
Diurnal: Having a repeating pattern or cycle 24 hours long.
DLCO: The diffusing capacity of the lungs for carbon monoxide. The ability
of the lungs to transfer carbon monoxide from the alveolar air into the
pulmonary capillary blood.
Dorsal hyphosis: Abnormal curvative of the spine; hunch-back.
Dose: The quantity of a substance to be taken all at one time or in
fractional amounts within a given period; also the total amount of a
pollutant delivered or concentration per unit time times time.
Dose-response curve: A curve on a graph based on responses occurring
in a system as a result of a series of stimuli intensities or doses.
Dry deposition: The processes by which matter is transferred to ground
from the atmosphere, other than precipitation; includes surface ab-
sorption of gases and sedimentation, Brownian diffusion and impaction
of particles.
Dyeing: A process of coloring fibers, yarns, or fabrics with either
natural or synthetic dyes.
Dynamic calibration: Testing of a monitoring system using a continuous
sample stream of known concentration.
Dynamic compliance (C.. ): Volume change per unit of transpulmonary
pressure minus tneypressure of pulmonary resistance during airflow.
Dynel: A trademark for a modacrylic staple fiber spun from a copolymer
of acrylonitrile and vinyl chloride. It has high strength, quick-
drying properties, and resistance to alkalies and acids.
Dyspepsia: Indigestion, upset stomach.
Dyspnea: Shortness of breath; difficulty or distress in breathing; rapid
breathing.
Ecosystem: The interacting system of a biological community and its
environment.
Eddy: A current of water or air running contrary to the main current.
G-17
-------
Edema: Pressure of excess fluid in cells, intercellular tissue or cavities
of the body.
Elastomer: A synthetic rubber product which has the physical properties
of natural rubber.
Electrocardiogram: The graphic record of the electrical currents that
initiate the heart's contraction.
Electrode: One of the two extremities of an electric circuit.
Electrolyte: A non-metallic electric conductor in which current, is carried
by the movement of ions; also a substance which displays these qualities
when dissolved in water or another solvent.
Electronegativity: Measure of affinity for negative charges or electrons.
Electron microscopy: A technique which utilizes a focused beam of electrons
to produce a high-resolution image of minute objects such as particu-
late matter, bacteria, viruses, and DNA.
Electronic excitation energy: Energy associated in the transition of
electrons from their normal low-energy orbitals or orbitals of higher
energy.
Electrophilic: Having an affinity fbr electrons.
Electrophoresis: A technique by which compounds can be separated from a
complex mixture by their attraction to the positive or negative
pole of an applied electric potential.
Eluant: A liquid used in the process of elution.
Elute: To perform an elution.
Elution: Separation of one material from another by washing or by dissolving
one in a solvent in which the other is not soluble.
Elutriate: To separate a coarse, insoluble powder from a finer one by
suspending them in water and pouring off the finer powder from the
upper part of the fluid.
Emission spectrometry: A rapid analytical technique based on measurement
of the characteristic radiation emitted by thermally or electrically
excited atoms or ions.
Emphysema: An anatomic alteration of the lung, characterized by abnormal
enlargement of air spacers distal to the terminal bronchioles, due
to dilation or destructive changes in the alveolar walls.
Emphysematous lesions: A wound or injury to the lung as a result of
emphysema.
G-18
-------
Empirical modeling: Characterization and description of a phenomena
based on experience or observation.
Encephalitis: Inflammation of the brain.
Endoplasmic reticulum: An elaborate membrane structure extending from the
nuclear membrane or eucaryotie cells to the cytoplasmic membrane.
Endothelium: A layer of flat cells lining especially blood and lymphatic
vessels.
Entropy: A measure of disorder or randomness in a system. Low entropy
is associated with highly ordered systems.
Enzyme: Any of numerous proteins produced by living cells which catalyze
biological reactions.
Enzyme Commission (EC): The International Commission on Enzymes, established
in 1956, developed a scheme of classification and nomenclature under
which each enzyme is assigned an EC number which identifies it by
function.
Eosinophils: Leukocytes (white blood cells) which stain readily with the
dye, eosin.
Epidemiology: A study of the distribution and determinants of disease
in human population groups.
Epidermis: The outermost living layer of cells of any organism.
Epididymal fat pads: The fatty tissue located near the epididymis. The
epididymis is the first convoluted portion of the excretory duct
of the testis.
Epiphyte: A plant growing on another plant but obtaining food from the
atmosphere.
Epithelial: Relating to epithelium, the membranous cellular layer which
covers free surfaces or lines tubes or cavities of an animal body,
which encloses, protects, secretes, excretes and/or assimilates.
Erosion corrosion: Acceleration or increase in rate of deterioration
or attack on a metal because of relative movement between a corrosive
fluid and the metal surface. Characterized by grooves, gullies, or
waves in the metal surface.
Erythrocyte: A mature red blood cell.
Escherichia coli: A short, gram-negative, rod-shaped bacteria common
to the human intestinal tract. A frequent cause of infections in
the urogenital tract.
G-19
-------
Esophageal: Relating to the portion of the digestive tract between the
pharynx and the stomach.
Estrus: That portion or phase of the sexual cycle of female animals
characterized by willingness to permit coitus.
Estrus cycle: The series of physiologic uterine, ovarian and other
changes that occur in higher animals.
Etiolation: Paleness and/or altered development resulting from the
absence of light.
Etiology: The causes of a disease or condition; also, the study of
causes.
Eucaryotic: Pertaining to those cells having a well-defined nucleus
surrounded by a double-layered membrane.
Euthrophication: Elevation of the level of nutrients in a body of water,
which can contribute to accelerated plant growth and filling.
Excited state: A state of higher electronic energy than the ground state,
usually a less stable one.
Expiratory (maximum) flow rate: The maximum rate at which air can be
expelled from the lungs.
Exposure level: Concentration of a contaminant to which an individual
or a population is exposed.
Extinction coefficient: A measure of the space rate of diminution, or
extinction, of any transmitted light, thus, it is the attenuation
coefficient applied to visible radiation.
Extramedullary hematopoiesis: The process of formation and development
of the various types of blood cells and other formed elements not
including that occurring in bone marrow.
Extravasate: To exclude from or pass out of a vessel into the tissues;
applies to urine, lymph, blood and similar fluids.
Far ultraviolet: Radiation in the range of wavelengths from 100 to 190
nanometers.
Federal Reference Method (FRM): For NO,,, the EPA-approved analyzers based
on the gas-phase chemiluminescent measurement principle and associated
calibration procedures; regulatory specifications prescribed in Title
40, Code of Federal Regulations, Part 50, Appendix F.
Fenestrae: Anatomical aperatures often closed by a membrane.
Fiber: A fine, threadlike piece, as of cotton, jute, or asbestos.
G-20
-------
Fiber-reactive dye: A water-soluble dyestuff which reacts chemically
with the cellulose in fibers under alkaline conditions; the dye
contains two chlorine atoms which combine with the hydroxyl groups of
the cellulose.
Fibrin: A white insoluble elastic filamentous protein derived from fibrino-
gen by the action of thrombin, especially in the clotting of blood.
Fibroadenoma: A benign neoplasm derived from glandular epithelium, in-
volving proliferating fibroblasts, cells found in connective tissue.
Fibroblast: An elongated cell with cytoplasmic processes present in
connective tissue, capable of forming collagen fibers.
Fibrosis: The formation of fibrous tissue, usually as a reparative or
reactive process and not as a normal constituent of an organ or
tissue.
Flocculation: Separation of material from a solution or suspension by
reaction with a flocculant to create fluffy masses containing the
material to be removed.
Fly ash: Fine, solid particles of noncombustible ash carried out of a
bed of solid fuel by a draft.
Folded-path optical system: A long (e.g. 8-22 m) chamber with multiple
mirrors at the ends which can be used to reflect an infrared beam through
an ambient air sample many times; a spectrometer can be used with such
a system to detect trace pollutants at very low levels.
Forced expiratory flow (FEF): The rate at which air can be expelled from
the lungs; see expiratory flow rate.
Forced expiratory volume (FEV): The maximum volume of air that can be
expired in a specific time interval when starting from maximal
inspiration.
Forced vital capacity (FVC): The greatest volume of air that can be
exhaled from the lungs under forced conditions after a maximum
inspiration.
Fractional threshold concentration: The portion of the concentration
at which an event or a response begins to occur, expressed as a
fraction.
Free radical: Any of a variety of highly-reactive atoms or molecules
characterized by having an unpaired electron.
Fritted bubbler: A. porous glass device used in air pollutant sampling
systems to introduce small bubbles into solution.
G-21
-------
Functional residual capacity: The volume of gas remaining in the lungs
at the end of a normal expiration. It is the sum of expiratory
reserve volume and residual volume.
Gas exchange: Movement of oxygen from the alveoli into the pulmonary
capillary blood as carbon dioxide enters the alveoli from the blood.
Gas chromatography (GC): A method of separating and analyzing mixtures
of chemical substances. A flow of gas causes the components of a
mixture to migrate differentially from a narrow starting zone in a
special porous, insoluble sorptive medium. The pattern formed by
zones of separated pigments and of colorless substances in this
process is called a chromatogram, and can be analyzed to obtain the
concentration of identified pollutants.
Gas-liquid chromatography: A method of separating and analyzing volatile
organic compounds, in which a sample is vaporized and swept through
a column filled with solid support material covered with a nonvolatile
liquid. Components of the sample can be identified and their con-
centrations determined by analysis of the characteristics of their
retention in the column, since compounds have varying degrees of
solubility in the liquid medium.
Gastric juice: A thin watery digestive fluid secreted by glands in the
mucous membrane of the stomach.
Gastroenteritis: Inflammation of the mucous membrane of stomach and
intestine.
Genotype: The type of genes possessed by an organism.
Geometric mean: An estimate of the average of a distribution. Specifically,
the nth root of the product of n observations.
Geometric standard deviation: A measure of variability of a distribution.
It is the antilogarithm of the standard deviation of the logarithms
of the observations.
Globulins (a, b, q): A family of proteins precipitated from plasma (or
serum) by half-saturation with ammonium sulfate, or separable by
electrophoresis. The main groups are the a, b, q fractions, differ-
ing with respect to associated lipids and carbohydrates and in their
content of antibodies (immunoglobulins).
Glomular nephrotic syndrome: Dysfunction of the kidneys characterized
by excessive protein loss in the urine, accumulation of body fluids
and alteration in albumin/globulin ratio.
Glucose: A sugar which is a principal source of energy for man and other
organisms.
Glucose-6-phosphate dehydrogenase: An enzyme (EC 1.1.1.49) catalyzing
the dehydrogenation of glucose-6-phosphate to 6-phosphogluconolactone.
G-22
-------
Glutamic-oxaloacetic transaminase (SCOT): An enzyme (EC 2.6.1.1) whose
serum level increases in myocardial infarction and in diseases in-
volving destruction of liver cells. Also known as aspartate
aminotransferase.
Glutamic-pyruvic transaminase (SGPT): Now known as alanine aminotransferase
(EC 2.6.1.2), the serum levels of this enzyme are used in liver function
tests.
Glutathione (GSH): A tripeptide composed of glycine, cystine, and glutamic
acid.
Glutathione peroxidase: An enzyme (EC 1.11.1) which catalyzes the destruction
of hydroperoxides formed from fatty acids and other substances.
Protects tissues from oxidative damage. It is a selenium-containing
protein.
Glutathione reductase: The enzyme (EC 1.6.4.2) which reduces the oxidized
form of glutathione.
Glycolytic pathway: The biochemical pathway by which glucose is con-
verted to lactic acid in various tissues, yielding energy as a
result.
Glycoside: A type of chemical compound formed from the condensation of
a sugar with another chemical radical via a hemiacetal linkage.
Goblet cells: Epithelial cells that have been distended with mucin and when
this is discharged as mucus, a goblet-shaped shell remains.
Golgi apparatus: A membrane system involved with secretory functions
and transport in a cell. Also known as a dictyosome.
Grana: The lamellar stacks of chlorophyll-containing material in plant
chloroplasts.
Griege carpet: A carpet in its unfinished state, i.e. before it has
been scoured and dyed. The term also is used for woven fabrics
in the unbleached and unfinished state.
Ground state: The state of minimum electronic energy of a molecule or
atom.
Guanylate cyclase (GC): An enzyme (EC 4.6.2.1) catalyzing the trans-
formation of guanosine triphosphate to guanosine 3':5'-cyclic phosphate.
H-Thymidine: Thymine deoxyribonucleoside: One of the four major nucleosides
in DMA. H-thymidine has been uniformly labeled with tritium, a radio-
active form of hydrogen.
Haze: Fine dust, smoke or fine vapor reducing transparency of air.
G-23
-------
Hemagglutination: The agglutination of red blood cells. Can be used as
as a measurement of antibody concentration.
Hematocrit: The percentage of the volume of a blood sample occupied by
cells.
Hematology: The medical specialty that pertains to the blood and blood-
forming tissues.
Hemochromatosis: A disease characterized by pigmentation of the skin
possibly due to inherited excessive absorption of iron.
Hemoglobin (Hb): The red, respiratory protein of the red blood cells,
hemoglobin transports oxygen from the lungs to the tissues as oxy-
hemoglobin (HbO™) and returns carbon dioxide to the lungs as hemoglobin
carbamate, completing the respiratory cycle.
Hemolysis: Alteration or destruction of red blood cells, causing hemoglobin
to be released into the medium in which the cells are suspended.
Hepatectomy: Complete removal of the liver in an experimental animal.
Hepatic: Relating to the liver.
Hepatocyte: A liver cell.
Heterogeneous process: A chemical reaction involving reactants of more
than one phase or state, such as one in which gases are absorbed into
aerosol droplets, where the reaction takes place.
Heterologous: A term referring .to donor and recipient cellular elements
from different organisms, such as red blood cells from sheep and
alveolar macrophage from rabbits.
Hexose monophosphate shunt: Also called the phosphogluconate oxidative
pathway of glucose metabolism which affords a total combustion of
glucose independent of the citric acid cycle. It is the important
generator of NADPH necessary for synthesis of fatty acids and the
operation of various enzymes. It serves as a source of ribose and
4- and 7-carbon sugars.
High-volume sampler (Hi-vol): Device for taking a sample of the particulate
content of a large amount of,air, by drawing air through a fiber filter
at a typical rate of 2,000 m /24 hr (1.38 m /min), or as high as 2,880
mV24 hr (2 mVmin).
Histamine: An amine derived from the amino acid, histidine. It is a
powerful stimulant of gastric secretion and a constrictor of bronchial
smooth muscle. It is a vasodilator and causes a fall in blood
pressure.
G-24
-------
Homogenate: Commonly refers to tissue ground into a creamy consistency
in which the cell structure is disintegrated.
Host defense mechanism: Inherent means by which a biologic organism
protects itself against infection, such as antibody formation,
macrophage action, ciliary action, etc.
Host resistance: The resistance exhibited by an organism, such as man,
to an infecting agent, such as a virus or bacteria.
Humoral: Relating to the extracellular fluids of the body, blood and
lymph.
Hybrid: An organism descended from parents belonging to different
varieties or species.
Hydrocarbons: A vast family of compounds containing carbon and hydrogen
in various combinations; found especially in fossil fuels. Some
contribute to photochemical smog.
Hydrolysis: Decomposition involving splitting of a bond and addition
of the H and OH parts of water to the two sides of the split bond.
Hydrometeor: A product of the condensation of atmospheric water vapor (e.g.
fog, rain, hail, snow).
Hydroxyproline: An ami no acid found among the hydrolysis products of
collagen.
Hygroscopic: Pertaining to a marked ability to accelerate the condensation
of water vapor.
Hyperplasia: Increase in the number of cells in a tissue or organ ex-
cluding tumor formation.
Hyperplastic: Relating to hyperplasia; an increase in the number of
cells.
Hypertrophy: Increase in the size of a tissue element, excluding tumor
formation.
Hypertension: Abnormally elevated blood pressure.
Hypolimnia: Portions of a lake below the thermocline, in which water
is stagnant and uniform in temperature.
Hypoxia: A lower than normal amount of oxygen in the air, blood or tissues
G-25
-------
Immunoglobulin (Ig): A class of structurally related proteins consist-
ing of two pairs of polypeptide chains. Antibodies are Ig's and
'all Ig's probably function as antibodies.
Immunoglobulin A (IgA): A type of antibody which comprises approximately
10 to 15 percent of the total amount of antibodies present in normal
serum.
Immunoglobulin G (IgG): A type of antibody which comprises approximately
80 percent of the total amount of antibodies present in normal serum.
Subfractions of IgG are fractions G,, and G2-
Immunoglobulin M (IgM): A type of antibody which comprises approximately
5 to 10 percent of the total amount of antibodies present in normal
serum.
Impaction: An impinging or striking of one object against another; also,
the force transmitted by this act.
Impactor: An instrument which collects samples of suspended particulates
by directing a stream of the suspension against a surface, or into a
liquid or a void.
Index of proliferation: Ratio of promonocytes to polymorphic monocytes
in the blood.
Infarction: Sudden insufficiency of arterial or venous blood supply
due to emboli, thrombi, or pressure.
Infectivity model: A testing system in which the susceptibility of
animals to airborne infectious agents with and without exposure to air
pollutants is investigated to produce information related to the
possible effects of the pollutant on man.
Inflorescence: The arrangement and development of flowers on an axis;
also, a flower cluster or a single flower.
Influenza A?/Taiwan Virus: An infectious viral disease, believed to
have originated in Taiwan, characterized by sudden onset, chills,
fevers, headache, and cough.
Infrared: Light invisible to the human eye, bgtween the wavelengths
of 7x10 and 10 m (7000 and 10,000,000 A).
Infrared laser: A device that utilizes the natural oscillations of atoms
or molecules to generate coherent electromagnetic radiation in the
infrared region of the spectrum.
Infrared spectrometer: An instrument for measuring the relative amounts
of radiant energy in the infrared region of the spectrum as a function
of wavelength.
G-26
-------
Ingestion: To take in for digestion.
In situ: In the natural or original position.
Instrumental averaging time: The time over which a single sample or
measurement is taken, resulting in a measurement which is an average
of the actual concentrations over that period.
Insult: An injury or trauma.
Intercostal: Between the ribs, especially of a leaf.
Interferant: A substance which a measurement method cannot distinguish
completely from the one being measured, which therefore can cause some
degree of false response or error.
Interferon: A macromolecular substance produced in response to infection
with active or inactivated virus, capable of inducing a state of
resistance.
Intergranular corrosion: A type of corrosion which takes place at and
adjacent to grain boundaries, with relatively little corrosion of
the grains.
Interstitial edema: An accumulation of an excessive amount of fluids
in a space within tissues.
Interstitial pneumonia: A chronic inflammation of the interstitial tissue
of the lung, resulting in compression of air cells.
Intraluminal mucus: Mucus that collects within any tubule.
Intraperitoneal injection: An injection of material into the serous
sac that lines the abdominal cavity.
In utero: Within the womb; not yet born.
In vitro: Refers to experiments conducted outside the living organism.
In vivo: Refers to experiments conducted within the living organism.
Irradiation: Exposure to any form of radiation.
Ischemia: Local anemia due to mechanical obstruction (mainly arterial
narrowing) of the blood supply.
Isoenzymes: Also called isozymes. One of a group of enzymes that are
very similar in catalytic properties, but may be differentiated by
variations in physical properties, such as isoelectric point or
electrophoretic mobility. Lactic acid dehydrogenase is an example
of an enzyme having many isomeric forms.
G-27
-------
Isopleth: A line on a map or chart connecting points of equal value.
Jacobs-Hochheiser method: The original Federal Reference Method for NO-,
currently unacceptable for air pollution work.
Klebsiella pneumoniae: A species of rod-shaped bacteria found in soil,
water, and in the intestinal tract of man and other animals. Certain
types may be causative agents in pneumonia.
Kyphosis: An abnormal curvature of the spine, with convexity backward.
Lactate: A salt or ester of lactic acid.
Lactic acid (lactate) dehydrogenase (LDH): An enzyme (EC 1.1.1.27) with
many isomeric forms which catalyzes the oxidation of lactate to
pyruvate via transfer of H to NAD. Isomeric forms of LDH in the
blood are indicators of heart damage.
Lamellar bodies: Arranged in plates or scales. One of the characteristics
of Type II alveolar cells.
Lavage fluid: Any fluid used to wash out hollow organs, such as the lung.
Lecithin: Any of several waxy hygroscopic phosphatides that are widely
distributed in animals and plants; they form colloidal solutions in
water and have emulsifying, wetting and hygroscopic properties.
Legume: A plant with root nodules containing nitrogen fixing bacteria.
Lesion: A wound, injury or other more or less circumscribed pathologic
change in the tissues.
Leukocyte: Any of the white blood cells.
Lewis base: A base, defined in the Lewis acid-base concept, is a sub-
stance that can donate an electron pair.
Lichens: Perennial plants which are a combination of two plants, an alga
and a fungus, growing together in an association so intimate that they
appear as one.
Ligand: Those molecules or anions attached to the central atom in a
complex.
Light-fastness: The ability of a dye to maintain its original color under
natural or indoor light.
Linolenic acid: An unsaturated fatty acid essential in nutrition.
Lipase: An enzyme that accelerates the hydrolysis or synthesis of fats
or the breakdown of lipoproteins.
G-28
-------
Lipids: A heterogeneous group of substances which occur widely in bio-
logical materials. They are characterized as a group by their
extractabi1ity in nonpolar organic solvents.
Lipofuscin: Brown pigment granules representing 1ipid-containing residues
of lysosomal digestion. Proposed to be an end product of lipid
oxidation which accumulates in tissue.
Lipoprotein: Complex or protein containing lipid and protein.
Loading rate: The amount of a nutrient available to a unit area of body
of water over a given period of time.
Locomotor activity. Movement of an organism from one place to another
of its own volition.
Long-path!ength infrared absorption: A measurement technique in which a
system of mirrors in a chamber is used to direct an infrared beam
through a sample of air for a long distance (up to 2 km); the amount
of infrared absorbed is measured to obtain the concentrations of
pollutants present.
Lung compliance (C,): The volume change produced by an increase in a
unit change in pressure across the lung, i.e., between the pleural
surface and the mouth.
Lycra: A spandex textile fiber created by E. I. du Pont de Nemours & Co.,
Inc., with excellent tensile strength, a long flex life and high
resistance to abrasion and heat degradation. Used in brassieres,
foundation garments, surgical hosiery, swim suits and military and
industrial uses.
Lymphocytes: White blood cells formed in lymphoid tissue throughout the
body, they comprise about 22 to 28 percent of the total number of
leukocytes in the circulating blood and function in immunity.
Lymphocytogram: The ratio, in the blood, of lymphocyte with narrow
cytoplasm to those with broad cytoplasm.
Lysosomes: Organelles found in cells of higher organisms that contain
high concentrations of degradative enzymes and are known to destroy
foreign substances that cells engulf by pinocytosis and phyocytosis.
Believed to be a major site where proteins are broken down.
Lysozymes: Lytic enzymes destructive to cell walls of certain bacteria.
Present in some body fluids, including tears and serum.
Macaca speciosa: A species of monkeys used in research.
Macrophage: Any large, ameboid, phagocytic cell having a nucleus without
many lobes, regardless of origin.
G-29
-------
Malaise: A feeling of general discomfort or uneasiness, often the first
indication of an infection or disease.
Malate dehydrogenase: An enzyme (EC 1.1.1.37) with at least six isomeric
forms that catalyze the dehydrogenation of ma late to oxaloacetate
or its decarboxylation (removal of a CCL group) to pyruvate. Malate,
oxaloacetate, and pyruvate are intermediate components of biochemical
pathways.
Mannitol: An alcohol derived from reduction of the sugar, fructose.
Used in renal function testing to measure glomerular (capillary)
fiItration.
Manometer: An instrument for the measurement of pressure of gases or
vapors.
Mass median diameter (MMD): Geometric median size of a distribution of
particles based on weight.
Mass spectrometry (MS): A procedure for identifying the various kinds of
particles present in a given substance, by ionizing the particles
and subjecting a beam of the ionized particles to an electric or
magnetic field such that the field deflects the particles in angles
directly proportional to the masses of the particles.
Maximum flow (V ): Maximum rate or expiration, usually expressed at
50 or 25 percent of vital capacity.
Maximum mid-expiratory flow rate (MMFR): The mean rate of expiratory gas
flow between 25 and 75 percent of the forced expiratory vital capacity.
Mean (arithmetic): The sum of observations divided by sample size.
Median: A value in a collection of data values which is exceeded in
magnitude by one-half the entries in the collection.
Mesoscale: Of or relating to meteorological phenomena from 1 to 100
kilometers in horizontal extent.
Messenger RNA: A type of RNA which conveys genetic information encoded
in the DNA to direct protein synthesis.
Metaplasia: The abnormal transformation of an adult, fully differentiated
tissue of one kind into a differentiated tissue of another kind.
Metaproterenol: A bronchodilator used for the treatment of bronchial
asthma.
Metastases: The shifting of a disease from one part of the body to another;
the appearance of neoplasms in parts of the body remote from the seat
of the primary tumor.
G-30
-------
Meteorology: The science that deals with the atmosphere and its phenomena.
Methemoglobin: A^form of hemoglobin in which the normal reduced state
of iron (Fe ) has been ox^oMzed to Fe . It contains oxygen in
firm union with ferric (Fe ) iron and is not capable of exchanging
oxygen in normal respiratory processes.
Methimazole: An anti-thyroid drug similar in action to propylthiouracil.
Methyltransferase: Any enzyme transferring methyl groups from one compound
to another.
Microcoulometric: Capable of measuring millionths of coulombs used in
electrolysis of a substance, to determine the amount of a substance
in a sample.
Microflora: A small or strictly localized plant.
Micron: One-millionth of a meter.
Microphage: A small phagocyte; a polymorphonuclear leukocyte that is
phagocytic.
Millimolar: One-thousandth of a molar solution. A solution of one-
thousandth of a mole (in grams) per liter.
Minute volume: The minute volume of breathing; a product of tidal volume
times the respiratory frequency in one minute.
Mitochondria: Organdies of the cell cytoplasm which contain enzymes
active in the conservation of energy obtained in the aerobic part
of the breakdown of carbohydrates and fats, in a process called
respiration.
Mobile sources: Automobiles, trucks and other pollution sources which are
not fixed in one location.
Modacrylic fiber: A manufactured fiber in which the fiber-forming sub-
stance is any long chain synthetic polymer composed of less than 85
percent but at least 35 percent by weight of acrylonitrite units.
Moeity: One of two or more parts into which something is divided.
Mole: The mass, in grams, numerically equal to the molecular weight of
a substance.
Molecular correlation spectrometry: A spectrophotometric technique which
is used to identify unknown absorbing materials and measure their
concentrations by using preset wavelengths.
Molecular weight: The weight of one molecule of a substance obtained
by adding the gram-atomic weights of each of the individual atoms
in the substance.
G-31
-------
Monocyte: A relatively large mononuclear leukocyte, normally constituting
3 to 7 percent of the leukocytes of the circulating blood.
Mordant: A substance which acts to bind dyes to a textile fiber of fabric.
Morphological: Relating to the form and structure of an organism or any
of its parts.
Moving average: A procedure involving taking averages over a specific
period prior to and including a year in question, so that successive
averaging periods overlap; e.g. a three-year moving average would
include data from 1967 through 1969 for the 1969 average and from
1968 through 1970 for 1970.
Mucociliary clearance: Removal of materials from the upper respiratory
tract via ciliary action.
Mucociliary transport: The process by which mucus is transported, by
ciliary action, from the lungs.
Mucosa: The mucous membrane; it consists of epithelium, lamina propria
and, in the digestive tract, a layer of smooth muscle.
Mucous membrane: A membrane secreting mucus which lines passages and
cavities communicating with the exterior of the body.
Murine: Relating to mice.
Mutagen: A substance capable of causing, within an organism, biological
changes that affect potential offspring through genetic mutation.
Mutagenic: Having the power to cause mutations. A mutation is a change
in the character of a gene (a sequence of base pairs in DNA) that
is perpetuated in subsequent divisions of the cell in which it occurs.
Myocardial infarction: Infarction of any area of the heart muscle usually
as a result of occlusion of a coronary artery.
Nares: The nostrils.
Nasopharyngeal: Relating to the nasal cavity and the pharynx (throat).
National Air Surveillance Network (NASN): Network of monitoring stations
for sampling air to determine extent of air pollution; established
jointly by federal and state governments.
Near ultraviolet: Radiation of the wavelengths 2000-4000 Angstroms.
Necrosis: Death of cells that can discolor areas of a plant or kill
the entire plant.
Necrotic: Pertaining to the pathologic death of one or more cells, or
of a portion of tissue or organ, resulting from irreversible damage.
6-32
-------
Neonate: A newborn.
Neoplasm: An abnormal tissue that grows more rapidly than normal; synonymous
with tumor.
Neoplasia: The pathologic process that results in the formation and
growth of a tumor.
Neutrophil: A mature white blood cell formed in bone marrow and released
into the circulating blood, where it normally accounts for 54 to 65
percent of the total number of leukocytes.
Ninhydrin: An organic reagent used to identify amino acids.
Nitramine: A compound consisting of a nitrogen attached to the nitrogen
of amine.
Nitrate: A salt or ester of nitric acid (NO ~).
Nitrification: The principal natural source of nitrate in which ammonium
(NH.+) ions are oxidized to nitrites by specialized microorganisms.
Other organisms oxidize nitrites to nitrates.
Nitrite: A salt or ester of nitrous acid (N02~).
Nitrocellulose: Any of several esters of nitric acid formed by its action
on cellulose, used in explosives, plastics, varnishes and rayon;
also called cellulose nitrate.
Nitrogen cycle: Refers to the complex pathways by which nitrogen-containing
compounds are moved from the atmosphere into organic life, into the
soil, and back to the atmosphere.
Nitrogen fixation: The metabolic assimilation of atmospheric nitrogen by
soil microorganisms, which becomes available for plant use when the
microorganisms die; also, industrial conversion of free nitrogen into
combined forms used in production of fertilizers and other products.
Nitrogen oxide: A compound composed of only nitrogen and oxygen. Components
of photochemical smog.
Nitrosamine: A compound consisting of a nitrosyl group connected to the
nitrogen of an amine.
Nitrosation: Addition of a nitrosyl group.
N-Nitroso compounds: Compounds carrying the functional nitrosyl group.
Nitrosyl: A group composed of one oxygen and one nitrogen atom (-N=0).
Nitrosylhemoglobin (NOHb): The red, respiratory protein of erythrocytes
to which a nitrosyl group is attached.
G-33
-------
N/P Ratio: Ratio of nitrogen to phosphorous dissolved in lake water,
important due to its effect on plant growth.
Nucleolus: A small spherical mass of material within the substance of the
nucleus of a eel 1.
Nucleophi1ic: Having an affinity for atomic nuclei; electron-donating.
Nucleoside: A compound that consists of a purine or pyrimidine base com-
bined with deoxyribose or ribose and found in RNA and DNA.
S'-Nucleotidase: An enzyme (EC 3.1.3.5) which hydrolyzes nucleoside 5'-
phosphates into phosphoric acid (hLPCL) and nucleosides.
Nucleotide: A compound consisting of a sugar (ribose or deoxyribose),
a base (a purine or a pyrimidine), and a phosphate; a basic structural
unit of RNA and DNA.
Nylon: A generic name chosen by E. I. du Pont de Nemours & Co. , Inc.
for a group of protein-like chemical products classed as synthetic
linear polymers; two main types are Nylon 6 and Nylon 66.
Occlusion: A point .which an opening is closed or obstructed.
Olefin: An open-chain hydrocarbon having at least one double bond.
Olfactory: Relating to the sense of smell.
Olfactory epithelium: The inner lining of the nose and mouth which contains
neural tissue sensitive to smell.
01igotrophic: A body of water deficient in plant nutrients; also generally
having abundant dissolved oxygen and no marked stratification.
Oribitals: Areas of high electron density in an atom or molecule.
Orion: An acrylic fiber produced by E. I. du Pont de Nemours and Co., Inc.,
based on a polymer of acrylonitrite; used extensively for outdoor
uses, it is resistant to chemicals and withstands high temperatures.
Osteogenic osteosarcoma: The most common and malignant of bone sarcomas
(tumors). It arises from bone-forming cells and affects chiefly
the ends of long bones.
Ovarian primordial follicle: A spheroidal cell aggregation in the ovary
in which the primordial oocyte (immature female sex cell) is surrounded
by a single layer of flattened follicular cells.
Oxidant: A chemical compound which has the ability to remove electrons
from another chemical species, thereby oxidizing it; also, a substance
containing oxygen which reacts in air to produce a new substance, or
one formed by the action of sunlight on oxides of nitrogen and hydro-
carbons.
G-34
-------
Oxidation: An ion or molecule undergoes oxidation by donating electrons.
Oxidative deamination: Removal of the NhL group from an ami no compound
by reaction with oxygen.
Oxidative phosphorylation: The mitochondria! process by which "high-
energy" phosphate bonds form from the energy released as a result of
the oxidation of various substrates. Principally occurs in the tri-
carboxylic acid pathway.
Oxyhemoglobin: Hemoglobin in combination with oxygen. It is the form
of hemoglobin present in arterial blood.
Ozone layer: A layer of the stratosphere from 20 to 50 km above the
earth's surface characterized by high ozone content produced by ultra-
violet radiation.
Ozone scavenging: Removal of 03 from ambient air or plumes by reaction with
NO, producing NOp and 0^.
Paired electrons: Electrons having opposite intrinsic spins about their
own axes.
Parenchyma: The essential and distinctive tissue of an organ or an ab-
normal growth, as distinguished from its supportive framework.
Parenchymal: Referring to the distinguishing or specific cells of a
gland or organ.
Partial pressure: The pressure exerted by a single component in a mixture
of gases.
Particulates: Fine liquid or solid particles such as dust, smoke, mist,
fumes or smog, found in the air or in emissions.
Pascal: A unit of pressure in the International System of Units. One
pascal is equal to 7.4 x 10 torr. The pascal is equivalent to one
newton per square meter.
Pathogen: Any virus, microorganism, or other substance causing disease.
Pathophysiological: Derangement of function seen in disease; alteration
in function as distinguished from structural defects.
Peptide bond: The bond formed when two ami no acids react with each other.
Percentiles: The percentage of all observations exceeding or preceding
some point; thus, 90th percentile is a level below which will fall 90
percent of the observations.
Perfusate: A liquid, solution or colloidal suspension that has been passed
over a special surface or through an appropriate structure.
G-35
-------
Perfusion: Artificial passage of fluid through blood vessels.
Permanent-press fabrics: Fabrics in which applied resins contribute to the
easy care and appearance of the fabric and to the crease and seam
flat-
ness by reacting with the cellulose on pressing after garment
manufacture.
Permeation tube: A tube which is selectively porous to specific gases.
Peroxidation: Refers to the process by which certain organic compounds
are converted to peroxides.
Peroxyacetyl nitrate (PAN): Pollutant created by action of sunlight on
hydrocarbons and NO in the air; an ingredient of photochemical smog.
pH: A measure of the acidity or alkalinity of a material, liquid, or solid.
pH is represented on a scale of 0 to 14 with 7 being a neutral state,
0 most acid, and 14 most alkaline.
Phagocytosis: Ingestion, by cells such as macrophages, of other cells,
bacteria, foreign particles, etc.; the cell membrane engulfs solid or
liquid particles which are drawn into the cytoplasm and digested.
Phenotype: The observable characteristics of an organism, resulting from
the interaction between an individual genetic structure and the
environment in which development takes place.
Phenylthiourea: A crystalline compound, C,HgN2S, that is bitter or tasteless
depending on a single dominant gene in the tester.
Phlegm: Viscid mucus secreted in abnormal quantity in the respiratory passages.
Phosphatase: Any of a group of enzymes that liberate inorganic phosphate
from phosphoric esters (E.G. sub-subclass 3.1.3).
Phosphocreatine kinase: An enzyme (EC 2.7.3.2) catalyzing the formation of
creatine and ATP, its breakdown is a source of energy in the contraction
of muscle; also called creatine phosphate.
Phospholipid: A molecule consisting of IJpid and phosphoric acid group(s).
An example is lecithin. Serves as an important structural factor
in biological membranes.
Photochemical oxidants: Primary ozone, NO-, PAN with lesser amounts of
other compounds formed as products of atmospheric reactions involving
organic pollutants, nitrogen oxides, oxygen, and sunlight.
Photochemical smog: Air pollution caused by chemical reaction of various
airborne chemicals in sunlight.
Photodissociation: The process by which a chemical compound breaks down into
simpler components under the influence of sunlight or other radiant energy.
G-36
-------
Photolysis: Decomposition upon irradiation by sunlight.
Photomultiplier tube: An electron multiplier in which electrons released
by photoelectric emission are multiplied in successive stages by
dynodes that produce secondary emissions.
Photon: A quantum of electromagnetic energy.
Photostationary: A substance or reaction which reaches and maintains a
steady state in the presence of light.
Photosynthesis: The process in which green parts of plants, when exposed to
light under suitable conditions of temperature and water supply, produce
carbohydrates using atmospheric carbon dioxide and releasing oxygen.
Phytotoxic: Poisonous to plants.
Phytoplankton: Minute aquatic plant life.
Pi 7t bonds: Bonds in which electron density is not symmetrical about a
line joining the bonded atoms.
Pinocytotic: Refers to the cellular process (pinocytosis) in which the cyto-
plasmic membrane forms invaginations- in the form of narrow,channels
leading into the cell. Liquids can flow into these channels and the
membrane pinches off pockets that are incorporated into the cytoplasm
and digested.
Pitting: A form of extremely localized corrosion that results in holes in
the metal. One of the most destructive forms of corrosion.
Pituary: A stalk-like gland near the base of the brain which is attached
to the hypothalmus. The anterior portion is a major repository for
for hormones that control growth, stimulate other glands, and regulate
the reproductive cycle.
Placenta: The organ in the uterus that provides metabolic interchange between
the fetus and mother.
Plasmid: Replicating unit, other than a nucleus gene, that contains
nucleoprotein and is involved in various aspects of metabolism in
organisms; also called paragenes.
Plasmolysis: The dissolution of cellular components, or the shrinking
of plant cells by osmotic loss of cytoplasmic water.
Plastic: A plastic is one of a large group of organic compounds synthesized
from cellulose, hydrocarbons, proteins or resins and capable of being
cast, extruded, or molded into various shapes.
Plasticizer: A chemical added to plastics to soften, increase malleability
or to make more readily deformable.
G-37
-------
Platelet (blood): An irregularly-shaped disk with no definite nucleus;
about one-third to one-half the size of an erythrocyte and containing
no hemoglobin. Platelets are more numerous than leukocytes, numbering
from 200,000 to 300,000 per cu. mm. of blood.
Plethysmograph: A device for measuring and recording changes in volume of
a part, organ or the whole body; a body plethysmograph is a chamber
apparatus surrounding the entire body.
Pleura: The serous membrane enveloping the lungs and lining the walls of
the chest cavity.
Plume: Emission from a flue or chimney, usually distributed stream-like
downwind of the source, which can be distinguished from the surrounding
air by appearance or chemical characteristics.
Pneumonia (interstitial): A chronic inflammation of the interstitial tissue
of the lung, resulting in compression of the air cells. An acute, infec-
tious disease.
Pneumonocytes: A nonspecific term sometimes used in referring to types of
cells characteristic of the respiratory part of the lung.
Podzol: Any of a group of zonal soils that develop in a moist climate,
especially under coniferous or mixed forest.
Point source: A single stationary location of pollutant discharge.
Polarography: A method of quantitative or qualitative analysis based on
current-voltage curves obtained by electrolysis of a solution with
steadily increasing voltage.
Pollution gradient: A series of exposure situations in which pollutant con-
centrations range from high to low.
Polyacrylonitrile: A polymer made by reacting ethylene oxide and hydrocyanic
acid. Dyne! and Orion are examples.
Polyamides: Polymerization products of chemical compounds which contain
amino (-NHp) and carboxyl (-COOH) groups. Condensation reactions
between the groups form amides (-CONHL). Nylon is an example of
a polyamide.
Polycarbonate: Any .of various tough transparent thermoplastics characterized
by high impact strength and high softening temperature.
Polycythemia: An increase above the normal in the number of red cells in the
blood.
Polyester fiber: A man-made or manufactured fiber in which the fiber-
forming substance is any long-chain synthetic polymer composed of
at least 85 percent by weight of an ester of a dihydric alcohol and
terephthalic acid. Dacron is an example.
G-38
-------
Polymer: A large molecule produced by linking together many like molecules.
Polymerization: In fiber manufacture, converting a chemical monomer (simple
molecule) into a fiber-forming material by joining many like molecules
into a stable, long-chain structure.
Polymorphic monocyte: Type of leukocyte with a multi-lobed nucleus.
Polymorphonuclear leukocytes: Cells which represent a secondary non-
specific cellular defense mechanism. They are transported to the lungs
from the bloodstream when the burden handled by the alveolar macrophages
is too large.
Polysaccharides: Polymers made up of sugars. An example is glycogen which
consists of repeating units of glucose.
Polystyrene: A thermoplastic plastic which may be transparent, opaque,
or translucent. It is light in weight, tasteless and odorless, it
also is resistant to ordinary chemicals.
Polyurethane: Any of various polymers that contain NHCOO linkages and are
used especially in flexible and rigid foams, elastomers and resins.
Pores of Kohn: Also known as interalveolar pores; pores between air cells.
Assumed to be pathways for collateral ventilation.
Precipitation: Any of the various forms of water particles that fall from
the atmosphere to the ground, rain, snow, etc.
Precursor: A substance from which another substance is formed; specifically,
one of the anthropogenic or natural emissions or atmospheric constituents
which reacts under sunlight to form secondary pollutants comprising
photochemical smog.
Probe: In air pollution sampling, the tube or other conduit extending
into the atmosphere to be sampled, through which the sample passes
to treatment, storage and/or analytical equipment.
Proline: An amino acid, CcHgNCL, that can be synthesized from glutamate
by animals.
Promonocyte: An immature monocyte not normally seen in the circulating
blood.
Proteinuria: The presence of more than 0.3 gm of urinary protein in a
24-hour urine collection.
Pulmonary: Relating to the lungs.
Pulmonary edema: An accumulation of excessive amounts of fluid in the lungs.
G-39
-------
Pulmonary lumen: The spaces in the interior of the tubular elements of
the lung (bronchioles and alveolar ducts).
Pulmonary resistance: Sum of airway resistance and viscous tissue resistance.
Purine bases: Organic bases which are constituents of DNA and RNA, including
adenine and guanine.
Purulent: Containing or forming pus.
Pyrimidine bases: Organic bases found in DNA and RNA. Cytosine and
thymine occur in DNA and cytosine and uracil are found in RNA.
QRS: Graphical representation on the electrocardiogram of a complex of three
distinct waves which represent the beginning of ventricular contraction.
Rainout: Removal of particles and/or gases from the atmosphere by their
involvement in cloud formation (particles act as condensation nuclei,
gases are absorbed by cloud droplets), with subsequent precipitation.
Rayleigh scattering: Coherent scattering in which the intensity of the
light of wavelength g, scattered in any direction making an angle
with the incident direction, is.directly proportional to 1 + cos r
and inversely proportional to g .
Reactive dyes: Dyes which react chemically with cellulose in fibers under
alkaline conditions. Also called fiber reactive or chemically
reactive dyes.
Reduction: Acceptance of electrons by an ion or molecule.
Reference method (RM): For N02, an EPA-approved gas-phase chemiluminescent
analyzer and associated calibration techniques; regulatory specifications
are described in Title 40, Code of Federal Regulations, Part 50,
Appendix F. Formerly, Federal Reference Method.
Residual capacity: The volume of air remaining in the lungs after a maximum
expiratory effort; same as residual volume.
Residual volume (RV): The volume of air remaining in the lungs after a
maximal expiration. RV = TLC - VC
Resin: Any of various solid or semi-solid amorphous natural organic sub-
stances, usually derived from plant secretions, which are soluble in
organic solvents but not in water; also any of many synthetic substances
with similar properties used in finishing fabrics, for permanent press
shrinkage control or water repellency.
Ribosomal RNA: The most abundant RNA in a cell and an integral constituent
of ribosomes.
G-40
-------
Ribosomes: Discrete units of RNA and protein which are instrumental in the
synthesis of proteins in a cell. Aggregates are called polysomes.
Runoff: Water from precipitation, irrigation or other sources that flows
over the ground surface to streams.
Sclerosis: Pathological hardening of tissue, especially from overgrowth
of fibrous tissue or increase in interstitial tissue.
Selective leaching: The removal of one element from a solid alloy by
corrosion processes.
Septa: A thin wall dividing two cavities or masses of softer tissue.
Seromucoid: Pertaining to a mixture of watery and mucinous material such
as that of certain glands.
Serum antiprotease: A substance, present in serum, that inhibits the activity
of proteinases (enzymes which destroy proteins).
Sigma (s) bonds: Bonds in which electron density is symmetrical about a
line joining the bonded atoms.
Silo-filler's disease: Pulmonary lesion produced by oxides of nitrogen
produced by fresh silage.
Single breath nitrogen elimination rate: Percentage rise in nitrogen fraction
per unit of volume expired.
Single breath nitrogen technique: A procedure in which a vital capacity
inspiration of 100 percent oxygen is followed by examination of nitrogen
in the vital capacity expirate.
Singlet state: The highly-reactive energy state of an atom in which certain
electrons have unpaired spins.
Sink: A reactant with or absorber of a substance.
Sodium arsenite: Na^AsO-, used with sodium hydroxide in the absorbing solu-
tion of a 24-hour integrated manual method for f*^-
Sodium dithionite: A strong reducing agent (a supplier of electrons).
Sodium metabisulfite: Na,S90(., used in absorbing solutions of N0? analysis
methods. * L 5
Sorb: To take up and hold by absorption or adsorption.
G-41
-------
Sorbent: A substance that takes up and holds another by absorption or
adsorption.
Sorbitol dehydrogenase: An enzyme that interconverts the sugars, sorbitol
and fructose.
Sorption: The process of being sorbed.
Spandex: A manufactured fiber in which the fiber forming substance is a
long chain synthetic elastomer composed of at least 85 percent of a
segmented polyurethane.
Spectrometer: An instrument used to measure radiation spectra or to deter-
mine wavelengths of the various radiations.
Spectrophotometry: A technique in which visible, UV, or infrared radiation
is passed through a substance or solution and the intensity of light
transmitted at various wavelengths is measured to determine the spectrum
of light absorbed.
Spectroscopy: Use of the spectrometer to determine concentrations of an
air pollutant.
Spermatocytes: A cell destined to give rise to spermatozoa (sperm).
Sphingomyelins: A group of phospholipids found in brain, spinal cord, kidney
and egg yolk.
Sphygomenometer: An apparatus, consisting of a cuff and a pressure gauge,
which is used to measure blood pressure.
Spirometry: Also called pneometry. Testing the air capacity of the lungs
with a pneometer.
Spleen: A large vascular organ located on the upper left side of the abdominal
cavity. It is a blood-forming organ in early life. It is a storage
organ for red corpuscles and because of the large number of macrophages,
acts as a blood filter.
Sputum: Expectorated matter, especially mucus or mucopurulent matter expec-
torated in diseases of the air passages.
Squamous: Scale-like, scaly.
Standard deviation: Measure of the dispersion of values about a mean
value. It is calculated as the positive square root of the average of
the squares of the individual deviations from the mean.
i
Standard temperature and pressure: 0°C, 760 mm mercury.
Staphylococcus aureus: A spherically-shaped, infectious species of bacteria
found especially on nasal mucous membrane and skin.
G-42
-------
Static lung compliance (C, . ,): Measure of lung's elastic recoil (volume
change resulting from inange in pressure) with no or insignificant air-
flow.
Steady state exposure: Exposure to air pollutants whose concentration
remains constant for a period of time.
Steroids: A large family of chemical substances comprising many hormones and
vitamins and having large ring structures.
Stilbene: An aromatic hydrocarbon C,4H1? used as a phosphor and in making
dyes.
Stoichiometric factor: Used to express the conversion efficiency of a non-
quantitative reaction, such as the reaction of N02 with azo dyes in air
monitoring methods.
Stoma: A minute opening or pore (plural is stomata).
Stratosphere: That region of the atmosphere extending from 11 km above the
surface of the earth to 50 km. At 50 km above the earth temperature
rises to a maximum of 0 C.
Streptococcus pyogenes: A species of bacteria found in the human mouth,
throat and respiratory tract and in inflammatory exudates, blood stream,
and lesions in human diseases. It causes formation of pus or even fatal
septicemias.
Stress corrosion cracking: Cracking caused by simultaneous presence of
tensile stress and a specific corrosive medium. The metal or alloy is
virtually unattached over most of its surface, while fine cracks progress
through it.
Strong interactions: Forces or bond energies holding molecules together.
Thermal energy will not disrupt the formed bonds.
Sublobular hepatic necrosis: The pathologic death of one or more cells, or
of a portion of the liver, beneath one or more lobes.
Succession: The progressive natural development of vegetation towards
a climax, during which one community is gradually replaced by others.
Succinate: A salt of succinic acid involved in energy production in the
citric acid cycle.
Sulfadiazine: One of a group of sulfa drugs. Highly effective against
pneumococcal, staphlococcal, and streptococcal infections.
Sulfamethazine: An antibacterial agent of the sulfonamide group, active
against homolytic streptococci, staphytococci, pneumococci and meningococci
Sulfanilimide: A crystalline sulfonamide (CgHgNpOpS), the amide of sulfanilic
acid and parent compound of most sulfa arugs.
G-43
-------
Sulfhydryl group: A chemical radical consisting of sulfur and hydrogen
which confers reducing potential to the chemical compound to which it is
attached (-SH).
Sulfur dioxide (SOp): Colorless gas with pungent odor released primarily from
burning of fossil fuels, such as coal, containing sulfur.
Sulfur dyes: Used only on vegetable fibers, such as cottons. They are
insoluble in water and must be converted chemically in order to be
soluble. They are resistant (fast) to alkalies and washing and fairly
fast to sunlight.
Supernatant: The clear or partially clear liquid layer which separates
from the hbmogenate upon centrifugation or standing.
Surfactant: A substance capable of altering the physiochemical nature of
surfaces, such as one used to reduce surface tension of a liquid.
Symbiotic: A close association between two organisms of different species in
which at least one of the two benefits.
Synergistic: A relationship in which the combined action or effect of two
or more components is greater than that of the components acting separately.
Systolic: Relating to the rhythmical contraction of the heart.
Tachypnea: Very rapid breathing.
12
Terragram (Tg): One million metric.tons, 10 grams.
Teratogenesis: The disturbed growth processes resulting in a deformed
fetus.
Teratogenic: Causing or relating to abnormal development of the fetus.
Threshold: The level at which a physiological or psychological effect begins
to be produced.
Thylakoid: A membranous lamella of protein and lipid in plant chloroplasts
where the photochemical reactions of photosynthesis take place.
Thymidine: A nucleoside (C,nH, .NpO,-) that is composed of thymine and
deoxyribose; occurs as a structural part of DNA.
Tidal volume (V,): The volume of air that is inspired or expired in a single
breath during regular breathing.
Titer: The standard of strength of a volumetric test solution. For example,
the titration of a volume of antibody-containing serum with another
volume containing virus.
G-44
-------
Tocopherol: a-d-tocopherol is one form of Vitamin E prepared synthetically.
The a form exhibits the most biological activity. It is an antioxidant
and retards rancidity of fats.
Torr: A unit of pressure sufficient to support a 1 mm column of mercury;
760 torr = 1 atmosphere.
Total lung capacity (TLC): The sum of all the compartments of the lung, or
the volume of air in the lungs at maximum inspiration.
Total suspended particulates (TSP): Solid and liquid particles present in
the atmosphere.
Trachea: Commonly known as the windpipe, a cartilaginous air tube extending
from the larnyx (voice box) into the thorax (chest) where it divides,
serving as the entrance to each of the lungs.
Transaminase: Aminotransferase; an enzyme transferring an ami no group from
an a-amino acid to the carbonyl carbon atom of an a-keto acid.
Transmissivity (UV): The percent of ultraviolet radiation passing through a
a medium.
Transmittance: The fraction of the radiant energy entering an absorbing
layer which reaches the layer's further boundary.
Transpiration: The process of the loss of water vapor from plants.
Triethanolamine: An amine, (HOChLCH^^N, used in the absorbing solution
of one analytical method for NCL.
Troposphere: That portion of the atmosphere in which temperature decreases
rapidly with altitude, clouds form, and mixing of air masses by convection
takes place. Generally extends to about 7 to 10 miles above the earth's
surface.
Type 1 epithelial cells: Squamous cells which provide a continuous lining
to the alveolar surface.
Type I pneumonocytes: Pulmonary surface epithelial cells.
Type II pneumonocytes: Great alveolar cells.
Ultraviolet: Light invisible0to the human eye of wavelengths between 4x10
and 5x!0"9 m (4000 to BOA).
Urea-formaldehyde resin: A compound composed of urea and formaldehyde in
an arrangement that conveys thermosetting properties.
G-45
-------
Urobilinogen: One of the products of destruction of blood cells; found in
the liver, intestines and urine.
Uterus: The womb; the hollow muscular organ in which the impregnated ovum
(egg) develops into the fetus.
Vacuole: A minute space in any tissue.
Vagal: Refers to the vagus nerve. This mixed nerve arises near the medulla
oblongata and passes down from the cranial cavity to supply the larynx,
lungs, heart, esophagus, stomach, and most of the abdominal viscera.
Valence: The number of electrons capable of being bonded or donated by
an atom during bonding.
Van Slyke reactions: Reaction of primary amines, including ami no acids,
with nitrous acid, yielding molecular nitrogen.
Variance: A measure of dispersion or variation of a sample from its
expected value; it is usually calculated as the square root a sum of
squared deviations about a mean divided by the sample size.
Vat dyes: Dyes which have a high degree of resistance to fading by light,
NO and washing. Widely used on cotton and viscose rayon. Colors are
brilliant and of almost any shade. The name was originally derived
from their application in a vat.
Venezuelan equine encephalomyelitis: A form of equine encephalomyelitis
found in parts of South America, Panama, Trinidad, and the United States,
and caused by a virus. Fever, diarrhea, and depression are common. In
man, there is fever and severe headache after an incubation period of 2
to 5 days.
Ventilatory volume (V^): The volume of gas exchanged between the lungs and
the atmosphere that occurs in breathing.
Villus: A projection from the surface, especially of a mucous membrane.
Vinyl chloride: A gaseous chemical suspected of causing at least one type
of cancer. It is used primarily in the manufacture of polyvinyl
chloride, a plastic.
Viscose rayon: Filaments of regenerated cellulose coagulated from a solution
of cellulose xanthate. Raw materials can be cotton 1 inters or chips
of spruce, pine, or hemlock.
o
Visible region: Light between the wavelengths of 4000-8000 A.
Visual range: The distance at which an object can be distinguished from
background.
Vital capacity: The greatest volume of air that can be exhaled from the
lungs after a maximum inspiration.
G-46
-------
Vitamin E: Any of several fat-soluble vitamins (tocopherols), essential
in nutrition of various vertibrates.
Washout: The capture of gases and particles by falling raindrops.
Weak interactions: Forces, electrostatic in nature, which bind atoms and/or
molecules to each other. Thermal energy will disrupt the interaction.
Also called van der Waal's forces.
Wet deposition: The process by which atmospheric substances are returned
to earth in the form of rain or other precipitation.
Wheat germ lipase: An enzyme, obtained from wheat germ, which is capable
of cleaving a fatty acid from a neutral fat; a lipolytic enzyme.
X-ray fluorescence spectrometry: A nondestructive technique which utilizes
the principle that every element emits characteristic x-ray emissions
when excited by high-energy radiation.
Zeolites: Hydrous silicates analogous to feldspars, occurring in lavas
and various soils.
Zooplankton: Minute animal life floating or swimming weakly in a body of water,
G-47
------- |