-------
90.0
Philadelphia-WRAC
1.0 10.0
Aerodynamic Diameter, ft (Mm)
100.0
90.0
Phoenix-WRAC
E
^3.
«
Q"
Mode
1
2
3
MMAD oa %Mass
0.188 1.54 22.4
1.70 1.90 13.8
16.4 2.79 63.9
1.0 10.0
Aerodynamic Diameter, ft (Mm)
100.0
Figure 3-21. Impactor size distribution measurement generated by Lundgren et al. with
the Wide Range Aerosol Classifier: (a) Philadelphia and (b) Phoenix. Note
the presence of more coarse mode particles in the size range 1 to 2.5 /j,m, in
the dryer environment of Phoenix.
Source: Adapted from Lundgren and Hausknecht, 1982b.
major portion of the coarse fraction of PM10. An attempt has been made to fit the distribution
with three, log-normal distributions. In this case, the fit is poor. In the Phoenix case the
accumulation mode cannot be defined other than that the MMAD is below 0.2 //m. The coarse
particle fractions are very wide suggesting the possibility of two or more modes (Figure 3-24).
The material between 1 and 2.5 //m is not a new mode but an indication of either an artifact due
to particle bounce, or a long-lasting tail of the coarse distribution.
The existing size-distribution data were recently reviewed by Lundgren and Burton (1995),
with emphasis on the coarse mode. They concluded that the coarse mode could be reasonably
well described by a lognormal distribution with a mass median aerodynamic diameter (MMAD)
of 15 to 25 //m and a mode spread (og) of approximately two. This allows one to calculate, for a
freshly-generated coarse mode aerosol, that about 1% of the
3-160
-------
mass would be less than 2.5 //m and only about 0.1% would be less than 1.0 //m in diameter.
This conclusion is confirmed by data from Whitby in which a wind change allowed a
measurement of fresh coarse mode aerosol (National Research Council, 1979). As can be seen
in Figure 3-22, the intermodal mass, 1.0 to 2.5 //m, was not affected, even though the mass at 20
//m increased substantially.
Hunter-Liggett
9-14-72
0.1 1 2.5
Geometric Diameter, D, , |jm
10
Figure 3-22. Example of aged and fresh coarse mode particle size distributions. A sudden
wind change brought fresh wind-blown dust to the sampler, operated as part
of the South Coast Air Quality Study. Note that there is only a very small
change in the intermodal mass, 1.0 to 2.5 /^m diameter, although there is a
major increase in the mass between 2.5 and 10 /^m in diameter.
Source: National Research Council (1979).
3-161
-------
Another extensive set of studies covering the full size range, but limited to the Chicago
area, has been reported by Noll and coworkers (Lin et al., 1993, 1994). They used an Andersen
impactor for smaller particles and a Noll Rotary Impactor for larger particles. Results of Lin et
al. also indicate a bimodal mass distribution. For the shorter time interval measurements (8 or
16 h), the average MMAD for the fine mode was 0.42 //m, with a og around two. The average
MMAD of the coarse mode was 26±8 //m, with a og varying from 2.0 to 3.5. As shown in
Figure 3-23, the results of Noll and coworkers (Lin et al., 1993, 1994) also suggest that in some
instances little coarse mode material is found in the intermodal region, 1.0 to 2.5 //m. Lin et al.
(1993) combined material on the 0.65 to 1.0 //m and the 1.0 to 2.0 //m stages before weighing.
Therefore, the MMAD of the accumulation mode is not as well defined as it might be, and could
be smaller than that given by the fitting process. Therefore, these results cannot be used to show
that some fine PM is found above 1.0 //m. When fitted to two log-normal distributions the fit is
poor and the coarse mode is very wide. The fit with three log-normal distributions is used to
show the possibility of particle bounce or a second mode within the coarse particle size range
contributing to mass in the intermodal (1-2.5 //m) region.
3.7.6 Intermodal Region
3.7.6.1 Coarse Mode
The question then arises, what portion of the coarse mode material found in the intermodal
region is real and what portion is artifact? As discussed in Section 3.3.3.2.4, the optical size
may differ from the geometric or aerodynamic size. Optical counters are normally calibrated
with latex particles, or other particles of a specific refractive index. Atmospheric particles with
different refractive indices would be incorrectly sized if the difference in refractive index
resulted in a difference in the amount of light scattered by the particles (Wilson et al., 1988; Liu
et al., 1992; Hering and McMurry, 1991). For particle counters using lasers, particles of
different sizes within the 0.5 to 1.0 //m range may give the same light scattering (Hering and
McMurry, 1991; Kim 1995).
In the case of impactors, it is possible that an artifact may arise from particle bounce, from
fragmentation of larger agglomerates, or from loosening of material from other surfaces by
impacting particles. The problem of particle bounce in impactors has been treated
3-162
-------
20.0
10.0
20.0
10.0
1.0 10.0
Aerodynamic Diameter,
100.0
(Mm)
1.0 10.0
Aerodynamic Diameter,
100.0
(Mm)
40.0
o>
o
30.0
20.0
ra
o
15.0
0.1 1.0 10.0
Aerodynamic Diameter, Et (him)
0.0'
100.0 0.1 1.0 10.0 100.0
Aerodynamic Diameter, Efe (|jm)
Figure 3-23. Size distributions reported by Noll and co-workers from the Chicago area
using an Anderson impactor for the smaller particles and a Noll Rotary
Impactor for the larger particles.
Source: Lin et al. (1993).
theoretically and practically in many studies (Wang and John, 1987, 1988). Most workers coat
the coarse particle stages with a grease or oil to reduce bounce. However, as the surface
becomes covered with aerosols, a particle may impact another particle instead of the surface and
either bounce to a lower stage or cause deagglomeration and reentrainment of previously
collected particles (John et al., 1991; John and Sethi, 1993). As impactor collection plates
become loaded or as inlet upper size cut surfaces become dirty, the magnitude of the effect
increases (Ranade et al., 1990; John and Wang, 1991). One result is a lowering of the effective
cut point of the inlet and the impactor stages. Thus, it is uncertain how much of the mass found
in the intermodal size range is real and how much is due to artifacts.
3-163
-------
There are several reasons to believe, however, that some of the intermodal mass may be
real. For example, Lundgren and Burton (1995) point out that the lifetime of particles in the
atmosphere is a strong function of their aerodynamic size. Thus, while freshly generated coarse
mode aerosol may have a MMAD of 20 //m, with only 1% below 2.5 //m, as the aerosol ages the
larger particles will rapidly fall out, leaving a distribution enriched with particles in the small-
size tail of the distribution.
A second explanation has to do with the possible multimodal nature of dust generated by
wind or vehicular traffic. A study by the U.S. Army (Pinnick et al., 1985) measured the size
distribution of dust generated by heavy vehicles driven on unpaved roadways in the arid
southwestern United States. A variety of light-scattering instruments were used and were
recalibrated for the refractive index of the soil particles. The occurrence of strong surface winds
(gusts of 15 to 20 m s"1) during the study permitted, in addition to the vehicular-generated dust,
some measurements of windblown dust. There were some differences between sandy soil and
silty soil, and between dust generated by vehicular traffic and by wind. However, all situations
produced a bimodal size distributions. The upper mode had an MMAD that ranged from 35 to
53 //m, with og from 1.37 to 1.68. Of particular interest, however, was a second mode having an
MMAD that varied from 6.2 to 9.6 jim, with a og from 1.95 to 2.20. (One measurement in silty
soil had an MMAD of 19.4 //m.) The MMADs of the smaller coarse particle modes are
significantly smaller than those coarse mode MMADs observed by Lundgren or Noll. An
example of vehicular generated dust is shown in Figure 3-24. Note that the differential mass is
plotted on a logarithmic scale. These results suggest that in arid areas, significant soil material,
generated by traffic or wind, may be found in the intermodal region.
A third reason for believing that the intermodal mass is real has to do with the relative size
efficiency of particle removal equipment used on power plants and other large industrial
facilities. Older control devices, which may still be used in some applications, allow significant
particle mass to escape. Overall mass efficiencies are approximately 80% for cyclones and 94%
for scrubbers. Modern control devices have very high overall efficiencies, 99.2% for
electrostatic precipitators (ESP) and 99.8% for baghouses. However, all of these devices have
efficiencies for coarse particles that decrease with decreasing size. Efficiencies typically reach a
minimum between 0.1 and 1 //m and increase for particles
3-164
-------
o
^
Q
0)
O
c
o
o
(0
m
E
0>
o
n
o
Q.
fl>
O)
n
O)
10a
10
10J
_ 5 ton truck (8-
101
speed)
0.1 0.2 0.4
^ 2 4 W 20 40
Geometric Diameter, D , urn
100 200
Figure 3-24. Size distribution of dust generated by driving a truck over an unpaved test
track. "Error bars" show the range of distributions from individual tests.
The curves shown are log-normal fits to the average distribution. The
original data were plotted as log radius but have been replotted versus log
diameter. The shaded bar between lines at diameters of 1.0 and 2.5 fj,m
indicates how the smaller size mode of this dust can contribute to the
intermodal mass found in arid areas (see Figures 3-21 and 3-23).
Source: Pinnick et al. (1985).
smaller than 0.1 //m. Thus, although most of the particulate mass is captured, the particles that
do escape are in the smaller size range. Data from U.S. EPA, plotted in Figure 3-25, (U.S.
Environmental Protection Agency, 1995) show the size distribution of fly ash from a pulverized
coal power plant and the size distribution of the material escaping from the various control
devices. The small-size tail of the coarse mode escapes preferentially and may possibly
contribute material to the intermodal region.
Cheng et al. (1985) reported experimental measurements from an atmospheric fluidized-
bed coal combustor. Size distribution measurements, made downstream of a cyclone and again
downstream from baghouse filtration of the material leaving the cyclone, are shown in
3-165
-------
^0.9
o>°-8
° 0.7
fo,
0.3
(A
« 0.2
3 0.1
0.0
Q"
1.4
O)
° 1.2
<
* 1.0
M
0.8
ra 0.4
<0.2
0.0
„ 0.6
Q
O) 0.5
-------
0.05 0.1 1.0 2.5
Stokes diameter,
400 p.
300
.0200
0.05 0.1 1.0 2.5
Stokes diameter, \im
Figure 3-26. Size distributions from a fluidized-bed, pulverized coal combustor, (a) after
initial cleanup by a cyclone collector and (b) after final cleanup by a
baghouse.
Source: Cheng et al. (1985).
A fourth piece of evidence comes from studies in which measurements are made of the
elemental composition of PM25 and PM10 or the coarse fraction of PM10. Elements
representative of soil type material have been found in the PM2 5 fraction. In a study in
Philadelphia that used dichotomous samplers, an amount of soil-type material equal to 5% of the
coarse mode fraction of PM10 was found in the PM25 fraction (Dzubay et al., 1988). Since the
virtual impactor used in the dichotomous sampler minimizes particle bounce and reintrainment,
this would appear to be the small-size tail of the coarse mode in the 1 to 2.5 jim size range.
5-167
-------
Similar results have been reported from the IMPROVE network (Eldred et al., 1994).
Elemental analysis suggested that soil-derived material, equal to 20% of the coarse fraction of
the PM10 sample, was found in the PM2 5 sample.
Thus, one can conclude that coarse mode material is found in the intermodal region. There
are reasons to suspect that a portion of this material is an artifact but that a portion is real coarse
mode material having an aerodynamic diameter between 1.0 and 2.5 jim. In either event, this
can lead to a misunderstanding of the source of the particles, to inappropriate control strategies,
or to exposure studies that fail to differentiate correctly between fine and coarse particles.
3.7.6.2 Fine Mode
This section discusses the source of fine mode material found in the intermodal region.
Early particle-counting data suggested that, with a few exceptions, significant mass of fine
mode material would not be found above 1 //m (see Figures 3-13, 3-18, 3-19, and 3-20).
However, impactor studies, on some occasions, have observed significant mass on stages with a
cut point of 1 (j,m. In some instances, the MMAD of the fine mode was as large as 1 //m (John
et al., 1990). The change in relative humidity produced by a few degrees change in temperature
can significantly modify the MMAD of an ambient aerosol size distribution. As the RH
approaches 100%, accumulation mode aerosols, with dry sizes below 1.0 //m in diameter, may
grow larger than 2.5 //m in diameter, be rejected by PM2 5 samples, and be counted as coarse
particles.
Before examining additional field data demonstrating the effect of relative humidity on
particle size, it is useful to review some basic information on the interaction of water vapor with
the components of fine particles. Sulfuric acid (H2SO4) is a hygroscopic substance. When
exposed to water vapor a H2SO4 droplet will absorb water vapor and grow in size until an
equilibrium exists between the liquid water concentration in the H2SO4 solution droplet and the
water vapor concentration in the air. The amount of water in the sulfuric acid droplet will
increase and decrease smoothly as the RH increases and decreases. Ammonium sulfate,
(NH4)2SO4, however, is deliquescent. If initially a crystal in dry air, it will remain a crystal until
a certain RH is reached; at this point it will absorb water and become a solution droplet. The RH
at which this happens, ~ 80% RH in the case of
3-168
-------
NH4)2SO4, is called the deliquescent point. At RH's above the deliquescent point the (NH4)2SO4
droplets are hygroscopic, gaining or losing water reversibly as the RH increases or decreases. If
the RH decreases below the deliquescent point the solution droplet becomes supersaturated and
unstable to crystallization. However, sub-micron sized droplets will remain supersaturated until
a significantly lower RH, known as the crystallization or efflorescent point is reached. The
crystallization point decreases with decreasing droplet size and decreasing purity (Whitby,
1984). Thus, for a deliquescent substance, a plot of droplet diameter or water content as a
function of RH will have two lines, one for increasing RH and another for decreasing RH. An
example of this phenomenon, known as hysteresis, is shown in Figure 3-27. Table 3-16 shows
the RH at the deliquescent and crystallization points for some compounds found in the
atmosphere.
Much experimental and theoretical effort has gone into understanding this process. The
basic theory was elucidated by Hanel (1976). Much experimental work has been done on
atmospheric species (e.g., Tang and Munkelwitz, 1977, 1993; Richardson and Spann, 1984).
The electrodynamic balance, by which single particles can be studied, has advanced the
understanding of particle-water vapor equilibrium, especially for particles in metastable states,
e.g., the supersaturated solution particles which are responsible for the hysteresis loop shown in
Figure 3-27 (Cohen et al., 1987a,b; Chan et al., 1992; Kim et al., 1994). Ammonium nitrate,
NH4NO3, because of its volatility, is difficult to handle but has been studied successfully by
Richardson and Hightower (1987). The aerosol equilibria models developed by Seinfeld and co-
workers allow calculation of the water content of bulk solution as a function of relative humidity
(Kim and Seinfeld). The model SCAPE (Kim et al., 1993a,b) has been used to estimate the
contribution of water to suspended aerosol mass in the California South Coast Air Basin using
particle composition data from the 1987 Southern California Air Quality Study (Meng et al.,
1995). From midnight to early morning, when the temperature is low and relative humidity is
high, water was usually the predominant aerosol substance. Paniculate water in the winter was
estimated to be considerably larger than in the summer at each of the four sites studied.
The water content of a sub-micron sized droplet, and therefore its size, depends not only on
the dry size but is a result of a balance between surface tension and solute concentration (Li et
al., 1992). Pure water is in equilibrium with its vapor when the RH
3-169
-------
2.0
o
Q.
Q
1.5
1.0
H
30
50 70
RH, %
8
7
6
4
3
2
1
90
o
Q.
"3.
Figure 3-27. Particle growth curves showing fully reversible hygroscopic growth of
sulfuric acid (H2SO4) particles, deliquescent growth of ammonium sulfate
[(NH4)2 SO4] particles at about 80% relative humidity (RH), hygroscopic
growth of ammonium sulfate solution droplets at RH greater than 80%, and
hysteresis (the droplet remains supersaturated as the RH decreases below
80%) until the crystallization point is reached.
Source: National Research Council (1993) adapted from Tang (1980).
TABLE 3-16. RELATIVE HUMIDITY OF DELIQUESCENCE AND
CRYSTALLIZATION FOR SEVERAL ATMOSPHERIC SALTS3
Compound
(NH4)2S04
NH4HSO4
NH4NO3
NaCl
Deliquescence
79.9 ±0.5
39.0±0.5b
61.8
75. 3 ±0.1
Crystallization0
37 ±2
42
aTaken from Tang and Munkelwitz (1993) unless otherwise indicated.
bTang and Munkelwitz (1977).
°Shaw and Rood (1990) and references therein.
3-170
-------
equals 100% and is therefore, stable, i.e. the rate of evaporation equals the rate of condensation.
The water in a solution will be in equilibrium with water vapor at a lower water vapor
concentration because the presence of solute molecules or ions lower the rate of evaporation.
Therefore, a solution will absorb water and become more dilute, increasing the water vapor
concentration needed for equilibrium until the solution water vapor concentration required for
equilibrium matches the ambient water vapor concentration or RH. As the droplet size decreases
the surface tension increases and the vapor pressure of water required to maintain equilibrium
increases. Therefore, the smaller the dry size of the particle, the less the amount of growth as
RH increases.
Theoretical calculations of the growth of various sizes of ammonium sulfate particles and
an experimental verification of such calculations, using a simulation of the humidification
process in the human lung, are shown in Figure 3-28. Note the very rapid increase in the amount
of water and in the diameter of the aerosol particle as the relative humidity approaches 100%
RH. Considering the difficulty of measuring relative humidity accurately between 99 and 100%,
theory and experiment are in reasonable agreement. As can be seen the effect of surface tension
is most important for particles with dry size less than 100 nm (0.1 //m). This phenomenon may
be of importance in considering the biological effect of water-soluble pollutants. Accumulation
mode particles will be diluted when exposed to humidification in the lungs. Ultrafine or nuclei
mode particles will not be diluted as much. In the atmospheric aerosol the number distribution
will almost always be dominated by particles below 100 nm (see Section 3.1.2). However,
aerosols generated in the laboratory for exposure studies probably lack the smaller particles
found in the atmosphere. This provides a hypothesis for the difference in effects observed in
epidemiological studies and laboratory exposure studies. The importance of this more
concentrated, ultrafine droplet component of the atmospheric aerosol may have been neglected
because most of the experimental studies of hygroscopicity have used near-micron-sized
particles. However, in the modeling of deposition of hygroscopic particles, workers, such as
Martonen (1993), have corrected the experimental curves of particle size as a function of RH,
based on measurements of near micron-sized particles, to account for the effects of surface
tension on ultrafine particles.
3-171
-------
u. 4
Theoretical Prediction at 22 °C
00000 Experimental Measurements
1216
125
50 100 150
NH4 HSO4 Dry Particle Diameter (nm)
200
I I
Theoretical Prediction at 22 °C
O O O O O Experimental Measurements
-125
". 4
is
U.
I
O
I I I I I I I I I I I I I
50 100 150
(NH4)2 SO4 Dry Particle Diameter (nm)
200
Figure 3-28. Theoretical predictions and experimental measurements of growth of
NH4HSO4 and (NH4)2SO4 particles at relative humidity between 95 and
100%.
Source: Li et al. (1992).
3-172
-------
In addition to the laboratory studies discussed above there are some measurements on the
effect of RH changes on atmospheric aerosol. McMurry and co-workers have made use of a
Tandem Differential Mobility Analyzer (TDMA) system (Rader and McMurry, 1986) to
measure the change in particle size with changes in relative humidity at Claremont, CA, as part
of the Southern California Air Quality Study (SCAQS) (McMurry and Stolzenberg, 1989) and at
the Grand Canyon National Park, AZ, as part of the Navajo Generating Station Visibility Study
(Zhang et al., 1993; Pitchford and McMurry, 1994). One mobility analyzer is used to isolate a
narrow size distribution. After humidification the size distribution of this fraction is measured.
An example is shown in Figure 3-29. Note that Figure 3-29 is a number size distribution not a
mass size distribution. Particle growth with increasing RH is evident. However, between 70 and
91% RH the distribution splits into less-hygroscopic and more-hygroscopic components.
Pitchford and McMurry (1994) attribute this splitting to external mixing, i.e. there are two
relatively distinct classes of particles, both containing some soluble and some non-soluble
material, with the more hygroscopic component containing significantly more soluble and
hygroscopic material. A summary of the results of these studies is given in Table 3-17 (Zhang et
al., 1993). The difference in growth rates may be due both to size and to variation in
composition as a function of size. The lower growth factor for 0.2 //m particles in Claremont
relative to the Grand Canyon may be due to a higher concentration of non-soluble organic
material in Claremont.
While there is a significant amount of information on the hygroscopic properties of
inorganic compounds, much less is known about the hygroscopic properties of organic
components of the atmospheric aerosol. Saxena et al. (1995) have examined the hygroscopic
properties of several organic species and noted that water soluble organics may be hygroscopic
or deliquescent. Using concurrent cascade impactor samples, they determined the composition
of the Grand Canyon and Claremont aerosol, whose size distribution as a function of relative
humidity was discussed above. They compared the observed water content at the higher relative
humidity with the water content calculated for the inorganic components. They concluded "that
the aggregate hygroscopic properties of inorganic particles are altered when organics are also
present. Furthermore, the alterations can be positive or negative. The findings are consistent
with the expectation that organics are
3-173
-------
o
o
o
.0
E
Initial Relative Humidity 53% Rl
Final Relative Humidity
• 7% RH
0 28% RH
49% RH
70% RH
91% RH
0.35
0.5
Diameter,
0.6
0.7
0.8
Figure 3-29. Tandem Differential Mobility Analyzer measurements of the sensitivity of
particle size to relative humidity at Claremont, CA. Particle number
concentrations varied during the measurement, therefore changes in relative
size with humidity are meaningful but changes in number concentration are
not.
Source: McMurry and Stolzenberg (1989).
predominantly secondary (and thus likely to be hydrophilic) in nonurban areas and
predominantly primary (and hence hydrophobic) in urban areas".
Some experimental examples of the significant effect of relative humidity on ambient
aerosol size distributions are shown in Figure 3-30 (Lowenthal et al., 1995). In this work,
impactor collection and ion chromatographic analysis were used to measure sulfate size
distributions over short enough periods to demonstrate the effects of changing relative
humidities. The results suggest that the lognormal distribution is preserved as relative humidity
increases, but that the MMAD increases. This effect is especially pronounced as the relative
humidity approaches 100%.
3-174
-------
TABLE 3-17. SUMMARY OF HYGROSCOPIC GROWTH FACTORS3
Dry Size (//m)
0.05
0.2
0.4
0.5
Dry Size (//m)
0.05
0.10
0.20
0.30
0.40
1987 SCAQS, Claremont,
More Hygroscopic Peak
D/90 ± 3% Rff)
Dr(0% RH)
1.14±0.05
1.23 ±0.08
1.63 ±0.11
1.59 ±0.08
1990 NGS Visibility Study, Grand
More Hygroscopic Peak
D/89 ± 4% RJD
Dr(0% RH)
1.36 ±0.08
1.42 ±0.08
1.49±0.11
1.51 ±0.09
1.43 ±0.10
CA
Less Hygroscopic Peak
D/87 ± 2% RH^)
Dr(0% RH)
1.03 ±0.03
1.02 ±0.02
1.04 ±0.05
1.07 ±0.03
Canyon, AZ
Less Hygroscopic Peak
D/89 ± 4% Rff)
Dr(0% RH)
1.14±0.10
1.17±0.09
1.17±0.10
1.14±0.10
1.07 ±0.03
aValues are mean ± standard deviations.
-------
1=
c, 2
5 1
o
•o
• RH = 99% 8/12/90, 0200 hr
+ RH < 50%
-
Sulfate Size Distributions + +
+
0.01
0.1 1
Diameter (urn)
RH = 95% 8/4/90, 0200 hr
+ RH < 50%
IE
£ 2
o>
o
O)
_o
s 1
Sulfate Size Distributions
0.01
0.1 1
Diameter (um)
10
Figure 3-30. Example of growth in particle size due primarily to increases in relative
humidity from Uniontown, PA.
Source: Lowenthalet al. (1995).
3-176
-------
There are also studies of the behavior of ambient aerosols as the relative humidity is
reduced by heating the sampled air. Shaw and Rood (1990) report a study using a heated
integrating nephelometer in which crystallization RHs of 4 to 67% were observed. Similar
studies in Washington, D.C. by Fitzgerald et al. (1982) found no evidence of crystallization or
efflorescence when RH was reduced to 30% RH.
Further experimental evidence of the effect of decreasing relative humidity on aerosol size
distribution is provided by impactor data reported by Berner (1989) and is shown in Figure 3-31.
One impactor sampled aerosol in its humidified state directly from the atmosphere. The inlet of
a second impactor was warmed ~1 °C above the ambient temperature of ~5 °C in order to
evaporate most of the particle-bound water before collecting the aerosol. The water and other
volatile material in both the "wet" and the "dry" samples would evaporate in the laboratory prior
to weighing the impactor stages. As can be seen, in the ambient air most of the non-volatile
mass was above 1.0 //m with significant amounts above 2.5 //m. However, after heating the size
of the aerosol was reduced so that most of the non-volatile mass was below 1.0 //m. Berner
treated the distributions as monomodal and derived growth factors of 4.9 for fog and 4.1 for
haze. If the observations are treated as multimodal, good bimodal, or as shown in Figure 3-31,
trimodal fits are obtained. This splitting into "more" and "less" hygroscopic modes at high
relative humidity has been observed by McMurry and co-workers (McMurry and Stolzenberg,
1989; Zhang et al., 1993) (Figure 3-29) and Lowenthal et al. (1995) (Figure 3-30). In some
cases, reported by Pitchford and McMurry (1994), splitting into three modes of varying
hygroscopicity was observed. However, the separation into two "more" hygroscopic modes may
represent, as suggested by Berner, variations in relative humidity extremes during different parts
of the overnight sampling period.
In measuring light scattering with the integrating nephelometer, the aerosol community has
been very concerned about the difference in relative humidity and temperature in the ambient air
and in the volume of air in which particle scattering is actually measured (Covert et al., 1972;
Fitzgerald et al., 1982). Temperature differences between the measurement volume and
ambient air of 1 or 2 °C can change the relative humidity and change the observed light
scattering. Great efforts have been made to minimize this temperature
3-177
-------
80.0
Bologna Haze, Wet (Berner, 1989)
E
_a
oS
O)
o
40.0
0.0
0.01
Mode MMAD og VMass
1 0.204 1.69 9.9
2 1.95 1.97 23.5
3 3.50 2.65 66.5
=>X
"Si
^
fe
X*=
\
\
S*
V
^=
0.1 1.0
Aerodynamic Diameter, Qe
10.0
100.0
100.0
O)
o
50.0
Bologna Haze, Dry (Berner, 1989)
Mode MMAD
1 0.130
2 0.589
3 1.65
% Dry mass lost
upon heating
15.8%
1.01
0.1 1.0
Aerodynamic Diameter, Qe
10.0
100.0
70.0
35.0-
ra
o
Bologna Fog, Wet (Bemer, 1989)
Mode MMAD og %Mass
1 0.310 2.09 30.8
2 1.34 1.93 36.4
3 5.31 1.91 32.8
1.01
0.1 1.0 10.0
Aerodynamic Diameter, Qe (Mm)
100.0
200.0
3100.0-
o>
o
0.0
0
Bologna Fog, Dry (Bemer, 1989)
Mode MMAD oq %Mass
1 0.145 1.39 17.8
2 0.524 1.36 65.4
3 1.56 1.32 13.9
% Dry mass lost
upon heating
10.9%
01
0.1 1.0 10.0
Aerodynamic Diameter, Qe (Mm)
100.0
Figure 3-31. Mass size distribution of non-volatile aerosol material. The aerosol was
collected at ambient conditions, "wet", or after evaporation of water, "dry".
Source: Berner (1989).
3-178
-------
difference. However, researchers have not been nearly as careful in considering temperature and
relative humidity effects when measuring size distribution, either with impactors or particle
counters, even though effects have been reported in the early literature (Wagman et al., 1967;
Sverdrup and Whitby, 1980).
A recent paper by Cass and coworkers (Eldering et al., 1994) provides some insight into
how differences in RH resulting from heating can cause differences between particle-counting
distributions and impactor distributions. Particle size distributions were obtained by counting
particles by mobility (electrical aerosol analyzer) and light scattering (optical particle counter).
An example is shown in Figure 3-32. Almost no particles were found between 1.0 and 2.5 jim
diameter. When these particle number data were converted to total expected light scattering,
they agreed with measurements made by a heated, but not an unheated, integrating
nephelometer; and when converted to expected mass, agreed with filter measurements of dry
mass. Eldering et al. (1994) conclude that even the moderate heating occurring in mobility and
optical counters was enough to change the size of the particles, especially when the ambient air
was close to 100% RH. It seems likely that most particle counting systems produce some
heating of the aerosol, and thus some reduction of the measured particle size from that existing
in the ambient air. On the other hand, if particle-size measuring devices were located in air
conditioned or heated trailers or laboratories, the temperature of the sampled air would be
changed and the measured particle size distribution would be different from that existing in the
ambient air (Sverdrup and Whitby, 1980).
During the high relative humidities that occur at nighttime, growth of hygroscopic
components can result in the growth of some fine mode aerosol to diameters greater than 1.0 jim
and perhaps even above 2.5 |im. As can be seen in Figure 3-28, dry ammonium sulfate particles
having a dry diameter of 0.5 jim will grow to «2.5 |im at a relative humidity between 99 and
100%. When the relative humidity actually reaches 100%, the particles will continue to grow to
maintain the relative humidity at 100%, and eventually become fog droplets that are large
enough to be collected in the fraction larger than 2.5 jim. Ammonium sulfate particles with dry
sizes greater than 0.5 //m would also grow into the larger than 2.5 //m size range.
3-179
-------
nuu.uu
~ 80.00 -
"E
o
*E 60.00 -
3.
Q.
Q
g> 40.00 -
5
20.00 -
Of\f\
.00
August 27, 1987
> i
! 1
. .1
Claremont Case B !
2400-0
COO Df>T
500 PoT
.r
j
0500-0900 PST
0900-1
- 1300-1
too DOT
olIU ro 1
TOO DOT
1700-2400 PST
UfUl-ltti
0.01 0.
—
1
•-ti1
—
^
i
T
.1
it
I
I
I L":::"::"::
\\-.
t? \
Geometric Diameter,
1.0 1
P ,M"i
Figure 3-32. Example of particle-counting volume distribution obtained in Claremont,
CA. Compare to Figures 3-14 and 3-31. Heating of the sampled air by the
mobility and optical counters are believed to have resulted in a distribution
representative of a lower than ambient relative humidity.
Source: Eldering et al. (1994).
The addition of water to hygroscopic particles, discussed in the previous section, is a
reversible process. Particles absorb water and grow as RH increases; as RH decreases some of
the particle-bound water evaporates and the particles shrink. However, the large amount of
liquid water associated with hygroscopic particles at high relative humidity provides a medium
for liquid phase transformation process. A number of atmospheric process, which convert SO2
to sulfate or NOX to nitrate, can take place in water solutions but not in the gas phase. These
processes are not reversible but lead to an accumulation of sulfate or nitrate and lead to an
increase in the dry size of the particle. Of course as more sulfate or nitrate is added to the
particle it will absorb more water so that the wet size will also increase.
The first observation and clear discussion of these combined effects of relative humidity on
growth and SO2 conversion to sulfate are given by Hering and Friedlander (1982) as shown in
3-180
-------
Table 3-18. Using a low pressure impactor, they observed that days with higher relative
humidity had higher sulfate concentration and higher MMAD's compared to days with lower
relative humidity. Hering and Friedlander (1982) named the small mode the condensation mode
and suggested that it was formed by the gas phase conversion of SO2 to sulfate and subsequent
nucleation, coagulation, and growth by condensation. They named the larger mode the droplet
mode and discussed possible formation mechanisms. This mode is now believed to result from
the reaction of SO2 in fog or cloud droplets (Meng and Seinfeld, 1994).
TABLE 3-18. COMPARISON OF SULFATE CONCENTRATION
AND MASS MEAN DIAMETERS OF AEROSOLS FOR DAYS
WITH HIGHER AND LOWER RELATIVE HUMIDITY
Minimum RH, %
Maximum RH, %
Sulfate concentration, //g/m3
Mass median aerodynamic diameter, //m
Low RH Days
17-35
45-68
3 -9
0.20 ±0.02
High RH Days
26-66
69 - 100
3 -52
0.54 ±0.07
Source: Hering and Friedlander (1982).
In a series of papers McMurry and co-workers make use of the aerosol growth law,
originally developed by Heisler and Friedlander (1977), to study the mechanism and rates of
sulfate formation in ambient air (McMurry et al., 1981; McMurry and Wilson, 1982, 1983).
They were able to apportion growth to condensation and droplet mechanisms and observed
droplet growth in particles up to 3 //m in diameter.
A process of aerosol growth due to increasing relative humidity (Figure 3-33) has also
been utilized by Cahill et al. (1990) to explain observations of sulfate size changes during the
1986 Carbonaceous Species Methods Comparison Study in Glendora, CA. Cahill used a
DRUM sampler to measure sulfate in nine size ranges. By tracking the mass of sulfate in the
0.56 to 1.15 |im size range Cahill et al. could follow the expansion and contraction of aerosol
particles containing sulfate. Because of the relative high time resolution of the DRUM sampler
(4 h except for an 8-h increment each night from midnight to 8 a.m.),
3-181
-------
E
3.
0.56 /j,m. The
approximate trajectories followed during each day by the Dae>0.56 /u,m sulfur
size fraction are shown for period P and period F. Note that even when the
humidities are low, 30 to 50 %, the period P aerosols remain coarser by a
factor of three than those of period F. The water content incorporated in the
aerosols during the 0000- to 0800-h time periods is lost only slowly, giving a
strong hysteresis effect in sulfur size.
Source: Cahill et al. (1990).
Cahill et al. (1990) could follow this process as the relative humidity increased during the night
and decreased during the day. These data indicate that during the "Poor Period" (low visibility)
particles grow as relative humidity increases. However, they did not return to the smaller size
observed during the "Fair Period" (good visibility). This could be due to a combination of
growth due to reaction of SO2 to sulfate within the particles or failure of the droplet to crystallize
thus maintaining particle-bound water in a supersaturated state.
3-182
-------
John et al. (1990), in studies in the Los Angeles area, observed a number of sulfate size
distributions with MMAD near 1.0 jim. A histogram of the sulfate MMADs from his study is
shown in Figure 3-34. John et al. (1990) have provided a qualitative explanation to account for
these large MMADs for fine mode aerosol. In analyzing their data John et al. plotted sulfate
mass as a function of sulfate MMAD and found two distinct regions, as shown in Figure 3-35.
Distributions with particles near 0.2 jim diameter are probably still dry; the particles have not
reached their deliquescent point. As the relative humidity increases they reach their
deliquescent point and grow rapidly into the 0.5 to 0.7 //m size range. During the formation of
fog, the hygroscopic particles act as fog condensation nuclei, and with relative humidity at
100%, grow into 1 to 10 //m fog droplets. Sulfur dioxide dissolves in the fog droplets and is
rapidly oxidized to sulfate by atmospheric oxidants such as H2O2 or O3, or by catalysis by Fe or
Mn. These particles lose some of their water as the relative humidity decreases below 100%
RH, but will have substantially more sulfate than prior to activation. Similar processes occur in
clouds (Schwartz, 1984a, 1986a).
This type of process probably accounts for the large size of the fine mode observed in
Vienna (Berner et al., 1979; Berner and Liirzer, 1980). Winter and summer size distributions are
shown in Figure 3-36. Berner et al. reported that fog occurred during the night time during the
winter study. In this European study, as in American studies, instances of fine mode size
distributions with MMADs near or above 1 |im seem to occur only when fog or very high
relative humidity conditions have been present. Two log-normal distributions are fit to the
accumulation mode to suggest the separation, at high relative humidity, into hygroscopic and
hydrophobic components. No distribution was fit to the coarse mode because only a fraction of
the coarse size range was measured.
Similar results have been observed in sampling with dichotomous samplers. A large
humidity driven shift of normally fine mode material into the coarse mode was observed by
Keeler et al. (1988). In the extreme case, 60% of the SO=4 and 50% of the PM2 5 mass was
shifted to the coarse fraction. Such occurrences were not rare, occurring in 12 out of 83 several-
hour sampling periods.
In an analysis of data from the IMPROVE network Cahill and co-workers (Eldred et al.,
1994) report that 20% of the total sulfate is found in the coarse fraction of PM10. Studies in
Philadelphia using dichotomous samplers have also reported that 20% of the total
3-183
-------
>
o
35
30
25
20
15
10
5
0
Summer All Sites SQ =
(a)
0.1 1 10
Aerodynamic Mode Diameter (|jm)
o
o
tt
•o
o
400
300
200
a 100
Summer All Sites SQ =
(b)
0.1 1
Aerodynamic Mode Diameter (urn)
10
Figure 3-34. Data from the South Coast Air Quality Study (John et al., 1990). Plots show
(a) frequency of sulfate modes of various sizes as a function of mode
diameter and (b) average sulfate mode concentration as a function of mode
diameter. Note that although there are only a few instances when the mode
diameter is near 1.0 /j,m, it is these situations that give rise to the highest
sulfate concentrations. Modes with diameters above 2.5 //m may be due to
collection of fog droplets containing sulfate or reaction of SO2 in liquid
droplets of NaCl due to NaCl sea spray droplets in which SO2 has dissolved
and reacted to form sulfate and release HC1 gas.
3-184
-------
1,000 -
o-
o>
o
3
a
o>
o
c
o
o
o>
•o
o
100 -
0.1 1
Aerodynamic Mode Diameter (|jm)
Figure 3-35. Log-log plot of sulfate mode concentration versus aerodynamic mode
diameter from Claremont, CA, during the summer SCAQS (John et al.,
1990). The solid lines have slopes corresponding to mode concentration
increasing with the cube of the mode diameter. A transition between the two
modes is believed to occur at approximately the sulfate mode concentration
indicated by the horizontal dashed line.
sulfate is found in the coarse fraction (Dzubay et al., 1988). Cahill and coworkers suggest that
sulfate particles may grow larger than 2.5 jim in diameter and thus be sampled in the PM10
fraction but not the PM2 5 fraction. It is possible for SO2 to react with basic carbonate coarse
particles to form a sulfate coating or to dissolve in wet NaCl particles, from oceans, lakes, or salt
placed on streets to dissolve ice, and be converted to sulfate with the release of HC1. However,
there also is substantial evidence that some fine sulfate, and therefore possibly other fine mode
material, may be found in the size range above 1.0 jam and even
3-185
-------
40.0
Vienna, Summer
10
Aerodynamic Diameter,
50.0
Vienna, Winter
10
Aerodynamic Diameter, R , |jm
Figure 3-36. Typical results of size-distribution measurements taken with a Berner
impactor in a Vienna street with heavy automotive traffic:
(a) measurements taken during summer at three different elevations, (b)
measurements taken during winter at three different elevations, fog was
frequently present during the winter sampling period.
Source: Berner and Lilrzer (1980).
above 2.5 jim diameter, due to the growth of hygroscopic particles at very high relative
humidity.
These observations, indicating that, during near 100% relative humidity conditions,
significant amounts of normally fine mode material will be found in the coarse fractions (>2.5
//m diameter), have broader implications than selection of a cut point to separate fine and coarse
3-186
-------
particles. Such shifts could cause problems for receptor modeling using chemical mass balance
or factor analysis, for interpretation of exposure data in epidemiological studies, and in estimated
removal of particulate matter by deposition.
3.7.7 Conclusions
This review of atmospheric particle-size-distributions was undertaken to provide
information which could be used to determine what cut-point; 1.0 //m, 2.5 //m, or something in
between; would give the best separation between the fine and coarse particle modes. The data
do not provide a clear or obvious answer. Depending on conditions, a significant amount of
either fine or coarse mode material may be found in the intermodal region between 1.0 and 3
//m. However, the analysis does demonstrate the important role of relative humidity in
influencing the size of the fine particle mode and indicates that significant fine mode material is
found above 1.0 //m only during periods of very high relative humidity.
Thus, a PM25 sample will contain most of the fine mode material, except during periods of
RH near 100 %. However, especially in conditions of low RH, it may contain 5 to 20 % of the
coarse mode material below 10 //m in diameter. A PMj 0 sample will prevent misclassification
of coarse mode material as fine but under high RH conditions will result in some of the fine
mode material being misclassified as coarse.
A reduction in RH, either intentionally or inadvertently, will reduce the size of the fine
mode. A sufficient reduction in RH will yield a dry fine particle mode with very little material
above 1.0 //m. However, reducing the RH by heating will result in loss of semivolatile
components such as ammonium nitrate and semivolatile organic compounds. No information
was found on techniques designed to remove particle-bound water without loss of other
semivolatile components.
3.8 SUMMARY
Atmospheric particulate matter (PM) refers to solid or liquid particles suspended in air.
The term atmospheric aerosol refers to both the suspended particles and the air (including
gaseous pollutants) in which the particles are suspended. However, the term aerosol is
3-187
-------
frequently used to refer only to the suspended particles. The terms particulate matter and
particles will be used most frequently in this document.
Particulate matter is not a single pollutant but rather a mixture of many classes of
pollutants. The components of PM differ in sources; formation mechanisms; composition; size;
and chemical, physical, and biological properties. Particle diameters span more than four orders
of magnitude, ranging from a few nanometers (nm) to one hundred micrometers (//m). Because
of this wide size range, plots of particle-size distribution are almost always plotted versus the
logarithm of the particle diameter. Diameter usually refers to the aerodynamic diameter, defined
as the diameter of a spherical particle with an equal settling velocity but a density of 1 g/cm3.
This normalizes particles of different shapes and densities.
One of the most fundamental divisions of atmospheric particles is the naturally occurring
separation into a fine particle mode and a coarse particle mode as shown in Figure 3-3. The
terms fine mode particles and coarse mode particles are used to refer to particles in the fine or
coarse particle distributions. The two distributions overlap between 1 and 3 //m aerodynamic
diameter.
Particles may also be defined by the size cut of the collection or measuring device. A
frequently used descriptor is the 50% cut point. This is the aerodynamic diameter at which the
efficiency of the device for particle collection is 50%. As particles increase in size above the
50% cut point, they are collected with decreasing efficiency, eventually reaching 0%; as
particles decrease in size below the 50% cut point, they are collected with increasing efficiency,
eventually reaching 100%. The indicator for the current particle standard is PM10 (i.e. particles
with a 50% cut point of 10 //m aerodynamic diameter). However, PM10 contains some particles
larger than 10 //m and does not contain all particles below 10 //m. Fine is also used to refer to
particles with an upper cut point of 3.5, 2.5 (PM25), 2.1, or 1.0 //m. Coarse is also used to refer
to particles between 2.5 and 10 //m (PM(10_2 5)) or particles collected by the high volume sampler
as well as the entire coarse mode.
Size fractions may also be characterized in terms of their entrance into various
compartments of the body. Thus, inhalable particles enter the respiratory tract, including the
head airways. Thoracic particles travel past the larynx and reach the lung airways and the gas-
exchange regions of the lung. Respirable particles reach the gas-exchange region of the lung.
3-188
-------
PM10 is an indicator of thoracic particles; PM25 is an indicator of fine mode particles; and PM(10.
25) is an indicator of the thoracic component of coarse mode particles.
The fine and coarse particle distributions are frequently approximated by log-normal
distributions. However, finer distinctions can be made. The fine particles consist of a nuclei
mode, composed of particles recently formed from gases, and an accumulation mode, into which
the nuclei grow and accumulate (Figure 3-6). Ultrafine particles, defined in this document as
distributions with mass median diameters below 0.1 //m, are associated with the nuclei mode
(Figures 3-1, 3-2, and 3-13). In the presence of fogs or clouds, the accumulation mode may split
into a smaller, less hygroscopic mode and a larger droplet mode. The latter is formed by gases
dissolving in the fog or cloud droplets, reacting, and forming particles when the water of the
droplets evaporates (Figure 3-14). There may also be several modes within the coarse particle
distribution or mode but these are usually less distinct.
The terms primary and secondary, anthropogenic and biogenic, outdoor and indoor
microenvironment have significant applications to particulate matter. Primary fine particles are
emitted from sources, either directly as particles or as vapors which rapidly condense to form
particles. Primary coarse particles are usually formed by mechanical processes. Secondary fine
particles are formed within the atmosphere as the result of gas-phase or aqueous-phase chemical
reactions. Anthropogenic particles may be formed by primary or secondary processes.
Similarly, biogenic particles include primary particles of biological origin, including
bioallergens, as well as secondary particles formed from biogenic precursors such as terpenes
emitted into the atmosphere. The term outdoor refers to community atmospheres. These are the
atmospheres which are usually monitored for particulate matter. Indoor microenviroments
include homes, apartments, schools, office buildings and other indoor work places, large
enclosed areas such as malls, vehicles used for commuting, etc.
Some general classes of particles, such as organic particles, can occur not only as fine or
coarse particles, but can be of either anthropogenic and biogenic origin, and can be produced
both in outdoor and indoor microenvironments. Organic particles also can be present in air as
primary fine particles from combustion processes or as secondary fine particles formed as a
result of atmospheric reactions involving higher molecular weight volatile anthropogenic
3-189
-------
alkenes and aromatics or from the atmospheric reactions of volatile biogenic compounds such as
terpenes. Therefore, there is considerable overlap for chemical species among the categories
listed above.
A substantial fraction of the fine particle mass, especially during the warmer months of the
year, is secondary PM, formed as a result of atmospheric reactions. Such reactions involve the
gas phase conversion of SO2 to H2SO4 by OH radicals and aqueous-phase reactions of SO2 with
H2O2, O3, or O2 (catalyzed by Fe and Mn). The NO2 portion of NOX can be converted to HNO3
by reaction with OH radicals during the day. During nighttime NO2 is converted into HNO3 by a
series of reactions involving O3 and the nitrate radical (NO3). Both H2SO4 and HNO3 react with
atmospheric ammonia (NH3). Gaseous NH3 reacts with gaseous HNO3 to form paniculate
NH4NO3. Gaseous NH3 reacts with H2SO4 to form acidic HSO4 and neutral (NH4)2SO4. A
number of volatile organic compounds can react with O3 and/or OH radical to form fine organic
particles. In addition, acid gases such as SO2 and HNO3 may react with coarse particles such as
CaCO3 and NaCl to form coarse particles of different chemical composition.
The concentrations of OH radicals, O3, and H2O2, formed by gas phase reactions involving
volatile organic compounds and NOX, depend on the concentrations of the reactants, and on
meteorological conditions including temperature, solar radiation, wind speed, mixing volume
and passage of high pressure systems. Therefore, formation of a substantial fraction of fine
particles can depend on the gas phase reactions which also produce O3 and a variety of other
volatile products.
The fine particle fraction, in addition to SO4 and NO3, contains elemental carbon (EC),
organic carbon (OC), H+ (hydrogen ions or acidity) and a number of metal compounds at lower
concentrations. Species such as SO4 , NO3 and some organic species are associated with
substantial amounts of particle-bound water. NH4NO3 is in equilibrium with HNO3 and NH3 so
it can vaporize from particles. Organic particles can also be in equilibrium with their vapor.
Such species are called semi-volatile. A number of trace elements including, but not necessarily
limited to, Pb, Zn, Ni, Cd, Na, Cl, Br, Se and As have been measured in the PM2 5 fraction of
fine particles. The coarse particles are largely composed of the crustal elements Si, Ca, Al, and
Fe. However, a considerable number of elements are found in both the fine and coarse fractions.
Chemical reactions of SO2 and NOX within plumes are an important source of FT, SO4 and
NO3. These conversions can occur by gas-phase and aqueous-phase mechanisms.
3-190
-------
In point-source plumes emitting SO2 and NOX, the gas-phase chemistry depends on plume
dilution, sunlight and background volatile organic compounds mixed into the diluting plume.
For the conversion of SO2 to H2SO4, the gas-phase rate in such plumes during summer midday
conditions in the eastern United States typically varies between 1 and 3% h"1 but in the cleaner
western United States rarely exceeds 1% h"1. For the conversion of NOX to HNO3, the gas-phase
rates appear to be approximately three times faster than the SO2 conversion rates. Winter rates
for SO2 conversion were approximately an order of magnitude lower than the summer rates.
The contribution of aqueous-phase chemistry to particle formation in point-source plumes
is highly variable, depending on the availability of the aqueous phase (wetted aerosols, clouds,
fog, and light rain) and the photochemically generated gas-phase oxidizing agents, especially
H2O2 for SO2 chemistry. The in-cloud conversion rates of SO2 to SO4 can be several times
larger than the gas-phase rates given above. Overall, it appears that SO2 oxidation rates to SO4
by gas-phase and aqueous-phase mechanisms may be comparable in summer, but aqueous phase
chemistry may dominate in winter.
In the western United States, markedly higher SO2 conversion rates have been reported in
smelter plumes than in power plant plumes. The conversion is predominantly by a gas-phase
mechanism. This result is attributed to the lack of NOX in smelter plumes. In power plant
plumes NO2 depletes OH and competes with SO2 for OH.
In urban plumes, the upper limit for the gas-phase SO2 conversion rate appears to be about
5% h"1 under the more polluted conditions. For NO2, the rates appear to be approximately three
times faster than the SO2 conversion rates. Conversion rates of SO2 and NOX in background air
are comparable to the peak rates in diluted plumes. Neutralization of H2SO4 formed by SO2
conversion increases with plume age and background NH3 concentration. If the NH3
concentrations are more than sufficient to neutralize H2SO4 to (NH4)2SO4, the HNO3 formed
from NOX conversions may be converted to NH4NO3.
The lifetimes of particles vary with size. Coarse particles can settle rapidly from the
atmosphere within hours, and normally travel only short distances. However, when mixed high
into the atmosphere as in dust storms the smaller sized coarse mode particles may have longer
lives and travel distances. Nuclei mode particles rapidly grow into the accumulation mode.
However, the accumulation mode does not grow into the coarse mode. Accumulation-mode fine
particles are kept suspended by normal air motions and have very low deposition rates to
3-191
-------
surfaces. They can be transported thousands of km and remain in the atmosphere for a number
of days. Both accumulation-mode and nuclei-mode (or ultrafine) particles have the ability to
penetrate deep into the lungs. Dry deposition rates are expressed in terms of a deposition
velocity which varies as the particle size, reaching a minimum between 0.1 and 1.0 //m
aerodynamic diameter. Accumulation-mode particles are removed from the atmosphere
primarily by cloud processes. Fine particles, especially particles with a hygroscopic component,
grow as the relative humidity increases, serve as cloud condensation nuclei, and grow into cloud
droplets. If the cloud droplets grow large enough to form rain, the particles are removed in the
rain. Falling rain drops impact coarse particles and remove them. Ultrafine or nuclei mode
particles are small enough to diffuse to the falling drop and be removed. Falling rain drops,
however, are not effective in removing accumulation-mode particles.
There are many reasons for wanting to collect fine and coarse particles separately.
However, because fine-mode particles and coarse-mode particles overlap in the size range
between 1.0 and 3 //m diameter, it is not clear what 50% cut point will give the best separation.
A review of atmospheric particle-size-distribution data did not provide a clear or obvious
answer. Depending on conditions, a significant amount of either fine or coarse mode material
may be found in the intermodal region between 1.0 and 3 //m. However, the analysis of the
existing data did demonstrate the important role of relative humidity in influencing the size of
the fine particle mode and indicated that significant fine mode material is found above 1.0 //m
only during periods of very high relative humidity.
Thus, a PM25 sample will contain most of the fine mode material, except during periods of
RH near 100 %. However, especially in conditions of low RH, it may contain 5 to 20 % of the
coarse mode material below 10 //m in diameter. A PMj 0 sample will prevent misclassification
of coarse mode material as fine but under high RH conditions will result in some of the fine
mode material being misclassified as coarse.
A reduction in RH, either intentionally or inadvertently, will reduce the size of the fine
mode. A sufficient reduction in RH will yield a dry fine particle mode with very little material
above 1.0 //m. However, techniques to reduce the RH without loss of semivolatile components
such as ammonium nitrate and semivolatile organic compounds have not yet been developed.
3-192
-------
REFERENCES
Adamson, A. W. (1976) Physical chemistry of surfaces New York, NY: John Wiley & Sons, Inc.
Akhter, M. S.; Chughtai, A. R.; Smith, D. M. (1984) Reaction of hexane soot with NO2/N2O4. J. Phys. Chem. 88:
5334-5342.
Akhter, M. S.; Chughtai, A. R.; Smith, D. M. (1985) The structure of hexane soot I: spectroscopic studies. Appl.
Spectrosc. 39: 143-153.
Aleic-Juretic, A.; Cvitas, T.; Klasinc, L. (1990) Heterogeneous poly cyclic aromatic hydrocarbon degradation with
ozone on silica gel carrier. Environ. Sci. Technol 24: 62-66.
Alkezweeny, A.; Powell, D. (1977) Estimation of transformation rate of SO2 to SO4 from atmospheric concentration
data. Atmos. Environ. 11: 179-182.
Allen, D. T.; Palen, E. J.; Haimov, M. I.; Hering, S. V.; Young, J. R. (1994) Fourier transform infrared spectroscopy of
aerosol collected in a low pressure impactor (LPI / FTIR): method development and field calibration. Aerosol
Sci. Technol. 21: 325-342.
Allwine, K. J. (1993) Atmospheric dispersion and tracer ventilation in a deep mountain valley. J. Appl. Meteorol. 32:
1017-1037.
Alofs, D. J.; Hagen, D. E.; Trueblood, M. B. (1989) Measured spectra of the hygroscopic fraction of atmospheric
aerosol particles. J. Appl. Meteorol. 28: 126-136.
Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the United States.
Atmos. Environ. 17: 2131-2165.
Altshuller, A. P. (1987) Potential contribution of sulfate production in cumulus cloud droplets to ground level particle
sulfur concentrations. Atmos. Environ. 21: 1097-1105.
American Conference of Governmental Industrial Hygienists (ACGIH). (1994) Appendix D: particle size-selective
sampling criteria for airborne particulate matter. In: 1994-1995 threshold limit values for chemical substances
and physical agents and biological exposure indices. Cincinnati, OH: American Conference of Governmental
Industrial Hygienists; pp. 43-46.
Anderson, J. R.; Aggett, F. J.; Buseck, P. R.; Germani, M. S.; Shattuck, T. W. (1988) Chemistry of individual aerosol
particles from Chandler, Arizona, an arid urban environment. Environ. Sci. Technol. 22: 811-818.
Andreae, M. O.; Charlson, R. J.; Bruynseels, F.; Storms, H.; Van Grieken, R.; Maenhaut, W. (1986) Internal mixture of
sea salt, silicates, and excess sulfate in marine aerosols. Science (Washington, DC) 232: 1620-1623.
Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function of
relative humidity. Environ. Sci. Technol. 27: 857-865.
Appel, B. R.; Tokiwa, Y.; Hsu, J.; Kothny, E. L.; Hahn, E. (1985) Visibility as related to atmospheric aerosol
constituents. Atmos. Environ. 19: 1525-1534.
Appel, B. R.; Cheng, W.; Salaymeh, F. (1989) Sampling of carbonaceous particles in the atmosphere—II. Atmos.
Environ. 23: 2167-2175.
Arimoto, R.; Duce, R. A.; Ray, B. J.; Unni, C. K. (1985) Atmospheric trace elements at Enewetak Atoll: 2. transport to
the ocean by wet and dry deposition. J. Geophys. Res. [Atmos.] 90: 2391-2408.
3-193
-------
Arnts, R. R.; Gay, B. W., Jr. (1979) Photochemistry of some naturally emitted hydrocarbons. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; EPA report no.
EPA-600/3-79-081. Available from: NTIS, Springfield, VA; PB80-131980.
Atkinson, R.; Aschmann, S. M. (1987) Kinetics of the gas-phase reactions of alkylnaphthalenes with O3, N2O5 and OH
radicals at 298±2 K. Atmos. Environ. 21: 2323-2326.
Avissar, R.; Pielke, R. A. (1989) A parameterization of heterogeneous land surfaces for atmospheric numerical models
and its impact on regional meteorology. Mon. Weather Rev. 117: 2113-2136.
Ayers, G. P.; Larson, T. V. (1990) Numerical study of droplet size dependent chemistry in oceanic, wintertime stratus
clouds at southern mid-latitudes. J. Atmos. Chem. 11: 143-167.
Baek, S. O.; Field, R. A.; Goldstone, M. E.; Kirk, P. W.; Lester, J. N.; Perry, R. (1991) A review of atmosphenc
polycyclic aromatic hydrocarbons: sources, fate and behavior. Water Air Soil Pollut. 60: 279-300.
Bagnold, R. A. (1941) The physics of blown sand and desert dunes. London, United Kingdom: Methuen & Co.
Baker, M. B.; Latham, J. (1979) The evolution of droplet spectra and the rate of production of embryonic raindrops in
small cumulus clouds. J. Atmos. Sci. 36: 1612-1615.
Baldwin, A. C. (1982) Heterogeneous reactions of sulfur dioxide with carbonaceous particles. Int. J. Chem. Kinet. 14:
269-277.
Barnard, W. R.; Gatz, D. F.; Stensland, G. J. (1987) Chemical characterization of aerosols emitted from vehicle traffic
on unpaved roads. Presented at: 80th annual meeting of the Air Pollution Control Association; June; New
York, NY. Pittsburgh, PA: Air Pollution Control Association; paper no. 87-14.3.
Barnard, W. R.; Stansland, G. J.; Gatz, D. F. (1988) Flux of alkaline materials from unpaved roads in the southwestern
United States. In: Particulate matter/fugitive dusts: measurement and control in western arid regions.
Proceedings of an Air Pollution Control Association meeting. Pittsburgh, PA: Air Pollution Control
Association; pp. 27-39.
Barrie, L. A. (1985) Scavenging ratios, wet deposition, and in-cloud oxidation: an application to the oxides of sulphur
and nitrogen. J. Geophys. Res. [Atmos.] 90: 5789-5799.
Barrie, L. A.; Georgii, H. W. (1976) An experimental investigation of the absorption of sulphur dioxide by water drops
containing heavy metal ions. Atmos. Environ. 10: 743-749.
Barth, M. C. (1994) Numerical modeling of sulfur and nitrogen chemistry in a narrow cold-frontal rainband: the impact
of meteorological and chemical parameters. J. Appl. Meteorol. 33: 855-868.
Barth, M. C.; Hegg, D. A.; Hobbs, P. V. (1992) Numerical modeling of cloud and precipitation chemistry associated
with two rainbands and some comparisons with observations. J. Geophys. Res. [Atmos.] 97: 5825-5845.
Bassett, M.; Seinfeld, J. H. (1983) Atmospheric equilibrium model of sulfate and nitrate aerosols. Atmos. Environ. 17:
2237-2252.
Bassett, M. E.; Seinfeld, J. H. (1984) Atmospheric equilibrium model of sulfate and nitrate aerosols - II. particle size
analysis. Atmos. Environ. 18: 1163-1170.
Behymer, T. D.; Hites, R. A. (1988) Photolysis of polycyclic aromatic hydrocarbons adsorbed on fly ash. Environ. Sci.
Technol. 22: 1311-1319.
3-194
-------
Benkovitz, C. M; Berkowitz, C. M; Easter, R. C.; Nemesure, S.; Wagener, R.; Schwartz, S. E. (1994) Sulfate over the
North Atlantic and adjacent continental regions: evaluation for October and November 1986 using a
three-dimensional model driven by observation-derived meteorology. J. Geophys. Res. [Atmos.] 99:
20,725-20,756.
Benner, C. L.; Eatough, D. J.; Eatough, N L.; Bhardwaja, P. (1991) Comparison of annular denuder and filter pack
collection of HNO3 (g), HNO2 (g), SO2 (g), and particulate-phase nitrate, nitrite and sulfate in the south-west
desert. Atmos. Environ. Part A 25: 1537-1545.
Bergstrom, R. W.; Seigneur, C.; Babson, B. L.; Holman, H.-Y.; Wojcik, M. A. (1981) Comparison of the observed and
predicted visual effects caused by power plant plumes. Atmos. Environ. 15: 2135-2150.
Bemer, A. (1989) Haze and its relation to atmospheric accumulation aerosol. Sci. Total Environ. 86: 251-263.
Bemer, A.; Lilrzer, C. (1980) Mass size distributions of traffic aerosols at Vienna. J. Phys. Chem. 84: 2079-2083.
Bemer, A.; Lilrzer, Ch.; Pohl, F.; Preining, O.; Wagner, P. (1979) The size distribution of the urban aerosol in Vienna.
Sci. Total Environ. 13: 245-261.
Bernstein, D. M.; Rahn, K. A. (1979) New York Summer Aerosol Study: trace element concentrations as a function of
particle size. In: Kneip, T. J.; Lippmann, M., eds. The New York Summer Aerosol Study, 1976. Ann. N. Y.
Acad. Sci. 322: 87-97.
Bidleman, T. F. (1988) Atmospheric processes. Environ. Sci. Technol. 22: 361-367.
Biggins, P. D. E.; Harrison, R. M. (1980) Chemical speciation of lead compounds in street dusts. Environ. Sci. Technol.
14: 336-339.
Bj0rseth, A.; Olufsen, B. S. (1983) Long-range transport of polycyclic aromatic hydrocarbons. In: Bj0rseth, A., ed.
Handbook of polycyclic aromatic hydrocarbons. New York, NY: Marcel Dekker, Inc.; pp. 507-524.
Bj0rseth, A.; Lunde, G.; Lindskog, A. (1979) Long-range transport of polycyclic aromatic hydrocarbons. Atmos.
Environ. 13: 45-53.
Bott, A. (1991) On the influence of the physico-chemical properties of aerosols on the life cycle of radiation fogs.
Boundary Layer Meteorol. 56: 1-31.
Bott, A.; Carmichael, G. R. (1993) Multiphase chemistry in a microphysical radiation fog model—a numerical study.
Atmos. Environ. Part A 27: 503-522.
Bower, K. N.; Choularton, T. W. (1993) Cloud processing of the cloud condensation nucleus spectrum and its
climatological consequences. Q. J. R. Meteorol. Soc. 119: 655-679.
Bower, K. N.; Hill, T. A.; Coe, H.; Choularton, T. W. (1991) SO2 oxidation in an entraining cloud model with explicit
microphysics. Atmos. Environ. Part A 25: 2401-2418.
Boyce, S. D.; Hoffmann, M. R. (1984) Kinetics and mechanism of the formation of hydroxymethanesulfonic acid at
lowpH. J. Phys. Chem. 88: 4740-4746.
Braaten, D. A.; Paw U, K. T. (1992) A stochastic particle resuspension and deposition model. In: Schwartz, S. E.; Slinn,
W. G. N., eds. Precipitation scavenging and atmosphere-surface exchange: proceedings of the fifth
international conference, v. 2: the Semonin volume: atmosphere-surface exchange processes; July 1991;
Richland, WA. Washington, DC: Hemisphere Publishing Corporation; pp. 1143-1152.
Braaten, D. A.; Paw U, K. T.; Shaw, R. H. (1989) Particles resuspension in a turbulent boundary layer—observed and
modeled. J. Aerosol Sci. 21: 613-628.
3-195
-------
Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1991) Indoor and outdoor concentrations of inorganic acidic
aerosols and gases. J. Air Waste Manage. Assoc. 41: 171-181.
Briggs, G. A.; Binkowski, F. S. (1985) Research on diffusion in atmospheric boundary layers: a position paper on
status and needs. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Sciences
Research Laboratory; EPA report no. EPA/600/3-85/072. Available from: NTIS, Springfield, VA;
PB86-122587.
Brosset, C.; Andreasson, K.; Perm, M. (1975) The nature and possible origin of acid particles observed at the Swedish
west coast. Atmos. Environ. 9: 631-642.
Brown, N. J.; Dod, R. L.; Mowrer, F. W.; Novakov, T.; Williamson, R. B. (1989) Smoke emission factors from
medium-scale fires: part 1. Aerosol Sci. Technol. 10: 2-19.
Buat-Menard, P.; Duce, R. A. (1986) Precipitation scavenging of aerosol particles over remote marine regions. Nature
(London) 321:508-510.
Burkhard, E. G.; Dutkiewicz, V. A.; Husain, L. (1994) A study of SO^ SO\ and trace elements in clear air and clouds
above the midwestern United States. Atmos. Environ. 28: 1521-1533.
Burton, R. M.; Lundgren, D. A. (1987) Wide range aerosol classifier: a size selective sampler for large particles.
Aerosol Sci. Technol. 6: 289-301.
Burtscher, H. (1992) Measurement and characteristics of combustion aerosols with special consideration of
photoelectric charging and charging by flame ions. J. Aerosol Sci. 23: 549-595.
Burtscher, H.; Leonardi, A.; Steiner, D.; Baltensperger, U.; Weber, A. (1993) Aging of combustion particles in the
atmosphere - results from a field study in Zurich. Water Air Soil Pollut. 68: 137-147.
Cadle, S. H.; Dasch, J. M. (1988) Wintertime concentrations and sinks of atmospheric particulate carbon at a rural
location in northern Michigan. Atmos. Environ. 22: 1373-1381.
Cahill, T. A.; Surovik, M.; Wittmeyer, I. (1990) Visibility and aerosols during the 1986 Carbonaceous Species Methods
Comparison Study. Aerosol Sci. Technol. 12: 149-160.
Calvert, J. G.; Su, F.; Bottenheim, J. W.; Strausz, O. P. (1978) Mechanism of the homogeneous oxidation of sulfur
dioxide in the troposphere. In: Sulfur in the atmosphere: proceedings of the international symposium;
September 1977; Dubrovnik, Yugoslavia. Atmos. Environ. 12: 197-226.
Calvert, J. G.; Lazrus, A.; Kok, G. L.; Heikes, B. G.; Walega, J. G.; Lmd, J.; Cantrell, C. A. (1985) Chemical
mechanisms of acid generation in the troposphere. Nature (London) 317: 27-35.
Cambray, R. S. (1989) Radioactive fallout in air and rain: results to the end of 1987. Harwell, United Kingdom: Atomic
Energy Establishment; report no. AERE-R 13226.
Cambray, R. S.; Cawse, P. A.; Garland, J. A.; Gibson, J. A. B.; Johnson, P.; Lewis, G. N. J.; Newton, D.; Salmon, L.;
Wade, B. O. (1987) Observations on radioactivity from the Chernobyl accident. Nucl. Energy (Br. Nucl.
Energy Soc.) 26: 77-101.
Carpenter, S. B.; Montgomery, T. L.; Leavitt, J. M.; Colbaugh, W. C.; Thomas, F. W. (1971) Principal plume dispersion
models: TVA power plants. J. Air Pollut. Control Assoc. 21: 491-495.
Carras, J. N.; Williams, D. J. (1981) The long-range dispersion of a plume from an isolated point source. Atmos.
Environ. 15: 2205-2217.
3-196
-------
Carras, J. N.; Williams, D. J. (1988) Measurements of relative oy up to 1800 km from a single source. Atmos. Environ.
22: 1061-1069.
Carter, E. J.; Borys, R. D. (1993) Aerosol-cloud chemical fractionation: enrichment factor analysis of cloud water. J.
Atmos. Chem. 17: 277-292.
Cass, G. R. (1979) On the relationship between sulfate air quality and visibility with examples in Los Angeles. Atmos.
Environ. 13: 1069-1084.
Cass, G. R.; Shair, F. H. (1984) Sulfate accumulation in a sea breeze/land breeze circulation system. J. Geophys. Res.
[Atmos.] 89: 1429-1438.
Cass, G. R.; Boone, P. M.; Macias, E. S. (1982) Emissions and air quality relationships for atmospheric carbon particles
in Los Angeles. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric life cycle. New York,
NY: Plenum Press; pp. 207-243.
Cawse, P. A. (1981) Trace elements in the atmosphere of the United Kingdom. Presented at: ESNA meeting;
September-October; Aberdeen, Scotland.
Chamberlain, A. C. (1983) Deposition and resuspension. In: Pruppacher, H. R.; Semonin, R. G.; Slinn, W. G. N, eds.
Precipitation scavenging, dry deposition, and resuspension, v. 2: dry deposition and resuspension: proceedings
of the fourth international conference; November-December 1982; Santa Monica, CA. New York, NY:
Elsevier; pp. 731-751.
Chameides, W. L. (1984) The photochemistry of a remote marine stratiform cloud. J. Geophys. Res. [Atmos.] 89:
4739-4755.
Chameides, W. L.; Davis, D. D. (1982) The free radical chemistry of cloud droplets and its impact upon the
composition of rain. J. Geophys. Res. C: Oceans Atmos. 87: 4863-4877.
Chan, W. H.; Tang A. J. S.; Chung, D. H. S.; Lusis, M. A. (1986) Concentration and deposition of trace metals in
Ontario - 1982. Water Air Soil Pollut. 29: 373-389.
Chan, C. K.; Flagan, R. C.; Seinfeld, J. H. (1992) Water activities of NH4NO3/(NH4)2SO4 solutions. Atmos. Environ.
Part A 26: 1661-1673.
Chang, S. G.; Brodzinsky, R.; Gundel, L. A.; Novakov, T. (1982) Chemical and catalytic properties of elemental
carbon. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric life cycle. New York, NY:
Plenum Press; pp. 158-181.
3-197
-------
Chang, J. S.; Middleton, P. B.; Stockwell, W. R.; Walcek, C. J.; Pleim, J. E.; Lansford, H. H.; Madromch, S.;
Binkowski, F. S.; Seaman, N. L.; Stauffer, D. R. (1991) The regional acid deposition model and engineering
model. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
Program. (State of science and technology report no. 4).
Chapman, E. G.; Luecken, D. J.; Dana, M. T.; Easter, R. C.; Hales, J. M.; Laulainen, N. S.; Thorp, J. M. (1987)
Inter-storm comparisons from the OSCAR high-density network experiment. Atmos. Environ. 21: 531 -549.
Cheng, Y.-S.; Carpenter, R. L.; Barr, E. B.; Hobbs, C. H. (1985) Size distribution of fine particle emissions from a
steam plant with a fluidized-bed coal combustor. Aerosol Sci. Technol. 4: 175-189.
Ching, J. K. S.; Alkezweeny, A. J. (1986) Tracer study of vertical exchange by cumulus clouds. J. Clim. Appl.
Meteorol. 25: 1702-1711.
Choularton, T. W.; Wicks, A. J.; Downer, R. M.; Gallagher, M. W.; Penkett, S. A.; Bandy, B. J.; Dollard, G. J.; Jones,
B. M. R.; Davies, T. D.; Gay, M. J.; Tyler, B. J.; Fowler, D.; Cape, J. N.; Hargreaves, K. J. (1992) A field study
of the generation of nitrate in a hill cap cloud. Environ. Pollut. 75: 69-73.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W. (1992)
PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A 26: 3335-3354.
Chow, J. C.; Watson, J. G.; Pntchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993a) The DRI
thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere; April
1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W. (1993b)
PM10 and PM25 compositions in California's San Joaquin Valley. Aerosol Sci. Technol. 18: 105-128.
Chow, J. C.; Watson, J. G.; Fujita, E. M.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994) Temporal and spatial
variations of PM25 and PM10 aerosol in the Southern California Air Quality Study. Atmos. Environ. 28:
2061-2080.
Chughtai, A. R.; Jassim, J. A.; Peterson, J. H.; Stedman, D. H.; Smith, D. M. (1991) Spectroscopic and solubility
characteristics of oxidized soots. Aerosol Sci. Technol. 15: 112-126.
Clark, T. L.; Cohn, R. D. (1990) The Across North America Tracer Experiment (ANATEX) model evaluation study.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Assessment Laboratory; EPA report no. EPA/600/3-90/051. Available from: NTIS, Springfield, VA;
PB90-261454.
Clark, P. A.; Fletcher, I. S.; Kallend, A. S.; McElroy, W. J.; Marsh, A. R. W.; Webb, A. H. (1984) Observations of
cloud chemistry during long-range transport of power plant plumes. Atmos. Environ. 18: 1849-1858.
Clarke, A. D. (1989) Aerosol light absorption by soot in remote environments. Aerosol Sci. Technol. 10: 161-171.
Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14: 479-488.
Clarke, A. D. (1993) Atmospheric nuclei in the Pacific midtroposphere—their nature, concentration, and evolution. J.
Geophys. Res. [Atmos.] 98: 20,633-20,647.
Clarke, A. G.; Karani, G. N. (1992) Characterisation of the carbonate content of atmospheric aerosols. J. Atmos. Chem.
14: 119-128.
3-198
-------
Clarke, A. G.; Radojevic, M. (1987) Oxidation of SO2 in rainwater and its role in acid rain chemistry. Atmos. Environ.
21: 1115-1123.
Clarke, J. F.; Clark, T. L.; Ching, J. K. S.; Haagenson, P. L.; Husar, R. B.; Patterson, D. E. (1983) Assessment of model
simulation of long-distance transport. Atmos. Environ. 17: 2449-2462.
Clarke, A. D.; Weiss, R. E.; Charlson, R. J. (1984) Elemental carbon aerosols in the urban, rural and remote-marine
troposphere and in the stratosphere: interferences from light absorption data and consequences regarding
radiative transfer. Sci. Total Environ. 36: 97-102.
Cobourn, W. G.; Husar, R. B.; Husar, J. D. (1978) Continuous in situ monitoring of ambient particulate sulfur using
flame photometry and thermal analysis. In: Husar, R. B.; Lodge, J. P., Jr.; Moore, D. J., eds. Sulfur in the
atmosphere: proceedings of the international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos.
Environ. 12: 89-98.
Cocks, A. T.; McElroy, W. L.; Wallis, P. G. (1982) The oxidation of sodium sulphite solutions by hydrogen peroxide.
Central Electricity Research Laboratories; report no. RD/L/2215N81.
Cofer, W. R., Ill; Stevens, R. K.; Wmstead, E. L.; Pinto, J. P.; Sebacher, D. L; Abdulraheem, M. Y.; Al-Sahafi, M.;
Mazurek, M. A.; Rasmussen, R. A.; Cahoon, D. R.; Levine, J. S. (1992) Kuwaiti oil fires: compositions of
source smoke. J. Geophys. Res. [Atmos.] 97: 14,521-14,525.
Cohen, M. D.; Flagan, R. C.; Seinfeld, J. H. (1987a) Studies of concentrated electrolyte solutions using the
electrodynamic balance. 1. Water activities for single-electrolyte solutions. J. Phys. Chem. 91: 4563-4574.
Cohen, M. D.; Flagan, R. C.; Seinfeld, J. H. (1987b) Studies of concentrated electrolyte solutions using the
electrodynamic balance. 2. Water activities for mixed-electrolyte solutions. J. Phys. Chem. 91: 4575-4582.
Colbeck, I.; Appleby, L.; Hardman, E. J.; Harrison, R. M. (1990) The optical properties and morphology of
cloud-processed carbonaceous smoke. J. Aerosol Sci. 21: 527-538.
Collett, J. L., Jr.; Oberholzer, B.; Mosimann, L.; Staehelin, J.; Waldvogel, A. (1993a) Contributions of cloud processes
to precipitation chemistry in mixed phase clouds. Water Air Soil Pollut. 68: 43-57.
Collett, J., Jr.; Oberholzer, B.; Staehelin, J. (1993b) Cloud chemistry in Mt. Rigi Switzerland: dependence on drop size
and relationship to precipitation chemistry. Atmos. Environ. Part A 27: 33-42.
Colvile, R. N.; Choularton, T. W.; Gallagher, M. W.; Wicks, A. J.; Downer, R. M.; Tyler, B. J.; Storeton-West, K. J.;
Fowler, D.; Cape, J. N.; Dollard, G. J.; Davies, T. J.; Jones, B. M. R.; Penkett, S. A.; Bandy, B. J.; Burgess, R.
A. (1994) Observation on Great Dun Fell of the pathways by which oxides of nitrogen are converted to nitrate.
Atmos. Environ. 28: 397-408.
Countess, R. J.; Wolff, G. T.; Cadle, S. H. (1980) The Denver winter aerosol: a comprehensive chemical
characterization. J. Air Pollut. Control Assoc. 30: 1194-1200.
Coutant, R. W.; Brown, L.; Chuang, J. C.; Riggin, R. M.; Lewis, R. G. (1988) Phase distribution and artifact formation
in ambient air sampling for polynuclear aromatic hydrocarbons. Atmos. Environ. 22: 403-409.
Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease-coated denuders for
collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
Covert, D. S.; Heintzenberg, J. (1984) Measurement of the degree of internal/external mixing of hygroscopic
compounds and soot in atmospheric aerosols. Sci. Total Environ. 36: 347-352.
3-199
-------
Covert, D. S.; Charlson, R. J.; Ahlquist, N. C. (1972) A study of the relationship of chemical composition and humidity
to light scattering by aerosols. J. Appl. Meteorol. 11: 968-976.
Covert, D. S.; Heintzenberg, J.; Hansson, H.-C. (1990) Electro-optical detection of external mixtures in aerosols.
Aerosol Sci. Technol. 12: 446-456.
Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary layer.
J. Geophys. Res. [Atmos.] 97: 20,581-20,589.
Cronn, D. R.; Charlson, R. J.; Knights, R. L.; Crittenden, A. L.; Appel, B. R. (1977) 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.
Cume, L. A.; Stafford, T. W.; Sheffield, A. E.; Klouda, G. A.; Wise, S. A.; Fletcher, R. A.; Donahue, D. J.; Ml, A. J.
T.; Linick, T. W. (1989) Microchemical and molecular dating. Radiocarbon 31: 448-463.
Cume, L. A.; Sheffield, A. E.; Riederer, C.; Gordon, G. E. (1994) Improved atmospheric understanding through
exploratory data analysis and complementary modeling: the urban K-Pb-C system. Atmos. Environ. 28:
1359-1369.
Daisey, J. M. (1987) Chemical composition of inhalable particulate matter—seasonal and intersite comparisons. In:
Lioy, P. J.; Daisey, J. M., eds. Toxic air pollution: a comprehensive study of non-criteria air pollutants.
Chelsea, MI: Lewis Publishers, Inc.; pp. 45-65.
DAlmeida, G. A. (1989) Desert aerosol: characteristics and effects on climate. In: Leinen, M.; Samthein, M.
Paleoclimatology and paleometeorology: modern and past patterns of global atmospheric transport. Boston,
MA: Kluwer Academic Publishers; pp. 311-338. (NATO ASI series. Series C, mathematical and physical
sciences: v. 282).
Dana, M. T.; Drewes, D. R.; Glover, D. W.; Hales, J. M. (1976) Precipitation scavenging of fossil-fuel effluents.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences Research
Laboratory; report no. EPA-600/4-76-031. Available from: NTIS, Springfield, VA; PB-256649.
Danielsen, E. F. (1961) Trajectories: isobaric, isentropic and actual. J. Meteorol. 18: 479-486.
Dasch, J. M.; Cadle, S. H. (1989) Atmospheric carbon particles in the Detroit urban area: wintertime sources and sinks.
Aerosol Sci. Technol. 10: 236-248.
Daum, P. H. (1988) Processes determining cloudwater composition: inferences from field measurements. In: Unsworth,
M. H.; Fowler, D., eds. Acid deposition at high elevation sites. Dordrecht, The Netherlands: Kluwer; pp.
139-153.
Daum, P. H.; Kelly, T. J.; Schwartz, S. E.; Newman, L. (1984a) Measurements of the chemical composition of
stratiform clouds. Atmos. Environ. 18: 2671-2684.
Daum, P. H.; Schwartz, S. E.; Newman, L. (1984b) Acidic and related constituents in liquid water stratiform clouds. J.
Geophys. Res. [Atmos.] 89: 1447-1458.
Daum, P. H.; Kelly, T. J.; Strapp, J. W.; Leaitch, W. R.; Joe, P.; Schemenauer, R. S.; Isaac, G. A.; Anlauf, K. G.; Wiebe,
H. A. (1987) Chemistry and physics of a winter stratus cloud layer: a case study. J. Geophys. Res. [Atmos.] 92:
8426-8436.
Daum, P. H.; Kleinman, L. L; Hills, A. J.; Lazrus, A. L.; Leslie, A. C. D.; Busness, K.; Boatman, J. (1990) Measurement
and interpretation of concentrations of H2O2 and related species in the upper midwest during summer. J.
Geophys. Res. [Atmos.] 95: 9857-9871.
3-200
-------
Daum, P. H.; Al-Sunaid, A.; Busness, K. M.; Hales, J. M.; Mazurek, M. (1993) Studies of the Kuwait oil fire plume
dunng midsummer 1991. J. Geophys. Res. [Atmos.] 98: 16,809-16,827.
Daumer, B.; Niessner, R.; Klockow, D. (1992) Laboratory studies of the influence of thin organic films on the
neutralization reaction of HjSO,, aerosol with ammonia. J. Aerosol Sci. 23: 315-325.
Davidson, C. I.; Wu, Y.-L. (1990) Dry deposition of particles and vapors. In: Lindberg, S. E.; Page, A. L.; Norton, S.
A., eds. Acidic precipitation: v. 3, sources, deposition, and canopy interactions. New York, NY:
Springer-Verlag; pp. 103-216.
De Bock, L. A.; Van Malderen, H.; Van Grieken, R. E. (1994) Individual aerosol particle composition variations in air
masses crossing the North Sea. Environ. Sci. Technol. 28: 1513-1520.
Delumyea, R. D.; Kalivretenos, A. (1987) Elemental carbon and lead content of fine particles from American and
French cities of comparable size and industry, 1985. Atmos. Environ. 21: 1643-1647.
Dennis, R. L.; McHenry, J. N; Barchet, W. R.; Binkowski, F. S.; Byun, D. W. (1993) Correcting RADM's sulfate
underprediction: discovery and correction of model errors and testing the corrections through comparisons
against field data. Atmos. Environ. Part A 27: 975-997.
De Raat, W. K.; Bakker, G. L.; de Meijere, F. A. (1990) Comparison of filter materials used for sampling of mutagens
and polycyclic aromatic hydrocarbons in ambient airborne particles. Atmos. Environ. Part A 24: 2875-2887.
De Santis, F.; Allegrini, I. (1992) Heterogeneous reactions of SO2 and NO2 on carbonaceous surfaces. Atmos. Environ.
Part A 26: 3061-3064.
De Valk, J. P. J. M. M. (1994) A model for cloud chemistry: a comparison between model simulations and observations
in stratus and cumulus. Atmos. Environ. 28: 1665-1678.
Dickerson, R. R.; Huffman, G. J.; Luke, W. T.; Nunnermacker, L. J.; Pickering, K. E.; Leslie, A. C. D.; Lindsey, C. G.;
Slinn, W. G. N.; Kelly, T. J.; Daum, P. H.; Delany, A. C.; Greenberg, J. P.; Zimmerman, P. R.; Boatman, J. F.;
Ray, J. D.; Stedman, D. H. (1987) Thunderstorms: an important mechanism in the transport of air pollutants.
Science (Washington, DC) 235: 460-465.
Dittenhoefer, A. C.; de Pena, R. G. (1978) A study of production and growth of sulfate particles in plumes from a
coal-fired power plant. Atmos. Environ. 12: 297-306.
Dlugi, R. (1989) Chemistry and deposition of soot particles in moist air and fog. Aerosol Sci. Technol. 10: 93-105.
Dlugi, R.; Glisten, H. (1983) The catalytic and photocatalytic activity of coal fly ashes. Atmos. Environ. 17: 1765-1771.
Dod, R. L.; Brown, N. J.; Mowrer, F. W.; Novakov, T.; Williamson, R. B. (1989) Smoke emission factors from
medium-scale fires: part 2. Aerosol Sci. Technol. 10: 20-27.
Doyle, G. I; Tuazon, E. C.; Graham, R. A.; Mischke, T. M.; Winer, A. M.; Pitts, J. N., Jr. (1979) Simultaneous
concentrations of ammonia and nitric acid in a polluted atmosphere and their equilibrium relationship to
particulate ammonium nitrate. Environ. Sci. Technol. 13: 1416-1419.
Draxler, R. R. (1982) Measuring and modeling the transport and dispersion of Krypton-85 1500 km from a point
source. Atmos. Environ. 16: 2763-2776.
Draxler, R. R. (1984) Diffusion and transport experiments. In: Randerson, D., ed. Atmospheric science and power
production. Washington, DC: U.S. Department of Energy, Office of Scientific and Technical Information; pp.
367-422; report no. DOE/TIC-27601. Available from: NTIS, Springfield, VA; DE84-005177.
3-201
-------
Draxler, R. R.; Dietz, R.; Lagomarsino, R. I; Start, G. (1991) Across North America Tracer Experiment (ANATEX):
sampling and analysis. Atmos. Environ. Part A 25: 2815-2836.
Drewes, D. R.; Hales, J. M. (1982) SMICK—a scavenging model incorporating chemical kinetics. Atmos. Environ. 16:
1717-1724.
Duce, R. A.; Mohnen, V. A.; Zimmerman, P. R.; Grosjean, D.; Cautreels, W.; Chatfield, R.; Jaenicke, R.; Ogren, J. A.;
Pellizzari, E. D.; Wallace, G. T. (1983) Organic material in the global troposphere. Rev. Geophys. Space Phys.
21: 921-952.
Durham, J. L.; Wilson, W. E.; Ellestad, T. G.; Willeke, K.; Whitby, K. T. (1975) Comparison of volume and mass
distributions for Denver aerosols. Atmos. Environ. 9: 717-722.
Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, I.; Sheffield, A. E.; Courtney, W. J. (1988) A composite receptor
method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
Eatough, D. J.; Richter, B. E.; Eatough, N. L.; Hansen, L. D. (1981) Sulfur chemistry in smelter and power plant
plumes in the western U.S. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility:
measurements and model components: proceedings of the symposium; November 1980; Grand Canyon
National Park, AZ. Atmos. Environ. 15: 2241-2253.
Eatough, D. J.; Chnstensen, J. J.; Eatough, N. L.; Hill, M. W.; Major, T. D.; Mangelson, N. F.; Post, M .E.; Ryder, J. F.;
Hansen, L. D.; Meisenheimer, R. G.; Fischer, J. W. (1982) Sulfur chemistry in a copper smelter plume. Atmos.
Environ. 16: 1001-1015.
Eatough, D. J.; Arthur, R. J.; Eatough, N. L.; Hill, M. W.; Mangelson, N. F.; Richter, B. E.; Hansen, L. D.; Cooper, J. A.
(1984) Rapid conversion of SO2(g) to sulfate in a fog bank. Environ. Sci. Technol. 18: 855-859.
Eatough, D. J.; Sedar, B.; Lewis, L.; Hansen, L. D.; Lewis, E. A.; Farber, R. J. (1989) Determination of semivolatile
organic compounds in particles in the Grand Canyon area. Aerosol Sci. Technol. 10: 438-449.
Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993)
A multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
material in the atmosphere. Atmos. Environ. Part A 27: 1213-1219.
3-202
-------
Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical composition of
fine particulate semi-volatile organic material in urban environments using diffusion denuder technology. In:
Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution and human mortality
and morbidity, part II; Janaury 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
Eisele, F. L.; Bradshaw, J. D. (1993) The elusive hydroxyl radical: measuring OH in the atmosphere. Anal. Chem. 65:
927A-939A.
Eisele, F. L.; Tanner, D. J. (1993) Measurement of the gas phase concentration of H2SO4 and methane sulfonic acid and
estimates of F^SC^ production and loss in the atmosphere. J. Geophys. Res. [Atmos.] 98: 9001-9010.
Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ. 28:
2733-2749.
Eldred, R. A.; Cahill, T. A.; Flocchim, R. G. (1994) Comparison of PM10 and PM^ aerosols in the IMPROVE network.
In: Proceedings of an international specialty conference on aerosol and atmospheric optics: radiative balance
and visual air quality, volume A. Pittsburgh, PA: Air and Waste Management Association. Air Waste:
submitted.
Eltgroth, M. W.; Hobbs, P. V. (1979) Evolution of particles in the plumes of coal-fired power plants—II. A numerical
model and comparisons with field measurements. Atmos. Environ. 13: 953-975.
Er-El, J.; Peskin, R. L. (1981) Relative diffusion of constant-level balloons in the southern hemisphere. J. Atmos. Sci.
38: 2264-2274.
Erickson, R. E.; Yates, L. M.; Clark, R. L.; McEwen, D. (1977) The reaction of sulfur dioxide with ozone in water and
its possible atmospheric significance. Atmos. Environ. 11: 813-817.
Facchini, M. C.; Fuzzi, S.; Lind, J. A.; Fierlinger-Oberlinninger, H.; Kalina, M.; Puxbaum, H.; Winiwarter, W.; Arends,
E.G.; Wobrock, W.; Jaeschke, W.; Bemer, A.; Kruisz, C. (1992) Phase-partitioning and chemical reactions of
low molecular weight organic compounds in fog. Tellus Ser. B 44B: 533-544.
Falerios, M.; Schild, K.; Sheehan, P.; Paustenbach, D. (1992) Airborne concentrations of trivalent and hexavalent
chromium from contaminated soils at unpaved partially paved commercial/industrial sites. J. Air Waste
Manage. Assoc. 42: 40-48.
Ferber, G. J.; Telegadas, K.; Heffter, J. L.; Dickson, C. R.; Dietz, R. N.; Krey, P. W. (1981) Demonstration of a
long-range atmospheric tracer system using perfluorocarbons. Silver Springs, MD: National Oceanic and
Atmospheric Administration; NOAA report no. NOAA TM ERL ARL-101.
Ferber, G. J.; Heffter, J. L.; Draxler, R. R.; Lagomarsmo, R. J.; Thomas, F. L.; Dietz, R. N.; Benkovitz, C. M. (1986)
Cross-Appalachian Tracer Experiment (CAPTEX-83) final report. Silver Springs, MD: National Oceanic and
Atmospheric Administration; NOAA report no. NOAA TM ERL ARL-142.
Fernandez de la Mora, J. (1986) Inertia and interception in the deposition of particles from boundary layers. Aerosol
Sci. Technol. 5: 261-266.
Fernandez de la Mora, J.; Friedlander, S. K. (1982) Aerosol and gas deposition to fully rough surfaces: filtration model
for blade-shaped elements. Int. J. Heat Mass Transfer 25: 1725-1735.
Finlayson-Pitts, B. J.; Pitts, J. N., Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques. New
York, NY: John Wiley & Sons; pp. 783-1007.
3-203
-------
Fitzgerald, J. W. (1973) Dependence of the supersaturation spectrum of CCN on aerosol size distribution and
composition. Presented at: the second international workshop on condensation and ice nuclei; August 1970;
Fort Collins, CO. J. Atmos. Sci. 30: 628-634.
Fitzgerald, J. W.; Hoppel, W. A.; Vietti, M. A. (1982) The size and scattering coefficient of urban aerosol particles at
Washington, DC as a function of relative humidity. J. Atmos. Sci. 39: 1838-1852.
Flossmann, A. I.; Hall, W. D.; Pruppacher, H. R. (1985) A theoretical study of the wet removal of atmospheric
pollutants. Part I: The redistribution of aerosol particles captured through nucleation and impaction scavenging
by growing cloud drops. J. Atmos. Sci. 42: 583-606.
Foitzik, L. (1950) Zur meteorologischen Optik von Dunst und Nebel: Teil II [The meteorological optics of haze and
fog: part II]. Z. Meteorol. 4: 321-329.
Foreman, W. T.; Bidleman, T. F. (1990) Semivolatile organic compounds in the ambient air of Denver, Colorado.
Atmos. Environ. Part A 24: 2405-2416.
Forkel, R.; Seidl, W.; Dlugi, R.; Deigele, E. (1990) A one-dimensional numerical model to simulate formation and
balance of sulfate during radiation fog events. J. Geophys. Res. [Atmos.] 95: 18,501-18,515.
Forrest, J.; Newman, L. (1973) Sampling and analysis of atmospheric sulfur compounds for isotope ratio studies.
Atmos. Environ. 7: 561-573.
Forrest, J.; Newman, L. (1977a) Further studies on the oxidation of sulfur dioxide in coal-fired power plant plumes.
Atmos. Environ. 11: 465-474.
Forrest, J.; Newman, L. (1977b) Oxidation of sulfur dioxide in the Sudbury smelter plume. Atmos. Environ.
11:517-520.
Forrest, J.; Garber, R.; Newman, L. (1979) Formation of sulfate, ammonium and nitrate in an oil-fired power plant
plume. Atmos. Environ. 13: 1287-1297.
Forrest, J.; Garber, R. W.; Newman, L. (1981) Conversion rates in power plant plumes based on filter pack data: the
coal-fired Cumberland plume. Atmos. Environ. 15: 2273-2282.
Freeman, D. J.; Cattell, F. C. R. (1990) Woodbuming as a source of atmospheric poly cyclic aromatic hydrocarbons.
Environ. Sci. Technol. 24: 1581-1585.
Frick, G. M.; Hoppel, W. A. (1993) Airship measurements of aerosol size distributions, cloud droplet spectra, and trace
gas concentrations in the marine boundary layer. Bull. Am. Meteorol. Soc. 74: 2195-2202.
Fried, A.; Henry, B. E.; Calvert, J. G.; Mozurkewich, M. (1994) The reaction probability of N2O5 with sulfuric acid
aerosols at stratospheric temperatures and compositions. J. Geophys. Res. [Atmos.] 99: 3517-3532.
Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition. J.
Aerosol Sci. 1:295-307.
Friedlander, S. K.; Turner, J. R.; Hering, S. V. (1986) A new method for estimating dry deposition velocities for
atmospheric aerosols. J. Aerosol Sci. 17: 240-244.
Fung, C. S.; Misra, P. K.; Bloxam, R.; Wong, S. (1991) A numerical experiment on the relative importance of H2O2 and
O3 in aqueous conversion of SO2 to SO42". Atmos. Environ. Part A 25: 411-423.
Fuzzi, S.; Castillo, R. A.; Jiusto, J. E.; Lala, G. G. (1984) Chemical composition of radiation fog water at Albany, New
York, and its relationship to fog microphysics. J. Geophys. Res. [Atmos.] 89: 7159-7164.
3-204
-------
Fuzzi, S.; Orsi, G.; Nardini, G.; Facchini, M. C.; McLaren, S.; McLaren, E.; Mariotti, M. (1988) Heterogeneous
processes in the Po Valley radiation fog. J. Geophys. Res. [Atmos.] 93: 11,141-11,151.
Gagosian, R. B.; Peltzer, E. T.; Merrill, J. T. (1987) Long-range transport of terrestrially derived lipids in aerosols from
the south Pacific. Nature (London) 325: 800-803.
Galloway, J. N.; Thornton, J. D.; Norton, S. A.; Volchok, H. L.; McLean, R. A. N. (1982) Trace metals in atmospheric
deposition: a review and assessment. Atmos. Environ. 16: 1677-1700.
Garland, J. A. (1979) Resuspension of particulate matter from grass and soil. Harwell, United Kingdom: Atomic
Energy Research Establishment; report no. AERE-R9452. Available from: NTIS, Springfield, VA.;
AERE-R9452.
Gatz, D. F. (1977) Scavenging ratio measurements in METROMEX. In: Semonin, R. G.; Beadle, R. W., eds.
Precipitation scavenging (1974): proceedings of a symposium; October 1974; Champaign, IL. Oak Ridge, TN:
Energy Research and Development Administration; pp. 71-87. Available from: NTIS, Springfield, VA;
CONF-741003. (ERDA symposium series no. 41).
Gebhart, K. A.; Malm, W. C.; Day, D. (1994) Examination of the effects of sulfate acidity and relative humidity on
light scattering at Shenandoah National Park. Atmos. Environ. 28: 841-849.
Gerde, P.; Scholander, P. (1989) Mass transfer rates of poly cyclic aromatic hydrocarbons between micron-size particles
and their environment—theoretical estimates. Environ. Health Perspect. 79: 249-258.
Germani, M. S.; Buseck, P. R. (1991) Automated scanning electron microscopy for atmospheric particle analysis. Anal.
Chem. 63: 2232-2237.
Gervat, G. P.; Clark, P. A.; Marsh, A. R. W.; Choularton, T. W.; Gay, M. J. (1988) Controlled chemical kinetic
experiments in cloud: a review of the CERL/UMIST Great Dun Fell project. In: Unsworth, M. H.; Fowler, D.,
eds. Acid deposition at high elevation sites: proceedings of the NATO advanced research workshop;
September 1986; Edinburgh, United Kingdom. Dordrecht, The Netherlands: Kluwer Academic Publishers; pp.
283-298. (NATO ASI series C: mathematical and physical sciences v. 252).
Gery, M. W.; Fox, D. L.; Jeffries, H. E.; Stockburger, L.; Weathers, W. S. (1985) A continuous stirred tank reactor
investigation of the gas-phase reaction of hydroxyl radicals and toluene. Int. J. Chem. Kinet. 17: 931-955.
Gery, M. W.; Fox, D. L.; Kamens, R. M.; Stockburger, L. (1987) Investigation of hydroxyl radical reactions with
o-xylene and m-xylene in a continuous stirred tank reactor. Environ. Sci. Technol. 21: 339-348.
Ghan, S. J.; Chuang, C. C.; Penner, J. E. (1993) A parameterization of cloud droplet nucleation, part I: single aerosol
type. Atmos. Res. 30: 198-221.
Gifford, F. A., Jr. (1961) Use of routine meteorological observations for estimating atmospheric dispersion. Nucl. Saf.
2:47-51.
Gill, P. S.; Graedel, T. E.; Wechsler, C. J. (1983) Organic films on atmospheric aerosol particles, fog droplets, cloud
droplets, raindrops and snowflakes. Rev. Geophys. Space Phys. 21: 903-920.
Gillani, N. V. (1985) Modeling of chemical transformations of SOX and NOX in the polluted atmosphere—an overview
of approaches and current status. In: De Wispelaere, C., ed. Air pollution modeling and its application IV. New
York, NY: Plenum Publishers; pp. 163-192.
Gillani, N. V. (1986) Ozone formation in pollutant plumes: a reactive plume model with arbitrary crosswind resolution.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Sciences Research
Laboratory; EPA report no. EPA/600/3-86/051. Available from: NTIS, Springfield, VA; PB86-236973.
3-205
-------
Gillani, N. V.; Husar, R. B. (1976) Synoptic-scale haziness over the eastern US and its long range transport. Presented
at: 4th National AMS/SAF conference on fire and forest meteorology; Boston, MA.
Gillani, N. V.; Pleim, J. E. (1995) Subgrid-scale features of anthropogenic emissions of VOC and NOx in the context of
regional Eulerian models. Atmos. Environ.: in press.
Gillani, N. V.; Wilson, W. E. (1980) Formation and transport of ozone and aerosols in power plant plumes. Ann. N. Y.
Acad. Sci. 338: 276-296.
Gillani, N. V.; Wilson, W. E. (1983) Gas-to-particle conversion of sulfur in power plant plumes—II. observations of
liquid-phase conversions. Atmos. Environ. 17: 1739-1752.
Gillani, N. V.; Husar, R. B.; Husar, J. D.; Patterson, D. E.; Wilson, W. E., Jr. (1978) Project MISTT: kinetics of
particulate sulfur formation in a power plant plume out to 300 km. Atmos. Environ. 12: 589-598.
Gillani, N. V.; Kohli, S.; Wilson, W. E. (1981) Gas-to-particle conversion of sulfur in power plant plumes—
I. parametrization of the conversion rate for dry, moderately polluted ambient conditions. Atmos. Environ. 15:
2293-2313.
Gillani, N. V.; Colby, J. A.; Wilson, W. E. (1983) Gas-to-particle conversion of sulfur in power plant plumes—III.
parameterization of plume-cloud interactions. Atmos. Environ. 17: 1753-1763.
Gillani, N. V.; Shannon, J. D.; Patterson, D. E. (1984) Transport processes. In: Altshuller, A. P.; Linthurst, R. A., eds.
The acidic deposition phenomenon and its effects: critical assessment review papers, v. 1: atmospheric
sciences. Washington, DC: U.S. Environmental Protection Agency, Office of Research and Development;
report no. EPA/600/8-83/016AF. Available from: NTIS, Springfield, VA; PB85-100030.
Gillani, N. V.; Schwartz, S. E.; Leaitch, W. R.; Strapp, J. W.; Isaac, G. A. (1995) Field observations in continental
stratiform clouds: partitioning of cloud particles between droplets and unactivated interstitial aerosols. J.
Geophys. Res. [Atmos.] 100: 18,687-18,706.
Gillette, D. A. (1974) On the production of soil wind erosion aerosols having the potential for long range transport. J.
Rech. Atmos. 8: 735-744.
Gillette, D. (1980) Major contributions of natural primary continental aerosols: source mechanisms. In: Kniep, T. J.;
Lioy, P. J., eds. Aerosols: anthropogenic and natural, sources and transport. Ann N. Y. Acad. Sci. 338:
348-358.
Gillette, D. A.; Hanson, K. J. (1989) Spatial and temporal variability of dust production caused by wind erosion in the
United States. J. Geophys. Res. [Atmos.] 94: 2197-2206.
Gillette, D.; Nagamoto, C. (1992) [Size distribution and single particle composition for two dust storms in Soviet
Central Asia in September 1989 and size distributions and chemical composition of local soil]. In: Golitsyn, G.,
ed. Soviet-American experiment for the investigation of arid aerosols. St. Petersburg, USSR: Typhoon; pp.
130-140.
Gillette, D. A.; Passi, R. (1988) Modeling dust emission caused by wind erosion. J. Geophys. Res. [Atmos.]
93: 14,233-14,242.
Gillette, D. A.; Sinclair, P. C. (1990) Estimation of suspension of alkaline material by dust devils in the United States.
Atmos. Environ. Part A 24: 1135-1142.
Gillette, D. A.; Walker, T. R. (1977) Characteristics of airborne particles produced by wind erosion of sandy soil, high
plains of west Texas. Soil Sci. 123: 97-110.
3-206
-------
Gillette, D. A.; Clayton, R. N.; Mayeda, T. K.; Jackson, M. L.; Sridhar, K. (1978) Tropospheric aerosols from some
major dust storms of the southwestern United States. J. Appl. Meteorol. 17: 832-845.
Gillette, D. A.; Stensland, G. J.; Williams, A. L.; Barnard, W.; Gatz, D.; Sinclair, P. C.; Johnson, T. C. (1992) Emissions
of alkaline elements calcium, magnesium, potassium, and sodium from open sources in the contiguous United
States. Global Biogeochem. Cycles 6: 437-457.
Gleason, J. F.; Smha, A.; Howard, C. J. (1987) Kinetics of the gas-phase reaction HOSO2 + O2 - HO2 + SO3. J. Phys.
Chem. 91:719-724.
Godish, T. (1985) Air quality. Chelsea, MI: Lewis Publishers, Inc.
Godowitch, J. M. (1989) Evaluation and sensitivity analyses results of the MESOPUFF II model with CAPTEX
measurements. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research
and Exposure Assessment Laboratory; EPA report no. EPA/600/3-89/056. Available from: NTIS, Springfield,
VA;PB89-198253.
Goldberg, E. D. (1985) Black carbon in the environment: properties and distribution. New York, NY: John Wiley &
Sons. (Metcalf, R. L.; Stumm, W., eds. Environmental science and technology series).
Gordon, G. E. (1991) Airborne particles on global and regional scales. Environ. Sci. Technol. 25: 1822-1828.
Gordon, R. J.; Trivedi, N. J.; Singh, B. P.; Ellis, E. C. (1988) Characterization of aerosol organics by diffuse reflectance
Fourier transform infrared spectroscopy. Environ. Sci. Technol. 22: 672-677'.
Goschnick, J.; Fichtner, M.; Lipp, M.; Schuricht, J.; Ache, H. J. (1993) Depth-resolved chemical analysis of
environmental microparticles by secondary mass spectrometry. Appl. Surf. Sci. 70/71: 63-67.
Graedel, T. E.; Goldberg, K. I. (1983) Kinetic studies of raindrop chemistry: 1. inorganic and organic processes. J.
Geophys. Res. C: Oceans Atmos. 88: 10,865-10,882.
Graedel, T. E.; Weschler, C. J. (1981) Chemistry within aqueous atmospheric aerosols and raindrops. Rev. Geophys.
Space Phys. 19: 505-539.
Graedel, T. E.; Hawkins, D. T.; Claxton, L. D. (1986) Atmospheric chemical compounds: sources, occurrence, and
bioassay. Orlando, FL: Academic Press, Inc.
Granat, L.; Soderland, R. (1975) Atmospheric deposition due to long and short distance sources with special reference
to wet and dry deposition of sulfur. Stockholm, Sweden: Stockholm University, Meteorological Institute;
report no. MISU-A32.
Gray, H. A. (1986) Control of atmospheric fine primary carbon particle concentration [Ph.D. thesis]. Pasadena, CA:
California Institute of Technology; EQL report no. 23.
Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1984) Elemental and organic carbon particle
concentrations: a long-term perspective. Sci. Total Environ. 36: 17-25.
Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1986) Characteristics of atmospheric organic
and elemental carbon particle concentrations in Los Angeles. Environ. Sci. Technol. 20: 580-589.
Grell, G. A.; Dudhia, J.; Stauffer, D. R. (1994) A description of the fifth-generation Penn State/NCAR mesoscale model
(MM5). Boulder, CO: National Center for Atmospheric Research; technical note no. NCAR/TN-398+STR.
Grosjean, D. (1977) Aerosols. In: Ozone and other photochemical oxidants. Washington, DC: National Academy of
Sciences; pp. 45-125.
3-207
-------
Grosjean, D. (1984a) Particulate carbon in Los Angeles air. Sci. Total Environ. 32: 133-145.
Grosjean,D. (1984b) Photooxidation of 1 -heptane. Sci. Total Environ. 37: 195-211.
Grosjean, D. (1985) Reactions of o-cresol and nitrocresol with NOX in sunlight and with ozone-nitrogen dioxide
mixtures in the dark. Environ. Sci. Technol. 19: 968-974.
Grosjean, D. (1992) In situ organic aerosol formation during a smog episode: estimated production and chemical
functionality. Atmos. Environ. Part A 26: 953-963.
Grosjean, D.; Friedlander, S. K. (1975) Gas-particle distribution factors for organic and other pollutants in the Los
Angeles atmosphere. J. Air Pollut. Control Assoc. 25: 1038-1044.
Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols. Atmos.
Environ. 23: 1733-1747.
Gundel, L. A.; Guyot-Sionnest, N. S.; Novakov, T. (1989) A study of the interaction of NO2 with carbon particles.
Aerosol Sci. Technol. 10: 343-351.
Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the phase
distributions of semi-volatile poly cyclic aromatic hydrocarbons using annular denuders. Atmos. Environ. 29:
1719-1733.
Hahn, J. (1980) Organic constituents of natural aerosols. Ann. N. Y. Acad. Sci. 338: 359-376.
Hales, J. M. (1991) Atmospheric processes research and process model development. Section 4. Precipitation
scavenging. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
Program; pp. 2-109 - 2-168. (State of science and technology report no. 2).
Hales, J. M.; Dana, M. T. (1979) Precipitation scavenging of urban pollutants by convective storm systems. J. Appl.
Meteorol. 18:294-316.
Hales, J. M.; Hicks, B. B.; Miller, J. M. (1987) The role of research measurement networks as contributors to federal
assessments of acid deposition. Bull. Am. Meteorol. Soc. 68: 216-225.
Hall, F. P., Jr.; Duchon, C. E.; Lee, L. G.; Hagan, R. R. (1973) Long-range transport of air pollution: a case study,
August 1970. Mon. Weather Rev. 101: 404-411.
Hamilton, R. S.; Mansfield, T. A. (1991) Airborne particulate elemental carbon: its sources, transport and contribution
to dark smoke and soiling. Atmos. Environ. Part A 25: 715-723.
Hampton, C. V.; Pierson, W. R.; Schuetzle, D.; Harvey, T. M. (1983) Hydrocarbon gases emitted from vehicles on the
road. 2. Determination of emission rates from diesel and spark-ignition vehicles. Environ. Sci. Technol. 17:
699-708.
Hanel, G. (1976) The properties of atmospheric aerosol particles as functions of the relative humidity at thermodynamic
equilibrium with the surrounding moist air. In: Landsberg, H. E.; Van Mieghem, J., eds. Advances in
geophysics: v. 19. New York, NY: Academic Press; pp. 73-188.
Hansen, A. D. A.; Novakov, T. (1989) Real-time measurements of the size fractionation of ambient black carbon
aerosols at elevated humidities. Aerosol Sci. Technol. 10: 106-110.
Hansen, A. D. A.; Rosen, H. (1990) Individual measurements of the emission factor of aerosol black carbon in
automobile plumes. J. Air Waste Manage. Assoc. 40: 1654-1657.
3-208
-------
Hansen, A. D. A.; Banner, W. H.; Novakov, T. (1991) Sulfur dioxide oxidation in laboratory clouds. Atmos. Environ.
Part A 25: 2521-2530.
Harrison, L. (1985) The segregation of aerosols by cloud-nucleating activity. Part II: Observation of an urban aerosol.
J. dim. Appl. Meteorol. 24: 312-321.
Harrison, R. M.; Kitto, A.-M. N. (1992) Estimation of the rate constant for the reaction of acid sulphate aerosol with
NH3 gas from atmospheric measurements. J. Atmos. Chem. 15: 133-143.
Harrison, R. M.; Msibi, I. M. (1994) Validation of techniques for fast response measurement of HNO3 and NH3 and
determination of the [NH3][HNO3] concentration product. Atmos. Environ. 28: 247-255.
Harrison, R. M.; Pio, C. A. (1983) A comparative study of the ionic composition of rainwater and atmospheric aerosols:
implications for the mechanism of acidification of rainwater. Atmos. Environ. 17: 2539-2543.
Harrison, R. M.; Laxen, D. P. H.; Wilson, S. J. (1981) Chemical associations of lead, cadmium, copper, and zinc in
street dusts and roadside soils. Environ. Sci. Technol. 15: 1378-1383.
Hart, K. M.; Pankow, J. F. (1994) High-volume air sampler for particle and gas sampling. 2. Use of backup filters to
correct for the adsorption of gas-phase poly cyclic aromatic hydrocarbons to the front filter. Environ. Sci.
Technol. 28: 655-661.
Hasan, H.; Dzubay, T. G. (1987) Size distributions of species in fine particles in Denver using a microorifice impactor.
Aerosol Sci. Technol. 6: 29-39.
Hatakeyama, S.; Tanonaka, T.; Weng, J.; Bandow, H.; Takagi, H.; Akimoto, H. (1985) Ozone-cyclohexene reaction in
air: quantitative analysis of particulate products and the reaction mechanism. Environ. Sci. Technol. 19:
935-942.
Hatakeyama, S.; Ohno, M.; Weng, J.; Takagi, H.; Akimoto, H. (1987) Mechanism for the formation of gaseous and
particulate products from ozone-cycloalkene reactions in air. Environ. Sci. Technol. 21: 52-57.
Hatakeyama, S.; Izumi, K.; Fukuyama, T.; Akimoto, H. (1989) Reactions of ozone with a-pinene and p-pinene in air:
yields of gaseous and particulate products. J. Geophys. Res. [Atmos.] 94: 13,013-13,024.
Hatakeyama, S.; Izumi, K.; Fukuyama, T.; Akimoto, H.; Washida, N. (1991) Reactions of OH with a-pinene and
P-pinene in air: estimate of global CO production from the atmospheric oxidation of terpenes. J. Geophys. Res.
[Atmos.] 96: 947-958.
Hegg, D. A. (1985) The importance of liquid-phase oxidation of SO2 in the troposphere. J. Geophys. Res. [Atmos.] 90:
3773-3779.
Hegg, D. A.; Hobbs, P. V. (1979) Some observations of particulate nitrate concentrations in coal-fired power plant
plumes. Atmos. Environ. 13: 1715-1716.
Hegg, D. A.; Hobbs, P. V. (1980) Measurements of gas-to-particle conversion in the plumes from five coal-fired
electric power plants. Atmos. Environ. 14: 99-116.
Hegg, D. A.; Hobbs, P. V. (1982) Measurements of sulfate production in natural clouds. Atmos. Environ.
16: 2663-2668.
Hegg, D. A.; Hobbs, P. V. (1983a) Errata [Hegg and Hobbs (1982)]. Atmos. Environ. 17: 1059.
Hegg, D. A.; Hobbs, P. V. (1983b) Author's reply: measurements of sulfate production in natural clouds. Atmos.
Environ. 17: 2632-2633.
3-209
-------
Hegg, D. A.; Hobbs, P. V. (1986) Sulfate and nitrate chemistry in cumuliform clouds. Atmos. Environ. 20: 901-909.
Hegg, D. A.; Hobbs, P. V. (1987) Comparisons of measured sulfate production due to ozone oxidation in clouds with a
kinetic rate equation. Geophys. Res. Lett. 14: 719-721.
Hegg, D. A.; Hobbs, P. V. (1988) Comparisons of sulfate and nitrate production in clouds on the mid-Atlantic and
Pacific Northwest Coasts of the United States. J. Atmos. Chem. 7: 325-333.
Hegg, D. A.; Larson, T. V. (1990) The effects of microphysical parameterization on model predictions of sulfate
production in clouds. Tellus Ser. B 42B: 272-284.
Hegg, D. A.; Hobbs, P. V.; Radke, L. F. (1984) Measurements of the scavenging of sulfate and nitrate in clouds.
Atmos. Environ. 18: 1939-1946.
Hegg, D. A.; Radke, L. F.; Hobbs, P. V. (1990) Particle production associated with marine clouds. J. Geophys. Res.
[Atmos.] 95: 13,917-13,926.
Hegg, D. A.; Radke, L. F.; Hobbs, P. V. (1991) Measurements of Aitken nuclei and cloud condensation nuclei in the
marine atmosphere and their relation to the DMS-cloud-climate hypothesis. J. Geophys. Res. [Atmos.] 96:
18,727-18,733.
Hegg, D. A.; Yuen, P.-F.; Larson, T. V. (1992) Modeling the effects of heterogeneous cloud chemistry on the marine
particle size distribution. J. Geophys. Res. [Atmos.] 97: 12,927-12,933.
Heintzenberg, J.; Winkler, P. (1984) Elemental carbon in the urban aerosol: results of a seventeen month study in
Hamburg, FRG. Sci. Total Environ. 36: 27-38.
Heintzenberg, J.; Ogren, J. A.; Noone, K. J.; Gardneus, L. (1989) The size distribution of submicrometer particles
within and about stratocumulus cloud droplets on Mt. Areskutan, Sweden. Atmos. Res. 24: 89-101.
Heisler, S. L.; Friedlander, S. K. (1977) Gas-to-particle conversion in photochemical smog - aerosol growth laws and
mechanisms for organics. Atmos. Environ. 11: 157-168.
Hering, S. V.; Friedlander, S. K. (1982) Origins of aerosol sulfur size distributions in the Los Angeles basin. Atmos.
Environ. 16: 2647-2656.
Hering, S. V.; McMurry, P. H. (1991) Optical counter response to monodisperse atmospheric aerosols. Atmos.
Environ. Part A 25: 463-468.
Hering, S. V.; Lawson, D. R.; Allegrini, I.; Febo, A.; Perrino, C.; Possanzini, M.; Sickles, J. E., II; Anlauf, K. G.;
Wiebe, A.; Appel, B. R.; John, W.; Ondo, J.; Wall, S.; Braman, R. S.; Sutton, R.; Cass, G. R.; Solomon, P. A.;
Eatough, D. J.; Eatough, N. L.; Ellis, E. C.; Grosjean, D.; Hicks, B. B.; Womack, J. D.; Horrocks, J.; Knapp, K.
T.; Ellestad, T. G.; Paur, R. I; Mitchell, W. J.; Pleasant, M.; Peake, E.; MacLean, A.; Pierson, W. R.;
Brachaczek, W.; Schiff, H. L; Mackay, G. L; Spicer, C. W.; Stedman, D. H.; Winer, A. M.; Biermann, H. W.;
Tuazon, E. C. (1988) The nitric acid shootout: field comparison of measurement methods. Atmos. Environ. 22:
1519-1539.
Hicks, B. B. (1984) Dry deposition processes. In: Altshuller, A. P.; Linthurst, R. A., eds. Acidic deposition
phenomenon and its effects: critical assessment review papers, volume I, atmospheric sciences. Washington,
DC: U.S. Environmental Protection Agency, Office of Research and Development; pp. 7-1 to 7-70; EPA report
no. EPA/600/8-83/016AF. Available from: NTIS, Springfield, VA; PB85-100030.
Hicks, B. B.; Meyers, T. P. (1989) Atmosphere-surface exchange processes. Presented at: 82nd annual meeting of the
Air & Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 89-113.4.
3-210
-------
Hicks, B. B.; Wesely, M. L.; Lindberg, S. E.; Bromberg, S. M., eds. (1986) Proceedings of the NAPAP workshop on
dry deposition; March; Harpers Ferry, WV. Oak Ridge, TN: National Oceanic and Atmospheric
Administration.
Hicks, B. B.; Draxler, R. R.; Dodge, M.; Hales, J. M.; Albritton, D. L.; Schwartz. S. E.; Meyers, T. P.; Fehsenfeld, F. C.;
Tanner, R. L.; Vong, R. J. (1991) Atmospheric processes research and process model development. In: Irving,
P. M., ed. Acidic deposition: state of science and technology, volume I: emissions, atmospheric processes, and
deposition. Washington, DC: The U.S. National Acid Precipitation Assessment Program. (State of science and
technology report no. 2).
Hildemann, L. M. (1990) A study of the origin of atmospheric organic aerosols [Ph.D. thesis]. Pasadena, CA:
California Institute of Technology. Available from: University Microfilms, Ann Arbor, MI; order no.
AAC90-14378.
Hildemann, L. M.; Russell, A. G.; Cass, G. R. (1984) Ammonia and nitric acid concentrations in equilibrium with
atmospheric aerosols: experiment vs theory, Atmos. Environ. 18: 1737-1750.
Hildemann, L. M.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1993) Mathematical modeling of urban organic
aerosol: properties measured by high-resolution gas chromatography. Environ. Sci. Technol. 27: 2045-2055.
Hildemann, L. M.; Klinedinst, D. B.; Klouda, G. A.; Currie, L. A.; Cass, G. R. (1994a) Sources of urban contemporary
carbon aerosol. Environ. Sci. Technol. 28: 1565-1576.
Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994b) Seasonal trends in Los Angeles ambient
organic aerosol observed by high-resolution gas chromatography. Aerosol Sci. Technol. 20: 303-317.
Hill, T. A.; Choularton, T. W.; Penkett, S. A. (1986) A model of sulphate production in a cap cloud and subsequent
turbulent deposition onto the hill surface. Atmos. Environ. 20: 1763-1771.
Hinds, W. C. (1982) Aerosol technology. New York, NY: John Wiley and Sons.
Hobbs, P. V.; Bowdle, D. A.; Radke, L. F. (1985) Particles in the lower troposphere over the high plains of the United
States. Part I: size distributions, elemental compositions and morphologies. J. Clim. Appl. Meteorol. 24:
1344-1356.
Hoffmann, M. R.; Boyce, S. D. (1983) Catalytic autooxidation of aqueous sulfur dioxide in relationship to atmospheric
systems. Adv. Environ. Sci. Technol. 12: 147-189.
Hoffmann, M. R.; Calvert, J. G. (1985) Chemical transformation modules for Eulerian acid deposition models: volume
II, the aqueous-phase chemistry. Research Triangle Park, NC: U.S. Environmental Protection Agency,
Atmospheric Sciences Research Laboratory; EPA report no. EPA/600/3-85/036. Available from: NTIS,
Springfield, VA; PB85-198653/AS.
Hoffmann, M. R.; Edwards, J. O. (1975) Kinetics of the oxidation of sulfite by hydrogen peroxide in acidic solution. J.
Phys. Chem. 79: 2096-2098.
Hoffmann, M. R.; Jacob, D. J. (1984) Kinetics and mechanisms of the catalytic oxidation of dissolved sulfur dioxide in
aqueous solution: an application to nighttime fog water chemistry. In: Calvert, J. G., ed. SO2, NO and NO2
oxidation mechanisms: atmospheric considerations. Boston, MA: Butterworth Publishers; pp. 101-172.
(Teasley, J. I., ed. Acid precipitation series: v. 3).
Hoffmann, H. E.; Kuehnemann, W. (1979) Comparison of the results of two measuring methods determining the
horizontal standard visibility with the visual visibility range. Atmos. Environ. 13: 1629-1634.
3-211
-------
Holsen, T. M; Noll, K. E.; Fang, G.-C.; Lee, W.-J.; Lin, J.-M.; Keeler, G. J. (1993) Dry deposition and particle size
distributions measured during the Lake Michigan Urban Air Toxics Study. Environ. Sci. Technol. 27:
1327-1333.
Hong, M.-S.; Carmichael, G. R. (1986a) An investigation of sulfate production in an orographic storm using a detailed
transport/chemistry model coupled with a detailed cloud scavenging model. Atmos. Environ. 20: 1989-1997.
Hong, M.-S.; Carmichael, G. R. (1986b) Examination of a subgrid-scale parameterization for the transport of pollutants
in a nonprecipitating cumulus cloud ensemble. Atmos. Environ. 20: 2205-2217.
Hoppel, W. A. (1988) The role of non-precipitating cloud cycles and gas-to-particle conversion in the maintainence of
the submicron aerosol size distribution over the tropical oceans. In: Hobbs, P. V.; McCormick, M. P., eds.
Aerosols and climate. Hampton, VA: A. Deepak Publishers; pp. 9-19.
Hoppel, W. A.; Frick, G. M. (1990) Submicron aerosol size distributions measured over the tropical and South Pacific.
Atmos. Environ. Part A 24: 645-659.
Hoppel, W. A.; Frick, G. M.; Larson, R. E. (1986) Effect of nonprecipitating clouds on the aerosol size distribution in
the marine boundary layer. Geophys. Res. Lett. 13: 125-128.
Hoppel, W. A.; Fitzgerald, J. W.; Frick, G. M.; Larson, R. E.;, et al. (1990) Aerosol size distributions and optical
properties found in the marine boundary layer over the Atlantic Ocean. J. Geophys. Res. [Atmos.] 95:
3659-3686.
Hoppel, W. A.; Frick, G. M.; Fitzgerald, J.; Larson, R. E. (1994) Marine boundary layer measurements of new particle
formation and the effects nonprecipitating clouds have on aerosol size distribution. J. Geophys. Res. [Atmos.]
99: 14,443-14,459.
Horvath, H.; Noll, K. E. (1969) The relationship between atmospheric light scattering coefficient and visibility. Atmos.
Environ. 3: 543-550.
Hosker, R. P., Jr.; Lindberg, S. E. (1982) Review: atmospheric deposition and plant assimilation of gases and particles.
Atmos. Environ. 16: 889-910.
Hov, O.; Isaksen, I. S. A. (1981) Generation of secondary pollutants in a power plant plume: a model study. Atmos.
Envrron. 15: 2367-2376.
Huang, P.-F.; Turpin, B. J.; Pipho, M. J.; Kittelson, D. B.; McMurry, P. H. (1994) Effects of water condensation and
evaporation on diesel chain-agglomerate morphology. J. Aerosol Sci. 25: 447-459.
Hubbe, J. M. (1989) Assessing procedures for estimating deposition rates for pollutants of interest. Sacramento, CA:
California Air Resources Board, San Joaquin Valley Air Pollution Study Agency; San Joaquin Valley Air
Quality Study technical support study no. 8.
Hudischewskyj, A. B.; Seigneur, C. (1989) Mathematical modeling of the chemistry and physics of aerosols in plumes.
Environ. Sci. Technol. 23: 413-421.
Huntzicker, J. J.; Hoffman, R. S.; Ling, C.-S. (1978) Continuous measurement and speciation of sulfur-containing
aerosols by flame photometry. Atmos. Environ. 12: 83-88.
Huntzicker, J. J.; Gary, R. A.; Ling, C.-S. (1980) Neutralization of sulfuric acid aerosol by ammonia. Environ. Sci.
Technol. 14: 819-824.
Huntzicker, J. J.; Johnson, R. L.; Shah, J. J.; Gary, R. A. (1982) Analysis of organic and elemental carbon in ambient
aerosols by a thermal-optical method. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon - atmospheric
3-212
-------
life cycle: proceedings of an international symposium; October 1980; Warren, MI. New York, NY: Plenum
Press; pp. 79-88.
Husain, L.; Dutkiewicz, V. A.; Hussain, M. M; Khwaja, H. A.; Burkhard, E. G.; Mehmood, G.; Parekh, P. P.; Canelli,
E. (1991) A study of heterogeneous oxidation of SO2 in summer clouds. J. Geophys. Res. [Atmos.] 96:
18,789-18,805.
Husar, R. B.; Gillani, N. V.; Husar, J. D. (1976) Particulate sulfur formation in power plant: urban and regional plumes.
Presented at: Symposium on aerosol science and technology at the 82nd national meeting of the AIChE;
August-September; Atlantic City, MD.
Husar, R. B.; Patterson, D. E.; Husar, J. D.; Gillani, N. V.; Wilson, W. E., Jr. (1978) Sulfur budget of a power plant
plume. In: Husar, R. B.; Lodge, J. P., Jr.; Moore, D. J., eds. Sulfur in the atmosphere: proceedings of the
international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos. Environ. 12: 549-568.
Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
haziness: a summary. Atmos. Environ. 15: 1919-1928.
Huss, A., Jr.; Lim, P. K.; Eckert, C. A. (1982a) Oxidation of aqueous sulfur dioxide. 1. Homogeneous manganese(II)
and iron(II) catalysis at low pH. J. Phys. Chem. 86: 4224-4228.
Huss, A., Jr.; Lim, P. K.; Eckert, C. A. (1982b) Oxidation of aqueous sulfur dioxide. 2. High-pressure studies and
proposed reaction mechanisms. J. Phys. Chem. 86: 4229-4233.
Ibusuki, T.; Takeuchi, K. (1987) Sulfur dioxide oxidation by oxygen catalyzed by mixtures of manganese(II) and
iron(III) in aqueous solutions at environmental reaction conditions. Atmos. Environ. 21: 1555-1560.
Irwin, J. S. (1983) Estimating plume dispersion—a comparison of several sigma schemes. J. Clim. Appl. Meteorol. 22:
92-114.
Isaksen, I. S. A.; Hov, O.; Hesstvedt, E. (1978) Ozone generation over rural areas. Environ. Sci. Technol.
12: 1279-1284.
Izumi, K.; Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of NOX.
Atmos. Environ. Part A 24: 1433-1441.
Izumi, K.; Murano, K.; Mizuochi, M.; Fukuyama, T. (1988) Aerosol formation by the photooxidation of cyclohexene
in the presence of nitrogen oxides. Environ. Sci. Technol. 22: 1207-1215.
Jackson, M. L.; Gillette, D. A.; Damelsen, E. F.; Blifford, I. H.; Bryson, R. A.; Syers, J. K. (1973) Global dustfall
during the quaternary as related to environments. Soil Sci. 116: 135-145.
Jacob, D. J. (1986) Chemistry of OH in remote clouds and its role in the production of formic acid and
peroxymonosulfate. J. Geophys. Res. [Atmos.] 91: 9807-9826.
Jacob, D. J.; Hoffmann, M. R. (1983) A dynamic model for the production of H+, NO3", and SO42" in urban fog. J.
Geophys. Res. C: Oceans Atmos. 88: 6611-6621.
Jacob, D. J.; Waldman, J. M.; Munger, J. W.; Hoffmann, M. R. (1985) Chemical composition of fogwater collected
along the California coast. Environ. Sci. Technol. 19: 730-736.
Jacob, D. J.; Munger, J. W.; Waldman, J. M.; Hoffmann, M. R. (1986a) The H2SO4-HNO3-NH3 system at high
humidities and in fogs: 1. spatial and temporal patterns in the San Joaquin Valley of California. J. Geophys.
Res. [Atmos.] 91: 1073-1088.
3-213
-------
Jacob, D. J.; Waldman, J. M; Munger, J. W.; Hoffmann, M. R. (1986b) The H2SO4-HNO3-NH3 system at high
humidities and in fogs: 2. comparison of field data with thermodynamic calculations. J. Geophys. Res.
[Atmos.] 91: 1089-1096.
Jaffrezo, J.-L.; Colin, J.-L. (1988) Rain-aerosol coupling in urban area: scavenging ratio measurement and
identification of some transfer processes. Atmos. Environ. 22: 929-935.
Japar, S. M.; Brachaczek, W. W.; Gorse, R. A., Jr.; Norbeck, J. M.; Pierson, W. R. (1986) The contribution of elemental
carbon to the optical properties of rural atmospheric aerosols. Atmos. Environ. 20: 1281-1289.
Jassim, J. A.; Lu, H. P.; Chughtai, A. R.; Smith, D. M. (1986) The origin of infrared absorption bands of soot in the
region 750 to 900 cm'1. Appl. Spectrosc. 40: 113-116.
John, W.; Sethi, V. (1993) Threshold for resuspension by particle impaction. Aerosol Sci. Technol. 19: 69-79.
John, W.; Wang, H.-C. (1991) Laboratory testing method for PM-10 samplers: lowered effectiveness from particle
loading. Aerosol Sci. Technol. 14: 93-101.
John, W.; Wall, S. M.; Ondo, J. L.; Winklmayr, W. (1990) Modes in the size distributions of atmospheric inorganic
aerosol. Atmos. Environ. Part A 24: 2349-2359.
John, W.; Winklmayr, W.; Wang, H.-C. (1991) Particle deagglomeration and reentrainment in a PM-10 sampler.
Aerosol Sci. Technol. 14: 165-176.
Johnson, R. L.; Shah, J. J.; Gary, R. A.; Huntzicker, J. J. (1981) An automated thermal-optical method for the analysis
of carbonaceous aerosol. In: Macias, E. S.; Hopke, P. K., eds. Atmospheric aerosol: source/air quality
relationships; based on a symposium jointly sponsored by the Divisions of Nuclear Chemistry and Technology
at the 180th national meeting of the American Chemical Society; August 1980; Las Vegas, NV. Washington,
DC: American Chemical Society; pp. 223-233. (Comstock, M. J., ed. ACS symposium series: 167).
Johnson, C. A.; Sigg, L.; Zobrist, J. (1987) Case studies on the chemical composition of fogwater: the influence of local
gaseous emissions. Atmos. Environ. 21: 2365-2374.
Joos, F.; Baltensperger, U. (1991) A field study on chemistry, S(IV) oxidation rates and vertical transport during fog
conditions. Atmos. Environ. Part A 25: 217-230.
Junge, C. E. (1963) Air chemistry and radioactivity. New York, NY: Academic Press. (Van Mieghem, J.; Hales, A. L.,
eds. International geophysics series: v. 4).
Junge, C. E. (1977) Basic considerations about trace constituents in the atmosphere as related to the fate of global
pollutants. In: Suffet, I. H., ed. Fate of pollutants in the air and water environments: part I, mechanism of
interaction between environments and mathematical modeling and the physical fate of pollutants, papers from
the 165th national American Chemical Society meeting; April 1975; Philadelphia, PA. New York, NY: John
Wiley & Sons; pp. 7-25. (Advances in environmental science and technology: v. 8).
Junge, C.; McLaren, E. (1971) Relationship of cloud nuclei spectra to aerosol size distribution and composition. J.
Atmos. Sci. 28: 382-390.
Kadowaki, S. (1994) Characterization of carbonaceous aerosols in the Nagoya urban area. 2. Behavior and origin of
particulate w-alkanes. Environ. Sci. Technol. 28: 129-135.
Kahl, J. D.; Samson, P. J. (1986) Uncertainly in trajectory calculations due to low resolution meteorological data. J.
Clim. Appl. Meteorol. 25: 1816-1831.
Kahl, J. D.; Samson, P. J. (1988) Trajectory sensitivity to rawinsonde data resolution. Atmos. Environ. 22: 1291-1299.
3-214
-------
Kahl, J. D.; Schnell, R. C.; Sheridan, P. J.; Zak, B. D.; Church, H. W.; Mason, A. S.; Heffter, J. L.; Hams, J. M. (1991)
Predicting atmospheric debris transport in real-time using a trajectory forecast model. Atmos. Environ. Part A
25: 1705-1713.
Kamens, R. M.; Jeffries, H. E.; Gery, M. W.; Wiener, R. W.; Sexton, K. G.; Howe, G. B. (1981) The impact of
a-pinene on urban smog formation: an outdoor smog chamber study. Atmos. Environ. 15: 969-981.
Kamens, R. M.; Gery, M. W.; Jeffries, H. E.; Jackson, M.; Cole, E. I. (1982) Ozone-isoprene reactions: product
formation and aerosol potential. Int. J. Chem. Kinet. 14: 955-975.
Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile poly cyclic
aromatic hydrocarbons. Environ. Sci. Technol. 29: 43-50.
Kaplan, I. R.; Gordon, R. J. (1994) Non-fossil-fuel fine-particle organic carbon aerosols in Southern California
determined during the Los Angeles Aerosol Characterization and Source Apportionment Study. Aerosol Sci.
Technol. 21:343-359.
Karamchandani, P.; Venkatram, A. (1992) The role of non -precipitating clouds in producing ambient sulfate during
summer: results from simulations with the Acid Deposition and Oxidant Model (ADOM) Atmos. Environ. Part
A 26: 1041-1052.
Katrinak, K. A.; Rez, P.; Buseck, P. R. (1992) Structural variations in individual carbonaceous particles from an urban
aerosol. Environ. Sci. Technol. 26: 1967-1976.
Katrinak, K. A.; Rez, P.; Perkes, P. R.; Buseck, P. R. (1993) Fractal geometry of carbonaceous aggregates from an
urban aerosol. Environ. Sci. Technol. 27: 539-547.
Kawamura, K.; Ng, L.-L.; Kaplan, I. R. (1985) Determination of organic acids (Cj-C10) in the atmosphere, motor
exhausts, and engine oils. Environ. Sci. Technol. 19: 1082-1086.
Keeler, G. J. ; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity on
dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
Keeler, G. I; Japar, S. M.; Brachaczek, W. W.; Gorse, R. A., Jr.; Norbeck, J. M.; Pierson, W. R. (1990) The sources of
aerosol elemental carbon at Allegheny Mountain. Atmos. Environ. Part A 24: 2795-2805.
Kelly, N. A. (1987) The photochemical formation and fate of nitric acid in the metropolitan Detroit area: ambient,
captive-air irradiation and modeling results. Atmos. Environ. 21: 2163-2177.
Kelly, T. J.; Schwartz, S. E.; Daum, P. H. (1989) Detectability of acid producing reactions in natural clouds. Atmos.
Environ. 23: 569-583.
Kerminen, V. M.; Wexler, A. S. (1994a) Particle formation due to SO2 oxidation and high relative humidity in the
remote marine boundary layer. J. Geophys. Res. [Atmos.] 99: 25607-25625.
Kerminen, V.-M.; Wexler, A. S. (1994b) Post-fog nucleation of HjSC^-HjO particles in smog. Atmos. Environ. 28:
2399-2406.
Khlystov, A.; ten Brink, H. M.; Wyers, G. P. (1993) Hygroscopic growth rates of aerosols at high relative humidity.
Petten, The Netherlands: Netherlands Energy Research Foundation ECN; report no. ECN-C-93-01 1.
Kim, Y. J. (1995) Response of the Active Scattering Aerosol Spectrometer Probe (ASASP-100X) to particles of
different chemical composition. Aerosol Sci. Technol. 22: 33-42.
3-215
-------
Kim, D.; Hopke, P. K. (1988) Classification of individual particles based on computer-controlled scanning electron
microscopy data. Aerosol Sci. Technol. 9: 133-151.
Kim, Y. P.; Seinfeld, J. H. (1995) Atmospheric gas-aerosol equilibrium: III. Thermodynamics of crustal elements Ca2+,
K+, andMg2+. Aerosol Sci. Technol. 22: 93-110.
Kim, M. G.; Yagawa, K.; Inoue, H.; Lee, Y. K.; Shirai, T. (1990) Measurement of tire tread in urban air by
pyrolysis-gas chromatography with flame photometric detection. Atmos. Environ. Part A 24: 1417-1422.
Kim, Y. P.; Seinfeld, J. H.; Saxena, P. (1993a) Atmospheric gas-aerosol equilibrium: I. thermodynamic model. Aerosol
Sci. Technol. 19: 157-181.
Kim, Y. P.; Seinfeld, J. H.; Saxena, P. (1993b) Atmospheric gas-aerosol equilibrium: II. analysis of common
approximations and activity coefficient calculation methods. Aerosol Sci. Technol. 19: 182-198.
Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993c) Vertical distribution of atmospheric
aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A 27: 1351-1362.
Kim, Y. P.; Pun, B. K.-L.; Chan, C. K.; Flagan, R. C.; Seinfeld, J. H. (1994) Determination of water activity in
ammonium sulfate and sulfuric acid mixtures using levitated single particles. Aerosol Sci. Technol.
20: 275-284.
Kitsa, V.; Lioy, P. J. (1992) Near field dispersion of mechanically resuspended dust from an unpaved road.
Transactions AWMA 1: 199-210.
Kitsa, V.; Lioy, P. J.; Chow, J. C.; Watson, J. G.; Shupack, S.; Howell, T.; Sanders, P. (1992) Particle-size distribution
of chromium: total and hexavalent chromium in inspirable, thoracic, and respirable soil particles from
contaminated sites in New Jersey. Aerosol Sci. Technol. 17: 213-229.
Klee, A. J. (1984) Source control - municipal solid waste incinerators. In: Calvert, S.; Englund, H. M., eds. Handbook
of air pollution technology. New York, NY: John Wiley & Sons; pp. 513-550.
Klouda, G. A.; Currie, L. A.; Verkouteren, R. M.; Einfeld, W.; Zak, B. D. (1988) Advances in microradiocarbon dating
and the direct tracing of environmental carbon. J. Radioanal. Nucl. Chem. 123: 191-197.
Koenig, L. R.; Murray, F. W. (1976) Ice-bearing cumulus cloud evolution: numerical simulation and general
comparison against observations. J. Appl. Meteorol. 15: 747-762.
Kohler, H. (1936) The nucleus in and the growth of hygroscopic droplets. Trans. Faraday Soc. 32: 1152-1161.
Kolb, C. E.; Jayne, J. T.; Worsnop, D. R.; Molina, M. J.; Meads, R. F.; Viggiano, A. A. (1994) Gas phase reaction of
sulfur trioxide with water vapor. J. Am. Chem. Soc. 116: 10314-10315.
Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988a) Evaluation
of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci. Technol. 22:
1463-1468.
Koutrakis, P.; Wolfson, J. M.; Spengler, J. D. (1988b) An improved method for measuring aerosol strong acidity:
results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmos. Environ. 22: 157-162.
Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric aerosol
sulfates as a function of the relative humidity. J. Geophys. Res. [Atmos.] 94: 6442-6448.
3-216
-------
Koutrakis, P.; Thompson, K. M; Wolfson, J. M.; Spengler, J. D.; Keeler, G. J.; Slater, J. L. (1992) Determination of
aerosol strong acidity losses due to interactions of collected particles: results from laboratory and field studies.
Atmos. Environ. Part A 26: 987-995.
Krieger, M. S.; Kites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds. Environ.
Sci. Technol. 26: 1551-1555.
Krieger, M. S.; Kites, R. A. (1994) Measurement of poly chlorinated biphenyls and poly cyclic aromatic hydrocarbons in
air with a diffusion denuder. Environ. Sci. Technol. 28: 1129-1133.
Kunen, S. M.; Lazrus, A. L.; Kok, G. L.; Heikes, B. G. (1983) Aqueous oxidation of SQ by hydrogen peroxide. J.
Geophys. Res. C: Oceans Atmos. 88: 3671-3674.
Kuo, Y.-H.; Skumanich, M.; Haagenson, P. L.; Chang, J. S. (1985) The accuracy of trajectory models as revealed by
the observing system simulation experiments. Mon. Weather Rev. 113: 1852-1867.
Lamb, B.; Guenther, A.; Gay, D.; Westberg, H. (1987) A national inventory of biogenic hydrocarbon emissions.
Atmos. Environ. 21: 1695-1705.
Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and evaluation of
a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air. Environ. Sci.
Technol. 22: 941-947.
Langner, J.; Rodhe, H. (1991) A global three-dimensional model of the tropospheric sulfur cycle. J. Atmos. Chem. 13:
225-263.
Larson, S. M.; Cass, G. R. (1989) Characteristics of summer midday low-visibility events in the Los Angeles area.
Environ. Sci. Technol. 23: 281-289.
Larson, S. M.; Cass, G. R.; Gray, H. A. (1989) Atmospheric carbon particles and the Los Angeles visibility problem.
Aerosol Sci. Technol. 10: 118-130.
Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
results from field studies. Environ. Sci. Technol. 28: 957-964.
Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
Lazrus, A. L.; Haagenson, P. L.; Kok, G. L.; Huebert, B. J.; Kreitzberg, C. W.; Likens, G. E.; Mohnen, V. A.; Wilson,
W. E.; Winchester, J. W. (1983) Acidity in air and water in a case of warm frontal precipitation. Atmos.
Environ. 17: 581-591.
Leaitch, W. R.; Strapp, J. W.; Wiebe, H. A.; Isaac, G. A. (1983) Measurements of scavenging and transformation of
aerosol inside cumulus. In: Pruppacher, H. R.; Semonin, R. G.; Slinn, W. G. N., eds. Precipitation scavenging,
dry deposition, and resuspension: proceedings of the fourth international conference, v. 1: precipitation
scavenging; November-December 1982; Santa Monica, CA. New York, NY: Elsevier; pp. 53-69.
Leaitch, W. R.; Strapp, J. W.; Wiebe, H. A.; Anlauf, K. G.; Isaac, G. A. (1986a) Chemical and microphysical studies of
nonprecipitating summer cloud in Ontario, Canada. J. Geophys. Res. [Atmos.] 91: 11,821-11,831.
Leaitch, W. R.; Strapp, J. W.; Isaac, G. A.; Hudson, J. G. (1986b) Cloud droplet nucleation and cloud scavenging of
aerosol sulphate in polluted atmospheres. Tellus Ser. B 38B: 328-344.
Leaitch, W. R.; Bottenheim, J. W.; Strapp, J. W. (1988) Possible contribution of N2O5 scavenging to HNO3 observed in
winter stratiform cloud. J. Geophys. Res. [Atmos.] 93: 12,569-12,584.
3-217
-------
Lee, Y.-N.; Schwartz, S. E. (1983) Kinetics of oxidation of aqueous sulfur(IV) by nitrogen dioxide. In: Pruppacher, H.
R.; Semonin, R. G.; Slinn, W. G. N., eds. Precipitation scavenging, dry deposition, and resuspension:
proceedings of the fourth international conference, v. 1: precipitation scavenging; November-December 1982;
Santa Monica, CA. New York, NY: Elsevier; pp. 453-470.
Leone, J. A.; Flagan, R. C.; Grosjean, D.; Seinfeld, J. H. (1985) An outdoor smog chamber and modeling study of
toluene-NOx photooxidation. Int. J. Chem. Kinet. 17: 177-216.
Lewis, R. S.; Been, W. M. (1994) Kinetics of the reaction of nitric oxide with oxygen in aqueous solutions. Chem. Res.
Toxicol. 7: 568-574.
Lewis, C. W.; Baumgardner, R. E.; Stevens, R. K.; Claxton, L. D.; Lewtas, J. (1988) Contribution of woodsmoke and
motor vehicle emissions to ambient aerosol mutagenicity. Environ. Sci. Technol. 22: 968-971.
Li, W.; Montassier, N.; Hopke, P. K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol Sci.
Technol. 17: 25-35.
Li, S.-M.; Anlauf, K. G.; Wiebe, H. A. (1993) Heterogeneous nighttime production and deposition of particle nitrate at
a rural site in North America during summer 1988. J. Geophys. Res. [Atmos.] 98: 5139-5157.
Liang, C. S. K.; Waldman, J. M. (1992) Indoor exposures to acidic aerosols at child and elderly care facilities. Indoor
Air 2: 196-207.
Liebsch, E. J.; de Pena, R. G. (1982) Sulfate aerosol production in coal-fired power plant plumes. Atmos. Environ. 16:
1323-1331.
Ligocki, M. P.; Pankow, J. F. (1989) Measurements of the gas/particle distributions of atmospheric organic compounds.
Environ. Sci. Technol. 23: 75-83.
Ligocki, M. P.; Salmon, L. G.; Fall, T.; Jones, M. C.; Nazaroff, W. W.; Cass, G. R. (1993) Charactenstics of airborne
particles inside southern California museums. Atmos. Environ. Part A 27: 697-711.
Lin, X.; Chameides, W. L. (1991) Model studies of the impact of chemical inhomogeneity on SO2 oxidation in warm
stratiform clouds. J. Atmos. Chem. 13: 109-129.
Lin, J.-M.; Fang, G.-C.; Holsen, T. M.; Noll, K. E. (1993) A comparison of dry deposition modeled from size
distribution data and measured with a smooth surface for total particle mass, lead and calcium in Chicago.
Atmos. Environ. Part A 27: 1131-1138.
Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
atmosphere. Aerosol Sci. Technol. 20: 239-252.
Lind, J. A.; Lazrus, A. L. (1983) Aqueous phase oxidation of sulfur IV by some organic peroxides [abstract]. EOS
Trans. 64: 670.
Lindberg, S. E. (1982) Factors influencing trace metal, sulfate and hydrogen ion concentrations in rain. Atmos.
Environ. 16: 1701-1709.
Linton, R. W.; Loh, A.; Natusch, D. F. S.; Evans, C. A., Jr.; Williams, P. (1976) Surface predominance of trace
elements in airborne particles. Science (Washington, DC) 191: 852-854.
Lioy, P. J.; Daisey, J. M. (1986) Airborne toxic elements and organic substances. Environ. Sci. Technol. 20: 8-14.
Lioy, P. J.; Watson, J. G., Jr. ; Spengler, J. D. (1980) APCA specialty conference workshop on baseline data for
inhalable particulate matter. J. Air Pollut. Control Assoc. 30: 1126-1130.
3-218
-------
Lippmann, M. (1980) Size distributions in urban aerosols. In: Kneip, T. J.; Lioy, P. J., eds. Aerosols: anthropogenic and
natural, sources and transport. Ann. N. Y. Acad. Sci. 338: 1-12.
Liu, P. S. K.; Leaitch, W. R.; Strapp, J. W.; Wasey, M. A. (1992) Response of particle measuring systems airborne to
ASASP and PCASP to NaCl and latex particles. Aerosol Sci. Technol. 16: 83-95.
Liu, P. S. K.; Leaitch, W. R.; Macdonald, A. M.; Isaac, G. A.; Strapp, J. W.; Wiebe, H. A. (1993) Sulphate production
in summer cloud over Ontario, Canada. Tellus Ser. B 45B: 368-389.
Lowenthal, D. H.; Rogers, C. F.; Saxena, P.; Watson, J. G.; Chow, J. C. (1995) Sensitivity of estimated light extinction
coefficients to model assumptions and measurement errors. Atmos. Environ. 29: 751-766.
Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air
pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
Lundgren, D. A.; Hausknecht, B. J. (1982a) Ambient aerosol mass distribution of 1 to 100 ,um particles in five cities.
Presented at: 75th annual meeting of the Air Pollution Control Association; June; New Orleans, LA. Pittsburgh,
PA: Air Pollution Control Association; paper no. 82-45.4.
Lundgren, D. A.; Hausknecht, B. J. (1982b) Ambient aerosol size distribution determination using a mobile wide range
aerosol classifier. Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory; EPA contract no. CR-808606-01-1.
Lundgren, D. A.; Hausknecht, B. J.; Burton, R. M. (1984) Large particle size distribution in five U.S. cities and the
effect on a new ambient particulate matter standard (PM10). Aerosol Sci. Technol. 7: 467-473.
Lusis, M. A.; Anlauf, K. G.; Barrie, L. A.; Wiebe, H. A. (1978) Plume chemistry studies at a northern Alberta power
plant. Atmos. Environ. 12: 2429-2437.
Maahs, H. G. (1983) Kinetics and mechanism of the oxidation of S(IV) by ozone in aqueous solution with particular
reference to SO2 conversion in nonurban tropospheric clouds. J. Geophys. Res. C: Oceans Atmos. 88:
10,721-10,732.
Macias, E. S.; Zwicker, J. O.; White, W. H. (1981) Regional haze case studies in the southwestern U.S.—II. source
contributions. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility: measurements and
model components: proceedings of the symposium; November 1980; Grand Canyon National Park, AZ.
Atmos. Environ. 15: 1987-1997.
Makhonko, K. P. (1986) Wind dust lift from the grass surface. Meteorol. Gidrol. 10: 61-65.
Malm, W. C.; Gebhart, K. A.; Molenar, J.; Cahill, T.; Eldred, R.; Huffman, D. (1994) Examining the relationship
between atmospheric aerosols and light extinction at Mount Rainier and North Cascades National Parks.
Atmos. Environ. 28: 347-360.
Mamane, Y.; Stevens, R. K.; Dzubay, T. G. (1990) On the sources of fine and coarse carbon particles in the Great
Lakes. J. Aerosol Sci. 21(suppl. 1): S353-S356.
Mansoori, B. A.; Johnston, M. V.; Wexler, A. S. (1994) Quantitation of ionic species in single microdroplets by on-line
laser desorption/ionization. Anal. Chem. 66: 3681-3687.
Martin, L. R. (1984) Kinetic studies of sulfite oxidation in aqueous solution. In: Calvert, J. G., ed. SO2, NO and NO2
oxidation mechanisms: atmospheric considerations. Boston, MA: Butterworth Publishers; pp. 63-100.
(Teasley, J. I., ed. Acid precipitation series: v. 3).
3-219
-------
Martin, L. R. (1994) Aqueous sulfur(IV) oxidation revisited. In: Nriagu, J. O.; Simmons, M. S., eds. Environmental
oxidants. New York, NY: John Wiley & Sons, Inc.; pp. 221-268. (Nriagu, J. O., ed. Advances in
environmental science and technology: v. 28).
Martin, L. R.; Damschen, D. E. (1981) Aqueous oxidation of sulfur dioxide by hydrogen peroxide at low pH. Atmos.
Environ. 15: 1615-1621.
Martin, L. R.; Hill, M. W. (1987a) The iron catalyzed oxidation of sulfur: reconciliation of the literature rates. Atmos.
Environ. 21: 1487-1490.
Martin, L. R.; Hill, M. W. (1987b) The effect of ionic strength on the manganese catalyzed oxidation of sulfur(IV).
Atmos. Environ. 21: 2267-2270.
Martin, L. R.; Hill, M. W.; Tai, A. F.; Good, T. W. (1991) The iron catalyzed oxidation of sulfur(IV) in aqueous
solution: differing effects of organics at high and low pH. J. Geophys. Res. [Atmos.] 96: 3085-3097.
Martonen, T. B. (1993) Mathematical model for the selective deposition of inhaled pharmaceuticals. J. Pharm. Sci. 82:
1191-1199.
Mason, A. S.; Gifford, F. A. (1992) Atmospheric tracer dispersion over a 24-h time span. Atmos. Environ. Part A 26:
3203-3205.
Mazurek, M. A.; Simoneit, B. R. T. (1984) Characterization of biogenic and petroleum-derived organic matter in
aerosols over remote, rural and urban areas. In: Keith, L. H., ed. Identification and analysis of organic
pollutants in air: [papers from a symposium at the 184th national American Chemical Society meeting];
September 1982; Kansas City, MO. Woburn, MA: Butterworth Publishers; pp. 353-370.
Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1989) Interpretation of high-resolution gas chromatography /mass
spectrometry data aquired from atmospheric organic aerosol samples. Aerosol Sci. Technol. 10: 408-420.
Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Biological input to visibility-reducing aerosol particles in the
remote arid southwestern United States. Environ. Sci. Technol. 25: 684-694.
McArdle, J. V.; Hoffmann, M. R. (1983) Kinetics and mechanism of the oxidation of aquated sulfur dioxide by
hydrogen peroxide at low pH. J. Phys. Chem. 87: 5425-5429.
McHenry, J. N; Dennis, R. L. (1991) Partitioning of the sulfate budget into gas- and aqueous-phase components in the
regional acid deposition model (RADM). In: Proceedings of the 7th joint AMS-AWMA conference:
applications of air pollution meteorology; New Orleans, LA; pp. 143-147.
McHenry, J. N.; Dennis, R. L. (1994) The relative importance of oxidation pathways and clouds to atmospheric
ambient sulfate production as predicted by the Regional Acid Deposition Model. J. Appl. Meteorol.
33: 890-905.
McMahon, T. A.; Denison, P. J. (1979) Empirical atmospheric deposition parameters—a survey. Atmos. Environ. 13:
571-585.
McMurry, P. H.; Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ. 13:
1635-1651.
McMurry, P. H.; Grosjean, D. (1985) Photochemical formation of organic aerosols: growth laws and mechanisms.
Atmos. Environ. 19: 1445-1451.
McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
aerosols. Atmos. Environ. 23: 497-507.
3-220
-------
McMurry, P. H.; Wilson, J. C. (1982) Growth laws for the formation of secondary ambient aerosols: implications for
chemical conversion mechanisms. Atmos. Environ. 16: 121-134.
McMurry, P. H.; Wilson, J. C. (1983) Droplet phase (heterogeneous) and gas phase (homogeneous) contributions to
secondary ambient aerosol formation as functions of relative humidity. J. Geophys. Res. 88: 5101-5108.
McMurry, P. H.; Zhang, X. Q. (1989) Size distributions of ambient organic and elemental carbon. Aerosol Sci.
Technol. 10: 430-437.
McMurry, P. H.; Rader, D. J.; Stith, J. L. (1981) Studies of aerosol formation in power plant plumes — I. growth laws
for secondary aerosols in power plant plumes: implications for chemical conversion mechanisms. In: White,
W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility: measurements and model components:
proceedings of the symposium; November 1980; Grand Canyon National Park, AZ. Atmos. Environ. 15:
2315-2327.
McMurry, P. H.; Zhang, X.; Lee, C.-T. (1996) Issues in aerosol measurement. J. Geophys. Res. [Atmos.]: in press.
McNider, R. T.; Moran, M. D.; Pielke, R. A. (1988) Influence of diurnal and inertial boundary -layer oscillations on
long-range dispersion. Atmos. Environ. 22: 2445-2462.
Meagher, J. F.; Stockburger, L.; Bailey, E. M.; Huff, O. (1978) The oxidation of sulfur dioxide to sulfate aerosols in the
plume of a coal-fired power plant. Atmos. Environ. 12: 2197-2203.
Meagher, J. F.; Stockburger, L., Ill; Bonanno, R. J.; Bailey, E. M.; Luria, M. (1981) Atmospheric oxidation of flue
gases from coal-fired power plants — a comparison between conventional and scrubbed plumes. Atmos.
Environ. 15: 749-762.
Meagher, J. F.; Olszyna, K. J.; Weatherford, F. P.; Mohnen, V. A. (1990) The availability of LL^ and O3 for aqueous
phase oxidation of SO2. The question of linearity. Atmos. Environ. Part A 24: 1825-1829.
Meng, Z.; Seinfeld, J. H. (1994) On the source of the submicrometer droplet mode of urban and regional aerosols.
Aerosol Sci. Technol. 20: 253-265.
Meng, Z.; Seinfeld, J. H.; Saxena, P.; Kim, Y. P. (1995) Contribution of water to particulate mass in the South Coast
Air Basin. Aerosol Sci. Technol. 22: 111-123.
Michaud, D.; Baril, M.; Perrault, G. (1993) Characterization of airborne dust from cast iron foundries by
physico-chemical methods and multivariate statistical analyses. Air Waste 43: 729-735.
Milford, J. B.; Davidson, C. I. (1985) The sizes of particulate trace elements in the atmosphere — a review. J. AirPollut.
Control Assoc. 35: 1249-1260.
Milford, J. B.; Davidson, C. I. (1987) The sizes of particulate sulfate and nitrate in the atmosphere — a review. JAPCA
37: 125-134.
Miller, D. F.; Alkezweeny, A. J. (1980) Aerosol formation in urban plumes over Lake Michigan. In: Kneip, T. J.; Lioy,
P. J., eds. Aerosols: anthropogenic and natural, sources and transport. Ann. N Y. Acad. Sci. 338: 219-232.
Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size considerations
for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc. 29: 610-615.
Mohnen, V. A.; Kadlecek, J. A. (1989) Cloud chemistry research at Whiteface Mountain. Tellus Ser. B 41B: 79-91.
Moran, M. D. (1992) Numerical modeling of mesoscale atmospheric dispersion [Ph.D. dissertation]. Ft. Collins, CO:
Colorado State University, Department of Atmospheric Science.
3-221
-------
Morandi, M. T. (1985) Development of source apportionment models for inhalable particulate matter and its extractable
organic fractions in urban areas of New Jersey [Ph.D. dissertation]. Syracuse, NY: New York University.
Available from: University Microfilms, Ann Arbor, MI; order no. AAC85-10516.
Morandi, M. T.; Lioy, P. J.; Daisey, J. M. (1991) Comparison of two multivariate modeling approaches for the source
apportionment of inhalable particulate matter in Newark, NJ. Atmos. Environ. Part A 25: 927-937.
Moyers, J. L.; Ranweiler, L. E.; Hopf, S. B.; Korte, N. E. (1977) Evaluation of particulate trace species in southwest
desert atmosphere. Environ. Sci. Technol. 11: 789-795.
Mozurkewich, M.; Calvert, J. G. (1988) Reaction probability of N2O5 on aqueous aerosols. J. Geophys. Res. [Atmos.]
93: 15,889-15,896.
Mueller, P. K.; Fung, K. K.; Heisler, S. L.; Grosjean, D.; Hidy, G. M. (1982) Atmospheric particulate carbon
observations in urban and rural areas of the United States. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate
carbon: atmospheric life cycle: proceedings of an international symposium; October 1980; Warren, MI. New
York, NY: Plenum Press; pp. 343-370.
Mulhbaier, J. L.; Williams, R. L. (1982) Fireplaces, furnaces and vehicles as emission sources of particulate carbon. In:
Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric life cycle. New York, NY: Plenum Press;
pp. 185-205.
Munger, J. W.; Jacob, D. J.; Waldman, J. M.; Hoffmann, M. R. (1983) Fogwater chemistry in an urban atmosphere. J.
Geophys. Res. C: Oceans Atmos. 88: 5109-5121.
Munger, J. W.; Jacob, D. J.; Hoffmann, M. R. (1984) The occurrence of bisulfite-aldehyde addition products in fog-
and cloudwater. J. Atmos. Chem. 1: 335-350.
Munger, J. W.; Collett, J., Jr.; Daube, B., Jr.; Hoffmann, M. R. (1990) Fogwater chemistry at Riverside, California.
Atmos. Environ. Part B 24: 185-205.
Mylonas, D. T.; Allen, D. T.; Ehrman, S. H.; Pratsinis, S. E. (1991) The sources and size distributions of organonitrates
in Los Angeles aerosol. Atmos. Environ. Part A 25: 2855-2861.
Nair, S. K.; Peters, L. K. (1989) Studies on non-precipitating cumulus cloud acidification. Atmos. Environ.
23: 1399-1423.
National Research Council. (1979) Airborne particles. Baltimore, MD: University Park Press.
National Research Council. (1983) Poly cyclic aromatic hydrocarbons: evaluation of sources and effects. Washington,
DC: National Academy Press.
National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
National Academy Press. 3v.
Newman, L. (1981) Atmospheric oxidation of sulfur dioxide: a review as viewed from power plant and smelter plume
studies. Atmos. Environ. 15: 2231-2239.
Newman, L.; Forrest, J.; Manowitz, B. (1975a) 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.
Newman, L.; Forrest, J.; Manowitz, B. (1975b) The application of an isotopic ratio technique to a study of the
atmospheric oxidation of sulfur dioxide in the plume from a coal fired power plant. Atmos. Environ.
9: 969-977.
3-222
-------
Nicholson, K. W. (1988) The dry deposition of small particles: a review of experimental measurements. Atmos.
Envrron. 22: 2653-2666.
Nielsen, T. (1984) Reactivity of poly cyclic aromatic hydrocarbons toward nitrating species. Environ. Sci. Technol. 18:
157-163.
Nielsen, T.; Ramdahl, T.; Bj0rseth, A. (1983) The fate of airborne polycyclic organic matter. Environ. Health Perspect.
47: 103-114.
Noone, K. I; Ogren, J. A.; Hallberg, A.; Hansson, H.-C.; Wiedensohler, A.; Swietlicki, E. (1992a) A statistical
examination of the chemical differences between interstitial and scavenged aerosol. Tellus Ser. B 44B:
581-592.
Noone, K. J.; Ogren, J. A.; Hallberg, A.; Heintzenberg, J.; Strom, J.; Hansson, H.-C.; Svenningsson, B.; Wiedensohler,
A.; Fuzzi, S.; Facchini, M. C.; Arends, B. G.; Bemer, A. (1992b) Changes in aerosol size- and phase
distributions due to physical and chemical processes in fog. Tellus Ser. B 44B: 489-504.
Novakov, T. (1984) The role of soot and primary oxidants in the atmospheric chemistry. Sci. Total Environ. 36: 1-10.
Novakov, T.; Penner, J. E. (1993) Large contribution of organic aerosols to cloud-condensation-nuclei concentrations.
Nature (London) 365: 823-826.
Noxon, J. F. (1983) NO3 and NO2in the mid-Pacific troposphere. J. Geophys. Res. C: Oceans Atmos.
88: 11,017-11,021.
Nunes, T. V.; Pio, C. A. (1993) Carbonaceous aerosols in industrial and coastal atmospheres. Atmos. Environ. Part A
27: 1339-1346.
Oberdorster, G. (1995) Lung particle overload: implications for occupational exposures to particles. Regul. Toxicol.
Pharmacol. 27: 123-135.
O'Brien, R. J.; Crabtree, J. H.; Holmes, J. R.; Hoggan, M. C.; Bockian, A. H. (1975) Formation of photochemical
aerosol from hydrocarbons: atmospheric analysis. Environ. Sci. Technol. 9: 577-582.
O'Dell, R. A.; Taheri, M.; Kabel, R. L. (1977) A model for uptake of pollutants by vegetation. J. Air Pollut. Control
Assoc. 27: 1104-1109.
Ogren, J. A.; Charlson, R. J. (1992) Implications for models and measurements of chemical inhomogeneities among
cloud droplets. Tellus Ser. B 44B: 208-225.
Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pritchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett, R. L.
(1988) Compositions of particles from selected sources in Philadelphia for receptor modeling applications.
JAPCA38: 1392-1402.
Olson, T. M.; Hoffman, M. R. (1989) Hydroxyalkylsulfonate formation: its role as a S(IV) reservoir in atmospheric
water droplets. Atmos. Environ. 23: 985-997.
Otani, Y.; Wang, C. S. (1984) Growth and deposition of saline droplets covered with a monolayer of surfactant.
Aerosol Sci. Technol. 3: 155-166.
Pack, D. H.; Ferber, G. I; Heffter, J. L.; Telegadas, K.; Angell, J. K.; Hoecker, W. H.; Machta, L. (1978) Meteorology
of long-range transport. Atmos. Environ. 12: 425-444.
3-223
-------
Palen, E. I; Allen, D. T.; Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1992) Fourier transform infrared
analysis of aerosol formed in the photo-oxidation of isoprene and p-pinene. Atmos. Environ. Part A 26:
1239-1251.
Pandis, S. N.; Seinfeld, J. H. (1989a) Sensitivity analysis of a chemical mechanism for aqueous-phase atmospheric
chemistry. J. Geophys. Res. [Atmos.] 94: 1105-1126.
Pandis, S. N.; Seinfeld, J. H. (1989b) Mathematical modeling of acid deposition due to radiation fog. J. Geophys. Res.
[Atmos.] 94: 12,911-12,923.
Pandis, S. N.; Seinfeld, J. H.; Pilinis, C. (1990a) The smog-fog-smog cycle and acid deposition. J. Geophys. Res.
[Atmos.] 95: 18,489-18,500.
Pandis, S. N.; Seinfeld, J. H.; Pilinis, C. (1990b) Chemical composition differences in fog and cloud droplets of
different sizes. Atmos. Environ. Part A 24: 1957-1969.
Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of isoprene
and P-pinene. Atmos. Environ. Part A 25: 997-1008.
Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992a) Secondary organic aerosol formation and transport.
Atmos. Environ. Part A 26: 2269-2282.
Pandis, S. N.; Seinfeld, J. H.; Pilinis, C. (1992b) Heterogeneous sulfate production in an urban fog. Atmos. Environ.
Part A 26: 2509-2522.
Pandis, S. N.; Wexler, A. S.; Seinfeld, J. H. (1993) Secondary organic aerosol formation and transport—II. predicting
the ambient secondary organic aerosol size distribution. Atmos. Environ. Part A 27: 2403-2416.
Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
particulate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
Pankow, J. F. (1988) Gas phase retention volume behavior of organic compounds on the sorbent
poly(oxy-w-terphenyl-2'-5'-ylene). Anal. Chem. 60: 950-958.
Pankow, J. F. (1989) Overview of the gas phase retention volume behavior of organic compounds on polyurethane
foam. Atmos. Environ. 23: 1107-1111.
Pankow, J. F. (1991) Commons-intercept and single compound regressions of gas-particle partitioning data vs. l/T.
Atmos. Environ. Part A 25: 2229-2239.
Pankow, J. F. (1994a) An absorption model of gas/particle partitioning of organic compounds in the atmosphere.
Atmos. Environ. 28: 185-188.
Pankow, J. F. (1994b) An absorption model of the gas/aerosol partitioning involved in the formation of secondary
organic aerosol. Atmos. Environ. 28: 189-193.
Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
Pankow, J. F.; Bidleman, T. F. (1992) Interdependence of the slopes and intercepts from log-log correlations of
measured gas-particle partitioning and vapor pressure -1. theory and analysis of available data. Atmos Environ.
Part A 26: 1071-1080.
Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
semivolatile organic compounds to urban particulate matter. Environ. Sci. Technol. 27: 2220-2226.
3-224
-------
Fairish, D. D.; Fahey, D. W.; Williams, E. J.; Liu, S. C.; Trainer, M; Murphy, P. C.; Albritton, D. L.; Fehsenfeld, F. C.
(1986) Background ozone and anthropogenic ozone enhancement at Niwot Ridge, Colorado. J. Atmos. Chem.
4: 63-80.
Pasquill, F. (1974) Atmospheric diffusion. New York, NY: John Wiley and Sons.
Pasquill, F.; Smith, F. B. (1983) Atmospheric diffusion. 3rd ed. New York, NY: John Wiley.
Pastuszka, J. S.; Kwapulinski, J. (1988) The change in mass size distribution of aerosol near dumps as a result of
resuspension of dust. Presented at: 81st annual meeting of Air Pollution Control Association; June; Dallas, TX.
Pittsburgh, PA: Air Pollution Control Association; paper no. 88-151.3.
Patrinos, A. A. N. (1985) The impact of urban and industrial emissions on mesoscale precipitation quality. J. Air Pollut.
Control Assoc. 35:719-727.
Patrinos, A. A. N.; Brown, R. M. (1984) Mesoscale wetfall chemistry around Philadelphia during frontal storms.
Geophys. Res. Lett. 11: 561-564.
Patrinos, A. A. N; Leach, M. J.; Brown, R. M.; Tanner, R. L.; Binkowski, F. S. (1989) An acid rain study in the
Washington, D.C. area. J. Appl. Meteorol. 28: 948-968.
Patterson, E. M.; Gillette, D. A. (1977a) Commonalities in measured size distributions for aerosols having a
soil-derived component. J. Geophys. Res. 82: 2074-2082.
Patterson, E. M.; Gillette, D. A. (1977b) Measurements of visibility vs. mass-concentration for airborne soil particles.
Atmos. Environ. 11: 193-196.
Patterson, R. K.; Wagman, J. (1977) Mass and composition of an urban aerosol as a function of particle size for several
visibility levels. J. Aerosol Sci. 8: 269-279.
Peirson, D. H.; Cawse, P. A.; Salmon, L.; Cambray, R. S. (1973) Trace elements in the atmospheric environment.
Nature (London) 241: 252-256.
Penkett, S. A. (1972) Oxidation of SO2 and other atmospheric gases by ozone in aqueous solution. Nature (London)
240: 105-106.
Penkett, S. A.; Jones, B. M. R.; Brice, K. A.; Eggleton, A. E. J. (1979) The importance of atmospheric ozone and
hydrogen peroxide in oxidising sulphur dioxide in cloud and rainwater. Atmos. Environ. 13: 123-137.
Penner, J. E.; Eddleman, H.; Novakov, T. (1993) Towards the development of a global inventory for black carbon
emissions. Atmos. Environ. Part A 27: 1277-1295.
Peters, K.; Eiden, R. (1992) Modelling the dry deposition velocity of aerosol particles to a spruce forest. Atmos.
Environ. Part A 26: 2555-2564.
Peterson, J. T.; Junge, C. E. (1971) Sources of particulate matter in the atmosphere. In: Matthews, W. H.; Kellogg, W.
W.; Robinson, G. D., eds. Man's impact on the climate. Cambridge, MA: MIT Press.
Pickle, T.; Allen, D. T.; Pratsinis, S. E. (1990) The sources and size distributions of aliphatic and carbonyl carbon in
Los Angeles aerosol. Atmos. Environ. Part A 24: 2221-2228.
Pielke, R. A.; et al. (1992) A comprehensive mesoscale modeling system - RAMS. Meteorol. Atmos. Phys. 49: 69-91.
Pierson, W. R.; Brachaczek, W. W. (1983) Particulate matter associated with vehicles on the road. II. Aerosol Sci.
Technol. 2: 1-40.
3-225
-------
Pierson, W. R.; Brachaczek, W. W. (1990) Dew chemistry and acid deposition in Glendora, California, during the 1986
Carbonaceous Species Methods Comparison Study. Aerosol Sci. Technol. 12: 8-27.
Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M; Norbeck, J. M. (1986) On the acidity of dew. J.
Geophys. Res. [Atmos.] 91: 4083-4096.
Pierson, W. R.; Brachaczek, W. W.; Japar, S. M.; Cass, G. R.; Solomon, P. A. (1988) Dry deposition and dew
chemistry in Claremont, California, during the 1985 Nitrogen Species Methods Comparison Study. Atmos.
Environ. 22: 1657-1663.
Pilinis, C.; Seinfeld, J. H. (1987) Continued development of a general equilibrium model for inorganic multicomponent
atmospheric aerosols. Atmos. Environ. 21: 2453-2466.
Pilinis, C.; Seinfeld, J. H. (1988) Development and evaluation of an Eulerian photochemical gas-aerosol model. Atmos.
Environ. 22: 1985-2001.
Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ. 23: 1601-1606.
Pinnick, R. G.; Fernandez, G.; Hinds, B. D.; Bruce, C. W.; Schaefer, R. W.; Pendleton, J. D. (1985) Dust generated by
vehicular traffic on unpaved roadways: sizes and infrared extinction characteristics. Aerosol Sci. Technol. 4:
99-121.
Pinnick, R. G.; Fernandez, G.; Martinez-Andazola, E.; Hinds, B. D.; Hansen, A. D. A.; Fuller, K. (1993) Aerosol in the
arid southwestern United States: measurements of mass loading, volatility, size distribution, absorption
characteristics, black carbon content, and vertical structure to 7 km above sea level. J. Geophys. Res. [Atmos.]
98: 2651-2666.
Pires, M.; Rossi, M. J.; Ross, D. S. (1994) Kinetic and mechanistic aspects of the NO oxidation by O2 in aqueous phase.
Int. J. Chem. Kmet. 26: 1207-1227.
Pitchford, M. L.; McMurry, P. H. (1994) Relationship between measured water vapor growth and chemistry of
atmospheric aerosol for Grand Canyon, Arizona, in winter 1990. Atmos. Environ. 28: 827-839.
Pitts, J. N, Jr.; Atkinson, R.; Sweetman, J. A.; Zielinska, B. (1985a) The gas-phase reaction of naphthalene with N2O5
to form nitronaphthalenes. Atmos. Environ. 19: 701-705.
Pitts, J. N., Jr.; Zielinska, B.; Sweetman, J. A.; Atkinson, R.; Winer, A. M. (1985b) Reaction of adsorbed pyrene and
perylene with gaseous N2O5 under simulated atmospheric conditions. Atmos. Environ. 19: 911-915.
Pitts, J. N, Jr.; Paur, H.-R.; Zielinska, B.; Arey, J.; Winer, A. M.; Ramdahl, T.; Mejia, V. (1986) Factors influencing the
reactivity of poly cyclic aromatic hydrocarbons adsorbed on filters and ambient POM with ozone.
Chemosphere 15: 675-685.
Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition processes. In:
Irving, P. M., ed. Acidic depostion: state of science and technology, volume I: emissions, atmospheric
processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment Program. (State
of science and technology report no. 1).
Pleim, J.; Venkatram, A.; Yamartino, R. J. (1984) The dry deposition model, v. 4. Rexdale, ON, Canada: Ontario
Ministry of the Environment, ADOM/TADAP Model Development Program.
Post, J. E.; Buseck, P. R. (1985) Quantitative energy-dispersive analysis of lead halide particles from the Phoenix urban
aerosol. Environ. Sci. Technol. 19: 682-685.
3-226
-------
Pratsinis, S.; Ellis, E. C.; Novakov, T.; Friedlander, S. K. (1984) The carbon containing component of the Los Angeles
aerosol: source apportionment and contributions to the visibility budget. J. Air Pollut. Control Assoc. 34:
643-650.
Pratsinis, S. E.; Zeldin, M. D.; Ellis, E. C. (1988) Source resolution of the fine carbonaceous aerosol by principal
component-stepwise regression analysis. Environ. Sci. Technol. 22: 212-216.
Pruppacher, H. R.; Klett, J. D. (1980) Microphysics of clouds and precipitation. Boston, MA: D. Reidel Publishing
Company.
Pueschel, R. F.; Van Valin, C. C.; Castillo, R. C.; Kadlecek, J. A.; Ganor, E. (1986) Aerosols in polluted versus
nonpolluted air masses: long-range transport and effects on clouds. J. Clim. Appl. Meteorol. 25: 1908-1917.
Rabano, E. S.; Castillo, N. T.; Torre, K. J.; Solomon, P. A. (1989) Speciation of arsenic in ambient aerosols collected in
Los Angeles. JAPCA 39: 76-80.
Rader, D. J.; McMurry, P. H. (1986) Application of the tandem differential mobility analyzer to studies of droplet
growth or evaporation. J. Aerosol Sci. 17: 771-787.
Radke, L. F.; Hobbs, P. V. (1991) Humidity and particle fields around some small cumulus clouds. J. Atmos. Sci. 48:
1190-1193.
Radke, L. F.; Brock, C. A.; Lyons, J. H.; Hobbs, P. V.; Schnell, R. C. (1989) Aerosol and lidar measurements of hazes
in mid-latitude and polar airmasses. Atmos. Environ. 23: 2417-2430.
Raga, G. B.; Jonas, P. R. (1993) On the link between cloud-top radiative properties and sub-cloud aerosol
concentrations. Q. J. R. Meteorol. Soc. 119: 1419-1425.
Ramdahl, T. (1983) Poly cyclic aromatic ketones in environmental samples. Environ. Sci. Technol. 17: 666-670.
Ramdahl, T.; Becher, G.; Bjrs0eth, A. (1982) Nitrated poly cyclic aromatic hydrocarbons in urban air particles. Environ.
Sci. Technol. 16: 861-865.
Ramdahl, T.; Bj0rseth, A.; Lokensgard, D.; Pitts, J. N., Jr. (1984) Nitration of polycyclic aromatic hydrocarbons
adsorbed to different carriers in a fluidized bed reactor. Chemosphere 13: 527-534.
Ranade, M. B. (Arun); Woods, M. C.; Chen, F.-L.; Purdue, L. J.; Rehme, K. A. (1990) Wind tunnel evaluation of PM10
samplers. Aerosol Sci. Technol. 13: 54-71.
Rao, A. K.; Whitby, K. T. (1978) Non-ideal collection characteristics of inertial impactors—I. single-stage impactors
and solid particles. J. Aerosol Sci. 9: 77-86.
Rau, J. A. (1989) Composition and size distribution of residential wood smoke particles. Aerosol Sci. Technol. 10:
181-192.
Reeks, M. W.; Reed, J.; Hall, D. (1988) On the resuspension of small particles by a turbulent flow. J. Phys. D: Appl.
Phys. 21: 574-589.
Richards, L. W. (1983) Comments on the oxidation of NO2 to nitrate - day and night. Atmos. Environ. 17: 397-402.
Richards, L. W.; Anderson, J. A.; Blumenthal, D. L.; Brandt, A. A.; McDonald, J. A.; Waters, N; Macias, E. S.;
Bhardwaja, P. S. (1981) The chemistry, aerosol physics, and optical properties of a western coal-fired power
plant plume. Atmos. Environ. 15: 2111-2134.
3-227
-------
Richards, L. W.; Anderson, J. A.; Blumenthal, D. L.; McDonald, J. A. (1982a) Nitrogen and sulfur chemistry and
aerosol formation in plumes. Presented at: 75th annual meeting of the Air Pollution Control Association; June;
New Orleans, LA. Pittsburgh, PA: Air Pollution Control Association; paper no. 82-24.6.
Richards, L. W.; Anderson, J. A.; Blumenthal, D. L.; McDonald, J. A. (1982b) Comparison of the optical properties of
particles in smelter and western power plant plumes [poster presentation]. Presented at: 75th annual meeting of
the Air Pollution Control Association; June; New Orleans, LA. Pittsburgh, PA: Air Pollution Control
Association; poster presentation no. 82-7P.5
Richards, L. W.; Blumenthal, D. L.; Bergstrom, R. W.; Seigneur, C.; Smith, T. B.; Macias, E. S.; Hermg, S. V.;
Bhardwaja, P. S.; Wilson, W. E. (1982c) VISTTA authors' reply. Atmos. Environ. 16: 2363-2365.
Richards, L. W.; Anderson, J. A.; Blumenthal, D. L.; McDonald, J. A.; Macias, E. S.; Hering, S. V.; Wilson, W. E., Jr.
(1985) Chemical, aerosol and optical measurements in the plumes of three midwestern coal-fired power plants.
Atmos. Environ. 19: 1685-1704.
Richards, L. W.; Blanchard, C. L.; Blumenthal, D. L. (1991) Navajo generating station visibility study, draft #2. Santa
Rosa, CA: Sonoma Technology, Inc.; final report STI-90200-1124-FRD2.
Richardson, C. B.; Hightower, R. L. (1987) Evaporation of ammonium nitrate particles. Atmos. Environ. 21: 971-975.
Richardson, C. B.; Spann, J. F. (1984) Measurement of the water cycle in a levitated ammonium sulfate particle. J.
Aerosol Sci. 15: 563-571.
Roberts, D.; Williams, D. (1979) The kinetics of oxidation of sulphur dioxide within the plume from a sulphide smelter
in a remote region. Atmos. Environ. 13: 1485-1499.
Roelofs, G. J. H. (1992a) Drop size dependent sulfate distribution in a growing cloud. J. Atmos. Chem. 14: 109-118.
Roelofs, G. J. H. (1992b) On the drop and aerosol size dependence of aqueous sulfate formation in a continental
cumulus cloud. Atmos. Environ. Part A 26: 2309-2321.
Roelofs, G. J. H. (1993) A cloud chemistry sensitivity study and comparison of explicit and bulk cloud model
performance. Atmos. Environ. Part A 27: 2255-2264.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci. Technol.
27:636-651.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
Sci. Technol. 27: 1892-1904.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol. 27:
2700-2711.
Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993d) Quantification of urban
organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part A
27: 1309-1330.
3-228
-------
Rogge, W. F.; Hildemann, L. M; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
Rolph, G. D.; Draxler, R. R. (1990) Sensitivity of three-dimensional trajectories to the spatial and temporal densities of
the wind field. J. Appl. Meteorol. 29: 1043-1054.
Rood, M. J.; Larson, T. V.; Covert, D. S.; Ahlquist, N. C. (1985) Measurement of laboratory and ambient aerosols with
temperature and humidity controlled nephelometry. Atmos. Environ. 19: 1181-1190.
Rood, M. J.; Shaw, M. A.; Larson, T. V.; Covert, D. S. (1989) Ubiquitous nature of ambient metastable aerosol. Nature
(London) 337: 537-539.
Rounds, S. A.; Pankow, J. F. (1990) Application of a radial diffusion model to describe gas/particle sorption kinetics.
Environ. Sci. Technol. 24: 1378-1386.
Rubel, G. O.; Gentry, J. W. (1984) Measurement of the kinetics of solution droplets in the presence of adsorbed
monolayers: determination of water accommodation coefficients. J. Phys. Chem. 88: 3142-3148.
Sakugawa, H.; Kaplan, I. R.; Tsai, W.; Cohen, Y. (1990) Atmospheric hydrogen peroxide: does it share a role with
ozone in degrading air quality? Environ. Sci. Technol. 24: 1452-1462.
Satsumabayashi, H.; Kurita, H.; Yokouchi, Y.; Ueda, H. (1989) Mono- and di-carboxylic acids under long-range
transport of air pollution in central Japan. Tellus Ser. B 4IB: 219-229.
Satsumakayashi, H.; Kurita, H.; Yokouchi, Y.; Ueda, H. (1990) Photochemical formation of particulate dicarboxylic
acids under long-range transport in central Japan. Atmos. Environ. Part A 24: 1443-1450.
Savoie, D. L.; Prospero, J. M.; Nees, R. T. (1987) Washout ratios of nitrate, non-sea-salt sulfate and sea-salt on Virginia
Key, Florida and on American Samoa. Atmos. Environ. 21: 103-112.
Saxena, P. (1994) Authors' response to comments on "Coupling thermodynamic theory with measurements to
characterize acidity of atmospheric particles." Aerosol Sci. Technol. 20: 287-288.
Saxena, V. K.; Hendler, A. H. (1983) In-cloud scavenging and resuspension of cloud active aerosols during winter
storms over Lake Michigan. In: Pruppacher, H. R.; Semonin, R. G.; Slinn, W. G. N., eds. Precipitation
scavenging, dry deposition, and resuspension, v. 1: precipitation scavenging: proceedings of the fourth
international conference; November-December 1982; Santa Monica, CA. New York, NY: Elsevier; pp. 91-101.
Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the literature
and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem.: accepted.
Saxena, P.; Seigneur, C. (1987) On the oxidation of SO2 to sulfate in atmospheric aerosols. Atmos. Environ.
21: 807-812.
Saxena, P.; Mueller, P. K.; Kim, Y. P.; Seinfeld, J. H.; Koutrakis, P. (1993) Coupling thermodynamic theory with
measurements to characterize acidity of atmospheric particles. Aerosol Sci. Technol. 19: 279-293.
Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
atmospheric particles. J. Geophys. Res. [Atmos.]: in press.
Saylor, R. D. (1989) A study of mesoscale acid deposition in the lower Ohio River Valley [Ph.D. dissertation].
Lexington, KY: University of Kentucky, Chemical Engineering Department.
Schlesinger, R. B. (1994) Comments on "Coupling thermodynamic theory with measurements to characterize acidity of
atmospheric particles" by P. Saxena et al. Aerosol Sci. Technol. 20: 285-286.
3-229
-------
Schmidt, J. A.; Andren, A. W. (1980) The atmospheric chemistry of nickel. In: Nriagu, J. O., ed. Nickel in the
environment. New York, NY: John Wiley & Sons, Inc.; pp. 93-135.
Schroeder, W. H.; Dobson, M.; Kane, D. M.; Johnson, N. D. (1987) Toxic trace elements associated with airborne
particulate matter: a review. JAPCA 37: 1267-1285.
Schuetzle, D.; Cronn, D.; Crittenden, A. L.; Charlson, R. J. (1975) Molecular composition of secondary aerosol and its
possible origin. Environ. Sci. Technol. 9: 838-845.
Schwartz, S. E. (1979) Residence times in reservoirs under non-steady-state conditions: application to atmospheric SO2
and aerosol sulfate. Tellus 31: 530-547.
Schwartz, S. E. (1984a) Gas-aqueous reactions of sulfur and nitrogen oxides in liquid-water clouds. In: Calvert, J. G.,
ed. SO2, NO and NO2 oxidation mechanisms: atmospheric considerations. Boston, MA: Butterworth
Publishers; pp. 173-208. (Teasley, J. I., ed. Acid precipitation series: v. 3).
Schwartz, S. E. (1984b) Gas- and aqueous-phase chemistry of HO2 in liquid water clouds. J. Geophys. Res. [Atmos.]
89: 11,589-11,598.
Schwartz, S. E. (1986a) Chemical conversions in clouds. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P., eds.
Aerosols: research, risk assessment and control strategies, proceedings of the second U.S.-Dutch international
symposium; May 1985; Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 349-375.
Schwartz, S. E. (1986b) Mass-transport considerations pertinent to aqueous-phase reactions of gases in liquid-water
clouds. In: Jaeschke, W., ed. Chemistry of multiphase atmospheric systems. Berlin, Federal Republic of
Germany: Springer-Verlag; pp. 415-471. (NATO advanced science institutes series G: ecological sciences v.
6).
Schwartz, S. E. (1987) Both sides now: the chemistry of clouds. Ann. N. Y. Acad. Sci. 502: 83-144.
Schwartz, S. E. (1988) Mass-transport limitation to the rate of in-cloud oxidation of SO2: re-examination in the light of
new data. Atmos. Environ. 22: 2491-2499.
Schwartz, S. E. (1989) Acid deposition: unraveling a regional phenomenon. Science (Washington, DC) 243: 753-763.
Schwartz, S. E.; Freiberg, J. E. (1981) Mass-transport limitation to the rate of reaction of gases in liquid droplets:
application to oxidation of SO2 in aqueous solutions. Atmos. Environ. 15:1129-1144.
Schwartz, S. E.; Newman, L. (1978) Processes limiting oxidation of sulfur dioxide in stack plumes. Environ. Sci.
Technol. 12: 67-73.
Schwartz, S. E.; Newman, L. (1983) Measurements of sulfate production in natural clouds [discussion of Hegg and
Hobbs (1982)]. Atmos. Environ. 17: 2629-2632.
Schwartz, S. E.; White, W. H. (1983) Kinetics of reactive dissolution of nitrogen oxides into aqueous solution. In:
Schwartz, S. E., ed. Trace atmospheric constituents: properties, transformations, and fates. New York, NY:
John Wiley & Sons; pp. 1-116. (Advances in environmental science and technology: v. 12).
Scott, B. C.; Laulainen, N. S. (1979) On the concentration of sulfate in precipitation. J. Appl. Meteorol. 18: 138-147.
Sega, K.; Fugas, M. (1984) Seasonal and spatial differences in mass concentration levels and particle size distribution of
aerosols over an urban area. Atmos. Environ. 18: 2433-2437.
Sehmel, G. A. (1973) Particle resuspension from an asphalt road caused by car and truck traffic. Atmos. Environ. 7:
291-309.
3-230
-------
Sehmel, G. A. (1980) Particle and gas dry deposition: a review. Atmos. Environ. 14: 983-1011.
Sehmel, G. A. (1984) Deposition and resuspension. In: Randerson, D., ed. Atmospheric science and power production.
Washington, DC: U.S. Department of Energy, Office of Scientific and Technical Information; report no.
DOE/TIC-27601; pp. 533-583. Available from: NTIS, Springfield, VA; DE84-005177.
Seidl, W. (1989) Ionic concentrations and initial S(IV)-oxidation rates in droplets during the condensational stage of
cloud. Tellus Ser. B 4IB: 32-50.
Seigneur, C. (1982) A model of sulfate aerosol dynamics in atmospheric plumes. Atmos. Environ. 16: 2207-2228.
Seigneur, C.; Constantinou, E. (1995) Chemical kinetic mechanism for atmospheric chromium. Environ. Sci. Technol.
29: 222-231.
Seigneur, C.; Saxena, P. (1984) A study of atmospheric acid formation in different environments. Atmos. Environ. 18:
2109-2124.
Seigneur, C.; Saxena, P. (1988) A theoretical investigation of sulfate formation in clouds. Atmos. Environ.
22: 101-115.
Seigneur, C.; Wegrecki, A. M. (1990) Mathematical modeling of cloud chemistry in the Los Angeles basin. Atmos.
Environ. Part A 24: 989-1006.
Seigneur, C.; Saxena, P.; Roth, P. M. (1984) Computer simulations of the atmospheric chemistry of sulfate and nitrate
formation. Science (Washington, DC) 225: 1028-1030.
Seinfeld, J. H.; Flagan, R. C.; Petti, T. B.; Stern, J. E.; Grosjean, D. (1987) Aerosol formation in aromatic
hydrocarbon-NOx systems. Pasadena, CA: California Institute of Technology; project no. AP-6.
Sergides, C. A.; Jassim, J. A.; Chughtai, A. R.; Smith, D. M. (1987) The structure of hexane soot. Part III: Ozonation
studies. Appl. Spectrosc. 41: 482-492.
Shah, J. J.; Johnson, R. L.; Heyerdahl, E. K.; Huntzicker, J. J. (1986) Carbonaceous aerosol at urban and rural sites in
the United States. J. AirPollut. Control Assoc. 36: 254-257.
Shao, Y.; Raupach, M. R.; Findlater, P. A. (1993) Effect of saltation bombardment on the entrainment of dust by wind.
J. Geophys. Res. [Atmos.] 98: 12,719-12,726.
Shaw, M. A.; Rood, M. J. (1990) Measurement of the crystallization humidities of ambient aerosol particles. Atmos.
Environ. Part A 24: 1837-1841.
Sheridan, P. J.; Schnell, R. C.; Kahl, J. D.; Boatman, J. F.; Garvey, D. M. (1993) Microanalysis of the aerosol collected
over south-central New Mexico during the ALIVE field experiment, May-December 1989. Atmos. Environ.
Part A 27: 1169-1183.
Shinn, J. H.; Homan, D. N.; Gay, D. D. (1983) Plutonium aerosol fluxes and pulmonary exposure rates during
resuspension from bare soils near a chemical separation facility. In: Pruppacher, H. R.; Semonin, R. G.; Slinn,
W. G. N., eds. Precipitation scavenging, dry deposition, and resuspension, v. 2: dry deposition and
resuspension: proceedings of the fourth international conference; November-December 1982; Santa Monica,
CA. New York, NY: Elsevier; pp. 1131-1143.
Sievering, H.; Van Valin, C. C.; Barrett, E. W.; Pueschel, R. F. (1984) Cloud scavenging of aerosol sulfur: two case
studies. Atmos. Environ. 18: 2685-2690.
3-231
-------
Sievering, H.; Boatman, J.; Galloway, J.; Keene, W.; Kim, Y.; Luria, M.; Ray, J. (1991) Heterogeneous sulfur
conversion in sea-salt aerosol particles: the role of aerosol water content and size distribution. Atmos. Environ.
Part A 25: 1479-1487.
Sievering, H.; Gorman, E.; Kim, Y.; Ley, T.; Seidl, W.; Boatman, J. (1994) Heterogeneous conversion contribution to
the sulfate observed over Lake Michigan. Atmos. Environ. 28: 367-370.
Simoneit, B. R. T. (1984) Organic matter of the troposphere—III. characterization and sources of petroleum and
pyrogenic residues in aerosols over the western United States. Atmos. Environ. 18: 51-67.
Simoneit, B. R. T. (1986) Characterization of organic constituents in aerosols in relation to their origin and transport: a
review. Presented at: the 2nd workshop on chemistry analysis of hydrocarbons in the environment; November
1985; Barcelona, Spain. Int. J. Environ. Anal. Chem. 23: 207-237.
Simoneit, B. R. T. (1989) Organic matter of the troposphere - V: application of molecular marker analysis to biogenic
emissions into the troposphere for source reconciliations. Presented at: the international symposium on
biosphere-atmosphere exchange; March 1986; Mainz, Federal Republic of Germany. J. Atmos. Chem. 8:
251-275.
Simoneit, B. R. T.; Mazurek, M. A. (1982) Organic matter of the troposphere—II. natural background of biogenic lipid
matter in aerosols over the rural western United States. Atmos. Environ. 16: 2139-2159.
Sisler, J. F.; Malm, W. C. (1994) The relative importance of soluble aerosols to spatial and seasonal trends of impaired
visibility in the United States. Atmos. Environ. 28: 851-862.
Sisterson, D. L.; Bowersox, V. C.; Olsen, A. R.; Meyers, T. P.; Olsen, A. R.; Vong, R. L. (1991) Deposition
monitoring: methods and results. In: Irving, P. M., ed. Acidic deposition: state of science and technology,
volume I: emissions, atmospheric processes, and depostition. Washington, DC: The U.S. National Acid
Precipitation Assessment Program. (State of science and technology report no. 6).
Slinn, W. G. N (1983) Air-to-sea transfer of particles. In: Liss, P. S.; Slinn, W. G. N, eds. Air-sea exchange of gases
and particles. Boston, MA: D. Reidel Publishing Company; pp. 299-405. (NATO ASI series. Series C,
mathematical and physical sciences; no. 108).
Sloane, C. S. (1982) Visibility trends—II. mideastern United States 1948-1978. Atmos. Environ. 16: 2309-2321.
Sloane, C. S.; Groblicki, P. J. (1981) Denver's visibility history. Atmos. Environ. 15: 2631-2638.
Sloane, C. S.; Watson, J.; Chow, I; Pritchett, L.; Richards, L. W. (1991) Size-segregated fine particle measurements by
chemical species and their impact on visibility impairment in Denver. Atmos. Environ. Part A 25: 1013-1024.
Smith, T. B. (1981) Some observations of pollutant transport associated with elevated plumes. Atmos. Environ. 15:
2197-2203.
Smith, D. M.; Akhter, M. S.; Jassim, J. A.; Sergides, C. A.; Welch, W. F.; Chughtai, A. R. (1989) Studies of the
structure and reactivity of soot. Aerosol Sci. Technol. 10: 311-325.
Snider, J. R.; Vali, G. (1994) Sulfur dioxide oxidation in winter orographic clouds. J. Geophys. Res. [Atmos.] 99:
18,713-18,733.
Spann, J. F.; Richardson, C. B. (1985) Measurement of the water cycle in mixed ammonium acid sulfate particles.
Atmos. Environ. 19: 819-825.
3-232
-------
Spengler, J. D.; Keeler, G. J.; Koutrakis, P.; Ryan, P. B.; Raizenne, M; Franklin, C. A. (1989) Exposures to acidic
aerosols. In: Symposium on the health effects of acid aerosols; October 1987; Research Triangle Park, NC.
Environ. Health Perspect. 79: 43-51.
Spicer, C. W. (1977a) Photochemical atmospheric pollutants derived from nitrogen oxides. Atmos. Environ.
11: 1089-1095.
Spicer, C. W. (1977b) The fate of nitrogen oxides in the atmosphere. In: Pitts, J. N., Jr.; Metcalf, R. L.; Lloyd, A. C.,
eds. Advances in environmental science and technology: v. 4. New York, NY: John Wiley and Sons, Inc.; pp.
163-261.
Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative humidity-dependent
bounce and density for atmospheric particles using the DMA-impactor technique. Atmos. Environ. 28:
1739-1746.
Steiner, D.; Burtscher, H.; Gross, H. (1992) Structure and disposition of particles from a spark-ignition engine. Atmos.
Environ. Part A 26: 997-1003.
Stelson, A. W.; Seinfeld, J. H. (1981) Chemical mass accounting of urban aerosol. Environ. Sci. Technol. 15: 671-679.
Stelson, A. W.; Seinfeld, J. H. (1982a) Relative humidity and temperature dependence of the ammonium nitrate
dissociation constant. Atmos. Environ. 16: 983-992.
Stelson, A. W.; Seinfeld, J. H. (1982b) Relative humidity and pH dependence of the vapor pressure of ammonium
nitrate-nitric acid solutions at 25 °C. Atmos. Environ. 16: 993-1000.
Stem, J. E.; Flagan, R. C.; Grosjean, D.; Seinfeld, J. H. (1987) Aerosol formation and growth in atmospheric aromatic
hydrocarbon photooxidation. Environ. Sci. Technol. 21: 1224-1231.
Stevens, R. K.; Dzubay, T. G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods applied to the
determination of the origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ. 18:
261-272.
Stewart, D. A.; Liu, M.-K. (1981) Development and application of a reactive plume model. Atmos. Environ.
15: 2377-2393.
Storey, J. M. E.; Pankow, J. F. (1992) Gas-particle partitioning of semi-volatile organic compounds to model
atmospheric particulate materials—I. Sorption to graphite, sodium chloride, alumina, and silica particles under
low humidity conditions. Atmos. Environ. Part A 26: 435-443.
Strapp, J. W.; Leaitch, W. R.; Anlauf, K. G.; Bottenheim, J. W.; Joe, P.; Schemenauer, R. S.; Wiebe, H. A.; Isaac, G. A.;
Kelly, T. J.; Daum, P. H. (1988) Winter cloud water and air composition in Central Ontario. J. Geophys. Res.
[Atmos.] 93: 3760-3772.
Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
Technol. 26: 2507-2517.
Suh, H. H.; Allen, G. A.; Aurian-Blajeni, B.; Koutrakis, P.; Burton, R. M. (1994a) Field method comparison for the
characterization of acid aerosols and gases. Atmos. Environ. 28: 2981-2989.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994b) The relationship between airborne acidity and ammonia in indoor
environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
3-233
-------
Sutherland, A. J. (1967) Proposed mechanism for sediment entrainment by turbulent flows. J. Geophys. Res.
72:6183-6194.
Svenningsson, B.; Hansson, H.-C.; Wiedensohler. A.; Noone, K.; Ogren, J.; Hallberg, A.; Calvile, R. (1994)
Hygroscopic growth of aerosol particles and its influence on nucleation scavenging in cloud: experimental
results from Kleiner Feldberg. J. Atmos. Chem. 19: 129-152.
Sverdrup, G. M.; Whitby, K. T. (1980) The effect of changing relative humidity on aerosol size distribution
measurements. In: Hidy, G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The
character and origins of smog aerosols: a digest of results from the California Aerosol Characterization
Experiment (ACHEX). New York, NY: John Wiley & Sons; pp. 527-558. (Advances in environmental science
and technology: v. 10).
Sviridenkov, M. A.; Gillette, D. A.; Isakov, A. A.; Sokolik, I. N; Smirnov, V. V.; Belan, B. D.; Pachenko, M. V.;
Andronova, A. V.; Kolomiets, S. M.; Zhukov, V. M.; Zhukovsky, D. A. (1993) Size distributions of dust
aerosol measured during the Soviet-American experiment in Tadzhikistan, 1989. Atmos. Environ. Part A 27:
2481-2486.
TM
TSI. (1993) Model 390045 DISTFIT aerosol data fitting software. St. Paul and Golden Valley, MN: TSI, Inc. and
Chimera Software.
Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science; pp.
153-167.
Tang, I. N.; Munkelwitz, H. R. (1977) Aerosol growth studies—III. ammonium bisulfate aerosols in a moist
atmosphere. J. Aerosol Sci. 8: 321-330.
Tang, I. N.; Munkelwitz, H. R. (1993) Composition and temperature dependence of the deliquescence properties of
hygroscopic aerosols. Atmos. Environ. Part A 27: 467-473.
Tang, I. N.; Munkelwitz, H. R. (1994) Water activities, densities, and refractive indices of aqueous sulfates and sodium
nitrate droplets of atmospheric importance. J. Geophys. Res. [Atmos.] 99: 18,801-18,808.
Tang, I. N.; Munkelwitz, H. R.; Davis, J. G. (1978) Aerosol growth studies - IV. phase transformation of mixed salt
aerosols in a moist atmosphere. J. Aerosol Sci. 9: 505-511.
Tang, I. N.; Wong, W. T.; Munkelwitz, H. R. (1981) The relative importance of atmospheric sulfates and nitrates in
visibility reduction. Atmos. Environ. 15: 2463-2471.
Tani, B.; Siegel, S.; Johnson, S. A.; Kumar, R. (1983) X-ray diffraction investigation of atmospheric aerosols in the
0.3-1.0 //m aerodynamic size range. Atmos. Environ. 17: 2277-2283.
Tao, Y.; McMurry, P. H. (1989) Vapor pressures and surface free energies of C14-C18 monocarboxylic acids and C5
and C6 dicarboxylic acids. Environ. Sci. Technol. 23: 1519-1523.
Taylor, G. I. (1922) Diffusion by continuous movements. Proc. London Math. Soc. 20: 196-212.
Taylor, G. R. (1989) Sulfate production and deposition in midlatitude continental cumulus clouds. Part II: Chemistry
model formulation and sensitivity analysis. J. Atmos. Sci. 46: 1991-2007.
Ten Brink, H. M.; Schwartz, S. E.; Daum, P. H. (1987) Efficient scavenging of aerosol sulfate by liquid-water clouds.
Atmos. Environ. 21: 2035-2052.
3-234
-------
Thibodeaux, L. J.; Nadler, K. C.; Valsaraj, K. T.; Reible, D. D. (1991) The effect of moisture on volatile organic
chemical gas-to-particle partitioning with atmospheric aerosols—competitive adsorption theory predictions.
Atmos. Environ. Part A 25: 1649-1656.
Thomas, P.; Vogt, S. (1990) Mesoscale atmospheric dispersion experiments using tracer and tetroons. Atmos. Environ.
Part A 24: 1271-1284.
Thompson, K. M.; Koutrakis, P.; Brauer, M.; Spengler, J. D.; Wilson, W. E.; Burton, R. M. (1991) Measurements of
aerosol acidity: sampling frequency, seasonal variability, and spatial variation. Presented at: 84th annual
meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada.
Pittsburgh, PA: Air & Waste Management Association; paper no. 91-89.5.
Thurston, G. D.; Gorczynski, J. E., Jr.; Jaques, P.; Currie, J.; He, D. (1992) An automated sequential sampling system
for particulate acid aerosols: description, characterization, and field sampling results. J. Exposure Anal.
Environ. Epidemiol. 2: 415-428.
Tuazon, E. C.; Atkinson, R.; Plum, C. N.; Winer, A. M.; Pitts, J. N., Jr. (1983) The reaction of gas phase N2O5 with
water vapor. Geophys. Res. Lett. 10: 953-956.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of primary
and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and elemental
carbon. Aerosol Sci. Technol. 12: 161-171.
Turpin, B. J.; Huntzicker, J. J.; Larson, S. M.; Cass, G. R. (1991) Los Angeles summer midday particulate carbon:
primary and secondary aerosol. Environ. Sci. Technol. 25: 1788-1793.
Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
compounds. Environ. Sci. Technol. 27: 2441-2449.
Twohy, C. H.; Austin, P. H.; Charlson, R. J. (1989) Chemical consequences of the initial diffusional growth of cloud
droplets: a clean marine case. Tellus Ser. B 41B: 51-60.
Twomey, S. (1959) The nuclei of natural cloud formation: part II: the supersaturation in natural clouds and the
variation of cloud droplet concentration. Geofis. Pura Appl. 43: 243-249.
U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
U.S. Environmental Protection Agency. (1986a) Health assessment document for nickel and nickel compounds.
Research Triangle Park, NC: Environmental Criteria and Assessment Office; report no. EPA/600/8-83/012FF.
Available from: NTIS, Springfield, VA; PB86-232212.
U.S. Environmental Protection Agency. (1986b) Air quality criteria for lead. Research Triangle Park, NC: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
3-235
-------
U.S. Environmental Protection Agency. (1995) Compilation of air pollutant emission factors, v. I: stationary point and
area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards; sections 1.1
and 1.2; report no. AP-42.
Uliasz, M. (1993) The atmospheric mesoscale dispersion modeling system. J. Appl. Meteorol. 32: 139-149.
Valerio, F.; Bottino, P.; Ugolini, D.; Cimberle, M. R.; Tozzi, G. A.; Frigerio, A. (1984) Chemical and photochemical
degradation of polycyclic aromatic hydrocarbons in the atmosphere. Sci. Total Environ. 40: 169-188.
Valsaraj, K. T.; Thoma, G. J.; Reible, D. D.; Thibodeaux, L. J. (1993) On the enrichment of hydrophobic organic
compounds in fog droplets. Atmos. Environ. Part A 27: 203-210.
Van Borm, W. A.; Adams, F. C.; Maenhaut, W. (1989) Characterization of individual particles in the Antwerp aerosol.
Atmos. Environ. 23: 1139-1151.
Vanderpool, R. W.; Lundgren, D. A.; Marple, V. A.; Rubow, K. L. (1987) Cocalibration of four large-particle
impactors. Aerosol Sci. Technol. 7: 177-185.
Van Doren, J. M.; Watson, L. R.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. (1990) Temperature
dependence of the uptake coefficients of FTNO3, HC1, and N2O5 by water droplets. J. Phys. Chem. 94:
3265-3272.
Van Grieken, R.; Xhoffer, C. (1992) Microanalysis of individual environmental particles. J. Anal. At. Spectrom. 7:
81-88.
Van Valin, C. C.; Luria, M.; Ray, J. D.; Boatman, J. F. (1990) Hydrogen peroxide and ozone over the northeastern
United States in June 1987. J. Geophys. Res. [Atmos.] 95: 5689-5695.
Venkataraman, C.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons and elemental
carbon. 2. Ambient measurements and effects of atmospheric processes. Environ. Sci. Technol. 28: 563-572.
Venkataraman, C.; Hildemann, L. M. (1994) Factors affecting the distribution of primary semivolatile organic
compounds of urban aerosol particles. Environ. Sci. Technol.: submitted.
Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons and
elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci. Technol. 28:
555-562.
Wagman, J.; Lee, R. E., Jr.; Axt, C. J. (1967) Influence of some atmospheric variables on the concentration and particle
size distribution of sulfate in urban air. Atmos. Environ. 1: 479-489.
Walcek, C. J.; Stockwell, W. R.; Chang, J. S. (1990) Theoretical estimates of the dynamic, radiative and chemical
effects of clouds on tropospheric trace gases. Atmos. Res. 25: 53-69.
Waldman, J. M.; Lioy, P. J.; Thurston, G. D.; Lippmann, M. (1990) Spatial and temporal patterns in summertime
sulfate aerosol acidity and neutralization within a metropolitan area. Atmos. Environ. Part B 24: 115-126.
Waldman, J. M.; Koutrakis, P.; Allen, G. A.; Thurston, G. D.; Burton, R. M.; Wilson, W. E. (1995) Human exposures
to particle strong acidity. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air
pollution and human mortality and morbidity, part II; January 1994; Irvine, CA. Inhalation Toxicol. 7:
657-669.
Wall, S. M.; John, W.; Ondo, J. L. (1988) Measurement of aerosol size distributions for nitrate and major ionic species.
Atmos. Environ. 22: 1649-1656.
3-236
-------
Wang, C.; Chang, J. S. (1993) A three-dimensional numerical model of cloud dynamics, microphysics, and chemistry:
4. cloud chemistry and precipitation chemistry. J. Geophys. Res. [Atmos.] 98: 16,799-16,808.
Wang, H.-C.; John, W. (1987) Comparative bounce properties of particle materials. Aerosol Sci. Technol. 7: 285-299.
Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactor for sampling inorganic ions. Aerosol Sci.
Technol. 8: 157-172.
Wang, S.-C.; Paulson, S. E.; Grosjean, D.; Flagan, R. C.; Seinfeld, J. H. (1992a) Aerosol formation and growth in
atmospheric organic/NOx systems—I. Outdoor smog chamber studies of C7- and C8- hydrocarbons. Atmos.
Environ. Part A 26: 403-420.
Wang, S.-C.; Flagan, R. C.; Seinfeld, J. H. (1992b) Aerosol formation and growth in atmospheric organic/NOx systems
—II. aerosol dynamics. Atmos. Environ. Part A 26: 421-434.
Warner, T. T. (1981) Verification of a three-dimensional transport model using tetroon data from Project STATE.
Atmos. Environ. 15: 2219-2222.
Warren, R. S.; Birch, P. (1987) Heavy metal levels in atmospheric particulates, roadside dust and soil along a major
urban highway. Sci. Total Environ. 59: 253-256.
Watson, J. G.; Chow, J. C. (1992) Data bases for PM10 and PM25 chemical compositions and source profiles.
Transactions AWMA 1:61-91.
Watson, J. G.; Chow, J. C.; Lurmann, F. W.; Musarra, S. P. (1994a) Ammonium nitrate, nitric acid, and ammonia
equilibrium in wintertime Phoenix, Arizona. Air Waste 44: 405-412.
Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994b) Chemical
mass balance source apportionment of PM10 during the Southern California Air Quality Study. Aerosol Sci.
Technol. 21: 1-36.
Weathers, K. C.; Likens, G. E.; Bormann, F. H.; Bicknell, S. H.; Bormann, B. T.; Daube, B. C., Jr.; Eaton, J. S.;
Galloway, J. N.; Keene, W. C.; Kimball, K. D.; McDowell, W. H.; Siccama, T. G.; Smiley, D.; Tarrant, R. A.
(1988) Cloudwater chemistry from ten sites in North America. Environ. Sci. Technol. 22: 1018-1026.
Weber, R. J.; McMurry, P. H.; Eisele, F. L.; Tanner, D. J. (1995) Measurement of expected nucleation precursor species
and 5 to 500 nm diameter particles at Mauna Loa, Hawaii. J. Atmos. Sci. 52: 2242-2257.
Wesely, M. L.; Hicks, B. B. (1977) Some factors that affect the deposition rates of sulfur dioxide and similar gases on
vegetation. J. Air Pollut. Control Assoc. 27: 1110-1116.
Wexler, A. S.; Seinfeld, J. H. (1990) The distribution of ammonium salts among a size and composition dispersed
aerosol. Atmos. Environ. Part A 24: 1231-1246.
Wexler, A. S.; Seinfeld, J. H. (1992) Analysis of aerosol ammonium nitrate: departures from equilibrium during
SCAQS. Atmos. Environ. Part A 26: 579-591.
Wexler, A. S.; Eldermg, A.; Pandis, S. N.; Cass, G. R.; Seinfeld, J. H.; Moon, K. C.; Hermg, S. (1992) Modeling
aerosol processes and visibility based on the SCAQS data. Final report. Sacramento, CA: California Air
Resources Board; contract no. A932-054.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
Whitby, K. T. (1980) Aerosol formation in urban plumes. Ann. N. Y. Acad. Sci. 338: 258-275.
3-237
-------
Whitby, K. T. (1984) Physical and optical behavior of sulfate-nitrate aerosols in equilibrium with atmospheric water
vapor, ammonia and nitric acid. In: Ruhnke, L. H.; Deepak, A., eds. Hygroscopic aerosols: technical
proceedings of the workshop on hygroscopic aerosols in the planetary boundary layer; April 1982; Vail, CO.
Hampton, VA: A. Deepak Publishing; pp. 45-63.
Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy, G. M.;
Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of smog aerosols:
a digest of results from the California Aerosol Characterization Experiment (ACHEX). New York, NY: John
Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and technology: v. 9).
Whitby, K. T.; Husar, R. B.; Liu, B. Y. H. (1972) The aerosol size distribution of Los Angeles smog. J. Colloid
Interface Sci. 39: 177-204.
Whitby, K. T.; Charlson, R. E.; Wilson, W. E.; Stevens, R. K. (1974) The size of suspended particle matter in air.
Science (Washington, DC) 183: 1098-1099.
White, W. H.; Macias, E. S. (1989) Carbonaceous particles and regional haze in the western United States. Aerosol Sci.
Technol. 10: 111-117.
White, W. H.; Anderson, J. A.; Blumenthal, D. L.; Husar, R. B.; Gillam, N. V.; Husar, J. D.; Wilson, W. E., Jr. (1976)
Formation and transport of secondary air pollutants: ozone and aerosols in the St. Louis urban plume. Science
(Washington, DC) 194: 187-189.
White, W. H.; Patterson, D. E.; Wilson, W. E. (1983) Urban exports to the nonurban troposphere: results from Project
MISTT. J. Geophys. Res. 88: 10,745-10,752.
White, W. H.; Seigneur, C.; Hemold, D. W.; Eltgroth, M. W.; Richards, L. W.; Roberts, P. T.; Bhardwaja, P. S.; Conner,
W. D.; Wilson, W. E., Jr. (1985) Predicting the visibility of chimney plumes: an intercomparison of four
models with observations at a well-controlled power plant. Atmos. Environ. 19: 515-528.
Wiedensohler, A.; Aalto, P.; Covert, D.; Heintzenberg, J.; McMurry, P. H. (1994) Intercomparison of four methods to
determine size distributions of low-concentration (~ 100 cm"3), ult:
illustrative data from the Artie. Aerosol Sci. Technol. 21: 95-109.
determine size distributions of low-concentration (~ 100 cm"3), ultrafine aerosols (3 < Dp < 10 nm) with
Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. Air Pollut. Control Assoc. 25:
529-534.
Willetts, B. (1992) [Personal communication]. Aberdeen, Scotland: University of Aberdeen.
Williams, D. J.; Carras, J. N.; Milne, J. W.; Heggie, A. C. (1981) The oxidation and long-range transport of sulphur
dioxide in a remote region. Atmos. Environ. 15: 2255-2262.
Wilson, W. E. (1978) Sulfates in the atmosphere: a progress report on project MISTT. Atmos. Environ. 12: 537-547.
Wilson, W. E., Jr. (1981) Sulfate formation in point source plumes: a review of recent field studies. Atmos. Environ.
15:2573-2581.
Wilson, J. C.; McMurry, P. H. (1981) Studies of aerosol formation in power plant plumes—II. secondary aerosol
formation in the Navajo generating station plume. In: White, W. H.; Moore, D. J.; Lodge, J. P., eds. Plumes and
visibility: measurements and model components: proceedings of the symposium; November 1980; Grand
Canyon National Park, AZ. Atmos. Environ. 15: 2329-2339.
Wilson, W. E.; Suh, H. H. (1996) Fine and coarse particles: concentration relationships relevant to epidemiological
studies. J. Air Waste Manage. Assoc.: accepted.
3-238
-------
Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. I; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher, R.
A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors sulfate
dispersion experiment: summary of EPA measurements. J. AirPollut. Control Assoc. 27: 46-51.
Wilson, J. C.; Gupta, A.; Whitby, K. T.; Wilson, W. E. (1988) Measured aerosol light scattering coefficients compared
with values calculated from EAA and optical particle counter measurements: improving the utility of the
comparison. Atmos. Environ. 22: 789-793.
Winklmayr, W.; Wang, H.-C.; John, W. (1990) Adaptation of the Twomey algorithm to the inversion of cascade
impactor data. Aerosol Sci. Technol. 13: 322-331.
Wolff, G. T. (1981)Particulate elemental carbon in the atmosphere. J. AirPollut. Control Assoc. 31: 935-938.
Wolff, G. T.; Korsog, P. E. (1985) Estimates of the contributions of sources to inhalable particulate concentrations in
Detroit. Atmos. Environ. 19: 1399-1409.
Wolff, G. T.; Lioy, P. J.; Wight, G. D.; Meyers, R. E.; Cederwall, R. T. (1977) An investigation of long-range transport
of ozone across the midwestern and eastern United States. Atmos. Environ. 11: 797-802.
Wolff, G. T.; Groblicki, P. J.; Cadle, S. H.; Countess, R. J. (1982) Particulate carbon at various locations in the United
States. In: Wolff, G. T.; Klimsch, R. L., eds. Particulate carbon: atmospheric life cycle. New York, NY: Plenum
Press; pp. 297-315.
Wolff, G. T.; Ruthkosky, M. S.; Stroup, D. P.; Korsog, P. E. (1991) A characterization of the pnncipal PM-10 species
in Claremont (summer) and Long Beach (fall) during SCAQS. Atmos. Environ. Part A 25: 2173-2186.
Wu, P.-M.; Okada, K. (1994) Nature of coarse nitrate particles in the atmosphere—a single particle approach. Atmos.
Environ. 28: 2053-2060.
Yamartino, R. J.; Scire, J. S.; Carmichael, G. R.; Chang, Y. S. (1992) The CALGRID mesoscale photochemical grid
model—I. Model formulation. Atmos. Environ. Part A 26: 1493-1512.
Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
Zafiriou, O. C.; Gagosian, R. B.; Peltzer, E. T.; Alford, J. B.; Loder, T. (1985) Air-to-sea fluxes of lipids at Enewetak
Atoll. J. Geophys. Res. [Atmos.] 90: 2409-2423.
Zhang, X.; McMurry, P. H. (1991) Theoretical analysis of evaporative losses of adsorbed or absorbed species during
atmospheric aerosol sampling. Environ. Sci. Technol. 25: 456-469.
Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from a-pinene and
p-pinene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
27: 1593-1607.
Zhang, J.; Wilson, W. E.; Lioy, P. J. (1994) Indoor air chemistry: formation of organic acids and aldehydes. Environ.
Sci. Technol. 28: 1975-1982.
Zimmerman, P. R. (1979) Testing of hydrocarbon emissions from vegetation, leaf litter and aquatic surfaces, and
development of a methodology for compiling biogenic emission inventories: final report. Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; EPA report
no. EPA-450/4-79-004. Available from: NTIS, Springfield, VA; PB-296070.
3-239
-------
Zwozdziak, J. W.; Zwozdziak, A. B. (1990) Sulphate aerosol production in the upper parts of the Sudety Range,
Poland. J. Aerosol Sci. 21(suppl. 1): S369-S372.
3-240
-------
4. SAMPLING AND ANALYSIS METHODS FOR
PARTICULATE MATTER AND ACID DEPOSITION
4.1 INTRODUCTION
Assessment of health risks associated with airborne aerosols implies that measurements be
made defining the aerosol characteristics, concentrations and exposures that contribute to, or
simply correlate with, adverse health effects. The proper selection of an aerosol sampling or
analysis methodology to accomplish such measurements requires that rationales be applied that
consider how the resulting data will be used and interpreted, in addition to the data quality
required. As an example, treatment of a sample to remove particle-associated liquid water,
either by heating the sample during the collection process or by equilibrating the sample at a low
relative humidity subsequent to collection, may lead to changes in the character of the collected
particles, relative to the dispersed particles, in addition to the removal of water (e.g. Meyer et al.,
1995). Similarly, integrated collection of acidic fine aerosols, without selectively removing the
larger, more basic particles, will cause neutralization (i.e., modification) of the sample on the
substrate (Stevens et al., 1978). The same logic applies to the selective removal of gas phase
components during sampling that might react with the deposited aerosol sample, in a manner
inconsistent with naturally occurring transformation processes. The assumption that fixed-
location measurements are representative of inhalation exposure implies that the effects of local
spatial and temporal gradients are understood and appropriately applied to the sampler siting
criteria (Spengler et al., 1994). Development of relationships between aerosol characteristics
and health or ecological responses requires that the aerosol sampling and analysis processes are
truly representative and adequately defined.
The application of sampling and analytical systems for aerosols must recognize that
particles exist modally as size distributions generated by distinctively different source categories
and having distinctly different chemistries, as discussed in Chapter 3. Two important reasons for
making size-specific aerosol measurements are (a) to relate the in situ aerosol character to the
potential deposition sites, and thus toxicity, of the respiratory system, and (b) separation of the
size distribution modes to identify sources, transformation processes or aerosol chemistry. The
interpretation of particle size must be made based on the diameter definition inherent in the
4-1
-------
measurement process. Since the respiratory system classifies particles of unknown shapes and
densities based on aerodynamic diameter, elucidation of aerosol relationships with health
responses requires that sampling techniques either incorporate inertial aerodynamic sizers or
provide mechanisms to accurately convert the measured diameters (e.g., optical) to an
aerodynamic basis. All particle diameters described in this chapter are aerodynamic, unless
otherwise specified.
Friedlander (1977) provided the descriptive matrix shown in Figure 4-1 for placing
measurement techniques that define aerosol characteristics into perspective, in terms of their
particle sizing capabilities, resolution times and chemical identification attributes. This approach
defined these characteristics by resolution (single particle or greater), discretizing ability, and
averaging process. The author notes that the "perfect" aerosol sampler would characterize
particle size with "perfect" resolution, determine the chemistry of each particle "perfectly", and
operate in real-time with no "lumping" of classes. These characteristics could be amended in
"real-world" terms by suggesting that the "perfect" sampler would also have minimal cost and
operator intervention. Also, if the aerosol measurement design goal is to mimick the respiratory
system, physiological averaging characteristics must be considered. Size-specific, integrated
aerosol measurements have improved significantly and their capabilities are better characterized
since the 1987 PM10 NAAQS, but a "perfect" aerosol sampling system has not been devised. As
discussed below, the methodologies required to adequately define the performance specifications
of aerosol samplers have yet to be devised.
Many recent developmental efforts in aerosol measurement technologies have addressed
the need to perfect the chemical characterization of reactive or volatile species collected on
filtration substrates (e.g., Lamb et al., 1980; Koutrakis et al., 1988). Some of the most
significant recent advances in aerosol measurement technologies have come in the form of
analysis system "protocols", rather than individual pieces of hardware. Recognizing that there is
no single "perfect" sampler, these protocols attempt to merge several aerosol sampling and
analysis technologies into an adaptable and analytically versatile system. System attributes
typically include one or more size-specific aerosol inlets, subsequent fractionators to separate the
fine and coarse particle modes, and denuders and/or sequential filter packs to selectively account
for reactive gas phase species. Examples include EPA's
4-2
-------
Instrument
Resolution
Size
Time Chemical
Composition
Quantity
Measured
(Integrand x I^J1 )
Perfect Single
Particle Counter
Analyzer
Optical Single
Particle Counter
Electrical
Mobility
Analyzer
gdnid
Condensation
Nuclei
Counter
jgdvdn, =
Impactor
gdnid
mpactor
Chemical
Analyzer
9 "j drl
Whole Sample
Chemical Analyzer
}g n} dn,
dv
Key:
/
Resolution of single particle level
Discretizing process
Averaging process
Figure 4-1. Characteristics of aerosol measurement instruments.
Source: Friedlander (1977).
Versatile Air Pollution Sampler (VAPS) (Conner et al., 1993), the Southern California Air
Quality Study (SCAQS) sampler (Fitz et al., 1989) and the Interagency Monitoring of
PROtected Visual Environments (IMPROVE) sampler (Malm et al., 1994).
Recognizing that personal exposure concentrations for aerosols may differ from classical
outdoor fixed-location measurements has produced much smaller and less obtrusive samplers
4-
-------
using the same sizing techniques for application indoors, or even to be worn on the body during
normal activities. Miniaturization of aerosol separators stretches the limits of current
technologies to maintain required sampling precisions and accuracies. One of the most
significant limitations imposed by the low flowrates inherent in personal exposure samplers is
the extremely small sample size available for chemical analysis.
This chapter briefly describes the technical capabilities and limitations of aerosol sampling
and analytical procedures in Sections 4.2 and 4.3, respectively, focusing on (1) those that were
used to collect data supporting other sections in this document, (2) those supporting the existing
PM10, TSP1 and Pb regulations, (3) those that were used to support health and welfare response
studies, (4) those having application in development of a possible fine particle standard, and (5)
discussing the attributes of several new technologies. The discussion of aerosol separation
technologies is divided between (a) devices used to mimic the larger particle (>10 //m)
penetration rationales for the upper airways, and (b) those devices generally used to mimic
smaller particle penetration (< 10 //m) to the thoracic regions. These device descriptions are
followed by sampling considerations for their applications. The applications of performance
specifications to define these measurement systems for regulatory purposes are discussed, along
with a number of critical observations suggesting that the current specification process does not
always ensure the accuracy or representativeness necessary in the field. The EPA program
designating PM10 reference and equivalent sampling systems is then briefly described, along
with a current list of designated devices. Selected measurement systems used to provide more
detailed characterization of aerosol properties for research studies are discussed, with a focus on
the determination of particle size distributions.
Aerosol sampling systems for specialty applications, including automated samplers,
personal exposure samplers and the sampling systems used in aerosol apportionment studies are
briefly described. The chapter then presents a short section (4.4) on sampling and analysis of
bioaerosols Nevalainen et al. (1992). Also, Nevalainen et al. (1993), and Qian et al. (1995)
provide excellent summaries of the principles involved in bioaerosol sampling and the most
commonly used techniques.
'Subsequent identifications in this chapter: "TSP" for Total Suspended Particulates by high volume sampler, "PM10" for
the fraction less than 10 um, "fine" for the fraction less than 2.5 um.
4-4
-------
4.2 SAMPLING FOR PARTICULATE MATTER
4.2.1 Background
The development of relationships between airborne particulate matter and human or
ecological effects requires that the aerosol2 measurement process be accurately, precisely and
representatively defined. Improvements in sampling methodologies since the 1982 Air Quality
Criteria Document for Particulate Matter and Sulfur Oxides (U.S. Environmental Protection
Agency, 1982)3 was released, have resulted from improved sensor technologies, and more
importantly, a better understanding of the aerosol character in situ4. Additionally, health studies
and atmospheric chemistry research in the past decade have focused more closely on smaller,
better-defined aerosol size fractions of known integrity, collected specifically for subsequent
chemical characterization.
The system of aerosols in ambient air is a continuum of particle sizes in a gas phase carrier
formed as the summation of all size distributions produced by individual sources and secondary
transformations. Portions of the composite distributions are often found to exist lognormally
(Baron and Willeke, 1993; see also Chapter 3, Section 3.3.3). Aerosol systems also exist as a
continuum of particle "ages", resulting from loss and transformation mechanisms such as
agglomeration, settling, volatilization, gas-particle reaction, and rain-out affecting freshly
generated particles. The chemical compositions of the various portions (modes) of the aerosol
size distribution are more discreet, and sampling strategies must consider a specific range of
sizes for a given chemical class. The constantly changing character of the atmosphere (or of
indoor air) places a premium on sampling strategies both to collect representative aerosol
samples from the air and to protect their integrity until analyzed.
The 1982 Criteria Document provided basic descriptions of many aerosol measurement
techniques still used today. These included both older optically-based techniques, such as
"Black Smoke" or "British Smoke" (BS) or "coefficient of haze" (COH) methods and certain
other now lesser used gravimetric methods, that are only briefly mentioned here but not
Consistent with recent literature (e.g., see Willeke and Baron, 1993), the term "aerosol" will refer to the continuum of
suspended particles and the carrier gas.
3Referred to in the text subsequently as an entity as the "1982 Criteria Document".
4The in situ characteristics of particles in the ambient air medium can be substantially modified by the sampling and
analysis processes. For example, a particle counter which draws particles through a restrictive or heated inlet before they
reach the sensing volume, may perceive the particle properties (e.g. scattering coefficients, size distributions) differently
from those that existed in the ambient.
4-5
-------
described in detail. Instead the reader is referred to the earlier Criteria Document (U.S.
Environmental Protection Agency, 1982) for more information on those methods not extensively
covered here. This section mainly highlights the more recent peer-reviewed research on aerosol
measurement technologies since 1982 and notes salient points that should be considered in their
application. The aerosol sampling section is not intended to be an exhaustive treatise, but is
structured to highlight important concepts and technologies relevant to the development of
aerosol measurement/response relationships, or supporting existing and potential EPA aerosol
regulations. Ancillary reference texts, describing basic aerosol mechanics (e.g., Hinds, 1982;
Reist, 1984) and applied aerosol mechanics and measurements (e.g., Willeke and Baron, 1993;
Hering, 1989; Lundgren et al., 1979; Liu, 1976) should be consulted for more fundamental
details.
4.2.2 Large Particle Separators
4.2.2.1 Cutpoint Considerations
The collection of an aerosol sample is defined by the penetration characteristics of the
inlet, overlaid on the existing in situ size distribution. Cooper and Guttrich (1981) describe this
process mathematically, and they estimate the influences of non-ideal penetration characteristics.
Miller et al. (1979) described the considerations for the possible selection of 15 //m (designated
"inhalable") as a standard for size-selective particle sampling with upper airway respiratory
deposition as the primary consideration. The selection of the most appropriate aerodynamic
criteria for ambient aerosol sampling was only partially resolved by the 1987 EPA designation
(U.S. Environmental Protection Agency, 1987) of a 10 //m (PM 10) cutpoint. The "ideal" PM10
inlet was referenced to the thoracic penetration model of Lippmann and Chan (1979). Ogden
(1992) noted that the standardization for aerosol cutpoint sizes and separation sharpness is still
under debate across settings (ambient air, occupational) and across national and international
governmental entities. As shown in Figure 4-2 (from Jensen and O'Brien, 1993), the
international conventions for cutpoints have been roughly categorized as Respirable, Thoracic
and Inhalable (previously, Inspirable). These cutpoints are related to the penetration,
respectively, to the gas exchange region of the lung, the larynx, and the nasal/oral plane. The
influences of physiological variables on these
4-6
-------
100
80
60
40
? 20
JJ 0/100
c
.2 80
o
JE 60
LU
C 40
O
o 20
3 0/100
O
80
60
40
20
0
Inhalable
^^ Proposed ISO (1992)
. ISO (1983)
"Thoracic
^™ ACGIH (1994)
^"Proposed ISO(1992)
ISO (1983)
Respirable
ACGIH (1994)
Proposed ISO(1992)
ISO (1983)
BMRC (1959)
0.1 1 10 100
Aerodynamic Diameter (pm)
Figure 4-2. American Conference of Governmental Industrial Hygienists (ACGIH),
British Medical Research Council (BMRC), and International Organization
for Standardization (ISO) size-selective sampling criteria.
Source: Jensen and O'Brien (1993).
4-7
-------
outpoints are described by Soderholm (1989). The British Standard EN 481 (CEN [European
Committee for Standardization], 1993) describes size fraction definitions for workplace aerosol
sampling, and identifies inhalable "conventions" relative to thoracic, respirable, extra-thoracic
and tracheobronchial penetration (but not necessarily deposition) in the respiratory system.
They define a thoracic cumulative lognormal distribution with a median of 11.64 //m and a
geometric standard deviation of 1.5, such that 50% of airborne particles with Da = 10 //m are in
the thoracic region. The American Conference of Governmental and Industrial Hygienists
(ACGIH, 1994) also adopted these convention definitions. Owen et al. (1992) provides an
extensive list of the outdoor and indoor particles by type and source category that are found in or
overlap these ranges. Willeke et al. (1992) describe the sampling efficiencies and test
procedures for bioaerosol monitors.
The concept of using an inlet or separator that has the same sampling (penetration)
characteristics as portions of the respiratory system has been discussed by a number of
researchers, including Marple and Rubow (1976), Lippmann and Chan (1979), Vincent and
Mark (1981), Soderholm (1989), Liden and Kenny (1991) and John and Wall (1983). They
describe sampler design considerations for matching penetration models for respirable, thoracic
and inhalable fractions that have been proposed by a number of governing bodies. Since all
models proposed for the same fraction do not necessarily coincide, given the variability and
differences in interpretation of respiratory system data, Soderholm (1989) proposed compromise
conventions for each fraction. Watson et al. (1983), Wedding and Carney (1983), and van der
Meulen (1986) mathematically evaluated the influences of inlet design parameters on collection
performance relative to proposed sampling criteria. These analyses suggested that factors such
as extremes in wind speed and coarse particle concentration could pose significant problems in
meeting performance specifications.
An analysis of the human head as an aerosol sampler was discussed by Ogden and Birkett
(1977), who noted that breathing is an anisokinetic sampling process. The concept of a "total
inhalable" fraction that passes the oral and nasal entry planes was refined by Mark and Vincent
(1986) with the development of a personal aerosol sampling inlet that mimicked this penetration
as a function of aerodynamic size. The inlet was designated the IOM for the Institute for
Occupational Medicine in Edinburgh, Scotland, where it was developed with the cutpoint as a
function of wind speed and aerosol type shown in Figure 4-3. The total
4-8
-------
120
100
"£•
"> 80
0
c
'jo
it eo
u
O)
c
I 4.
n
CO
20
'
I
g /'_
-• — O A 9m/s /
'""V*"*-. O \ -V
""""'*"" ^ A ^\ '
^ • ^\ '1-
O ^'^.. _ 4m/s . J
M. T \ P
• x
V D V-^^ ^v5xV
r-^J^fs^*
A 4 ~
- imis
O
+ o
, , 1 , ,
1 10 100
Aerodynamic Diameter (um)
Inhalable
Convention
Solid Particles
Aloxite
• 1m/s
A 2m/s
• 4m/s
^ 6m/s
V 9m/s
Sodium Fluorescei
O 1m/s
A 2m/s
d 4m/s
O 6m/s
^ 9m/s
Droplets
npuc
u c n o
O 6m/s
+ 9m/s
Figure 4-3. Sampling efficiency of IOM ambient inhalable aerosol sampler for three different types of test aerosol.
Source: Mark et al. (1992).
-------
inhalable approach has been adopted by the International Standards Organization (ISO, 1993),
European Committee for Standardization (CEN, 1993) and by the American Conference on
Governmental and Industrial Hygienists (ACGIH, 1985; ACGIH, 1994) for workplace aerosol
sampling. The ACGIH (1985) reference provides a detailed rationale for the selection of various
cut sizes. The total inhalable fraction using the IOM inlet was selected for a total human
exposure study (Pellizzari et al., 1995) to provide the total body burden for metals (lead and
arsenic) by the air exposure route.
Similar thoracic penetration conventions have been adopted by ISO, CEN, ACGIH and
EPA, each with D50 values of 10.0 //m (ISO, 1993; CEN, 1993; ACGIH, 1994; and U.S. EPA,
1987). The EPA definition was based primarily on the data of Chan and Lippmann (1980). The
exact shapes of each efficiency curve were mathematically defined by Soderholm (1989) and are
slightly different for each convention.
The respirable conventions have had D50 values ranging from 3.5 to 5.0 //m, but a
compromise convention has been accepted internationally by several organizations. It has a D50
of 4.0 (j,m (Soderholm, 1989). ISO (1993) calls this the "healthy adult respirable convention".
Liden and Kenny (1992) discuss the performance of currently available respirable samplers.
EPA's emphasis on the 2.5 //m cutpoint was more closely associated with separating the fine and
coarse atmospheric aerosol modes, rather than mimicking a respiratory deposition convention.
The exact location of this minimum in the atmospheric size distribution is currently under
debate. It is noteworthy that ISO (1993) defines a "high risk" respirable convention which is
claimed to relate to the deposition of particles in the lungs of children and adults with certain
lung diseases. The respirable "high risk" convention has a D50 of 2.4 //m, so it could be
identified closely with the EPA samplers having a cutpoint of 2.5 //m.
The PM10 size fraction has become nearly universal for ambient air sampling in the U.S.,
with the implementation of the 1987 standard (U.S. Environmental Protection Agency, 1987).
The setting of performance specifications, even with their limitations, has provided a more
consistent PM10 data base, with better definition of the data quality. As additional information
becomes available on the sources of biases in aerosol collection methodologies, further
characterizations of older methods may be needed to better define the quality of collected data.
Factors that affect bias, and especially representativeness, should be identified and their
influences determined as a function of particle size. As examples, Appel et al. (1984) studied
4-10
-------
gas/particle and particle/substrate interactions for sulfates and nitrates, volatilization losses of
nitrates were reported by Zhang and McMurry (1992), while losses for organics were reported
by Eatough et al. (1993). Because of the prevalence of these chemical classes in the fine
fraction, the effect of the losses on larger fractions (e.g., PM10, TSP) would be proportionately
smaller and can now be estimated. The losses of larger particles through aerosol inlet sampling
lines (Anand et al., 1992) has a substantial influence on PM10 coarse fraction samples. This was
demonstrated for the British smoke shade sampler inlet line by McFarland et al. (1982). Inlet
losses would be expected to play only a minor role in sampling the fine particle fraction (<2.5
//m). Biases in concentration for samplers with large particle cutpoints are exacerbated by the
large amount of mass present near the cutpoints and the steep slope of mass versus aerodynamic
size. Thus, small changes in cutpoint can give significant and hard-to-predict mass biases.
4.2.2.2 Total Suspended Particulates
The TSP high volume sampler has remained essentially unchanged since the sampler's
identification as a reference ambient sampling device in 1971 (Federal Register, 1971). The
sampling performance (e.g., wind speed and direction sensitivity) was described in detail in the
1982 Criteria Document, and the TSP sampler was shown by McFarland and Ortiz (1979) to
collect particles with aerodynamic diameters exceeding 40 //m. More importantly, its particle
collection characteristics were shown to be significantly sensitive to wind speed (2 to 24 km/h)
and wind direction. Only minor technical updates have been incorporated in commercially
available units, such as in the types of available sequence and elapsed timers (mechanical,
electronic) and in the types of flow controllers (mass flow, volumetric). Also, cassettes are now
available that protect the fragile glass or quartz fiber filters during handling and transport. Size
fractionating inlets for smaller size cutpoints (e.g., 2.5, 6.0 and 10.0 //m) and cascade impactors
have been developed. Similar to the Pb strategy of using the TSP high volume sampler to
collect a "total" sample, asbestos sampling utilizes an aerosol inlet that attempts to collect a
"total" sample, by using an open-faced filter holder with a conductive inlet cowling. Baron
(1993) discusses the potential anisokinetic problems that can occur with such a simple inlet, but
notes that the small Stokes number for typical asbestos fibers provides efficiencies close to
100%.
4-11
-------
4.2.2.3 Total Inhalable Particles
The toxicity of contaminants such as lead poses health concerns as total body burdens,
suggesting that penetration of all aerosols inhaled into the nose and mouth must be considered,
rather than just thoracic penetration. The TSP sampler for atmospheric lead is thought (Federal
Register, 1978) to more closely capture this larger size fraction than would a PM10 counterpart,
but was not specifically designed to mimic inhalability. The ISO "inhalable" draft sampling
convention (ISO, 1993) is intended to apply to such situations, defining collection of all particles
passing the oral/nasal entry planes. The total inhalable cutpoint is currently available only in a
personal sampler version. Mark and Vincent (1986) described the development of an inhalable
particle inlet (designated as the IOM) meeting the ISO (1992), CEN (1993) and ACGIH (1994)
conventions for inspirable dust. This inlet was improved by Upton et al. (1992) and tested by
Mark et al. (1992) and shown to satisfy the ACGIH criteria for wind speeds of 0.5 and 1.0 m/s.
4.2.2.4 PM10
The penetration of ambient aerosols through a size-fractionating inlet to the collection
substrate must be characterized over the ranges of operating conditions (meteorology and aerosol
types) that may be encountered. The range of conditions currently required by EPA PM10
performance specifications was given in U.S. Environmental Protection Agency (1987). Ranade
et al. (1990) and John and Wall (1983) described the required testing, which specifies a
controlled flow wind tunnel, monodispersed fluorescently-tagged wet and dry aerosols, and an
isokinetic nozzle aerosol sampling reference to determine aerodynamic penetration through
candidate PM10 inlets.
Marple and Rubow (1976) placed inertial impactors on the inlet of an optical particle
counter to provide an aerodynamic calibration of the optical readout for non-ideal particles.
Buettner (1990) noted that an aerodynamically calibrated optical particle counter could in turn
be used to test the sampling performance of other devices only if the particle shape and
refractive index of the test aerosol were consistent between calibrations. Maynard (1993) used
this approach to determine the penetration of a respirable cyclone to polydisperse glass micro-
spheres, using the TSI, Inc. Aerodynamic Particle Sizer (APS). John and Wall (1983) noted that
inaccurate inlet sizing results may be obtained using poly-disperse AC test dust, as the result of
agglomeration. Kenny and Liden (1991) used the APS to characterize personal sampler inlets
4-12
-------
and observed that, on theoretical grounds, calm air sampling would be expected to provide unity
aspiration efficiencies for particles below about 8 //m. Tufto and Willeke (1982) used an optical
particle counter (OPC) to monitor monodisperse aerosols in a wind tunnel setting to determine
the performance of aerosol sampling inlets relative to an isokinetic nozzle. Yamada (1983)
proposed using electron microscopy to determine the size distributions of poly dispersed particles
using manual counting techniques before and after a candidate aerosol separator. Penetration
data from this technique were found to be significantly less precise and more difficult to
interpret compared with data for the same separators using fluorometric methods.
The aerosol cutpoint performance of two PM10 samplers that have met the EPA
performance specifications is illustrated (see Figure 4-4) by the data for the Andersen 321A and
Wedding IP10 high volume sampler inlets at 8 km/h from Ranade et al. (1990). The data show
that the cutpoint requirements, defined as a D50 of 10.0 //m ± 0.5 //m and mimicking a modeled
cutpoint sharpness (og), were met for each of the tested wind speeds. These performance results
were verified by repeating the tests in wind tunnels located at two other research facilities. A
diagram (U.S. Environmental Protection Agency, 1992) of the two-stage Sierra-Andersen PM10
high volume sampler inlet with a design flowrate of 1.13 m3/min is shown in Figure 4-5. The
buffer chamber of this inlet serves to dampen the particle-laden air stream passing through two
sets of acceleration nozzles, which deposit particles larger than PM10 on internal collection
surfaces. The PM10 fraction is typically collected by a glass fiber filter. An oiled impaction
shim was incorporated into the first stage fractionator of the 321A to minimize reentrainment of
deposited particles during field sampling. This modified version (Sierra-Andersen 321B) was
designated as an EPA reference method for PM10 in 1987. A subsequent single-stage
fractionator (Sierra-Andersen 1200) was developed5 and designated as an EPA reference
method, with a D50 of 9.5 //m and a hinged design to facilitate cleaning and oiling of the oiled
impaction shim.
5Graseby-Andersen, Inc., Atlanta, GA.
4-13
-------
Wedding Iffo
Model 321A
3 4 5 6 7 8 910 15 20
Aerodynamic Diameter (pm)
Figure 4-4. Liquid particle sampling effectiveness curves with solid particle points
superimposed for the Wedding IP10 (•) and the Andersen Samplers
Model 321A inlets at 8 km/h.
Source: U.S. Environmental Protection Agency (1992).
A diagram of the cyclone-based Wedding6 PM10 high volume sampler inlet (U.S.
Environmental Protection Agency, 1990) with a design flowrate of 1.13 m3/min is shown in
Figure 4-6. This inlet uses an omni-directional cyclone to accelerate the particle-laden air
stream to deposit particles larger than PM10 on an oiled collection surface. Two additional turns
are made to alter the flow into a downward trajectory toward the collection filter. A brush is
used to clean the deposited aerosol from the absorber surface through an access port. This inlet
was designated as an EPA reference method for PM10 in 1987.
6Wedding and Associates, Fort Collins, CO.
4-14
-------
Buffer Chamber
Air Flow
Acceleration Nozzle
Impaction Chamber
Acceleration Nozzle
Impaction Chamber
Vent Tubes
Filter Cassette
Filter
Filter Support
Screen
Motor Inlet
Figure 4-5. Two-stage Sierra Andersen PM10 sampler.
Source: U.S. Environmental Protection Agency (1992).
-------
Housing
Deflector
Spacing
Maintenance Access Port
Vanes
Vane
Assembly
Base
Insect
Screen
A •
Inner _ \
Tube
Absorber
No-Bounce
Surface
Protective
Housing
Aerodynamic
Inlet
Pathway
Aerodynamic Flow
Deflector
Outer Tube
Figure 4-6. Sampling characteristics of two-stage size-selective inlet for liquid aerosols.
Source: U.S. Environmental Protection Agency (1992).
4-16
-------
The aerosol collection performances for 16.67 1pm PM10 inlets for the dichotomous
sampler are described by Wedding et al. (1982) and McFarland and Ortiz (1984) and are
illustrated by the penetration data in Figure 4-7. The variability of the performance as a function
of wind speed for the Andersen 321A PM10 inlet is shown in Figure 4-8 from data by McFarland
et al. (1984). This is a dramatic improvement over the variability shown by the TSP high
volume sampler (McFarland and Ortiz, 1979) for the same wind speed range. An attempt to
simplify the complexity and improve the availability of wind tunnels to test PM10 inlets was
addressed by Teague et al. (1992), who describe a compact tunnel 6 m long by 1.2 m high that is
capable of testing inlets against the EPA PM10 specifications.
Watson and Chow (1993) noted that the EPA PM10 performance specifications allowed a
tolerance range around the D50 that permitted inlets to be undesirably "fine tuned" to provide a
cutpoint on the lower or upper end of the range. Since a significant amount of mass in the
atmospheric aerosol may be associated with particles in the allowable tolerance range, a
"reduction" in reported concentrations could be achieved by simply using a lower (e.g., 9.6 //m)
cutpoint inlet that is still within the acceptable D50 range. The biases between acceptable
samplers have been apparent in the data from field aerosol comparison studies (e.g., Rodes et al.,
1985; Purdue et al., 1986; Thanukos et al., 1992). Most of the reported biases between samplers
were less than 10%, although some differences greater than 30% were reported. The data
suggested that the collection efficiency of the high volume sampler PM10 inlets based on
cyclonic separation (Wedding, 1985) were consistently lower, while those based on low velocity
impaction (McFarland et al., 1984) were consistently higher. Sweitzer (1985) reported results of
a field comparison of these two high volume sampler types at an industrial location and reported
average biases of 15%. It was noted that this amount of bias was unacceptable for compliance
monitoring and more stringent performance requirements should be used. Rodes et al. (1985)
observed that the PM10 concentration data from the dichotomous sampler (regardless of the inlet
design) gave the most predictable results.
Wang and John (1988) were critical of the EPA PM10 performance specification on
allowable particle bounce (U.S. Environmental Protection Agency, 1987), stating that the
criteria can lead to a 30% overestimation of mass under worst-case conditions. In a related
paper, John et al. (1991) reported that although reentrainment by air flow alone of particles
4-17
-------
100
80
60
d>
o
o
£ 40
20
4 6 8 10 20
Aerodynamic Particle Diameter (pm)
40
Figure 4-7. Penetration of particles for 16.67 1pm dichotomous sampler PM10 inlets.
Source: McFarland et al. (1984).
deposited in an aerosol inlet is typically negligible, reentrainment caused from subsequent
particle deagglomeration caused by "bombardment" can be substantial. John and Wang (1991)
suggested that particle loading on oiled deposition surfaces can bias the collection
4-18
-------
100
80
-60
c
o
^M
re
4-1
O
c
o! 40
20
o 2 km/h
A 8 km/h
n 24 km/h
4 6 8 10 20
Aerodynamic Particle Diameter
Figure 4-8. Collection performance variability illustrating the influence of wind speed for
the Andersen 321A PM,n inlet.
10
Source: McFarland et al. (1984).
(2.2%/gram deposited) and strongly suggested that periodic cleaning and re-oiling should be
required for PM10 inlets. Ozkaynak et al. (1993) observed that immediately after inlets of the
4-19
-------
Wedding (1985) design were cleaned, an underestimation (compared to the dichotomous
sampler) occurred of 14%. This bias was followed by a steady "recovery" period of 2 days, until
the expected performance returned. They also observed a strong influence of diurnal
temperature change on the ratios of concentrations between the Wedding (1985) design samplers
and other PM10 samplers. This influence could not be attributed to a physical phenomenon.
The EPA PM10 performance specification program should be considered successful (John
and Wall, 1983) in providing consistent aerosol collection results during field sampling.
As noted by Thanukos et al. (1992), the cases of greatest concern were those where the measured
concentrations were near an exceedance level. Wiener et al. (1994) noted that EPA was
scrutinizing the current performance of designated reference and equivalent sampling methods
for PM10 in light of reassessment of the existing standard. A review of the current PM10
performance requirements and possible amendments of the existing specifications may be
appropriate, given the information base now available.
Laboratory and field testing reported in the literature since 1987 suggest that the EPA PM10
Federal Reference Method (FRM) specifications and test requirements have not adequately
controlled the differences observed in collocated ambient PM10 sampling. The most significant
performance flaws have combined to produce excessive (up to 60%) mass concentration biases.
These biases apparently resulted from the combined factors of (1) allowing a cutpoint tolerance
(10 ± 0.5 //m), (2) an inadequate restriction placed on internal particle bounce, and (3) a
degradation of particle separation performance as certain technology PM10 inlets became soiled.
Particle bounce or soiling problems have not been reported for the PM10 inlets for the
dichotomous sampler.
A cutpoint tolerance of ±0.5 //m was required to account for expected differences between
different wind tunnel laboratories testing the same hardware. The between-sampler bias from
this tolerance limit alone is predictable and should provide PM10 concentration differences
significantly less than ±10% in most cases. Particle bounce allowances are not as predictable,
but design practices (primarily surface coatings with viscous oil, as suggested by John et al.
[1991]) to minimize the penetration caused by bounce and resuspension have been shown to be
very effective when properly serviced. The influences of internal surface soiling on PM10 inlet
performance were not recognized when the FJAM was established in 1987, but were found to
have severe consequences for some separation technologies. The magnitude of biases from
4-20
-------
soiling is also not readily predicted, but can be ameliorated by not allowing the inlet to become
excessively dirty during operation by routine cleaning prior to sampling.
Although the EPA test procedures have not been formally amended since 1987, the
manufacturers of the designated PM10 reference methods (see section 4.2.6) have voluntarily
modified their hardware designs and instruction procedures to accommodate particle bounce and
soiling concerns. The SA-321b and SA-321c PM10 inlets were voluntarily withdrawn from the
market by the vendor because of excessive biases attributed to particle bounce. The
manufacturer now sells the SA1200 inlet which provides oiled surfaces to eliminate particle
bounce and access screws to facilitate cleaning. The manufacturer also amended the instruction
manuals to require a routine cleaning schedule. Similarly, the manufacturer for the Wedding
PM10 inlet now provides an access port in the inlet and a cleaning procedure that can be applied
prior to the collection of each sample. Based on our current understanding of the PM10 sampling
process, it could be expected that sampling systems can be designed and concentration
measurements made that are within 10% of the true concentrations.
4.2.3 Fine Particle Separators
4.2.3.1 Cutpoint Considerations
Although a particle separation at 2.5 //m has been utilized by the dichotomous sampler for
a number of years, the 1987 standard reassessment (U.S. Environmental Protection Agency,
1987) did not specifically require routine monitoring for fine particles. It has become apparent
(see Chapters 8 and 12) that certain health and ecological responses are most strongly correlated
with fine particles, significantly smaller than 10 //m, and their related chemistry. Since the mass
of a particle is proportional to the cube of its diameter, larger particles (especially above 10 //m)
can totally dominate the mass of PM10 and TSP samples. The 2.5 //m cutpoint generally occurs
near a minimum in the mass distribution, minimizing mass concentration differences between
samplers with cutpoint biases. The development of control strategies based on mass
concentrations from a smaller cutpoint standard must be carefully constructed, especially if large
particle interference problems (e.g., particle bounce) cannot be appropriately minimized.
Practical considerations would be the time and expense required to develop separators with
1.0 //m cutpoints that meet required specifications, conduct validation testing, and retrofit
existing samplers. A virtual impaction "trichotomous" sample was described by Marple and
4-21
-------
Olson (1995) that uses a PM10 inlet and separators for both 2.5 and 1.0 //m outpoints. They also
noted that technology was not a limiting factor in providing a fine particle separator. Given the
body of data available at 2.5 //m, a focused effort may prove practical that defines the
characteristics of the particle mass and chemistry between 1.0 and 2.5 //m. This would add to
the technical knowledge base, allow interpretive corrections between cutpoints to be made, and
permit continued sampling at 2.5 //m with a minimum of additional resources.
4.2.3.2 Virtual Impactors
The dichotomous sampler utilizes virtual impaction to separate the fine (<2.5 //m) and
coarse (2.5 to 10 //m) fractions into two separate flowstreams (see, for example, Novick and
Alvarez, 1987) for collection on filters. The calibration of a nominal 2.5 //m impactor,
including wall loss data, is shown in Figure 4-9 (from Loo and Cork, 1988). The current
separator design was shown to provide a relatively sharp cutpoint with minimal internal losses.
A virtual impactor has been designed with a 1.0 //m cutpoint (Marple et al., 1989), and for
cutpoints as small as 0.12 //m (Sioutas et al., 1994). After a cross-channel correction factor for
the coarse mode is applied, the mass concentrations of each fraction and the total mass (using a
PM10 inlet) can be determined gravimetrically. An inherent consideration with virtual separation
is contamination of the coarse fraction by a portion of the fine fraction, equivalent to the ratio of
the coarse channel flow to the total flow (typically 10%). Although a straightforward
mathematical correction can account for the particle mass between channels, this can influence
subsequent chemical and physical characterizations, if significant differences exist between the
chemistry of each fraction (e.g., acidic fine fraction and basic coarse fraction). Stevens et al.
(1993) utilized this limited addition of fine particles to the coarse fraction to advantage in the
SEM analysis of samples collected on Nuclepore
4-22
-------
100 -
20
Figure 4-9. Aerosol separation and internal losses for a 2.5-/j,m dichotomous sampler
virtual impactor.
Source: Loo and Cork (1988).
filters. Keeler et al. (1988) showed that the growth of fine aerosols at elevated relative
humidities can significantly alter the ratio of fine to coarse collection for the dichotomous
sampler. During early morning periods when the humidity approached 100%, an apparent loss
of up to 60% of the fine mass (to the coarse channel) was observed. Keeler et al. (1988)
concluded that analyzing only the fine fraction of the measured aerosol may not be appropriate,
especially for short integration intervals.
A high volume (1.13 m3/min) virtual impactor assembly was developed by Marple, et al.
(1990) that can be placed on an existing high volume sampler to permit larger total collections
than the dichotomous sampler for chemical speciation by size fraction. By placing a number of
4-23
-------
virtual impactors in parallel, a separation can be achieved at higher flows, while reducing the
total pressure drop. Marple et al. (1993) provide a list of commercially available virtual
impactors by flowrate and available cutpoints. They also note that virtual separators inherently
concentrate the particles in the coarse fraction (typically by a factor of 10), making them useful
as pre-concentrators for sensors with marginal sensitivities. John et al. (1983) found that an
oiled Nuclepore filter with a nominal 8 //m porosity could provide a D50 cutpoint of 2.5 //m,
similar to that of a virtual impactor, if operated at the appropriate face velocity and for a
sampling period short enough to minimize overloading.
4.2.3.3 Cyclones
Cyclones have been used as aerosol separators in personal exposure sampling in
occupational settings for many years. Lippmann and Chan (1979) summarized the cyclones for
sampling aerosol sizes below 10 //m and noted that the aerosol penetration through a cyclone can
be designed to closely mimic respiratory deposition. An intercomparison of three cyclone-based
personal exposure samplers under occupational conditions (concentrations typically > 1 mg/m3)
was described by Groves et al. (1994). They reported that even though the cyclones were
reportedly designed to mimic similar respirable conventions, biases as large as a factor of two
were noted, possibly attributable to overloading problems. Marple et al. (1993) provided a list
of commercially available air sampling cyclones, by sampling flowrate and D50 range. Cyclones
can be used individually or in a cascade arrangement to provide a size distribution. Hartley and
Breuer (1982) describe methods to reduce biases when using a 10 mm (diameter) personal air
sampling cyclone, especially as related to cutpoint shifts caused by flowrate changes. Saltzman
(1984) provided a similar analysis for atmospheric sampling cyclones. Sass-Kortsak et al.
(1993) observed that substantial uniformity-of-deposition problems can occur on the filters
downstream of personal sampling cyclones. Wedding and Weigand (1983) used a cyclone
within a high volume aerosol inlet to provide a PM6 0 cutpoint for ambient sampling that did not
allow penetration of particles greater than 10.0 //m.
The simplicity of cyclones has prompted their use as inlets and subsequent separators in
samplers designed to fractionate the aerosol sample for chemical analysis. The "Enhanced
Method" employed by EPA for sampling acidic aerosols uses a glass cyclone with a 2.5 //m
cutpoint as the sampler inlet (U.S. EPA, 1992). The percent collection as a function of
4-24
-------
aerodynamic diameter is shown in Figure 4-10 (Winberry et al., 1993). The modest outpoint
sharpness exhibited by some cyclones should be considered when attempting to separate particle
size fractions that may interact chemically. Hering et al. (1990) describe several validated
aerosol systems for sampling carbonaceous particles that utilize cyclones with 2.5 //m cutpoints
to sample the fine fraction on either Teflon or quartz substrates. Spagnolo and Paoletti (1994)
describe a dual cyclone ambient aerosol sampler with a 15 //m inlet (described by Liu and Piu,
1981). This sampler was designed to collect a 20 to 15 //m fraction, a 20 to 4.0 //m fraction, and
a 0 to 2.5 //m fraction. Malm et al. (1994) describe a sampling system with a PM10 inlet and
three parallel channels following a 2.5 //m cutpoint cyclone that was used for the 40 site
IMPROVE network. Over 120,000 fine particle filter substrates of Teflon®, nylon and quartz
were collected for chemical analysis over a 6 year period.
4.2.3.4 Impactors
Impactors have been developed for a wide range of cutpoints and flowrates. In cascade
arrangements (see Section 4.2.7.1.1) with a characterized inlet, impactors provide particle
distribution information over a range of aerodynamic sizes. Impactors used as components of
inlets or as in-line fractionators stop and retain the aerosol on a surface (e.g., oil-soaked, sintered
metal or glass) that provides consistent performance (primarily minimal bounce) over the entire
sampling interval. Recovery and analysis of the deposited particles in these situations are
usually not considerations. Koutrakis et al. (1990) described the design of 2.1 //m cutpoint
impactor for a single stage annular denuder system that exhibited internal losses of less than 3%.
Marple (1978) described the use of multiple nozzle impactors in a single stage to emulate
selected respiratory penetration curves.
Marple et al. (1993) noted that the three primary limitations of impactors are particle
bounce, overloading of collection stages and interstage losses. Particles can bounce from a stage
after impaction if the surface forces are not adequate for their retention. Wang and John (1988)
described the effects of surface loading and relative humidity on particle bounce and growth, and
they noted that if less than 6% of the impact area was covered by deposited
4-25
-------
to
100
80
60
o
^
o
_
o 40
O
20
b
2 2.5 4 6
Aerodynamic Diameter (pm)
8
10
Figure 4-10. Percent collection as a function of aerodynamic diameter for the U.S. Environmental Protection Agency
enhanced method glass cyclone.
Source: Winberry et al. (1993).
-------
particles, particle-to-particle collisions (and bounce) could be neglected. They also showed that
ammonium sulfate aerosol growth with increasing humidity resulted in a 25% shift in cutpoint as
the relative humidity increased to 64%. Biswas et al. (1987) showed that, especially in low
pressure zones, the relative humidity and temperature can change rapidly within a cascade
impactor, potentially altering cutpoints and losses. Wang and John (1988) in subsequent work
did not observe these shifts, noting that the transit time in a jet is only on the order of 10 //s.
Turner and Hering (1987) noted that the stage substrate materials (Mylar®, stainless steel and
glass) with the same grease (Vaseline®) could produce substantially different particle adhesion
characteristics. Vanderpool et al. (1987) showed that using glass fiber filters as impactor
surfaces can produce drastically reduced performance as compared to a greased substrate
(see Figure 4-11). Markowski (1987) suggested that adding a duplicate (same cutpoint) serial
impactor stage can permit reasonable bounce and re-entrainment corrections to be made.
4.2.4 Sampling Considerations
4.2.4.1 Siting Criteria
Selection of aerosol sampling locations is partially guided by siting criteria under the 1987
PM10 regulation (U.S. Environmental Protection Agency, 1987), which provided limited
guidance for Pb and PM10 samplers. The details behind these guidelines for PM10 are provided
in a guidance document (U.S. Environmental Protection Agency, 1987), which relates physical
and chemical characteristics of aerosols to the spatial scales (regional, urban, neighborhood,
middle and micro) required to define the influences of sources on various populations. Guidance
was also provided on the influences of nearby point, line and area sources on sampling location
as a general function of particle size. Only limited information was noted to be available on
specific influences of local obstructions and topography (e.g., trees, buildings) on measured
aerosol concentrations. The primary focus was establishment of the degree that a sampling
location was representative of a specific scale.
The high purchase cost, and occasionally physical size, of aerosol samplers have restricted
the number of sampling sites used in air monitoring studies. This may pose problems if the
selected sites are not truly representative of the exposures for the populations at risk. To address
the biases resulting from too few aerosol samplers in a field study, a
4-27
-------
100
~ 80
•3 60
UJ
o 40
o
o
o
20
4th Impactor
O Greased Substrate
D Glass-Fiber Filter
5 10 20 40
Aerodynamic Particle Diameter (urn)
Figure 4-11. Performance of glass fiber filters compared to greased substrate.
Source: Vanderpoolet al. (1987).
"saturation" sampler approach has been used, utilizing an inexpensive, miniature and
battery-powered PM10 sampler that can be deployed at a large number of sites. Phillips et al.
(1994) reported application of this approach, using 15 PM10 saturation samplers in conjunction
with one dichotomous sampler to study the contribution of diesel emissions to total PM levels in
Philadelphia. Although the mean for PM10 concentrations of the saturation samplers was
essentially identical to that of the dichotomous sampler, the saturation data showed site-to-site
mean differences of up to 30 //g/m3.
4.2.4.2 Averaging Time/Sampling Frequency
The collection frequency for samples to support the EPA PM10 NAAQS has typically been
on an every-6th-day schedule. Shaw et al. (1982) raised a statistically-based concern that
4-28
-------
infrequent collection increases the coefficient of variation about the overall mean concentration
value; that is, the variability of computed fine mass concentration means increased as the square
root of the number of intervals between individual measurements. Symanski and Rappaport
(1994), using time series analyses, described the influences of autocorrelation and non-stationary
behavior in occupational settings on concentration distributions constructed from infrequent
sampling. They recommended a random sampling design where a sufficient number of locations
are sampled repeatedly over an adequate period of time to account for the full range of exposure
possibilities. Hornung and Reed (1990) described a method of estimating non-detectable (or
missing) values to lessen variance about the estimate of the geometric mean, by assuming that
the concentration distribution is log-normal.
Insufficient sample collections can be remedied by more frequent operation of manual
samplers. The recent PM10 equivalency designations (see section 4.2.5) of two beta gauge
samplers and the TEOM sampler can provide the necessary information, with hourly rather than
daily resolution. The initial cost of an automated sampler is typically 2-3 times that of a manual,
single channel PM10 sampler, but can be offset by savings in operator labor costs. If inherent
biases described in section 4.2.3.4 for the beta and TEOM samplers can be overcome (and they
are field reliable), these approaches should prove very useful in routine regulatory and research
monitoring studies. Potential also exists for the integrating nephelometer to be an acceptable
exceedance monitor7, using site specific calibrations relating the measured scattering coefficient,
bsp, to fine aerosol mass concentrations (e.g., Larson et al., 1992).
Another consideration for defining sampling intervals is the setting of start and stop clock
times. Daily 24-h sampling is most often done from midnight-to-midnight, but occasionally
from noon-to-noon to either reduce the number of samplers required or to reduce operator
burden. Sampling locations with highly variable diurnal aerosol concentration patterns (e.g.,
from night time wood smoke influence or day time traffic dust), or marked differences between
week days and weekend days may require special consideration. These influences can be
especially significant for <24-h sampling periods.
7A Pollutant Standard Index (PSI) monitor used to estimate when a pre-determined exceedance level has been reached
or exceeded, to potentially trigger the operation of an equivalent PM10 gravimetrically-based sampler.
4-29
-------
4.2.4.3 Collection Substrates
The selection of a filtration substrate for integrated collection of particles must be made
with some knowledge of the expected particle characteristics and a pre-determined analytical
protocol. The expected sampled size distribution places a requirement on the porosity of the
filter media to effectively trap a reasonably high percentage of the particles with a minimum of
pressure drop. The most common filter types used in air sampling are fiber and membrane.
Fiber filters tend to be less expensive than membrane filters, have low pressure drops, and have
high efficiencies for all particle sizes. They are most commonly available in glass fiber, Teflon
coated glass fiber and quartz materials. Membrane filters retain the particles on the surface for
non-depth analyses (e.g., X-Ray Fluorescence), can have specific porosity's, and are available in
a wide variety of materials. Teflon is a popular membrane material because of its inertness, but
is 2 to 4 times as expensive as more common materials. Liu et al. (1978) summarize the
effective penetration characteristics as a function of particle size and pressure drops for a wide
variety of fiber and membrane filters. The selection of filter diameter for a given flowrate
influences the face velocity and the loading capacity before the pressure drop becomes
unacceptable. A 47mm filter provides a surface area that is 60% larger than that of a 37mm
filter. Polycarbonate filters with well defined porosities (e.g., Nuclepore®) have been used in
"stacked" arrangements as fine particle separators. John et al. (1983) describe using an 8 //m
porosity filter in series with a back-up filter to effectively provide a 3.5 //m separation of fine
and coarse particles in a small, inexpensive package. Samplers based on this principle were
widely used in the early 1980's (Cahill et al., 1990) and their performance under field conditions
was shown to be equivalent to later cyclone based PM2 5 samplers in the IMPROVE network.
The reactivities of filter substrates with the aerosol have been reported extensively.
A common problem with glass fiber filters used on high volume samplers is the basic pH of the
glass material and its effective conversion of SO2 to particulate sulfates (e.g., Pierson et al.,
1976). Appel et al. (1984) also reported similar conversions of nitrogen oxides to particulate
nitrates on glass fiber filters. Witz et al. (1990) reported losses of particulate nitrates, chlorides
and ammonium (19, 51 and 65%, respectively) from quartz fiber filters during storage. No
significant losses of sulfates were reported from quartz filters. Similarly, Zhang and McMurry
(1992) reported the anomalous loss of fine particle nitrates from Teflon filters and noted that
predictive loss theories were insufficiently accurate to permit corrections. Lipfert (1994) also
4-30
-------
observed that nitrate artifacts on glass fiber filters were difficult to quantify on a routine basis.
Measurements of paniculate nitrate using nylon filters by the IMPROVE protocols show,
however, that such effects are minor except in California (Malm et al., 1994). Eatough et al.
(1993) found significant losses of particulate organic compounds on quartz filters due to
volatilization, such that ambient concentrations of particulate carbon may be underestimated
substantially. Lipfert (1994) investigated filter artifacts in a field study in New York and
concluded that positive sulfate artifacts inflated PM10 values from glass fiber filters by 6 //g/m3.
It was noted that the combination of sulfate and nitrate artifacts on glass fiber filters may inflate
TSP measurements by as much as 10 to 20 //g/m3.
4.2.4.4 Chemical Speciation Sampling
The collection of aerosol samples for chemical speciation analysis adds another dimension
to the complexity of the sampling protocol (also see Section 4.3). The simplest approach utilizes
a characterized inlet or separator to define a size fraction, provides an aerosol collection
substrate compatible with the analytical technique, and collects an adequate quantity of sample
for analysis. This approach is applicable for relatively nonreactive and stable components such
as heavy metals. An important consideration is the potential reactivity of the sampling substrate
with either the collected aerosols or the gas phase. Appel et al. (1984) predicted effects of filter
alkalinity on conversion of acid gases to sulfates and nitrates and provided an upper limit
estimate for artifact sulfate formation (added mass) for TSP high volume sampling of 8-15
Mg/m3 for a 24-h sample.
Analyses for semi-volatile organics found in both the particle and vapor phases must be
collected by adding a vapor trap (e.g., polyurethane foam plug) downstream of the sampling
filter. Arey et al. (1987) noted that this arrangement of sequential sampling reservoirs may
account for the total mass of organics, but not accurately describe their phase distribution in situ,
due to "blow-off from the filter during sampling. Van Vaeck et al. (1984) measured the
volatilization "blow-off losses of organic species from cascade impactor sampling to be up to
30%, while the loss of total mass was only 10%. McDow and Huntzicker (1990) characterized
the face velocity dependence for organic carbon sampling and provided correction models, based
on adsorption losses to a backup filter. Turpin et al. (1994) examined organic aerosol sampling
artifacts and highlighted the distinction between "organic carbon" and individual organic species.
4-31
-------
They observed that organic carbon sampled from the atmosphere is unlikely to attain equilibrium
between that in the gas phase and that adsorbed on a quartz fiber back-up filter. They also noted
that under typical sampling conditions, adsorption is the dominant artifact in the sampling of
particulate organic carbon, and longer sampling periods reduce the percentage of collected
material that is adsorbed vapor. It was recommended that collection of aerosols for carbon
analyses be made on a pre-fired quartz filter, with estimates of the adsorption artifact made from
a quartz filter placed behind a Teflon filter in a parallel sampler.
For more highly reactive and unstable species, the recognition of the in situ character of
the aerosol in the air must be identified and preserved during all facets of the sampling process
to provide a representative and accurate sample. Durham et al. (1978) described a denuder to
remove sulfur dioxide while sampling for sub-micron aerosols. Spicer and Schumacher (1979)
observed that many artifact reactions may occur if stripping of nitric acid, sulfuric acid and
ammonia is not performed during speciated aerosol sampling. Appel et al. (1988a) described the
various loss mechanisms that apply to the aerosol and vapor phases while sampling for nitric
acid. They noted that residence time, surface material compositions, and conditioning prior to
sampling were the predominant variables affecting transmission efficiency.
The determination of strong acidity for atmospheric aerosols (U.S. Environmental
Protection Agency, 1992) describes an "enhanced" method that recognizes the inter-relationships
between the vapor and aerosol phases for each constituent and the potential interferences. An
inlet cyclone or impactor is used to provide a 2.5 //m cutpoint to exclude the higher pH aerosols
found in the coarse fraction of PM10. As shown in Figure 4-12, denuders are used in the
flowstream which selectively remove gas phase components with minimal, characterized losses
of aerosol. Ye et al. (1991) determined the aerosol losses through an 10 1pm annular denuder
system as a function of particle size. They noted that total particle losses were less than a few
percent whether the denuders were coated or uncoated. Also, using parallel annular denuders,
Forrest et al. (1982) found aerosol losses of only 0.2 to 2.2% for 0.3 to 0.6 //m particles and 4 to
5% for 1 to 2 //m particles.
4-32
-------
Filter Pac
Coated
Filters
fNa2C03(#2)X^S
^ Na2C03(#1) k*
Lv-Muic Mciui
Teflon Filter^^J
d4
d3
d2
Coupler (Typical) — .
Coupler
d1
/ impactor
.. ^ H
ajttttttttttttttttttttttttt^
**• «*"
-
J
Citric Acid
^i^^^^
*-
•^ — .
^
*«
•x
n
O
o
en
(0
s
•*~-li>
^___--^
t-~_-
T
n
O
0
CM
(0
z
•*— -^
»- -*
>-
o
(0
x
*— -^
A
NO2
HN03
NH3
,NH4,804>NOa
T
NH3
1
t
NO2> HNO2> SO2
HCL, HNO2
HNO3> SO2
1
T
HNO3
S02
Figure 4-12. Schematic diagram of an annular denuder system.
Source: U.S. Environmental Protection Agency (1992).
4-33
-------
Filter packs have been developed, consisting of a sandwich of filters and collection media
of various types in series, to collect aerosols and selectively trap gases and aerosol volatilization
products. Benner et al. (1991) described an annular denuder sampling system using Teflon and
nylon filter packs and annular denuders to quantitatively collect the distributed ammonium
nitrate, nitric acid and ammonia in the vapor and aerosol species. They observed that volatile
nitrates were 71% ± 27% of the total nitrates during the day and 55% ± 30% at night in arid,
southwestern U. S. locations. Masia et al. (1994) described the anomalous uptake of ammonia
on the nylon filters, which were expected to collect only the gas phase nitric acid. Wang and
John (1988) reported volatilization losses of ammonium nitrate in the Berner impactor of 7%
under hot, dry (18% Rh) conditions.
Vossler et al. (1988) reported the results of improvements in an annular denuder system,
including Teflon coating of the internal glass surfaces. They found an apparent particle bounce
problem with the cyclone inlets (with or without Teflon coating) and proposed adding an
additional in-line, greased impactor. John et al. (1988) found that anodized aluminum surfaces
absorb nitric acid efficiently and irreversibly. Several method comparison studies have been
reported for systems utilizing annular denuder/filter pack technologies, including Harrison and
Kitto (1990), Sickles et al. (1990), and Benner et al. (1991).
4.2.4.5 Data Corrections/Analyses
Aerosol concentration data are reported in units of mass per volume (e.g., //g/m3). The
current EPA regulations for sampling TSP, PM10 and Pb require that sampler flowrates be
controlled and the sampled volumes be standardized to 760 mm Hg and 25 °C. These
requirements may pose problems in the interpretation of concentrations from aerosol samplers.
Wedding (1985) notes that the flowrate through inertial impactors should be maintained at
"local" temperatures and pressures to retain the separator's aerodynamic calibration. Mass flow
controllers may significantly affect the separator flow velocity during large diurnal temperature
changes, excessively biasing the resulting cutpoint diameter.
Subsequent correction of the sampled aerosol volume to "standard" conditions by
mathematically compensating for average meteorological conditions may improperly report the
aerosol concentration measurement. If the rationale for aerosol sampling was to mimic
respiratory penetration (which occurred at local conditions), a correction after-the-fact may not
4-34
-------
be appropriate. These corrections are typically small (less than a few percent) except in
locations at higher altitudes and those with large diurnal or seasonal temperature changes. The
basis for mandating flowrate controller performance for aerosol samplers is sound, but the
subsequent requirements for concentration corrections for temperature and pressure are complex.
Although the issue of sampled volume correction for local temperature and pressure is beyond
the scope of this document, the scientific bases should be reassessed for aerosol sampling to
determine if this requirement is consistent with EPA goals.
The matching of aerosol measurement capabilities with data quality requirements is
discussed by Baron and Willike (1993). They note that although aerosol sampler precision can
be determined from collocated measurements, field sampling accuracy is more difficult to
define. Generation of mono- or polydisperse calibration aerosols are rarely done in field settings
because of the complexity of the calibration process. Typically, only the aerosol sampler
flowrate accuracy is determined in the field. Biases between the means from collocated aerosol
samplers using different separation techniques, may result from sampler operational errors, or
from inadequacies in determining the performance specifications during laboratory testing.
4.2.5 Performance Specifications
4.2.5.1 Approaches
A significant step in the standardization process for aerosol sampling was the EPA
definition (U.S. Environmental Protection Agency, 1987) of the PM10 size fraction, based on the
aerodynamic diameter of particles capable of penetrating to the thoracic region of the respiratory
system. This definition was followed by implementation of the PM10 provisions of EPA's
Ambient Air Monitoring Reference and Equivalent Methods regulation (U.S. Environmental
Protection Agency, 1987). The format of the latter regulation included adoption of performance
specifications for aerosol samplers, based on controlled wind tunnel testing with mono-dispersed
aerosols. Controlled laboratory testing is followed by limited field testing, including tests of
candidate equivalent methods to demonstrate comparability to designated reference methods.
The stringency of the field testing to elucidate potential sampling biases is strongly influenced
by the local sampling site environment, including factors such as wind speed, nearby point
sources, and the probability of fugitive dust events.
4-35
-------
This approach was chosen, rather than the design specification approach taken in 1971
(Federal Register, 1971), which identified the high volume sampler and associated operational
procedures as the reference method for Total Suspended Particulates (TSP). The 1971
regulation had no provisions for the use of alternative or equivalent methods, and subsequent to
this design designation, significant problems of the TSP high volume sampler, such as wind
speed and direction dependency (McFarland et al., 1979) and off-mode collection (Sides and
Saiger, 1976), were reported. These inherent biases complicated the interpretation of TSP
concentration data (U.S. Environmental Protection Agency, 1982) and weakened correlations
with other measures. The problems were estimated to have induced biases of less than 10% for
most situations, but occasionally as high as 30%. The subsequent development of aerosol testing
programs for size selective aerosol samplers (e.g., McFarland and Ortiz, 1979; Wedding, 1980;
John and Wall, 1983; Ranade et al., 1990; Hall et al., 1992) more rapidly identified weaknesses
in existing technologies and facilitated the development of better methods.
No reference standard exists for aerosol concentration measurements in air. The
calibration of aerosol samplers relies primarily on characterizations under controlled conditions
of the sampler sub-systems, including the size selective inlet, sample conditioning and
transmission system, the flow control system, and, if used, subsequent size separators, sample
collection and storage elements, and sensors and associated electronics. Although the precision
of an aerosol sampler is readily obtained by using replicate, collocated samplers, the accuracy
can only be estimated by comparison with either designated "reference" samplers or with
computations of expected aerosol mass collections. Performance specification limits are used to
control the overall aerosol sampling accuracy. As noted by John and Wall (1983) the selection
of a comprehensive list of sampling elements requiring inclusion and the setting of the
performance limits for each element is a difficult task, especially when the range of "real-world"
sampling situations is considered.
Performance specifications were utilized for the PM10 standard to allow the broadest
spectrum of measurement technologies, hopefully encouraging the development of new and
better methods. A research program was implemented by EPA in parallel with preparation and
review of the 1982 Criteria Document to identify the critical specifications and understand the
inter-relationships among the parameters influencing the aerosol sampling process. Studies of
the influences of factors such as wind velocity, particle character, flow rate stability, particle
4-36
-------
bounce and wall losses on precision and accuracy substantially advanced the science of large
particle sampling. The performance specification approach was a significant improvement over
the design specification approach used for the TSP high volume sampler, in that it fostered the
development of new information and technologies and provided for the use of alternative
methods. In retrospect, the primary weakness of the design specification approach for the TSP
reference method was not the process per se, but the technical inadequacy of the development
and testing program that produced the high volume sampler design.
The utilization of a performance specification approach requires that a minimum level of
knowledge be available about the measurement process and the associated test procedures.
Some significant drawbacks subsequently observed in the performance specification approach
for PM10 included the complexity, expense and scarcity of aerosol wind tunnel test facilities, and
the difficulty in defining comprehensive specifications that considered all of the nuances of
aerosol sampling. Wind tunnel evaluation and limited field tests do not always identify sampler
related problems encountered during extended periods of ambient sampling (e.g., John and
Wang, 1991). Future performances tests should ideally include extended field testing, for
example, to evaluate performance in different geographic regions and seasons, as well as under
different meteorological conditions.
4.2.5.2 Performance Testing
Since the 1982 Criteria Document (U.S. Environmental Protection Agency, 1982a),
aerosol sampling research studies have identified numerous factors that influence the precision
and accuracy of samplers in both wind tunnel and field performance testing. Rodes et al. (1985),
Purdue et al. (1986), and Cook et al. (1995) showed, in field evaluations under a variety of
sampling situations, that PM10 samplers meeting the EPA performance specifications provide
aerosol concentration measurements with a precision of 10% or less when samplers of the same
model were compared. However, significant biases were evident when different types of
samplers were compared. The Andersen SA-321A PM10 sampler was found to collect an
average of 58% more mass than a collocated Wedding PM10 sampler (Perdue et al., 1986). This
was partly attributed to the (predicted) bias associated with cutpoint differences between the
inlets. A more significant bias (not predicted) was associated with degraded performances in
opposite directions (Andersen over-sampling, Wedding under-sampling) due to soiling of the
4-37
-------
separators during extended sampling periods. Rodes et al. (1985) noted that sampler precisions
(coefficients of variation) were better than ±10%, with several samplers better than ±5%. Cook
et al. (1995) reported good agreement (variability less than 15%) among several types of PM25
samplers. Other sampler types showed significant biases. Under the conditions of the study,
high concentrations of NH4NO3 and organic carbon (winter in Bakerfield, CA), samplers which
heated the collected particles to 30 °C or 50 °C during sampling gave lower mass values than
filter samples which were collected at ambient conditions and equilibrated for 24 hours at 23 ± 3
°C and 40 ± 5% relative humidity. Coefficient of Haze (COH) measurements by an American
Iron and Steel (AISI) tape sampler and light scattering (bscat) measured by an intergrating
nephelometer heated to 17 °C correlate well with PM25 measurments (COH, r = 0.82 to 0.91;
bscat,r = 0.91 to 0.98).
Mark et al. (1992) reviewed the attributes of wind tunnel testing, and noted that tests using
controlled conditions are a necessity to determine whether an aerosol sampler meets a basic set
of established performance specifications. Hollander (1990) suggested that sampler performance
criteria should be evaluated in controlled outdoor tests, given the inability of wind tunnels to
accurately mimic the influences of outdoor meteorological conditions on sampling. The current
EPA PM10 performance testing requires field tests to demonstrate sampler precision and flow
rate stability, and the comparability of equivalent methods to designated reference methods. The
stringency of such tests are highly dependent on the sampling location chosen, local aerosol
sources, the existing meteorology and the season.
Kenny and Liden (1991) noted that the EPA PM10 sampler performance specifications
(U.S. Environmental Protection Agency, 1987) provided inadequate consideration for defining
the uncertainty in each parameter, and they suggested that bias mapping approaches be
considered. Bias mapping relates the allowable precision of a parameter to the critical values of
expected bias that just meet the specifications. A similar but less robust procedure is used in the
EPA performance specifications. Botham et al. (1991) recommended that the wind tunnel test
system duplicate the expected field sampling scenarios as closely as possible, including
characteristic flow obstructions. They described the wind tunnel testing of personal aerosol
samplers mounted on an anthropogenically consistent (e.g., breathing, heated) mannequin.
Hoffman et al. (1988) and John et al. (1991) described the adverse influence of internal surface
soiling on aerosol collection performance during extended field operation, and noted that the
4-38
-------
existing EPA PM10 performance specifications only considered clean samplers. Mark et al.
(1992) noted that even though wind tunnel performance testing cannot exactly emulate outdoor
turbulence scales, testing in the controlled tunnel environment is a necessity to adequately
characterize particle samplers.
Significant new innovations in aerosol sensing technologies that meet the PM10
performance specification and have earned designations as equivalent methods (see
Section 4.2.6) have occurred since the 1982 Criteria Document. These indirect8 methods include
automated beta attenuation monitors (e.g., Merrifield, 1989; Wedding and Weigand, 1993), and
the automated Tapered Element Oscillating Microbalance (TEOM®) technology (Patashnick
and Rupprecht, 1991). The TEOM® sampler does not use gravimetric analysis on a balance, but
computes mass based on the frequency shift as particles are deposited on an oscillating element.
These designations added automated sampling capabilities to the previously all-manual list of
sampling methods. Recent field tests of both the beta and TEOM methodologies suggest that
biases compared to gravimetrically-based samplers may exist that were not identified by the
EPA performance test requirements. Arnold et al. (1992) provide data suggesting that the mass
concentration data from a Wedding beta gauge averaged 19% lower than a collocated Wedding
PM10 gravimetric sampler. Several researchers reported that the TEOM can yield mass
concentrations that are either lower or higher than those observed in reference method
measurements (Hering, et al., 1994; Meyer, et al., 1992; Meyer et al., 1995). The TEOM
operates at an elevated temperature (30 °C or 50 °C) during the collection and measurement
process in order to ensure the removal of liquid water associated with particles. In the reference
method, the particle-associated water is removed during an equilibration period in a specified
temperature and relative humidity range. Both techniques are subject to loss of semivolatile
materials such as NH4NO3 and some organic components. The TEOM may lose semivolatile
material that is volatilized due to the higher than ambient sampling temperatures. The reference
method may lose semivolatile material during sampling (if concentrations decrease or
temperature increases during the sampling period). The reference method is also subject to loss
of semivolatile materials during equilibration and storage prior to weighing. These processes, in
areas or times during which semivolatile aerosol components are a significant component of the
8An alternate technology used instead of direct gravimetric analysis to infer mass concentrations from developed
relationships.
4-39
-------
ambient aerosol mass, can cause either technique to yield a significant under-estimation of the
mass of paniculate matter in the ambient air. This also applies to some degree to any integrated
sample collected on a substrate. Devising comprehensive performance specifications and test
procedures for aerosol samplers, given the complexities of aerosol chemistry, physics, and
mechanics, is a demanding task.
The size-selective, gravimetrically-based, 24-h manual aerosol concentration measurement
has been the mainstay of compliance sampling for at least two decades. Although several new
sensor technologies have been designated as Equivalent methods for PM10 by EPA, no superior
technology has been developed that is a better reference method than that based on collection of
a discreet aerosol sample followed by gravimetric analysis. Improvements have been made since
1982 in the accuracy and precision of integrated, manual aerosol sampling. Some of the most
significant advances have occurred in aerosol size separation technologies, improved
performance characterization test methods, and speciation sampling techniques.
As discussed by Lippmann (1993), there may be no threshold for health responses down to
the lowest aerosol concentrations. This implies that the precision and lower detection limit
requirements will continue to be important for aerosol measurements across the concentration
spectrum. These factors become even more critical as the size fraction of interest becomes
smaller and fewer total particles are collected. At low concentrations (especially with small size
fractions), normally insignificant factors can become important contributors to biases. Witz
et al. (1990) reported rapid and substantial losses of nitrates, chlorides and ammonium ion (19,
65 and 51%, respectively) from quartz high volume sampler filters during storage periods of one
week prior to analyses. Transformations can also occur on glass fiber substrates during
sampling, as reported by Sickles and Hodson (1989) for the rapid conversion of collected nitrites
to nitrates in the presence of ozone. Zhang and McMurry (1992) showed that nearly complete
evaporative losses of Fine particle nitrate can occur during sampling on Teflon filters. Lioy
et al. (1988), in a study using PM10 samplers, reported 25 to 34% lower concentration values
resulting from losses of glass fibers from the filter to the filter holder gasket during sampling.
Feeney et al. (1984) reported weight gains in Teflon filters used in contaminated ring cassettes,
that posed significant problems for light aerosol loadings. Grinshpun et al. (1993) suggest that if
unavoidable changes in the aerosol occur during sampling, development of a model that permits
back-calculation of the in situ characteristics can be considered.
4-40
-------
4.2.6 Reference and Equivalent Method Program
Ambient air PM10 measurements are used (among other purposes) to determine whether
defined geographical areas are in attainment or non-attainment with the National Ambient Air
Quality Standards (NAAQS) for PM10. These measurements are obtained by the States in their
state and local air monitoring station (SLAMS) networks as required under 40 CFR Part 58.
Further, Appendix C of Part 58 requires that the ambient air monitoring methods used in these
EPA-required SLAMS networks must be methods that have been designated by the EPA as
either reference or equivalent methods.
Monitoring methods for particulate matter (i.e., PM10) are designated by the EPA as
reference or equivalent methods under the provisions of 40 CFR Part 53, which was amended in
1987 to add specific requirements for PM10 methods. Part 53 sets forth functional specifications
and other requirements that reference and equivalent methods for each criteria pollutant must
meet, along with explicit test procedures by which candidate methods or samplers are to be
tested against those specifications. General requirements and provisions for reference and
equivalent methods are also given in Part 53, as are the requirements for submitting an
application to the EPA for a reference or equivalent method determination. The distinction
between reference and equivalent methods is a technical one. On one hand, it provides for
detailed, explicit specification of a selected measurement technology for reference methods. On
the other hand, it allows alternative (including innovative and potentially improved)
methodologies for equivalent methods, based only on meeting specified requirements for
functional performance and for comparability to the reference method. For purposes of
determining attainment or non-attainment with the NAAQS, however, the distinction between
reference and equivalent methods is largely, if not entirely, immaterial.
Under the Part 53 requirements, reference methods for PM10 must be shown to use the
measurement principle and meet the other specifications set forth in 40 CFR 50, Appendix J
(Code of Federal Regulations, 1991). They must also include a PM10 sampler that meets the
requirements specified in Subpart D of 40 CFR 53. Appendix J specifies a measurement
principle based on extracting an air sample from the atmosphere with a powered sampler that
incorporates inertial separation of the PM10 size range particles followed by collection of the
PM10 particles on a filter over a 24-h period. The average PM10 concentration for the sample
period is determined by dividing the net weight gain of the filter over the sample period by the
4-41
-------
total volume of air sampled. Other specifications are prescribed in Appendix J for flow rate
control and measurement, flow rate measurement device calibration, filter media characteristics
and performance, filter conditioning before and after sampling, filter weighing, sampler
operation, and correction of sample volume to EPA reference temperature and pressure. Also,
sampler performance requirements in Subpart D of Part 53 include wind tunnel tests for
"sampling effectiveness" (the efficacy of the PM10 particle size separation capability) at each of
three wind speeds and "50 percent cutpoint" (the accuracy of the primary 10-micron particle size
separation). Field tests for sampling precision and flow rate stability are also specified. In spite
of the instrumental nature of the sampler, this method is basically a manual procedure, and all
designated reference methods for PM10 are therefore defined as manual methods.
Equivalent methods for PM10, alternatively, need not be based on the measurement
principle specified in Appendix J nor meet the other Appendix J requirements. Instead,
equivalent methods must meet the "sampler" performance specifications set forth in Subpart D
of Part 53 and demonstrate comparability to a reference method as required by Subpart C of Part
53. The provisions of Subpart C specify that a candidate equivalent method must produce PM10
measurements that agree with measurements produced by collocated reference method samplers
at each of two field test sites. For this purpose, agreement means a regression slope of 1 ± 0.1, a
regression intercept of 0 ± 5 |ig/m3, and a correlation >0.97. These requirements allow virtually
any type of PM10 measurement technique, and therefore an equivalent method for PM10 may be
either a manual method or a fully automated instrumental method (i. e., analyzer).
As of this writing, the EPA has designated seven reference methods and three equivalent
methods for PM10, as listed in Table 4-1. The reference methods include four methods featuring
high-volume samplers from two manufacturers, with one using a cyclone-type size separator and
the others using an impaction-type separator. The other reference methods include a low-
volume sampler (from a third manufacturer), a low-volume sampler featuring a secondary size
separation at 2.5 microns (dichotomous sampler), and a medium-volume, non-commercial
sampler. The three designated equivalent methods are all automated PM10 analyzers and include
two operating on the beta-attenuation principle and one based on a tapered element oscillating
microbalance (TEOM™). It should be noted that although these latter three automated PM10
analyzers may be capable of providing continuous or semi-continuous PM10 concentration
4-42
-------
measurements, only 24-h average PM10 measurements are recognized as approved under their
equivalent method designations.
4.2.7 Determination of Size Distribution
The determination of aerosol size distributions can be a powerful research tool when
studying source contributions and transformation processes. A number of techniques are
available as described by texts such as Willeke and Baron (1993) to make near real-time, single
particle aerosol measurement in addition to cascade impactors.
4.2.7.1 Cascade Impactors
In cascade applications, the aerosol is impacted and trapped onto a series of removable,
coated substrates (e.g., greased foils), including a final total stage collection on a filter for
gravimetric analysis. Marple et al. (1993) list over 30 single stage and cascade impactors that
are either commercially available or still commonly used. The design and calibration of a
miniature eight-stage cascade impactor for personal air sampling in occupational settings is
described by Rubow et al. (1987), operating at 2.0 1pm. Evaluations of the most commonly used
cascade impactor systems have been reported by Vaughan (1989) for the Andersen MK1 and
MK2 7-stage cascade impactors, Marple et al. (1991) for the 10-stage Micro-Orifice Uniform
Deposit Impactor (MOUDI), and Wang and John (1988) and Hillamo and
4-43
-------
TABLE 4-1. U.S. ENVIRONMENTAL PROTECTION AGENCY-DESIGNATED REFERENCE
AND EQUIVALENT METHODS FOR PM,n
Method No.
Identification
Description
Type
Date
RFPS-1087-062
RFPS-1287-063
RFPS-1287-064
RFPS-1287-065
Wedding & Associates PM Critical
Flow High- Volume Sampler.
High^volume (1.13 mVmin) sampler with cyclone-
type PM10 inlet; 203 x 254 cm (8 x 10 in) filter.
Manual reference
method
Sierra-Andersen or General Metal High-volume (1.13 mVmin) sampler with impaction- Manual reference
Works Model 1200 PM10 High-
Volume Air Sampler System
Sierra-Andersen or General Metal
Works Model 321-B PM10 High-
Volume Air Sampler System
Sierra-Andersen or General Metal
Works Model 321-C PM10 High-
Volume Air Sampler System
type PM10 inlet; 203 x 254 cm (8 x 10 in) filter.
method
High-volume (1.13 mVmin) sampler with impaction- Manual reference
type PM10 inlet; 203 x 254 cm (8 x 10 in) filter. (No method
longer available.)
High-volume (1.13 mVmin) sampler with impaction- Manual reference
type PM10 inlet; 203 x 254 cm (8 x 10 in) filter. (No method
longer available.)
10/06/87
12/01/87
12/01/87
12/01/87
RFPS-0389-07 1 Oregon DEQ Medium Volume PM
Sampler
Non-commercial medium -volume (110 L/min) Manual reference
sampler with impaction-type inlet and automatic filter method
change; two 47-mm diameter filters.
3/24/89
RFPS-0789-073
EQPM-0990-076
Low-volume (16.7 L/min) sampler with impaction-
type PM10 inlet; additional particle size separation at
2.5 micron, collected on two 37-mm diameter filters.
Manual reference
method
7/27/89
Sierra-Andersen Models SA241 or
SA241M or General Metal Works
Models G241 and G241M PM10
Dichotomous Samplers
Andersen Instruments Model FH62I- Low-volume (16.7 L/min) PM10 analyzers using Automated equivalent 9/18/90
impaction-type PM10 inlet, 40 mm filter tape, and beta method
attenuation analysis.
N PM10 Beta Attenuation Monitor
-------
TABLE 4-1 (cont'd). U.S. ENVIRONMENTAL PROTECTION AGENCY-DESIGNATED REFERENCE
AND EQUIVALENT METHODS FOR PM,n
Method No.
Identification
Description
Type
Date
EQPM-1090-079
EQPM-0391-081
Rupprecht & Patashnick TEOM
Series 1400 and Series 1400a
PM10 Monitors
Wedding & Associates PM
Gauge Automated Particle
Sampler
Beta
Low-volume (16.7 L/min) PM10 analyzers using
impaction-type PM10 inlet, 12.7 mm diameter filter,
and tapered element oscillating microbalance
analysis.
Low-volume (16.7 L/min) PM10 analyzer using cy-
clone-type PM10 inlet, 32 mm filter tape, and beta
attenuation analysis.
Automated equivalent 10/29/90
method
Automated equivalent 3/5/91
method
RFPS-0694-098
Rupprecht & Patashnick Partisol
Model 2000 Air Sampler
Low-volume (16.7 L/min) PM10 samplerwith impac-
tion-type inlet and 47 mm diameter filter.
Manual reference
method
7/11/94
-------
Kauppinen (1991) for the 6-stage Berner, low pressure cascade impactor. The smallest particle
stages of these impactors can have very small diameter jets and/or very low total pressures to
achieve the sub-micron separations. The MOUDI impactor has 2000 holes on the lowest
cutpoint stage. Raabe et al. (1988) describe an 8 stage cascade slit impactor with slowly rotating
impactor drums instead of flat plates. This arrangement, in combination with a PIXIE analyzer,
permitted aerodynamic sizing of elemental components, with temporal resolution. The skill and
care required in the operation of cascade impactors suggests that they are research rather than
routine samplers.
The importance of the aerosol calibration of a cascade impactor is illustrated by Vaughan
(1989) in Figure 4-13, which compares the experimental data with the manufacturer's
calibrations and indicates biases as large as 1.0 //m. Marple et al. (1991) provided a similar type
of stage calibration for the MOUDI impactor and included data on the internal particle losses
(see Figure 4-14). These loss data showed that an improperly designed inlet to the impactor,
combined with the inertial and interception losses of the larger particle sizes, can substantially
bias the first stage collections. This was also demonstrated for the inlet to the Andersen
impactor by McFarland et al. (1977). Cascade impactors that cover wide particle size ranges
inherently require design compromises among competing factors, including cutpoint sharpness,
internal stage losses and the physical size of the device.
Cascade impactors can be used to construct distributions of mass and speciated constituents
as a function of aerodynamic diameter. These distributions can be constructed graphically or
using matrix inversion techniques. Marple et al. (1993) notes that impactor stage calibrations
which do not demonstrate sharp cutoffs can cause significant between-stage sizing errors if not
accommodated. John et al. (1990) measured distributions over the 0.08 to 16 //m range for mass
and inorganic ions for several sites in Southern California. They identified the standard coarse
mode and two separate, previously unreported modes in the 0.1 to 1.0 //m range. This latter
range was referred to by Whitby (1978) as a single "accumulation" mode. John et al. (1990)
described a "condensation" mode at 0.2 ±0.1 //m containing gas phase reaction products, and a
"droplet" mode at 0.7 ± 0.2 //m which grows from the "condensation" mode by the addition of
water and sulfates. Fang et al. (1991) described the effects of flow-inducted relative humidity
changes on the sizing of acid aerosols in the MOUDI impactor. They noted that it may not be
possible to measure size
4-46
-------
0.1
/
Measured
calibration
Manufacturer's
' calibration
pactor - Stage
0.2 0.3 0.5 0.7 1 23 5710
Aerodynamic Diameter (|jm)
20 30 ' 50'70 100
Figure 4-13. Measured calibration of the Andersen Cascade Dupactor as compared to that supplied by the manufacturer.
Source: Vaughan (1989).
-------
J^.
I
oo
100
80
(0
o
u
£40
(0
a.
20
Liquid Particles
Solid Particles
10
Stage—3
Cut-Points
8
StageNumber
6
1 Inlet
T -i-rm
0.01
0.1
1
10
100
Aerodynamic Particle Diameter (urn)
Figure 4-14. Internal losses for the MOUDI impactor.
Source: Marple et al. (1991).
-------
distributions of small (less than about 0.2 to 0.5 //m) particles with impactors at relative
humidities exceeding 80%.
4.2.7.2 Single Particle Samplers
Aerosol size distribution data are useful for studies of particle transport and transformation
processes, source characterization, and particle sizing and collection device performance. In
addition to cascade impactors, a number of real time or near real time sizing instruments are
available and described in texts such as Willeke and Baron (1993). While cascade impactors
provide distributions in terms of aerodynamically sized mass, single particle sampling devices
can produce optically sized distributions as a function of particle number (count), with surface
area and volume distributions computed during the data reduction, assuming spherical particles.
Particle density and shape information as a function of size are required to convert from volume
distributions to an estimated mass basis. Individual particle sizing and counting instruments are
generally limited to a particle detection range of a decade or so, but several devices can overlap
to cover the range of approximately 0.001 to 10 //m. The principle of detection of an instrument
restricts the particle sizes which can be detected. For example, instruments using electrical
mobility analysis are limited to particle sizes less than about 1 //m. Optical methods are
typically used to measure particles larger than about 0.1 to 0.3 //m. Inlet and transport system
losses of coarse particles above about 2 //m, prior to the sensing volume, must be factored into
reported size distributions.
The three most commonly used single particle sampler types are aerodynamic particle
sizers, electrical mobility analyzers and optical particle counters (OPC's). Aerodynamic particle
sizers use laser doppler anemometry to measure the velocity of particles in a jet. The
acceleration of the particle is related to the aerodynamic particle diameter. This technique is
typically applied to particles larger than about 0.5 //m. In electrical mobility analysis, aerosol
with a known charge distribution flows through an electric field. The particles migrate
according to their mobility which can be related to size. The original TSI electrical aerosol
analyzer (EAA) performed this separation in an integrated manner over the total size
distribution and detected the particles by unipolar diffusion charging. A more versatile
approach, the differential mobility analyzer or DMA (Liu et al., 1978), is able to examine a
narrow slice of the size distribution in an equilibrium charge state, detected by a condensation
4-49
-------
nucleus counter (CNC). Differential mobility analyzers have been employed in pairs (Tandem
Differential Mobility Analyzer, or TDMA) to examine both (a) particle characteristics such as
NH3 and H2SO4 reaction rates (McMurry et al., 1983) and (b) the sensitivity of the size
distributions of Los Angeles aerosol to relative humidity (McMurry and Stolzenburg, 1989).
The latter research used the first DMA to select particles of known mobility from the input
aerosol, a humidification system to condition the selected particles, and the second DMA to
determine mobility changes. Optical particle counters pass a jet of aerosol through an optical
system. Light scattered from individual particles is detected and the signal in processed in a
multi-channel analyzer. Discreet signals are counted and sorted by intensity and by optical size.
One example of a forward-scattering counter with an open sensing volume (for use on aircraft) is
the Particle Measuring Systems, Inc., FSSP-300, which can provide high resolution (31 channel)
count distributions over the size range of 0.3 to 20 //m (Rader and O'Hern, 1993). Gebhart
(1993) described currently available OPC's and their counting efficiencies over a range of
diameters.
Single particle samplers have common considerations, as dicussed below.
Calibration: They are calibrated with reference aerosol either by the manufacture or by the
user. If the properties of the aerosol measured are quite different than the calibration, the
indicated size distribution may be quite different than actual distribution. Brockman et al.
(1988) demonstrated that the APS calibration can vary significantly with the type of test aerosol
and showed substantial response biases between oleic acid and polystyrene latex spheres above
10 (j,m. Wang and John (1989) described a procedure to correct the APS response for aerosol
particle density. Particle shape can also provide serious sizing errors, and specific calibrations
are needed for particles with shape factors significantly different from unity (spherical). Yeh
(1993) commented that the calculated geometric standard deviations (og) determined by the
EAA and DMA are generally larger than 1.3, even if the correct value is significantly closer to
unity. Woskie et al. (1993) observed, as did Willeke and Degarmo (1988), that optical particle
counting devices must be appropriately calibrated using realistic aerosols, especially for low
concentration applications. Harrison and Harrison (1982) suggested that the ratio of fine particle
mass concentration to optical scattering extinction will be more variable when a significant
contribution is made by irregular (shaped) particles - an event likely to occur when the mean
mass diameter exceeds 1 //m.
4-50
-------
Particle Concentration Effects: Gebhart (1993) noted that the response of single particle
counters may be influenced by extremely high particle concentrations. Wake (1989) and
Heitbrink et al. (1991) described the coincidence problems of the APS when sampling high total
particle concentrations, especially for sizes greater than 1 //m. Baron et al. (1993) reported that
the concentration levels giving 1% coincidence in an aerodynamic particle sizer for 0.8, 3 and 10
//m particles, respectively, are the relatively low values of 558, 387 and 234 particles/cm3.
Optical particle counters can experience coincidence errors (two particles are detected as a single
particle) and counter saturation at high particle concentrations. Hinds and Kraske (1986)
described the performance of the PMS, Inc. LAS-X and noted a sizing accuracy of ±2 channel
widths, with coincidence errors of less than 10% for concentrations below 10,000 particles/cm3.
Clearly, typical particle concentrations found in the atmosphere may produce significant errors if
sample dilution is not utilized.
4.2.8 Automated Sampling
Automated methods to provide measures of aerosol concentrations in the air have existed
for decades in an attempt to provide temporal definition of suspended particles and enhance
every-sixth-day sampling schedules with a minimum labor expense. Arnold et al. (1992)
collected daily 24-h PM10 samples with an automated monitor and noted that 80% of the highest
10 daily concentrations between 1989 and 1990 were not encountered by the every-sixth-day
sampling schedule. Some of the automated samplers (e.g., British Smoke Shade and AISI tape
samplers) described in the 1982 Criteria Document were indicator measures of aerosol
concentration, using calibrations relating aerosol concentrations to reflected or absorbed light.
Tape samplers were used in the U. S. primarily as exceedance (index) monitors.
The beta attenuation and integrating nephelometer techniques described in the 1982
Criteria Document primarily were research methods. Since that time, the beta gauge sampling
approach has been refined and a new approach, based on the Tapered Element Oscillating
Microbalance (TEOM) principle, has been developed. Samplers based on these techniques have
been designated as equivalent methods for PM10.
Although one could be readily constructed, there are presently no commercially available,
automated high volume (> 1 m3/min flowrate) aerosol samplers, excluding the possibility of the
timed operation of an array of manual samplers. The physical size of such a sampling system
4-51
-------
using 8x10 inch filters is impractical. The dichotomous sampler is currently the only low
volume, gravimetrically-based sampler commercially available in an automated version.
4.2.8.1 Smoke Shade (British Smoke, Black Smoke)
Historically, the British smoke shade sampler was one of the earliest ambient PM sampling
devices to be developed and to gain widespread use as an automated optical PM monitoring
method. Key features and limitations of the British or black smoke (BS) method were discussed
in EPA's 1982 Criteria Document. As indicated in Chapters 3 and 14 of that Criteria Document,
the BS method typically involves use of a sampler that draws ambient air through an inverted
funnel and approximately 3m of plastic tubing to deposit collected particles on white filter paper.
The amount of PM deposited during a given time period (e.g., 1-h during severe episodes, or
more typically, 24-h) is determined by measuring the blackness of the stain on the filter paper.
An automated version of the sampler can collect daily samples sequentially for up to eight days.
It is important to note, as described in the 1982 Criteria Document, that the BS method and
its variations (e.g., the OECD version) in routine use typically employ standard monitoring
equipment with a D50 cutpoint=4.5 //m, which mainly allows fine-mode particles and small
coarse mode particles (some ranging up to ~8 to 10 //m) to be collected. Thus, regardless of
whether larger particles are present in the atmosphere, the BS method collects predominately
small particles. Also, the BS method neither directly measures mass nor determines chemical
composition of the collected PM. Rather, it measures light absorption of particles as indicated
by reflectance from the stain formed by the particles collected on the filter paper, which depends
both on the density of the stain, or amount of PM collected, and the optical properties of the
collected PM. Smoke particles composed of elemental carbon, found in incomplete fossil-fuel
combustion products, typically make the greatest contribution to the darkness of the stain,
especially in urban areas. Thus, the amount of elemental carbon, but not organic carbon, present
in the stain tends to be most highly correlated with BS reflectance readings. Other nonblack,
noncarbon particles also have optical properties such that they can affect the reflectance
readings, although their contribution to optical absorption is usually negligible.
Since the relative proportions of atmospheric carbon and noncarbon PM can vary greatly
from site to site or from one time to another at the same site, the same absolute BS reflectance
reading can be associated with markedly different amounts (or mass) of collected particles or, in
4-52
-------
unusual circumstances, even with markedly different amounts of carbon. Site-specific
calibrations of reflectance readings against actual mass measurements obtained by collocated
gravimetric monitoring devices are therefore necessary to obtain estimates of atmospheric
concentrations of particulate matter based on the BS method. A single calibration curve relating
mass or atmospheric concentration (in //g/m3) of parti culate matter to BS reflectance readings
obtained at a given site may serve as a basis for crude estimates of the levels of PM (mainly
small particles) at that site over time, so long as the chemical composition and relative
proportions of elemental carbon and noncarbon PM do not change substantially. However, the
actual mass or smoke concentrations present at a particular site may differ markedly (by factors
of two or more) from the values calculated from a given reflectance reading on either of the two
most widely used standard curves (the British and OECD standard smoke curves)9. Thus, great
care must be taken in interpreting the meaning of any BS value reported in terms of//g/m3,
especially as employed in the British and other European epidemiological studies discussed in
Chapter 12 of this document.
There has existed long standing interest with regard to relationships between ambient PM
concentrations indexed by BS readings (based on conversion of reflectance values to estimated
Mg/m3 concentrations by means of standard calibration curves) and those obtained by gravimetric
methods. The 1982 Criteria Document noted that Ball and Hume (1977) and Waller (1963)
found that such relationships are site, season, and particle-source dependent. Also, Lee et al.
(1972) noted, from collocated TSP hi-vol and smoke shade sampler comparisons made at
various sites in England, that the overall correlation coefficients between these measurements for
all sites was 0.618. However, the individual coefficients ranged from 0.936 (good correlation)
to 0.072 (no correlation). Bailey and Clayton (1980) showed that smoke shade measurements
correlated more closely with soot (elemental carbon) content than with gravimetric mass. Other
work by Paschel and Egner (1981) and Clayton and Wallin (1982) showed consistently higher
TSP values than BS readings (converted to //g/m3) from collocated samplers in various U.S. and
9For this reason, smoke data reported in ,ug/m3 based on either the British or OECD Standard curve are appropriately
interpreted in terms of "nominal" //g/m3 smoke units and cannot be accepted as accurate estimates of airborne PM mass
unless corroborated by local site-specific gravimetric calibrations. In other words, unless based on local site-specific
calibrations, smoke readings in //g/m3 cannot yield quantitative estimates of atmospheric PM concentrations. In the
absence of such calibrations, smoke readings only allow for rough qualitative (i.e., <; =; or >) comparisons of amounts
of PM present at a given time versus another time at the same site and do not permit meaningful comparisons between
PM levels at different geographic areas having airborne PM of different chemical composition (especially in terms of
relative proportions of elemental carbon).
4-53
-------
U.K. cities, respectively (as would be expected given that the BS measurements of fine and
small coarse mode particles typically represent only some fraction of the wider range of particles
sampled by TSP measurements). Clayton and Wallin (1982), not surprisingly, also found widely
variable ratios of TSP to BS readings from different U.K. cities reflecting the varying
proportions of small particles present in the total ambient mix of particles at different sites.
Likewise, varying (site- and season-dependent) relationships between BS measurements and
ambient PM measurements made by various gravimetric methods have been reported in the
Federal Republic of Germany (Laskus, 1983) and in the semi-arid climate of Baghdad, Iraq
(Kanbour et al., 1990). Lastly, Muir and Laxton (1995) reported that, for Bristol (a moderate
size U.K. city), daily average BS (averaged over six urban background sites) appears to be a
reasonable predictor of daily average PM10 and daily 1-h peak PM10 values; but different
relationships apply for winter versus summer, indicating that BS and PM10 measure different
components of airborne PM (i.e., BS may be a better index of fine-mode particles than PM10,
which has a D50 cutpoint oFlO //m).
Only limited examples exist of derivation of models of interrelationships between BS
readings and gravimetric measurements for particular time periods in a given location. For
example, see Mage (1995) for discussion of an empirical model relating BS to TSP values
during London winters of the 1950s and 1960s.
4.2.8.2 Coefficient of Haze (AISI/ASTM Tape Sampler)
The 1982 Criteria Document also described a second type of automated optical PM
measurement methods. Developed before 1940, the American Iron and Steel Institute (AISI)
light transmittance method is similar in approach to the BS technique and has been employed for
routine monitoring in some American cities. The instrument collects particles with a D50
cutpoint of=5.0 //m aerodynamic diameter and uses an air intake similar to that of the BS
method. Ambient PM is collected on a filter-paper tape that is periodically advanced to allow
accumulation of another stain. Opacity of the stain is determined by transmittance of light
through the deposited material and the tape. The results are expressed in terms of optical density
or coefficient of haze (CoH) units per 1,000 linear feet of air sampled (rather than in mass units).
Readings in CoH units are somewhat more responsive to noncarbon particles than are BS
measurements; but, again, the AISI method neither directly measures mass nor determines
4-54
-------
chemical composition of the PM collected. Any attempt to relate CoHs to Mg/m3 requires site-
specific calibration of CoH readings against mass measurements determined by a collocated
gravimetric device, but the accuracy of such mass estimates are still subject to question.
Few attempts have been reported on calibration of COH measurements versus results from
collocated gravimetric devices. One notable attempt (Ingram, 1969; Ingram and Golden, 1973)
was reported for New York City, but the results are of very limited applicability to New York
City aerometric data of the 1960's. Also, Regan et al. (1979) showed that CoH readings
correlate favorably with gravimetric measurements limited to smaller particle sizes. Edwards
(1980) and Edwards et al. (1983) have also shown that BS reflectance measurements can be
related to the absorption coefficient of the atmosphere and that BS measurements can be
converted to approximate CoH measurements made by AISI tape sampler using the absorption
coefficient relationships. As several investigators noted, (e.g., Lodge, et al., 1981), if a
relationship could be developed between optical and gravimetric measurements, it would be site
specific, but still variable because of seasonal and long-term differences in the sources of
collected particle size fractions and their carbon content.
4.2.8.3 TEOM® Sampler
The Tapered Element Oscillating Microbalance (R & P, Inc.) sensor, as described by
Patashnick and Rupprecht (1991), consists of an oscillating tapered tube with a filter on its free
end (see the diagram in Figure 4-15). The change in mass of the filter and collected aerosol
produces a shift in the oscillation frequency of the tapered tube that is directly related to mass.
Rupprecht et al. (1992) suggested that the filter can be archived after sampling for
4-55
-------
Flow-
Flow
Sampling Head
Heated Air Inlet
Filter Cartridg
Tapered Element
Electronic
Feedback System
'
Microprocessor
to Flow Controller
Figure 4-15. Rupprecht and Patashnick TEOM® sampler.
Source: Patashnick and Rupprecht (1991).
4-56
-------
subsequent analysis. The sampler inlet has a PM10 outpoint and operates at 16.67 1pm. A flow
splitter samples a 3 1pm portion of this flow to be filtered. Since the fraction of volatile species
(e.g., water, nitrates, organics) in the aerosol is a function of ambient temperature, the TEOM®
sampler heats the inlet air stream to a constant 30 or 50 °C to keep moisture in the vapor phase.
The mass transducer is also heated to 50 °C to stabilize the measurement process. Operation
with the flow stream heated to a lower temperature (e.g., 30 °C) is possible, but care must be
taken to avoid moisture condensation that will confound the measurement. The transducer is
also heated to 50 °C to stabilize the mass measurement. A factory calibration regression is used
to electronically correct the computed mass from the TEOM® sampler to that measured by a
reference PM10 sampler.
Although several studies (e.g. Patashnick and Rupprecht, 1991; Kalthoff and Grumpier,
1990) have shown consistent and linear relationships between the TEOM® sampler and
gravimetric PM10 samplers, a number of studies have shown biases under certain conditions.
Several researchers, including Cahill et al. (1994), Hering (1994) and Meyer et al. (1992) have
reported that the modification of the aerosol by the elevated operating temperature appears to
have a significant effect (loss) on mass concentration. Meyer et al. (1992) collocated a TEOM®
sampler with an PM10 SA1200 gravimetric sampler in Mammoth Lakes, CA during a winter
heating season (heavy wood stove usage). The regressions between the TEOM® sampler and
PM10 sampler gave strong correlations (r2 > 0.98), with slopes of 0.55 for operation at 50 °C,
and 0.66 for operation at 30 °C. The negative bias of the TEOM was attributed primarily to
losses of semi-volatile organics from the filter. Cahill et al. (1994) reported that the TEOM®
sampler showed biases on the order of 30% low and poor correlations with PM10 samplers in
dry, dusty conditions. The reasons for this discrepancy were unknown. The field comparison
data of Patashnick and Rupprecht (1990) showed near unity (1 ± 0.06) regression slopes for the
TEOM with the Wedding IP10 and Sierra-Andersen dichotomous samplers in El Paso, TX and
Birmingham, AL. Since aerosol composition is highly dependent on local sources and
meteorology, volatilization losses could be expected to be site- and season-dependent. This
could significantly affect the rigor of collocated field sampling. A WESTAR (1995) council
report summarizes the relationships between TEOM® monitors and other direct gravimetric
samplers in at least 10 states in the western U.S. This report concluded that on average the
TEOM® sampler concentrations were 21.8% lower than other collocated PM10 samplers for
4-57
-------
concentrations > 50 //g/m3. This would significantly affect the TEOM® sampler's ability to be
used as a "trigger" monitor for control strategy plans. More data are needed to determine the
implications of these problems on the ability of the TEOM® sampler to be used in a regulatory
setting. Although it is clear that the TEOM® sampler can provide PM10 data comparable to the
existing reference method samplers, the specific field sampling conditions where excessive bias
might be expected to occur have not been completely defined. A portion of the bias is
undoubtedly due to concomitant variabilities in the associated gravimetric measurements.
4.2.8.4 Beta Gauge
The Andersen FH 62I-N beta attenuation sampler was described by Merrifield (1989) and
uses a 30 mCi Krypton-85 source and detector to determine the attenuation caused by deposited
aerosols on a filter (see diagram in Figure 4-16). To improve the stability over time, a reference
reading is periodically made of a foil with an attenuation similar to that of the filter and collected
aerosol. The Wedding beta attenuation sampler was described by Wedding and Weigand (1993)
and uses a 100 mCi 14C source. Both samplers have inlets with a PM10 cutpoint, with the
Andersen sampler operating at 16.67 1pm and the Wedding at 18.9 1pm. The filter material is
contained on a roll and advances automatically on a time sequence, or when a preset aerosol
loading is reached. An automatic beta gauge sampler was also described by Spagnolo (1989),
using a 15 //m inlet and a 14C source. The calibration of a beta gauge is site specific, and a
calibration regression must be processed electronically to provide accurate mass readings.
Rupprecht et al. (1992) suggested that the closer link between deposited mass and frequency
shift for the TEOM principle should provide less site-specific response, compared to the aerosol
compositional sensitivity of the beta gauge technique.
Arnold et al. (1992) provided data over a 2 year period in Denver, CO for the mass
concentration regression data from a Wedding beta gauge, showing a range of correlations
(r2 from 0.72 to 0.86), varying by sampler and season. The authors suggested that installation of
a newer technology beta gauge accounted for the higher correlations, but noted that unexplained
outliers resulted in poorer than expected results. The regression slopes between the two sampler
types showed that the beta gauge averaged 19% lower than a
4-58
-------
Measuring Chamber
Compensation Chamber
Chamber for Dust Precipitatic
and Measurement
30 m Ci KR-85 Source
Filter Feed Spool
Filter Takeup
High-Voltage Power Supply
n
mperature / Pressure
Rotary Vane Pump
Bit
I/O
50-Pin
Connector
V24/RS232
Figure 4-16. Andersen beta gauge sampler.
-------
collocated Wedding PM10 gravimetric sampler. It should be noted that the Wedding PM10 inlet
has typically been reported (see Section 4.2.2.4) to be 10 to 15% lower in collocated field tests
with Sierra-Andersen PM10 inlets. A WESTAR (1995) council report summarizes the
relationships between beta gauge monitors and other direct gravimetric samplers in at least five
states in the western U.S. This report concluded that on average beta gauge concentrations were
8.6% lower than other collocated PM10 samplers for concentrations > 20 //g/m3. Field data from
Wedding and Weigand (1993) at two sites (Fort Collins, CO and Cleveland, OH) using the same
samplers produced regressions exhibiting strong correlations (r2 = 0.99) with no apparent outliers
and a composite slope of 1.00. Arnold et al. (1992) operated the PM10 high volume samplers
on the required every-6th-day schedule and the beta attenuation monitors continuously, and
noted that only 22.5% of the exceedance days, as measured by the beta monitor, were
operational days for the high volume samplers.
4.2.8.5 Nephelometer
The integrating nephelometer is commonly used as a visibility monitor; it measures the
light scattered by aerosols, integrated over as wide a range of angles as possible. A schematic
diagram of the integrating nephelometer is shown in Figure 4-17 (from Hinds, 1982). The
measured scattering coefficient of particles, bsp, can be summed with the absorption coefficient,
bap, and the comparable coefficients for the gas phase to compute the overall atmospheric
extinction coefficient, bext. Methods for estimating absorption and extinction for atmospheric
particles are discussed in 8.2.2. The atmospheric extinction has been related to visibility as
visual range. The particle scattering coefficient is dependent upon particle size, index of
refraction and illumination wavelength, as shown by Charlson et al. (1968) in Figure 4-18, while
the absorption coefficient is relatively independent of size. The field calibration of
nephelometers has historically been based on the refractive index of Freon-12 (and occasionally
carbon dioxide), but newer calibration procedures using atomized sugar aerosols have been
proposed (Horvath and Kaller, 1994) as more environmentally conscious. Nephelometry over a
narrow wavelength band or at a selected wavelength can be applied to measure the laser light
scattered from a volume of aerosol containing a number of
4-60
-------
Power
Supply
Flash Tube
Power Supply
Aerosol
Outlet
hotomultiplie
Tube
Scattering
Volume
Collimating
Disks
Aerosol
Inlet
Amplifier
Recorder
Clean Air
Purge
Figure 4-17. Integrating nephelometer.
-------
10
CO
E
o
_
E
o
J^.
I
to
-o
n
o
M
w
ja
ra
10
-2
\ I I I
n n r
Scattering
10
-1
10
4.00
Diameter (pm)
Figure 4-18. Particle-scattering coefficient per volume concentration as a function of particle size for spherical particles of
refractive index 1.5 illuminated by 550 nm light.
Source: Charlsonet al. (1968).
-------
particles. Gebhart (1993) described devices such as the MIE, Inc.10. MINTRAM, often used in
portable applications to estimate real-time aerosol concentrations. Cantrell et. al. (1993) showed
that MINIRAM calibration was significantly different for diesel and mine aerosols. Woskie
et al. (1993) described the performance of a MINIRAM (using the manufacturer's calibration)
against gravimetric borate concentrations for particles as large as 30 //m, and found significant
biases (a regression slope = 4.48). This bias was expected, since the large mass median particle
diameters were substantially outside the respirable particle range recommended by the
manufacturer.
The relative insensitivity of the nephelometer to particles above ~2 //m results in poor
correlations with PM10 mass. Larson et al. (1992) showed strong correlations (r2 = 0.945)
between bsp and fine fraction mass (see Figure 4-19) for a woodsmoke impacted neighborhood
near Seattle, WA, with a slope of 4.89 m2/g. They noted that this slope fell within the range of
values reported by others and was predicted by Mie scattering theory. The slope of the Larson
et al. (1992) data could be compared with other site-specific calibrations, such as the data of
Waggoner and Weiss (1980), which gave a composite slope of 3.13 m2/g, characterized by the
authors as representative of a "wide range" of sites. Lewis (1981) provided an analysis of the
relationships of the features of the ambient size distribution to bsp. The inlet air stream to the
nephelometers for the latter data was heated from 5 to 15 °C above background. Rood et al.
(1987) conducted a controlled comparison of the influence of aerosol properties on bsp in
Riverside, CA and reported a regression slope against fine mass (defined as less than 2.0 //m) of
2.1 m2/g with an r2 value of 0.92. In this experiment, the relative humidity for bsp determinations
was controlled to less than 35% and the gravimetric filter substrate was nylon. The authors
attributed the smaller than normal slope reading to possible nitrate evaporation from the filtered
aerosol and artifact reactions with the nylon substrate material. Thomas et al. (1993)
demonstrated that the influence of relative humidity on the relationship between photometer
response and collocated gravimetric particle concentrations can be predicted.
The data scatter in Figure 4-19 (if assumed to be typical of such comparisons) would
suggest that fine particle mass concentration estimates from bsp values were typically within 5 to
7 //g/m3 of the gravimetrically determined values. To be useful as a surrogate measure
'"Bedford, MA.
4-63
-------
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
Lake Forest Park
Weekly Average Values
January 17,1991 to December 19, 1991
Slope = 4.89 m /g
R2 = 0.945
10
15
20
25
30
35
40
45
PM2.5(ugm-3)
Figure 4-19. Correlation of bsp and fine fraction mass.
Source: Larson et al. (1992).
for mass concentration, the site-specific nephelometer calibration should be valid for a wide
range of situations, especially during episodes where the concentration levels approach or exceed
an action limit. The scattergram of bsp versus fine particle mass provided by Rood et al. (1987)
showed much greater variability, with a given bsp value providing an estimated 20 to 25 //g/m3
concentration range. They noted that metastable H2O contributed 5 to 20% of the total particle
light scattering coefficient, especially during the late afternoon and early evening. The
4-64
-------
precisions and biases of the dependent and independent variables between bsp and fine mass
concentration are not constants, since at least one factor - moisture content of the aerosol -
affects both measures. The gravimetric sample filters are typically equilibrated to a specific
relative humidity range (e.g., 40 to 60%) to normalize the tare weighings.
Sloane (1986) and others have noted that light scattering from particles is not solely a
function of mass but are also very dependent on a summation of the scattering coefficients of
each species. The scattering cross section of a particle is dependent on the water content, and,
hence, the relative humidity in situ. Pre-heating of the inlet air of the nephelometer normalizes
the response to water content, but biases the reading relative to the in situ case. Sloane (1986)
also gave the computed and measured scattering coefficients for ammonium sulfate and noted
that chemical interactions can cause a two-fold variation in scattering response to a change in the
mass of hygroscopic constituents. It was also observed that the light scattering efficiency of an
aerosol such as ammonium acid sulfate is not a constant, but varies with the overall aerosol
composition. Eldering et al. (1994) developed and validated a predictive model for bsp in
Southern California. This model used composite size distributions constructed from a TSI, Inc.11
EAA, a PMS, Inc.12 LAS-X and a Climet, Inc.13 multi-channel OPC, and filter-based estimates
of refractive indices for ammonium sulfate, ammonium nitrate, organic carbon, elemental carbon
and residual aerosol mass concentrations as independent variables. The quality of their
comparisons with nephelometer data suggested that this approach could be used to test models
that predict visual range from source emissions. Further research is needed to determine the
effectiveness of the integrating nephelometer as a predictor of fine particle mass concentrations.
4.2.9 Specialized Sampling
4.2.9.1 Personal Exposure Sampling
The application of aerosol measurement technologies to smaller and less obtrusive
samplers have resulted in devices used as fixed-location indoor aerosol samplers and personal
exposure monitors (PEMs) worn on the body to estimate exposure. The reduction in physical
size of personal aerosol sampling systems to reduce participant burden sometimes results in
"Minneapolis, MN.
12Boulder, CO.
"Redlands, CA.
4-65
-------
poorer aerosol collection performance as compared to the outdoor counterparts. Wiener and
Rodes (1993) noted that personal sampling systems generally have poorer precisions than
outdoor aerosol samplers, due to the smaller sampler collections (from lower flowrates) and
poorer flow controllers. Ozkaynak et al. (1993) reported that the precisions of collocated PEMs
in the PTEAM study operating at 4.01 pm for a 12-h period were 3 to 4% (RSD). Wallace et al.
(1994) reported biases for the Particle Total Exposure Assessment Methodology study averaging
a factor of two between personal exposure measurements and fixed location PM10
concentrations. He was unable to completely account for the biases, but attributed portions to
proximity to indoor sources, a difference in inlet cutpoints (11.7 //m versus 10.0 //m) and the
collection of aerosols from the "personal cloud" caused by body dander. Rodes et al. (1991)
showed that the ratio of personal to indoor aerosol measurements for the EPA PTEAM study
appeared to be log-normally distributed with a median value of 1.98 and an unexpectedly high
value of 3.7 at the 90th ("most exposed") percentile. Ingham and Yan (1994) suggested that the
performance of a personal aerosol sampling inlet in an isolated mode (without mounting on a
representative humanoid bluff body) can result in substantial under-sampling for larger particles.
The relationship between measured aerosol exposure at some external location on the body and
actual uptake through oral and nasal entry is very complex.
Buckley et al. (1991) described the collection efficiency of an MSP, Inc.14 personal aerosol
sampler at 4.0 1pm as shown in Figure 4-20. They evaluated this sampler in a field comparison
study with collocated PM10 high volume and dichotomous samplers. The precision for the
personal sampler was found to be very good (CV = ±3.2%) with strong correlations (r2 = 0.970)
with the dichotomous samplers. Lioy et al. (1988) described a similar comparison for a 10 1pm
Air Diagnostics and Engineering, Inc.15 indoor air sampler, with a PM10 inlet characterized by
Marple et al. (1987). Correlations against the PM10 dichotomous sampler were also described
as very strong (r2 > 0.970), but noted a substantial bias caused by the loss of fragments from
indoor air sampler's glass fiber filters. They recommended that exposure studies using samplers
that collect small total volumes should utilize filters with greater integrity, such as Teflon.
Colome et al. (1992) describe an
"Minneapolis, MN.
15Naples, ME.
4-66
-------
100
80
.2 60
o
m
>^
UJ
40
0
O
o
20
1
10
Aerodynamic Particle Diameter (|jm)
Figure 4-20. Collection efficiency of the MSP personal aerosol sampler inlet.
Source: Buckley et al. (1991).
indoor/outdoor sampling study using an impactor characterized by Marple et al. (1987) with a
PM10 cutpoint that had duplicate impactors with the same cutpoint in series. This sequential
arrangement, in combination with a coating of 100 //I of light oil, was used to minimize particle
bounce at 4.0 1pm for 24 h period.
Personal aerosol sampler systems have typically been characterized as burdensome
(excessive weight, size, noise). The success of passive detector badges for gaseous pollutants
has recently prompted research into passive aerosol samplers. Brown et al. (1994) described a
prototype aerosol sampler utilizing electrostatic charge to move the particles to a collection
substrate. They noted that preliminary results are encouraging, but the effective sampling rate
and size-selectivity of the sampler was dependent on the electrical mobility of the aerosol. This
posed calibration problems for real aerosols with a distribution of electrical mobility's.
4-67
-------
Hollander (1992) described a passive pulsed-corona sampler that has similar collection
characteristics as a PM10 inlet, with only modest wind speed dependence.
The performance characterization of PEMs has been considered for occupational settings
by Kenny and Liden (1989), who reviewed the ACGIH, National Institute for Occupational
Safety and Health (NIOSH), and EPA PM10 aerosol sampler performance programs. They
proposed that an international consensus be reached on the basic principles underlying the
experimental protocols for testing personal samplers, as an essential prerequisite to the setting of
standards. An ISO working group has made progress in developing such a consensus (Kenny,
1992). As EPA becomes more focused on exposure assessment and personal exposure sampling,
it will become even more important for the agency to consider establishing performance
specifications for personal aerosol samplers.
Models have become powerful tools in understanding aerosol behavior in the vicinity of
personal exposure samplers. This is demonstrated by particle trajectory models that can predict
the influences of the geometries and flow field on aerosol capture and losses (e.g., Okazaki and
Willeke, 1987, Ingham and Yan, 1994, and Tsai and Vincent, 1993). These models have not
only permitted more rapid design changes to accommodate new cutpoints and flowrates, but
have added insights as to the influence of air flow obstructions on sampling efficiencies.
Vincent and Mark (1982) suggested that there is a critical particle trajectory that determines
whether a particle is sampled or rejected by an inlet worn on the body. An extension of this
model applicable to personal exposure sampling by Ingham and Yan (1994) suggested that
testing the performance of a personal aerosol sampling inlet in an isolated mode (without
mounting the inlet on a representative bluff body) can result in under-sampling for larger
particles by a factor of two. Validation of this model may explain a portion of the bias reported
by Wallace et al. (1994) between personal and indoor sampler measurements.
4.2.9.2 Receptor Model Sampling
Receptor modeling has become an established tool to relate ambient concentrations of
pollutants to major source categories, by apportioning the components in collected ambient
aerosol samples using complimentary source "signatures". Various approaches developed for
constructing source/receptor relationships were described by Henry et al. (1984), who also
provided a review of modeling fundamentals. They listed the advantages and disadvantages of
4-68
-------
multivariate models and discussed multi-collinearity problems associated with the presence of
two or more sources with nearly identical signatures. Javitz et al. (1988) described the basic
Chemical Mass Balance (CMB) approach and showed the influence of the variance in
identifying a component in the source signature sample on the projected apportionment. Dzubay
et al. (1984) described aerosol source and receptor collection schemes that permitted the
separation of ambient samples into fine and coarse fractions for mass, elemental and volatile
carbon, and metals analyses. Stevens and Pace (1984) suggested the addition of Scanning
Electron Microscopy to permit additional categorization using x-ray diffraction analysis. The
most widely used aerosol receptor model is the EPA CMB 7.0 model described by Watson et al.
(1990). This paper describes the structure of the model and computer code and the data
requirements to evaluate the validity of the estimates. Numerous papers have been published
describing the applications of receptor models to the apportionment of the sources of aerosols,
with the receptor modeling conference summary by Watson et al. (1989b) descriptive of the
state-of-the-art.
Stevens et al. (1993) described (see Figure 4-21) a modified dichotomous sampler with a
PM10 inlet, two Fine channels operating at 15 1pm and one coarse channel operating at 2.0 1pm,
designated as the Versatile Air Pollution Sampler (VAPS). The additional fine fraction channel
permitted sampling on a 47 mm Teflon filter for elemental analysis and a 47-mm quartz filter for
carbon speciation (elemental and volatile). A Nuclepore filter was used on the Coarse channel
for Scanning Electron Microscopy (SEM) evaluation and energy dispersive x-ray diffraction
analysis for selected particles.
4.2.9.3 Particle Acidity
An emphasis was placed on sampling sulfuric acidic aerosols in the 1982 Criteria
Document. This was followed by a number of research efforts (e.g., Ferm, 1986; Koutrakis
et al., 1988; Pierson et. al., 1989) to identify and study the in situ rate reactions, develop
sampling strategies to representatively remove the acid particle from the air, identify the
co-existing reactive species (e.g., ammonia, nitric acid, aerosol sulfates and nitrates), and protect
the collected aerosol prior to analysis. A "Standard" and an "Enhanced" method
were subsequently described (U.S. Environmental Protection Agency, 1992) for the
4-69
-------
32 l/min
J^.
I
o
25-um Cut Annular Denuder Teflon Filter
Receiver Jet Collects Mass, H+,
(VAPS Body) SO2,HNO3,HCI Elemental Composition
Accelerator
Jet
VAPS Impactor
Impactor Press
#47 FP
Adapter
PUF Adapter
with Quick
Disconnect to
Vacuum Pump
PUF Trap
80 mm * 32 mm
Figure 4-21. Modified dichotomous sampler (VAPS).
Source: Stevens et al. (1993).
-------
determination of aerosol acidity (titratable H+) using annular denuder technology. The
"Standard" method did not account for potential interferences from nitric acid, ammonium
nitrate aerosol, or other ammonium salts. The "Enhanced" method added an additional denuder
prior to filtration, with nylon and treated glass fiber backup filters to account for these species.
These sampling technologies utilized either an inlet impactor or a cyclone with 2.5 //m cutpoints
to sample the fine fraction. This technology has recently been extended to other reactive
aerosol systems, including semi-volatile organics (e.g., Vossler et al., 1988). Bennett et al.
(1994) describe a PM2 5 cyclone-based, filter pack sampling system designed for fine particle
network sampling and acidity measurements, as part of the Acid MODES program. The sampler
operated at 8.8 1pm, and was designed to selectively remove ammonia, speciate gas and particle
phase sulfur compounds, as well as collect gas phase nitric acid. An intercomparison of 18 nitric
acid measurement methods was reported by Hering et al. (1988), who noted that measurements
differed by as much as a factor of four and biases increased as nitric acid loadings increased. In
general the filter pack systems reported the highest acidity measurements, while the denuder-
difference techniques reported significantly lower measurements. Benner et al. (1991) in a
comparison of the SCENES filter pack sampler with a denuder-based sampler found excellent
agreement between sampler types for both nitric acid and total nitrates. They attributed the close
agreement to limited positive artifact formation, since the test field site had high nitric acid gas
to particulate nitrate ratios. John et al. (1988) noted that internal aluminum sampler surfaces
denude nitric acid, and describe the design of an aluminum denuder for the inlet of a
commercially available dichotomous sampler to quantitatively remove nitric acid for extended
periods.
Brauer et al. (1989) describe the design of a miniature personal sampler to collect acid
aerosols and gases. A significant finding was the lower than expected personal acidity levels,
attributed to the "personal cloud" production of ammonia by the body. Personal exposure levels
of acid aerosols were reported to be lower than indoor measurements.
4.2.10 Measurement Method Comparisons
4.2.10.1 Nitrate
Methods for measuring particle nitrate and gaseous nitric acid were compared in the field
as part of the 1985 Nitrogen Species Methods Comparison Study conducted over an 8-day
4-71
-------
period in the summer of 1985 in Claremont, CA (Hering et al., 1988). Particle nitrate methods
included sampling with filter packs (teflon and nylon filters operated in series), sampling with
nylon or impregnated filters operated downstream of a denuder to remove vapor nitric acid
(Possanzini et al., 1983; Shaw et al., 1982; Appel et al., 1981), and sampling with an impactor
(Wall et al., 1988). Results from that study showed that the precision for identical samplers was
about 4% (Anlauf et al., 1988; Solomon et al., 1988). Denuded nylon filter methods were used
in 6 different samplers operated by 4 different groups (Appel et al., 1988; John et al., 1988;
Pierson et al., 1988; Solomon et al., 1988). Data from these 6 methods show no systematic bias
among samplers. The average measurement precision (coefficient of variation) was 11%.
Impactor results were also in agreement with that from the denuded nylon filters (Wall et al.,
1988). In contrast, fine particle nitrate values from teflon filter of the filter packs were 43 to
59% lower than those measured by denuded nylon filters, with higher discrepancies for longer
sampling times (Soloman et al., 1988). The lower results on filter pack sampling are due to the
volatilization of nitrate particles from the filter. The vaporized nitrate is measured as nitric acid
on the backup filter (Hering et al., 1988; Solomon et al., 1988). To summarize, sampling with
denuded nylon filters or with impactors gave equivalent values for fine particle nitrate, whereas
teflon filter sampling was biased low due to the volatilization losses.
The results of the 1985 Nitrogen Species Methods Comparison Study were confirmed by
data collection as part of the 1987 Southern California Air Quality Study (Chow et al., 1994). In
this study, sampling times were 4 to 7 h. Samples were retrieved immediately, within
30 minutes of the end of sampling. Fine particle samples were collected by teflon filters, by
denuded nylon filters and by impactors. Results, stratified by time of day and season, are
illustrated in Figures 4-22 and 4-23 for central Los Angeles, CA and Claremont, CA,
respectively. Losses from the teflon filters are greatest in the summer, especially for daytime
samples (10 a.m. to 2 p.m., and 2 p.m. to 6 p.m.). Over 11 summer sampling days at 8 basin
locations for Claremont, CA, an average of 79% or 9.9|ig/m3 of the fine particle nitrate was
volatilized from the teflon filters for summer daytime sampling. For nighttime and morning
samples, 40% was lost. The percentage losses are smaller for winter samples, but the absolute
magnitude remains high at 8.9 |ig/m3 for daytime samples. Impactor data are in much closer
agreement with those from the denuded nylon filter than the teflon filter.
4-72
-------
o
51
0)
I-
IO
40 T
O)
3 SOt
0)
75
£ 20"
10"
Central Los Angeles: Summer
80 T
• Summer: Night
O Summer: Morninc
O Summer: Day
• 1:1 Line
10 20 30
PM2.5 Denuded Nylon Filter Nitrate (|jg/m )
Central Los Angeles: Winter
40
Winter: Night
• Winter: Morniru
O Winter: Day
•1:1 Line
10 20 30 40 50 60 70
PM2.5 Denuded Nylon Filter Nitrate (|jg/m )
Figure 4-22. Comparison of PM2 5 nitrate mass measurements from Teflon® filter versus
denuded nylon filter sample collection for Los Angeles, CA.
Source: Chow, et al. (1994).
4-73
-------
Filter Comparisons for Claremont: PM2.5 Nitrate
9 Summer: Night
• Summer: Morninc
O Summer: Day
•1:1 Line
40 T
• 30
re
i
o
(0
Q.
20"
£ 10--
c
i_
a
m
10 20 30
PM2.5 Denuded Nylon Filter Nitrate (ug/m )
Impactor Comparison for Claremont: PM2.5 Nitrate
• Summer: Night
O Summer: Morninc
O Summer: Day
• 1:1 Line
10 20 30 40
PM2.5 Denuded Nylon Filter Nitrate (ug/m )
Figure 4-23. Comparison of PM2 5 nitrate mass measurements from Teflon® filter versus
denuded nylon filter sample collection for Claremont, CA.
Source: Chow, et al. (1994).
4-74
-------
4.2.10.2 Carbonaceous Particulate Matter
Methods for measuring carbonaceous aerosol, classified as either "organic" or "black"
carbon, were compared in a similar study conducted in the summer of 1986 in Glendora, CA
(Hering et al., 1990). In that study, analytical methods were compared, as were differences in
simultaneous ambient sampling of PM25 aerosol with quartz filters, adsorption-corrected quartz
filters and two types of impactors. The results showed generally good agreement among
analytical methods for total carbon, with 5 of the 6 laboratories reporting values within 9% of
each other. In contrast, ambient sampling results showed variations among methods. Quartz
filter results, whether or not corrected for carbon vapor adsorption were within 40% of each
other. Concentrations from impactors, exclusive of after-filter, were lower than the mean from
the filter samplers by as much as 50%. Addition of the after-filter carbon brought impactor
values to within 10% of the mean, but the lack of "black" carbon on these after-filters leads to
the conclusion that vapor adsorption led to a positive bias for quartz filter sampling on these
days. Similar results were found for the 1987 Southern California Air Quality Study, for which
impactor measurements of carbon were systematically lower than filter measurements (Chow, et
al., 1994).
4.3 ANALYSIS OF PARTICULATE MATTER
The interest in the composition of aerosol particles lies in the areas of: (1) explaining and
inventorying the observed mass, (2) establishing the effect of aerosols on health and welfare, and
(3) attributing ambient aerosols to pollution sources. While any compositional measurement
will address one or more of these goals, certain methods excel for specific tasks. In general, no
single method can measure all chemical species, and comprehensive aerosol characterization
programs use a combination of methods to address complex needs. This allows each method to
be optimized for its objective, rather than be compromised to achieve goals unsuitable to the
technique. Such programs also greatly aid quality assurance objectives, since confidence may be
placed in the accuracy of a result when it is obtained by two or more methods on different
substrates and independent samplers.
In the sections that follow, some of the more commonly used methods that address the
goals stated above are described. The sections are designed to be illustrative rather than
4-75
-------
exhaustive, since new methods are constantly appearing as old methods are being improved.
These chemical analysis methods for the following section are divided into four categories:
(1) mass, (2) elements, (3) water-soluble ions, and (4) organics. Material balance comparing the
sum of the chemical species to the PM mass concentrations show that elements, water soluble
ions, and organic and elemental carbon typically explain 65 to 85% of the measured mass and
are adequate to characterized the chemical composition of measured mass for filter samples
collected in most urban and non-urban areas. Some of these chemical analysis methods are non-
destructive, and these are preferred because they preserve the filter for other uses. Methods
which require destruction of the filter are best performed on a section of the filter to save a
portion of the filter of other analyses or as a quality control check on the same analysis method.
Table 4-2 identifies the elements and chemical compounds commonly found in air using these
methods with typical detection limits.
Less common analytical methods, which are applied to a small number of specially-taken
samples, include isotopic abundances (Jackson, 1981; Currie, 1982; Hirose and Sugimura,
1984); mineral compounds (Davis, 1978, 1980; Schipper et al., 1993); and functional groups
(Mylonas et al., 1991; Palen et al., 1992; 1993; Allen et al., 1994). Recent advances in infrared
optics and detectors have resulted in the quantitative determination of the major functional
groups (e.g., sulfate, nitrate, aliphatic carbons, carbonyl carbons, organonitrates, and alcohols)
in the atmospheric aerosol (Allen et al., 1994). The advantages of functional analysis in source
apportionment are that the number of functional groups is much less than the number of organic
compounds to be classified. The cited references provide information on sampling and analysis
methods for these highly-specialized methods.
The following section focuses on:
Physical analysis of elements and single particle size, shape, and composition,
Wet chemical analysis of anions and cations, and
Organic analysis of organic compounds and elemental/organic carbon.
4-76
-------
TABLE 4-2. INSTRUMENTAL DETECTION LIMITS FOR
PARTICLES ON FILTERS
Minimum Detection Limit in ng/m3a
Be
Na
Mg
Al
Si
P
S
Cl
K
Ca
Sc
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Ga
As
Se
Br
Rb
Sr
Y
Zr
Mo
Pd
Ag
Cd
In
Sn
ICP/
AESb'd
0.06
NA
0.02
20
3
50
10
NA
NA
0.04
0.06
0.3
0.7
2
0.1
0.5
1
2
0.3
1
42
50
25
NA
NA
0.03
0.1
0.6
5
42
1
0.4
63
21
AA
Flameb'd
2d
0.2d
0.3
30
85
100,000
NA
NA
2d
ld
50
95
52
2
1
4
6d
5
4
1
52
100
100
NA
NA
4
300
1000
31
10
4
1
31
31
AA
Furnaceb
0.05
<0.05
0.004
0.01
0.1
40
NA
NA
0.02
0.05
NA
NA
0.2
0.01
0.01
0.02
0.02
0.1
0.02
0.001
NA
0.2
0.5
NA
NA
0.2
NA
NA
0.02
NA
0.005
0.003
NA
0.2
INAAb'f
NAh
2
300
24
NA
NA
6,000
5
24
94
0.001
65
0.6
0.2
0.12
4
0.02
NA
30
o
J
0.5
0.2
0.06
0.4
6
18
NA
NA
NA
NA
0.12
4
0.006
NA
PIXE8
NA
60
20
12
9
8
8
8
5
4
NA
3
o
5
2
2
2
NA
1
1
1
1
1
1
1
2
2
NA
3
5
NA
NA
NA
NA
NA
XRFC
NA
NA
NA
5
3
3
2
5
3
2
NA
2
1
1
0.8
0.7
0.4
0.4
0.5
0.5
0.9
0.8
0.6
0.5
0.5
0.5
0.6
0.8
1
5
6
6
6
8
ICb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
ACb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
TORb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
TABLE 4-2 (cont'd). INSTRUMENTAL DETECTION LIMITS FOR
4-77
-------
PARTICLES ON FILTERS
Minimum Detection Limit in ng/m3a
Species
Sb
I
Cs
Ba
La
Au
Hg
Tl
Pb
Ce
Sm
Eu
Hf
Ta
W
Th
U
Cl-
N03-
so;
NH;
oc
EC
ICP/
AESb'd
31
NA
NA
0.05
10
2.1
26
42
10
52
52
0.08
16
26
31
63
21
NA
NA
NA
NA
NA
NA
AA Flameb'd
31
NA
NA
8d
2,000
21
500
21
10
NA
2,000
21
2,000
2,000
1,000
NA
25,000
NA
NA
NA
NA
NA
NA
AA
Fumaceb
0.2
NA
NA
0.04
NA
0.1
21
0.1
0.05
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
INAAb'f
0.06
1
0.03
6
0.05
NA
NA
NA
NA
0.06
0.01
0.006
0.01
0.02
0.2
0.01
NA
NA
NA
NA
NA
NA
NA
PIXE8
NA
NA
NA
NA
NA
NA
NA
NA
3
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
XRFC
9
NA
NA
25
30
2
1
1
1
NA
NA
NA
NA
NA
NA
NA
1
NA
NA
NA
NA
NA
NA
ICb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
50
50
50
NA
NA
NA
ACb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
50
NA
NA
TORb
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
100
100
^Minimum detection limit is three times the standard deviation of the blank for a filter of 1 mg/cm2 areal
density.
ICP/AES = Inductively Coupled Plasma with Atomic Emission Spectroscopy.
AA = Atomic Absorption Spectrophotometry.
PIXE = Proton Induced X-ray Emissions Spectrometry.
XRF = Non-Dispersive X-ray Fluorescence Spectrometry.
INAA = Instrumental Neutron Activation Analysis.
1C = Ion Chromatography.
AC = Automated Colorimetry.
TOR = Thermal Optical Reflectance.
bConcentration is based on the extraction of 1/2 of a 47 mm quartz-fiber filter in 15 ml of deionized-distilled
water, with a nominal flow rate of 20 L/min for 24-h samples.
°Concentration is based on 13.8 cm2 deposit area for a 47 mm ringed teflon-membrane filter, with a nominal
flow rate of 20 L/min for 24-h samples with 100 sec radiation time.
dHarman(1989).
"Fernandez de la Mora (1989).
fOlmez (1989).
8Eldredetal. (1993).
hNot Available.
4-78
-------
4.3.1 Mass Measurement Methods
Particulate mass concentration is the most commonly made measurement on aerosol
samples. It is used to determine compliance with PM10 standards and to select certain samples
for more detailed, and more expensive, chemical analyses. As noted in Section 2, the beta
attenuation and inertial microbalance methods have been incorporated into in situ measurement
systems which acquire real-time mass measurements. Gravimetric analysis is used almost
exclusively to obtain mass measurements of filters in a laboratory environment. The U.S.
Environmental Protection Agency (1976) has published detailed procedures for mass analyses
associated with 20.32 cm x 25.40 cm fiber filters, but the guidance for other types of filters used
for chemical analyses is less well documented.
Gravimetry measures the net mass on a filter by weighing the filter before and after
sampling with a balance in a temperature- and relative humidity-controlled environment. PM10
reference methods require that filters be equilibrated for 24 h at a constant (within ±5%) relative
humidity between 20 and 40% and at a constant (within ±3 °C) temperature between 15 and 30
°C. These are intended to minimize the liquid water associated with soluble compounds and to
minimize the loss of volatile species. Nominal values of 30% RH and 15 to 20 °C best conserve
the particle deposits during sample weighing.
Balances used to weigh 20.32 cm x 25.40 cm filters from high volume PM10 samples must
have a sensitivity of at least 100 //g. Balances used for medium volume PM10 samples should
have a sensitivity of at least 10 //g, and those used for low-volume PM10 samples should have a
sensitivity of at least 1 //g. Modifications to the balance chamber are sometimes needed to
accommodate filters of different sizes. All filters, even those from high-volume PM10 samplers,
should be handled with gloved hands when subsequent chemical analyses are a possibility.
Balance calibrations should be established before and after each weighing session using
Class M and Class S standards, and they should be verified with a standard mass every 10 filters.
Approximately one out often filters should be re-weighed by a different person at a later time.
These re-weights should be used to calculate the precision of the measurement as outlined by
Watson etal. (1989a).
Feeney et al. (1984) examined the gravimetric measurement of lightly loaded membrane
filters and obtained excellent precision and accuracy. The sensitivity of the electrobalance is
4-79
-------
about ±0.001 mg, though tolerances on re-weights of Teflon-membrane filters are typically
±0.010 mg. The main interference in gravimetric analysis of filters results from electrostatic
effects. Engelbrecht et al. (1980) found that residual charge on a filter could produce an
electrostatic interaction between the filter on the pan and the metal casing of the electrobalance.
This charge can be removed by exposing the filter to a radioactive polonium source before and
during sample weighing.
Beta attenuation methods have been applied in the laboratory as well as in the field, and
the results are comparable to those of gravimetric measurements. The precision of beta-gauge
measurements has been shown to be ±5 //g/m3 or better for counting intervals of one minute per
sample, which translates into ±32 //g/filter for 37 mm diameter substrates. This is substantially
higher than the ±6 //g/filter precision determined by gravimetric analysis using an electrobalance
(Feeney et al., 1984). Jaklevic et al. (1981) found equivalent accuracy and precision for both
techniques as they were used in that study. Courtney et al. (1982) found beta attenuation and
gravimetric mass measurements to differ by less than ±5%. Patashnick and Rupprecht (1991)
examine results from TEOM samplers operated alongside filter-based PM10 samplers, and Shimp
(1988) reports comparisons with beta attenuation field monitors; these comparisons all show
good agreement for mass measurements.
4.3.2 Physical Analysis
The most common interest in elemental composition derives from concerns about health
effects and the utility of these elements to trace the sources of suspended particles. Instrumental
neutron activation analysis (INAA), photon-induced x-ray fluorescence (XRF), particle-induced
x-ray emission (PIXE), atomic absorption spectrophotometry (AAS), inductively-coupled
plasma with atomic emission spectroscopy (ICP/AES), and scanning electron microscopy with
x-ray fluorescence (SEM/XRF) have all been applied to elemental measurements of aerosol
samples. AAS and ICP/AES are also appropriate for ion measurements when the particles are
extracted in deionized-distilled water (DDW). Since air filters contain very small particle
deposits (20 to 100 //g/cm2), preference is given to methods that can accommodate small sample
sizes. XRF and PIXE leave the sample intact after analysis so that it can be submitted to
additional examinations by other methods. Excellent agreement was found for the
4-80
-------
intercomparison of elements acquired form the XRF and PIXE analyses (Cahill, 1980). The
analytical measurement specifications of air filter samples for the different elemental analysis is
shown in Table 4-2.
4.3.2.1 X-Ray Fluorescence of Trace Elements
In x-ray fluorescence (XRF) (Dzubay and Stevens, 1975; Hammerle and Pierson, 1975;
Jaklevic et al., 1977; Torok and Van Grieken, 1994), the filter deposit is irradiated by high
energy x-rays that eject inner shell electrons from the atoms of each element in the sample.
When a higher energy electron drops into the vacant lower energy orbital, a fluorescent x-ray
photon is released. The energy of this photon is unique to each element, and the number of
photons is proportional to the concentration of the element. Concentrations are quantified by
comparing photon counts for a sample with those obtained from thin-film standards of known
concentration.
XRF methods can be broadly divided into two categories: wavelength dispersive x-ray
fluorescence (WDXRF), which utilizes crystal diffraction for observation of fluorescent x-rays,
and energy dispersive x-ray fluorescence (EDXRF), which uses a silicon semiconductor
detector. The WDXRF method is characterized by high spectral resolution, which minimizes
peak overlaps. It requires high power excitation to overcome low sensitivity, resulting in
excessive sample heating and potential degradation. Conversely, EDXRF features high
sensitivity but less spectral resolution, requiring complex spectral deconvolution procedures.
XRF methods can be further categorized as direct/filtered excitation, where the x-ray beam
from the tube is optionally filtered and then focused directly on the sample, or secondary target
excitation, where the beam is focused on a target of material selected to produce x-rays of the
desired energy. The secondary fluorescent radiation is then used to excite the samples. The
direct/filtered approach has the advantage of delivering higher incident radiation flux to the
sample for a given x-ray tube power, since about 99% of the incident energy is lost in a
secondary fluorescence. However, the secondary fluorescence approach, produces a more nearly
monochromatic excitation that reduces unwanted scatter from the filter, thereby yielding better
detection limits.
XRF is usually performed on Teflon-membrane filters for a variety of trace elements. A
typical XRF system is schematically illustrated in Figure 4-24. The x-ray output stability should
4-81
-------
be within ±0.25% for any 8-h period within a 24-h duration. Typically, analyses are controlled,
spectra are acquired, and elemental concentrations are calculated by software on a computer that
is interfaced to the analyzer. Separate XRF analyses are conducted on each sample to optimize
detection limits for the specified elements. A comparison of the minimum detectable limits of
Teflon-membrane and quartz-fiber filters is listed in Table 4-3. Figure 4-25 shows an example
of an XRF spectrum.
Three types of XRF standards are used for calibration, performance testing, and auditing:
(1) vacuum-deposited thin-film elements and compounds (Micromatter); (2) polymer films
(Dzubay et al., 1981); and (3) National Institute of Science and Technology (NIST, formerly
NBS) thin-glass films. The thin film standards cover the largest number of elements and are
used to establish calibration curves, while the polymer film standards are used to verify the
accuracy of the thin film standards. The NIST standards are used to validate the accuracy of the
calibration curves. NIST produces the definitive standard reference materials, but these are only
available for the species of aluminum, silicon, calcium, iron, cobalt, copper, manganese, and
uranium (SRM 1832), and silicon, potassium, titanium, iron, zinc, and lead (SRM 1833). One or
more separate Micromatter thin-film standards are used to calibrate the system for each element.
Sensitivity factors (number of x-ray counts per //g/cm2 of the element) are determined for
each excitation condition. These factors are then adjusted for absorption of the incident and
emitted radiation in the thin film. These sensitivity factors are plotted as a function of atomic
number and a smooth curve is fitted to the experimental values. The calibration sensitivities are
then read from these curves for the atomic numbers of each element in each excitation condition.
NIST standards are analyzed on a periodic basis to verify the sensitivity factors. A multi-layer
thin film standard prepared by Micromatter is analyzed with each set of samples to check the
stability of the instrument response. When deviations from specified values are greater than
±5%, the system should be re-calibrated.
The sensitivity factors are multiplied by the net peak intensities yielded by ambient
samples to obtain the //g/cm2 deposit for each element. The net peak intensity is obtained
4-82
-------
Sampl
X-ray exci
.Characteristic Silicon detector
\-rays / FET
"• /preamp
•=!>
Pulse
processoi
Secondary
target
^ Anode
Electron bean
Signal
processing
Analog-to|
digital
converter!
X-ray tube
Data output
I
I
I
^B
Mini-
compute
Data
handling
Figure 4-24. Schematic of a typical X-ray fluorescence system.
4-83
-------
TABLE 4-3. MINIMUM DETECTABLE LIMITS3 FOR X-RAY FLUORESCENCE
ANALYSIS OF AIR FILTERS
Element
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
Ga
As
Se
Br
Rb
Sr
Y
Zr
Mo
Pd
Ag
Cd
In
Condition
Numberd
5
5
5
5
4
4
4
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
4
1
1
1
1
Quartz-Fiber
Filterb
Protocol QA-
A ng/cm2 e
NAf
NA
NA
40s
30
40
100
50
20
8
7
15
5
4
4
6
8
9
5
5
5
8
8
10
20
20
20
25
30
Teflon Membrane Filter0
Protocol A
ng/cm2 d
10
6.3
5.6
5.0
10
6.1
4.5
2.9
2.5
1.9
1.6
1.5
0.88
0.89
1.1
1.1
1.9
1.6
1.2
1.0
1.0
1.1
1.3
1.7
2.7
11
12
12
13
Protocol B
ng/cm2
7.2
4.4
4.0
3.5
7.4
4.3
3.2
2.1
1.7
1.4
1.1
1.1
0.62
0.63
0.76
0.76
1.4
1.1
0.86
0.72
0.68
0.78
0.92
1.2
1.9
7.6
8.6
8.6
9.5
Protocol C
ng/cm2
3.6
2.2
2.0
1.8
3.7
2.2
1.6
1.0
0.87
0.67
0.56
0.54
0.31
0.31
0.38
0.38
0.68
0.56
0.43
0.36
0.34
0.39
0.46
0.59
0.95
3.8
4.3
4.3
4.8
Protocol D
ng/cm2
2.5
1.4
1.4
1.2
2.6
1.5
1.1
0.73
0.62
0.48
0.40
0.38
0.22
0.22
0.27
0.27
0.48
0.39
0.31
0.25
0.24
0.28
0.33
0.42
0.67
2.7
3.0
3.0
3.4
4-84
-------
TABLE 4-3 (cont'd). MINIMUM DETECTABLE LIMITS3 FOR X-RAY
FLUORESCENCE ANALYSIS OF AIR FILTERS
Element
Condition
Numberd
Quartz-Fiber
Filterb
Protocol QA-
A ng/cm2 e
Teflon Membrane Filter0
Protocol A
ng/cm2 d
Protocol B
ng/cm2
Protocol C
ng/cm2
Protocol D
ng/cm2
Sn
1
40
17
12
6.2
4.4
Sb
Ba
La
Au
Hg
Tl
Pb
U
1
1
1
2
2
2
2
2
50
170
190
NA
20
NA
14
NA
18
52
62
3.1
2.6
2.5
3.0
2.3
13
37
44
2.2
1.8
1.8
2.2
1.7
6.4
18
22
1.1
0.91
0.88
1.1
0.83
4.5
13
16
0.77
0.65
0.62
0.76
0.59
TVTDL defined as three times the standard deviation of the blank for a filter of 1 mg/cm2 areal density.
bAnalysis times are 100 sec. for Conditions 1 and 4, and 400 sec. for Conditions 2 and 3. Actual MDL's for
quartz filters vary from batch to batch due to elemental contamination variability.
"Standard protocol, developed at the Desert Research Institute, University and Community College System of
Nevada, Reno, NV, analysis times are 100 sec. for Conditions 1, 4 and 5, and 400 sec. for Conditions 2 and 3
for Protocol A; 200 sec. for Conditions 1, 4 and 5 and 800 sec. for Conditions 2 and 3 for Protocol B; 800
sec. for Conditions 1,4 and 5 and 3,200 sec. for Conditions 2 and 3 for Protocol C; and 1600 sec. for
Conditions 1, 4 and 5 and 6400 sec. for Conditions 2 and 3 for Protocol D.
dCondition 1 is direct mode excitation with a primary excitation filter of 0.15 mm thick Mo. Tube voltage is
50 KV and tube current is 0.6 mA. Condition 2 is direct mode excitation with a primary excitation filter of
0.13 mm thick Rh. Tube voltage is 35 KV and tube voltage is 2.0 mA. Condition 3 uses Ge secondary target
excitation with the secondary excitation filtered by a Whatman 41 filter. Tube voltage is 30 KV and tube
current is 3.3 mA. Condition 4 uses Ti secondary target excitation with the secondary excitation filtered by
3.8 //m thick mylar film. Tube voltage is 30 KV and tube current is 3.3 mA. Condition 5 uses direct mode
excitation with a primary excitation filter consisting of 3 layers of Whatman 41 filters. Tube voltage is 8 KV
and tube current os 0.6 mA. Multi-channel analyzer energy range is 0 to 40 KeV for condition 1, 0 - 20 KeV
for condition 2, and 0 to 10 KeV for conditions 3, 4, and 5.
"Typical exposed area is 406 cm2 for standard high-volume filters; 6.4 cm2 for 37 mm ringed Teflon-membrane
filters; and 13.8 cm2 for 47 mm ringed Teflon-membrane filters.
Information not available.
8For condition 4.
4-85
-------
26-Oct-1992 18:09:56
SJTT046
Vert= 2000 counts Disp= 1
Preset= 100 sees
Comp= 2 Elapsed= 400 sees
San Jose, 1/21/92, PM 10
18:01 - 06:00
Excitation Condition 3
0.320 Range= 10.230 keV 10.230 >•
Integral 0 = 243425
5
Figure 4-25. Example of an X-ray fluorescence spectrum.
Source: Chow and Watson (1994).
10
by: (1) subtracting background radiation; (2) subtracting spectral interferences; and
(3) adjusting for x-ray absorption.
XRF analysis of air particulate samples has had widest application to samples collected on
membrane-type filters such as Teflon- or polycarbonate-membrane filter substrates. These
membrane filters collect the deposit on their surfaces, which eliminates biases due to absorption
of x-rays by the filter material. These filters also have a low areal density which minimizes the
scatter of incident x-rays, and their inherent trace element content is very low. Quartz-fiber
filters used for high-volume aerosol sampling do not exhibit these features. As noted earlier,
blank elemental concentrations in quartz-fiber filters that have not undergone acceptance testing
can be several orders of magnitude higher than the concentrations in the particulate deposits.
4-86
-------
The concentrations vary substantially among the different types of quartz-fiber filters and even
within the same filter type and manufacturing lot. Blank impurity concentrations and their
variabilities decrease the precision of background subtraction from the XRF spectral data,
resulting in higher detection limits. Impurities observed in various types of glass- and quartz-
fiber filters include aluminum, silicon, sulfur, chlorine, potassium, calcium, iron, nickel, copper,
zinc, rubidium, strontium, molybdenum, barium, and lead. Concentrations for aluminum,
silicon, phosphorus, sulfur, and chlorine cannot be determined for quartz-fiber filters because of
the large silicon content of the filters.
Quartz-fiber filters also trap particles within the filter matrix, rather than on the surface.
This causes absorption of X rays within the filter fibers yielding lower concentrations than
would otherwise be measured. The magnitude of this absorption increases exponentially as the
atomic number of the measured element decreases and varies from sample to sample.
Absorption factors generally are "1.2" or less for iron and heavier elements, but can be from "2"
to "5" for sulfur.
Quartz-fiber filters are much thicker than membrane filters resulting in an increased
scattering of x-rays and a consequent increase in background and degradation of detection limits.
The increased x-ray scatter also overloads the x-ray detector which requires samples to be
analyzed at a lowered x-ray intensity. These effects alone can result in degradation of detection
limits by up to a factor of 10 with respect to Teflon-membrane substrates.
Larger particles collected during aerosol sampling have sufficient size to cause absorption
of x-rays within the particles. Attenuation factors for fine particles (PM2 5, particles with
aerodynamic diameters equal to or less than 2.5 //m) are generally negligible (Criss, 1976), even
for the lightest elements, but these attenuations can be significant for coarse fraction particles
(particles with aerodynamic diameters from 2.5 to 10 //m). Correction factors for XRF have
been derived using the theory of Dzubay and Nelson (1975) and should be applied to coarse
particle measurements.
4.3.2.2 Particle Induced X-Ray Emission of Trace Elements
Particle Induced X-Ray Emission (PIXE) is another form of elemental analysis based on
the characteristics of x-rays and the nature of x-ray detection (Cahill et al., 1987; 1989). PIXE
uses beams of energetic ions, consisting of protons at an energy level of 2 to 5 MeV, to create
4-87
-------
inner electron shell vacancies. As inner electron shell atomic vacancies are filled by outer
electrons, the emitted characteristics of x-rays can be detected by wavelength dispersion (which
is scattering from a crystal) or by energy dispersion (which involves direct conversion of x-rays).
The development of focusing energetic proton beams (proton microprobes) has expanded the
application of PIXE from environmental and biological sciences to geology and material
sciences. Figure 4-26 illustrates a typical PIXE setup in a thin target mode (Cahill, et al., 1989).
PIXE analysis is often used for impactor samples or small filter substrates, since proton beams
can be focused to a small area with no loss of sensitivity (Cahill and Wakabayashi, 1993).
Very thick filters or thick particle deposits on filter substrates scatter the excitation protons
and lower the signal-to-noise ratio for PIXE. X-ray analysis methods, such as PIXE and XRF,
require particle size diameter corrections (for low atomic number targets) associated with a
spherical particle of a given diameter (typically particles with aerodynamic diameters >2.5 //m)
and compositions typical in ambient aerosol studies. These analyses also require correction for
sample loadings that reflect the passage of x-rays through a uniform deposit layer. Procedures
for instrument calibration, spectrum process, and quality assurance are similar to those
documented in Section 4.3.1.2 for XRF.
PIXE analysis can provide information on one of the widest range of elements in a single
analysis, since x-ray results require two or three separate anodes. However, attempts to improve
sensitivity of PIXE analysis may result in damage to Teflon-membrane filters. Recent
developments (Malm et al., 1994) using PIXE analysis at moderate sensitivity plus single anode
XRF analysis at high sensitivity for transition/heavy metals have achieved the minimum
detectable limits of less than 0.01 ng/m3. With the addition of hydrogen analysis (a surrogate for
organic matter), almost all gravimetric mass concentrations can be explained (Cahill, et al.,
1987).
XRF and PIXE are the most commonly used elemental analysis methods owing to their
nondestructive multi-element capabilities, relatively low cost, high detection limits, and
preservation of the filter for additional analyses. XRF sometimes needs to be supplemented with
INAA when extremely low detection limits are needed, but the high cost of INAA precludes this
method from being applied to large numbers of samples. AAS is a good
-------
oo
VO
PIXE-2
(Fe-Mo)
PIXE-1
(Na-Mn)
Figure 4-26. Schematic of a PIXE/PESA analysis system.
-------
alternative for water-soluble species, especially for low atomic number. ICP/AES analysis is a
viable alternative, but it is less desirable because of the sample extraction elements such as
sodium and magnesium, but it requires large dilution factors to measure many different elements
expense and the destruction of the filter.
4.3.2.3 Instrumental Neutron Activation Analysis of Trace Elements
Instrumental neutron activation analysis (INAA) (Dams et al., 1970; Zoller and Gordon,
1970; Olmez, 1989; Ondov and Divita, 1993) basically involves irradiation of a thin membrane
filter sample in the core of a nuclear reactor for periods ranging from a few minutes to several
hours. Bombardment of the sample with neutrons induces a nuclear reaction of the stable
isotopes in the sample. The energies of the gamma rays emitted by the decay of this induced
radioactivity are used to identify them, and therefore, their parents. With the use of prepared
elemental standards, the amount of parent element in the sample can be determined since the
intensity of these gamma rays are proportional to their number.
The gamma-ray spectra of radioactive species are usually collected with a high resolution
germanium detector utilizing commercially available amplifiers and multi-channel analyzers.
Typical detector efficiencies range from 10 to 40% relative to a 3 x 3 in. sodium iodide detector.
Detector system resolution, measured as the full-width at half-maximum for Table 4-4, the 1,332
KeV gamma-ray peak of 60Co, should be less than 2.3 KeV in order to provide adequate
resolution between isotopes of neighboring energies.
In order to obtain a full suite of elemental analysis results (often over 40 elements),
multiple counting periods and irradiations are performed on the same sample (e.g., two
irradiations would produce elements separated into short- and long-lived decay products). An
example of the elements determined from multiple irradiations and counting periods and the
irradiation, cooling, and counting times used for ambient particulate samples collected on
Teflon-membrane filter material are summarized in Table 4-4 (Divita, 1993). These irradiations
were performed at the 20-MW NIST Research Reactor operated at 15-MW (neutron flux of 7.7
x 1013 and 2.7 x 1013 neutron/cm2 x s).
The power of INAA is that it is not generally subject to interferences like XRF or PIXE
due to a much better ratio of gamma ray peak widths to total spectral width, by a factor of about
20. INAA does not quantify some of the abundant species in ambient
4-90
-------
TABLE 4-4. INSTRUMENTAL NEUTRON ACTIVATION ANALYSIS COUNTING
SCHEME AND ELEMENTS MEASURED
Counting Period
Short-Lived 1
Short-Lived 2
Long-Lived 1
Irradiation Time Cooling
Time
10 min 5 min
20min
4-6 h 3-4 days
Counting Time
5 min
20 min
6-8 h
Elements Measured
Mg, Al, S, Ca, Ti, V, Cu
Na, Mg, Cl, K, Ca, Mn, Zn,
Ga, Br, Sr, In, I, Ba
Na, K, Ga, As, Br, Mo, Cd,
Sb, La, Nd, Sn, Yb, Lu, W,
Au, U
Long-Lived 2
30 days 12-24 h Sc, Cr, Fe, Co, Zn, Se, Sr,
Ag, Sb, Cs, Ba, Ce, Nd, Eu,
Gd, Tb, Lu, Hf, Ta, Th
particulate matter such as silicon, nickel, tin, cadmin, mercury, and lead. While INAA is
technically nondestructive, sample preparation involves folding the samples tightly and sealing it
in plastic, and the irradiation process makes the filter membrane brittle and radioactive. These
factors limit the use of the sample for subsequent analyses by other methods. The technique also
suffers from the fact that a nuclear reactor is usually used as a source of neutrons. However,
since the advent of high-resolution gamma-ray detectors, individual samples can be analyzed for
numerous elements simultaneously, most at remarkably trace levels without the need for
chemical separation. This greatly diminishes the danger of contamination due to excessive
sample handling and introduction of chemical reagents used for separation procedures.
4.3.2.4 Microscopy Analysis of Particle Size, Shape, and Composition
Morphological and chemical features of particles can be used to identify the sources and
transport mechanism of airborne particles. The chemical analysis of individual particles
allows the attribution of specific pollution sources more straightforward while the abundance of
a specific group is a representative of the source strength. Both light (optical) and scanning
electron microscopy have been applied in environmental studies to examine the
4-91
-------
single particles (e.g., Casuccio et al., 1983; Bruynseels et al., 1988; Van Borm and Adams, 1988;
Van Borm et al., 1989; Cornille et al., 1990; Hopke and Casuccio, 1991; Turpin et al., 1993a).
Light microscopy has been used for providing particle size information regarding the
morphology of microscopic features (Crutcher, 1982). The practical resolution of optical
microscopes is limited by the wavelengths associated with light of the visible spectrum. When
features of interest occur in micron and submicron size ranges, detailed resolution cannot be
obtained. The practical resolution of light microscopy is typically 1 to 2 //m (Meyer-Arendt,
1972).
The use of accelerated electrons in electron microscopy (a) allows for the formation of
magnified images and an increased depth of field and (b) provides the resolution of a few
angstroms (10"4 //m). Electron microscopy has now evolved to include: (1) the transmission
electron microscope (TEM); (2) the scanning electron microscope (SEM), and; (3) the scanning
transmission electron microscope (STEM) (Hearle et al. 1972; Lee et al., 1979; Lee and Fisher,
1980; Lee and Kelly, 1980; Lee et al., 1981; Johnson et al., 1981; Mclntyre and Johnson, 1982;
Casuccio et al., 1983; Wernisch, 1985, 1986; Kim et al., 1987; Kim and Hopke, 1988; Dzubay
andMamane, 1989; Schamber, 1993).
The SEM and STEM use accelerated electrons to strike the sample. As the electron beam
strikes the samples, various signals (e.g., secondary, backscattered, and Anger electrons,
characteristic x-rays, photons, and cathodoluminescence) are generated. These signals can be
collected to provide highly detailed information on a point-by-point basis. The secondary
electron signal yields a sample image with three-dimensional prospective, high depth of field,
and illuminated appearance. Back scattered electron images are used to separate phases
containing elements of different atomic number.
The information obtained from light and scanning microscopy analyses are usually
considered to be qualitative, due to the limited number of particles counted. To achieve a
quantitative analysis, a sufficient number of particles must be properly sized and identified by
morphology and/or chemistry to represent the entire sample. The selection of filter media,
optimal particle loadings, and sample handling methods are also of importance. In this manner,
the microscopic characteristics can be directly and reliably related to the bulk or macroscopic
properties of the sample (Casuccio et al., 1983).
4-92
-------
Microscopic analysis requires a high degree of skill and extensive quality assurance to
provide quantitative information. The techniques is complex and expensive when quantitative
analysis is required. The evolution of computer technology has allowed for quantitative analysis
of particle samples of an entire population of features. With advanced pattern recognition
methods, data from individual particle features can be sorted and summarized by size and
composition, permitting improved quantitative source apportionment (Bruynseels et al., 1988;
Hopke and Casuccio, 1991). Casuccio et al. (1983) summarized the pros and cons of automatic
scanning electron microscopy.
Recent development of the SEM/XRF allows analysis of elemental compositions and
morphological information on small quantities of material (Bruynseels et al., 1988). Coupled
with statistical data analysis, computer controlled scanning electron microscopy shows great
promise for identifying and quantifying complex pollution sources in the field of receptor
modeling source apportionment (e.g., Griffin and Goldberg, 1979; Janocko et al., 1982; Johnson
et al., 1982; Massart and Kaufman, 1983; Hopke, 1985; Derde et al., 1987, Saucy et al., 1987;
Mamane, 1988; Dzubay andMamane, 1989).
4.3.3 Wet Chemical Analysis
Aerosol ions refer to chemical compounds that are soluble in water. The water-soluble
portion of suspended particles associates itself with liquid water in the atmosphere when relative
humidity increases, thereby changing the light scattering properties of these particles. Different
emissions sources may also be distinguished by their soluble and non-soluble fractions. Gaseous
precursors can also be converted to their ionic counterparts when they interact with chemicals
impregnated on the filter material.
Several simple ions, such as soluble sodium, magnesium, potassium, and calcium are best
quantified by atomic absorption spectrometry (AAS) as described above. In practice, AAS has
been very useful for measuring water-soluble potassium and sodium, which are important in
apportioning sources of vegetative burning and sea salt, respectively. Polyatomic ions such as
sulfate, nitrate, ammonium, and phosphate must be quantified by other methods such as ion
chromatography (1C) and automated colorimetry (AC). Simple ions, such as chloride,
chromium III, and chromium IV, may also be measured by these methods along with the
polyatomic ions.
4-93
-------
All ion analysis methods require filters to be extracted in DDW and then filtered to remove
the insoluble residue. The extraction volume needs to be as small as possible, lest the solution
become too dilute to detect the desired constituents. Each square centimeter of filter should be
extracted in no more than 2 ml of solvent for typical sampler flow rates of 20 to 30 L/min and
sample durations of 24 h. This often results in no more than 20 ml of extract that can be
submitted to the different analytical methods, thereby giving preference to those methods which
require only a small sample volume. Sufficient sample deposit must be acquired to account for
the dilution volume required by each method.
When other analyses are to be performed on the same filter, the filter must first be
sectioned using a precision positioning jig attached to a paper cutter. For rectangular filters
(typically 20.32 cm by 25.40 cm), a 2.0 cm by 20.32 cm wide strip is cut from the center two-
thirds of the filter. Circular filters of 25-, 37-, and 47-mm diameters are usually cut in half for
these analyses, so the results need to be multiplied by two to obtain the deposit on the entire
filter. Filter materials that can be easily sectioned without damage to the filter or the deposit
must be chosen for these analyses.
4.3.3.1 Ion Chromatographic Analysis for Chloride, Nitrate, and Sulfate
Ion chromatography (1C) can be used for both anions (fluoride [F~], chloride [Cl~], nitrite
[NO^, bromide [Br"], nitrate [NO3, phosphate [PO^3], sulfate [SOJ]) and cations (soluble
potassium [K+], ammonium [NH^], soluble sodium [Na+]) with separate columns. Applied to
aerosol samples, the anions are most commonly analyzed by 1C with the cations being analyzed
by a combination of atomic absorption spectrophotometry (AAS) and automated colorimetry
(AC) (U.S. EPA, 1994). In 1C (Small et al., 1975; Mulik et al., 1976; Butler et al., 1978) the
sample extract passes through an ion-exchange column that separates the ions in time for
individual quantification, usually by a electroconductivity detector. Figure 4-27 shows a
schematic representation of the 1C system. Prior to detection, the column effluent enters a
suppressor column where the chemical composition of the eluent is altered, resulting in a lower
background conductivity. The ions are identified by their elution/retention times and are
quantified by the conductivity peak area or peak height. 1C is especially desirable for particle
samples because it provides results for several ions with a single analysis and it uses a small
portion of the filter extract with low detection limits.
4-94
-------
Delivery Moduk
Chromatography Moduh
Detector Moduli
Eluent
Reservoir
o
A.
Pump
Sample
Injector
Guard
Column
Separator
Column
Suppressor
Device
Conductivity
Cell
fwastej
Figure 4-27. Schematic representation of an ion chromatography system.
4-95
-------
Water-soluble chloride (Cl"), nitrate (NCQ, and sulfate (SO4) are the most commonly measured
anions in aerosol samples. Figure 4-28 shows an example of an 1C anion chromatogram. 1C
analyses can be automated by interfacing to an automatic sampler that can conduct unattended
analysis of as many as 400 samples (Tejada et al., 1978).
18,000-
15,500-
13,000-
10,500-
8,000-
5,500-
3,000-
500-
-2,000-
0.
Fluoride
I
Chloride
Nitrite
| Nitrate
I 1 \ f\ Ou Sulfate
I |\ |\ J\ Ph^fhate S\
^
)0 5.00 10.00
Minutes
Figure 4-28. Example of an ion chromatogram showing the separation of fluoride,
chloride, nitrite, nitrate, phosphate, and sulfate ions.
Several independent quality assurance (QA) standards should be used to check the
calibration curve. The standards that are traceable to NIST simulated rainwater standards are:
Environmental Resource Associates (ERA, Arvada, CA) custom standards containing the anions
measured at a concentration of 100 //g/ml, ERA Waste Water Nutrient Standard, ERA Waste
Water Mineral Standard, and Alltech individual standards at 200 //g/ml. The QA standards are
diluted in DDW to concentrations that are within the range of the calibration curve.
Calibration curves are performed weekly. Chemical compounds are identified by matching
the retention time of each peak in the unknown sample with the retention times of
4-96
-------
peaks in the chromatograms of the standards. The QA standards are analyzed at the beginning
of each sample run to check calibrations. A DDW blank is analyzed after every 20 samples and
a calibrations standard is analyzed after every 10 samples. These quality control (QC) checks
verify the baseline and calibration respectively.
4.3.3.2 Automated Colorimetric Analysis for Ammonium, Nitrate, and Sulfate
Automated Colorimetry (AC) applies different colorimetric analyses to small sample
volumes with automatic sample throughput. The most common ions measured are ammonium,
chloride, nitrate, and sulfate (Butler et al., 1978; Fung et al., 1979). Since 1C provides multi-
species analysis for the anions, ammonium is most commonly measured by AC.
The AC system is illustrated schematically in Figure 4-29. The heart of the automated
colorimetric system is a peristaltic pump, which introduces air bubbles into the sample stream at
known intervals. These bubbles separate samples in the continuous stream. Each sample is
mixed with reagents and subjected to appropriate reaction periods before submission to a
colorimeter. The ion being measured usually reacts to form a colored liquid. The liquid
absorbance is related to the amount of the ion in the sample by Beer's Law. This absorbance is
measured by a photomultiplier tube through an interference filter specific to the species being
measured.
The standard AC technique can analyze -60 samples per hour per channel, with minimal
operator attention and relatively low maintenance and material costs. Several channels can be
set up to simultaneously analyze several ions. The methylthymol-blue (MTB) method is applied
to analyze sulfate. The reaction of sulfate with MTB-barium complex results in free ligand,
which is measured colorimetrically at 460 nm. Nitrate is reduced to nitrite that reacts with
sulfanilamide to form a diazo compound. This compound is then reacted to an azo dye for
colorimetric determination at 520 nm. Ammonium is measured with the indophenol method.
The sample is mixed sequentially with potassium sodium tartrate, sodium phenolate, sodium
hypochlorite, sodium hydroxide, and sodium nitroprusside. The reaction results in a blue-
colored solution with an absorbance measured at 630 nm. The system determines carry-over by
analysis of a low concentration standard
4-97
-------
Sampler
Heater
(optional)
Mixing Coils-
Mixing
Cell
Flow
Cell
Optical
Filter
Photomultiplier
Detector
Reagent Line #1
Reagent Line #2
Sample Line
Reagent Line #3
Reagent Line #4
Reagent Line #5
Reagent Line #6
Peristatic
Pump
Figure 4-29. Schematic of a typical automated colorimetric system.
following a high concentration. The percent carry-over is then automatically calculated and can
be applied to the samples analyzed during the run.
Intercomparison studies between AC and 1C have been conducted by Butler et al. (1978)
and Fung et al. (1979). Butler et al. (1978) found excellent agreement between sulfate and
nitrate measurements by AC and 1C. The accuracy of both methods is within the experimental
errors, with higher blank values observed for AC techniques. Comparable results were also
obtained between the two methods by Fung et al. (1979). The choice between the two methods
for sample analysis is dictated by sensitivity, scheduling, and cost constraints.
Two milliliters of extract in sample vials are placed in an autosampler that is controlled by
a computer. Five standard concentrations (e.g., (NH4)2SO4, Na2SO4, NaNO3) are prepared from
American Chemical Society reagent-grade chemicals following the same procedure as that for
1C standards. Each set of samples consists of two DDW blanks to establish a baseline, five
calibration standards and a blank, then sets often samples followed by analysis of one of the
standards and a replicate from a previous batch. The computer control allows additional analysis
4-98
-------
of any filter extract to be repeated without the necessity of loading the extract into more than one
vial.
4.3.3.3 Atomic Absorption Spectrophotometric (AAS) and Inductive Coupled Plasma
Atomic Emission Spectro (ICP/AES) Photometry Analyses for Trace Elements
In atomic absorption Spectrophotometric (AAS) analysis (Fernandez de la Mora, 1989), the
sample is first extracted in a strong solvent to dissolve the solid material; the filter or a portion
thereof is also dissolved during this process. A few milliliters of this extract are introduced into
a flame where the elements are vaporized. Most elements absorb light at certain wavelengths in
the visible spectrum, and a light beam with wavelengths specific to the elements being measured
is directed through the flame to be detected by a monochrometer. The light absorbed by the
flame containing the extract is compared with the absorption from known standards to quantify
the elemental concentrations. AAS requires an individual analysis for each element, and a large
filter or several filters are needed to obtain concentrations for a large number of the elements
specified in Table 4-3. AAS is a useful complement to other methods, such as XRF and PIXE,
for species such as beryllium, sodium, and magnesium that are not well-quantified by XRF and
PIXE. Airborne particles are chemically complex and do not dissolve easily into complete
solution, regardless of the strength of the solvent. There is always a possibility that insoluble
residues are left behind and soluble species may co-precipitate on them or on container walls.
In inductive coupled plasma atomic emission Spectrophotometric (ICP/AES), (Lynch et al.,
1980; Harman, 1989), the dissolved sample is introduced into an atmosphere of argon gas seeded
with free electrons induced by high voltage from a surrounding Tesla coil. The high
temperatures in the induced plasma raise valence electrons above their normally stable states.
When these electrons return to their stable states, a photon of light is emitted which is unique to
the element which was excited. This light is detected at specified wavelengths to identify the
elements in the sample. ICP/AES acquires a large number of elemental concentrations using
small sample volumes with acceptable detection limits for atmospheric samples. As with AAS,
this method requires complete extraction and destruction of the sample.
4-99
-------
4.3.4 Organic Analysis
4.3.4.1 Analysis of Organic Compounds
Organic compounds comprise a major portion of airborne particles in the atmosphere, thus
contributing to visibility degradation, and affecting the properties of clouds into which these
particles are scavenged. Specific groups of organic compounds (e.g., polycyclic aromatic
hydrocarbons, PAHs) have also been implicated in human health effects. However, due to the
very complex composition of the organic fraction of atmospheric aerosols, the detailed
composition and atmospheric distributions of organic aerosol constituents are still not well
understood.
Sampling techniques for atmospheric paniculate matter have been extensively investigated,
resulting in the development of collection methods suspended in a wide range of sizes. Particles
are most frequently collected on glass or quartz-fiber filters that have been specially treated to
achieve low "carbon blanks". Ambient organic particulate matter has also been collected on a
variety of particle sizing devices, such as low pressure impactors and Micro Oriface Uniforms
Deposit Impactors("MOUDI"). Very recently, diffusion denuder based samplers have been used
as well (Tang et al., 1994). However, the task of sampling organic compounds in airborne
particles is complicated by the fact that many of these compounds have equilibrium vapor
pressures (gaseous concentrations) that are considerably larger than their normal ambient
concentrations. This implies a temperature- and concentration-dependent distribution of such
organics between particulate and vapor phases. It also suggests that artifacts may occur due to
volatilization during the sampling process (Coutant et al., 1988). Such volatilization would
cause the under-estimation of the particle-phase concentrations of organics. Conversely, the
adsorption of gaseous substances on deposited particles or on the filter material itself, a process
driven by the lowered vapor pressure over the sorbed material, would lead to over-estimation of
the particle-phase fraction (Bidleman et al., 1986; Ligocki and Pankow, 1989; McDow and
Huntzicker, 1990). In addition, several studies have suggested that chemical degradation of
some organics may occur during the sampling procedure (Lindskog et al., 1985; Arey et al.,
1988; Parmar and Grosjean, 1990).
The partitioning of semi-volatile organic compounds (SOC) between vapor and particle
phases has received much attention (Cautreels and Cauwenberghe, 1978; Broddin et al.,
4-100
-------
1980; Hampton et al., 1983; Ligocki and Pankow, 1989; Cotham and Bidleman, 1992; Lane
et al., 1992; Kaupp and Umlauf, 1992; Pankow, 1992; Turpin et al., 1993b, 1996). Most
estimates of partition have relied on high-volume (hi-vol) sampling, using a filter to collect
particles followed by a solid adsorbent trap to collect the gaseous portion of SOC (e.g., Kaupp
and Umlauf, 1992, Foreman and Bidleman, 1990). Kaupp and Umlauf (1992) recently reported
that this approach, although not absolutely free from sorption and desorption artifacts, produces
reliable results. The maximum differences observed between hi-vol filter-solid adsorbent
sampling and impactor sampling (the latter believed to be less susceptible to these sampling
artifacts) did not exceed a factor of two.
There is good theoretical and experimental evidence that use of a diffusion denuder
technique significantly improves measurements of vapor-particle phase partitioning (Coutant
et al., 1988, 1989, 1992; Lane et al., 1988). However, at the present state of their technological
development, the reliability of denuders for investigation of atmospheric partitioning of non-
polar SOC needs to be improved, as suggested by contradictions in published field data (e.g.,
Kaupp and Umlauf, 1992). Gundel et al. (1992) used a proprietary XAD-4-coated tube for
vapor collection, followed by filter collection of organic aerosol particles and a sorbent bed to
quantitatively retain desorbed (volatilized) organic vapors. Denuders that remove ozone from
the air before it reaches the filter reduce the potential for artifact formation on the captured
particulate material during sampling (Williams and Grosjean, 1990).
Since the organic fraction of airborne particulate matter is typically a complex mixture of
hundreds to thousands of compounds distributed over many organic functional groups, its
chemical analysis is an extremely difficult task (Appel et al., 1977; Simoneit, 1984; Flessel
et al., 1991; Hildemann et al., 1991; Li and Kamens, 1993; Rogge et al., 1993a, 1993b, 1993c).
Analyses of organics generally begin with solvent extraction of the particulate sample. A variety
of solvents and extraction techniques have been used in the past. One common method is
sequential extraction with increasingly polar solvents, which typically separates the organic
material into nonpolar, moderately polar, and polar fractions (Daisey et al., 1982). This step is
usually followed by further fractionation using open-column liquid chromatography and/or high-
performance liquid chromatography (HPLC) in order to obtain several less complicated fractions
(e.g., Schuetzle and Lewtas, 1986; Atkinson et al., 1988).
4-101
-------
These fractions can then be analyzed further with high resolution capillary-column gas
chromatography (GC), combined with mass spectrometry (GC/MS), Fourier transform infrared
(GC/FTIR/MS) or other selective detectors.
Much of the recent work on the identification of nonpolar and semi-polar organics in
airborne samples has used bioassay-directed chemical analysis (Scheutzle and Lewtas, 1986),
and has focused on identification of fractions and compounds that are most likely to be of
significance to human health. In particular, PAHs and their nitro-derivatives (nitroarenes)
attracted considerable attention due to their mutagenic and, in some cases, carcinogenic
properties. More than 100 PAHs have been identified in the PM2 5 fraction of ambient
paniculate matter (Lee et al., 1981). While most of the nitroarenes found in ambient particles
are also present in primary combustion-generated emissions, some are formed from their parent
PAH in the atmospheric nitration reactions (e.g., Arey et al., 1986; Zielinska et al., 1989,
Ramdahl et al., 1986).
Little work has been done to date to chemically characterize the polar fraction in detail,
even though polar material accounts for up to half the mass and mutagenicity of soluble ambient
paniculate organic matter (Atherholt et al., 1985; Gundel et al., 1994). Until recently the polar
fraction has remained analytically intractable, since very polar and labile species interact with
conventional fractionation column packing materials and cannot be recovered quantitatively.
Recently, very polar paniculate organic matter has been successfully fractionated using
cyanopropyl-bonded silica (Gundel et al., 1994), with good recovery of mass and mutagenicity
(Kado et al., 1989). However, ambient paniculate polar organic material cannot be analyzed
with conventional GC/MS because of GC column losses resulting from adsorption, thermal
decomposition, and chemical interactions. New analytical techniques, such as HPLC/MS and
MS/MS, need to be applied if the chemical constituents of polar paniculate organic matter are to
be identified and quantified.
Most of the recent work on the identification of paniculate organic matter has focused on
mutagenic and carcinogenic compounds that are of significance to human health. Relatively
little work has been done to characterize individual compounds or classes of compounds that are
specific to certain sources of organic aerosol. In urban and rural atmospheres, as well as in the
remote troposphere, organic composition corresponding to chemical source profiles for of plant
waxes, resin residues, and long-chain hydrocarbons
4-102
-------
from petroleum residues have been found (e.g., Gagosian et al., 1981; Simoneit, 1984; Mazurek
etal., 1987, 1989, 1991; Simoneit et al., 1991). However, a variety of smaller, multi-functional
compounds characteristic of gas-to-particle conversion have also been observed (e.g., Finlayson-
Pitts and Pitts, 1986). These compounds tend to be present in the polar fraction of ambient
organic aerosol particles, having been formed from atmospheric chemical reactions of less polar
precursors. Little is currently known about the chemical composition of this polar fraction due
to the serious analytical difficulties mentioned above.
4.3.4.2 Analysis of Organic and Elemental Carbon
Three classes of carbon are commonly measured in aerosol samples collected on
quartz-fiber filters: (1) organic, volatile, or non-light absorbing carbon; (2) elemental or light-
absorbing carbon; and 3) carbonate carbon. Carbonate carbon (i.e., K2CO3, Na2CO3, MgCO3,
CaCO3) can be determined on a separate filter section by measurement of the carbon dioxide
(CO2) evolved upon acidification (Chow et al., 1993b; Johnson et al., 1981). Though progress
has been made in the quantification of specific organic chemical compounds in suspended
particles (e.g., Rogge et al., 1993a,b,c), sampling and analysis methods have not yet evolved for
use in practical monitoring situations.
Many methods have been applied to the separation of organic and elemental carbon in
ambient and source particulate samples (Mueller et al., 1971; Lin et al., 1973; Gordon, 1974;
Grosjean, 1975; Smith et al., 1975; Appel et al., 1976, 1979; Kukreja and Bove, 1976; Dod
et al., 1979; Johnson and Huntzicker, 1979; Macias et al., 1979; Malissa, 1979; Weiss et al.,
1979; Cadle et al., 1980a; Johnson et al., 1981b; Daisey et al., 1981; Novakov, 1982; Cadle and
Groblicki, 1982; Gerber, 1982; Huntzicker et al., 1982; Stevens et al., 1982; Wolff et al., 1982;
Japar et al., 1984; Chow et al., 1993b). Comparisons among the results of the majority of these
methods show that they yield comparable quantities of total carbon in aerosol samples, but the
distinctions between organic and elemental carbon are quite different (Countess, 1990; Hering
etal., 1990).
The definitions of organic and elemental carbon are operational and reflect the method and
purpose of measurement. Elemental carbon is sometimes termed "soot", "graphitic carbon," or
"black carbon." For studying visibility reduction, light-absorbing carbon is a more useful
concept than elemental carbon. For source apportionment by receptor models,
4-103
-------
several consistent but distinct fractions of carbon in both source and receptor samples are
desired, regardless of their light-absorbing or chemical properties. Differences in ratios of the
carbon concentrations in these fractions form part of the source profile that distinguishes the
contribution of one source from the contributions of other sources.
Light-absorbing carbon is not entirely constituted by graphitic carbon, since there are many
organic materials that absorb light (e.g., tar, motor oil, asphalt, coffee). Even the "graphitic"
black carbon in the atmosphere has only a poorly developed graphitic structure with abundant
surface chemical groups. "Elemental carbon" is a poor but common description of what is
measured. For example, a substance of three-bond carbon molecules (e.g., pencil lead) is black
and completely absorbs light, but four-bond carbon in a diamond is completely transparent and
absorbs very little light. Both are pure, elemental carbon.
Chow et al. (1993b) document several variations of the thermal (T), thermal/optical
reflectance (TOR), thermal/optical transmission (TOT), and thermal manganese oxidation
(TMO) methods for organic and elemental carbon. The TOR and TMO methods have been most
commonly applied in aerosol studies in the United States.
The TOR method of carbon analysis developed by Huntzicker et al. (1982) has been
adapted by several laboratories for the quantification of organic and elemental carbon on quartz-
fiber filter deposits. While the principle used by these laboratories is identical to that of
Huntzicker et al. (1982), the details differ with respect to calibration standards, analysis time,
temperature ramping, and volatilization/combustion temperatures. In the TOR method (Chow
et al., 1993b), a filter is submitted to volatilization at temperatures ranging from ambient to
550°C in a pure helium atmosphere, then to combustion at temperatures between 550 to 800°C
in a 2% oxygen and 98% helium atmosphere with several temperature ramping steps. The
carbon that evolves at each temperature is converted to methane and quantified with a flame
ionization detector. The reflectance from the deposit side of the filter punch is monitored
throughout the analysis. This reflectance usually decreases during volatilization in the helium
atmosphere owing to the pyrolysis of organic material. When oxygen is added, the reflectance
increases as the light-absorbing carbon is combusted and removed. Organic carbon is defined as
that which evolves prior to re-attainment of the original reflectance, and elemental carbon is
defined as that which evolves after the original reflectance has been re-attained. By this
definition, "organic carbon" is actually organic carbon that does not
4-104
-------
absorb light at the wavelength (632.8 nm) used, and "elemental carbon" is light-absorbing
carbon (Chow et al., 1993b). The TOT method applies the same thermal/optical carbon analysis
method except that transmission through instead of reflectance off of the filter punch is
measured. Thermal methods apply no optical correction and define elemental carbon as that
which evolves after the oxidizing atmosphere is introduced.
The TMO method (Fung, 1990) uses manganese dioxide (MnO2), present and in contact
with the sample throughout the analysis, as the oxidizing agent, and temperature is relied upon to
distinguish between organic and elemental carbon. Carbon evolving at 525°C is classified as
organic carbon, and carbon evolving at 850°C is classified as elemental carbon.
Carbon analysis methods require a uniform filter deposit because only a small portion of
each filter is submitted to chemical analysis. The blank filter should be white for light reflection
methods, and at least partially transparent for light transmission methods. The filter must also
withstand very high temperatures without melting during combustion.
Since all organic matter contains hydrogen as the most common elemental species, analysis
of hydrogen by proton elastic scattering analysis (PESA) has been developed by Cahill et al.
(1989). A correction must be made for hydrogen in sulfates and nitrates, but since the analysis is
done in a vacuum, water is largely absent. PESA has excellent sensitivity which is
approximately 20 times better than combustion techniques. This method requires knowledge of
the chemical state of sulfates, nevertheless, reasonable agreement was found as compared to the
combustion techniques.
4.3.4.3 Organic Aerosol Sampling Artifacts
Sampling artifacts contribute to inaccuracies in mass measurements of particulate organic
matter collected by filtration. They can generally be classified into two types: (1) adsorption on
filters or collected particulate matter of organic gases normally in the vapor phase causes
particulate organic mass to be overestimated, and (2) volatilization of collected organic material
during sampling leads to an underestimate of particulate organic mass. These artifacts can cause
significant errors in particle mass measurements in areas where a large fraction of the particulate
mass is organic.
4-105
-------
Vaporization Artifact
Significant loss of organic mass from filter samples occurs when clean air or nitrogen is
pumped through them after collection (Commins, 1962; Rondia, 1965; Van Vaeck et al., 1984).
This has frequently been referred to as "blow-off or "volatilization artifact" (Broddin et al.,
1980; Konig et al., 1980; Van Vaeck et al., 1984). Van Vaeck et al. (1984) found up to 70% of
some n-alkanes volatilized from the filter on exposure to a clean air stream. Coutant et al.
(1988) reported that the amount of fluoranthene and pyrene lost through the volatilization
artifact for a set of ambient samples ranged from 7 to 62% and 16 to 83%, respectively.
Eatough et al. (1989) concluded that 40 to 80% of the organic material was lost after collection
from samples at Hopi Point in the Southwestern United States. It has been proposed that an
upper limit for the volatilization artifact is reached if the concentration of the volatilizing species
reaches its equilibrium vapor concentration in the air exiting the filter, but that actual loss from
the filter can be substantially lower because of slow volatilization kinetics or strong adsorption
on particulate matter (Pupp et al., 1974). The volatilization artifact has been offered as a
possible explanation for frequently observed variations in concentrations of parti culate organic
matter with flow rate, face velocity and sampling period duration (Delia Fiorentina et al., 1975;
Appel et al., 1979; Schwartz et al., 1981). An increase in pressure drop across the filter during
sampling can also promote volatilization artifact if enough paniculate matter is collected (Van
Vaeck et al., 1984). However, pressure drop does not appear to explain artifact behavior under
typical sampling conditions if the pressure drop across the filter does not change during
sampling (McDow and Huntzicker, 1990; Turpin et al., 1994).
Adsorption Artifact
Other workers have been more concerned with adsorption of the gas-phase organics.
Cadle et al. (1983) reported that adsorbed vapor accounted for an average of 15% of the organic
carbon collected on quartz fiber filters. In the recent Carbonaceous Species Methods
Intercomparison Study it was estimated that organic vapor adsorption on filters caused organic
aerosol concentrations to be overestimated by 14 to 53% (Hering et al., 1990). Significant
adsorption of organic vapors has also been observed on backup filters from a variety of different
primary aerosol sources (Hildemann et al., 1991). The adsorption
4-106
-------
artifact appears to be inversely related to participate organic matter concentration, so that artifact
correction becomes more important at lower concentrations of particulate organic matter as
shown in Figure 4-30 (McDow and Huntzicker, 1990). Adsorption artifact also varies with face
velocity (McDow and Huntzicker, 1990; Turpin et al., 1993b) and sampling duration (McDow
and Huntzicker, 1993), and significant amounts of adsorbed vapor volatilizes when clean air
flows across the filter (McDow and Huntzicker, 1993). Because of this, it is not possible to
distinguish between adsorption and volatilization artifacts either by blowing clean air across a
filter or by a simple comparison of variations of collected organic mass with face velocity or
sampling duration. Adsorption occurs to a greater extent on filters which have already collected
organics on the filter surface during sampling than on clean filters not previously used for
sampling, suggesting that the filter becomes an increasingly better adsorbent as adsorbed vapors
build up on the filter (Gotham and Bidleman, 1992).
The following compounds have been observed to be adsorbed on quartz or glass fiber
filters: n-alkanes (Eichmann et al., 1979; Hart and Pankow, 1990), PAH (Ligocki and Pankow,
1989), and formaldehyde (Klippel and Warneck, 1980). Appel et al. (1989) analyzed backup
filters for carbonate and ruled out carbon dioxide as a major contributor to adsorption artifact in
Los Angeles on the basis of these analyses.
Artifact Correction
Appel et al. (1989) advocated a simple backup filter correction procedure described by
Equation 4-1:
Cp = QQJ-QQ2 (4-1)
where Cp is artifact corrected parti culate concentration, QQ1 represents the mass collected on
filter QQ1 and QQ2 represents the mass collected on downstream backup filter QQ2
(Figure 4-31). In some cases a modified backup filter correction procedure described by
Equation 4-2 appears to be more accurate (McDow and Huntzicker, 1990):
Cp = QQ1 ~ TQ2 (4-2)
4-107
-------
50
40
o
'5 30
0)
o
o
20
10
4 6 8 10 12
Uncorrected OC (|jgC/rfl )
14
16
Figure 4-30. Percent correction for vapor adsorption on quartz fiber filters for
submicrometer particle sampling at a face velocity of 40 cm s-1 for
13 samples in Portland, OR.
Source: McDow and Huntzicker (1990).
where Cp is artifact corrected particulate concentration, QQ1 represents the mass collected on
filter QQ1 in Figure 4-31, and TQ2 represents the mass collected from filter TQ2, the backup
filter behind a Teflon filter in a parallel sampling port.
Several approaches have been used to attempt to determine the relative importance of the
adsorption and volatilization artifacts. Using quartz fiber denuders to remove vapors upstream
of filter samples, Appel et al., (1989) found 59% and Fitz (1990) found 80% on average of the
organic mass adsorbed on the backup filter could be removed by the denuder, indicating that the
41% or 20% of the organic mass adsorbed on the backup filter was volatilized from the collected
paniculate matter.
4-108
-------
Inlet
Inlet
] QQ1
TQ1
] QQ2
] TQ2
Figure 4-31. Two types of filter series used for adsorption artifact corrections. QQ1 is a
quartz fiber filter, and QQ2 is a quartz fiber backup filter to a quartz filter.
TQ1 is a Teflon membrane filter, and TQ2 is a quartz fiber backup filter to
a Teflon filter.
Source: McDow and Huntzicker (1990).
McDow and Huntzicker (1990) used Equation 4-3 to correct for adsorption artifacts in
samples simultaneously collected at three different face velocities. They found that in four
experiments more than 80% of the observed difference in organic carbon mass was eliminated
by this correction procedure. In contrast, if the organic carbon mass on the backup filter was
added to that of the front filter the difference between samples collected at different face
velocities was significantly greater. This suggests that adsorption artifact is more likely to
account for observed face velocity differences than volatilization artifact.
Eatough et al. (1989, 1993) felt that both the adsorption and the volatilization artifacts
were important. Eatough concluded that the backup filter, either QQ2 or TQ2 in Figure 4-32,
would adsorb both organic material from the gas phase and organic vapors volatilized from the
collected particulate matter. In order to obtain a correct measure of the
4-109
-------
Sampler 1
DENUDER-
FILTER
QZn
QZi.a
GIF
1.1
Sampler 2
FILTER-
DENUDER
CIF2
Legend
Diffusion
Denuder
Quartz
Filter
I 1 Sorbent
I 1 Filter
Figure 4-32. Schematic of the BYU Organic Sampling System. Sampler 1 (denuder/filter)
and sampler 2 (filter/denuder).
Source: Eatough (1995).
4-110
-------
organic particulate matter, present in the ambient air in participate form, it would be necessary to
eliminate the adsorption artifact and add back the volatilization artifact. Accordingly, Eatough
collected particulate matter using two parallel sampling trains described in Figure 4-32 (Eatough
et al., 1989, 1993; Eatough, 1995) and proposed as an artifact correction equation:
Cp = Q1,1 + Ql,2 + CIF1J/E - CIF2J/E (4-3)
where: Cp is artifact corrected particulate concentration; Ql,l and Ql,2 are the organic carbon
masses collected on the first and second filters following the denuder in sampler 1 of Figure 4-
32, respectively; GIF 1,1 and CIF2,1 are the carbon masses collected on the sorbent samplers,
carbon impregnated filters (GIF) in samplers 1 and 2 in Figure 4-32, respectively; and E is the
vapor collection efficiency of the denuder. Eatough (1995) demonstrated that the denuder, made
from carbon impregnated filter paper (GIF), removed all of the gas phase organic that could be
adsorbed on the quartz fiber filter material. Thus, the organic material on Q 1,2 would be due to
the volatilization artifact only and Q2,2 - Ql,2 would give an indication of the adsorption
artifact (assuming independent adsorption of both artifacts). Any organic material volatilized
from the organic particles collected on Q 1,1 and not adsorbed on Q 1,1 or Q 1,2 would be
adsorbed on GIF 1,1. While the denuder is 100% efficient in removing organic material that
would adsorb on quartz fiber filters, it is not 100% efficient for adsorbing the organic material
that would be adsorbed by the carbon impregnated filters. Therefore, assuming that all of the
organic material vaporized from particles collected on Q2,l would be adsorbed on Q2,l, Q2,2 or
the denuder in Sampler 2, CIF2,1 may be used to correct CIF2,2 for any organic material which
passed through the denuder on sampler 1 and was adsorbed on GIF 1,2. Since the carbon
impregnated filters in the denuders are not 100% efficient they are each corrected for their
efficiency (measured separately by comparing the organic mass on several carbon impregnated
filters in series).
Several types of samplers have also been designed to reduce sampling artifacts. Van
Vaeck et al. (1984) designed a sampler which automatically replaced filters after short time
intervals. This prevented large increases in pressure drop across the filter observed during the
relatively long sampling periods they typically used. Several denuder systems have also
4-111
-------
been designed to reduced sampling artifacts (Appel, et al., 1989; Coutant et al., 1988; Eatough et
al., 1989, 1993; Fitz, 1990). Turpin et al. (1993b) developed a laminar flow separator, which
also utilizes differences in diffusion rates between vapors and particles to reduce sampling
artifacts.
Little is known concerning the chemical species responsible for sampling artifacts, with the
exception of the few species reported here. Volatile organic compounds (VOCs) such as
formaldehyde make a contribution to the adsorption artifact. Semi-volatile organic compounds
(SVOCs), those compounds such as n-alkanes and polycyclic aromatic hydrocarbons, which are
generally distributed between the vapor phase and particulate matter in the atmosphere, play a
role in both types of artifacts.
Equilibrium partitioning of SVOCs between condensed phase, vapor phase and adsorbed
phase depends on their temperature- dependent vapor pressure, the surface area of the collection
material, and their concentration. (Section 3.3.3; Junge, 1977; Yamasaki et al., 1982; Pankow,
1987). Some examples of possible causes of SVOC phase equilibrium shifts leading to sampling
artifacts are (1) changes in temperature, either if the air temperature changes during sampling, if
the sampler is cooled or heated, or if samples are allowed to stand in room air with a different
temperature than during sampling, (2) changes in surface area, either in ambient aerosol surface
area, or the increase in available surface area for adsorption experienced when an SVOC
encounters additional filter surface area, (3) changes in SVOC concentration, which can also
occur during sampling or after sample collection if samples are exposed to room air. Thus
SVOCs can vaporize during the temperature and relative humidity conditioning prescribed by
the Federal Reference Method for measuring particulate mass.
Conclusions
The following conclusions can be drawn from this literature review. (1) There is general
agreement that sampling artifacts can cause significant errors in the measurement of particulate
organic mass. (2) Disagreement exists about whether adsorption artifact or volatilization artifact
are the most important sampling artifact to consider. It is not clear to what extent disagreements
between studies are caused by differences in the aerosol sampled, sampling procedures used, or
interpretation of sampling results. (3) Little is known about the causes of sampling artifacts or
the individual species involved. (4) Sampling artifacts may be strongly influenced by changes in
4-112
-------
temperature or organic vapor concentration during sampling and storage. Procedures which do
not take these factors into consideration are likely to contribute to sampling artifact errors. (5)
Organic aerosol sampling artifacts can cause significant errors in particle mass measurements in
areas where a large fraction of the particulate mass is organic.
4.3.5 Methods Validation
The use of multiple methods and parallel samplers achieves both optimum performance
and quality assurance. While this has been a part of major research studies since the 1970s, its
extension to long-term monitoring of aerosols was most extensively introduces in the SCENES
and IMPROVE visibility programs (Eldred and Cahill, 1984). The concept was labeled,
"Integral Redundancy," and was recently adopted by the United Nation's Global Atmospheric
Watch Program.
The internal consistency checks applied to the IMPROVE network are listed as follows:
(1) Mass (gravimetric) is compared to the sum of all elements on the Teflon-membrane
filter of Channel A (PIXE, PESA, XRF analysis; Internally XRF and PIXE are
compared for elements around iron). This was pioneered in the SCENES program and
is now the standard practice for many aerosol studies.
(2) Sulfate, by ion chromatography on Channel B's nylon filter, after an acidic vapor
denuder, is compared to sulfur (X3) from Channel A's Teflon-membrane filter by
PIXE. Agreement is excellent, except for summer.
(3) Organic matter, by combustion on Channel C's quartz-fiber filter stack, is compared to
organic matter via PESA analysis of hydrogen on Channel A's Teflon-membrane
filter. This is an exceptionally sever test due to the nature of organics. These
comparisons are made for every IMPROVE analysis, yielding about 25,000
comparisons to date (Malm et al., 1994).
These types of data validation checks should be carried out in every PM measurement
program to ensure the accuracy, precision, and validity of the chemical analysis data.
4-113
-------
4.4 BIOAEROSOLS SAMPLING AND ANALYSIS
4.4.1 Analytical Methods
Because of the complexity of the particles contained in the term "bioaerosols" no single
analytical method is available that will allow assessment of all of the potential biologically-
derived particle in an aerosol. Table 4-5 is an overview of the available analytical methods,
examples of the kinds of agents detected, and some sampling considerations.
TABLE 4-5. OVERVIEW OF ANALYTICAL METHODS
Bioassay
Chemical assays
Molecular
techniques
Kinds of Agents
Examples
Sampling
Considerations
Culture
Microscopy
Immunoassay
culturable
organisms
recognizable
particles
agents that
stimulate antibodies
fungal spores, yeasts,
bacteria, viruses (rarely
used)
pollen, fungal spores,
bacteria
allergens, aflatoxin, glucan
viability must be
protected
good optical
quality is required
agents must be
elutable from
agents exerting
observable effects
in a biological
system
chemicals with
recognized
characteristics
DNA or RNA-
containing particles
endotoxin, cytotoxins
trichothecene toxins
specific organisms
sampling medium.
Activity must be
preserved
same as
immunoassay
same as
immunoassay
A good principle to guide the kind of analysis for use in detecting a particular bioaerosol is
to use the approach that best characterizes the agent of disease rather than the
4-114
-------
agent-bearing particle. Thus, while culture is appropriate where infectious disease is of concern,
or where you know that allergens are only released as a spore germinates, it is likely to be only a
poor indicator for mycotoxin exposure. Culture always underestimates actual levels of any
viable aerosol because no culture conditions are appropriate for all cells. The extent of the
underestimate can be very large if an aerosol is damaged or consists primarily of non-living
cells. The reason culture is not the best approach for evaluating mycotoxins is because it is
unlikely that viability is a necessary requirement for mycotoxin release from spores (although
this has not been studied).
Microscopy allows direct counts of identifiable particles. Light microscopy will reveal
particles as small as 1.5 //m reliably. Identification of the type of particle requires either some
morphological characteristic unique to the particle, or some factor that can be labelled with a
visible dye. Most pollens and many fungal spores can be placed in relevant groupings by
microscopy alone. Bacteria, on the other hand, can only be counted. Specific techniques to
enhance visibility based on specific immune responses or DNA polymerization techniques have
yet to be developed.
Immunoassays detect the actual agent of hypersensitivity disease. Two types are
commonly used: one based on a mixture of polyclonal antibodies that detects a relatively wide
range of allergens, and the other based on monoclonal antibodies that detects only the single
allergen to which the antibody is detected. Endotoxin is measured using a bioassay that involves
dose-dependant clotting of lysate from the amoebocytes of horseshoe crabs. This is not only an
agent-specific assay, but actually measures biological activity of the endotoxin rather than
simply the number of molecules.
4.4.2 Sample Collection Methods
Bioaerosol particles follow the principles of physics like any other particle type, and are
collected from aerosols by equipment that use these common physical principles. Bioaerosol
sampling devices were recently reviewed in depth by Macher et al., 1995. The most commonly
used bioaerosol samplers are suction sieve impactors that collect particles directly on culture
media. The second most commonly used types are slit impactors that collect particles either on
rotating plates of agar, or on grease-coated surfaces. Rotating arm impactors are often used for
the collection of pollen in clinical allergy practices across the country (American Academy of
4-115
-------
Allergy and Immunology, 1994). In addition to the impactors, bioaerosols are also collected
using filtration, either with filters mounted in cassettes or on large sheets of filter material
mounted in high-volume suction samplers. Liquid impingers are also used under research
conditions.
Analysis of culture plate samples is more or less restricted to static culture, although one
group has developed a procedure for suspending the catch in a liquid, and using dilution culture
to increase the upper level of sensitivity. For static culture, the maximum number of fungal
colonies on a 100 mm petri plate that does not result in inhibition between colonies is about 30.
The number of bacteria is much higher (-100). Sieve plate impactors have a limited number of
sites available for deposition, so that above some given number than depends on the number of
holes in the sieve plate, multiple impactions occur. For biological aerosols, this means that only
one colony of one organism is likely to appear at each site although several different kinds of
organisms might have been collected. Rotating slit culture samplers do not present this
constraint, although the upper limit to prevent competition losses remains in effect.
Analysis of samples collected on greased surfaces is generally limited to microscopy,
although some attempts have been to transfer allergens to nitrocellulose membranes and analyze
by immunoassay (immunoblotting). Filtration samples can be analyzed by culture, microscopy,
and by elution followed by immuno- or bioassay. Obviously these are the most versatile
devices. However, cultural counts made from filter collections may severely underestimate
actual levels because of desiccation on the filter. Microscopic analysis requires large numbers of
particles on the filter, so that, unless long sampling times are used, the sensitivity can be poor.
Filter collections have been the choice for samples to be analyzed by immunoassay (e.g., cat
allergens) and bioassay (e.g., Endotoxin).
4.5 SUMMARY
Though much of the discussion in the preceding sub-sections has been specific to different
sampling and analysis methods, several generalizations can be drawn.
First, it is found that samples taken to determine compliance with air quality standards are
often used for other purposes, such as source apportionment, personal exposure, and chemical
4-116
-------
characterization. Different sampling systems that are more closely coupled to the intended
analysis methods may be needed to attain additional monitoring objectives.
Second, though off-site gravimetric analysis of filter samples is straightforward and
relatively inexpensive in terms of equipment, more costly in-situ monitors have the potential to
provide higher time resolution, more frequent sampling intervals, and possibly reduced
manpower requirements. The mass concentrations obtained may not always be comparable
between the manual and automated methods, owing to differences in particle volatilization and
liquid water content of off-site and in situ measurements.
Third, technology is now proven and available to measure the major chemical components
of suspended particles, e.g. many separate elements, organic carbon, elemental carbon, sulfate,
nitrate, ammonium, and H+ ions. With reasonable assumptions regarding oxide and
hydrocarbon forms, most of the measured mass at many locations can be accounted for by these
chemical measurements. This technology could be applied more routinely than it has in the past
to better characterize particles to determine compliance with future air quality standards.
Fourth, since ambient particle size distributions contain fine and coarse particle modes,
with a minimum between them in the 1 to 3 //m size range, shifts in inlet cut-points near the
2.5 (j,m size range are not expected to have a large effect on the mass collected owing to the low
proportion of particles with sizes near this cut-point. This contrasts to the sensitivity of PM10
mass concentrations to small shifts in the cutpoints of PM10 inlets, where the maximum of the
coarse mode occurs between 6 and 25 //m (Lundgren and Burton, 1995).
Fifth, concentrations of volatile chemicals (such as ammonium nitrate or certain organic
compounds) and liquid water may change during sampling, during sample transport and storage,
and during sample analysis. Liquid water may be removed by lowering the relative humidity
surrounding the sample by heating the sampled air stream, or by selectively denuding the
airstream of water vapor. Several sampling systems involving diffusion denuders and absorbing
substrates operating in series and in parallel have been demonstrated to quantify volatilized
particles, but these are not practical for sustained, long-term monitoring on limited budgets.
Finally, collocated studies show substantial differences between mass concentration
measurements acquired by different sampling systems. They also show differences for similar
sampling systems for which procedures are somewhat different. Inlet maintenance, filter
handling and storage, laboratory analyses, and quality control procedures are just as important
4-117
-------
variables as sampler design in explaining these differences. Inlet characteristics and particle
volatilization properties are the most important variables that cause mass concentrations to
differ. The lack of common calibration standards is one of the major reasons for differences
between certain chemical analysis results.
This chapter also briefly describes the technical capabilities and limitations of specific
aerosol sampling procedures, focusing on those that (1) were used to collect data supporting
other sections in this document, (2) supported the existing PM10, TSP and Pb regulations, and (3)
have application in development of a possible fine particle standard. The discussion of aerosol
separation technologies is divided between devices used to mimic the larger particle penetration
rationales for the upper airways, and those used to mimic smaller particle penetration to the sub-
thoracic regions. The applications of performance specifications to define these measurement
systems for regulatory purposes are discussed with observations suggesting that the current
specification process has not always assured the necessary sampling accuracy. Particle sampling
systems for specialty applications, including automated samplers and personal exposure
monitors, are briefly described.
4.5.1 PM10 Sampling
Laboratory and field testing reported in the literature since 1987 suggest that the EPA
specifications and test requirements for PM10 samplers have not adequately controlled the
differences observed in collocated ambient sampling. The most significant performance flaws
have combined to produce mass concentration biases as large as 60%. These biases appear to
have resulted from the combined factors of (1) allowing a cutpoint tolerance of 10 ± 0.5 jim,
(2) placing an inadequate restriction on internal particle bounce, and (3) allowing a degradation
of particle separation performance as certain technology PM10 inlets became soiled. The
between-sampler bias from a ±0.5 jim tolerance limit is predictable and should provide PM10
concentration differences significantly less than ±10% in almost all cases. Design practices
(primarily surface coatings with viscous oil) to minimize the penetration caused by bounce and
resuspension have been shown to be very effective. The magnitude of biases from soiling events
can be accommodated by not allowing the inlet to become excessively dirty during operation
4-118
-------
through routine cleaning prior to sampling. Particle bounce or soiling problems have not been
reported for the PM10 inlets for the dichotomous sampler.
Based on the current understanding of the PM10 sampling process, it could be expected that
sampling systems can now be designed and concentration measurements made that are within
10% of the true concentration. This range poses the greatest concern where the measured
concentrations are near a standard exceedance level. A review by EPA of the current PM10
performance requirements and possible amendments of the existing specifications may be
appropriate, given the information base now available.
4.5.2 Fine Particle Sampling
The technology is available to provide an accurate Fine particle cutpoint (e.g. 1.0 or
2.5 jim) for routine sampling. Virtual impactors and cyclones have been shown to be the most
trouble-free and versatile methodologies. The exclusion of larger particles using a scalping inlet
eliminates many of the transport and loss problems encountered during PM10 sampling. The
absence of the Coarse particle fraction, however, exaggerates the problems inherent with Fine
particle chemistry, such as particle-substrate interactions and sublimation losses. Although it
could be expected that Fine particle mass concentration measurements can be made within 10%
of the true concentration, accurate chemical speciation may require more comprehensive
sampling system components, including gas stream denuders and sequential filter packs.
4.5.3 Concentration Corrections to Standard Conditions
The appropriateness of the correction of particulate concentrations to a reference
temperature and pressure is currently under review at EPA. Aerodynamic sampling requires
incorporation of local conditions to provide the correct velocities for accurate particle size
separation. Correcting the collection volume to standard conditions may improperly influence
interpretations of the developed relationships between particle concentration and adverse health
responses. It appears to be more appropriate to compute particle concentrations at site
conditions and provide temperature and barometric pressure data subsequently, as needed for
data interpretation.
4-119
-------
4.5.4 Performance Versus Design Specifications for Sampling Systems
The current EPA PM10 Reference and Equivalent Method program established in 1987 is
based on providing the necessary data quality by using sampling performance specifications.
Several research studies have recently reported that key elements of the sampling process were
inadequately considered when the original performance specifications were developed. The
observations from these controlled studies have been bolstered by reviews of field data from
collocated PM10 samplers that showed substantial biases under certain conditions. The particle
sampling process is complex. Obtaining an acceptable bias level using performance standards is
difficult, but not impossible, if the appropriate developmental research is identified and
implemented. The alternative approach of defining sampling systems by design specifications
seems attractive, but may ultimately pose more problems than are solved without producing
better quality data. Additionally, specification of a sampling system by design would have the
undesirable attribute of virtually eliminating further new technology research. The approach for
specifying particle sampling systems is currently under review at EPA.
4.5.5 Automated Sampling
The performances of two sampling methods that are currently designated as Equivalent
PM10 methods by EPA - beta attenuation and the TEOM sampler - have been evaluated
extensively in field settings. Although acceptable comparisons with EPA Reference sampling
methods are reported in collocated field studies, attention must be paid to situations where
significant biases existed. These biases have been attributed to a number of factors, but focused
on the treatment of the particle sample during and after collection. The presence of highly
reactive or unstable particles at sampling locations in the western U.S. appears to cause the
greatest concern, because of a higher proportion of these species. These bias issues are
significant because they complicate the use of automated samplers as "triggers" for control
strategy actions, and they question the adequacy of the existing performance specifications for
equivalent PM10 sampling systems.
4-120
-------
4.5.6 PM Samplers for Special Applications
Reviews of typical U.S. personal activity patterns suggest that personal exposure sampling
for particulates should also be considered in developing population risk assessments. Relatively
unobtrusive personal sampling systems have been designed for a number of particle size
cutpoints, and recent studies suggest that acceptable accuracies and precisions are possible. The
collection of particle size distribution data can assist in identifying paniculate sources and
subsequent studies of particle transport and fate. Well characterized cascade impactors are
available that cover the aerodynamic size range from at least 0.1 to 10 jim. More automated
optical systems are also available, providing either optical or aerodynamic diameter ranges from
about 0.5 to 10 jim. Source apportionment sampling systems are available to assist in relating
the chemical attributes of ambient particulate matter to the chemical "signatures" from various
source categories. This is accomplished by using sampling system components and collection
substrates designed to collect specific chemical classes (e.g., a suite of individual metals,
speciated carbon) in defined particle size categories.
4-121
-------
REFERENCES
Allen, D. T.; Palen, E. J.; Haimov, M. I.; Hering, S. V.; Young, J. R. (1994) Fourier transform infrared spectroscopy of
aerosol collected in a low pressure impactor (LPI/FTIR): method development and field calibration. Aerosol
Sci. Technol. 21: 325-342.
American Academy of Allery & Immunology. (1994) Aeroallergen Monitoring Network 1994 pollen and spore report.
Milwaukee, WI: American Academy of Allery & Immunology.
American Conference of Governmental Industrial Hygienists (ACGIH). (1985) Particle size-selective sampling in the
workplace: report of the ACGIH technical committee on air sampling procedures. Cincinnati, OH: American
Conference of Governmental Industrial Hygienists.
American Conference of Governmental Industrial Hygienists (ACGIH). (1994) 1994-1995 threshold limit values for
chemical substances and physical agents and biological exposure indices. Cincinnati, OH: American
Conference of Governmental Industrial Hygienists.
Anand, N. K.; McFarland, A. R.; Kihm, K. D.; Wong, F. S. (1992) Optimization of aerosol penetration through
transport lines. Aerosol Sci. Technol. 16: 105-112.
Anlauf, K. G.; MacTavish, D. C.; Wiebe, H. A.; Schiff, H. I.; Mackay, G. I. (1988) Measurement of atmospheric nitric
acid by the filter method and comparisons with the tuneable diode laser and other methods. Atmos. Environ.
22: 1579-1586.
Appel, B. R.; Colodny, P.; Wesolowski, J. J. (1976) Analysis of carbonaceous materials in southern California
atmospheric aerosols. Environ. Sci. Technol. 10: 359-363.
Appel, B. R.; Hoffer, E. M.; Haik, M.; Wall, S. M.; Kothny, E. L.; Knights, R. L.; Wesolowski, J. J. (1977)
Characterization of organic particulate matter. Final report. Sacramento, CA: California Air Resources Board;
report no. ARB-R-5-682-77-72. Available from: NTIS, Springfield, VA; PB-279209.
Appel, B. R.; Hoffer, E. M.; Kothny, E. L.; Wall, S. M.; Haik, M.; Knights, R. L. (1979) Analysis of carbonaceous
material in southern California atmospheric aerosols. 2. Environ. Sci. Technol. 13: 98-104.
Appel, B. R.; Tokiwa, Y.; Haik, M. (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
Appel, B. R.; Tokiwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact particulate sulfate and nitrate formation on filter
media. Atmos. Environ. 18: 409-416.
Appel, B. R.; Povard, V.; Kothny, E. L. (1988a) Loss of nitric acid within inlet devices intended to exclude coarse
particles during atmospheric sampling. Atmos. Environ. 22: 2535-2540.
Appel, B. R.; Tokiwa, Y.; Kothny, E. L.; Wu, R.; Povard, V. (1988b) Evaluation of procedures for measuring
atmospheric nitric acid and ammonia. Atmos. Environ. 22: 1565-1573.
Appel, B. R.; Cheng, W.; Salaymeh, F. (1989) Sampling of carbonaceous particles in the atmosphere—II. Atmos.
Environ. 23: 2167-2175.
Arey, J.; Zielinska, B.; Atkinson, R.; Winer, A. M.; Ramdahl, T.; Pitts, J. N., Jr. (1986) The formation of nitro-PAH
from the gas-phase reactions of fluoranthene and pyrene with the OH radical in the presence of NOr Atmos.
Environ. 20: 2339-2345.
4-122
-------
Arey, J.; Zielinska, B.; Atkinson, R.; Winer, A. M. (1987) Poly cyclic aromatic hydrocarbon and nitroarene
concentrations in ambient air during a wintertime high-NO^ episode in the Los Angeles basin. Atmos. Environ.
21: 1437-1444.
Arey, J.; Zielinska, B.; Atkinson, R.; Winer, A. M. (1988) Formation of nitroarene s during ambient high-volume
sampling. Environ. Sci. Technol. 22: 457-462.
Arnold, S.; Hague, W.; Pierce, G.; Sheetz, R. (1992) The use of beta gauge monitors for PSI and every day
SIP monitoring: an overview of the Denver experience. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and
nontraditional particulate source controls, an A&WMA/EPA international specialty conference, v. I; January;
Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 13-23. (A&WMA transactions
series no. 22).
Atherholt, T. B.; McGarrity, G. J.; Louis, J. B.; McGeorge, L. J.; Lioy, P. J.; Daisey, J. M.; Greenberg, A.; Darack, F.
(1985) Mutagenicity studies of New Jersey ambient air particulate extracts. In: Waters, M. D.; Sandhu, S. S.;
Lewtas, J.; Claxton, L.; Strauss, G.; Nesnow, S., eds. Short-term bioassays in the analysis of complex
environmental mixtures IV. New York, NY: Plenum Press; pp. 211 -231. (Environmental science research: v.
32).
Atkinson, R.; Arey, J.; Winer, A. M.; Zielinska, B. (1988) A survey of ambient concentrations of selected poly cyclic
aromatic hydrocarbons (PAH) at various locations in California [final report]. Sacramento, CA: California Air
Resources Board; contract no. A5-185-32.
Bailey, D. L. R.; Clayton, P. (1980) The measurement of suspended particulate and carbon concentrations in the
atmosphere using standard smoke shade methods. Stevenage, Hertfordshire, United Kingdom: Warren Spring
Laboratory; report no. LR 325 (AP).
Ball, D. J.; Hume, R. (1977) The relative importance of vehicular and domestic emissions of dark smoke in greater
London in the mid-1970s, 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.
Baron, P. A. (1993) Measurement of asbestos and other fibers. In: Willeke, K.; Baron, P. A., eds. Aerosol
measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 560-590.
Baron, P. A.; Willeke, K. (1993) Aerosol fundamentals. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement:
principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 8-22.
Baron, P. A.; Mazumder, M. K.; Cheng, Y. S. (1993) Direct-reading techniques using optical particle detection. In:
Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles, techniques, and applications. New York, NY:
Van Nostrand Reinhold; p. 381-409.
Bartley, D. L.; Breuer, G. M. (1982) Analysis and optimization of the performance of the 10 mm cyclone. Am. Ind.
Hyg. Assoc. J. 43: 520-528.
Benner, C. L.; Eatough, D. J.; Eatough, N L.; Bhardwaja, P. (1991) Comparison of annular denuder and filter pack
collection of HNO3 (g), HNO2 (g), SO2 (g), and particulate-phase nitrate, nitrite and sulfate in the south-west
desert. Atmos. Environ. Part A 25: 1537-1545.
Bennett, R. L.; Stockburger, L.; Barnes, H. M. (1994) Comparison of sulfur measurements from a regional fine particle
network with concurrent acid modes network results. Atmos. Environ. 28: 409-419.
Bidleman, T. F.; Billings, W. N.; Foreman, W. T. (1986) Vapor-particle partitioning of semivolatile organic
compounds: estimates from field collections. Environ. Sci. Technol. 20: 1038-1043.
4-123
-------
Biswas, P.; Jones, C. L.; Flagan, R. C. (1987) Distortion of size distributions by condensation and evaporation in
aerosol instruments. Aerosol Sci. Technol. 7: 231-246.
Botham, R. A.; Hughson, G. W.; Vincent, J. H.; Mark, D. (1991) The development of a test system for investigating the
performances of personal aerosol samplers under actual workplace conditions. Am. Ind. Hyg. Assoc. J. 52:
423-427.
Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
Technol. 23: 1408-1412.
Brockmann, J. E.; Yamano, N.; Lucero, D. (1988) Calibration of the aerodynamic particle sizer 3310 (APS-3310) with
polystyrene latex monodisperse spheres and oleic acid monodisperse particles. Aerosol Sci. Technol. 8:
279-281.
Broddin, G.; Cautreels, W.; Van Cauwenberghe, K. (1980) On the aliphatic and poly aromatic hydrocarbon levels in
urban and background aerosols from Belgium and the Netherlands. Atmos. Environ. 14: 895-910.
Brown, R. C.; Wake, D.; Thorpe, A.; Hemingway, M. A.; Roff, M. W. (1994) Preliminary assessment of a device for
passive sampling of airborne particulate. Ann. Occup. Hyg. 38: 303-318.
Bruynseels, F.; Storms, H.; Van Grieken, R.; Van der Auwera, L. (1988) Characterization of North Sea aerosols by
individual particle analysis. Atmos. Environ. 22: 2593-2602.
Buckley, T. J.; Waldman, J. M.; Freeman, N. C. G.; Lioy, P. J.; Marple, V A.; Turner, W. A. (1991) Calibration,
intersampler comparison, and field application of a new PM-10 personal air-sampling impactor. Aerosol Sci.
Technol. 14: 380-387.
Buettner, H. (1990) Measurement of the size of fine nonspherical particles with a light-scattering particle counter.
Aerosol Sci. Technol. 12: 413-421.
Butler, F. E.; lungers, R. H.; Porter, L. F.; Riley, A. E.; Toth, F. J. (1978) Analysis of air particulates by ion
chromatography: comparison with accepted methods. In: Sawicki, E.; Mulik, J. D.; Wittgenstein, E., eds. Ion
chromatographic analysis of environmental pollutants. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.;
pp. 65-76.
CEN (European Committee for Standardization). (1993) Workplace atmospheres—size fraction definitions for
measurement of airborne particles. London, United Kingdom: British Standards Institution; publication no. BS
EN 481: 1993; BS 6069: section 3.5: 1993.
Cadle, S. H.; Groblicki, P. J. (1982) An evaluation of methods for the determination of organic and elemental carbon in
particulate samples. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric life cycle. New
York, NY: Plenum Press; pp. 89-109.
Cadle, S. H.; Groblicki, P. J.; Stroup, D. P. (1980a) Automated carbon analyzer for particulate samples. Anal. Chem.
52: 2201-2206.
Cadle, S. H.; Nebel, G. J.; Williams, R. L. (1980b) Measurements of unregulated emissions from General Motors'
light-duty vehicles. SAE Trans. 88: 2381-2401.
Cadle, S. H.; Groblicki, P. J.; Mulawa, P. A. (1983) Problems in the sampling and analysis of carbon particles. Atmos.
Environ. 17: 593-600.
Cahill, T. A. (1980) Proton microprobes and particle-induced X-ray analytical systems. Annu. Rev. Nucl. Part. Sci. 30:
211-252.
4-124
-------
Cahill, T. A.; Wakabayashi, P. (1993) Compositional analysis of size-segregated aerosol samples. In: Newman, L., ed.
Measurement challenges in atmospheric chemistry. Washington, DC: American Chemical Society; pp.
211-228. (Advances in chemistry series no. 232).
Cahill, T. A.; Eldred, R. A.; Kusko, B. H.; Feeney, P. I; Malm, W. C. (1987) Concentrations of natural hydrocarbon
particles at National Park Service sites as derived from mass/hydrogen/sulfur correlations. In: Bhardwaja, P. S.,
ed. Visibility protection: research and policy aspects, an APCA international specialty conference; September
1986; Grand Teton National Park, WY. Pittsburgh, PA: Air Pollution Control Association; pp. 407-417.
(APCA transactions series no. TR-10).
Cahill, T. A.; Eldred, R. A.; Motallebi, N; Malm, W. C. (1989) Indirect measurement of hydrocarbon aerosols across
the United States by nonsulfate hydrogen-remaining gravimetric mass correlations. Aerosol Sci. Technol. 10:
421-429.
Cahill, T. A.; Eldred, R. A.; Feeney, P. I; Beveridge, P. I; Wilkinson, L. K. (1990) The stacked filter unit revisited. In:
Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA international specially conference; October
1989; Estes Park, CO. Pittsburgh, PA: Air & Waste Management Association; pp. 213-221. (A&WMA
transactions series no. TR-17).
Cahill, T. A.; Gill, T. E.; Gillette, D. A.; Gearhart, E. A.; Reid, J. S.; Yau, M.-L. (1994) Generation, characterization and
transport of Owens (Dry) Lake dusts [final report]. Sacramento, CA: California Air Resources Board; contract
no. A132-105.
Cantrell, B. K.; Williams, K. L.; Watts, W. F., Jr.; Jankowski, R. A. (1993) Mine aerosol measurement. In: Willeke, K.;
Baron, P. A., eds. Aerosol measurement: principles, techniques, and applications. New York, NY: Van
Nostrand Reinhold; pp. 591-611.
Casuccio, G. S.; Janocko, P. B.; Lee, R. J.; Kelly, J. F.; Dattner, S. L.; Mgebroff, J. S. (1983) The use of computer
controlled scanning electron microscopy in environmental studies. J. Air Pollut. Control Assoc. 33: 937-943.
Cautreels, W.; Van Cauwenberghe, K. (1978) Experiments on the distribution of organic pollutants between airborne
particulate matter and the corresponding gas phase. Atmos. Environ. 12: 1133-1141.
Chan, T. L.; Lippmann, M. (1980) Experimental measurements and empirical modelling of the regional deposition of
inhaled particles in humans. Am. Ind. Hyg. Assoc. J. 41: 399-409.
Charlson, R. J.; Ahlquist, N. C.; Horvath, H. (1968) On the generality of correlation of atmospheric aerosol mass
concentration and light scatter. Atmos. Environ. 2: 455-464.
Chow, J. C.; Watson, J. G. (1994) Contemporary source profiles for geological material and motor vehicle emissions
[final report]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards.
Chow, J. C.; Watson, J. G.; Ono, D. M.; Mathai, C. V. (1993a) PM10 standards and nontraditional particulate source
controls: a summary of the A&WMA/EPA international specialty conference. J. Air Waste Manage. Assoc. 43:
74-84.
Chow, J. C.; Watson, J. G.; Pntchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993b) The DRI
thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere; April
1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
4-125
-------
Chow, J. C.; Fujita, E. M; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994) Evaluation of filter-based
aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess. 30:
49-80.
Clayton, P.; Wallin, S. C. (1982) Comparison of ambient suspended particulate matter concentrations as measured by
the British smoke sampler and the high volume sampler. Atmos. Environ. 16: 2031-2032.
Code of Federal Regulations. (1991) Appendix J to Part 50 — reference method for the determination of particulate
matter as PM10 in the atmosphere. C. F. R. 40: §50.
Colome, S. D.; Kado, N. Y.; Jaques, P.; Kleinman, M. (1992) Indoor-outdoor air pollution relations: particulate matter
less than 10 //m in aerodynamic diameter (PM10) in homes of asthmatics. Atmos. Environ. Part A 26:
2173-2178.
Commins, B. T. (1962) Interim report on the study of techniques for determination of poly cyclic aromatic
hydrocarbons in air. In: Sawicki, E.; Cassel, K., Jr., eds. Analysis of carcinogenic air pollutants: [proceedings
of a symposium]; August 1961; Cincinnati, OH. Bethesda, MD: U.S. Department of Health, Education, and
Welfare, Public Health Service, National Cancer Institute; pp. 225-233. (National Cancer Institute monograph
no. 9).
Conner, T. L.; Miller, J. L.; Willis, R. D.; Kellogg, R. B.; Dann, T. F. (1993) Source apportionment of fine and coarse
particles in southern Ontario, Canada. Presented at: 86th annual meeting & exhibition of the Air & Waste
Management Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
93-TP-58.05.
Cook, J. P.; Oslund, W. E.; Frank, N. (1995) Evaluation of fine particulate samplers (PM^) in an area of volatile
constituents. In: Particulate matter: health and regulatory issues: proceedings of an international specialty
conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; pp. 277-296.
(A&WMA publication VIP-49).
Cooper, D. W.; Guttrich, G. L. (1981) A study of the cut diameter concept for interpreting particle sizing data. Atmos.
Environ. 15: 1699-1707.
Cornille, P.; Maenhaut, W.; Pacyna, J. M. (1990) Sources and characteristics of the atmospheric aerosol near Damascus,
Syria. Atmos. Environ. Part A 24: 1083-1093.
Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26: 469-478.
Countess, R. J. (1990) Interlaboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol. 12: 1 14-121.
Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a P gauge for aerosol mass
determinations. Environ. Sci. Technol. 16: 236-239.
Coutant, R. W.; Brown, L.; Chuang, J. C.; Riggin, R. M.; Lewis, R. G. (1988) Phase distribution and artifact formation
in ambient air sampling for polynuclear aromatic hydrocarbons. Atmos. Environ. 22: 403-409.
Coutant, R. W.; Callahan, P. J.; Kuhlman, M. R.; Lewis, R. G. (1989) Design and performance of a high-volume
compound annular denuder. Atmos. Environ. 23: 2205-221 1.
Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease -coated denuders for
collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
Criss, J. W. (1976) Particle size and composition effects in x-ray fluorescence analysis of pollution samples. Anal.
Chem. 48: 179-186.
4-126
-------
Crutcher, E. R. (1982) Light microscopy as an analytical approach to receptor modeling. In: A specialty conference on:
receptor models applied to contemporary pollution problems; October; Danvers, MA. Pittsburgh, PA: Air
Pollution Control Association; pp. 266-284. (SP-48).
Currie, L. A. (1982) Contemporary particulate carbon. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon:
atmospheric life cycle. New York, NY: Plenum Press; pp. 245-260.
Daisey, J. M.; McCaffrey, R. I; Gallagher, R. A. (1981) Poly cyclic aromatic hydrocarbons and total extractable
particulate organic matter in the Arctic aerosol. Atmos. Environ. 15: 1353-1363.
Daisey, J. M.; Hershman, R. J.; Kneip, T. J. (1982) Ambient levels of particulate organic matter in New York City in
winter and summer. Atmos. Environ. 16: 2161-2168.
Dams, R.; Robbins, J. A.; Rahn, K. A.; Winchester, J. W. (1970) Nondestructive neutron activation analysis of air
pollution particulates. Anal. Chem. 42: 861-867.
Davis, B. L. (1978) Additional suggestions for X-ray quantitative analysis of high-volume filter samples.
Atmos. Environ. 12: 2403-2406.
Davis, B. L. (1980)"Standardless" X-ray diffraction quantitative analysis. Atmos. Environ. 14: 217-220.
Delia Fiorentina, H.; De Wiest, F.; De Graeve, J. (1975) Determination par spectrometrie infrarouge de la matiere
organique non volatile associee aux particules en suspension dans 1'air—II. Facteurs influencant 1'indice
aliphatique. [Determination of the nonvolatile organic matter associated with suspended particles by infrared
spectrometry—II. Factors influencing the aliphatic index]. Atmos. Environ. 9: 517-522.
Derde, M.-P.; Buy dens, L.; Guns, C.; Massart, D. L.; Hopke, P. K. (1987) Comparison of rule-building expert systems
with pattern recognition for the classification of analytical data. Anal. Chem. 59: 1868-1871.
Divita, F., Jr. (1993) Size distributions and sources of submicrometer atmospheric particles in Washington, D.C. and
Philadelphia, PA [Ph.D. dissertation]. College Park, MD: University of Maryland. Available from: University
Microfilms, Ann Arbor, MI; publication no. AAC94-25018.
Dod, R. L.; Rosen, H.; Novakov, T. (1979) Atmospheric aerosol research annual report for 1977-1978. Berkeley, CA:
Lawrence Berkeley Laboratories; document no. 8696.
Durham, J. L.; Wilson, W. E.; Bailey, E. B. (1978) Application of an SO2-denuder for continuous measurement of
sulfur in submicrometric aerosols. Atmos. Environ. 12: 883-886.
Dzubay, T. G.; Mamane, Y. (1989) Use of electron microscopy data in receptor models forPM-10. Atmos. Environ.
23: 467-476.
Dzubay, T. G.; Nelson, R. O. (1975) Self absorption corrections for X-ray fluorescence analysis of aerosols. Adv.
X-RayAnal. 18:619-631.
Dzubay, T. G.; Stevens, R. K. (1975) Ambient air analysis with dichotomous sampler and X-ray fluorescence
spectrometer. Environ. Sci. Technol. 9: 663-668.
Dzubay, T. G.; Morosoff, N; Whitaker, G. L.; Yasuda, H. (1981) Evaluation of polymer films as standards for X-ray
fluorescence spectrometers. In: Electron microscopy and X-ray applications to environmental and occupational
health analysis. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.
Dzubay, T. G.; Stevens, R. K.; Balfour, W. D.; Williamson, H. J.; Cooper, J. A.; Core, J. E.; De Cesar, R. T.; Crutcher,
E. R.; Dattner, S. L.; Davis, B. L.; Heisler, S. L.; Shah, J. J.; Hopke, P. K.; Johnson, D. L. (1984)
Interlaboratory comparison of receptor model results for Houston aerosol. Atmos. Environ. 18: 1555-1566.
4-127
-------
Eatough, D. J. (1995) BOSS, the Brigham Young University Organic Sampling System: determination of particulate
carbonaceous material using diffision denuder sampling technology. In: Lane, D., ed. Gas/phase partition
measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers; in
press.
Eatough, D. J.; Sedar, B.; Lewis, L.; Hansen, L. D.; Lewis, E. A.; Farber, R. J. (1989) Determination of semivolatile
organic compounds in particles in the Grand Canyon area. Aerosol Sci. Technol. 10: 438-449.
Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993)
A multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
material in the atmosphere. Atmos. Environ. Part A 27: 1213-1219.
Edwards, J. D. (1980) A comparison of the British smoke shade method and the intergrating plate method
[Ph.D thesis]. Seattle, WA: University of Washington.
Edwards, J. D.; Ogren, J. A.; Weiss, R. E.; Charlson, R. J. (1983) Particulate air pollutants: a comparison of British
"Smoke" with optical absorption coefficient and elemental carbon concentration. Atmos. Environ. 17:
2337-2341.
Eichmann, R.; Neuling, P.; Ketseridis, G.; Hahn, J.; Jaenicke, R.; Junge, C. (1979) w-alkane studies in the troposphere
—I. Gas and particulate concentrations in North Atlantic air. Atmos. Environ. 13: 587-599.
Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ. 28:
2733-2749.
Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the United
States of America. Atmos. Environ. 28: 1009-1019.
Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1993) Comparison of independent measurements of sulfur and sulfate in the
IMPROVE network. Presented at: 86th annual meeting and exhibition of the Air & Waste Management
Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
93-RA-l 10.02.
Engelbrecht, D. R.; Cahill, T. A.; Feeney, P. J. (1980) Electrostatic effects on gravimetric analysis of membrane filters.
J. AirPollut. Control Assoc. 30: 391-392.
Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes on
size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
Technol. 14: 266-277.
Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30) 36: 8186-8201.
Federal Register. (1978) Reference method for the determination of lead in suspended particulate matter collected from
ambient air. F. R. (October 5) 43: 46258.
Feeney, P.; Cahill, T.; Olivera, J.; Guidara, R. (1984) Gravimetric determination of mass on lightly loaded membrane
filters. J. AirPollut. Control Assoc. 31: 376-378.
Perm, M. (1986) Concentration measurements and equilibrium studies of ammonium, nitrate and sulphur species in air
and precipitation [Ph.D. thesis]. Goteborg, Sweden: Goteborg University, Department of Inorganic Chemistry.
Available from: Library of Congress, Washington, DC; call no. QC879/.6/.F47/1986.
Fernandez de la Mora, J. (1986) Inertia and interception in the deposition of particles from boundary layers. Aerosol
Sci. Technol. 5: 261-266.
4-128
-------
Finlayson-Pitts, B. J.; Pitts, J. N., Jr. (1986) Atmospheric chemistry: fundamentals and experimental techniques. New
York, NY: John Wiley & Sons; pp. 783-1007.
Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
Fitz, D.; Chan, M.; Cass, G.; Lawson, D.; Ashbaugh, L. (1989) A multi-component size-classifying aerosol and gas
sampler for ambient air monitoring. Presented at: 82nd annual meeting of the Air & Waste Management
Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste Management Association; paper no. 89-140.1.
Flessel, P.; Wang, Y. Y.; Chang, K.-L; Wesolowski, J. I; Guirguis, G. N.; Kim, I.-S.; Levaggi, D.; Sm, W. (1991)
Seasonal variations and trends in concentrations of filter-collected poly cyclic aromatic hydrocarbons (PAH)
and mutagenic activity in the San Francisco Bay Area. J. Air Waste Manage. Assoc. 41: 276-281.
Foreman, W. T.; Bidleman, T. F. (1990) Semivolatile organic compounds in the ambient air of Denver, Colorado.
Atmos. Environ. Part A 24: 2405-2416.
Forrest, J.; Spandau, D. J.; Tanner, R. L.; Newman, L. (1982) Determination of atmospheric nitrate and nitric acid
employing a diffusion denuder with a filter pack. Atmos. Environ. 16: 1473-1485.
Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals of aerosol behavior. New York, NY: John Wiley &
Sons.
Fung, K. (1990) Particulate carbon speciation by MnO2 oxidation. Aerosol Sci. Technol. 12: 122-127.
Fung, K. K.; Heisler, S. L.; Price, A.; Nuesca, B. V.; Mueller, P. K. (1979) Comparison of ion chromatography and
automated wet chemical methods for analysis of sulfate and nitrate in ambient particulate filter samples. In:
Mulik, J. D.; Sawicki, E., eds. Ion chromatographic analysis of environmental pollutants: volume 2. Ann
Arbor, MI: Ann Arbor Science Publishers Inc.; pp. 203-209.
Gagosian, R. B.; Peltzer, E. T.; Zafiriou, O. C. (1981) Atmospheric transport of continentally derived lipids to the
tropical North Pacific. Nature (London) 291: 312-314.
Gebhart, J. (1993) Optical direct-reading techniques: light intensity systems. In: Willeke, K.; Baron, P. A., eds. Aerosol
measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 313-344.
Gerber, H. E. (1982) Optical techniques for the measurement of light absorption by particulates. In: Wolff, G. T.;
Klimisch, R. L., eds. Particulate carbon: atmospheric life cycle. New York, NY: Plenum Press; pp. 145-158.
Gordon, R. J. (1974) Solvent selection in extraction of airborne particulate matter. Atmos. Environ. 8: 189-191.
Griffin, J. J.; Goldberg, E. D. (1979) Morphologies and origin of elemental carbon in the environment. Science
(Washington, DC) 206: 563-565.
Grinshpun, S.; Willeke, K.; Kalatoor, S. (1993) A general equation for aerosol aspiration by thin-walled sampling
probes in calm and moving air. Atmos. Environ. Part A 27: 1459-1470.
Grosjean, D. (1975) Solvent extraction and organic carbon determination in atmospheric particulate matter: the organic
extraction-organic carbon analyzer (OE-OCA) technique. Anal. Chem. 47: 797-805.
Groves, W. A.; Hahne, R. M. A.; Levine, S. P.; Schork, M. A. (1994) A field comparison of respirable dust samplers.
Am. Ind. Hyg. Assoc. J. 55: 748-755.
Gundel, L. A.; Stevens, R. K.; Daisey, J. M.; Lee, V.; Mahanama, K. R. R. (1992) Annular denuders for sampling
semi-volatile polycyclic aromatic hydrocarbons and other organic species. Presented at: 11th annual meeting of
the American Association for Aerosol Research; October; San Francisco, CA.
4-129
-------
Gundel, L. A.; Mahanama, K. R. R.; Daisey, J. M. (1994) Fractionation of polar organic extracts of airborne participate
matter using cyanopropyl-bonded silica in solid-phase extraction. J. Chromatogr. 629: 75-82.
Hall, D. J.; Mark, D.; Upton, S. L. (1992) A new large wind tunnel for dust and aerosol studies. J. Aerosol Sci.
23(suppl. 1): S591-S594.
Hammerle, R. H.; Pierson, W. R. (1975) Sources and elemental composition of aerosol in Pasadena, Calif.,
by energy-dispersive X-ray fluorescence. Environ. Sci. Technol. 9: 1058-1068.
Hampton, C. V.; Pierson, W. R.; Schuetzle, D.; Harvey, T. M. (1983) Hydrocarbon gases emitted from vehicles on the
road. 2. Determination of emission rates from diesel and spark-ignition vehicles. Environ. Sci. Technol. 17:
699-708.
Harman, J. N., Chairman. (1989) Inductively coupled plasma emission spectroscopy. In: Lodge, J. P., Jr., ed. Methods
of air sampling and analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 89-92.
Harrison, L.; Harrison, H. (1982) A comparison of optical and aerodynamic aerosol sizes: the effects of
inhomogeneous particles. Atmos. Environ. 16: 293-300.
Harrison, R. M.; Kitto, A.-M. N. (1990) Field intercomparison of filter pack and denuder sampling methods for reactive
gaseous and particulate pollutants. Atmos. Environ. Part A 24: 2633-2640.
Hart, K. M.; Pankow, J. F. (1990) Comparison of w-alkane and PAH concentrations collected on quartz fiber and
Teflon membrane filters in an urban environment. J. Aerosol Sci. 21(suppl. 1): S377-S380.
Hearle, J. W. S.; Sparrow, J. T.; Cross, P. M. (1972) The use of the scanning electron microscope. New York, NY:
Pergamon Press.
Heitbrink, W. A.; Baron, P. A.; Willeke, K. (1991) Coincidence in time-of-flight aerosol spectrometers: phantom
particle creation. Aerosol Sci. Technol. 14: 112-126.
Henry, R. C.; Lewis, C. W.; Hopke, P. K.; Williamson, H. J. (1984) Review of receptor model fundamentals. Atmos.
Environ. 18: 1507-1515.
Hering, S. V., ed. (1989) Air sampling instruments for evaluation of atmospheric contaminants. 7th ed. Cincinnati, OH:
American Conference of Governmental Industrial Hygienists.
Hering, S. V.; Lawson, D. R.; Allegrini, I.; Febo, A.; Perrino, C.; Possanzini, M.; Sickles, J. E., II; Anlauf, K. G.;
Wiebe, A.; Appel, B. R.; John, W.; Ondo, J.; Wall, S.; Braman, R. S.; Sutton, R.; Cass, G. R.; Solomon, P. A.;
Eatough, D. J.; Eatough, N. L.; Ellis, E. C.; Grosjean, D.; Hicks, B. B.; Womack, J. D.; Horrocks, J.; Knapp, K.
T.; Ellestad, T. G.; Paur, R. I; Mitchell, W. J.; Pleasant, M.; Peake, E.; MacLean, A.; Pierson, W. R.;
Brachaczek, W.; Schiff, H. L; Mackay, G. L; Spicer, C. W.; Stedman, D. H.; Winer, A. M.; Biermann, H. W.;
Tuazon, E. C. (1988) The nitric acid shootout: field comparison of measurement methods. Atmos. Environ. 22:
1519-1539.
Henng, S. V.; Appel, B. R.; Cheng, W.; Salaymeh, F.; Cadle, S. H.; Mulawa, P. A.; Cahill, T. A.; Eldred, R. A.;
Surovik, M.; Fitz, D.; Howes, J. E.; Knapp, K. T.; Stockburger, L.; Turpin, B. J.; Huntzicker, J. J.; Zhang,
X.-Q.; McMurry, P. H. (1990) Comparison of sampling methods for carbonaceous aerosols in ambient air.
Aerosol Sci. Technol. 12: 200-213.
Hering, S. V.; Roberts, P. T.; Taylor, C. A. (1994) Particle measurements for the Children's Health Study: development
of a two-week sampler [abstract]. Presented at: Sixth conference of the International Society for Environmental
Epidemiology/fourth conference of the International Society for Exposure Analysis; September; Research
Triangle Park, NC. Chapel Hill, NC: University of North Carolina, School of Public Health; p. 260.
4-130
-------
Hildemann, L. M.; Markowski, G. R.; Cass, G. R. (1991) Chemical composition of emissions from urban sources of
fine organic aerosol. Environ. Sci. Technol. 25: 744-759.
Hillamo, R. E.; Kauppinen, E. I. (1991) On the performance of the Bemer low pressure impactor. Aerosol Sci. Technol.
14: 33-47.
Hinds, W. C. (1982) Aerosol technology. New York, NY: John Wiley and Sons.
Hinds, W. C.; Kraske, G. (1986) Performance of PMS model LAS-X optical particle counter. J. Aerosol Sci. 17: 67-72.
Hirose, K.; Sugimura, Y. (1984) Excess 228Th in the airborne dust: an indicator of continental dust from the east Asian
deserts. Earth Planet. Sci. Lett. 70: 110-114.
Hoffman, A. J.; Purdue, L. J.; Rehme, K. A.; Holland, D. M. (1988) 1987 PM10 sampler intercomparison study. In:
Mathai, C. V.; Stonefield, D. H., eds. PM-10: implementation of standards, an APCA/EPA international
specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp.
138-149. (APCA transactions series TR-13).
Hollander, W. (1990) Proposed performance criteria for samplers of total suspended particulate matter.
Atmos. Environ. Part A 24: 173-177.
Hollander, W. (1992) Ambient aerosol sampling without pumps. J. Aerosol Sci. 23(suppl. 1): S619-S622.
Hopke, P. K. (1985) Receptor modeling in environmental chemistry. New York, NY: John Wiley & Sons.
Hopke, P. K.; Casuccio, G. S. (1991) Scanning electron microscopy. In: Hopke, P. K., ed. Receptor modeling for air
quality management: v. 7. Amsterdam, The Netherlands: Elsevier; pp. 149-212.
Hornung, R. W.; Reed, L. D. (1990) Estimation of average concentration in the presence of nondetectable values. Appl.
Occup. Environ. Hyg. 5: 46-51.
Horvath, H.; Kaller, W. (1994) Calibration of integrating nephelometers in the post-halocarbon era. Atmos. Environ.
28: 1219-1223.
Huntzicker, J. J.; Johnson, R. L.; Shah, J. J.; Gary, R. A. (1982) Analysis of organic and elemental carbon in ambient
aerosols by a thermal-optical method. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon - atmospheric
life cycle: proceedings of an international symposium; October 1980; Warren, MI. New York, NY: Plenum
Press; pp. 79-88.
Ingham, D. B.; Yan, B. (1994) The effect of a cylindrical backing body on the sampling efficiency of a cylindrical
sampler. J. Aerosol Sci. 25: 535-541.
Ingram, W. T. (1969) Smoke curve calibration. New York, NY: New York University, School of Engineering and
Science; PHS contract no. PH-86-68-66.
Ingram, W. T.; Golden, J. (1973) Smoke curve calibration. J. Air Pollut. Control Assoc. 23: 110-115.
International Standards Organization (ISO). (1983) Air quality: particle size fraction definitions for health-related
sampling. Geneva, Switzerland: International Standards Organization; technical report ISO/TR7708-1983.
International Standards Organization (ISO). (1992) Air quality—particle size fraction definitions for health-related
sampling (draft revision of ISO/TR 7708:1983). Geneva, Switzerland: International Organization for
Standardization; draft international standard no. ISO/DIS 7708.
4-131
-------
International Standards Organization (ISO). (1993) Air quality—particle size fraction definitions for health-related
sampling (revision of ISO/TR 7708:1983). Geneva, Switzerland: International Organization for
Standardization; approved international standard no. ISO/DIS 7708.
Jaklevic, J. M.; Loo, B. W.; Goulding, F. S. (1977) Photon-induced X-ray fluorescence analysis using
energy-dispersive detector and dichotomous sampler. In: Dzubay, T. G., ed. X-ray fluorescence analysis of
environmental samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 3-18.
Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.; Loo, B. W. (1981) A p-gauge method applied to aerosol samples. Environ.
Sci. Technol. 15: 680-686.
Janocko, P. B.; Casuccio, G. S.; Dattner, S. L.; Johnson, D. L.; Crutcher, E. R. (1982) The El Paso airshed; source
apportionment using complementary analyses and receptor models. In: A specially conference on: receptor
models applied to contemporary pollution problems; October; Danvers, MA. Pittsburgh, PA: Air Pollution
Control Association; pp. 249-265. (SP-48).
Japar, S. M.; Szkarlat, A. C.; Gorse, R. A., Jr.; Heyerdahl, E. K.; Johnson, R. L.; Rau, J. A.; Huntzicker, J. J. (1984)
Comparison of solvent extraction and thermal-optical carbon analysis methods: application to diesel vehicle
exhaust aerosol. Environ. Sci. Technol. 18: 231-234.
Javitz, H. S.; Watson, J. G.; Robinson, N. (1988) Performance of the chemical mass balance model with simulated
local-scale aerosols. Atmos. Environ. 22: 2309-2322.
Jensen, P. A.; O'Brien, D. (1993) Industrial hygiene. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement:
principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 537-559.
John, W.; Wall, S. M. (1983) Aerosol testing techniques for size-selective samplers. J. Aerosol Sci. 14: 713-727.
John, W.; Wang, H.-C. (1991) Laboratory testing method for PM-10 samplers: lowered effectiveness from particle
loading. Aerosol Sci. Technol. 14: 93-101.
John, W.; Hering, S.; Reischl, G.; Sasaki, G.; Goren, S. (1983) Characteristics of Nuclepore filters with large pore size
—II. filtration properties. Atmos. Environ. 17: 373-382.
John, W.; Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
dichotomous sampler. Atmos. Environ. 22: 1627-1635.
John, W.; Wall, S. M.; Ondo, J. L.; Winklmayr, W. (1990) Modes in the size distributions of atmospheric inorganic
aerosol. Atmos. Environ. Part A 24: 2349-2359.
John, W.; Winklmayr, W.; Wang, H.-C. (1991) Particle deagglomeration and reentrainment in a PM-10 sampler.
Aerosol Sci. Technol. 14: 165-176.
Johnson, R. L.; Huntzicker, J. J. (1979) Analysis of volatilizable and elemental carbon in ambient aerosols.
In: Novakov, T., ed. Proceedings: carbonaceous particles in the atmosphere; March 1978; Berkeley, CA.
Berkeley, CA: University of California, Lawrence Berkeley Laboratory; pp. 10-13; report no. LBL-9037.
Johnson, D. L.; Mclntyre, B. L.; Fortmann, R.; Stevens, R. K.; Hanna, R. B. (1981a) A chemical element comparison of
individual particle analysis and bulk analysis methods. Scanning Electron Microsc. 1: 469-476.
Johnson, R. L.; Shah, J. J.; Gary, R. A.; Huntzicker, J. J. (1981b) An automated thermal-optical method for the analysis
of carbonaceous aerosol. In: Macias, E. S.; Hopke, P. K., eds. Atmospheric aerosol: source/air quality
relationships; based on a symposium jointly sponsored by the Divisions of Nuclear Chemistry and Technology
at the 180th national meeting of the American Chemical Society; August 1980; Las Vegas, NV. Washington,
DC: American Chemical Society; pp. 223-233. (Comstock, M. J., ed. ACS symposium series: 167).
4-132
-------
Junge, C. E. (1977) Basic considerations about trace constituents in the atmosphere as related to the fate of global
pollutants. In: Suffet, I. H., ed. Fate of pollutants in the air and water environments: part I, mechanism of
interaction between environments and mathematical modeling and the physical fate of pollutants, papers from
the 165th national American Chemical Society meeting; April 1975; Philadelphia, PA. New York, NY: John
Wiley & Sons; pp. 7-25. (Advances in environmental science and technology: v. 8).
Kado, N. Y.; Kozmicky, P. A.; Hsieh, D. P. H.; Gundel, L. A.; Daisey, J. M; Mahanama, K. R. R.; Schuetzle, D. (1991)
Mutagenic activity of the polar organic matter and its fractions from airborne particles. Presented at: AAAR
meeting; October; Traverse City, MI.
Kalthoff, K.; Grumpier, J. (1990) PM10 methods comparison test report [unpublished material]. Sacramento, CA:
California Air Resources Board, Instrumentation and Operations Support Section.
Kanbour, F. I.; Altai, F. A.; Yassin, S.; Harrison, R. M.; Kitto, A.-M. N. (1990) A comparison of smoke shade and
gravimetric determination of suspended particulate matter in a semi-arid climate (Baghdad, Iraq). Atmos.
Environ. Part A 24: 1297-1301.
Kaupp, H.; Umlauf, G. (1992) Atmospheric gas-particle partitioning of organic compounds: comparison of sampling
methods. Atmos. Environ. Part A 26: 2259-2267.
Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity on
dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
Kenny, L. C. (1992) Report of progress towards the development of a performance standard for aerosol sampling
instruments used in occupational hygiene. J. Aerosol Sci. 23: 773-779.
Kenny, L. C.; Liden, G. (1989) The application of performance standards to personal airborne dust samplers. Ann.
Occup. Hyg. 33:289-300.
Kenny, L. C.; Liden, G. (1991) A technique for assessing size-selective dust samplers using the APS and polydisperse
test aerosols. J. Aerosol Sci. 22: 91-100.
Kim, D.; Hopke, P. K. (1988) Classification of individual particles based on computer-controlled scanning electron
microscopy data. Aerosol Sci. Technol. 9: 133-151.
Kim, D. S.; Hopke, P. K.; Massart, D. L.; Kaufman, L.; Casuccio, G. S. (1987) Multivariate analysis of CCSEM auto
emission data. Sci. Total Environ. 59: 141-155.
Klippel, W.; Warneck, P. (1980) The formaldehyde content of the atmospheric aerosol. Atmos. Environ. 14: 809-818.
Konig, J.; Funcke, W.; Balfanz, E.; Grosch, B.; Pott, F. (1980) Testing a high volume air sampler for quantitative
collection of poly cyclic aromatic hydrocarbons. Atmos. Environ. 14: 609-613.
Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988) Evaluation
of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci. Technol. 22:
1463-1468.
Koutrakis, P.; Wolfson, J. M.; Brauer, M.; Spengler, J. D. (1990) Design of a glass impactor for an annular
denuder/filter pack system. Aerosol Sci. Technol. 12: 607-612.
Kukreja, V. P.; Bove, J. L. (1976) An enrichment method for polycyclic aromatic hydrocarbons (PAH's) collected on
glass fiber filters using hydrofluoric acid. J. Environ. Sci. Health Part All: 517-524.
Lamb, S. I.; Petrowski, C.; Kaplan, I. R.; Simoneit, B. R. T. (1980) Organic compounds in urban atmospheres: a review
of distribution, collection and analysis. J. Air Pollut. Control Assoc. 30:1098-1115.
4-133
-------
Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and evaluation of
a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air. Environ. Sci.
Technol. 22: 941-947.
Lane, D. A.; Johnson, N. D.; Hanley, M.-J. J.; Schroeder, W. H.; Ord, D. T. (1992) Gas- and particle-phase
concentrations of a-hexachlorocyclohexane, y-hexachlorocyclohexane, and hexachlorobenzene in Ontario air.
Environ. Sci. Technol. 26: 126-133.
Larson, T. V.; Yuen, P. F.; Maykut, N. N. (1992) Weekly composite sampling of PM10 for total mass and trace
elemental analysis. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source
controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA:
Air & Waste Management Association; pp. 39-50. (A&WMA transactions series no. 22).
Laskus, L. (1983) Comparative measurements of suspended particulates using several gravimetric methods and the
British smoke sampler. Sci. Total Environ. 31: 23-40.
Lee, R. J.; Fisher, R. M. (1980) Quantitative characterization of particulates by scanning and high voltage electron
microscopy. Washington, DC: National Bureau of Standards; special publication.
Lee, R. J.; Kelly, J. F. (1980) Overview of SEM-based automated image analysis. Scanning Electron Microsc. 1:
303-310.
Lee, R. E., Jr.; Caldwell, J. S.; Morgan, G. B. (1972) The evaluation of methods formeasuring suspended particulates
in air. Atmos. Environ. 6: 593-622.
Lee, R. J.; Fasiska, E. J.; Janocko, P.; McFarland, D.; Penkala, S. (1979) Electron-beam particulate analysis. Ind. Res.
Dev. (June): 105-108.
Lee, R. J.; Spitzig, W. A.; Kelly, J. F.; Fisher, R. M. (1981) Quantitative metallography by computer-controlled
scanning electron microscopy. J. Met. 33(3): 20-25.
Lewis, C. W. (1981) On the proportionality of fine mass concentration and extinction coefficient for bimodal size
distributions. Atmos. Environ. 15: 2639-2646.
Li, C. K.; Kamens, R. M. (1993) The use of poly cyclic aromatic hydrocarbons as source signatures in receptor
modeling. Atmos. Environ. Part A 27: 523-532.
Liden, G.; Kenny, L. C. (1991) Comparison of measured respirable dust sampler penetration curves with sampling
conventions. Ann. Occup. Hyg. 35: 485-504.
Liden, G.; Kenny, L. C. (1992) The performance of respirable dust samplers: sampler bias, precision and inaccuracy.
Ann. Occup. Hyg. 36: 1-22.
Ligocki, M. P.; Pankow, J. F. (1989) Measurements of the gas/particle distributions of atmospheric organic compounds.
Environ. Sci. Technol. 23: 75-83.
Lin, C.; Baker, M.; Charlson, R. J. (1973) Absorption coefficient of atmospheric aerosol: a method for measurement.
Appl. Opt. 12: 1356-1363.
Lindskog, A.; Brorstrom-Lunden, E.; Sjodin, A. (1985) Transformation of reactive PAH on particles by exposure to
oxidized nitrogen compounds and ozone. Environ. Int. 11: 125-130.
Lioy, P. J.; Wainman, T.; Turner, W.; Marple, V. A. (1988) An intercomparison of the indoor air sampling impactor
and the dichotomous sampler for a 10-//m cut size. JAPCA 38: 668-670.
4-134
-------
Lipfert, F. W. (1994) Filter artifacts associated with particulate measurements: recent evidence and effects on statistical
relationships. Atmos. Environ. 28: 3233-3249.
Lippmann, M. (1993) Environmental toxicology and exposure limits for ambient air. Appl. Occup. Environ. Hyg. 8:
847-858.
Lippmann, M.; Chan, T. L. (1979) Cyclone sampler performance. Staub Reinhalt. Luft 39: 7-11.
Liu, B. Y. H., ed. (1976) Fine particles: aerosol generation, measurement, sampling, and analysis. New York, NY:
Academic Press, Inc.
Liu, B. Y. H.; Pui, D. Y. H. (1981) Aerosol sampling inlets and inhalable particles. Atmos. Environ. 15: 589-600.
Liu, B. Y. H.; Pui, D. Y. H.; Rubow, K. L.; Kuhlmey, G. A. (1978) Progress report: research on air sampling filter
media. Minneapolis, MN: University of Minnesota, Mechanical Engineering Department, Particle Technology
Laboratory; publication no. 366.
Lodge, J. P., Jr.; Waggoner, A. P.; Klodt, D. T.; Grain, C. N. (1981) Non-health effects of airborne particulate matter.
Atmos. Environ. 15: 431-482.
Loo, B. W.; Cork, C. P. (1988) Development of high efficiency virtual impactors. Aerosol Sci. Technol. 9: 167-176.
Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air
pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
Lundgren, D. A.; Lippmann, M.; Harris, F. S., Jr.; Clark, W. E.; Marlow, W. H.; Durham, M. D., eds. (1979) Aerosol
measurement: [papers from a workshop]; March 1976; Gainesville, FL. Gainesville, FL: University Presses of
Florida.
Lynch, A. J.; McQuaker, N. R.; Brown, D. F. (1980) ICP/AES analysis and the composition of airborne and soil
materials in the vicinity of a lead/zinc smelter complex. J. Air Pollut. Control Assoc. 30: 257-260.
Macher, J.; Chatigny, M. A.; Burge, H. A. (1995) Sampling airborne microorganisms and aeroallergens. In: Air
sampling instruments. 8th ed. Cincinnati, OH: American Conference of Governmental Industrial Hygienists.
Macias, E. S.; Blumenthal, D. L.; Anderson, J. A.; Cantrell, B. K. (1979) Characterization of visibility-reducing
aerosols in the Southwest. Project VISTTA. Progress report no. 1 (final report). Research Triangle Park, NC:
U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory; EPA report no.
EPA-600/7-79-243. Available from: NTIS, Springfield, VA; PB80-139041.
Mage, D. T. (1995) The relationship between total suspended particulate matter (TSP) and British Smoke
measurements in London: development of a simple model. J. Air Waste Manage. Assoc. 45: 737-739.
Malissa, H. (1979) Some analytical approaches to the chemical characterization of carbonaceous particulates.
In: Novakov, T., ed. Proceedings: carbonaceous particles in the atmosphere; March 1978; Berkeley, CA.
Berkeley, CA: University of California, Lawrence Berkeley Laboratory; pp. 3-9; report no. LBL-9037.
Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
Mamane, Y. (1988) Estimate of municipal refuse incinerator contribution to Philadelphia aerosol—I. source analysis.
Atmos. Environ. 22: 2411-2418.
4-135
-------
Mark, D.; Vincent, J. H. (1986) A new personal sampler for airborne total dust in workplaces. Ann. Occup. Hyg. 30:
89-102.
Mark, D.; Lyons, C. P.; Upton, S. L.; Hall, D. J. (1992) A review of the rationale of current methods for determining
the performance of aerosol samplers. J. Aerosol Sci. 23(suppl. 1): S611-S614.
Markowski, G. R. (1987) On identifying and correcting for reentrainment in cascade impactor measurements. Aerosol
Sci. Technol. 7: 143-159.
Marple, V. A. (1978) Simulation of respirable penetration characteristics by inertial impaction. J. Aerosol Sci.
9: 125-134.
Marple, V. A.; Olson, B. A. (1995) A high volume PMj/PM^/PMj 0 trichotomous sampler. In: Particulate matter:
health and regulatory issues: proceedings of an international specialty conference; April; Pittsburgh, PA.
Pittsburgh, PA: Air & Waste Management Association; pp. 237-261. (A&WMA publication VIP-49).
Marple, V. A.; Rubow, K. L. (1976) Aerodynamic particle size calibration of optical particle counters. J. Aerosol Sci. 7:
425-433.
Marple, V. A.; Rubow, K. L.; Turner, W.; Spengler, J. D. (1987) Low flow rate sharp cut impactors for indoor air
sampling: design and calibration. JAPCA 37: 1303-1307.
Marple, V. A.; Liu, B. Y. H.; Olson, B. A. (1989) Evaluation of a cleanroom concentrating aerosol sampler.
In: Proceedings of the 35th annual meeting of the Institute of Environmental Sciences. Anaheim, CA: Institute
of Environmental Sciences; pp. 360-363.
Marple, V. A.; Liu, B. Y. H.; Burton, R. M. (1990) A high volume dichotomous sampler: design and evaluation.
Presented at: 83rd annual meeting & exhibition of the Air & Waste Management Association; June; Pittsburgh,
PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-84.2.
Marple, V. A.; Rubow, K. L.; Behm, S. M. (1991) A microorifice uniform deposit impactor (MOUDI): description,
calibration, and use. Aerosol Sci. Technol. 14: 434-446.
Marple, V. A.; Rubow, K. L.; Olson, B. A. (1993) Inertial, gravitational, centrifugal, and thermal collection techniques.
In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles, techniques, and applications. New York,
NY: Van Nostrand Remhold; pp. 206-232.
Masia, P.; Di Palo, V.; Possanzini, M. (1994) Uptake of ammonia by nylon filters in filter pack systems. Atmos.
Environ. 28: 365-366.
Massart, D. L.; Kaufman, L. (1983) The interpretation of analytical chemical data by the use of cluster analysis. New
York, NY: John Wiley & Sons (Elving, P. J.; Winefordner, J. D., eds. Chemical analysis: a series of
monographs on analytical chemistry and its applications: v. 65).
Maynard, A. D. (1993) Respirable dust sampler characterisation: efficiency curve reproducibility. J. Aerosol Sci.
24(suppl. 1): S457-S458.
Mazurek, M. A.; Simoneit, B. R. T.; Cass, G. R.; Gray, H. A. (1987) Quantitative high-resolution gas chromatography
and high-resolution gas chromatography/mass spectrometry analyses of carbonaceous fine aerosol particles.
Int. J. Environ. Anal. Chem. 29: 119-139.
Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1989) Interpretation of high-resolution gas chromatography/mass
spectrometry data aquired from atmospheric organic aerosol samples. Aerosol Sci. Technol. 10: 408-420.
4-136
-------
Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Biological input to visibility-reducing aerosol particles in the
remote arid southwestern United States. Environ. Sci. Technol. 25: 684-694.
McDow, S. R.; Huntzicker, J. J. (1990) Vapor adsorption artifact in the sampling of organic aerosol: face velocity
effects. Atmos. Environ. Part A 24: 2563-2571.
McDow, S. R.; Huntzicker, J. J. (1993) Vapor adsorption artifact in the sampling of organic aerosol. In: Winegar, E. D.;
Keith, L. H., eds. Sampling and analysis of airborne pollutants. Boca Raton, FL: Lewis Publishers, Inc.; pp.
191-208.
McFarland, A. R.; Ortiz, C. A. (1979) Aerosol characterization of ambient particulate samplers used in environmental
monitoring studies. College Station, TX: Texas A&M Research Foundation; contract no. 68-02-2720.
McFarland, A. R.; Ortiz, C. A. (1984) Characterization of Sierra-Andersen model 321a ,um size selective inlet for hi-vol
samplers. College Station, TX: Texas A&M University, Department of Civil Engineering, Air Quality
Laboratory; report no. 4716/01/02/84/ARM.
McFarland, A. R.; Wedding, J. B.; Cermak, J. E. (1977) Wind tunnel evaluation of a modified Andersen impactor and
an all weather sampler inlet. Atmos. Environ. 11: 535-539.
McFarland, A. R.; Ortiz, C. A.; Rodes, C. E. (1979) Characteristics of aerosol samplers used in ambient air monitoring.
Presented at: 86th national meeting of the American Institute of Chemical Engineers; April; Houston, TX.
College Station, TX: Texas A&M University.
McFarland, A. R.; Ortiz, C. A.; Rodes, C. E. (1982) Wind tunnel evaluation of the British smoke shade sampler. Atmos.
Environ. 16: 325-328.
McFarland, A. R.; Ortiz, C. A.; Bertch, R. W., Jr. (1984) A 10 ,um cutpoint size selective inlet for hi-vol samplers. J. Air
Pollut. Control Assoc. 34: 544-547.
Mclntyre, B. L.; Johnson, D. L. (1982) A particle class balance receptor model for aerosol apportionment in Syracuse,
NY. In: Dattner, S. L.; Hopke, P. K., eds. A specially conference on receptor models applied to contemporary
pollution problems; October; Danvers, MA. Pittsburgh, PA: Air Pollution Control Association. (SP-48).
McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
aerosols. Atmos. Environ. 23: 497-507.
McMurry, P. H.; Takano, H.; Anderson, G. R. (1983) Study of the ammonia (gas)-sulfuric acid (aerosol) reaction rate.
Environ. Sci. Technol. 17: 347-352.
Merrifield, T. (1989) The measurement of PM-10 by p-absorption. Presented at: 82nd annual meeting and exhibition of
the Air & Waste Management Association; June; Anaheim, CA. Pittsburgh, PA: Air & Waste Management
Association; paper no. 89-66A.2.
Meyer, M. B.; Lijek, J.; Ono, D. (1992) Continuous PM10 measurements in a woodsmoke environment. In: Chow, J. C.;
Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international
specially conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association; pp.
24-38. (A&WMA transactions series no. 22).
Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
particulate matter. In: Particulate matter: health and regulatory issues: proceedings of an international specialty
conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; pp. 204-210.
(A&WMA publication VIP-49).
Meyer-Arendt, J. R. (1972) Introduction to classical and modern optics. Englewood Cliffs, NJ: Prentice Hall, Inc.
4-137
-------
Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size considerations
for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc. 29: 610-615.
Mueller, P. K.; Mosley, R. W.; Pierce, L. B. (1971) Carbonate and noncarbonate carbon in atmospheric particles =
carbon, carbonate et non carbonate, dans les particules atmospherique. In: Englund, H. M.; Beery, W. T., eds.
Proceedings of the second international clean air congress; December 1970; Washington, DC. New York, NY:
Academic Press, Inc.; pp. 532-539.
Muir, D.; Laxen, D. P. H. (1995) Black smoke as a surrogate for PM10 in health studies? Atmos. Environ. 29: 959-962.
Mulik, J.; Puckett, R.; Williams, D.; Sawicki, E. (1976) Ion chromatographic analysis of sulfate and nitrate in ambient
aerosols. Anal. Lett. 9: 653-663.
Mylonas, D. T.; Allen, D. T.; Ehrman, S. H.; Pratsinis, S. E. (1991) The sources and size distributions of organonitrates
in Los Angeles aerosol. Atmos. Environ. Part A 25: 2855-2861.
Nevalainen, A.; Pastuszka, J.; Liebhaber, F.; Willeke, K. (1992) Performance of bioaerosol samplers: collection
characteristics and sampler design considerations. Atmos. Environ. Part A 26: 531-540.
Nevalainen, A.; Willeke, K.; Liebhaber, F.; Pastuszka, J.; Burge, H.; Henningson, E. (1993) Bioaerosol sampling. In:
Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles, techniques, and applications. New York, NY:
Van Nostrand Reinhold; pp. 471-492.
Novakov, T. (1982) Soot in the atmosphere. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric life
cycle. New York, NY: Plenum Press; pp. 19-41.
Novick, V. J.; Alvarez, J. L. (1987) Design of a multistage virtual impactor. Aerosol Sci. Technol. 6: 63-70.
Ogden, T. L. (1992) Inhalable and respirable dust - moving from aerosol science to legislation. J. Aerosol Sci.
23(suppl. 1): S445-S448.
Ogden, T. L.; Birkett, J. L. (1977) The human head as a dust sampler. In: Walton, W. H., ed. Inhaled particles IV, part
1: proceedings of an international symposium; September 1975; Edinburgh, Scotland. Oxford, United
Kingdom: Pergamon Press; pp. 93-105.
Olmez, I. (1989) Trace element signatures in groundwater pollution. In: Watson, J. G., ed. Receptor models in air
resources management: an international specialty conference; February 1988; San Francisco, CA. Pittsburgh,
PA: Air & Waste Management Association; pp. 3-11. (APCA transactions series: no. 14).
Olmez, L; Harman, J. N. (1989) Instrumental neutron activation analysis of atmospheric particulate matter. In: Lodge, J.
P., Jr., ed. Methods of air sampling and analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 143-150.
Ondov, J. M.; Divita, F., Jr. (1993) Size spectra for trace elements in urban aerosol particles by instrumental neutron
activation analysis. J. Radioanal. Nucl. Chem. 167: 247-258.
Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air. Atmos.
Environ. Part A 26: 2149-2162.
Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Spengler, J. (1993) The particle TEAM (PTEAM) study: analysis of the
data, volume III [draft final report]. Research Triangle Park, NC: U.S. Environmental Protection Agency; EPA
contract no. 68-02-4544.
Palen, E. J.; Allen, D. T.; Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1992) Fourier transform infrared
analysis of aerosol formed in the photo-oxidation of isoprene and p-pinene. Atmos. Environ. Part A 26:
1239-1251.
4-138
-------
Palen, E. I; Allen, D. T.; Pandis, S. N.; Paulson, S.; Seinfeld, J. H.; Flagan, R. C. (1993) Fourier Transform Infrared
analysis of aerosol formed in the photo-oxidation of 1-octene. Atmos. Environ. Part A 27: 1471-1477.
Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
particulate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
Pankow, J. F. (1992) Application of common y-intercept regression parameters for log Kp vs l/T for predicting
gas-particle partitioning in the urban environment. Atmos. Environ. Part A 26: 2489-2497.
Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ. Part A 24:
2695-2698.
Pashel, G.; Egner, D. (1981) 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. 15:
919-927.
Patashnick, H.; Rupprecht, E. G. (1990) Continuous PM-10 measurements using the tapered element oscillating
microbalance. Presented at: 83rd annual meeting and exhibition of the Air & Waste Management Association;
June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-84.1.
Patashnick, H.; Rupprecht, E. G. (1991) Continuous PM-10 measurements using the tapered element oscillating
microbalance. J. Air Waste Manage. Assoc. 41: 1079-1083.
Pellizzari, E.; Lioy, P.; Quackenboss, J.; Whitmore, R.; Clayton, A.; Freeman, N.; Waldman, J.; Thomas, K.; Rodes, C.;
Wilcosky, T. (1995) Population-based exposure measurements in EPA Region 5: Phase I field study in support
of the National Human Exposure Assessment Survey. J. Exposure Anal. Environ. Epidemiol. 5: 327-358.
Phillips, B.; Lumpkin, T.; Pleasant, M. (1994) Philadelphia diesel particulate matter monitoring study. In: Measurement
of toxic and related air pollutants: proceedings of the U.S. EPA/A&WMA international symposium; May;
Durham, NC. Pittsburgh, PA: Air Waste Manage. Assoc.; pp. 640-645. (A&WMA specialty conference
VIP-39).
Pierson, W. R.; Hammerle, R. H.; Brachaczek, W. W. (1976) Sulfate formed by interaction of sulfur dioxide with filters
and aerosol deposits. Anal. Chem. 48: 1808-1811.
Pierson, W. R.; Brachaczek, W. W.; Japar, S. M.; Cass, G. R.; Solomon, P. A. (1988) Dry deposition and dew
chemistry in Claremont, California, during the 1985 Nitrogen Species Methods Comparison Study. Atmos.
Environ. 22: 1657-1663.
Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Norbeck, J. M.; Keeler, G. J. (1989) Atmosphenc
acidity measurements on Allegheny Mountain and the origins of ambient acidity in the northeastern United
States. Atmos. Environ. 23: 431-459.
Possanzini, M.; Febo, A.; Liberti, A. (1983) New design of a high-performance denuder for the sampling of
atmospheric pollutants. Atmos. Environ. 17: 2605-2610.
Pupp, C.; Lao, R. C.; Murray, J. J.; Pottie, R. F. (1974) Equilibrium vapour concentrations of some polycyclic aromatic
hydrocarbons, As4O6 and SeO2 and the collection efficiencies of these air pollutants. Atmos. Environ. 8:
915-925.
Purdue, L. J.; Rodes, C. E.; Rehme, K. A.; Holland, D. M.; Bond, A. E. (1986) Intercomparison of high-volume PM10
samplers at a site with high particulate concentrations. J. Air Pollut. Control Assoc. 36: 917-920.
Qian, Y.; Willeke, K.; Ulevicius, V.; Grinshpun, S. A.; Donnelly, J. (1995) Dynamic size spectrometry of airborne
microorganisms: laboratory evaluation and calibration. Atmos. Environ. 29: 1123-1129.
4-139
-------
Raabe, O. G.; Braaten, D. A.; Axelbaum, R. L.; league, S. V.; Cahill, T. A. (1988) Calibration studies of the drum
impactor. J. Aerosol Sci. 19: 183-195.
Rader, D. J.; O'Hern, T. J. (1993) Optical direct-reading techniques: in situ sensing. In: Willeke, K.; Baron, P. A., eds.
Aerosol measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp.
345-380.
Ramdahl, T.; Zielinska, B.; Arey, J.; Atkinson, R.; Winer, A. M.; Pitts, J. N, Jr. (1986) Ubiquitous occurrence of
2-nitrofluoranthene and 2-nitropyrene in air. Nature (London) 321: 425-427.
Ranade, M. B. (Arun); Woods, M. C.; Chen, F.-L.; Purdue, L. I; Rehme, K. A. (1990) Wind tunnel evaluation of PM10
samplers. Aerosol Sci. Technol. 13: 54-71.
Regan, G. F.; Goranson, S. K.; Larson, L. L. (1979) Use of tape samplers as fine particulate monitors. J. Air Pollut.
Control Assoc. 29: 1158-1160.
Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
Rodes, C. E.; Holland, D. M.; Purdue, L. J.; Rehme, K. A. (1985) A field comparison of PM10 inlets at four locations. J.
Air Pollut. Control Assoc. 35: 345-354.
Rodes, C. E.; Kamens, R. M.; Wiener, R. W. (1991) The significance and characteristics of the personal activity cloud
on exposure assessment measurements for indoor contaminants. Indoor Air 2: 123-145.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci. Technol.
27:636-651.
Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993b) Quantification of urban
organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part A
27: 1309-1330.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
Sci. Technol. 27: 1892-1904.
Rondia, D. (1965) Sur la volatilite des hydrocarbures poly cy cliques [Volatility of poly cyclic hydrocarbons]. Int. J. Air
Water Pollut. 9: 113-121.
Rood, M. J.; Covert, D. S.; Larson, T. V (1987) Hygroscopic properties of atmospheric aerosol in Riverside,
California. Tellus Ser. B 39B: 383-397.
Rubow, K. L.; Marple, V. A.; Olin, J.; McCawley, M.A. (1987) A personal cascade impactor: design, evaluation and
calibration. Am. Ind. Hyg. Assoc. J. 48: 532-538.
Rupprecht, E.; Meyer, M.; Patashnick, H. (1992) The tapered element oscillating microbalance as a tool for measuring
ambient particulate concentrations in real time. J. Aerosol Sci. 23(suppl. 1): S635-S638.
Saltzman, B. E. (1984) Generalized performance characteristics of miniature cyclones for atmospheric particulate
sampling. Am. Ind. Hyg. Assoc. J. 45: 671-680.
Sass-Kortsak, A. M.; O'Brien, C. R.; Bozek, P. R.; Purdham, J. T. (1993) Comparison of the 10-mm nylon cyclone,
horizontal elutriator, and aluminum cyclone for silica and wood dust measurements. Appl. Occup. Environ.
Hyg. 8:31-37.
4-140
-------
Saucy, D. A.; Anderson, J. R.; Buseck, P. R. (1987) Cluster analysis applied to atmospheric aerosol samples from the
Norwegian Arctic. Atmos. Environ. 21: 1649-1657.
Schamber, F. H. (1993) SEM for nontraditional users. Am. Lab. 25(9): 34, 36-38.
Schipper, L. B., Ill; Chow, J. C.; Frazier, C. A. (1993) Particulate air toxic emission estimation of the PM10 fraction in
natural aggregate processing facilities. Presented at: 86th annual meeting and exhibition of the Air & Waste
Management Association; June; Denver, CO. Pittsburgh, PA: Air Waste Manage. Assoc.; paper no.
93-MP-6.03.
Schuetzle, D.; Lewtas, J. (1986) Bioassay-directed chemical analysis in environmental research. Anal. Chem.
58:1060A-1076A.
Schwartz, G. P.; Daisey, J. M.; Lioy, P. J. (1981) Effect of sampling duration on the concentration of particulate
organics collected on glass fiber filters. Am. Ind. Hyg. Assoc. J. 42: 258-263.
Shaw, R. W., Jr.; Stevens, R. K.; Bowermaster, J.; Tesch, J. W.; Tew, E. (1982) Measurements of atmospheric nitrate
and nitric acid: the denuder difference experiment. Atmos. Environ. 16: 845-853.
Shimp, D. R. (1988) Field comparison of beta attenuation PM-10 sampler and high-volume PM-10 sampler. In: Mathai,
C. V.; Stonefield, D. H., eds. PM10: implementation of standards, an APCA/EPA international specialty
conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 171-178.
(APCA transactions series TR-13).
Sickles, J. E., II; Hodson, L. L. (1989) Fate of nitrous acid on selected collection surfaces. Atmos. Environ.
23: 2321-2324.
Sickles, J. E., II; Grohse, P. M.; Hodson, L. L.; Salmons, C. A.; Cox, K. W.; Turner, A. R.; Estes, E. D. (1990)
Development of a method for the sampling and analysis of sulfur dioxide and nitrogen dioxide from ambient
air. Anal. Chem. 62: 338-346.
Sides, J. O.; Saiger, H. F. (1976) Effects of prolonged static exposures of filters to ambient air on high volume
sampling results. Topeka, KS: Department of Health and Environment.
Simoneit, B. R. T. (1984) Organic matter of the troposphere—III. characterization and sources of petroleum and
pyrogenic residues in aerosols over the western United States. Atmos. Environ. 18: 51-67.
Simoneit, B. R. T.; Sheng, G.; Chen, X.; Fu, J.; Zhang, J.; Xu, Y. (1991) Molecular marker study of extractable organic
matter in aerosols from urban areas of China. Atmos. Environ. Part A 25: 2111-2129.
Sioutas, C.; Koutrakis, P.; Olson, B. A. (1994) Development and evaluation of a low cutpoint virtual impactor. Aerosol
Sci. Technol. 21:223-235.
Sloane, C. S. (1986) Effect of composition on aerosol light scattering efficiencies. Atmos. Environ. 20: 1025-1037.
Small, H.; Stevens, T. S.; Bauman, W. C. (1975) Novel ion exchange chromatographic method using conductimetric
detection. Anal. Chem. 47: 1801-1809.
Smith, R. G.; Bryan, R. J.; Feldstein, M.; Locke, D. C.; Warner, P. O. (1975) Tentative method for the determination of
carbon monoxide (detector tube method). Health Lab. Sci. 12: 171-172.
Soderholm, S. C. (1989) Proposed international conventions for particle size-selective sampling. Ann. Occup. Hyg. 33:
301-320.
4-141
-------
Solomon, P. A.; Larson, S. M.; Fall, T.; Cass, G. R. (1988) Basinwide nitric acid and related species concentrations
observed during the Claremont Nitrogen Species Comparison Study. Atmos. Environ. 22: 1587-1594.
Spagnolo, G. S. (1989) Automatic instrument for aerosol samples using the beta-particle attenuation. J. Aerosol Sci. 20:
19-27.
Spagnolo, G. S.; Paoletti, D. (1994) Automatic system for three fractions sampling of aerosol particles in outdoor
environments. J. Air Waste Manage. Assoc. 44: 702-706.
Spengler, J.; Schwab, M.; Ryan, P. B.; Colome, S.; Wilson, A. L.; Billick, I; Becker, E. (1994) Personal exposure to
nitrogen dioxide in the Los Angeles Basin. J. Air Waste Manage. Assoc. 44: 39-47.
Spicer, C. W.; Schumacher, P. M. (1979) Particulate nitrate: laboratory and field studies of major sampling
interferences. Atmos. Environ. 13: 543-552.
Stevens, R. K.; Pace, T. G. (1984) Overview of the mathematical and empirical receptor models workshop (Quail Roost
II). Atmos. Environ. 18: 1499-1506.
Stevens, R. K.; Dzubay, T. G. ; Russwurm, G.; Rickel, D. (1978) Sampling and analysis of atmospheric sulfates and
related species. Atmos. Environ. 12: 55-68.
Stevens, R. K.; McClenny, W. A.; Dzubay, T. G.; Mason, M. A.; Courtney, W. J. (1982) Analytical methods to measure
the carbonaceous content of aerosols. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon: atmospheric
life cycle. New York, NY: Plenum Publishing Corp.; pp. 111-129.
Stevens, R.; Pinto, J.; Mamane, Y.; Ondov, J.; Abdulraheem, M.; Al-Majed, N; Sadek, M.; Gofer, W.; Ellenson, W.;
Kellogg, R. (1993) Chemical and physical properties of emissions from Kuwaiti oil fires. Water Sci. Technol.
27: 223-233.
Stott, W. R.; Chatfield, E. J. (1979) A precision SEM image analysis system with full feature EDXA characterization.
Scanning Electron Microsc. 2: 53-60.
Sweitzer, T. A. (1985) A field evaluation of two PM10 inlets in an industrialized area of Illinois. J. Air Pollut. Control
Assoc. 35: 744-746.
Symanski, E.; Rappaport, S. M. (1994) An investigation of the dependence of exposure variability on the interval
between measurements. Ann. Occup. Hyg. 38: 361-372.
Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles with a
diffusion denuder sampling system. Atmos. Environ. 28: 939-947.
league, S. V.; Raabe, O. G.; Eldred, R. A. (1992) A bench top wind tunnel for testing PM10 sampling inlets. In: Chow,
J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an A&WMA/EPA
international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management
Association; pp. 2-12. (A&WMA transactions series no. 22).
Tejada, S. B.; Zweidinger, R. B.; Sigsby, J. E., Jr.; Bradow, R. L. (1978) Modification of an ion chromatograph for
automated routine analysis: applications to mobile source emissions. In: Sawicki, E.; Mulik, J. D.; Wittgenstein,
E., eds. Ion chromatographic analysis of environmental pollutants. Ann Arbor, MI: Ann Arbor Science
Publishers, Inc.; pp. 111-124.
Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
4-142
-------
PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international specialty
conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 244-261.
(A&WMA transactions series no. 22).
Thomas, A.; Gebhart, J.; Milller, J. (1993) Effect of relative humidity on light scattering of atmospheric aerosols. J.
Aerosol Sci. 24(suppl. 1): S117-S118.
Torok, S. B.; Van Gneken, R. E. (1994) X-ray spectrometry. Anal. Chem. 66: 186R-206R.
Tsai, P.-J.; Vincent, J. H. (1993) Impaction model for the aspiration efficiencies of aerosol samplers at large angles with
respect to the wind. J. Aerosol Sci. 24: 919-928.
Tufto, P. A.; Willeke, K. (1982) Dynamic evaluation of aerosol sampling inlets. Environ. Sci. Technol. 16: 607-609.
Turner, J. R.; Hering, S. V. (1987) Greased and oiled substrates as bounce-free impaction surfaces. J. Aerosol Sci. 18:
215-224.
Turpin, B. J.; Huang, P.-F.; Roos, A.; McMurry, P. H. (1993a) Elemental analysis of single atmospheric particles
influencing visibility at the Grand Canyon. In: Bailey, G. W.; Rieder, C. L., eds. Proceedings of the 51st annual
meeting of the Microscopy Society of America; August; Cincinnati, OH. San Francisco, CA: San Francisco
Press, Inc.; pp. 1124-1125.
Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993b) Design and
evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
compounds. Environ. Sci. Technol. 27: 2441-2449.
Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. (1994) Investigation of organic aerosol sampling artifacts in the Los
Angeles basin. Atmos. Environ. 28: 3061-3071.
Turpin, B. J.; Huang, P.-F.; McMurry, P. H.; Hawkins, D. M.; Hering, S. V.; Casuccio, G. S. (1996) Description of a
cluster analysis technique to characterize single particle compositions and application to Grand Canyon
aerosols. Aerosol Sci. Technol.: submitted.
U.S. Environmental Protection Agency. (1976) Laboratory procedures for the analysis of ammonia in particulates
collected by means of HI-VOL samplers-Technicon autoanalyzer II procedures. Research Triangle Park, NC:
Analytical Chemistry Branch.
U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides: v. II.
Research Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office; report no. EPA-600/8-82-029bF. Available from: NTIS, Springfield, VA; PB84-120419.
U.S. Environmental Protection Agency. (1987) PM10 SIP development guideline. Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards; EPA report no.
EPA-450/2-86-001. Available from: NTIS, Springfield, VA; PB87-206488.
U.S. Environmental Protection Agency. (1992) Determination of the strong acidity of atmospheric fine-particles (<2.5
,uin) using annular denuder technology. Washington, DC: Atmospheric Research and Exposure Assessment
Laboratory; EPA report no. EPA/600/R-93/037.
U.S. Environmental Protection Agency. (1994) Guidelines for PM-10 sampling and analysis applicable to receptor
modeling. Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no.
EPA/452/R-94/009. Available from: NTIS, Springfield, VA; PB94-177441.
Upton, S. L.; Mark, D.; Hall, D. J. (1992) Improvements to the design and performance of an ambient inhalable aerosol
sampler. J. Aerosol Sci. 23(suppl. 1): S595-S598.
4-143
-------
Van Borm, W. A.; Adams, F. C. (1988) Cluster analysis of electron microprobe analysis data of individual particles for
source apportionment of air particulate matter. Atmos. Environ. 22: 2297-2307.
Van Borm, W. A.; Adams, F. C.; Maenhaut, W. (1989) Characterization of individual particles in the Antwerp aerosol.
Atmos. Environ. 23: 1139-1151.
Van der Meulen, A. (1986) The relationship between PM10 and thoracic particle sampling. J. Air Pollut. Control Assoc.
36: 383-387.
Vanderpool, R. W.; Lundgren, D. A.; Marple, V. A.; Rubow, K. L. (1987) Cocalibration of four large-particle
impactors. Aerosol Sci. Technol. 7: 177-185.
Van Vaeck, L.; Van Cauwenberghe, K.; Janssens, J. (1984) The gas-particle distribution of organic aerosol
constituents: measurements of the volatilisation artefact in hi-vol cascade impactor sampling. Atmos. Environ.
18:417-430.
Vaughan, N. P. (1989) The Andersen impactor: calibration, wall losses and numerical simulation. J. Aerosol Sci. 20:
67-90.
Vincent, J. H.; Mark, D. (1981) The basis of dust sampling in occupational hygiene: a critical review. Ann. Occup.
Hyg. 24: 375-390.
Vincent, J. H.; Mark, D. (1982) Applications of blunt sampler theory to the definition and measurement of inhalable
dust. Ann. Occup. Hyg. 26: 3-19.
Vossler, T. L.; Stevens, R. K.; Paur, R. J.; Baumgardner, R. E.; Bell, J. P. (1988) Evaluation of improved inlets and
annular denuder systems to measure inorganic air pollutants. Atmos. Environ. 22: 1729-1736.
WESTAR Council. (1995) Comparative data review of manual and continuous PM10 sampling methods. Seattle, WA:
U.S. Environmental Protection Agency, Region 10; July.
Waggoner, A. P.; Weiss, R. E. (1980) Comparison of fine particle mass concentration and light scattering extinction in
ambient aerosol. Atmos. Environ. 14: 623-626.
Wake, D. (1989) Anomalous effects in filter penetration measurements using the aerodynamic particle sizer
(APS 3300). J. Aerosol Sci. 20: 13-17.
Wall, S. M.; John, W.; Ondo, J. L. (1988) Measurement of aerosol size distributions for nitrate and major ionic species.
Atmos. Environ. 22: 1649-1656.
Wallace, L. A.; Duan, N.; Ziegenfus, R. (1994) Can long-term exposure distributions be predicted from short-term
measurements? Risk Anal. 14: 75-85.
Waller, R. E. (1963) Acid droplets in town air. Int. J. Air Water Pollut. 7: 773-778.
Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactor for sampling inorganic ions. Aerosol Sci.
Technol. 8: 157-172.
Wang, H.-C.; John, W. (1989) A simple iteration procedure to correct for the density effect in the aerodynamic particle
sizer. Aerosol Sci. Technol. 10: 501-505.
Watson, J. G.; Chow, J. C.; Shah, J. J.; Pace, T. G. (1983) The effect of sampling inlets on the PM-10 and PM-15 to
TSP concentration ratios. J. Air Pollut. Control Assoc. 33: 114-119.
4-144
-------
Watson, J. G.; Bowen, J. L.; Chow, J. C.; Rogers, C. F.; Ruby, M. G.; Rood, M. J.; Egami, R. T. (1989a) High-volume
measurement of size classified particulate matter. In: Lodge, J. P., Jr., ed. Methods of air sampling and analysis.
3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 427-439.
Watson, J. G.; Chow, J. C.; Mathai, C. V. (1989b) Receptor models in air resources management: a summary of the
APCA international specialty conference. JAPCA 39: 419-426.
Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q. (1990)
The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
Wedding, J. B. (1980) Ambient aerosol sampling: history, present thinking and a proposed inlet for inhalable
particulate matter. Presented at: 73rd annual meeting of the Air Pollution Control Association; June; Montreal,
PQ, Canada. Pittsburgh, PA: Air Pollution Control Association; paper no. 80-38.1.
Wedding, J. B. (1985) Errors in sampling ambient concentrations with instruments employing setpoint temperature
compensated mass flow transducers. Atmos. Environ. 19: 1219-1222.
Wedding, J. B.; Carney, T. C. (1983) A quantitative technique for determining the impact of non-ideal ambient sampler
inlets on the collected mass. Atmos. Environ. 17: 873-882.
Wedding, J. B.; Weigand, M. A. (1983) Preliminary communication: a thoracic particle inlet (d50 = 6 //m, do = 1 ij,m)
for the high-volume sampler. Atoms. Environ. 17: 1203-1204.
Wedding, J. B.; Weigand, M. A. (1993) An automatic particle sampler with beta gauging. J. Air Waste Manage. Assoc.
43: 475-479.
Wedding, J. B.; Weigand, M. A.; Carney, T. C. (1982) A 10 //m cutpoint inlet for the dichotomous sampler. Environ.
Sci. Technol. 16: 602-606.
Wedding, J. B.; Weigand, M. A.; Lodge, J. P., Jr. (1983) A thoracic particle inlet (d50 = 6 //m, d0 = 10 ,um) for the
high-volume sampler. Atmos. Environ. 17: 1203-1204.
Weiss, R. E.; Waggoner, A. P.; Charlson, R. J.; Thorsell, D. L.; Hall, J. S.; Riley, L. A. (1979) Studies of the optical,
physical & chemical properties of light absorbing aerosols. In: Novakov, T., ed. Proceedings: carbonaceous
particles in the atmosphere; March 1978; Berkeley, CA. Berkeley, CA: University of California, Laurence
Berkeley Laboratory; pp. 257-262; report no. LBL-9037.
Wernisch, J. (1985) Quantitative electron microprobe analysis without standard samples. X-ray Spectrom. 14: 109-119.
Wernisch, J. (1986) Application of different $ (pz) distributions in quantitative electron probe microanalysis without
standard samples. Radex Rundsch. 2/3: 110-113.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
Wiener, R. W.; Rodes, C. E. (1993) Indoor aerosols and aerosol exposure. In: Willeke, K.; Baron, P. A., eds. Aerosol
measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 659-689.
Wiener, R. W.; Rehme, K. A.; Eisner, A.; Vanderpool, R. W. (1994) Review and comparison of PM10 sampler
performance. Presented at: Fourth international aerosol conference; August; Los Angeles, CA.
Willeke, K.; Degarmo, S. J. (1988) Passive versus active aerosol monitoring. Appl. Ind. Hyg. 3: 263-266.
Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York, NY:
Van Nostrand Reinhold Publishers.
4-145
-------
Willeke, K.; Grinshpun, S. A.; Chang, C.-W.; Juozaitis, A.; Liebhaber, F.; Nevalainen, A.; Thompson, M. (1992) Inlet
sampling efficiency of bioaerosol samplers. J. Aerosol Sci. 23(suppl. 1): S651-S654.
Williams, E. L., II; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders. Environ. Sci.
Technol. 24: 811-814.
Winberry, W. T., Jr.; Forehand, L.; Murphy, N. T.; Ceroli, A.; Phinney, B.; Evans, A. (1993) Determination of reactive
acidic and basic gases and particulate matter in indoor air (annular denuder technique). In: Methods for
determination of indoor air pollutants: EPA methods. Park Ridge, NJ: Noyes Data Corporation.
Witz, S.; Eden, R. W.; Wadley, M. W.; Dunwoody, C.; Papa, R.; Torre, K. J. (1990) Rapid loss of particulate nitrate,
chloride and ammonium on quartz fiber filters during storage. J. Air Waste Manage. Assoc. 40: 53-61.
Wolff, G. T.; Groblicki, P. J.; Cadle, S. H.; Countess, R. J. (1982) Particulate carbon at various locations in the United
States. In: Wolff, G. T.; Klimsch, R. L., eds. Particulate carbon: atmospheric life cycle. New York, NY: Plenum
Press; pp. 297-315.
Woskie, S. R.; Shen, P.; Fmkel, M.; Eisen, E. A.; Smith, T. J.; Smith, R.; Wegman, D. H. (1993) Calibration of a
continuous-reading aerosol monitor (Miniram) to measure borate dust exposures. Appl. Occup. Environ. Hyg.
8: 38-45.
Yamada, Y. (1983) A new method for the determination of collection efficiency of an aerosol sampler by electron
microscopy. Atmos. Environ. 17: 369-372.
Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
Ye, Y.; Tsai, C.-J.; Pui, D. Y. H.; Lewis, C. W. (1991) Particle transmission characteristics of an annular denuder
ambient sampling system. Aerosol Sci. Technol. 14: 102-111.
Yeh, H.-C. (1993) Electrical techniques. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 410-426.
Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine particulate nitrates during sampling. Atmos. Environ. Part
A 26: 3305-3312.
Zielinska, B.; Arey, J.; Atkinson, R.; Winer, A. M. (1989) The nitroarenes of molecular weight 247 in ambient
particulate samples collected in southern California. Atmos. Environ. 23: 223-229.
Zoller, W. H.; Gordon, G. E. (1970) Instrumental neutron activation analysis of atmospheric pollutants utilizing Ge(Li)
y-ray detectors. Anal. Chem. 42: 257-265.
4-146
-------
5. SOURCES AND EMISSIONS OF
ATMOSPHERIC PARTICLES
5.1 INTRODUCTION
Unlike gaseous criteria pollutants (SO2, NO2, CO, O3), which are well defined
chemical entities, atmospheric particles comprise a complex mixture of chemical constituents.
Because of this fact, sources of each constituent of the atmospheric aerosol must be considered
in turn. Since particulate matter (PM) is composed of both primary and secondary constituents,
emissions of both the primary components and the gaseous precursors must be considered. The
chemical composition of ambient aerosols was treated in general terms in Chapter 3.
Information on ambient concentrations of particles of various sizes (PM10, PM25) and their
chemical composition, based on specific field studies, is presented in Chapter 6.
Tables 5-1A and 5-1B summarize anthropogenic and natural sources for the major primary
and secondary aerosol constituents of fine and coarse particles. Anthropogenic sources can be
further divided into stationary and mobile sources. Stationary sources include fuel combustion
for electrical utilities and industrial processes; construction and demolition; metals, minerals,
petrochemicals and wood products processing; mills and elevators used in agriculture; erosion
from tilled lands; waste disposal and recycling; and fugitive dust from paved and unpaved roads.
Mobile, or transportation related, sources include direct emissions of primary PM and secondary
PM precursors from highway and off-highway vehicles and nonroad sources. Also shown are
sources for precursor gases whose oxidation forms secondary particulate matter. In general, the
nature of sources of particulate matter shown in Table 5-1A is very different from that for
particulate matter shown in Table 5-IB. A large fraction of the mass in the fine size fraction is
derived from material that has been volatilized in combustion chambers and then recondensed to
form primary fine PM, or has been formed in the atmosphere from precursor gases as secondary
PM. Since precursor gases and fine particulate matter are capable of travelling great distances, it
is difficult to identify individual sources of constituents shown in Table 5-1 A. The PM
constituents shown in Table 5-IB
5-1
-------
TABLE 5-1A. CONSTITUENTS OF ATMOSPHERIC FINE PARTICLES (<2.5
AND THEIR MAJOR SOURCES
Sources
Primary
Secondary
Aerosol
species Natural Anthropogenic
SO4= Sea spray Fossil fuel
combustion
Natural
Oxidation of reduced
sulfur gases emitted
by the oceans and
wetlands; and SO2 and
H2S emitted by
volcanism and forest
fires
Anthropogenic
Oxidation of SO2
emitted from fossil fuel
combustion
Erosion,
re-entrainment
NO,
Minerals
NH/
Organic Wild fires
carbon
(OC)
Elemental Wild fires
carbon
Metals
Volcanic
activity
Motor vehicle exhaust
Fugitive dust; paved,
unpaved roads;
agriculture and
forestry
Motor vehicle exhaust
Open burning, wood
burning, cooking,
motor vehicle
exhaust, tire wear
Motor vehicle
exhaust, wood
burning, cooking
Fossil fuel
combustion, smelting,
brake wear
Oxidation of NOX
produced by soils,
forest fires, and
lighting
Emissions of NH3
from wild animals,
undisturbed soil
Oxidation of
hydrocarbons emitted
by vegetation,
(terpenes, waxes);
wild fires
Oxidation of NOX
emitted from fossil fuel
combustion; and in
motor vehicle exhaust
Emissions of NH3 from
animal husbandry,
sewage, fertilized land
Oxidation of
hydrocarbons emitted
by motor vehicles,
open burning, wood
burning
Bioaerosols Viruses,
bacteria
5-2
-------
TABLE 5-1B. CONSTITUENTS OF ATMOSPHERIC COARSE PARTICLES
(>2.5 ,um) AND THEIR MAJOR SOURCES
Sources
Aerosol species
Minerals
Metals
Miscellaneous
ions
Organic carbon
Organic debris
Bioaerosols
Primary
Natural
Erosion,
re-entrainment
Erosion,
re-entrainment,
organic debris
Sea spray
—
Plant, insect
fragments
Pollen, fungal
spores, bacterial
agglomerates
Secondary
Anthropogenic Natural Anthropogenic
Fugitive dust; paved, — —
unpaved road dust,
agriculture and forestry
— — —
Road salting — —
Tire and asphalt wear — —
— — —
_
have shorter lifetimes in the atmosphere, so their impacts tend to be more localized. Only major
sources for each constituent are listed in Tables 5-1A and 5-1B.
Natural sources of primary PM include windblown dust from undisturbed land, sea spray,
and plant and insect debris. The oxidation of a fraction of terpenes emitted by vegetation and
reduced sulfur species from anaerobic environments leads to secondary PM formation.
Ammonium (NH4+) ions which are crucial for regulating the pH of particles are derived from
emissions of ammonia (NH3) gas. Source categories for NH3 have been divided into emissions
from undisturbed soils (natural) and emissions which are related to human activities (e.g.,
fertilized lands, domestic and farm animal waste). It is difficult to describe emissions from
biomass burning as either natural or anthropogenic. Clearly, fuel wood burning is an
anthropogenic source of PM, whereas wildfires would be a natural source. Forest fires have
been included as a natural source, because of the lack of information on the amount of
prescribed burning or accidental fires caused by humans.
5O
-------
Similar considerations apply to the biogenic emissions of trace metals which may be remobilized
from anthropogenic inputs.
Although a large number of potential source contributions have been listed for particulate
matter and gaseous precursors in Tables 5-1A and 5-1B, it should be noted that emissions
inventories have been compiled for only a limited number of entries for either aerosol
constituents or source categories. The remainder of the chapter includes discussion of the
processes responsible for the most important sources of primary and secondary PM in Sections
5.2 and 5.3, respectively, followed by discussion of emissions estimates for the United States in
Section 5.4. Applications of emissions inventories and other techniques, such as receptor
modeling for inferring sources of ambient particulate matter, are then discussed in Section 5.5.
5.2 SOURCES OF PRIMARY PARTICULATE MATTER
This section discusses processes responsible for the emissions of primary particulate
matter. The order of sources roughly follows their estimated relative source strengths for the
United States to be presented in Section 5.4. Emissions of mineral particles produced as the
result of natural wind erosion and human activities are discussed in 5.2.1. Sources of primary
particulate matter produced by fossil fuel combustion and other stationary anthropogenic sources
are discussed in 5.2.2, while sources of secondary particulate matter are discussed in section 5.3.
Motor vehicle emissions are discussed in 5.2.3. Vegetation burning in woodstoves and forest
fires is discussed as a source of particulate matter in 5.2.4. Sea salt aerosol production, the
suspension of organic debris, and the production of trace metals by natural processes are
discussed in 5.2.5. Data for the chemical composition and particle size distribution for each of
these sources of particulate matter are included where available along with information about
techniques for measuring source compositions and emissions rates.
5.2.1 Wind Erosion and Fugitive Dust
Windblown dust constitutes a major component of the atmospheric aerosol, especially in
arid and semi-arid areas of the world. Windblown dust represents the largest single category
5-4
-------
in global emissions inventories, constituting about 50% of the total global source of primary and
secondary particulate matter (IPCC, 1995). Since the next major category is sea-salt aerosol
production, which is estimated to constitute about 40% of total emissions, it can be seen that
about 70% of non-sea-salt aerosol emitted is in the form of mineral dust. If one-half of the dust
is assumed to be emitted in the PM10 size range, then it can be seen that 54% of non-sea-salt
PM10 emitted globally is dust, less than about 10% of which originates in the United States.
Many areas of the western United States are classified as arid or semi-arid, potentially
leading to a larger contribution of dust to the mass of the ambient aerosol there compared to the
eastern United States. Large-scale dust events are generally associated with semi-arid regions
where marginal lands are used for agriculture and herding. During times of drought, the
denuded and broken soil surface is easily carried away, periodically forming "dust bowl"
conditions as in the midwestern U.S. (Prospero, 1995).
Emission rates of mineral aerosols are found to be strongly dependent on meteorological
parameters such as wind velocity and precipitation. Wind tunnel experiments (Bagnold, 1941;
Chepil, 1945) have shown that the motion of loose particles on the surface is initiated when the
surface wind stress (The wind stress acting on the surface is supplied by the downward transport
of momentum from the mean winds. In micrometeorological applications, u*, or the square root
of the ratio of the wind stress to the air density is used.) acting on erodible particles exceeds the
downward force of gravity and the interparticle cohesion forces acting on the particles. Particle
motion occurs when u* exceeds the threshold friction velocity, u*t, which is dependent on particle
properties. Values of u*t are strongly size dependent, with a minimum for particles having
diameters of about 60 jam (Bagnold, 1941). Individual smaller particles are held by cohesive
forces and larger particles are constrained by the force of gravity. Measurements of u*t are
available for a number of different soil types (e.g., Gillette et al., 1980).
Three types of particle motion were characterized in the early wind tunnel experiments:
suspension, saltation, and creeping. Suspension refers to the upward transport of dust
(d< 60 jim) by turbulent eddies; saltation to the horizontal motion of particles (60 < d <
2000 |im) which can reach heights of up to a meter above the surface before they fall back;
5-5
-------
creeping to particles too massive (d > 2000 jim) to be lifted from the surface so they roll along.
Because of strong cohesive forces in soil crusts and rock surfaces, particles are not
suspended directly by the transfer of momentum from the wind but by sandblasting and abrasion
by saltating particles. The impact of saltating grains then results in the emission of smaller
particles (Shao et al., 1993). The flux of saltating particles increases rapidly with wind speed,
and varies as (u*)2(u*-u*t). The size distribution of the suspended aerosol is then controlled by
the aerosol microphysical processes of coagulation and sedimentation.
Non-erodible elements on natural surfaces cut down on the surface area available for
erosion, and they take up wind momentum which would otherwise be available for erosion. Soil
moisture, salts, and organic matter mainly affect soil cohesion (e.g., Gillette et al., 1982) and
thus the size distribution of soil particle aggregates. Chepil (1956), Belly (1964), Bisal and
Hsieh (1966), and Svasek and Terwindt (1974) show that substantially greater wind forces are
needed when soil surface moisture is increased by less than 1% from its dry state. The moisture
content of soils will vary throughout the year depending on the frequency and intensity of
precipitation events, irrigation, and the relative humidity and temperature of the surrounding air.
Large amounts of rain falling during 1 mo of a year will not be as effective in stabilizing dust as
the same amount of rain interspersed at intervals throughout the year.
An operational difficulty arises because u* is derived from anemometers placed at a height
of 5 or 10 m above the surface and requires assumptions about the wind profile down to the
surface. The challenge is to derive values for wind stress acting on erodible elements (Alfaro
and Gomes, 1995) which are valid for large areas. Alfaro and Gomes (1995) have derived
relations between wind velocity measurements made at conventional heights and surface wind
stresses using radar imagery to characterize surface roughness. Surface roughness is determined
by the presence of vegetation, structures, rocks and boulders, topographic irregularities and
surface obstructions. Marticorena and Bergametti (1995) have developed parameterizations
including these physical considerations suitable for use in large scale models.
Apart from the large-scale, mean flow small-scale atmospheric vortices are also capable of
suspending dust. Dust devils, so-called because of the dust they entrain, may be found in arid
areas along roads or where the surface has been disturbed by human activity (Hall,
5-6
-------
1981; Snow and McClelland, 1990). Hall (1981) proposed that dust devils could constitute the
major source of suspended dust on hot summer days with light winds and convectively unstable
conditions, as an example in Pima Co., AZ demonstrates. Hall (1981) estimated that large scale
winds could raise 171 kg km"2 day"1 and motor vehicles could raise 48 kg km"2 day"1 on an
annually averaged basis, while dust devils could raise up to 250 kg km"2 day"1 of dust (in all size
ranges) on hot summer days. Atmospheric vortices are not a source component currently treated
in emissions inventories.
Apart from sources within the continental United States, an additional source of
windblown dust involves the long-range transport of dust from the Sahara desert westward
across the Atlantic Ocean. Individual dust storms have been tracked across the Atlantic, after
emerging from the northwest coast of Africa, to the east coast of the United States (Ott et al.,
1991). Saharan dust is carried into the Miami area, capable of producing dense hazes during the
summer (Prospero et al., 1987). While summertime monthly mean dust concentrations are about
10 |ig/m3 (Prospero et al., 1993), dust events are highly sporadic and of short duration. In a
one-year study of Saharan dust deposition in Miami, Prospero et al. (1987) found that 22% of
the annual deposition occurred in one day and 68% in rain events that occurred during two dust
episodes spread over a total of four days. Gatz (1995) has found evidence suggesting that
Saharan dust has reached as far as central Illinois in at least one episode which occurred during
the summer of 1979. Up to 20 |ig/m3 of the ambient aerosol may have originated in the Sahara
desert and the Sahel during this episode. These dust events are highly sporadic and more work
needs to be done to characterize the frequency, magnitude, and variability of these events.
Similar dust transport may also occur from the deserts of Asia across the Pacific Ocean
(Prospero, 1995), but it is not clear to what extent any of this dust reaches the United States (See
Chapter 6 for more information on long distance transport of dust particles into the United States
from Africa or Asia.)
The compositions of soils and average crustal material are shown in Table 5-2 (adapted
from Warneck, 1988). Two entries are shown as representations of average crustal material.
Differences from the mean soil composition shown can result from local geology and climate
conditions. Major elements in both soil and crustal profiles are Si, Al, and Fe which are found
in the form of various minerals. In addition, organic matter constitutes a few percent,
5-7
-------
TABLE 5-2. AVERAGE ABUNDANCES OF MAJOR
ELEMENTS IN SOIL AND CRUSTAL ROCK
Elemental Abundances (ppmw)
Element
Si
Al
Fe
Ca
Mg
Na
K
Ti
Mn
Cr
V
Co
Soil
(a)
330,000
71,300
38,000
13,700
6,300
6,300
13,600
4,600
850
200
100
8
Crustal
(b)
277,200
81,300
50,000
36,300
20,900
28,300
25,900
4,400
950
100
135
25
Rock
(c)
311,000
77,400
34,300
25,700
33,000
31,900
29,500
4,400
670
48
98
12
Source: (a) Vinogradov (1959); (b) Mason (1966); (c) Turekian (1971), Model A; as quoted in Warneck (1988).
on average, of soils. In general, the soil profile is similar to the crustal profiles, except for the
depletion of soluble elements such as Ca, Mg, and Na.
Because of the enormous difficulties encountered in developing theoretical estimates of
windblown dust emissions, most current estimates rely on the results of empirical studies. These
studies typically involve the placement of wind tunnels over natural surfaces and then measuring
emission rates and size distributions for different wind conditions. The emissions of fugitive
dust raised as the result of human activities are also extremely difficult to quantify. Fugitive
dust emissions arise from paved and unpaved roads, building construction and demolition,
storage piles, and agricultural tilling in addition to wind erosion.
Figure 5-1 shows examples of size distributions in dust from paved and unpaved roads,
agricultural soil, sand and gravel, and alkaline lake bed sediments which were measured in a
5-S
-------
100
ID
w
m
Paved Unpaved Agricultural Soil/Gravel
Road Dust Road Dust Soil
Alkaline
Lake Bed
<1.0|jm
<2.5|jm
10
OlSP
Figure 5-1. Size distribution of particles generated in a laboratory resuspension
chamber.
Source: Chow et al. (1994).
laboratory resuspension chamber as part of a study in California (Chow et al., 1994). This figure
shows substantial variation in particle size among some of these fugitive dust sources. The PMj 0
abundance (6.9%) in the alkaline lake bed dust is twice its abundance in paved and unpaved
road dust. Approximately 10% of TSP is in the PM25 fraction and approximately 50% of TSP is
in the PM10 fraction. The sand/gravel dust sample shows that 65% of the mass is in particles
larger than the PM10 fraction. The PM2 5 fraction of TSP is approximately 30% to 40% higher in
alkaline lake beds and sand/gravel than in the other soil types. The tests were performed after
seiving and with a short (<1 min) waiting period prior to sampling. It is expected that the
fraction of PMl 0 and PM2 5 would increase with distance from a fugitive dust emitter as the
larger particles deposit to the surface at a larger velocity than the smaller particles. Additional
data shown in Figure 5-2 (Houck et al., 1989, 1990) were obtained in a study characterizing
particle sources in California.
5-9
-------
100
80
60
a.
-------
Unpaved roads and other unpaved areas with vehicular activity are essentially unlimited
reservoirs of dust loading when vehicles are moving. These surfaces are always being disturbed,
and wind erosion seldom has an opportunity to increase their surface roughness sufficiently to
inhibit particle suspension. The U.S. EPA AP-42 emission factor (U.S. Environmental
Protection Agency, 1995a) for unpaved road dust emissions contains variables which account for
silt loading, mean vehicle speed, mean vehicle weight, mean number of wheels, and number of
days with detectable precipitation, to determine annual PM10 dust emissions for each
vehicle-kilometer traveled. Dust loadings on a paved road surface build up by being tracked out
from unpaved areas such as construction sites, unpaved roads, parking lots, and shoulders; by
spills from trucks carrying dirt and other particulate materials; by transport of dirt collected on
vehicle undercarriages; by wear of vehicle components such as tires, brakes, clutches, and
exhaust system components; by wear of the pavement surface; by deposition of suspended
particles from many emissions sources; and by water and wind erosion from adjacent areas.
Moisture causes dust to adhere to vehicle surfaces so that it can be carried out of unpaved roads,
parking lots, and staging areas. Carryout also occurs when trucks exit heavily watered
construction sites. This dust is deposited on paved roadway surfaces as it dries, where it is
available for suspension far from its point of origin. Fugitive dust emissions from paved roads
are often higher after rainstorms in areas where unpaved accesses are abundant, even though the
rain may have flushed existing dust from many of the paved streets.
The size distribution of samples of paved road dust obtained from a source characterization
study in California is shown in Figure 5-2. As might be expected, most of the emissions are in
the coarse size mode. The chemical composition of paved road dust obtained in Denver, CO,
during the winter of 1987-1988 is shown in Figure 5-3. The chemical composition of paved
road dust is much like an ambient PM10 sample, which consists of a complex mixture of
particulate matter from a wide variety of sources. Hopke et al. (1980) found that the inorganic
composition of urban roadway dust in samples from Urbana, IL could be described in terms of
contributions from natural soil, automobile exhaust, rust, tire wear, and salt. Automobile
contributions arose from exhaust emissions enriched in Pb; from rust as Fe; tire wear particles
enriched in Zn; brake linings enriched in Cr, Ba, and Mn; and cement particles derived from
roadways by abrasion. The complexity
5-11
-------
10
Chemical Compound
Figure 5-3. Chemical abundances for PM2 5 emissions from paved road dust in Denver,
CO. Solid bars represent fractional abundances, and the error bars represent
variability in species abundances. Error bars represent detection limits when
there are no solid bars.
Source: Watson and Chow (1994).
of paved road dust is also evident in the comparison of a paved road dust profile reported by
Chow et al. (1991) for Phoenix, AZ, with profiles from other geological sources in the area.
Chow et al. (1991) noted that the abundance of organic carbon in the profile was 11±9%, larger
and more variable than its abundance in profiles from agricultural land, construction sites, and
vacant lots. The presence of particles produced by automotive emissions, tire wear, organic
detritus, and engine oils may account for this enrichment for organic carbon. The abundances of
Pb and Br in Phoenix paved road dust were more than double the concentrations in the other
geological profiles, indicating the presence of tailpipe exhaust from vehicles burning leaded
5-12
-------
fuels. The contribution of tire wear could have been from 4 to 45% of that of motor vehicle
exhaust, based on the results of Pierson and Brachaczek (1974). Enrichments in species from
clutch and brake wear were not detectable in the Phoenix paved road dust profiles. These are
often composed of asbestos and/or semi-metal carbon composites. Cooper et al. (1987)
examined the elemental composition of semi-metal brake shoes and found abundances of-45%
Fe, -2% Cu, -0.5% Sn, -3% Ba, and -0.5% Mo. None of these species were found in the
Phoenix paved road dust profiles at levels significantly in excess of their abundances in other
geological sub-types.
Many fugitive dust sources are episodic rather than continuous emitters. Though
windblown dust emissions are low on an annual average, they can be quite large during those
few episodes when wind speeds are high. In Coachella Valley, CA, the South Coast Air Quality
Management District (1994) calculated 24-h emissions based on a worst windy day. On a day
when wind gust speeds exceeded 96 km/h, fugitive dust emissions could account for 20% of the
entire annual emissions. Since the rate of dust suspension varies as the cube of the wind speed
for large wind speeds, estimates of windblown dust emissions use highest wind speeds quoted in
National Weather Service Local Climatological Summaries. Construction activities are also
episodic in nature. Reeser et al. (1992) reported that fugitive dust emissions during wintertime
in Denver, CO, were 44% higher than those found in the annual inventory using standard
emissions inventory methods.
Finally, the spatial disaggregation for fugitive dust emissions is poorer than that for all
other source categories. Whereas most mobile sources are confined to established roadways and
most area sources are located in populated regions, suspendable dust sources are everywhere.
Most fugitive dust emissions are compiled on a county-wide basis and are not allocated to
specific fields, streets, unpaved roads, and construction sites possibly contributing to high
airborne PM concentrations. Several of these limitations may be impossible to overcome, but
many result from old methods being applied to the problem.
The inherent variability of fugitive dust emissions may preclude absolute emissions
estimates. Nevertheless, this examination of physical processes shows that better knowledge of
the locations of these emissions, the joint frequencies of activities and different meteorological
conditions, and more site specific measurements of key parameters could provide much better
absolute emissions rates than are now available.
5-13
-------
5.2.2 Stationary Sources
The combustion of fossil fuels, such as coal and oil, leads to the formation of both primary
and secondary particulate matter. Fossil fuels are mainly composed of a mixture of the remnants
of plant matter and surrounding soils which have been processed at elevated temperatures and
pressures over periods of up to three hundred million years. The process of coal formation
results in a matrix of high molecular weight, highly cross-linked polyaromatic carbocyclic and
heterocyclic ring compounds containing C, H, O, N, P, and S, and crustal materials. The
hydrogen, nitrogen and phosphorus contents of coal are lower than the original biomass,
reflecting losses by microbial utilization and thermal processing. Petroleum consists of long
chain straight and branched alkanes with high carbon numbers (i.e., C25-C50), alkenes and
aromatic hydrocarbons. The trace element content of these fuels reflects the trace element
content of the initial organic matter and soil, subsequent hydrothermal alteration during
diagenesis and industrial processing. Because of the inherent variability in each of these factors,
the trace element content of fossil fuels is highly variable.
Coal combustion in the high temperature combustion zones of power plants results in the
melting and volatilization of refractory crustal components, such as aluminosilicate minerals
which condense to form spherical fly ash particles. Fly ash is enriched with metals compared to
ordinary crustal material by the condensation of metal vapors. The sulfur content of fossil fuels
ranges from fractions of a percent to about 4%. The sulfur in the fuel is released primarily as
SO2 along with smaller amounts of sulfate. Ratios of sulfate S to total S range from about 1%
for modern coal fired power plants to several percent in residential, commercial and industrial
boilers (Goklany et al., 1984).
The elemental composition of primary parti culate matter emitted in the fine fraction from a
variety of power plants and industries in the Philadelphia area is shown in Table 5-3 as a
representative example of emissions from stationary fossil combustion sources (Olmez et al.,
1988). Entries for the coal fired power plant show that Si and Al followed by sulfate are the
major primary constituents produced by coal combustion, while fractional abundances of
elemental carbon were much lower and organic carbon species were not detected. Sulfate is the
major parti culate constituent released by the oil fired power plants examined in this study; and,
again, elemental and organic carbon are not among the major species emitted. Olmez et al.
(1988) also compared their results to a number of similar studies and concluded
5-14
-------
TABLE 5-3. COMPOSITION OF FINE PARTICLES RELEASED BY
VARIOUS STATIONARY SOURCES IN THE PHILADELPHIA AREA
Species
(Units)
C-v (%)
C-e (%)
NH4 (%)
Na (%)
Al (%)
Si (%)
P (%)
S (%)
SO4 (%)
Cl (%)
K (%)
Ca (%)
Sc (ppm)
Ti (%)
V (ppm)
Cr (ppm)
Mn (ppm)
Fe (%)
Co (ppm)
Ni (ppm)
Eddy stone Coal-
Fired Power
Plant
ND
0.89 ±0.12
1.89±0.19
0.31 ±0.03
14±2
21.8±1.6
0.62 ±0.10
3. 4 ±0.6
11. 9± 1.2
0.022 ±0.11
1.20 ±0.09
1.4±0.5
42 ±2
1.1±0.2
550 ±170
390 ± 120
290 ±15
7.6 ±0.4
93 ± 10
380 ±50
Oil-Fired Power Plants
N
o
5
3
3
3
9
9
9
o
J
3
9
3
3
3
3
3
o
J
3
o
5
9
Eddy stone
2.7 ±1.2
7.7 ±1.5
3.5±1.6
3.0 ±0.8
0.45 ±0.09
1.9±0.6
1.5 ±0.4
11±2
40 ±4
0.019 ±0.009
0.16 ±0.05
3.6±1.0
0.17 ±0.02
0.040 ± 0.044
11 500 ±3000
235 ±10
380 ± 40
1.6 ±0.2
790 ± 150
15000 ±5000
N
3
3
3
o
J
3
9
9
9
o
5
2
9
o
J
3
9
3
3
3
3
o
5
9
Schuylkill
0.75 ±0.63
0.22 ±0.17
3.7 ±1.7
3. 3 ±0.8
0.94 ±0.08
2.6 ±0.4
1.0 ±0.2
13 ±1
45 ±7
ND
0.21 ±0.03
2.3 ±1.0
0.47 ± 0.02
0.12 ±0.02
20000 ± 3000
230 ± 70
210 ±50
1.7 ±0.4
1100 ±200
19000 ±2000
N
4
4
4
o
J
3
11
11
11
4
11
o
J
3
11
3
3
3
3
3
11
Secondary
Al Plant
1.6±1.5
0.18±0.10
2.2 ±0.9
16.3 ±0.8
1.74 ±0.09
3.1 ±2.2
0.45 ± 0.27
3±4
5.9 ±2
21±4
10.9±1.5
0.12 ±0.09
0.092 ±0.039
0.024 ±0.003
36 ±7
410 ±20
120 ±15
0.31 ±0.02
13±2
300 ±100
N
2
2
2
1
1
2
2
2
2
1
2
2
1
2
1
1
1
1
1
2
Fluid Cat.
Cracker
ND
0.16 ±0.05
0.43 ± 0.22
0.38 ±0.05
6.8±1.2
9.8 ±20.0
ND
4.2 ±12.6
38 ±4
ND
0.031 ±0.005
0.030 ±0.004
2.7 ±0.4
0.38±0.1
250 ± 70
59 ±8
14±3
0.20 ±0.03
15±2
220 ± 30
N
3
3
3
3
9
9
3
9
9
3
o
J
3
3
o
6
9
3
9
Municipal
Incinerator
0.57 ±0.26
3. 5 ±0.2
0.36 ±0.07
6.6 ±3. 5
0.25 ±0.10
1.7 ±0.3
0.63 ±0.12
2.9 ±0.8
6.8 ±2.3
29 ±5
7.6 ±2.3
0.23 ±0.10
0.11 ±0.02
0.030 ±0.015
8.6 ±5. 3
99±31
165 ±40
0.22 ±0.05
3. 7 ±0.8
290 ± 40
N
4
4
4
o
5
3
10
10
10
4
3
10
10
1
10
2
3
3
3
o
6
10
-------
TABLE 5-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED
BY VARIOUS STATIONARY SOURCES IN THE PHILADELPHIA AREA
Oi
Species
(Units)
Cu (ppm)
Zn (%)
As (ppm)
Se (ppm)
Br (ppm)
Rb (ppm)
Sr (ppm)
Zr (ppm)
Mo (ppm)
Ag (ppm)
Cd (ppm)
In (ppm)
Sn (ppm)
Sb (ppm)
Cs (ppm)
Ba (ppm)
La (ppm)
Ce (ppm)
Nd (ppm)
Sm (ppm)
Eddy stone Coal-
Fired Power
Plant
290 ± 20
0.041 ±0.005
640 ± 80
250 ± 20
35 ±8
190 ±80
1290 ±60
490 ±190
170 ±60
ND
ND
0.71 ±0.04
ND
(a)
9.2 ±0.9
ND
120 ± 10
180 ± 10
80 ±26
23 ±2
Oil-Fired Power Plants
N
9
3
3
3
3
1
9
9
2
2
2
3
2
3
3
Eddy stone
980 ± 320
1.3 ±0.3
33 ±6
26 ±9
90 ±60
ND
160 ±50
140 ± 180
930 ±210
ND
ND
ND
320 ± 230
370 ±410
ND
1960 ±100
130 ±30
89 ±23
28 ±5
3.7 ±0.7
N
9
3
1
3
9
9
9
3
9
3
3
3
3
2
3
Schuylkill
1100 ±500
0.78 ±0.30
50 ±16
23 ±7
45 ±17
ND
280 ± 70
100 ±120
1500 ±300
ND
ND
ND
200 ± 80
1020 ±90
ND
2000 ± 500
450 ± 30
360 ± 20
230 ± 20
20.5 ±1.5
N
11
3
3
3
11
11
11
3
11
3
3
3
3
3
3
Secondary
Al Plant
450 ± 200
0.079 ± 0.006
15±6
66 ±3
630 ± 70
97 ±38
ND
ND
ND
ND
ND
ND
550 ± 540
6100 ±300
ND
ND
19 ±2
ND
ND
ND
N Fluid Cat. Cracker
2 14±8
1 0.0026 ± 0.0007
1 ND
1 15± 1
2 5.6±1.8
1 ND
36 ±6
130 ±50
ND
ND
ND
ND
2 ND
1 7.7 ±1.5
ND
290 ± 90
1 3300 ± 500
2700 ± 400
1800 ±250
170 ±20
N
9
3
3
9
9
2
3
2
3
3
3
3
Municipal
Incinerator
1300 ±500
10.4 ±0.5
64 ±34
42 ± 16
2300 ± 800
230 ± 50
87 ± 14
ND
240 ± 130
71 ±15
1200 ±700
4.9 ±1.4
6700 ±1900
1300± 1000
5.9 ±3.0
ND
1.1 ±0.5
ND
ND
ND
N
3
3
3
3
10
2
10
10
3
3
3
10
3
3
1
-------
TABLE 5-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED
BY VARIOUS STATIONARY SOURCES IN THE PHILADELPHIA AREA
Species
(Units)
Eu (ppm)
Gd (ppm)
Tb (ppm)
Yb (ppm)
Lu (ppm)
Hf(ppm)
Ta (ppm)
W (ppm)
Au (ppm)
Pb (%)
Th (ppm)
% mass
Eddy stone
Coal-Fired
Power Plant
5.1 ±0.5
ND
3. 3 ±0.3
10.3 ±0.5
ND
5. 8 ±0.8
ND
20 ±8
ND
0.041 ±0.004
24 ±2
24 ±2
Oil-Fired Power Plants
N
3
3
1
3
1
9
3
6
Eddy stone N
ND
ND
ND
ND
ND
0.39 ±0.07 1
ND
60 ±5 2
0.054 ±0.017 2
1.8±0.6 9
1.9±0.5 2
93. 5 ±2.5 6
Schuylkill N
0.65 ±0.23 3
ND
0.90 ±0.29 3
ND
ND
ND
ND
ND
ND
1.0 ±0.2 11
ND
96 ±2 6
Secondary
Al Plant
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.081 ±0.014
ND
81 ± 10
N Fluid Cat. Cracker
4.9 ±0.7
71 ± 10
8.9±1.3
3.7 ±0.4
0.59±0.17
0.99 ±0.08
0.56 ±0.10
ND
ND
2 0.0091 ±0.0021
6.2 ±0.7
2 97 ±2
N
3
3
3
3
3
3
3
9
3
7
Municipal
Incinerator
ND
ND
ND
ND
ND
ND
ND
ND
0.56 ±0.27
5.8±1.2
ND
89 ±2
N
3
10
7
N = Number of samples.
ND = Not detected.
The "% mass" entries give the average percentage of the total emitted mass found in the fine fraction.
(a) Omitted because of sample contamination.
Source: Adapted from Olmez et al. (1988).
-------
that their data could have much wider applicability to receptor model studies in other areas with
some of the same source types. The high temperature of combustion in power plants results in
the almost complete oxidation of the carbon in the fuel to CO2 and very small amounts of CO.
A number of trace elements are greatly enriched over crustal abundances (in different fuels),
such as Se in coal and V and Ni in oil. In fact, the higher V content of the fuel oil than in coal
could help account for the higher sulfate seen in the profiles from the oil-fired power plant
compared to the coal-fired power plant since V is known to catalyze the oxidation of reduced
sulfur species. Although Table 5-3 only gives values of the fine particle composition,
measurements of coarse particle composition were also reported by Olmez et al. (1988) which
were qualitatively similar.
The composition of the organic carbon produced by stationary sources has not been well
characterized. Information is available for the composition of poly cyclic aromatic
hydrocarbons, or PAH's (Daisey et al., 1986), while data for the composition of other classes of
organic compounds are sparse. In addition, the phase distribution of a number of PAH's and
other organic compounds will depend strongly on ambient atmospheric conditions. It may be
expected that the composition of emissions in systems operating at low temperatures (e.g.,
residential coal combustion) will reflect that of the unburned fuel.
Emissions from stationary sources are determined mainly by stack sampling with a variety
of techniques. All these techniques rely on measurements of stack flow rates and concentrations
of pollutants to determine emissions. Method 5 (Federal Register, 1977) consists of a sampling
train which is commonly used to measure emissions of various trace metals. The method is
cumbersome and is limited in the number of species that can be sampled. Based on the
realization that direct sampling of hot undiluted stack gases may not yield an accurate
representation of the chemical composition and size distribution of particles leaving the stack,
dilution sampling has also been used (e.g., Olmez et al., 1988). Condensation, coagulation, and
chemical reactions occur as stack gases are cooled and diluted. In dilution sampling, stack gases
are diluted with filtered ambient air in an attempt to partially simulate processes occurring in
upper portions of the stack and in the plume leaving the stack. Another advantage in the use of
dilution systems is that the same sampling substrates and analytical techniques used in ambient
sampling can be used. As a result, a wider variety of constituents can be sampled than in
conventional direct sampling techniques and biases resulting from the use of separate sampling
5-18
-------
systems in source apportionment studies are eliminated. Remote monitoring methods (e.g.,
differential optical absorption spectroscopy) have also been used to determine emissions of
species such as Hg. The size distribution of particles emitted by burning crude oil is shown in
Figure 5-2. As can be seen, almost all of the mass is in the fine fraction.
Apart from emissions in the combustion of fossil fuels, trace elements are emitted as the
result of various industrial processes such as steel and iron manufacturing and non-ferrous metal
production (e.g., for Pb, Cu, Ni, Zn, and Cd) as may be expected, emissions factors for various
trace elements are highly source-specific (Nriagu and Pacyna, 1988). Inspection of Table 5-3
reveals that the emissions from the catalytic cracker and the oil-fired power plant are greatly
enriched in rare-earth elements such as La compared to other sources.
Emissions from municipal waste incinerators are dominated by Cl arising mainly from the
combustion of plastics and metals that form volatile chlorides. The metals can originate from
cans or other metallic objects and some metals such as Zn and Cd are also additives in plastics or
rubber. Many elements such as S, Cl, Zn, Br, Ag, Cd, Sn, In, and Sb are enormously enriched
compared to their crustal abundances. A comparison of the trace elemental composition of
incinerator emissions in Philadelphia, PA (shown in Table 5-3) with the composition of
incinerator emissions in Washington D.C., and Chicago, IL, (Olmez et al., 1988) shows
agreement for most constituents to better than a factor of two. High levels of Hg associated with
emissions from medical waste incinerators from discarded thermometers, mercurials, mercury
batteries, etc., have been declining because of reductions in the use of Hg for medical purposes
(Walker and Cooper, 1992). A sizable fraction of the Hg may be particulate Hg(II) as opposed
to gas phase Hg°.
5.2.3 Mobile Sources
Particulate matter from motor vehicles originates from tailpipe exhaust and from friction
acting on individual components such as tires and brakes. Both diesel and gasoline fueled
vehicles are sources of primary and secondary parti culate matter. The rates of emission and the
composition of particles emitted by motor vehicles have been measured using dynamometers
with samples collected directly in the exhaust of individual vehicles (e.g., Lang et al., 1982) or at
the vents of inspection facilities (e.g., Watson et al., 1994a); or in tunnels and along open
roadways (e.g., Pierson and Brachaczek, 1983; Szkarlat and Japar, 1983). Each approach has its
5-19
-------
merits and limitations and each approach is best used to augment the other. The principal
components emitted by diesel and gasoline fueled vehicles are organic carbon (OC) and
elemental carbon (EC) as shown in Table 5-4. As can be seen, the variability among entries for
an individual fuel type is large and overlaps that found between different fuel types. On
average, the abundance of elemental carbon is larger than that of organic carbon in the exhaust
of diesel vehicles, while organic carbon is the dominant species in the exhaust of gasoline fueled
vehicles. There appears to be a tendency for emissions of elemental carbon to increase relative
to emissions of organic carbon for gasoline fueled vehicles as simulated driving conditions are
changed from a steady 55 km /hr to those in the Federal Test Procedures (FTP's). Also shown
are the results of sampling from mixed vehicle types along roadsides and in tunnels.
The results shown in Table 5-4 were obtained during the late 1980's, and, so, the results
may not be entirely representative of current vehicles. Examples of data for the trace element
composition of motor vehicle emissions obtained in Phoenix, AZ are shown in Table 5-5. SO2
emissions are also shown in relation to the mass of fine particles emitted. As can be seen, small
quantities of soluble ions such as SO4 and NH4+ are emitted. The ammonium may be emitted as
the result of an improperly functioning catalytic converter, or may simply be the result of
contamination during sample handling and analysis. Four fractions are given for the organic
carbon fraction and three for elemental carbon. These refer to abundances measured at different
temperatures in a thermographic analysis. Temperatures for OC1, OC2, OC3, and OC4 are 120
°C, 250 °C, 450 °C, and 550 °C, respectively; and, forECl, EC2, ECS, they are 550 °C, 700
°C, and 800 °C, respectively, in He/2% O2. The abundances of trace elements are all quite low,
with most being less than 1%. It is not clear what the source of the small amount of Pb seen in
the auto exhaust profile is. It is extremely difficult to find suitable tracers for automotive
exhaust since Pb has been removed from gasoline. However, it should also be remembered that
restrictions in the use of leaded gasoline have resulted in a dramatic lowering of ambient Pb
levels. Huang et al. (1994) attempted to identify marker elements in motor vehicle emissions,
based on sampling the exhaust of 49 automobiles. They proposed that the combination of Zn,
Br, and
5-20
-------
TABLE 5-4. FRACTIONAL ORGANIC AND ELEMENTAL CARBON
ABUNDANCES IN MOTOR VEHICLE EMISSIONS
Fuel Type
Diesel
Denver, COa
Los Angeles, CAa
Bakersfield, CAb
Phoenix, AZb
Unleaded gasoline
Denver, COa
Los Angeles, CAC
Los Angeles, CAa
Phoenix, AZb
Leaded gasoline
Denver, COa
Los Angeles, CAC
Los Angeles, CAa
Mixed (tunnel and roadside^
Denver, CO
Los Angeles, CAd
Phoenix, AZ
Organic Carbon
23 ± 8%
36 ± 3%
49 ± 13%
40 ± 7%
76 ± 29%
93 ± 52%
49 ± 10%
30 ± 12%
67 ± 23%
52 ± 4%
3 1 ± 20%
50 ± 24%
38 ±6%
39 ± 19%
Elemental Carbon
74 ±21%
52 ± 5%
43 ± 8%
33 ± 8%
18 ± 11%
5 ± 7%
39 ± %
14 ±8%
16 ±7%
13 ± 1%
15 ±2%
28 ± 19%
38 ± 5%
36 ± 11%
Ne
O
2
O
8
8
11
11
9
O
3
O
3
Sources
1,2
3, 4, 5, 6
7
8
1,2
3,4,5,6
3,4,5,6
8
1,2
3, 4, 5, 6
3, 4, 5, 6
1,2
3
8
Sources: (1) Watson et al. (1990a), (2) Watson et al. (1990b), (3) Cooper et al. (1987), (4) NBA (1990a),
(5) NBA (1990b), (6) NBA (1990c), and (7) Houck et al. (1989), cited in (8) Watson et al. (1994a).
Notes: (a) Modified Federal Test Procedures followed in dynamometer tests; (b) Roof monitoring at
inspection station; (c) 55 km/hr steady speed in dynamometer tests; (d) Rt. 1 tunnel at LA airport,
(e) N = Number of samples.
Sb could be used for this purpose. However, the relative abundances of these species in
automobile exhaust were shown to be highly variable, implying that other sources of these
elements may limit their usefulness as automotive tracers in some locations. To minimize
5-21
-------
TABLE 5-5. PHOENIX PM2, MOTOR VEHICLE EMISSIONS PROFILES (% MASS)
Chemical Species
N03-
so/-
NH4+
oc
OC1
OC2
OC3
OC4
EC
EC1
EC2
ECS
Al
Si
P
S
Cl
K
Ca
Ti
Cr
Mn
Fe
Cu
Zn
Sb
Ba
La
Pb
SO/
Auto
3. 9 ±2.9
2.3 ±1.3
1.7±1.0
30.1 ±12.3
11.3±3.5
9.2 ±6.8
4.6 ±2.2
3.5±1.5
13. 5 ±8.0
11.7 ±7.2
3.1 ±1.6
0.15 ±0.30
0.41 ±0.20
1.64 ±0.88
0.11 ±0.07
1.01 ±0.48
0.34 ±0.32
0.25 ±0.14
0.71 ±0.41
0.07 ±0.13
0.02 ±0.01
0.10 ±0.04
0.68 ±0.42
0.07 ± 0.06
0.27 ± 0.22
0.02 ±0.13
0.06 ±0.40
0.15±0.51
0.16 ±0.07
32.8 ±13. 9
Diesel
0.31 ±0.40
2.4 ±1.0
0.87 ±0.13
40.1 ±6.6
21.0±6.3
9.1 ±1.9
5.9±1.3
4.0±1.5
32.9 ±8.0
4.4±1.3
27. 9 ±5.6
0.69 ±0.82
0.17±0.12
0.46 ±0.18
0.06 ±0.06
1.24 ±0.28
0.03 ± 0.06
0.04 ±0.03
0.16 ±0.06
0.00 ±0.15
0.00 ±0.01
0.01 ±0.01
0.16 ±0.07
0.01 ±0.01
0.07 ± 0.02
0.01 ±0.14
0.14 ±0.47
0.18±0.59
0.01 ±0.03
66.9 ±24.0
Source: Watson et al. (1994a).
Note: Elemental abundances <0.01% (V, Co, Ni, Ga, As, Se, Br, Rb, Sr. Y, Zr, Mo, Pd, Ag, Cd, In, Sn, Au, Hg, Tl,
U) in XRF analyses excluded; OC = organic carbon; EC = elemental carbon.
"Relative to total PM2,.
5-22
-------
errors arising from the loss of Br from filters, samples should be analyzed as soon as possible
after collection (O'Connor et al., 1977).
The chemical mechanisms responsible for the formation of carbonaceous particles in diesel
engines are not well established but are thought to involve the intermediate formation of
polycyclic aromatic hydrocarbons, or PAH's (U.S. Environmental Protection Agency, 1993).
Elemental carbon particles may be formed by the polymerization of gaseous intermediates
adsorbed on a core of refractory material. The particles are in the form of chain or cluster
agglomerates at temperatures above 500 °C. At temperatures below 500 °C, high molecular
weight organic compounds condense on the carbon chain agglomerates. Roughly 10-40% of
particulate emissions from diesels are extractable into organic solvents (National Research
Council, 1982). In a typical profile, 50% of the extract is composed of aliphatic hydrocarbons
with 14-35 C atoms and alkyl substituted benzenes; 4% are PAH's and heterocycles; and about
6% are PAH oxidation products including a small fraction of nitro-PAH's. The highly polar
fraction of the organic emissions has not been as well characterized (Johnson, 1988). Inorganic
compounds such as sulfuric acid are also produced in diesel engines (Truex et al., 1980).
Particulate matter is also formed in internal combustion engines as the result of the
incomplete combustion of gasoline with contributions from engine oil. The particles consist
essentially of a solid carbon core with a coating of organic compounds, sulfate, and trace
elements. The composition of PAH's, oxy-PAH's and their alkyl homologues in tailpipe
emissions from gasoline fueled vehicles is similar to that produced in diesel engines (Behymer
and Kites, 1984). Particles produced by gasoline fueled vehicles range from 0.01 to 0.1 jim in
diameter with a peak at around 0.02 jim, while the majority of particles in diesel exhaust range
from 0.1 to 1.0 jim with a peak at around 0.15 |im (U.S. Environmental Protection Agency,
1993).
The concentrations of particulate matter and total hydrocarbons in the exhaust of gasoline
fueled vehicles were found to be roughly correlated with each other by Hammerle et al. (1992).
Emission factors for particulate matter in the exhaust of gasoline fueled vehicles range from
0.011 g/km for light duty vehicles to 0.12 g/km for heavy duty gasoline vehicles, and from 0.23
g/km in the exhaust of diesel passenger vehicles to 1.20 g/km for heavy duty diesel vehicles
(Radwan, 1995). These values are based on characteristics of the motor vehicle fleet in 1990.
5-23
-------
As mentioned before, the composition of automotive emissions is sampled using individual
vehicles on chassis dynamometers or by collecting aerosol samples along roadsides or in tunnels.
The control over operating characteristics by using dynamometers allows the development of
models which can predict emissions on the basis of variables such as vehicle make and age and
driving cycle. The representativeness of dynamometer test data can be questioned if volunteered
vehicles, as opposed to randomly selected vehicles, are sampled. In addition, measuring
emissions from individual vehicles is also costly and the sample numbers are usually small, as
reflected in the small number of samples shown in Table 5-4. Moreover, a number of driving
practices are not reflected in the Federal Test Procedures leading to significant underestimates of
emissions of CO and hydrocarbons (Calvert et al., 1993). It is still not clear what effects
superemitters and off-cycle driving practices have on particle emissions rates. If the relation
between particulate matter and hydrocarbon emissions discussed above is representative of the
vehicle fleet, the effects could be substantial. Hansen and Rosen (1990) measured the ratio of
light-absorbing carbon to CO2 in the exhausts of 60 gasoline fueled vehicles. They found a
factor of 250 difference between the highest and lowest ratios measured. Larger scale studies
designed to assess the variability of paniculate emissions from motor vehicles are lacking.
Roadside and tunnel measurements sample large numbers of vehicles of different types and
have demonstrated their potential for validating the predictions of emissions models. However,
the extent to which traffic conditions in the tunnel reflect those in the situation under study must
be defined for the results to be considered representative. The same considerations can be
extended to dynamometer studies and to open-road studies along road segments. Results from
some tunnel studies are of limited usefulness because they have been obtained under highway
driving conditions which may not be representative of the conditions found in most urban and
suburban areas. Additional uncertainties result from resuspended dust in using tunnel and
roadside studies to characterize motor vehicle emissions. However, methods are available for
estimating contributions from tire wear (Pierson and Brachaczek, 1974, 1976). Remote
measurements of elemental carbon in the exhaust plumes of individual vehicles (Hansen and
Rosen, 1990) can overcome many of these difficulties, but the method cannot yet be applied to
aerosol constituents other than elemental carbon.
5-24
-------
5.2.4 Biomass Burning
In addition to fossil fuels, biomass in the form of wood may be burned in forest fires or as
fuel for heating or cooking. At first glance these two broad categories might seem to serve to
distinguish between natural and anthropogenic sources. However, many forest fires result from
human intervention, either deliberately through prescribed burning in forest management or
accidentally through the improper disposal of flammable material or fugitive sparks
(e.g., Andreae, 1991). On the other hand, human intervention also suppresses lightning
triggered fires. Not enough data is available to assess the effects of humans on forest fires,
except for land clearing for agriculture. In contrast to the mobile and stationary sources
discussed earlier, emissions from biomass burning in woodstoves and forest fires are strongly
seasonal and can be highly episodic within their peak emissions seasons. Burning fuelwood is
confined mainly to the winter months and is acknowledged to be a major source of ambient air
particulate matter in the northwestern United States during the heating season. Forest fires
mainly occur during the driest seasons of the year in different areas of the country and are
especially prevalent during prolonged droughts.
An example of the composition of fine particles (PM25) produced by woodstoves is shown
in Figure 5-4. These data were obtained in Denver during the winter of 1987-1988 (Watson and
Chow, 1994). As was the case for motor vehicle emissions, organic and elemental carbon are
the major components of particulate emissions from wood burning. It should be remembered
that the relative amounts shown for organic carbon and elemental carbon will vary with the type
of stove, the stage of combustion and the type and condition of the fuelwood. Potassium (K) is
by far the major trace element found in woodstove emissions (Watson and Chow, 1994), making
it suitable for use as a tracer for vegetation burning (Lewis et al., 1988). Fine particles are
dominant in studies of wood burning emissions. For instance, the mass median diameter of
wood-smoke particles was found to be about 0.17 jim in a study of the emissions from burning
hardwood, softwood and synthetic logs (Dasch, 1982).
5-25
-------
10™I""™™"""3™""
Chemical Compound
Figure 5-4. Chemical abundances for PM2 5 emissions from wood burning in Denver, CO.
Solid bars represent fractional abundances, and the error bars represent
variability in species abundances. Error bars represent detection limits when
there are no solid bars.
Source: Watson and Chow (1994).
Measurements of aerosol composition, size distributions, and aerosol emissions factors
have been made in biomass burning plumes either on towers (Susott et al., 1991) or aloft on
fixed wing aircraft (e.g., Radke et al., 1991) or on helicopters (e.g., Cofer et al., 1988). As was
found for woodstove emissions, the composition of biomass burning emissions is strongly
dependent on the stage of combustion (i.e., flaming, smoldering, or mixed), and the type of
vegetation (e.g., forest, grassland, scrub). Over 90% of the dry mass in particulate biomass
burning emissions is composed of organic carbon (Mazurek et al., 1991). Ratios of organic
carbon to elemental carbon are highly variable ranging from 10:1 to 95:1, with the highest ratio
found for smoldering conditions and the lowest for flaming conditions. Ambient particle
concentrations were about two mg/m3 during the measurement period. Available measurements
suggest that K is by far the most abundant trace element in biomass burning plumes. Although
there is considerable inter-sample variation, results from tower samples also suggest that S, Cl,
5-26
-------
and K are highest during flaming stages, while Al, Si, Ca, and Fe tend to increase during the
smoldering phase (Susott et al., 1991). Emissions factors for total particulate emissions increase
by factors of two to four in going from flaming to smoldering stages in the individual fires
studied by Susott et al. (1991). These measurements were made when ambient particle
concentrations ranged from 15 to 40 mg/m3.
Particles in biomass burning plumes from a number of different fires were found to have
three distinguishable size modes, namely a nucleation mode, an accumulation mode, and a
coarse mode (Radke et al., 1991). Based on an average of 81 samples, approximately 70% of
the mass was found in particles < 3.5 jim in aerodynamic diameter. The fine particle
composition was found to be dominated by tarlike, condensed hydrocarbons and the particles
were usually spherical in shape. Additional information for the size distribution of particles
produced by vegetation burning was shown in Figure 5-2.
5.2.5 Sea-Salt Production and Other Natural Sources of Aerosol
Although sea-salt aerosol production is confined to salt water bodies, it is included here
because many marine aerosols can exert a strong influence on the composition of the ambient
aerosol in coastal areas. In some respects, the production of sea-salt aerosols is like that of
windblown dust in that both are produced by wind agitation of the surface. The difference
between the two categories arises because sea-salt particles are produced from the bursting of air
bubbles rising to the sea surface. Air bubbles are formed by the entrainment of air into the water
by breaking waves. The surface energy of a collapsing bubble is converted to kinetic energy in
the form of a jet of water which can eject drops above the sea surface. The mean diameter of the
jet drops is about 15% of the bubble diameter (Wu, 1979). Bubbles in breaking waves range in
size from a few |im to several mm in diameter. Field measurements by Johnson and Cooke
(1979) of bubble size spectra show maxima in diameters at around 100 jim, with the bubble size
distribution varying as (d/d0)"5 with d0 = 100 //m.
Since the water jet receives its water from the surface layer, which is enriched in organic
compounds, the aerosol drops are composed of this organic material in addition to sea salt (about
3.5% by weight in sea water). Na+ (30.7%),C1' (55.0%), SO4= (7.7%), Mg2+ (3.6%), Ca2+
(1.2%), K+ (1.1%), HCO3" (0.4%), and Br'(0.2%) are the major ionic species by mass in sea
water (Wilson, 1975). The composition of the marine aerosol also reflects the occurrence of
5-27
-------
displacement reactions which enrich sea-salt particles in SO4" and NO3" while depleting them of
Cl" and Br". As the drops travel upward above the water surface, they encounter lower relative
humidities and lose water until they come into equilibrium with their environment. The
resulting marine aerosol size distribution reflects the processes of coagulation, coalescence, and
sedimentation.
Seasalt is concentrated in the coarse size mode with a mass median diameter of about 7 jim
for samples collected in Florida, the Canary Islands and Barbados (Savoie and Prospero, 1982).
The size distribution of sulfate is distinctly bimodal. Sulfate in the coarse mode is derived from
sea water but sulfate in the submicron aerosol arises from the oxidation of dimethyl sulfide
(CH3SCH3) or DMS. DMS is produced during the decomposition of marine micro-organisms.
DMS is oxidized to MSA (methane sulfonic acid) a large fraction of which is oxidized to sulfate
(e.g., Hertel et al., 1994).
Apart from sea spray, other natural sources of particles include the suspension of organic
debris and volcanism. Particles are released from plants in the form of seeds, pollen, spores, leaf
waxes and resins, ranging in size from 1 to 250 jim (Warneck, 1988). Fungal spores and animal
debris such as insect fragments are also to be found in ambient aerosol samples in this size
range. While material from all the foregoing categories may exist as individual particles,
bacteria are usually found attached to other dust particles (Warneck, 1988). Smaller bioaerosol
particles include viruses, individual bacteria, protozoa, and algae (Matthias-Maser and Jaenicke,
1994). In addition to natural sources, other sources of bioaerosol include industry (e.g., textile
mills), agriculture, and municipal waste disposal (Spendlove, 1974).
Trace metals are emitted to the atmosphere from a variety of sources such as sea spray,
wind blown dust, volcanoes, wild fires and biotic sources (Nriagu, 1989). Biologically mediated
volatilization processes (e.g., biomethylation) are estimated to account for 30-50% of the
worldwide total Hg, As, and Se emitted annually, whereas other metals are derived principally
from pollens, spores, waxes, plant fragments, fungi, and algae. It is not clear, however, how
much of the biomethylated species are remobilized from anthropogenic inputs. Median ratios of
the natural contribution to globally averaged total sources for trace metals are estimated to be
0.39 (As), 0.15 (Cd), 0.59 (Cr), 0.44 (Cu), 0.41 (Hg), 0.35 (Ni), 0.04 (Pb), 0.41 (Sb), 0.58 (Se),
0.25 (V), and 0.34 (Zn), suggesting a not insignificant natural source for many trace elements. It
5-28
-------
should be noted though that these estimates are based on emissions estimates which have
uncertainty ranges of an order of magnitude.
5.3 SOURCES OF SECONDARY PARTICULATE MATTER (SULFUR
DIOXIDE, NITROGEN OXIDES, AND ORGANIC CARBON)
Secondary particulate matter is an important contributor to suspended particle mass.
Sulfate is formed by the oxidation of SO2, nitrate by the oxidation of NO2, and aerosol organic
carbon species by the oxidation of a large number of precursors. Thus, the formation rate of a
substantial fraction of aerosol mass is given by a complex function of both emission rates of
precursor gases and the rates of photochemical processes in the atmosphere. In order to use
precursor emissions estimates effectively, however, it is necessary to understand the nature of
the processes that cause them to convert to particulate matter. Mechanisms for the oxidation of
SO2 to SO4=, and NO2 to NO3", have been discussed in Chapter 3. Both species are oxidized
during daytime in the gas phase by hydroxyl (OH) radicals. At night, NOX is also oxidized to
nitric acid by a sequence of reactions initiated by O3, that include nitrate radicals (NO3) and
dinitrogenpentoxide (N2O5). SO2 is also oxidized by heterogeneous reactions occurring in films
of atmospheric particles and in cloud and fog droplets. Data for primary and secondary
components of aerosol mass at a number of locations across the United States can be found in
Chapter 6.
While the mechanisms and pathways for forming inorganic secondary particulate matter
are fairly well known, those for organic secondary aerosol are not well understood. Numerous
precursors participate in these conversions, and the rates at which these convert from gas to
particles are highly dependent on the concentrations of other pollutants and meteorological
conditions. Pandis et al. (1992) identified three mechanisms for secondary organic PM
formation: (1) condensation of oxidized end-products of photochemical reactions (e.g., ketones,
aldehydes, organic acids, and hydroperoxides); (2) adsorption of organic gases onto existing
solid particles (e.g., polycyclic aromatic hydrocarbons); and (3) dissolution of soluble gases
which can undergo reactions in particles (e.g., aldehydes). The first and third mechanisms are
expected to be of major importance during the summertime when photochemistry is at its peak.
The second pathway can be driven by diurnal and seasonal temperature and humidity variations
5-29
-------
at any time of the year. Turpin and Huntzicker (1991) and Turpin et al. (1991) provided strong
evidence that secondary PM formation occurs during periods of photochemical ozone formation
in Los Angeles.
Haagen-Smit (1952) first demonstrated that hydrocarbons irradiated in the presence of NOX
produce light scattering aerosols. Results of later studies summarized by Altshuller and Bufalini
(1965) indicated that aerosols are produced by the irradiation of mixtures of NOX and numerous
six-carbon and higher molecular weight acyclic and cyclic olefms and aromatic hydrocarbons.
Cyclic olefms were shown to be more effective in aerosol formation than acyclic olefms of
similar molecular weight by Stevenson et al. (1965). The possibility that aerosols might be
formed from biogenic hydrocarbon emissions was investigated by Went (1960) and Rasmussen
and Went (1965). Analyses of the aerosol produced from the photooxidation of a-pinene and
NOX mixtures indicated the presence of pinonic acid and norpinonic acid (Wilson et al., 1972).
Numerous smog chamber studies of the formation of secondary organic aerosol from the
photooxidation of terpene precursors have been performed since these earlier studies. A study of
the reaction of a-pinene and P-pinene with O3 by Hatakeyama et al. (1989) obtained aerosol
carbon yields (mass of aerosol carbon produced per mass of C reacted), or ACY's, of 18% and
14%, respectively, for HC levels ranging from 10-120 ppb C. In this study, pinonaldehyde,
pinenic acid, nor-pinonaldehyde, and nor-pinonic acid accounted for less than 10% of the
aerosol yield from the reaction of a-pinene. Hatakeyama et al. (1991) subsequently obtained
ACY's of 56 ± 4% and 79 ± 8% following the reaction of a-pinene and P-pinene, respectively,
for initial HC levels of 820-3170 ppb C and NOX levels of 210-2550 ppb. Pandis et al. (1991)
obtained ACY's ranging from 0.1 to 8% for the oxidation of P-pinene for HC levels ranging
from 20-250 ppb C and NOX levels ranging from 39 to about 700 ppb. Zhang et al. (1992)
obtained ACY's ranging from 0 to 5.3% for HC levels ranging from 37-582 ppb C and NOX
levels ranging from 31-380 ppb for the oxidation of a-pinene. Results from the above studies
showed that aerosol yields strongly depend on the initial concentration of terpenes and the ratio
of hydrocarbons (HC) to NOX in the reaction chamber. However, Hooker et al. (1985) did not
find a significant dependence of aerosol yield on initial HC abundance for HC levels ranging
from 3.1-50 ppb C. Their approach differed from that used in all of the above studies because
they used 14C-a-pinene. Of the 14C-a-pinene which reacted, 38-68% was found in aerosol
5-30
-------
products, 6-20% was found in gas phase products, and 11-29% was lost to the walls of their
reaction chamber.
After reaction of the a-pinene with OH radicals or O3, the radical product will add O2 to
form a peroxy radical. Zhang et al. (1992) proposed that the peroxy radical may react with NO
initiating a series of reactions forming pinonaldehyde, which may condense depending on its
concentration, or the peroxy radical may react with HO2 or other free radicals to form aerosol
products. The inhibition of the second pathway by the addition of NO was proposed by Zhang
et al. (1992) to explain the decrease of aerosol yield with added NO. They also suggested that
the dependence of aerosol yield on initial HC concentration arises because the concentration of
pinonaldehyde can more easily exceed its saturation value and the rate of formation of aerosol
products in the other pathway will also increase.
Pandis et al. (1991) found no aerosol products formed in the photooxidation of isoprene,
although they and Zhang et al. (1992) found that the addition of isoprene to reaction mixtures
increased the reactivity of the systems studied. Based on their experimental results and the high
ratio of terpene to NOX concentration ratios found in rural and remote areas, Zhang et al. (1992)
suggested that the upper limits for aerosol yields they obtained should be used in estimating the
aerosol yields from the oxidation of biogenic hydrocarbons.
The aerosol forming potentials of a wide variety of individual anthropogenic and biogenic
hydrocarbons were compiled by Pandis et al. (1992) based mainly on estimates made by
Grosjean and Seinfeld (1989) and data from Pandis et al. (1991) for p-pinene and Izumi and
Fukuyama (1990) for aromatic HC's. The estimates given by Pandis et al. (1992) were
converted to aerosol carbon yields below. Examples of compounds with zero ACY's are all Cr
C7 alkanes, all C2-C6 acyclic alkenes, benzene, and aldehydes; examples of compounds with
lowest ACY's (< 2.0%) are C8-C10 alkanes, C6-C8 cycloalkanes, C7-C9 acyclic alkenes, C5 cyclic
alkenes and p-xylene; examples of compounds with intermediate values (2.0%-4.0%) are Cn-C14
alkanes, C9-C10 cycloalkanes, alkyl benzenes other than p-xylene, C10-C13 alkenes and C6+
cycloalkenes; and examples of compounds with high values (>4.0%) are C15+ alkanes, Cn+
cycloalkanes, C14+ cyclic alkenes and monoterpenes.
Studies of the production of secondary OC in ambient air have focussed on the Los
Angeles Basin. Based on aerosol yields shown above, Pandis et al. (1991) suggested that about
1-4 tons day"1 of secondary OC in the Los Angeles basin is formed from the oxidation of
5-31
-------
monoterpenes which are emitted at the rate of 10-40 tons day"1. This estimate may be compared
to the secondary OC production rate of 7.5 tons day"1 estimated to result from the oxidation of
anthropogenic hydrocarbons which are emitted at the rate of 1200 tons day"1 (Grosjean and
Seinfeld, 1989). The overall yield of secondary OC from anthropogenic sources in this example
is about 0.6%. Pandis et al. (1991) also proposed that most of the secondary OC in highly
vegetated urban areas such as Atlanta is produced by the oxidation of monoterpenes.
As part of the Southern California Air Quality Study (SCAQS), Turpin and Huntzicker
(1991) measured elemental and organic carbon at Claremont, CA in the summer of 1987 with an
in situ carbon analyzer with 2 hour time resolution. During an air pollution episode centered on
August 28, 1987, airmass trajectories arriving at Claremont were directed eastward (i.e., inland
from the coast), allowing the entrainment of substantial hydrocarbon precursors during transit.
Peak OC concentrations (23 |ig/m3) and highest OC to EC ratios (4.6 ± 0.4) occurred together at
Claremont from about 1500 to 1700 PDT. In addition, correlations between EC and OC were
low throughout the day (R2 =0.38). Turpin and Huntzicker (1991) also measured OC and EC
concentrations at Long Beach in November of 1987 with the same instrumentation. On the basis
of these data, they suggested that OC to EC ratios of 2.2 ± 0.7 are characteristic of primary OC
in the Los Angeles area.
Pandis et al. (1992) constructed a Lagrangian trajectory model to simulate the chemical
formation, transport and deposition of secondary OC during the August episode. They used
estimates of aerosol yields from HC oxidation compiled by Grosjean and Seinfeld (1989),
updated as necessary (e.g., Pandis et al., 1991) along with estimates of daily emissions, to
predict that 28% of the peak secondary OC on Aug. 28 at Claremont resulted from the oxidation
of toluene, 38% from other aromatic HC's, 9% from biogenic HCs, 21% from alkanes and
cycloalkanes, and the remaining 4% from other species. Values were somewhat different on a
daily average basis (19% from toluene, 46% from other aromatic HC's, 16% from biogenic
HC's, 15% from alkanes, and 4% from alkenes). There was reasonable agreement with the data
of Turpin and Huntzicker (1991) throughout most of the day, but calculated peak secondary OC
levels ( ~5 |ig/m3) were about half those inferred by Turpin and Huntzicker (1991). A
combination of factors could have contributed to this underprediction including errors in
emissions, deposition rates, chemical reaction rate data and aerosol yields. In general, the
calculated secondary OC represented 15-27% of the daily average total OC at inland locations
5-32
-------
(Burbank, Claremont, Azusa, and Rubidoux) on August 28, 1987 and 5-19% of the average total
OC at coastal sites.
Attempts were made during SCAQS to determine the composition of the organic carbon
fraction of the ambient aerosol. Organic nitrates were measured on size segregated samples
collected on zinc selenide disks which were later analyzed by transmission FTIR by Mylonas et
al. (1991). Concentrations of organic nitrates in the particle phase ranged from 0.8 to 4.0 |ig/m3,
with maximum mass loadings in the 0.05 to 0.075 |im and 0.12 to 0.26 jim size ranges.
Concurrently, Pickle et al. (1990) used infrared spectroscopy to measure the total abundance of
compounds containing carbonyl groups and aliphatic compounds. Maximum absorption at
wavelengths characteristic of carbonyl groups was observed for particles in the 0.12 to 0.26 jim
and 0.5 to 1.0 jim size ranges. These results suggest that carbonyl compounds are largely of
secondary origin and that IR absorption by aliphatic compounds in particles smaller than 0.12
|im was correlated directly with automotive emissions.
Kao and Friedlander (1995) examined the statistical properties of a number of PM
components in the South Coast Air Basin. They found that the concentrations of non-reactive,
primary components of PM10 have approximately log normal frequency distributions and
constant values of geometric standard deviations (GSDs) regardless of source type and location
within their study area. However, aerosol constituents of secondary origin (e.g., SO4=, NH4+, and
NO3") were found to have much higher GSD's. Surprisingly, the GSD's of organic (1.87) and
elemental (1.74) carbon were both found to be within la (0.14) of the mean GSD (1.85) for non-
reactive primary species, compared to GSD's of 2.1 for sulfate, 3.5 for nitrate, and 2.6 for
ammonium. These results suggest that most of the OC seen in ambient samples is of primary
origin. Pinto et al. (1995) found similar results for data obtained during the summer of 1994.
Further studies are needed to determine if these relations are valid at other locations and to
determine to what extent the results might be influenced by the evaporation of volatile
constituents after sampling.
It must be emphasized that the inferences drawn from field studies in the Los Angeles
Basin are unique to that area and cannot be extrapolated to other areas of the country.
In addition, there is a high degree of uncertainty associated with all aspects of the calculation of
secondary OC concentrations which is compounded by the volatilization of OC during and after
sampling. Grosjean and Seinfeld (1989) derived a factor of five range in estimates of production
5-33
-------
rates of secondary OC in the Los Angeles area by comparison of results obtained from four
different methods. Aerosol yields from the oxidation of individual hydrocarbons reported by
different investigators vary by an order of magnitude (Grosjean and Seinfeld, 1989). Significant
uncertainties always arise in the interpretation of smog chamber data because of wall reactions.
Limitations also exist in extrapolating the results of smog chamber studies to ambient conditions
found in urban airsheds and forest canopies. Concentrations of terpenes and NOX are much
lower in forest canopies (Altshuller, 1983) than are commonly used in smog chamber studies.
The identification of aerosol products of terpene oxidation has not been a specific aim of field
studies, making it difficult to judge the results of model calculations of secondary OC formation.
Uncertainties may also arise because of the methods used to measure biogenic hydrocarbon
emissions. Khalil and Rasmussen (1992) found much lower ratios of terpenes to other
hydrocarbons (e.g., isoprene) in forest air than were expected, based on their relative emissions
strengths and rate coefficients for reaction with OH radicals and O3. They offered two
explanations, either the terpenes were being rapidly removed by some heterogeneous process or
emissions were artificially enhanced by feedbacks caused by the bag enclosures they used. If the
former consideration is correct, then the production of aerosol carbon from terpene emissions
could be substantial; if the latter is correct, then terpene emissions could have been
overestimated by the techniques used.
5.4 EMISSIONS ESTIMATES FOR PRIMARY PARTICULATE
MATTER AND SO2, NOX, AND VOCs IN THE UNITED STA
The emissions of a pollutant can be expressed by the following equation:
E ? A.Fi.(l-Ceff,i) (5-1)
where E is the total emissions rate from all sources; A; is the activity rate for source i; F; is the
emissions factor for the production of the pollutant by source i; and Ceffi is the fractional
efficiency of control devices used by source i. Activity rates relevant to the entries shown in
Tables 5-6 to 5-10 might refer to the electricity generated by power plants, the amount of coke
produced by a coke oven, the distance travelled by motor vehicles, the amount of biomass
5-34
-------
consumed by forest fires per year, etc. The mass of pollutant emitted per unit activity of a
source is then expressed in terms of an emissions factor (e.g., amount of NOX emitted per
kw-hour of energy generated or per vehicle mile travelled). Emissions factors are given in
compilations (e.g., AP-42 [U.S. Environmental Protection Agency, 1995a]) or are calculated by
emissions models, which include a number of variables which can affect emissions. Examples
include the U.S. Environmental Protection Agency's PARTS model for estimating particulate
motor vehicle emissions, and BEIS which is used to calculate emissions of hydrocarbons from
vegetation (Geron et al., 1994). The product of A; x F; yields an estimate of the uncontrolled
emissions from a particular source i. These are then multiplied by a factor incorporating the
effects of any control devices that might be used. It is acknowledged that control equipment
breaks down, and its efficiency might not be maintained over its lifetime of operation.
Therefore, the optimum efficiencies of control devices are multiplied by a rule effectiveness
factor. The default value for the rule effectiveness factor is taken to be 0.8 in the inventory
calculations, unless a better factor can be justified (U.S. Environmental Protection Agency,
1989). Equation 5-1 was used in the preparation of the emissions inventories shown in
Tables 5-6 through 5-10. Further details about collection and reporting methods may be found
in the National Emissions Inventory Trends data base (U.S. Environmental Protection Agency,
1994).
Table 5-6 shows the primary PM10 emissions estimated for the period of 1985 through
1993 using the National Emissions Inventory Trends data base (U.S. Environmental Protection
Agency, 1994). Emissions are shown in the original units used in their calculation. A short ton
is equal to 2,000 pounds or 9.08 x 10s gm. Between 1985 and 1993, PM10 emissions from
stationary and mobile sources decreased almost 10 percent. During this period, contributions
from highway vehicles decreased by 27 percent, reflecting emissions controls on diesel vehicles.
Contributions from industrial fuel production
5-35
-------
TABLE 5-6. NATIONWIDE PRIMARY PM,n EMISSION ESTIMATES FROM
10
MOBILE AND STATIONARY SOURCES, 1985 TO 1993
(Thousands short tons/year)
Source Category
Fuel Combustion - Electric Utilities
Fuel Combustion - Industrial
Fuel Combustion - Other
Chemical and Allied Product Manufacturing
Metals Processing
Petroleum and Related Industries
Other Industrial Processes
Solvent Utilization
Storage and Transport
Waste Disposal and Recycling
Highway Vehicles
Off-Highway
Total
1985
284
234
896
67
147
32
317
2
57
279
271
368
2,953
1986
289
231
902
68
137
31
321
2
56
275
265
372
2,949
1987
282
226
910
68
131
30
314
2
54
265
261
350
2,893
1988
278
230
918
73
141
29
314
2
54
259
256
387
2,942
1989
278
229
922
74
142
28
308
2
54
251
253
372
2,909
1990
291
228
930
74
140
28
306
2
54
242
239
372
2,907
1991
253
229
942
72
136
28
300
2
53
245
223
367
2,849
1992
255
223
819
75
137
27
303
2
53
246
210
379
2,729
1993
270
219
723
75
141
26
311
2
55
248
197
395
2,661
Note: The sums of sub-categories may not equal total due to rounding (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1994).
-------
TABLE 5-7. MISCELLANEOUS AND NATURAL SOURCE PRIMARY PM10 EMISSION ESTIMATES,
1985 TO 1993
(Thousands short tons/year)
Source Category
Fugitive Dust
Unpaved roads
Paved roads
Construction/mining and quarrying
Agriculture and Forestry
Agricultural crops
Agricultural livestock
Other Combustion
-------
TABLE 5-8. NATIONWIDE SULFUR OXIDES EMISSION ESTIMATES, 1984 TO 1993
Source Category
Fuel Combustion - Electric Utilities
Fuel Combustion - Industrial
Fuel Combustion - Other
Chemical and Allied Product
Manufacturing
Metals Processing
Petroleum and Related Industries
Other Industrial Processes
Y1 Solvent Utilization
OJ
oo
Storage and Transport
Waste Disposal and Recycling
Highway Vehicles
Off-Highway
Miscellaneous
Total
1984
16,023
2,723
728
229
1,387
707
923
0
0
25
445
198
9
23,396
1985
16,273
3,169
578
456
1,042
505
425
1
4
34
446
208
7
23,148
1986
15,701
3,116
611
432
888
469
427
1
4
35
449
221
7
22,361
1987
15,715
3,068
663
425
616
445
418
1
4
35
457
233
7
22,085
(Thousands
1988
15,990
3,111
660
449
702
443
411
1
5
36
468
253
7
22,535
short tons/year)
1989
16,218
3,086
623
440
657
429
405
1
5
36
480
267
7
22,653
1990
15,898
3,106
597
440
578
440
401
1
5
36
480
265
14
22,261
1991
15,78
4
3,139
608
442
544
444
391
1
5
36
478
266
11
22,14
9
1992
15,41
7
2,947
600
447
557
417
401
1
5
37
483
273
10
21,59
2
1993
15,83
6
2,830
600
460
580
409
413
1
5
37
438
278
11
21,88
8
Note: The sums of sub-categories may not equal the totals, due to rounding (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1994).
-------
TABLE 5-9. NATIONWIDE NOxa EMISSION ESTIMATES, 1984 TO 1993
(Thousands short tons/year)
Source Category
Fuel Combustion - Electric Utilities
Fuel Combustion - Industrial
Fuel Combustion - Other
Chemical and Allied Product Manufacturing
Metals Processing
Petroleum and Related Industries
Other Industrial Processes
Solvent Utilization
(^ Storage and Transport
^ Waste Disposal and Recycling
Highway Vehicles
Off-Highway
Miscellaneous
Total
1984
7,268
3,415
670
161
54
70
203
0
0
90
8,387
2,644
210
23,17
2
1985
6,916
3,209
701
374
87
124
327
2
2
87
8,089
2,734
201
22,85
3
1986
9,909
3,065
694
381
80
109
328
3
2
87
7,773
2,777
202
22,40
9
1987
7,128
3,063
710
371
76
101
320
3
2
85
7,662
2,664
203
22,38
6
1988
7,530
3,187
737
398
82
100
315
3
2
85
7,661
2,914
206
23,22
1
1989
7,607
3,209
730
395
83
97
311
3
2
84
7,662
2,844
205
23,25
0
1990
7,516
3,256
732
399
81
100
306
2
2
82
7,488
2,843
384
23,192
1991
7,482
3,309
745
401
79
103
298
2
2
81
7,373
2,796
305
22,977
1992
7,473
3,206
735
411
80
96
305
3
3
83
7,440
2,885
272
22,991
1993
7,782
3,176
732
414
82
95
314
3
3
84
7,437
2,966
296
23,402
"Emissions calculated as NO2.
Note: The sums of sub-categories may not equal total due to rounding (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1994).
-------
TABLE 5-10. NATIONWIDE VOLATILE ORGANIC COMPOUND EMISSION ESTIMATES, 1984 TO 1993
(Thousands short tons/year)
Source Category
Fuel Combustion - Electric Utilities
Fuel Combustion - Industrial
Fuel Combustion - Other
Chemical and Allied Product Manufacturing
Metals Processing
Petroleum and Related Industries
Other Industrial Processes
Solvent Utilization
Storage and Transport
Waste Disposal and Recycling
Highway Vehicles
Off-Highway
Miscellaneous
Total
1984
45
156
917
1,620
182
1,253
227
6,309
1,810
687
9,441
1,973
951
25,57
2
1985
32
248
508
1,579
76
797
439
5,779
1,836
2,310
9,376
2,008
428
25,41
7
1986
34
254
499
1,640
73
764
445
5,710
1,767
2,293
8,874
2,039
435
24,82
6
1987
34
249
482
1,633
70
752
460
5,828
1,893
2,256
8,201
2,038
440
24,33
8
1988
37
271
470
1,752
74
733
479
6,034
1,948
2,310
8,290
2,106
458
24,96
1
1989
37
266
452
1,748
74
731
476
6,053
1,856
2,290
7,192
2,103
453
23,73
1
1990
36
266
437
1,771
72
737
478
6,063
1,861
2,262
6,854
2,120
1,320
24,276
1991
36
270
426
1,778
69
745
475
6,064
1,868
2,217
6,499
2,123
937
23,508
1992
35
271
385
1,799
72
729
482
6,121
1,848
2,266
6,072
2,160
780
23,020
1993
36
271
341
1,811
74
720
486
6,249
1,861
2,271
6,094
2,207
893
23,312
Note: The sums of sub-categories may not equal total due to rounding (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1994).
-------
decreased by 6 percent, leading to an overall decrease of about 10% in emissions from all of
these categories from 1985 to 1993.
Table 5-7 shows PM10 emissions from natural and miscellaneous sources for 1985 to 1994.
Fugitive dust is the largest source in the miscellaneous category. No clear trend is evident in
overall fugitive dust emissions, because increases in paved road dust are offset by decreases in
the mining and quarrying and construction categories. The large year-to-year variability in wind
erosion reflects changes in precipitation and regional soil conditions. For instance, the values for
1993 reflect the flooding and extremely wet conditions that occurred in the midwestern United
States.
Tables 5-8 through 5-10 show nationwide emissions for sulfur dioxide, oxides of nitrogen,
and VOC's, which are all precursors for secondary aerosol formation, for the period from 1984
through 1993. Electric utilities account for the largest fraction of sulfur dioxide, nearly 70% of
total emissions in 1993 (Table 5-8). Estimates of sulfur dioxide emissions from industrial fuel
combustion increased by approximately 16% from 1984 to 1985, and decreased by 11% between
1985 and 1993. Sulfur dioxide emissions from chemical manufacturing doubled between 1984
and 1985, with emissions leveling off between 0.42 and 0.46 million short tons/year after 1985.
Sulfur dioxide emissions from highway vehicles were estimated to have increased by 8% from
1984 to 1989, then levelling off and then decreasing by about 10% from 1992 to 1993, reflecting
the introduction of regulations for the desulfurization of diesel fuel. Off-highway vehicle
emissions increased from 0.20 million short tons per year in 1984 to 0.28 million short tons per
year in 1993. Major sulfur dioxide emissions reductions were observed for petroleum
processing and other industrial processes, with decreases of 40% to 50% over the ten-year
period. In total, however, sulfur dioxide emissions estimates in 1993 decreased by 6% from
those given for 1984.
Table 5-9 shows no significant variations in total nitrogen oxides emissions over the
10-year period. Electric utility and motor vehicle emissions each account for about one-third of
total emissions. Emissions from (a) industrial and other fuel combustion and (b) from
off-highway vehicles each account for about one-sixth of total emissions. There is little change
in total emissions from 1984 to 1993. Moderate increases are seen in the electric utility,
industrial and other fuel combustion, and off-highway vehicles categories with much larger
5-41
-------
relative increases for chemical manufacturing and metals processing. These increases are offset
by decreases in fuel combustion by industry and on-highway vehicles.
Volatile organic compound (VOC) emissions in Table 5-10 are dominated by highway
vehicles and solvent use. These two sources together account for 50 to 60% of total emissions.
Off-highway vehicles, petroleum-related industries, chemical manufacturing, and petroleum
storage and transport account for most of the remaining amounts. VOC emissions from highway
vehicles were reduced between 1984 and 1993 by 35%, in spite of increased vehicle mileage.
Most of this decrease is due to the presumed effectiveness of emissions controls on newer
vehicles. VOC emissions from petroleum industries also were reduced by 43% between 1984
and 1993. Total VOC emissions decreased by 9% between 1984 and 1993. It should be noted
that emissions from natural sources are not reflected in the above discussion.
Although total emissions of gaseous precursors (SO2, NOX, and VOC's) are shown in
Tables 5-8, 5-9, and 5-10, it should be remembered that these values cannot be directly
translated into production rates of paniculate matter. Dry deposition and precipitation
scavenging of some of these gases can occur before they are oxidized to aerosols in the
atmosphere. In addition, some fraction of these gases are transported outside of the domain of
the continental United States before being oxidized. Likewise, emissions of these gases from
areas outside the United States can result in the transport of their oxidation products into the
United States. While the chemical oxidation of SO2 will lead quantitatively to the formation of
SO4=, the formation of aerosol from the oxidation of VOC's will be much less because only a
small fraction of VOC's react to form particles, and those that do have efficiencies less than 10%
(c.f. Section 5.3). The oxidation of NO2 will yield HNO3, some of which may dry deposit or be
scavenged by precipitation, and the remainder will form particulate nitrate.
Projections of future emissions of primary PM10, SO2, and NOX are shown in Table 5-11.
Controls mandated by the Clean Air Act Amendments of 1990 are expected to reduce PM10
emissions in nonattainment areas. However, because emissions in nonattainment areas constitute
a small subset of total emissions, overall emissions are projected as still likely to increase.
Fugitive dust sources contribute the major share of the increase. Changes in emissions after
1996 solely reflect activity level changes with the
5-42
-------
TABLE 5-11. PROJECTED TRENDS IN PARTICIPATE MATTER (PM10), SULFUR DIOXIDE (SO2),
AND OXIDES OF NITROGEN (NOY) EMISSIONS (106 short tons yr *)
PM,n Source Categories
Fuel Combustion"
1990
1993
1996
1999
2000
2002
2005
2008
2010
Natural"
4.36
1.98
4.36
4.36
4.36
4.36
4.36
4.36
4.36
Misc.a'b
36.3
37.9
43.6
48.5
49.8
51.8
54.9
57.4
59.0
Electric
Utilities
0.28
0.26
0.31
0.33
0.34
0.35
0.37
0.40
0.42
Industrial
0.24
0.23
0.21
0.20
0.20
0.19
0.19
0.18
0.18
Other
0.55
0.54
0.66
0.59
0.66
0.59
0.64
0.69
0.73
Mobile3
OS On-Road
0.90
0.91
0.89
0.93
0.94
0.97
1.01
1.04
1.06
0.36
0.32
0.15
0.13
0.12
0.13
0.13
0.13
0.12
Nonroad
0.37
0.40
0.44
0.47
0.48
0.50
0.53
0.55
0.56
Total
43.3
42.5
50.6
55.9
56.9
59.0
62.2
64.7
66.4
SO/
22.4
21.5
18.1
17.6
17.4
17.1
16.7
16.1
15.7
NO/
23.0
23.3
21.9
21.8
20.5
20.5
20.8
21.3
21.6
aSame categories as used in Tables 5-6 and 5-7.
bThe miscellaneous category includes fugitive dust from unpaved and paved roads, and other sources; wildfires and managed burning; and agricultural and
forestry related emissions.
COS refers to other stationary sources such as chemical manufacturing, metal processing, petroleum refining, other industrial processes, solvent utilization,
storage and transport, waste disposal and recycling.
dOnly total emissions are shown.
Source: U.S. Environmental Protection Agency (1995b).
-------
exception of on-road vehicles. Emission factors for on-road vehicles are expected to decrease
mainly because of stringent standards for diesel emissions. Diesel vehicle emissions are
expected to decrease nationwide by about 70% from 1990 to 2010 (U.S. Environmental
Protection Agency, 1993). This decrease results mainly from a roughly 90% decrease in
emissions factors which are partially offset by an increase in total diesel vehicle miles travelled.
As can be seen from Table 5-11, emissions from non-road sources (e.g., marine vessels,
railroads, aircraft, vehicles used in construction, industry, agriculture, airport services, and
landscaping) are projected to exceed those from on-road vehicles from 1990 to 2010.
Emissions of SO2 from fossil fuel combustion by electric utilities show an expected
continued decline through 2010. Emissions from all other categories in Table 5-7 show a slight
increase from 1993 to 2002 and then level off to the year 2010. Total NOX emissions show a
decrease of over 10% from 1993 to 2002, then increase by about 5% by the year 2010. This
pattern reflects projected emissions for the major categories of fuel combustion by electric
utilities and on-road vehicles.
Emissions of ammonia and ammonium are not included in the U.S. Environmental
Protection Agency inventories for criteria pollutants. Dentener and Crutzen (1994) have
constructed a global inventory of NH3 emissions. Anthropogenic sources (animals kept for
human use, fertilizer applications, and biomass burning) and natural sources (wild animals,
vegetation, and the oceans) were included. Emissions from sewage were not included, though.
Vegetation was found to be either a source or a sink for NH3 depending on ambient
concentrations and vegetation type. Animals kept for human use represent the largest single
source category. Highest emission rates in North America were found in the central United
States. Matthews (1994) found that about 75% of U.S. NH3 emissions from the application of
nitrogenous fertilizers occur in the central United States, with the remainder about evenly
divided between the eastern and western United States. Emissions of approximately 0.51 Tg
NH3-N yr"1 were calculated for the United States. The Dentener and Crutzen (1994) estimate of
NH3 emissions for North America of 5.2 Tg N yr"1 may be compared to a wet deposition rate of
NH4+ in the United States of 3 -4.5 Tg N yr"1, and three separate emission inventories yielding
values of 1.2, 8.8, and 2.8 Tg N yr"1 for the U.S. (Placet et al., 1991).
While emissions of organic carbon (OC) and elemental carbon (EC) are included implicitly
in the emissions inventories for PM10, it is still useful to consider independent estimates. Zhang
5-44
-------
et al. (1992) estimated the total production of secondary organic aerosol to be about 1.2 Tg yr"1
in the United States. Liousse et al. (1996) have constructed OC and EC emissions inventories
for use in a global scale chemical tracer model. They estimate OC emissions of 0.80 Tg OC yr"1
from live biomass combustion, 1.4 Tg OC yr"1 from fossil fuel combustion, and 0.59 Tg OC yr"1
from the oxidation of naturally emitted terpenes assuming a fractional aerosol yield of 5%.
Carbon values for OC sources have been multiplied by a factor of 1.2 to account for the presence
of oxidized species. EC emissions from the combustion of live biomass and fossil fuels are
estimated to be 0.11 Tg EC yr"1 and 0.30 Tg C yr"1, respectively. These estimates are roughly
8% of total particulate emissions shown in Tables 5-6 and 5-7. Comparisons of model results
with observations from the IMPROVE/NESCAUM network by Liousse et al. (1996) suggest
that both the OC and EC emissions derived for their model may be systematically
underestimated by at least a factor of two.
The regional nature of total primary parti culate matter emissions is illustrated in Figure 5-
5. At least 80% of the emissions in any single region arises from fugitive dust sources and wind
erosion. SO2 regional emissions are shown in Figure 5-6 as a reminder that they are highest in
the eastern United States and that the oxidation of SO2 to SO4= can constitute a substantial
fraction of the aerosol mass in the eastern United States. It can also be seen that the ratio of SO2
to primary PM10 emissions tends to be much higher in the eastern than in the western United
States.
Annual averages do not reflect the seasonality of certain emissions. Residential wood
burning in fireplaces and stoves, for example, is a seasonal practice which reaches its peak
during cold weather. Cold weather also affects motor vehicle exhaust paniculate emissions, both
in terms of chemical composition and emission rates (e.g., Watson et al., 1990b; Huang et al.,
1994). Planting, fertilizing, and harvesting are also seasonal activities. Forest fires occur mainly
during the local dry season and during periods of drought.
Several of the sources in Tables 5-6 through 5-10 are episodic rather than continuous in
nature. This is especially true of prescribed and structural fires and fugitive dust emissions.
Although windblown dust emissions are low on an annual average, they are likely to be quite
5-45
-------
Figure 5-5. Estimates of primary PM10 emissions by U.S. Environmental Protection
Agency region for 1992.
Units = 106 short tons/yr (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1993).
Figure 5-6. Estimates of sulfur dioxide emissions by U.S. Environmental Protection
Agency region for 1992.
Units = 106 short tons/yr (1 short ton = 9.08 x 105 gms).
Source: U.S. Environmental Protection Agency (1993).
5-46
-------
large during those few episodes when wind speeds are high. The transport of Saharan dust to the
continental United States is also highly episodic.
5.5 APPLICATIONS AND LIMITATIONS OF EMISSIONS
INVENTORIES AND RECEPTOR MODELS
This section examines requirements for the design and construction of emissions
inventories and potential areas of uncertainty and limitations in their use. Receptor modeling
methods to apportion sources to mass components in ambient aerosol measurements, and results
for a number of aerosol monitoring studies, will then be presented. Some general considerations
of the relative strengths and weaknesses of using emissions inventories and receptor models to
assign sources to particulate matter components in ambient samples will then be discussed.
Finally, results from specific receptor modeling studies in the eastern and western United States
will be discussed.
5.5.1 Uncertainties in Emissions Estimates
It is difficult to assign uncertainties quantitatively to entries in emissions inventories.
Methods that can be used to verify or place constraints on emissions inventories are sparse. In
general, the overall uncertainty in the emissions of a given pollutant includes contributions from
all of the terms on the right hand side of Eq. 5-1 (activity rates, emissions factors, and control
device efficiencies). Additional uncertainties can arise during the compilation of an emissions
inventory because of missing sources and arithmetical errors. The variability of emissions can
cause errors when annual average emissions are applied to applications involving shorter time
scales.
Activity rates for well-defined point sources (e.g., power plants) should have the smallest
uncertainty associated with their use, since accurate production records need to be kept. On the
other hand, activity rates for a number of areally dispersed fugitive sources are extremely
difficult to quantify. Emissions factors for easily measured fuel components which are
quantitatively released during combustion (e.g., CO2 and SO2) should be the most reliable.
Emissions of components formed during combustion are more difficult to characterize as the
5-47
-------
emissions rates are dependent on factors specific to individual combustion units and on
combustion stage (i.e., smoldering or active). Although the AP-42 emissions factors (U.S.
Environmental Protection Agency, 1995a) contain extensive information for a large number of
source types, these data are very limited in the number of sources sampled. The efficiency of
control devices is determined by their age, their maintenance history, and operating conditions.
It is virtually impossible to assign uncertainties in control device performance due to these
factors. It should be noted that the largest uncertainties occur for those devices which have the
highest efficiencies (>90%). This occurs because the efficiencies are subtracted from one and
small errors in assigning efficiencies can lead to large errors in emissions.
Ideally an emissions inventory should include all major sources of a given pollutant. This
may be an easy task for major point sources, but becomes problematic for poorly characterized
area sources. As an example, it was recently realized that meat cooking could be a significant
source of organic carbon (Hildemann et al., 1991). Further research is needed to better
characterize the sources of pollutants in order to reduce this source of uncertainty. Errors can
arise from the misreporting of data, and arithmetic errors can occur in the course of compiling
entries from thousands of individual sources. A quality assurance program is required to check
for outliers and arithmetic errors.
Because of the variability in emissions rates, there can be errors in the application of
inventories developed on an annually averaged basis (as are the inventories shown in Tables 5-6
to 5-10) to episodes occurring on much shorter time scales. As an example, most modeling
studies of air pollution episodes are carried out for periods of a few days.
Uncertainties in annual emissions were estimated to range from 4 to 9% for SO2 and from
6 to 11% for NOX in the 1985 NAPAP inventories for the United States (Placet et al., 1991).
Uncertainties in these estimates increase as the emissions are disaggregated both spatially and
temporally. The uncertainties quoted above are conservative estimates and refer only to random
variability about the mean, assuming that the variability in emissions factors was adequately
characterized and that extrapolation of emissions factors to sources other than those for which
they were measured is valid. The estimates do not consider the effects of weather or variations
in operating and maintenance procedures. Fugitive dust sources, as mentioned above, are
extremely difficult to quantify, and stated emission rates may only represent order-of-magnitude
estimates. As rough estimates, uncertainties in emissions
5-48
-------
estimates could be as low as 10% for the best characterized source categories, while emissions
figures for windblown dust should be regarded as order-of-magnitude estimates. Given (a)
uncertainties in the deposition of SO2 and its oxidation rate, (b) the variability seen in OC and
EC emissions from motor vehicles along with the findings from past verification studies for
NMHC and CO to NOX ratios, (c) ranges of values found among independent estimates for
emissions of individual species (NH3, OC), and (d) the predominance of fugitive emissions, PM
emissions rates should be regarded as order-of-magnitude estimates.
There have been few field studies designed to test emissions inventories observationally.
The most direct approach would be to obtain cross-sections of pollutants upwind and downwind
of major urban areas from aircraft. The computed mass flux through a cross section of the urban
plume can then be equated to emissions from the city chosen. This approach has been attempted
on a few occasions. Results have been ambiguous because of contributions from fugitive
sources, non-steady wind flows, and general logistic difficulties.
Greater success, albeit on a smaller scale, has been achieved in studies that tested
predictions of the State of California EMFAC emissions model. An ambient-air study in the Los
Angeles basin (Fujita et al., 1992) showed that motor vehicle emissions of CO and nonmethane
hydrocarbons (NMHC) were being systematically underpredicted in the emissions model by a
factor of about 2.5, assuming that NOX emissions were much better known; i.e., the CO to NOX
and NMHC to NOX ratios were underpredicted by the model. A study performed in a tunnel in
the Los Angeles basin (Ingalls, 1989; Pierson et al., 1990) showed that motor vehicle NOX
emission rates (g/mi) were predicted approximately correctly but that the CO and NMHC
emission rates were systematically underpredicted in the emissions model by factors of two to
three. Similar tests need to be performed for particulate matter emissions from motor vehicles.
A completely different approach to obtaining area-wide emissions of pollutants relies on
the construction of inversion algorithms applied in the context of atmospheric transport models
(Brown, 1993). Emissions of a pollutant that are required to produce a specified distribution of
surface concentrations are solved for by using model-derived transport and chemical loss terms.
Uncertainties in the emissions fields are then generated in terms of specified uncertainties in the
observed data and in the model transport and chemistry fields.
5-49
-------
A number of factors limit the ability of an emissions inventory driven, chemical tracer
model to determine the effects of various sources on particle samples obtained at a particular
location apart from uncertainties in the inventories given above. Air pollution model predictions
represent averages over the area of a grid cell, which in the case of the Urban Airshed Model
typically has been 25 km2 (5 km x 5 km). The contributions of sources to pollutant
concentrations at a monitoring site are strongly controlled by local conditions which cannot be
resolved by an Eulerian grid-cell model. Examples would be the downward mixing of tall stack
emissions and deviations from the mean flow caused by buildings. The impact of local sources
may not be accurately predicted, because their emissions would be smeared over the area of a
grid cell or if the local wind flow were in the wrong direction during sampling.
For these reasons, receptor models have been used to determine source contributions to
particulate matter at individual monitoring sites. Receptor models are strictly diagnostic in their
application and do not have the prognostic, or predictive, capability of chemical transport
models. In addition, receptor models have been developed for apportioning sources of primary
particulate matter and are not formulated to include the processes of secondary particulate matter
formation which are explicitly included in the chemical transport models.
5.5.2 Receptor Modeling Methods
Receptor models relate source contributions to ambient concentrations based on
composition analysis of ambient particulate samples. They depend on the assumption of mass
conservation and the use of a mass balance. As an example, assume that the total concentration
of particulate lead measured at a site can be considered to be the sum of contributions from a
number of independent sources,
""total = ""motor vehicles + ""soil + ""smelter + • • • (5-2)
Since most sources emit particles that contain a number of chemical elements or compounds, the
atmospheric concentration of an element can be considered to be the product of the abundance of
the element of interest (ng/mg) in the effluent and the mass concentration of particles from that
source in the atmosphere (mg/m3). For lead from motor vehicles, for example,
5-50
-------
motor vehicles 3Pl>, mv mv (5-3)
where apbmv is the abundance of lead in motor vehicle emissions, and^,v is the mass
concentration of motor vehicle emitted particles in the atmosphere. Extending this idea to m
chemical elements, n samples, and p independent sources,
*V = £ »* 4y (5-4)
where x;j is the ith elemental concentration measured in the jth sample (ng m"3), aik is the
gravimetric abundance of the ith element in material from the k* source (ng mg"1), and fkj is the
airborne mass concentration of material from the kth source contributing to the jth sample (mg
m"3). The fkj are the quantities to be determined from Equation 5-4. To distinguish the
contributions of one source type from another using receptor models, the chemical and physical
characteristics must be such that (1) they are present in different proportions in different source
emissions, (2) these proportions remain relatively constant for each source type, and (3) changes
in these proportions between source and receptor are negligible or can be empirically
represented.
A number of specialty conference proceedings, review articles, and books have been
published to provide greater detail about source apportionment receptor models (Cooper and
Watson, 1980; Watson et al., 1981; Macias and Hopke, 1981; Dattner and Hopke, 1982; Pace,
1986; Watson et al., 1989; Gordon, 1980, 1988; Stevens and Pace, 1984; Hopke, 1985, 1991;
Javitz et al., 1988). Watson et al. (1994b) present data analysis plans which include receptor
models as an integral part of visibility and PM10 source apportionment and control strategy
development.
The first step in attempting to relate ambient particulate matter measured at a particular
location to source contributions is typically data evaluation. The objectives for data evaluation
are: (1) to summarize the accuracy and precision of measurements; (2) to identify and
investigate extreme and inconsistent values; (3) to perform data comparisons and investigate
discrepancies; and (4) to estimate the equivalence of measurements of the same variable by
different methods. Even with the most stringent quality assurance, it is prudent to perform
several straightforward analyses to identify the presence of any discrepancies in atmospheric
5-51
-------
particulate data and to correct, flag, or eliminate them. Investigating the equivalence of different
measurement methods for the same variable is especially important for particulate chemical
measurements, which may show substantial differences in concentration depending upon how
they were made. Data evaluation activities include: (1) plotting and examining pollutant time
series data to identify spikes and outliers for investigation; (2) comparing the sum of chemical
species with PM10 mass measurements; and (3) comparing measurements of the same variables at
the same or nearby sites using different measurement devices and procedures.
After data evaluation the next step in an analysis of particulate air quality in a region is a
process that can be termed a descriptive air quality analysis. The objectives of a descriptive air
quality analysis are: (1) to identify similarities and differences in air quality at different
sampling sites; and (2) to depict temporal and spatial variations in particulate and gaseous
precursor measurements. Descriptive air quality analysis includes: (1) statistical summaries of
median and extreme values of air quality variables for different sites, episodes, and times of day;
(2) time series plots of PM10 and selected chemical components; (3) spatial pie plots of
particulate chemical composition; and (4) spatial and temporal correlations between PM10 and
chemical composition measurements. The product of this analysis is a quantitative overview of
particulate concentrations during the period of interest.
Performed at the same time as a descriptive air quality analysis, a descriptive
meteorological analysis is carried out to: (1) describe the spatial structure, temporal variability,
and statistical distribution of meteorological conditions; (2) describe the transport and mixing
patterns in the study domain; and (3) identify relationships between meteorology and
atmospheric particulate concentrations. Data normally available would include wind speed,
wind direction, temperature, relative humidity, and solar radiation at ground level and aloft (if
available).
Descriptive meteorological analysis activities include: (1) statistical summaries of
meteorological variables; (2) time series and spatial plots of meteorological variables, including
wind vectors, with examination for phenomena such as inter-basin transport, stagnation, slope
flows, convergence zones, and recirculation; (3) identification of layers and orographic
phenomena that change with elevation; (4) tabulations of fog occurrences,
5-52
-------
frequencies, locations, and intensities; and (5) meteorological descriptions and comparisons with
meteorology during high PM10 episodes from prior years.
The product of these analyses is a conceptual understanding of how meteorological
phenomena influence atmospheric particulate concentrations in a particular region.
The next step in receptor modeling for particulate matter is a source profile compilation.
The objectives of source profile compilation analysis are: (1) to combine profiles from
individual samples into composite profiles; and (2) to assign source profiles to source categories
based on their degree of similarity or difference. Data needed for this study are the chemical
measurements on samples from representative source types that are expected to contribute to
airborne particulate matter during study periods. Major source types include, for example: (1)
suspended geological material from roads and from agricultural and unpaved areas; (2) primary
particle exhaust from gasoline- and diesel-powered vehicles; (3) industrial sources; (4) residual
oil combustion; and (5) secondary ammonium sulfate and ammonium nitrate originating from
gaseous precursors. Source profile compilations include: (1) tables and plots of individual
profiles and their uncertainties; (2) calculation of averages and standard deviations for category
profiles; and (3) calculation of weighted composite profiles for source categories which are
found for the source apportionment modeling described below. It is important to emphasize that
source and ambient samples must be analyzed using the same protocols and methods (U.S.
Environmental Protection Agency, 1994).
The chemical mass balance (CMB) receptor model is the model most commonly used for
particulate matter source apportionment. The CMB model uses the chemical and physical
characteristics of gases and particles measured at source and receptor to both identify the
presence of, and quantify source contributions to, the receptor (Friedlander, 1973).
The CMB consists of an effective variance least-squares solution to the set of linear
equations (5-4) that expresses each concentration of a chemical species at a receptor site as a
linear sum of products of source profile species and source contributions. The source profile
species, i.e., the fractional amount of the species in the emissions from each source type, and the
receptor concentrations, with appropriate uncertainty estimates, serve as input data to the CMB
model. The output consists of: (1) the source contribution estimates of
5-53
-------
each source type; (2) the standard errors of these source contribution estimates; and (3) the
amount contributed by each source type to each chemical species. The model calculates values
for the contributions from each source type and the uncertainties associated with those values.
Input data uncertainties are used both to weight the importance of input data values in the
solution and to calculate the uncertainties of the source contributions. The CMB model
assumptions are: (1) compositions of source emissions are constant over the period of ambient
and source sampling; (2) chemical species do not react with each other, i.e., they add linearly;
(3) all sources with a potential for significantly contributing to the receptor have been identified
and their emissions have been characterized; (4) the source compositions are linearly
independent of each other; (5) the number of sources or source categories is less than or equal to
the number of chemical species; and (6) measurement uncertainties are random, uncorrelated,
and normally distributed. Assumptions 1 through 6 for the CMB model are fairly restrictive and
will never be completely satisfied in actual practice. Fortunately, the CMB model can tolerate
reasonable deviations from these assumptions, although these deviations increase the stated
uncertainties of the source contribution estimates.
The CMB modeling procedure requires: (1) identification of the contributing source types;
(2) selection of chemical species to be included; (3) estimation of the fraction of each of the
chemical species which is contained in each source type (i.e., the source compositions); (4)
estimation of the uncertainty in both ambient concentrations (including artifacts during sampling
and storage such as gas absorption or volatilization in filter samples) and source compositions;
(5) estimation of differential losses during transport from source to receptor; (6) solution of the
chemical mass balance equations; and (7) validation and reconciliation. Each of these steps
requires different types of data. Uncertainties in the modeling results can be noticeably reduced
by obtaining source profile measurements which correspond to the period of ambient
measurements (Glover et al., 1991; Dzubay et al., 1988; and Olmez et al., 1988). Stratifying
data according to wind direction can also increase the number of source types that can be
resolved as shown in the above studies.
Emissions inventories are examined to determine the types of sources that are most likely
to influence a receptor. These emissions inventories for particulate matter are
5-54
-------
frequently far from complete, however, and other measures are needed to infer the influence of
uninventoried sources. The Principal Components Analysis and Empirical Orthogonal Function
models described below can aid in this identification. Once these sources have been identified,
profiles acquired from similar sources can be examined to select the chemical species to be
measured. The more species measured, the better the precision of the CMB apportionment.
The Principal Components Analysis (PCA) receptor model classifies variables into groups
identifiable as causes of particulate matter levels measured at receptors. Typical causes are
emissions sources, chemical interactions, or meteorological phenomena. The PCA model uses
ambient concentrations of chemical species and meteorological data as inputs. PCA does not use
source emissions measurements, as does the CMB model, but it may require 50 or more
measurements of many species from different time periods at a single receptor site.
The PCA procedure is as follows: (1) select the chemical species and measurement cases
to be included; (2) calculate the correlation coefficients between the species; (3) calculate the
eigenvectors and eigenvalues of the correlation matrix; (4) select eigenvectors to be retained; (5)
rotate these eigenvectors into a more physically meaningful space; and (6) interpret the rotated
vectors as air pollution sources based on the chemical species with which they are highly
correlated. Freeman et al. (1989) describe the computer software and methods required to use
the PCA model for PM10 source assessment. See also Henry (1991).
The PCA model assumptions are: (1) compositions of source emissions are constant over
the period of ambient and source sampling; (2) chemical species concentrations add linearly; (3)
measurement errors are random and uncorrelated; (4) the case-to-case variability of actual source
contributions is much larger than the variability due to other causes, such as measurement
uncertainty or changes in source profiles due to process and fuel changes; (5) causes of
variability that affect all sources equally (such as atmospheric dispersion) have much smaller
effects than causes of variability for individual source types (such as wind direction or emission
rate changes); (6) the number of cases exceeds the number of variables in the PCA model to an
extent that statistical stability is achieved; and (7) eigenvector rotations are physically
meaningful.
5-55
-------
There are a number of examples of the application of PC A models. Photochemical factors
were found to influence particulate matter measurements from Los Angeles, CA, New York, NY
(Henry and Hidy, 1979), St. Louis, MO (Henry and Hidy, 1982), Lewisburg, WV (Lioy et al.,
1982), and Detroit, MI (Wolff et al., 1985a). These photochemical factors were consistently
associated with daily average and maximum ozone (O3), maximum temperatures, and absolute
humidity. The photochemical factors found for Los Angeles data (Henry and Hidy, 1979) were
highly correlated with daily maximum and minimum relative humidity measurements. Local
source factors were found for Salt Lake City (Henry and Hidy, 1982) and Los Angeles (Henry
and Hidy, 1979) and were highly correlated with sulfur dioxide (SO2) and the wind direction
frequency distributions. Dispersion/stagnation factors were found for St. Louis, Salt Lake City,
and Lewisburg. The variables correlated with the dispersion/stagnation factor were nitric oxide
(NO), nitrogen dioxide (NO2), wind speed at midnight and noon, average wind speed, morning
mixing height, maximum hourly precipitation, and average precipitation. PCA has also been
used to identify sources which may not be inventoried (Wolff and Korsog, 1985; Cheng et al.,
1988; Henry and Kim, 1989; Koutrakis and Spengler, 1987; Zeng and Hopke, 1989).
The PCA procedure as outlined above provides only a qualitative assessment of air
pollution sources. In some circumstances, however, the procedure can be extended to produce
quantitative estimates of the source impacts. For example, a chemical species strongly
associated with a single PCA group may be suitable as a source tracer for use in a subsequent
multiple linear regression receptor model (Kleinman et al., 1980)
The Empirical Orthogonal Function (EOF) receptor model is applied to a spatially dense
network of measurements to identify the locations of emissions sources and to estimate the net
fluxes (emissions minus deposition) of those pollutants. The EOFs manifest themselves as
isopleth maps of flux density. When a major point source is the emitter, such as a coal-fired
power plant, the EOFs have been shown (Gebhardt et al., 1990) to surround that source. EOFs
have been applied to air pollution measurements by Peterson (1970), Ashbaugh et al. (1984),
Wolff et al. (1985b), and Henry et al. (1990). Henry et al. (1990) were the first researchers to
place this method on a firm theoretical foundation and to demonstrate that EOFs reproduce the
net fluxes used as input to a dispersion model.
5-56
-------
In prior studies, the EOF model was applied to single chemical constituents, such as sulfur
dioxide, sulfate, and total particulate mass concentrations. In a recent study (Watson et al.,
1991), the EOF model was applied to the source contribution estimates calculated for each
sample from the CMB modeling described above. In this way, source-type specific patterns of
net flux were determined. For example, the geological source contributions may be represented
as a linear sum of EOFs which correspond to a dirt road, a construction site, and an area of
intense agricultural activity. The motor vehicle exhaust source contributions may be represented
by a linear sum of EOFs which correspond to a major highway, a large truck stop, or an urban
core area. The EOF model may also be applied to specific chemical species which are identified
as potential markers for uninventoried sources.
The EOF procedure is similar to the PCA procedure: (1) select the source contribution
estimates and measurement cases to be included; (2) calculate the covariance coefficients
between the species measured at the same time at several sites; (3) calculate the eigenvectors and
eigenvalues of the covariance matrix; (4) select eigenvectors to be retained; (5) rotate these
eigenvectors into a more physically meaningful space; and (6) interpolate between the values of
these eigenvectors to supply the net flux patterns. The main difference between PCA and EOF
is that PCA operates on many samples from a single site taken over an extended time period,
while EOF operates on many samples from many sites taken over a single time period.
The formulation of Henry et al. (1990), termed Source Identification Through Empirical
Orthogonal Functions (SITEOF), uses wind velocities as input in addition to the spatially
distributed source contribution estimates. The SITEOF assumptions are: (1) net fluxes of
spatially-distributed pollutants add linearly; (2) pollutants are homogeneously distributed
vertically in the mixed layer; (3) measurement errors are random and uncorrelated; (4) the
number of sampling sites exceeds the number of source locations to an extent that statistical
stability is achieved; and (5) measurement locations are located in positions to maximize spatial
gradients from major source emissions. The major unknown concerning the SITEOF model is
the extent to which assumptions 4 and 5 can be met in actual practice. Motor vehicle exhaust is
confined to specific areas (e.g., roads and parking lots), and it is a straightforward task to locate
monitors close to and far from these known locations. Fugitive dust, on the other hand, can be
emitted from many locations.
5-57
-------
The SITEOF model is one of a class of procedures referred to as "hybrid receptor models".
Such models make use of not only the ambient species concentration measurements that form
the input data for a pure receptor model, but in addition source emission rates or atmospheric
dispersion or transformation information characteristic of dispersion models. By exploiting
simultaneously the strengths of the two complementary approaches their individual weaknesses
should be minimized. A survey of hybrid receptor models is available (Lewis and Stevens,
1987).
Ashbaugh et al., (1985) developed the concept of the potential source contribution function
(PSCF) for performing the apportionment of secondary species, for combining air parcel back
trajectories from a receptor site with chemical data at the site to infer possible source locations.
The PSCF is an estimate of the conditional probability that a trajectory which passed through a
given cell in the emissions grid (g;j) contributed a concentration greater than some threshold
value to ambient concentrations at the receptor site. Gao et al. (1993) extended the PSCF
analysis to provide an apportionment of secondary species. By multiplying the PSCF by the
emissions rate in gj, an estimate of the maximum contribution of sources in gj to the
concentrations measured at the receptor site is obtained. Further research is needed to quantify
the uncertainties associated with this method. These uncertainties are related to unidentified
sources, background sources, emissions estimates at the time of calculation, the differential loss
of species (e.g., by deposition), and mixing of air parcels from different cells during transit from
source to receptor. Gao et al. (1993) have applied PSCF's along with emissions estimates to the
apportionment of SO2 and SO4 at sites in the South Coast Air Basin, and Cheng et al. (1996)
have also applied this technique to the apportionment of NOX and NH3 in this area.
The use of 14C isotopic analysis has been used to distinguish between fossil fuel and
biomass sources of carbon in aerosol samples. An example would be to determine the fraction
of ambient aerosol mass concentration in wintertime samples originating from woodburning.
This method has been particularly useful in validating less expensive receptor methods of
achieving the same goal (Wolff et al., 1981; Lewis et al., 1988).
The preceding sections have dealt with receptor models that rely on chemical information
obtained from bulk samples. It is worth noting that in addition there are powerful receptor
modeling methods which also use the morphology and composition of
5-58
-------
individual particles to provide clues to their source origin (Dzubay and Mamane, 1989).
Scanning electron microscopy (SEM) along with energy dispersive X-ray analysis (EDX) has
allowed the size distribution of particles to be characterized according to shape and elemental
composition. This technique has proven to be extremely useful for distinguishing between fly
ash and soil derived particles; both types of particles have similar composition, but fly ash
particles are spherical while soil particles are irregularly shaped.
Manually performing SEM/EDX analyses of the large number individual particles
necessary to characterize a size distribution is extremely time consuming. Automated methods
have been developed for the rapid characterization of the shapes of hundreds of particles in
aerosol samples (Xie et al., 1994a, 1994b). The morphology data can then be used along with
EDX data to assign particles to clusters related to specific source types (Van Espen, 1984).
5.5.3 Source Contributions to Ambient Particles Derived by Receptor
Models
Receptor modeling has been used for obtaining information about the nature of sources of
ambient aerosol samples. The results of several studies will be discussed to provide an
indication of different sources of particulate matter across the United States. First, results
obtained by using the CMB approach for estimating contributions to PM2 5 and PM(10_2 5) from
different source categories at monitoring sites in the United States east of the Mississippi River
will be discussed. Estimated contributions from a number of source categories to PM10 in
ambient samples, obtained mainly at sites west of the Mississippi River, will then be discussed.
Dzubay et al. (1988) estimated source category contributions to 24-hour PM2 5 and PM(10.
25) samples obtained by a dichotomous sampler at three widely separated sites in the
Philadelphia, PA area (NE airport in Philadelphia, PA; Camden, NJ; and a site about 30 km to
the SW of Camden, NJ) during the summer of 1982. They used a composite of CMB, multiple
linear regression (MLR), and wind trajectory receptor models. Source compositions shown in
Table 5-3 were obtained partly to provide input to this study (Olmez et al., 1988). Ambient
concentrations of individual species used by Dzubay et al. (1988) are shown in Table 6A-2a
(Chapter 6, Appendix A). Sulfate, associated NH4+ and water
5-59
-------
constituted about 70% of PM2 5. Since the mean fractional abundances of PM2 5 to PM10 was
0.75, it can be seen that sulfate components contributed approximately 53% of PM10. Coal- and
oil-fired power plants located east of the Mississippi River were found to contribute 50 ± 6%
and 11 ± 4% of PM2 5, by using Se as a tracer for coal combustion and V and Ni as tracers for oil
combustion, based on an MLR analysis.
The study was performed during a period (summer of 1982) when the Pb content of
gasoline was declining markedly, and so Pb could not be used as a unique tracer of motor
vehicle emissions. CMB was used to determine nonvehicular Pb, which was subtracted from the
measured Pb concentration to yield a tracer for vehicle exhaust. Motor vehicle exhaust was then
found to contribute about 8%, on average, to PM2 5. Local sources of sulfate (determined from
the MLR intercept) were found to contribute 13%, on average, with smaller contributions from
local industrial sources, incinerators, and crustal matter to PM2 5.
Crustal matter constituted about 76%, on average, of PM(10.25). Sulfate and associated
NH4+ and water constituted only about 7% of PM(10_2 5). Other contributions to PM(10_2 5) were
found to arise from sea-salt and incinerators. In a study of the Philadelphia aerosol in the
summer of 1994, Pinto et al. (1995) found close agreement with Dzubay et al. (1988) both in
measured concentrations and in the magnitude of the inferred fractional contribution of regional
sulfate sources.
Glover et al. (1991) estimated the contributions of different source categories to 24-hour
PM2 5 and PM(10_2 5) samples obtained with a dichotomous sampler at a site in Granite City, IL.
Again, sulfate was the major constituent of PM25, constituting from 59% of PM25 with SSW
winds to 86.6% of PM2 5 with NNW winds. Inferred contributions from specific source types
were also shown to be strongly dependent on wind direction. Inferred contributions from iron
works ranged from 3.4% with NNW winds to 16.4% with SSE winds. Inferred contributions
from a Pb smelter ranged from 2.8% with WNW winds to 11.6% with SSW winds. Inferred
contributions from other sources (e.g., motor vehicles, incinerators, other smelters, and soil)
were all typically a few per cent.
Sulfate was a relatively minor constituent (< 10%) of PM(10.25) samples. Major inferred
contributions were from iron works, ranging from 5.7% with WNW winds to 53.8% with ENE
winds; soil, ranging from 4.2% with WSW winds to 35.8% with ESE
5-60
-------
winds; street dust, ranging from 1.4% SSE winds to 45.6% with WNW winds; with generally
smaller contributions from the other sources listed for PM2 5.
These results demonstrate the different nature of PM25 and PM(10.25) sources (i.e., PM25
was derived from regional sources, while PM(10_2 5) was derived from local industries); the utility
of wind sectoring to help locate sources; and the need to obtain site-specific source composition
profiles. The use of site-specific source profiles instead of profiles culled from the literature
resulted in decreases in predicted error values, especially for fugitive dust.
Results obtained at a number of monitoring sites in the central and western United States
obtained by using the CMB model are shown in Table 5-12 for PM10. The sampling sites
represent a variety of different source characteristics within different regions of Arizona,
California, Colorado, Idaho, Illinois, Nevada and Ohio. Several of these are background sites,
specifically Estrella Park, Gunnery Range, Pinnacle Peak, and Corona de Tucson, AZ, and San
Nicolas Island, CA. Definitions of source categories also vary from study to study. In spite of
these differences, several features can be observed from the values in this table.
Fugitive dust (geological material) from roads, agriculture and erosion appears as a highly
variable contributor to PM10 at nearly all sampling sites shown in Table 5-12, contributing about
40% of the average PM10 mass concentration. The average fugitive dust source contribution is
highly variable among sampling sites within the same urban areas, as seen by differences
between the Central Phoenix (33 //g/m3) and Scottsdale (25 //g/m3) sites in Arizona, and it is
also seasonally variable, as evidenced by the summer and fall contributions at Rubidoux, CA.
These studies found that the source profiles for fugitive dust were chemically similar, even
though the dust came from different emitters, so that further apportionment into sub-categories
was not possible by the CMB model alone. Road sand often contains salts that allow it to be
distinguished from other fugitive dust sources. It is usually the only exposed fugitive dust
source when other sources are covered by snowpack. Dust from some construction activities and
cement plants can also be separated from other sources due to enrichments in calcium content of
these emissions, as seen in studies at Rubidoux, CA, Rillito, AZ (near cement plants), and
Pocatello, ID (near chemical and fertilizer production plants).
5-61
-------
TABLE 5-12. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
a\
,ug/m3
Sampling Site
Central Phoenix, AZ (Chow et al., 1991)
Craycroft, AZ (Chow et al., 1992a)
Hayden 1, AZ (Garfield) (Ryan et al., 1988)
Hayden 2, AZ (Jail) (Ryan et al., 1988)
Phoenix, AZ (Estrella Park) (Chow et al., 1991)
Phoenix, AZ (Gunnery Rg.) (Chow et al., 1991)
Phoenix, AX (Pinnacle Pk.) (Chow et al., 1991)
Rillito, AZ (Thanukos et al., 1992)
Scottsdale, AZ (Chow et al., 1991)
West Phoenix, AZ (Chow et al., 1991)
Bakersfield, CA(Magliano, 1988)
Bakerfield, CA (Chow et al., 1992b)
Crows Landing, CA (Chow et al., 1992b)
Fellows, CA (Chow et al., 1992b)
Fresno, CA (Magliano, 1988)
Fresno, CA (Chow et al., 1992b)
Indio, CA (Kim et al., 1992)
Kern Wildlife Refuge, CA (Chow et al., 1992b)
Long Beach, CA (Gray et al., 1988)
Long Beach, CA (Summer) (Watson et al., 1994b)
Long Beach, CA (Fall) (Watson et al., 1994b)
Riverside, CA (Chow et al., 1992c)
Rubidoux, CA (Gray et al., 1988)
Rubidoux, CA (Summer) (Watson et al., 1994b)
Rubidoux, CA (Fall) (Watson et al., 1994b)
Rubidoux, CA (Chow et al., 1992c)
San Nicolas Island, CA (Summer) (Watson et al.,
1994b)
Primary
Time Period Geological
Winter 1989-1990
Winter 1989-1990
1986
1986
Winter 1989-1990
Winter 1989-1990
Winter 1989-1990
1988
Winter 1989-1990
Winter 1989-1990
1986
1988-1989
1988-1989
1988-1989
1986
1988-1989
1988-1989
1986
Summer 1987
Fall 1987
1988
1986
Summer 1987
Fall 1987
1988
Summer 1987
33.0
13.0
5.0
21.0
37.0
20.0
7.0
42.7
25.0
30.0
27.4
42.9
32.2
29.0
17.1
31.8
33.0
15.1
20.7
11.1
11.3
32.6
43.1
34.9
19.2
48.0
1.6
Primary
Motor
Primary Vehicle
Construction Exhaust
0.0
0.0
2.0
4.0
0.0
0.0
0.0
13.8
0.0
0.0
3.0
1.6
0.0
1.4
0.7
0.0
3.0
2.0
0.0
0.0
0.0
0.0
4.0
4.5
16.1
0.0
0.0
25.0
8.3
0.0
0.0
10.0
5.5
2.9
1.2'
19.0
25.0
5.5
7.7
2.2
2.1
4.0
6.8
4.4
2.2
5.1
6.3
42.8
7.0
5.6'
17.3
30.3
10.2
0.9
Primary
Vegetative
Burning
2.3
0.0
0.0
0.0
0.9
0.0
1.0
0.0
7.4
10.0
9.6
6.5
3.4
3.4
9.2
5.1
7.1
4.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Secondary
Ammonium
Sulfate
0.2
0.7
4.0
4.0
1.6
1.0
0.9
0.0
0.6
0.4
5.6
5.5
2.8
5.1
1.8
3.6
3.6
3.3
8.0
10.9
3.8
4.8
6.4
9.5
2.1
5.3
3.7
Secondary Misc.
Ammonium Source
Nitrate
2.8
0.6
0.0
0.0
0.0
0.0
0.0
0.0
3.6
3.1
0.0
12.7
6.5
7.5
0.0
10.4
4.1
1.5
9.2
0.8
23.2
21.4
21.3
27.4
31.6
21.7
0.5
0.0
1.2
74.0
28.0
0.0
0.0
0.0
11.6
0.0
0.0
0.5
1.0
0.5
7.0
0.1
0.3
0.2
0.5
0.1
0.1
0.0
0.3
0.3
0.0
0.0
0.4
0.0
Misc. Misc. Misc. Measured
1 Source 2 Source Source PM10
3 4 Concentration
0.0
0.0
5.0"
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.5°
1.5°
1.4°
0.0
1.0°
1.0"
1.5°
2.0"
2.2"
2.7"
1.3"
1.0"
5.1"
1.1"
1.5"
4.3"
0.0
0.0
1.0e
1.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.6k
1.2"
1.4"
0.0
O.lk
0.0
0.7k
6.4k
0.0
0.0
1.1°
5.9k
0.0
0.0
5.7°
0.0
0.0
0.0
0.0
' 0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0 j
0.0
0.0
0.0
0.0 >
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
64.0
23.4
105.0
59.0
55.0
27.0
12.0
79.5
55.0
69.0
67.6
79.6
52.5
54.6
48.1
71.5
58.0
47.8
51.9
46.1
96.1
64.0
87.4
114.8
112.0
87.0
17.4
-------
TABLE 5-12 (cont'd). RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM10
,ug/m3
Sampling Site
Stockton, CA (Chow et al., 1992b)
Pocatello, ID (Houck et al., 1992)
S. Chicago, IL (Hopke et al., 1988)
S.E. Chicago, IL (Vermette et al., 1992)
Reno, NV (Chow et al., 1988)
Sparks, NV (Chow et al., 1988)
Follansbee, WV (Skidmore et al., 1992)
Mingo, OH (Skidmore et al., 1992)
Steubenville, OH (Skidmore et al., 1992)
"Smelter background aerosol.
Time Period
1989
1990
1986
1988
1986-1987
1986-1987
1991
1991
1991
Primary
Geological
34.4
8.3
27.2
14.7
14.9
15.1
10.0
12.0
8.3
Primary
Motor
Primary
Primary Vehicle Vegetative
Construction Exhaust
0.5 5.2
7.5 0.1
2.4 2.8
0.0 0.9
0.0 10.0
0.0 11.6
0.0 35.0
0.0 14.0
0.0 14.0
Burning
4.8
0.0
0.0
'0.0
1.9
13.4
0.0
4.1
0.8
Secondary Secondary Misc. Misc.
Misc.
Misc.
Measured
Ammonium Ammonium Source Source Source Source PM10
Sulfate
3.1
0.0
15.4
7.7
1.3
2.7
16.0
15.0
14.0
Residual oil combustion.
'Cement plant sources, including kiln stacks, gypsum pile, and kiln area.
"Copper ore.
""Copper tailings.
"Copper smelter building.
'Heavy-duty diesel exhaust emission.
background aerosol.
'Marine aerosol, road salt, and sea salt plus sodium nitrate.
'Motor vehicle exhaust from diesel and leaded gasoline.
Secondary organic carbon.
Biomass burning.
Primary crude oil.
NaCl + NaNO .
Lime.
Road sanding material.
Asphalt industry.
3
Nitrate 1
7.0 0.7
0.0 0.0
15.1
0.8
0.6 0.0
0.9 0.0
9.3
3.4
3.8
Regional sulfate.
Steel mills.
Refuse incinerator.
Local road dust, coal
Incineration.
Unexplained mass.
2
1.8"
0.0
2.2"
0.3"
0.0
0.0
0.0
11. Ox
5.0X
3
0.0k
84.1
0.0
1.1"
0.0
0.2
0.0
0.0
0.0
4
0.0
0.0
0.0
7.7s
0.0
0.0
0.0
0.0
0.0
Concentration
62.4
100.0
80.1
4f.O
30.0
41.0
66.0
60.0
46.0
yard road dust, steel haul road dust.
Phosphorus/phosphate industry.
-------
Dust sources constitute 88% of the annual average PM10 National Emissions Inventory
(U.S. Environmental Protection Agency, 1994), but they average more than 50% of the
contribution to average PM10 concentrations in only about 40% of the entries shown in Table 5-
12. The reasons for this apparent discrepancy are not clear. In addition to errors in inventories
or source apportionments, weather-related factors (wind speed and ground wetness) and the
dominance of local sources on spatial scales too small to be captured in inventories may be
involved. It should be remembered that dust emissions are widely dispersed and highly
sporadic. Dust particles also have short atmospheric residence times and as a result their
dominance in emissions inventories may not be reflected in samples collected near specific
sources. The contributions from primary motor vehicle exhaust, residential wood combustion,
and industrial sources would be underestimated if values from the National Emissions Inventory
Trends data base (U.S. Environmental Protection Agency, 1994) were used. Some of these
deficiencies, such as fugitive dust emissions, are probably intractable, and the best that can be
done is to estimate the uncertainties in these emissions and to value the data accordingly when
decisions are made.
In Table 5-12, primary motor vehicle exhaust contributions account for up to 40% of
average PM10 at many of the sampling sites. Vehicle exhaust contributions are also variable at
different sites within the same study area. The mean value and the variability of motor vehicle
exhaust contributions reflects the proximity of sampling sites to roadways and traffic conditions
during the time of sampling. Vegetation burning, which includes agricultural fires, wildfires,
prescribed burning, and residential wood combustion, was found to be significant at residential
sampling sites such as: Craycroft, Scottsdale, and West Phoenix, AZ; Fresno, Bakersfield, and
Stockton, CA; Sparks, NV; and Mingo, OH. The predominance of these contributions during
winter months and the local rather than regional coverage indicates that residential wood
combustion was the major sub-category, even though chemical profiles are too similar to
separate residential combustion from other vegetative burning sources. For example, Chow et
al. (1988) show substantial differences between the residential Sparks, NV and urban-
commercial Reno, NV burning contributions even though these sites are separated by less than
10 km. Sites near documented industrial activity show evidence of that activity, but not
necessarily involving primary particles emitted by point sources. Hayden, AZ, for example,
contains a large smelter, but the major smelter contributions appear to arise from fugitive
5-64
-------
emissions of copper tailings rather than stack emissions. Secondary sulfate contributions at
Hayden were low, even though SO2 emissions from the stack were substantial during the time of
the study. Fellows, CA is in the midst of oilfield facilities that burn crude oil for tertiary oil
extraction. These have been converted to natural gas since the 1988 to 1989 study period. The
Follansbee, WV, Mingo, OH, and Steubenville, OH sites are all close to each other in the Ohio
River Valley and show evidence of the widespread steel mill emissions in that area.
Marine aerosol is found, as expected, at coastal sites such as Long Beach (average 3.8% of
total mass), and San Nicolas Island (25%). These contributions are relatively variable and are
larger at the more remote sites. Individual values reflect proximity to local sources. Of great
importance are the contributions from secondary ammonium sulfate and ammonium nitrate in
the western United States. These are especially noticeable at sites in California's San Joaquin
Valley (Bakersfield, Crows Landing, Fellows, Fresno, Kern Wildlife, and Stockton) and in the
Los Angeles area.
In addition to these commonly measured components, it is possible that isotopic ratios in
source emissions may vary in an informative way with the nature of the combustion process and
with the geologic age and character of the source input material. Carbon-14, for example, has
been used to separate contemporary carbon due to vegetative burning from carbon emitted by
fossil fuel combustion (Currie et al., 1984). Organic compounds (Rogge et al., 1991, 1993a,
1993b, 1993c, 1993d, 1993e; Lowenthal et al., 1994; Hildemann et al., 1991, 1993) show great
promise for further differentiation among sources, but measurement methods need to be
standardized and made more cost-effective to take advantage of extended chemical source
profiles.
Several aspects of the data in Table 5-12 limit the generalizations that can be drawn from
it:
• The source contribution estimates for the receptor sites shown are for different years and
different times of year. The results, therefore, do not show the temporal variability
which may exist in relative source contributions and should not be used to infer
differences between sites.
• Samples selected for chemical analysis are often biased toward the highest PM10 mass
concentrations in these studies, so average source contribution estimates are probably
not representative of annual averages.
5-65
-------
Many studies were conducted during the late 1980s, when a portion of the vehicle fleet
still used leaded gasoline. While the lead and bromine in motor vehicle emissions
facilitated the distinction of motor vehicle contributions from other sources, it was also
associated with higher emission rates than vehicles using unleaded fuels. Lead has been
virtually eliminated from vehicle fuels.
Uncertainties of source contribution estimates are not usually reported with the average
values summarized in Table 5-12. Estimates of standard errors are calculated in source
apportionment studies, and typically range from 15 to 30% of the source contribution
estimate. They are much higher when the chemical source profiles for different sources
are highly uncertain or too similar to distinguish one source from another.
Different measurement sites within the same airshed show different proportions of
contributions from the same sources. Most often, the sites in close proximity to an
emitter show a much larger contribution from that emitter than sites that are distant from
that emitter, even by distances as short as 10 km (e.g., Chow et al., 1988; 1992c).
Given the mass, trace element, ion, and carbon components measured in source and
receptor samples in most of the studies from Table 5-12, greater differentiation among
sources (e.g., diesel and gasoline vehicle exhaust, meat cooking and other organic
carbon sources, different sources of fugitive dust, and secondary aerosol precursors) is
not possible for the studies shown in Table 5-12.
5.6 SUMMARY AND CONCLUSIONS
Ambient particulate matter contains both primary and secondary components. Due to the
complexity of the composition of ambient PM10, sources are best discussed in terms of individual
constituents of both primary and secondary PM10. Each of these constituents can have
anthropogenic and natural sources, as shown in Tables 5-1A and 5-1B. The distinction between
natural and anthropogenic sources is not always obvious. While windblown dust might seem to
be the result of natural processes, highest emission rates are associated with agricultural
activities in areas that are susceptible to periodic drought. Examples include the dust bowl
region of the midwestern United States and the Sahel of Africa. Most forest fires in the United
States may ultimately be of human origin, either through prescribed burning or accident.
Windblown dust from whatever source represents the largest single source of PM in U.S.
and global emissions inventories. Although dust emissions (88% of total U.S. PM10)
5-66
-------
are far in excess of any other source of primary or secondary PM10 in any region of the country,
measurements of soil constituents in ambient samples suggest that the overall contribution from
this source could be much lower. The reasons for this apparent discrepancy are not clear. In
addition to errors in inventories or source apportionments, weather-related factors (wind speed
and ground wetness) and the dominance of local sources on spatial scales too small to be
captured in inventories may be involved. It should be remembered that dust emissions are
widely dispersed and highly sporadic. Dust particles also have short atmospheric residence
times and, as a result, their dominance in emissions inventories may not be reflected in samples
collected near specific sources.
There is a great deal of spatial and temporal variability which is still not reflected in
emissions inventories. Apart from seasonal variability, many of the sources discussed in this
chapter are highly episodic even within their peak emissions seasons. Examples include the
long-range transport of Saharan dust to the United States, regional dust storms, volcanism, and
forest fires. Their spatial variability is also evident. Annual estimates for an area can easily be
exceeded in a few days by unusual events involving these sources. Less dramatic examples of
strong seasonal variability, such as wood burned for home heating in the northwestern United
States, may be the major source of winter PM there.
It might be thought that enough data are available to adequately characterize mobile and
stationary source emissions. However, data characterizing the variability of PM emissions from
mobile sources are quite sparse. Available data suggest that elemental carbon followed by
organic carbon species are the major components of diesel particulate emissions, while organic
carbon emissions are larger than elemental carbon emissions in the case of gasoline fueled
vehicles.
Emissions from biomass burning are also composed mainly of organic carbon species and
elemental carbon, although the ratio of organic carbon to elemental carbon is much higher than
in motor vehicle emissions. Power plant emissions are not significant sources of aerosol carbon.
The fractional yield of secondary organic carbon from the oxidation of natural and
anthropogenic hydrocarbons is highly uncertain. Yields from the oxidation of anthropogenic
hydrocarbons are probably less than a few percent, and larger yields are found in the oxidation
of terpenes emitted by vegetation.
5-67
-------
As seen in Table 5-1B, emissions of surface dust, organic debris, and sea spray are
concentrated mainly in the coarse fraction of PM10 ( > 2.5 jim aero. diam.). A small fraction of
this material is in the PM25 size range ( < 2.5 jim aero, diam., c.f. Figure 5-1). Nevertheless,
concentrations of crustal material can be appreciable especially during dust events. It should
also be remembered that all of the Saharan dust reaching the United States is in the PM2 5 size
range. Emissions from combustion sources (mobile and stationary sources, biomass burning) are
predominantly in the PM2 5 size range.
As shown in Table 5-6, estimated primary PM10 emissions decreased by about 10% from
1985 through 1993. A high degree of variability is evident for emissions from miscellaneous
(fugitive dust, biomass burning, and agriculture) and natural (wind erosion of natural surfaces)
categories shown in Table 5-7. Estimated SO2 emissions decreased by several per cent from
1984 through 1993 as shown in Table 5-8. Estimated emissions of NOX show little variation
over the same time period as shown in Table 5-9. Emissions of primary PM10 are projected to
increase to the year 2010 mainly because of increases in fugitive dust emissions, while emissions
of SO2 and NOX are expected to decrease over the same time period.
Uncertainties in emissions inventories are difficult to quantify. They may be as low as
10% for well-defined sources (e.g., for SO2) and may range up to a factor of 10 or so for
windblown dust. As a rule, total PM emissions rates should be regarded as order-of-magnitude
estimates. Because of the large uncertainty associated with emissions of suspended dust, trends
of total PM10 emissions should be viewed with caution and emissions from specific source
categories are best discussed on an individual basis.
Emissions inventories are generally not the most appropriate way to apportion material in
ambient samples. Receptor modeling has proven to be an especially valuable tool in this regard.
Compositional profiles developed for receptor modeling applications are perhaps the most
accessible and reliable means to characterize the composition of emissions. Quoted
uncertainties in source apportionments of constituents in ambient aerosol samples typically range
from 15 to 30%. Receptor modeling studies in the western United States have found that motor
vehicles and fugitive dust are major sources of PM10. Likewise, a limited number of studies in
the eastern United States have found that fossil fuel combustion and fugitive dust are major
sources of PM10. Techniques are currently being developed to use receptor
5-68
-------
modeling techniques along with ambient data to refine emissions inventory estimates. Because
of the site-specific nature of receptor modeling results, more rigorous methods for determining
site locations and methods for applying receptor model results to larger spatial scales are needed
for this purpose. Again, it should be emphasized that, because of limitations in receptor
modeling methods in treating secondary components, these efforts are more likely to be
successful for primary components, although it should be mentioned that methods are being
developed to apportion secondary constituents.
5-69
-------
REFERENCES
Alfaro, S. C.; Gomes, L. (1995) Improving the large-scale modeling of the saltation flux of soil particles in presence of
nonerodible elements. J. Geophys. Res. [Atmos.] 100: 16,357-16,366.
Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the United States.
Atmos. Environ. 17: 2131-2165.
Altshuller, A. P.; Bufalini, J. J. (1965) Photochemical aspects of air pollution: a review. Photochem. Photobiol. 4:
97-146.
Andreae, M. O. (1991) Biomass burning: its history, use, and distribution and its impact on environmental quality and
global climate. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
implications. Cambridge, MA: MIT Press; pp. 1-21.
Ashbaugh, L. L.; Myrup, L. O.; Flocchini, R. G. (1984) A principal component analysis of sulfur concentrations in the
western United States. Atmos. Environ. 18: 783-791.
Ashbaugh, L. L.; Malm, W. C.; Sadeh, W. Z. (1985) A residence time probability analysis of sulfur concentrations at
Grand Canyon National Park. Atmos. Environ. 19: 1263-1270.
Bagnold, R. A. (1941) The physics of blown sand and desert dunes. London, United Kingdom: Methuen & Co.
Behymer, T. D.; Kites, R. A. (1984) Similarity of some organic compounds in spark-ignition and diesel engine
particulate extracts. Environ. Sci. Technol. 18: 203-206.
Belly, P.-Y. (1964) Sand movement by wind. Washington, DC: U.S. Army Corps of Engineers, Coastal Engineering
Research Center; technical memorandum no. 1.
Bisal, F.; Hsieh, J. (1966) Influence of moisture on erodibility of soil by wind. Soil Sci. 102: 143-146.
Brown, M. (1993) Deduction of emissions of source gases using an objective inversion algorithm and a chemical
transport model. J. Geophys. Res. [Atmos.] 98: 12,639-12,660.
Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some reflections
on controlling vehicle emissions. Science 261: 37-45.
Cheng, M.-D.; Lioy, P. J.; Opperman, A. J. (1988) Resolving PM 10 data collected in New Jersey by various
multivariate analysis techniques. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10: implementation of standards,
an APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution
Control Association; pp. 472-483. (APCA transactions series TR-13).
Cheng, M.-D.; Gao, N; Hopke, P. K. (1996) Source apportionment study of nitrogen species measured in sourthern
California in 1987. J. Environ. Eng. (N. Y.) 122: in press.
Chepil, W. S. (1945) Dynamics of wind erosion. Soil Sci. 60: 305-320.
Chepil, W. S. (1956) Influence of moisture on erodibility of soil by wind. Soil Sci. Soc. Am. Proc. 20: 288-292.
Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal source
contributions to PM10 and PM25 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10: implementation
of standards, an APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA:
Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
5-70
-------
Chow, J. C.; Watson, J. G.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietnch, D. L.; Moon, D.; Sloane, C. (1991)
The 1989-90 Phoenix PM10 study, v. II: source apportionment [final report]. Phoenix, AZ: Arizona Department
of Environmental Quality; report no. 8931.6F1; Table 3-1.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hinsvark, B. A.; Prichett, L. C.; Neuroth, G. R. (1992a)
Wintertime PM10 and PM^ chemical compositions and source contributions in Tucson, Arizona. In: Chow, J.
C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an A&WMA/EPA
international specially conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management
Association; pp. 231-243. (A&WMA transactions series no. 22).
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W. (1992b)
PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A 26: 3335-3354.
Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
PM10 source contributions in Rubidoux, California. Atmos. Environ. Part A 26: 693-706.
Chow, J. C.; Watson, J. G.; Houck, J. E.; Pritchett, L. C.; Rogers, C. F.; Frazier, C. A.; Egami, R. T.; Ball, B. M. (1994)
A laboratory resuspension chamber to measure fugitive dust size distributions and chemical compositions.
Atmos. Environ. 28: 3463-3481.
Cofer, W. R.; Levine, J. S.; Sebacher, D. L; Winstead, E. L.; Riggan, P. J.; Brass, J. A.; Ambrosia, V. G. (1988) J.
Geophys. Res. [Atmos.] 93: 5207-5212.
Cooper, J. A.; Watson, J. G., Jr. (1980) Receptor oriented methods of air particulate source apportionment. J. Air Pollut.
Control Assoc. 30: 1116-1125.
Cooper, J. A.; Redline, D. C.; Sherman, J. R.; Valdovinos, L. M.; Pollard, W. L.; Scavone, L. C.; Badgett-West, C.
(1987) PM10 source composition library for the South Coast Air Basin: volume I, source profile development
documentation final report. El Monte, CA: South Coast Air Quality Management District; July 15.
Currie, L. D.; Gerlach, R. W.; Lewis, C. W.; Balfour, W. D.; Cooper, J. A.; Dattner, S. L.; De Cesar, R. T.; Gordon, G.
E.; Heisler, S. L.; Hopke, P. K.; Shah, J. J.; Thurston, G. D.; Williamson, H. J. (1984) Intel-laboratory
comparison of source apportionment procedures: results for simulated data sets. Atmos. Environ. 18:
1517-1537.
Daisey, J. M.; Cheney, J. L.; Lioy, P. J. (1986) Profiles of organic particulate emissions from air pollution sources:
status and needs for receptor source apportionment modeling. J. Air Pollut. Control Assoc. 36: 17-33.
Dasch, J. M. (1982) Particulate and gaseous emissions from wood-burning fireplaces. Environ. Sci. Technol.
16: 639-645.
Dattner, S. L.; Hopke, P. K. (1982) Receptor models applied to contemporary pollution problems: an introduction. In:
Frederick, E. R., ed. A specialty conference on: receptor models applied to contemporary pollution problems;
October; Danvers, MA. Pittsburgh, PA: Air Pollution Control Association; pp. 1-5. (APCA specially
conference proceedings SP-48).
Dentener, F. J.; Crutzen, P. J. (1994) A three-dimensional model of the global ammonia cycle. J. Atmos. Chem. 19:
331-369.
Dzubay, T. G.; Mamane, Y. (1989) Use of electron microscopy data in receptor models forPM-10. Atmos. Environ.
23: 467-476.
Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, L; Sheffield, A. E.; Courtney, W. J. (1988) A composite receptor
method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
5-71
-------
Federal Register. (1977) Method 5 - determination of particulate emissions from stationary sources. F. R. (August 18)
42: 41776-41783.
Freeman, D. L.; Chow, J. C.; Egami, R. T.; Watson, J. G. (1989) A receptor and dispersion modeling software package.
In: Watson, J. G., ed. Receptor models in air resources management: an international specialty conference;
February 1988; San Francisco, CA. Pittsburgh, PA: Air & Waste Management Association; pp. 243-268.
(APCA transactions series: no. 14).
Friedlander, S. K. (1973) Chemical element balances and identification of air pollution sources. Environ. Sci. Technol.
7: 235-240.
Fujita, E. M.; Croes, B. E.; Bennett, C. L.; Lawson, D. R.; Lurmann, F. W.; Main, H. H. (1992) Comparison of emission
inventory and ambient concentration ratios of CO, NMOG, and NOxin California's south coast air basin. J. Air
Waste Manage. Assoc. 42: 264-276.
Gao, N.; Cheng, M.-D.; Hopke, P. K. (1993) Potential source contribution function analysis and source apportionment
of sulfur species measured at Rubidoux, CA during the Southern California Air Quality Study, 1987. Anal.
Chim.Acta 277: 369-380.
Gatz, D. F. (1995) A large silicon-aluminum aerosol plume in central Illinois: Saharan Desert dust? Atmos. Environ.:
accepted.
Gebhart, K. A.; Lattimer, D. A.; Sisler, J. F. (1990) Empirical orthogonal function analysis of the particulate sulfate
concentrations measured during WHITEX. In: Mathai, C. V., ed. Visibility and fine particles:
an A&WMA/EPA international specialty conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air &
Waste Management Association; pp. 860-871. (A&WMA transactions series no. TR-17).
Geron, C. D.; Guenther, A. B.; Pierce, T. E. (1994) An improved model for estimating emissions of volatile organic
compounds from forests in the eastern United States. J. Geophys. Res. [Atmos.] 99: 12,773-12,791.
Gillette, D. A.; Adams, J.; Endo, A.; Smith, D.; Kihl, R. (1980) Threshold velocities for input of soil particles into the
air by desert soils. J. Geophys. Res. C: Oceans Atmos. 85: 5621-5630.
Gillette, D. A.; Adams, J.; Muhs, D. R.; Khil, R. (1982) Threshold friction velocities and rupture moduli for crusted
desert soils for the input of soil particles into the air. J. Geophys. Res. B 87: 9003-9015.
Glover, D. M.; Hopke, P. K.; Vermette, S. J.; Landsberger, S.; D'Auben, D. R. (1991) Source apportionment with site
specific source profiles. J. Air Waste Manage. Assoc. 41: 294-305.
Goklany, I. M.; Hoffnagle, G. F.; Brackbill, E. A. (1984) Acidic and total primary sulfates: development of emission
factors for major stationary combustion sources. J. AirPollut. Control Assoc. 34: 124-134.
Gordon, G. E. (1980) Receptor models. Environ. Sci. Technol. 14: 792-800.
Gordon, G. E. (1988) Receptor models. Environ. Sci. Technol. 22: 1132-1142.
Gray, H. A.; Landry, B.; Liu, C. S.; Henry, R. C.; Cooper, J. A.; Sherman, J. R. (1988) Receptor modeling for PM10
source apportionment in the South Coast Air Basin of California. In: Mathai, C. V.; Stonefield, D. H., eds.
PM-10: implementation of standards, an APCA/EPA international specialty conference; February; San
Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 399-418. (APCA transactions series
TR-13).
Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols. Atmos.
Environ. 23: 1733-1747.
5-72
-------
Haagen-Smit, A. J. (1952) Chemistry and physiology of Los Angeles smog. Ind. Eng. Chem. 44: 1342-1346.
Hall, F. F., Jr. (1981) Visibility reductions from soil dust in the western U.S. In: White, W. H.; Moore, D. J.; Lodge, J.
P., Jr., eds. Plumes and visibility: measurements and model components: proceedings of the symposium;
November 1980; Grand Canyon National Park, AZ. Atmos. Environ. 15: 1929-1933.
Hammerle, R. H.; Kormski, T. J.; Weir, J. E.; Chladek, E.; Glerczak, C. A.; Chase, R. E.; Hurley, R. G. (1992) Effect of
mileage accumulation on particulate emissions from vehicles using gasoline with methylcyclopentadienyl
manganese tricarbonyl. Presented at: International congress & exposition; February; Detroit, MI. Warrendale,
PA: Society of Automotive Engineers; SAE technical paper no. 920731.
Hansen, A. D. A.; Rosen, H. (1990) Individual measurements of the emission factor of aerosol black carbon in
automobile plumes. J. Air Waste Manage. Assoc. 40: 1654-1657.
Hatakeyama, S.; Izumi, K.; Fukuyama, T.; Akimoto, H. (1989) Reactions of ozone with a-pinene and p-pinene in air:
yields of gaseous and particulate products. J. Geophys. Res. [Atmos.] 94: 13,013-13,024.
Hatakeyama, S.; Izumi, K.; Fukuyama, T.; Akimoto, H.; Washida, N. (1991) Reactions of OH with a-pinene and
P-pinene in air: estimate of global CO production from the atmospheric oxidation of terpenes. J. Geophys. Res.
[Atmos.] 96: 947-958.
Henry, R. C. (1991) Multivariate receptor models. In: Hopke, P. K., ed. Receptor modeling for air quality management.
Amsterdam, The Netherlands: Elsevier; pp. 117-147.
Henry, R. C.; Hidy, G. M. (1979) Multivariate analysis of particulate sulfate and other air quality variables by principal
components—part I: annual data from Los Angeles and New York. Atmos. Environ. 13: 1581-1596.
Henry, R. C.; Hidy, G. M. (1982) Multivariate analysis of particulate sulfate and other air quality variables by principal
components—II: Salt Lake City, Utah and St. Louis, Missouri. Atmos. Environ. 16: 929-943.
Henry, R. C.; Kim, B. M. (1989) A factor analysis receptor model with explicit physical constraints. In: Watson, J. G.,
ed. Receptor models in air resources management: an international specialty conference; February 1988; San
Francisco, CA. Pittsburgh, PA: Air & Waste Management Association; pp. 214-225. (APCA transactions
series: no. 14).
Henry, R. C.; Wang, Y.-J.; Gebhart, K. A. (1990) The relationship between empirical orthogonal functions and sources
of air pollution. Atmos. Environ. Part A 24: 503-509.
Hertel, O.; Christensen, J.; Hov, 0, (1994) Modelling of the end products of the chemical decomposition of DMS in the
marine boundary layer. Atmos. Environ. 28: 2431-2449.
Hildemann, L. M.; Markowski, G. R.; Jones, M. C.; Cass, G. R. (1991) Submicrometer aerosol mass distributions of
emissions from boilers, fireplaces, automobiles, diesel trucks, and meat-cooking operations. Aerosol Sci.
Technol. 14: 138-152.
Hildemann, L. M.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1993) Mathematical modeling of urban organic
aerosol: properties measured by high-resolution gas chromatography. Environ. Sci. Technol. 27: 2045-2055.
Hooker, C. L.; Westberg, H. H.; Sheppard, J. C. (1985) Determination of carbon balances for smog chamber terpene
oxidation experiments using a 14C tracer technique. J. Atmos. Chem. 2: 307-320.
Hopke, P. K. (1985) Receptor modeling in environmental chemistry. New York, NY: John Wiley & Sons.
Hopke, P. K., ed. (1991) Receptor modeling for air quality management. New York, NY: Elsevier Science Publishing
Company, Inc. (Vandeginste, B. G. M.; Kvalheim, O. M., eds. Data handling in science and technology: v. 7).
5-73
-------
Hopke, P. K.; Lamb, R. E.; Natusch, D. F. S. (1980) Multielemental characterization of urban roadway dust. Environ.
Sci. Technol. 14: 164-172.
Hopke, P. K.; Wlaschin, W.; Landsberger, S.; Sweet, C.; Vermette, S. J. (1988) The source apportionment of PM10 in
south Chicago. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10: implementation of standards, an APCA/EPA
international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution Control
Association; pp. 484-494. (APCA transactions series TR-13).
Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M; Frazier, C. A. (1989)
Determination of particle size distribution and chemical composition of particulate matter from selected sources
in California: volume I and executive summary (final report). Sacramento, CA: California Air Resources
Board.
Houck, J. E.; Goulet, J. M.; Chow, J. C.; Watson, J. G.; Pritchett, L. C. (1990) Chemical characterization of emission
sources contributing to light extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA
international specially conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air & Waste Management
Association; pp. 437-446. (A&WMA transactions series no. TR-17).
Houck, J. E.; Rau, J. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM10 nonattainment area.
In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an
A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste
Management Association; pp. 219-230. (A&WMA transactions series no. 22).
Huang, X.; Olmez, L; Aras, N. K.; Gordon, G. E. (1994) Emissions of trace elements from motor vehicles: potential
marker elements and source composition profile. Atmos. Environ. 28: 1385-1391.
Ingalls, M. N. (1989) On-road vehicle emission factors from measurements in a Los Angeles area tunnel. Presented at:
82nd annual meeting & exhibition of the Air and Waste Management Association; June; Anaheim, CA.
Pittsburgh, PA: Air and Waste Management Association; paper no. 89-137.3.
Intergovernmental Panel on Climate Change. (1995) Climate change 1995: the science of climate change: draft
summary for policymakers (SPM) of the draft contribution of Working Group I to the IPCC second assessment
report. Geneva, Switzerland: World Meteorological Association, IPCC Secretariat; November.
Izumi, K.; Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of NOX.
Atmos. Environ. Part A 24: 1433-1441.
Javitz, H. S.; Watson, J. G.; Guertin, J. P.; Mueller, P. K. (1988) Results of a receptor modeling feasibility study.
JAPCA 38: 661-667.
Johnson, J. H. (1988) Automotive emissions. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the
automobile, and public health. Washington, D.C.: National Academy Press; pp. 39-76.
Johnson, B.; Cooke, R. C. (1979) Bubble populations and spectra in coastal waters: a photographic approach.
J. Geophys. Res. C: Oceans Atmos. 84: 3763-3766.
Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources. Environ.
Sci. Technol. 29: 19-28.
Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and ambient
concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan development in
the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source
5-74
-------
controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
Air & Waste Management Association; pp. 979-991. (A&WMA transactions series no. 22).
Kleinman, M. T.; Pasternack, B. S.; Eisenbud, M.; Kneip, T. J. (1980) Identifying and estimating the relative
importance of sources of airborne particulates. Environ. Sci. Technol. 14: 62-65.
Koutrakis, P.; Spengler, J. D. (1987) Source apportionment of ambient particles in Steubenville, OH using specific
rotation factor analysis. Atmos. Environ. 21: 1511-1519.
Lang, J. M.; Snow, L.; Carlson, R.; Black, F.; Zweidinger, R.; Tejada, S. (1982) Characterization of particulate
emissions from in-use gasoline-fueled motor vehicles. Warrendale, PA: Society of Automotive Engineers; SAE
technical paper no. 811186.
Lewis, C. W.; Stevens, R. K. (1987) Hybrid receptor models. In: Johnson, R. W.; Gordon, G. E.; Calkins, W.;
Elzerman, A. Z., eds. The chemistry of acid rain: sources and atmospheric processes; developed from
a symposium sponsored by the Divisions of Petroleum Chemistry, Inc., Nuclear Chemistry and Technology,
Environmental Chemistry, and Fuel Chemistry at the 191st meeting of the American Chemical Society; April
1986; New York, NY. Washington, DC: American Chemical Society; pp. 58-65. (ACS symposium series:
349).
Lewis, C. W.; Baumgardner, R. E.; Stevens, R. K.; Claxton, L. D.; Lewtas, J. (1988) Contribution of woodsmoke and
motor vehicle emissions to ambient aerosol mutagenicity. Environ. Sci. Technol. 22: 968-971.
Liousse, C.; Penner, J. E.; Chuang, C.; Walton, J. J.; Eddleman, H.; Cachier, H. (1996) A global three-dimensional
model study of carbonaceous aerosols. J. Geophys. Res. [Atmos.] 101: in press.
Lioy, P. J.; Mallon, R. P.; Lippmann, M.; Kneip, T. J.; Samson, P. J. (1982) Factors affecting the variability of
summertime sulfate in a rural area using principal component analysis. J. Air Pollut. Control Assoc. 32:
1043-1047.
Lowenthal, D. H.; Zielinska, B.; Chow, J. C.; Watson, J. G.; Gautam, M.; Ferguson, D. H.; Neuroth, G. R.; Stevens, K.
D. (1994) Characterization of heavy-duty diesel vehicle emissions. Atmos. Environ. 28: 731-743.
Macias, E. S.; Hopke, P. K., eds. (1981) Atmospheric aerosol: source/air quality relationships; based on a symposium
jointly sponsored by the Divisions of Nuclear Chemistry and Technology at the 180th national meeting of the
American Chemical Society; August 1980; Las Vegas, NV. Washington, DC: American Chemical Society.
(Comstock, M. J., ed. ACS symposium series: 167).
Magliano, K. L. (1988) Level 1 PM10 assessment in a California air basin. In: Mathai, C. V.; Stonefield, D. H., eds.
PM-10: implementation of standards, an APCA/EPA international specially conference; February; San
Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 508-517. (APCA transactions series
TR-13).
Marticorena, B.; Bergametti, G. (1995) Modeling the atmospheric dust cycle: I. design of a soil-derived dust emission
scheme. J. Geophys. Res. [Atmos.] 100: 16,415-16,430.
Mason, B. (1966) Principles of geochemistry. 3rd ed. New York, NY: John Wiley & Sons, Inc.
Matthews, E. (1994) Nitrogenous fertilizers: global distribution of consumption and associated emissions of nitrous
oxide and ammonia. Global Biogeochem. Cycles 8: 411-439.
Matthias-Maser, S.; Jaenicke, R. (1994) Examination of atmospheric bioaerosol particles with radii > 0.2 ,um.
J. Aerosol Sci. 25: 1605-1613.
5-75
-------
Mazurek, M. A.; Gofer, W. R., Ill; Levine, J. S. (1991) Carbonaceous aerosols from prescribed burning of a boreal
forest ecosystem. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
implications. Cambridge, MA: MIT Press; pp. 258-263.
Mylonas, D. T.; Allen, D. T.; Ehrman, S. H.; Pratsinis, S. E. (1991) The sources and size distributions of organonitrates
in Los Angeles aerosol. Atmos. Environ. Part A 25: 2855-2861.
NEA. (1990a) Enhancement of the south coast air basin source profile library for chemical mass balance receptor
model applications: volume I, final report. Beaverton, OR: NEA.
NEA. (1990b) Enhancement of the south coast air basin source profile library for chemical mass balance receptor
model applications: volume II, source profile library supplement. Beverton, OR: NEA.
NEA. (1990c) Enhancement of the south coast air basin source profile library for chemical mass balance receptor
model applications: volume III, appendices. Beverton, OR: NEA.
National Research Council. (1982) Diesel cars: benefits, risks and public policy: final report of the Diesel Impacts
Study Committee. Washington, DC: National Academy Press.
Nriagu, J. O. (1989) A global assessment of natural sources of atmospheric trace metals. Nature (London) 338: 47-49.
Nriagu, J. O.; Pacyna, J. M. (1988) Quantitative assessment of worldwide contamination of air, water and soils by trace
metals. Nature (London) 333: 134-139.
O'Connor, B. H.; Kerrigan, G. C.; Hinchliffe, P. (1977) The loss of Br from thin-film samples during X-ray
fluorescence analysis. X-Ray Spectrom. 6: 83-85.
Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pntchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett, R. L.
(1988) Compositions of particles from selected sources in Philadelphia for receptor modeling applications.
JAPCA38: 1392-1402.
Ott, S.-T.; Ott, A.; Martin, D. W.; Young, J. A. (1991) Analysis of a trans-Atlantic saharan dust outbreak based on
satellite and GATE data. Mon. Weather Rev. 119: 1832-1850.
Pace, T. G., ed. (1986) Receptor methods for source apportionment: real world issues and applications, an APCA
international specialty conference; March 1985; Williamsburg, VA. Pittsburgh, PA: Air Pollution Control
Association. (APCA transactions no. TR-5).
Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of isoprene
and p-pinene. Atmos. Environ. Part A 25: 997-1008.
Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and transport.
Atmos. Environ. Part A 26: 2269-2282.
Peterson, J. T. (1970) Distribution of sulfur dioxide over metropolitan St. Louis, as described by empirical
eigenvectors, and its relation to meteorological parameters. Atmos. Environ. 4: 501-518.
Pickle, T.; Allen, D. T.; Pratsinis, S. E. (1990) The sources and size distributions of aliphatic and carbonyl carbon in
Los Angeles aerosol. Atmos. Environ. Part A 24: 2221-2228.
Pierson, W. R.; Brachaczek, W. W. (1974) Airborne particulate debris from rubber tires. Presented at: Joint meeting of
the American Chemical Society, Rubber Division, and the Chemical Institute of Canada, Macromolecular
Science Division; May; Toronto, Ontario, Canada. Rubber Chem. Technol. 47: 1275-1299.
5-76
-------
Pierson, W. R.; Brachaczek, W. W. (1976) Particulate matter associated with vehicles on the road. Presented at: SAE
automotive engineering congress and exposition; February; Detroit, MI. Warrendale, PA: Society of
Automotive Engineers; SAE technical paper no. 760039.
Pierson, W. R.; Brachaczek, W. W. (1983) Particulate matter associated with vehicles on the road. II. Aerosol Sci.
Technol. 2: 1-40.
Pierson, W. R.; Gertler, A. W.; Bradow, R. L. (1990) Comparison of the SCAQS tunnel study with other on-road
vehicle emission data. J. Air Waste Manage. Assoc. 40: 1495-1504.
Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, Sz.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995) Chemical
characterization of the ambient aerosol in Philadelphia. Presented at: Particulate matter: health and regulatory
issues: an A&WMA international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
Management Association; submitted abstract.
Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition processes. In:
Irving, P. M., ed. Acidic depostion: state of science and technology, volume I: emissions, atmospheric
processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment Program. (State
of science and technology report no. 1).
Prospero, J. M. (1995) The atmospheric transport of particles to the ocean. In: Ittekkot, V.; Honjo, S.; Depetris, P. J.,
eds. SCOPE report: particle flux in the ocean. New York, NY: John Wiley & Sons, Ltd.; in press.
Prospero, J. M.; Nees, R. T.; Uematsu, M. (1987) Deposition rate of particulate and dissolved aluminum derived from
Saharan dust in precipitation at Miami, Florida. J. Geophys. Res. [Atmos.] 92: 14,723-14,731.
Prospero, J. M.; Savoie, D. L.; Arimoto, R.; Huang, F. (1993) Long-term trends in African mineral dust concentrations
over the western North Atlantic: relationship to North African rainfall. EOS Trans. Am. Geophys. Union 74:
146.
Radke, L. F.; Hegg, D. A.; Hobbs, P. V.; Nance, J. D.; Lyons, J. H.; Laursen, K. K.; Weiss, R. E.; Riggan, P. J.; Ward,
D. E. (1991) Particulate and trace gas emissions from large biomass fires in North America. In: Levine, J. S.,
ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge, MA: MIT Press;
pp. 209-224.
Radwan, C. (1995) Particulate emissions trends predicted in the PARTS model. In: Particulate matter: health and
regulatory issues: proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA:
Air & Waste Management Association; pp. 599-607. (A&WMA publication VIP-49).
Rasmussen, R. A.; Went, F. W. (1965) Volatile organic material of plant origin in the atmosphere. Proc. Natl. Acad.
Sci. U. S. A. 53:215-220.
Reeser, W. K.; Zimmer, R. A.; Cummins, P.; Briggs, K. R. (1992) An episodic PM10 emissions inventory for Denver,
Colorado. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an
A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste
Management Association; pp. 131-145. (A&WMA transactions series no. 22).
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci. Technol.
27:636-651.
5-77
-------
Rogge, W. F.; Hildemann, L. M; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol. 27:
2700-2711.
Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
Sci. Technol. 27: 1892-1904.
Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993e) Quantification of urban
organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part A
27: 1309-1330.
Ryan, W. M.; Badgett-West, C. R.; Holtz, D. R.; Peters, T. A.; Cooper, J. A.; Ono, D. (1988) Reconciliation of receptor
and dispersion modeling impacts of PM10 in Hay den, Arizona. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10:
implementation of standards, an APCA/EPA international specialty conference; February; San Francisco, CA.
Pittsburgh, PA: Air Pollution Control Association; pp. 419-429. (APCA transactions series TR-13).
Savoie, D. L.; Prospero, J. M. (1982) Particle size distribution of nitrate and sulfate in the marine atmosphere. Geophys.
Res. Lett. 9: 1207-1210.
Shao, Y.; Raupach, M. R.; Findlater, P. A. (1993) Effect of saltation bombardment on the entrainment of dust by wind.
J. Geophys. Res. [Atmos.] 98: 12,719-12,726.
Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PM10 air quality assessment for the Jefferson County, Ohio air
quality control region. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source
controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
Air and Waste Management Association; pp. 1016-1031. (A&WMA transactions series no. 22).
Snow, J. T.; McClelland, T. M. (1990) Dust devils at White Sands Missile Range, New Mexico: 1. temporal and spatial
distributions. J. Geophys. Res. [Atmos.] 95: 13,707-13,721.
South Coast Air Quality Management District. (1994) State implementation plan for PM10 in the Coechella Valley:
1994 "BACM" revision. Diamond Bar, CA: South Coast Air Quality Management District.
Spendlove, J. C. (1974) Industrial, agricultural, and municipal microbial aerosol problems. In: Current problems in
aerobiology: proceedings of the thirtieth general meeting of the Society for Industrial Microbiology; August
1973; Evanston, IL. Dev. Ind. Microbiol. 15: 20-27.
Stevens, R. K.; Pace, T. G. (1984) Overview of the mathematical and empirical receptor models workshop (Quail Roost
II). Atmos. Environ. 18: 1499-1506.
Stevenson, H. J. R.; Sanderson, D. E.; Altshuller, A. P. (1965) Formation of photochemical aerosols. Int. J. Air Water
Pollut. 9: 367-375.
Susott, R. A.; Ward, D. E.; Babbitt, R. E.; Latham, D. J. (1991) The measurement of trace emissions and combustion
characteristics for a mass fire. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and
biospheric implications. Cambridge, MA: MIT Press; pp, 245-257.
Svasek, J. N; Terwindt, J. H. J. (1974) Measurements of sand transport by wind on a natural beach. Sedimentology 21:
311-322.
5-78
-------
Szkarlat, A. C.; Japar, S. M. (1983) Optical and chemical properties of particle emissions from on-road vehicles. J. Air
Pollut. Control Assoc. 33: 592-597.
Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international specialty
conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 244-261.
(A&WMA transactions series no. 22).
Truex, T. I; Pierson, W. R.;McKee,D. E. (1980) Sulfate in diesel exhaust. Environ. Sci. Technol. 14: 1118-1121.
Turekian, K. K. (1971) Elements, geochemical distribution of. In: Encyclopedia of science and technlogy, v. 4. New
York, NY: McGraw-Hill Book Company; pp. 627-630.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
Turpin, B. J.; Huntzicker, J. J.; Larson, S. M.; Cass, G. R. (1991) Los Angeles summer midday particulate carbon:
primary and secondary aerosol. Environ. Sci. Technol. 25: 1788-1793.
U.S. Environmental Protection Agency. (1989) Procedures for estimating and applying rule effectiveness in post-1987
base year emission inventories for ozone and carbon monoxide state implementation plans. Research Triangle
Park, NC: Office of Air Quality Planning and Standards.
U.S. Environmental Protection Agency. (1993) Motor vehicle-related air toxics study. Ann Arbor, MI: Office of
Mobile Sources; EPA report no. EPA/420/R-93/005. Available from: NTIS, Springfield, VA;
PB93-182590/XAB.
U.S. Environmental Protection Agency. (1994) National air quality and emissions trends report, 1993. Research
Triangle Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA 454/R-94-026.
U.S. Environmental Protection Agency. (1995a) Compilation of air pollutant emission factors, volume I: stationary
point and area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
report no. AP-42.
U.S. Environmental Protection Agency. (1995b) National air pollutant emission trends, 1900 - 1994. Research Triangle
Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-95-011.
Van Espen, P. (1984) A program for the processing of analytical data (DPP). Anal. Chim. Acta 165: 31-49.
Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM10 source apportionment using local surface dust profiles:
examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source
controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA:
Air & Waste Management Association; pp. 262-278. (A&WMA transactions series no. 22).
Vinogradov, A. P. (1959) The geochemistry of rare and dispersed chemical elements in soils. 2nd ed. New York, NY:
Consultants Bureau, Inc.
Walker, B. L.; Cooper, C. D. (1992) Air pollution emission factors for medical waste incinerators. J. Air Waste
Manage. Assoc. 42: 784-791.
Warneck, P. (1988) Chemistry of the natural atmosphere. New York, NY: Academic Press, Inc.
Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ. 19:
241-266.
5-79
-------
Watson, J. G.; Henry, R. C.; Cooper, J. A.; Macias, E. S. (1981) The state of the art of receptor models relating ambient
suspended particulate matter to sources. In: Macias, E. S.; Hopke, P. K., eds. Atmospheric aerosol: source/air
quality relationships; based on a symposium jointly sponsored by the Divisions of Nuclear Chemistry and
Technology and Environmental Chemistry at the 180th national meeting of the American Chemical Society;
August 1980; Las Vegas, NV. Washington, DC: American Chemical Society; pp. 89-106. (Comstock, M. J.,
ed. ACS symposium series: 167).
Watson, J. G.; Chow, J. C.; Mathai, C. V. (1989) Receptor models in air resources management: a summary of the
APCA international specialty conference. JAPCA 39: 419-426.
Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Burns, S.; Ragazzi, R. A. (1990a) Composite source profiles
for particulate motor vehicle exhaust source apportionment in Denver, CO. In: Mathai, C. V., ed. Visibility and
fine particles: an A&WMA/EPA international specialty conference; October 1989; Estes Park, CO. Pittsburgh,
PA: Air & Waste Management Association; pp. 422-436. (A&WMA transactions series no. TR-17).
Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Ragazzi, R. A.; Bums, S. (1990b) Chemical source profiles
for particulate motor vehicle exhaust under cold and high altitude operating conditions. Sci. Total Environ. 93:
183-190.
Watson, J. G.; Chow, J. C.; Pace, T. G. (1991) Chemical mass balance. In: Hopke, P. K., ed. Data handling in science
and technology: v. 7, receptor modeling for air quality management. New York, NY: Elsevier Press; pp.
83-116.
Watson, J. G.; Chow, J. C.; Lowenthal, L. C.; Pntchett, C. A.; Frazier, C. A.; Neuroth, G. R.; Robbms, R. (1994a)
Differences in the carbon compostion of source profiles for diesel- and gasoline-powered vehicles. Atmos.
Environ. 28: 2493-2505.
Watson, J. G.; Chow, J. C.; Blumenthal, D. L.; Lurmann, F. W.; Hackney, R. J.; Magliano, K. L.; Pederson, J. R.; Neff,
W. D.; Roth, P. M.; Solomon, P. A.; Thuillier, R. H.; Ziman, S. D. (1994b) Planning for data analysis. In:
Solomon, P. A.; Silver, T. A., eds. Planning and managing regional air quality modeling and measurement
studies: a perspective through the San Joaquin Valley air quality study and AUSPEX. Boca Raton, FL: Lewis
Publishers; pp. 335-349.
Went, F. W. (1960) Organic matter in the atmosphere, and its possible relation to petroleum formation. Proc. Natl.
Acad. Sci. U. S. A. 46: 212-221.
Wilson, T. R. S. (1975) Salinity and major elements of sea water. In: Riley, J. P.; Skirrow, G., eds. Chemical
oceanography. 2nd ed. London, United Kingdom: Academic Press; pp. 365-413.
Wilson, W. E.; Schwartz, W. E.; Kinzer, G. W. (1972) Haze formation—its nature and origin. New York, NY:
Coordinating Research Council; report no. CRC-APRAC-CAPA-6-68-3. Available from: NTIS, Springfield,
VA;PB-212609.
Wolff, G. T.; Korsog, P. E. (1985) Estimates of the contributions of sources to inhalable particulate concentrations in
Detroit. Atmos. Environ. 19: 1399-1409.
Wolff, G. T.; Countess, R. J.; Grobhcki, P. J.; Ferman, M. A.; Cadle, S. H.; Muhlbaier, J. L. (1981) Visibility-reducing
species in the Denver "brown cloud"—II. sources and temporal patterns. Atmos. Environ. 15: 2485-2502.
Wolff, G. T.; Korsog, P. E.; Kelly, N A.; Ferman, M. A. (1985a) Relationships between fine particulate species,
gaseous pollutants and meteorological parameters in Detroit. Atmos. Environ. 19: 1341-1349.
Wolff, G. T.; Korsog, P. E.; Stroup, D. P.; Ruthkosky, M. S.; Morrissey, M. L. (1985b) The influence of local and
regional sources on the concentration of inhalable particulate matter in southeastern Michigan. Atmos. Environ.
19: 305-313.
5-80
-------
Wu, J. (1979) Sea spray in the atmospheric surface layer: review and analysis of laboratory and oceanic results. J.
Geophys. Res. [Atmos.] 84: 1683-1704.
Xie, Y.; Hopke, P. K.; Casuccio, G.; Henderson, B. (1994a) Use of multiple fractal dimensions to quantify airborne
particle shape. Aerosol Sci. Technol. 20: 161-168.
Xie, Y.; Hopke, P. K.; Casuccio, G.; Henderson, B. (1994b) Use of chain code histogram method to quantify airborne
particle shapes. Aerosol Sci. Technol. 21: 210-218.
Zeng, Y.; Hopke, P. K. (1989) Three-mode factor analysis: a new multivariate method for analyzing spatial and
temporal composition variation. In: Watson, J. G., ed. Receptor models in air resources management:
an international specialty conference; February 1988; San Francisco, CA. Pittsburgh, PA: Air & Waste
Management Association; pp. 173-189. (APCA transactions series: no. 14).
Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from a-pinene and
p-pmene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
5-81
-------
6. ENVIRONMENTAL CONCENTRATIONS
6.1 BACKGROUND, PURPOSE, AND SCOPE
This chapter summarizes the concentrations of particulate matter over the United States,
including the spatial, temporal, size and chemical aspects. The information needs for assessing
the major aerosol effects of concern are summarized in Table 6-1. The general approach
followed in preparing this chapter was to organize, evaluate, and summarize the existing large
scale aerosol data sets over the United States. Emphasis was placed on complete national
coverage as well as the fusion and reconciliation of multiple data sets.
Space is the main organizing dimension used to structure this chapter. Aerosol
concentration data are presented on global, continental, national, regional, and
sub-regional/urban scales. Within each spatial domain, the spatial-temporal structure, size, and
chemical composition are presented. An overview of the pattern analysis methods is given in the
remainder of Section 6.1. The presentation of aerosol patterns begins from the global and
continental perspective (Section 6.2). Next, nationwide aerosol patterns (Section 6.3) derived
from nonurban and urban PM10 and PM2 5 monitoring networks are examined. Section 6.3 also
includes a discussion of various measures of background PM25 and PM10. In Section 6.4 the
aerosol characteristics over seven subregions of the conterminous United States are examined in
more detail. The 10-year trends, seasonal patterns, relationships between PM2 5 and PM10, and
fine particle chemical composition are examined for each region. Section 6.5 focuses further on
the subregional and urban-scale aerosol pattern over representative areas of the United States.
Section 6.6 presents more detailed information on the chemical composition of the aerosol from
a number of intensive field studies. Section 6.7 deals with measurements of fine particle acidity.
Section 6.8 focuses on the concentration of ultrafine particles and Section 6.9 on the chemical
composition of ultrafine particles. Section 6.10 examines trends and relationships for PM25,
PM(10_2 5), and PM10 in data bases having long term data on both components.
Aerosol concentration data for the United States have been reported by many aerosol
researchers over the past decade. This chapter draws heavily on the contribution, of research
groups that have produced data, reports, and analyses of nonurban data. However, their maps,
charts, and computations have been re-done for consistency with urban data reports.
6-1
-------
6.1.1 Dimensionality and Structuring of the Aerosol Data Space
Aerosol concentration patterns contain endless detail and complexity in space, time, size,
and chemical composition. Aerosol samples from the conterminous United States reveal the
coexistence of sulfates, hydrogen ions, ammonium, organic carbon (OC), nitrates, elemental
carbon (EC [soot]), soil dust, sea salt, and trace metals. This chemically rich aerosol mixture
arises from the multiplicity of contributing aerosol sources, each having a unique chemical
mixture for the primary aerosol at the time of emission. The primary aerosol chemistry is
further enriched by the addition of species during atmospheric chemical processes. Finally, the
immensely effective mixing ability of the lower troposphere stirs these primary and secondary
particles into a mixed batch with varying degrees of homogeneity, depending on location and
time.
A major consideration in structuring the aerosol pattern analysis is that it has to be
consistent with the physical and chemical processes that determine the concentrations of the
aerosol. The concentration of paniculate matter (C), at any given location and time is
determined by the combined interaction of emissions (E), dilution (D), and chemical
transformation and removal processes (T),expressed as:
C=f(D,T,E)
Dilution, transformation/removal, and emissions are generic operators and can, in
principle, be determined from suitable measurements and models. However, for consideration
of aerosol pattern analysis it is sufficient to recognize and separate these three major causal
factors influencing the aerosol concentration pattern.
It is convenient to categorize the highly variable aerosol signal along the following major
dimensions: space, time, size and chemical composition. The dependence of concentration on
space and time is common to all pollutants. However, both the distribution with respect to
particle size as well as the chemical distribution within a given size range constitute unique
dimensions of particulate matter that are not present for other pollutants. The concentrations of
single-compound gaseous pollutants can be fully characterized by their spatial and temporal
pattern. This classification by dimensions is consistent with the size-chemical composition
distribution function introduced by Friedlander (1977). It could be said that particulate matter is
6-2
-------
a composite of hundreds of different substances exhibiting a high degree of spatial and temporal
variability.
6.1.2 Spatial Pattern and Scales
The spatial dimension covers the geographic scale and pattern of aerosols. Based on
consideration of emissions, meteorology, and political boundaries, the spatial dimension can be
broken into global, national, regional-synoptic, meso, urban, and local scales. Some of the
characteristics of these spatial scales are illustrated in Table 6-1.
TABLE 6-1. SPATIAL REGIONS AND SCALES
Global
Continent
10,000 -
50,000 km
National
Country
5,000 -
10,000 km
Regional
Multi-state
1,000-5,000
km
Meso
State
100-
1,000km
Urban
County
10- 100km
Local
City center
1 - 10 km
6.1.3 Temporal Pattern and Scales
The time dimension of aerosols extends over at least six different scales (Figure 6-1).
A significant, unique feature of the temporal domain is the existence of periodicities. The
secular time scale extends over several decades or centuries. Given climatic and chemical
stability of the atmosphere the main causes of secular concentration trends are changes in
anthropogenic emissions. Emissions, atmospheric dilution, as well as chemical/removal
processes, can be influenced by the seasonal cycle. The synoptic scale covers the duration of
-------
I—/I
Dilution |)( [chemistry/Remova = Concentratiot
Secular
Yearly
Weekly
Synoptic
Daily
Microscale
Minutes
Figure 6-1. Time scales for particle emissions.
synoptic meteorological events (3-5 days). Its role is primarily reflected in dilution and
chemical/removal processes. The daily cycle strongly influences the emissions, dilution, and
chemical/removal processes. Microscale defines variation of the order of an hour caused by
short-term atmospheric phenomena. In the analysis that follows we will emphasize secular
trends and yearly cycles, with some consideration of daily aerosol pattern. The microscale
patterns will be largely ignored.
6.1.4 Space-Time Relationships
The spatial and time scales of aerosol pattern are linked by the atmospheric residence time
of particles. Short residence times restrict the aerosol to a short transport distance from a source,
causing strong spatial and temporal gradients. Longer residence times yield more
6-4
-------
10
8 -
o>
S
of
u
c
n
4-1
CO
Q
•c
o
Q.
(0
Global
CO 2, CH4
Synoptic-
Regional
SOX,NOX,03
FineParticle
(<2 urn)
Mesoscale
NO, NO2, O3
Coarse
Particles
Microscale
Heavy Dust, Sand
(>20 urn)
10
10
102 103 104 105 106
Residence Time, seconds
10'
Figure 6-2. Relationship of spatial and temporal scales for coarse and fine particles.
uniform regional patterns caused by long range transport. The relationship between spatial and
temporal scales for coarse and fine particles is illustrated in Figure 6-2.
The aerosol residence time itself is determined by the competing rates of chemical
transformations and removal. Secondary aerosol formation tends to be associated with multi-
day long range transport because of the time delay necessary for the formation. For sulfates, for
example, the residence time is 3-5 days. For fine particles, 0.1 //m to several //m, the main
removal mechanism involves cloud processing, while coarse particles above 10 //m are deposited
by sedimentation. Ultrafine particles, below 0.1 //m, also rapidly coagulate to form particles in
the 0.1 to 1.0 //m size range. Another factor which must be considered is local turbulence. As a
consequence of low removal rates, aerosols in the 0.1-1.0 //m size range reside in the atmosphere
for longer periods than either smaller or larger particles (Figure 6-3).
6-5
-------
10
10
10'
10
8 T
6 . .
5 ..
I io4 J-
102 *
io1 J-
1
"10Z
/ Wet Removal
Coagulation Sedimentation
1°3 A
•• 10
•• 1
J-io-1
o>
1 .i
o>
o
c
o>
2
i/j
01
a:
10"4 10'3 10'2 10'1 10° 101 102 103
Radius, um >•
Figure 6-3. Residence time in the lower troposphere for atmospheric particles from 0.1 to
1.0 jim. ( — Background aerosol, 300 particles cm3; — continental aerosol,
15,000 particles cm3.)
Source: Jaenicke (1980).
If aerosols are lifted into the mid- or upper-troposphere their residence time will increase to
several weeks. Large scale aerosol injections into the stratosphere through volcanoes or deep
convection result in atmospheric residences of a month or two months for ash and > 2 years for
sulfates formed from SO2 oxidation.
In the context of the specific analysis that follows, the space-time-concentration
relationship in urban and mountainous areas is of particular importance (Figure 6-4). Urban
areas have strong spatial emission gradients and also may have corresponding concentration
gradients for directly emitted species, particularly in the winter under poor horizontal and
vertical transport conditions.
In mountainous regions, the strong concentration gradients are caused by both topography
that limits transport as well as the prevalence of emissions in valley floors. Strong
6-6
-------
Summer
Summer
rural
Winter
urban
rural
valley
mountain
rural
urban
rural
mountain
valley
mountain
Figure 6-4. Space-time relationship in urban and mountainous areas.
wintertime inversions tend to amplify the valley-mountain top concentration difference. Fog
formation also accelerates the formation of aerosols in valleys
6.1.5 Particle Size Distribution
The aerosol size distribution is of importance in quantifying both the formation
(generation) as well as the effects of aerosols. Condensation of gaseous substances during
combustion in the atmosphere generally produces fine particles below 1 //m in diameter.
Resuspension of soil dust and dispersion of sea spray produces coarse particles above 1 //m.
The size distribution of particles also influences both the atmospheric behavior and the
effects of aerosols. Atmospheric coagulation, cloud scavenging, and removal by impaction and
settling are strongly size dependent (Figure 6-3). The effects on human health depend on
size-dependent lung penetration. The effects of light scattering on visibility and climate are also
strongly dependent on particle size.
Measurements over the past decades (Whitby et al., 1972; Whitby, 1978) show that
atmospheric aerosols may be classified as fine mode particles or coarse mode particles. The size
distribution of atmospheric particles is discussed in Section 3.7. The sources, formation
mechanisms, and chemical compositions of these two aerosol modes are different. In general,
6-7
-------
the two aerosol size modes have independent spatial and temporal patterns as described
throughout this chapter. Coarse dust particles tend to be more variable in space and time and can
be suspended through natural or human activities. Fine particles during the warmer months of
the year are largely of secondary origin and their spatial-temporal pattern is more regional.
Notable exceptions are urban-industrial hotspots and mountain valleys where primary submicron
size smoke particles can prevail.
6.1.6 Aerosol Chemical Composition
The chemical composition of atmospheric aerosol is believed to influence the effect on
human health. While the causal mechanisms are not fully understood, the acidity,
carcinogenicity, and other forms of toxicity are chemical properties considered relevant to
human health.
The aerosol chemical composition has also become an important property for identifying
source types based on chemical "fingerprints" in the ambient aerosol. Since aerosols reside in
the atmosphere for days and weeks, there is a substantial amount of mixing that takes place
among the contributions of many sources. At any given "receptor" location and time, the aerosol
is a mixture of many source contributions each having a chemical signature for possible source
type identification.
Fine particles are generally composed of sulfates, hydrogen ions, ammonium, organics,
nitrates, elemental carbon (soot), as well as a portion of the trace metals (Section 6.6). Each
major chemical form has sub-species such as acidic and neutral sulfates, light and heavy
organics, ammonium and sodium nitrates, etc.
The chemical composition of coarse particles is dominated by the elements of the earth's
crust, Si, Al, Fe, and other elements commonly found in soil. Near industrial sources, coarse
particles may be contaminated by lead and other trace metals. At ocean shores, coarse particles
may consist of sea salt arising from breaking of waves. Both resuspended dust and sea salt are
primary particles, carrying the chemical signatures of their sources.
6-8
-------
6.2 GLOBAL AND CONTINENTAL SCALE AEROSOL PATTERN
There are two data sets which can be used to provide information on fine particle
concentration patterns on continental and global scales. Routine visibility distance observations,
recorded hourly at many U.S. airports by the U.S. Weather Service, provide an indication of fine
particle pollution over the United States. The visibility distance data have been converted to
aerosol extinction coefficients and used to access patterns and trends of aerosol pollution over
the United States (Husar et al., 1994; Husar and Wilson, 1993). Routine satellite monitoring of
backscattered solar radiation over the oceans by the Advanced Very High Resolution Radiometer
sensors on polar orbiting meteorological satellites provides a data set which can be used to give
an indication of aerosol over the world's oceans. These two data sets have been merged to
provide a global and continental perspective. The data analyses presented here were performed
for this Criteria Document and have not yet been published elsewhere.
Aerosol detection over the oceans is facilitated by the fact that the ocean reflectance at 0.6
|im is only 0.02. Hence, even small backscattering from aerosols produces a measurable aerosol
signal. The backscattering is converted to a vertically integrated equivalent aerosol optical
thickness assuming a shape for the aerosol size distribution or phase function. Clouds are
eliminated by a cloud mask, so the data are biased toward clear-sky conditions. The oceanic
aerosol maps represent a two-year average (July 1989-June 1991) prior to the eruption of Mt.
Pinatubo, when the stratosphere was unusually clear of aerosol. Consequently, the images
reflect mainly the spatial pattern of tropospheric aerosol.
A continental-scale perspective for North America is shown in Figure 6-5. Seasonal
depictions of the oceanic aerosol for the entire globe are shown in Figure 6-6. The average
aerosol map of Eastern North America for June, July and August (Figure 6-5) shows areas of
high optical depth over the Mid-Atlantic States and over the Atlantic Ocean. The aerosol
concentration over the oceans is highest near the coast and declines with distance from the coast.
This indicates that the aerosol is of continental origin and represents a plume originating in
eastern North America, heading north-east across the Atlantic ocean. This plume can also be
seen in the spring and summer season oceanic aerosol patterns shown in Figure 6-6.
6-9
-------
^r
March, April r May-
%r-r£
?
Q
/ \,.
fcm-l
_
—
—
—
—
D.n
D.14
D.ll
D.DB
0.05
X-
~\^--
^ v
continental - visibility
taa-1
!
—
—
—
—
D.2
D.n
D.14
D.ll
D.08
D-Q5
•July, August:f Sept.
C
continental - visibility
1 ' f^
oceanic - RVHRR satellite
Figure 6-5. Continental scale pattern of aerosols derived from visibility observations over
land and satellite monitoring over the oceans: North America.
6-10
-------
-/"*y > /•' *>
j-; ;•,. -Tf
t* 1000
240
210
180
150
120
90
60
30
*
r
^T^~"* -,,,_
' y- -• •• ,"•
• •- -I
u
-"-""-4,
r,
/
w' ^T^^Si plx,
" A ;A.. VU^fH ''•
\"" ' »J^" S)1l*wsl| '* "*«?'"* - \ \
;tij. "^^bt^'ij^- ^
t* 1000
SB 24°
i 180
™~t 150
-4 120
—j 90
•! so
•i »•
\
Figure 6-6. Global pattern of oceanic aerosols derived from satellite observations.
-------
The continental aerosol extinction coefficient data for the southwest coast of North
America indicate elevated aerosol extinction over southern California. The area includes the
hazy South Coast and the San Joaquin Valley air basins. It is interesting to note that somewhat
elevated aerosol optical thickness is also recorded over the Pacific near Southern California.
However, the low aerosol signal and the semi-quantitative satellite data preclude a clear cause
and effect association.
The seasonal aerosol pattern over the oceans reveals that the highest aerosol signal is found
near the tropics, where wind-blown dust and biomass burning in Africa and southern Asia
produce 5,000 km long aerosol plumes (Figure 6-6). Further aerosol belts which may be of
marine origin are observed just north of the Equator and at 30 to 60° latitudes in both
hemispheres. The backscattering in the summer hemispheres exceeds the winter values by a
factor of 5 to 10. There is a pronounced seasonality in each aerosol region (Figure 6-7); the
higher aerosol levels appear in the summer hemisphere although many continental and marine
regions show a spring maximum. Thus, the global tropospheric aerosol is a dynamic collection
of independent aerosol regions, each having unique sources and temporal patterns.
The seasonal oceanic aerosol maps show two distinctly different spatial patterns: aerosol
plumes originating from continents, and oceanic aerosol patches that are detached from the
continents. The continental aerosol plumes are characterized by high values near the coastal
areas and a decline with distance from the coast. The most prominent aerosol plume is seen over
the equatorial Atlantic, originating from West Africa and crossing the tropical Atlantic. It is the
well known Sahara dust plume. Additional continental plumes emanate form Southwest Africa,
Indonesia, China-Japan, Central America and eastern North America. Aerosols which may be
of marine origin dominate large zonal belts (30 to 60° N and S) in the summer hemispheres as
well as near the Equator. In summary, the global tropospheric aerosol is a collection of largely
independent aerosol regions, each having a bio-geochemically active source and unique spatial
temporal pattern.
Based on the above global and continental-scale observations, it can be concluded that the
continental plume from eastern North America is not as intense as those from other industrial
and non-industrial regions of the world. However, quantitative aerosol comparisons of global
regions are not available.
6-12
-------
400 -
310-
300 -
250
200
150 -
1 00 j
50 •
ft
y
T » i • P-~ «, -» WO
f ^^H| 21
11 <-1 140
/ \ 12°
,' \ 100
j/pl^*--N ^ Rn
- ,x*>*:i"~'* W Africa ^ 60
Ux" % 40
^1 20
n
2:OD£rCC?-Z:='OGLlrr>O I
NWPacic, fc
/" \
/ M \ N
f if' \A . x
/ •• ••»• \ 'w--'^ x
' // A\V"
1 / '"' *• ll ".
m/ •• *~^ -x \
^ NE Pacific ""Vl
m
z CD ct cc >• z; ^ CD CL >~ >• o
=' ~> UJ <— ' Q UJ «C UJ UJ '=•* O LU
™ ti_ 2U K£ *:> ^^ J <£ t^J C? ^y "^ u~ *:> ^ 2E SH*! ~ *% to * — •* ^ d
100 T
80 •
80 -
7CI !
60 j
iO -
40 •
30 -
20 •
10 -
-
2
120 -
100 -
80 •
eo-
40 •
1
20
i
ft
y
70 a
,xJ/N^ EC Pacific 60 " 1
-•xjrx "'x \ 50 ••
"*-\*\ 40 ••
''•n_ H.
—HI x
\\ 30 ••
\ J..-'^*""B
20 ""
10 »
n
New Zealand .
U
\ J
% S _|/:?
Ih » ^-x'«' /
\ \ W
% \ /^'t
% \ y/ 1
\», V1/ i
!i^x fj \
x**ifII^P-*;' SE Pacffi
4 4 4 4 1 4 1 } 1 1 (
1 t t 1 i t t < i t \ _.!_'_' - '
SO < CL < 5 => S Sb o § a < ft! < S 1 1 = 1 & o g a
Ix, *>^ §^ *v ***">* ^^ -=j" CO i-~' ~^f Ci " ^^ «£^=' - «^=' " ^** ***** "s~~ ^™"^
- e 12°
100
/ \ 80
P 'x->
/ * m'' en
/ /' \ \ iO
/ r' \ \
• ,« ./ x^-. 40
r" F "!ii %-,
>-' ,«** f%ft f»ft \ •*.
/ o ^ -m gO
Ji— * Shifted 6 months """«-,._
f ~n
N -f
IK^""^,
/ •
Jf m,,--®, \
C J&T """"m,
s^jr^Global "~«x ^
ll >%»-^'^H
' It JK-«/' ^^*
^iBit ^ M*e~"
Ws^^^.^W— «
S
T 1 T f T ¥ I T I I T 1
I 03 Ql CC >~ IS — 1 O CL H- "> O Z CD Cr CC >- 2 — ' i'r.< CL t— "> O
«C UJ «C Q... <( Z> — ? ff> LU f O UJ «C UJ «C CL -'C ~' =:' 3 UJ '=f O UJ
~' tj- 2 "C 5 "' """ < to O 2 o "^ u- 2 •*£ 2 ~' •' < w '-:' 2 o
Figure 6-7. Seasonal pattern of oceanic aerosols derived from satellite observations.
6-13
-------
6.3 U.S. NATIONAL AEROSOL PATTERN AND TRENDS
Our current understanding of the U.S. national aerosol pattern arises from nonurban,
regional background monitoring networks, the Interagency Monitoring of Protected Visual
Environments (IMPROVE) (Sisler et al., 1993) and the Northeast States for Coordinated Air
Use Management (NESCAUM) (Poirot et al., 1990, 1991), and from a mainly urban network,
the Aerometric Information Retrieval System (AIRS) (AIRS, 1995). The nonurban and urban
networks yield markedly different national patterns, particularly over the western United States.
For this reason the results from the two sets of observations are presented separately and the
differences between two networks are evaluated. The data analyses presented here were
performed for this Criteria Document and have not yet been published elsewhere.
6.3.1 Nonurban National Aerosol Pattern
Nonurban aerosol concentrations are measured at remote sites, away from urban-industrial
activities. Size-segregated aerosol mass and chemical composition data are available for 50
sites, through the IMPROVE (Joseph et al., 1987; Eldred et al., 1987, 1988, 1990; Eldred and
Cahill, 1994) and NESCAUM (Poirot et al., 1990, 1991; Flocchini et al., 1990) networks. These
are located mostly in national parks and wilderness areas. The PM10 and PM2 5 mass
concentrations are sampled and analyzed on separate filters. The sampling frequency is
generally twice a week (Wednesdays and Saturdays) for 24 hours. The PM2 5 samples are
analyzed for chemical composition which makes the data sets suitable for chemical mass balance
computations (e.g., Sisler et al., 1993; Malm et al., 1994b). The IMPROVE/NESCAUM aerosol
data are available from 1988 through 1993.
Measurements of PM are available from the IMPROVE/NESCAUM network at a smaller
number of sites compared to the number of sites for which measurements are available from the
AIRS network. The nonurban sites also have very different geographical distributions from
those sites in the urban network. Therefore, the ability to compare PM10 concentrations from the
nonurban and urban networks is severely limited by these factors.
The monthly distributions of chemical species, the chemical mass balances, obtained from
the measurements at nonurban sites are incomplete. Only sulfate, organics, soil, and soot
(elemental carbon) are considered. The contributions of hydrogen ion, water, trace metals and
6-14
-------
sea salt are not listed. The contribution of nitrate is included on a national basis, but not in the
subsequent discussion for regions.
The results of the national spatial and temporal pattern analysis are presented in quarterly
contour maps and monthly seasonal time charts. The contours drawn for the eastern United
States are derived from only 15 to 20 stations. As a consequence, these contour lines are to be
taken as guides to the eye and not as actual patterns. The quarters of the year are calendrical.
6.3.1.1 Nonurban PM2 5 Mass Concentrations
Maps of seasonal average nonurban PM2 5 concentrations are shown in Figure 6-8. The
maps divide the country roughly into eastern and western halves. The eastern United States is
covered by large, contiguous PM2 5 concentrations that range from 10 //g/m3 in Quarter 1, and
17 //g/m3 in Quarter 3. During the transition seasons (Quarters 2 and 4) the eastern U.S.
nonurban PM25 concentrations are at about 12 //g/m3. Within the eastern U.S., there are
subregions such as New England that have lower concentrations ranging between 8 and
12 //g/m3. During the third quarter, there is a wider range of geographic distribution of PM2 5
concentrations in the eastern United States than in other quarters of the year.
The lowest nonurban PM2 5 concentrations are measured over the central mountainous
western states. The low winter concentrations are at about 3 //g/m3, while the summer values are
around 6 //g/m3. Somewhat elevated PM2 5 concentrations are observed over the southwestern
border adjacent to Mexico as well as in California and the Pacific Northwest. The nonurban fine
particle mass clearly shows multiple aerosol regions over the conterminous U.S., each exhibiting
unique spatial and seasonal characteristics.
6.3.1.2 Nonurban Particulate Matter Coarse Mass Concentrations
In classifying size fractions of PM, PM10 refers to PM collected in a sampler with a
50% cutpoint of 10 |im aerodynamic diameter and PM25 to PM collected in a sampler with a
cutpoint of 2.5 jam aerodynamic diameter. PMCoarse or coarse will be used to refer to the PM
between the cutpoints of 2.5 and 10 jim, whether determined by subtracting a PM25 sample mass
from a PM10 sample mass or determined directly from the coarse particle channel of a
dichotomous sampler with a PM10 (or PM15) jim diameter upper cutpoint. Fine will also
6-15
-------
o\
Coars*
jjg/m'
If25
is 21
' 17
' 13
^.^ Q
*^^" 5
'i*ass
'•
f
• 25
21
17
13
9
5
C o ar s« »
Figure 6-8. Coarse mass concentration derived from nonurban IMPROVE/NESCAUM networks.
-------
be used to refer to PM2 5. PM2 5 is an indicator of the fine mode particle mass but it is not an
exact indicator, since PM2 5 may contain some coarse mode PM. Likewise, PMCoarse or coarse
refers to the inhalable fraction of the coarse mode, not the entire coarse mode. Under high
relative humidity conditions PMCoarse may contain some fine mode PM.
The nonurban coarse aerosol mass concentration in the size range 2.5 to 10 //m is given in
the seasonal maps in Figure 6-9. It is plotted on the same concentration scale as the nonurban
PM2 5 and PMCoarse maps to show that the nonurban coarse mass concentration is less than the
fine mass concentration over most of the country. The lowest nonurban coarse particle
concentration is recorded during the first, second, and fourth calendar quarters when virtually the
entire conterminous United States showed values <10 //g/m3. The industrialized Midwest,
adjacent to the Ohio River, shows low PMCoarse concentration (<10 //g/m3) comparable to the
relatively clean Rocky Mountains states. The highest nonurban coarse mass concentrations
appear during quarters 2 and 3. In quarter 2, the southwestern United States adjacent to the
Mexican border shows the highest nonurban coarse mass concentrations. In quarter 3, the
monitoring sites in Florida and Southern California exhibit high concentrations (>12 //g/m3).
6.3.1.3 Nonurban PM10 Mass Concentrations
Maps of seasonal average nonurban PM10 concentrations are shown in Figure 6-10. PM10
is the sum of the PM2 5 and PMCoarse. The spatial pattern from east to west, including the
delineation of aerosol regions, is generally similar to the PM25. However, the PM10
concentrations exceed the PM2 5 by up to a factor of two depending on region and season. The
sparseness of nonurban sites over large areas of the central United States limits the reliability of
profiles in these areas.
In the eastern U.S., PM10 concentrations range between 12 //g/m3 in Quarter 1 and
25 //g/m3 in Quarter 3. During the transition seasons (Quarters 2 and 4) the eastern U.S.
non-urban PM10 concentrations are about 15 //g/m3, except in New England. The lowest PM10
concentrations are measured over the central mountainous states, 5 //g/m3 in Quarter 1,10 //g/m3
in Quarter 3, and 7 //g/m3 during the transition seasons. Higher PM10 concentrations, between
10 and 20 //g/m3, were measured over the southwestern United States
6-17
-------
€«?««
ug/m3
»25
21
17
13
oo
•• 4-
- '" V-
Jf
•25
21
17
13
9
5
CO if Sf 0-«:"K>
Figure 6-9. Coarse mass concentration derived from nonurban IMPROVE/NESCAUM networks.
-------
M9/m3
25
— 21
17
13
9
5
VO
¥
Figure 6-10. PM10 mass concentration derived from nonurban IMPROVE/NESCAUM networks.
-------
as well as over the Pacific states from California to the Northwest than over the central
mountainous states.
6.3.1.4 PM2 5/PM10 Ratio at Nonurban Sites
The PM10 aerosol mass is composed of fine mass (PM25) and coarse mass, below 10//m
(Figure 6-10). Both the sources and the effects of fine particles differ markedly from those of
coarse particles. For this reason it is beneficial to examine the relative contribution of PM2 5 and
PM10 concentrations. Figure 6-11 shows the seasonal fine mass as a fraction of PM10.
Nationally, the fine fraction at nonurban sites ranges between 0.4 and 0.8. The highest fine
fraction is recorded east of the Mississippi River, where 60 to 70% of the PM10 mass is in
particles <2.5 //m in size. This is also the region that shows the highest PM10 concentrations;
thus, fine particles dominate the nonurban aerosol concentrations east of the Mississippi River.
The fine fraction exceeds the coarse fraction at the nonurban northwestern sites. The fine
fraction is the lowest in the southwestern United States (< 50%), particularly in the spring season
(Quarter 2).
Spatial and seasonal variation of the fine fraction is a further indication for the existence of
different aerosol regions over the conterminous U.S. This is further illuminated in Section 6.4
where the aerosol characteristics over different regions of the United States are discussed.
6.3.1.5 Nonurban Fine Particle Chemistry
The elemental composition of nonurban fine particles over the conterminous United States
is now reasonably well understood. The IMPROVE/NESCAUM network provides over
five years of aerosol mass and chemical composition data. The data from these networks allows
the chemical apportionment of the fine particle mass into aerosol types such as sulfates, organic
carbon, elemental carbon, and fine soil (Schichtel and Husar, 1991; Sisler et al., 1993, Sisler and
Malm, 1994). The quantification of these aerosol types is relevant to both the determination of
aerosol effects and source apportionment of particle mass. It should be emphasized that urban
areas, mountain valleys, and remote monitoring sites are likely to have different relative
concentrations of the aerosol types. Also, the quantification of semivolatile organic compounds,
nitrates, and other unstable species is subject to major uncertainties.
6-20
-------
!..«•*;:
oi Wi
*-nt
cd p-*i!1»a ¥
Figure 6-11. Fine fraction of PM10 derived from nonurban IMPROVE/NESCAUM networks.
-------
Much work remains in order to define the chemical, as opposed to the elemental, composition,
especially for organic compounds.
At nonurban eastern U.S. sites, a large fraction of the fine aerosols are composed of sulfate
and related species (ammonium ions, hydrogen ions, and associated water) and organic
compounds. In the northeastern and southeastern U.S., organic carbon appear to equal sulfate in
the fourth quarter of the year. In the southwestern U.S., wind blown dust is a major component
of fine mass while sulfate is less important (Schichtel and Husar, 1991).
Annually averaged fine particle sulfate, as ammonium sulfate; organic carbon; elemental
carbon; and nitrate, as ammonium nitrate, concentrations from the IMPROVE network across
the U.S. are shown in Figures 6-12 and 6-13 (Sisler et al., 1993; Malm et al., 1994b). The
station density is limited, especially in the eastern U.S. The contour lines in the annual average
maps are to be used as guides to the eye, rather than precise values. Concentrations of sulfate in
the eastern U.S. (Figure 6-12a) exceeds those over the mountainous western states by factor of
five or more. Elevated sulfate in excess of 1 //g/m3 is also reported over the Pacific coast states.
Sulfates typically contribute over 50% of the fine particulate mass in the eastern U.S., while
sulfates contribute 30% or below in the West.
Fine particle nitrates (Figure 6-12b) are highest in California, exceeding 4 |ig/m3 at most
sites. Their share of the fine mass at several California sites exceeds 20%. Organic carbon
concentrations (Figure 6-13a) are high over California and northwestern sites, as well as at the
eastern U.S. sites. Organic carbon contributes over 50% of the fine particle mass in the
Northwest, and about 30% throughout the eastern U.S. There is a high degree of uncertainty
associated with the measurement of particulate nitrate and organic carbon because of artifacts
arising from the adsorption of vapors or the loss of semivolatile materials. The elemental carbon
concentrations (Figure 6-13b) are significant over the Northwest and southern California, as well
as at the Washington, DC, site. Over most of the country elemental carbon is 5% or less of the
fine particle mass.
The chemical composition of PM10 and PM25 aerosols in the IMPROVE network (Eldred
et al., 1994b) revealed that the average coarse mass does not differ significantly between the
East and West; however, the fine mass is higher in the East. Also about 80% of
6-22
-------
to
1J
ijyi.'K.jBBj-Mi*™*" / - j 5 * t € 'Hjsjj,u%i; *=-m * i, „_ if * "*JJf1r .. ' "^nJ L, ^* ^ I \ j=Tirf,*"«*« . *•'•«.
JP™ ""»• asssssssssssssfc^1 ^ Iff * "iJ^lL „""" *•* jUU ffsSI ,s«/-. ** :^f= jfej ji ?*•=*&. "V. vWe*.««, I £****** «ij"*
i -sigi.iS*^ i Vs-^gT^-fa^^w f -s * -, j jg /•$ > |,g ,.* W**"'tI*- «w^s^t ^^^if,,__i t **••<% \ . t'-^yii : "*%'%* "UI **\ f
«yi ,.„, ^ r% ; I • i ± '• f ' "'* w <* i f * T1 j ,^.1 f^y^'1' 2?" )4* *, J% tf8^ • --*=-.-j ft s < 1 \ ' • "*a*!=i ====• *.„«, r^iiiis!f :iiiiiiii ; JjT %^ ^, ;^~; a*
kx^w^'l^T^iH -./A - /^ /'i/'L ^^if*17 *"r,*1'! ~I '-\ '*"••- *"--•"' .'^'"••sJifirSSv t I ^*itj"i^>. !' : V i "v.^
r\«r-»— \r^> ^ * > -r:
** -fesurSTr WV—,'/ ,' I™;;, ;( ^t'l^"'- ' -CM* ' \ l '—t j^? "'"i' *-**'
iK.|J
If,
'*?>l M9/m
C-.ni
^^KffljiaP'g'"
r-.
*«'*-,
NO,
*» y ~T
..... _
Wlf ^"*"*-
V/^'*1":-
«**c. ^?* :
" ' ""
_H
s^i
_^
*!
•
"%_ -J 1
> I
-TV
f-^ ' " ~~~*i !
1 I I
ff* \ I
T3 •» ? :
_ »•*"* *
\ j
H ^x.^Sr*-^
so;
Figure 6-12. Yearly average absolute and relative concentrations for sulfate and nitrate.
Source: Sisler et al. (1993) and Malm et al. (1994b).
% NO ;
-------
j -i \f\ zd-V*
%5L''' ^ '"''' '^If"t"iS->
*l«iMs>\ i ti£S|i,i,» f~"~- "•- ~-. :-. ,. /*,__>" *;*<>:?
'Mt '-IV'-aji: .in •«»/ i .'»:-" ••: V
. —- vq^n^;--j%»
^ .L'*^ J1 -^ f f.*J = '™'^cL ^ it^°*f
f"'
,•*;>
«a-~«s-
6*1
: »">« ^/'""T-iiri,.-..-' '
»,/,,- - «pa
.SlT* 'i^f |. I ||^
ittJ*
•- M*-J
Organic
Carbon, M9/ni
-.,
'
% Organic
Carbon
/ *•& fcf^*"*' f**tp8s^-^^""i*;'HT'"»."j^|^ '•."
"S*!"'i * ^**^! i-"™^i"~¥*B««*»*a,,fe
fU*| *fe||_ j f™^ ^_ , ^^. I "^"'^"' • 'i 1 1 '
*.*L i1', '•* X^y"SlIMP- Jfr-mim-iiMimTf-* " 3
'. fj-<.« '*- ^i Tt. ~jiji3K'fciJ||''fflr*v 3 i '. ''-f^ * „, „«*-!.«**" ""* "^*
^•JSJSTS I "* *»'"^3l'p|i ' " 4. l=i=, pas. - ^ |**3rt,«.«* ~ **"V*s.. -='
•*^|.! f '-a4**'1 i : » «j*-—i—K. r
''"tti ' --4", f' 1 -.-..V*
' I^!L4u—^^-^ ! C""/ 'uwJ^-'i,
'lft.1-.
i i r **fe
'-s /'^
»»"*•
ftr*-
; ;, i Elemental
• »i" i 3
_:-*_„„„„„;, Carbon, \iglm
.-""** i
% Elemental
Carbon
*;»"-.
Figure 6-13. Yearly average absolute and relative concentrations for organic carbon and elemental carbon.
Source: Sisler et al. (1993) and Malm et al. (1994b).
-------
soil elements and 20% of sulfur were found in the coarse fraction. Most trace elements were
found in the fine fraction, both in the East and in the West. The spatial and seasonal patterns in
particle concentrations and their relationships to optical extinction in the United States from the
IMPROVE network were summarized by Malm et al. (1994b).
In studying the regional patterns of nonurban trace metals in the IMPROVE network,
Eldred et al. (1994a) found a good correlation between selenium and sulfur at all sites in the
East. The correlation in the West is lower. Comparison of the S/Se ratios for summer and
winter shows that there is approximately twice the sulfur relative to selenium in summer
compared to winter. Se is a tracer for S emitted from coal-fired fossil fuel power plants; this
shift in S/Se from summer to winter is consistent with a substantial secondary photochemical
contribution to SO4 during the summer. Zinc is highest at the sites in the central East. It does
not correlate well with sulfur. Lead and bromine are relatively uniform, with slightly higher
mean concentrations in the East. There is poor correlation between lead and bromine. Copper
and arsenic are highest in the Arizona copper smelter region. Copper is also higher in the central
East.
Trends (1982 to 1992) of nonurban fine particle sulfur, zinc, lead, and soil elements were
reported by Eldred et al. (1994a) using the IMPROVE network data. They observe that in the
southwest, sulfur trends in spring, summer, and fall decreased, while most of the winter trends
increased. The trends in the Northwest increase slightly. The two eastern sites (Shenandoah and
Great Smoky Mountains) have increased almost 4% per year in summer, increased 1 to 3% in
spring and fall, and decreased 2% in winter. The annual increase was between 2 and 3%.
Generally, there were no significant trends in zinc and the soil elements. Lead at all sites
decreased sharply through 1986, corresponding to the shift to unleaded gasoline. The ten year
trends reported by Eldred et al. (1994b) have not been compared and reconciled with other
compatible data.
6.3.1.6 Seasonally of the Nonurban Chemistry
This section discusses the seasonality of size segregated chemical composition at
non-urban monitoring sites (IMPROVE/NESCAUM) over the entire U.S. (Figure 6-14).
6-25
-------
PM 2.5 Concentration - U.S.
IMPROVE/NESCAUM Data
Chemical Fine Mass Balance - U.S.
IMPROVE/NESCAUM Data
w
a
u_ 0.4
(C)
1989 Mar May Jul Sep Nov
Sulfate + OC + Soil +• EC
PM10, PM2.5 and PMC-U.S.
IMPROVE/NESCAUM Data
o
u
15,000
5,000
(b)
1989 Mar May Jul Sep Nov
^PM10 ^PM2.5 ^ PM Coarse
Chemical Tracers - U.S.
IMPROVE/NESCAUM Data
o
O
4,000
3,500
3,000
2,500
2,000
1,500
1,000
(d)
1989 Mar May Jul Sep Nov
Sulfur - Max = 4000 Selenium - Max = 4
Vanadium - Max = 10 S/Se - Max = 4000
Figure 6-14. Seasonal pattern of nonurban aerosol concentrations for the entire
United States: (a) monitoring locations; (b) PM10, PM2^ and PMCoarse
(PMC); (c) sulfate, soil, organic carbon (OC), and elemental carbon (EC)
fractions; and (d) tracers.
6-26
-------
The nationally aggregated average PM10, PM2 5 and PMCoarse is shown in Figure 6-14b
(See Section 6.3.1.2 for a definition of PMCoarse.). The nonurban PM10 concentration ranges
from 8 //g/m3 in the winter, December through February, to about 15 //g/m3 in June to August.
On the national scale the PM10 seasonality is clearly sinusoidal with a summer peak. Fine
particles over the nonurban conterminous United States account for about 50 to 60% of the PM10
mass concentration throughout the year. The coarse mass accounts for 40 to 50% throughout the
year. Hence, the fine-coarse aerosol ratio does not change dramatically for the average nonurban
aerosol.
The relative chemical composition of the aggregated nonurban aerosol pattern is shown in
Figure 6-14c, including sulfates, organic carbon, soil, and elemental carbon as a fraction of the
fine particle mass concentration. The Figure also shows the sum of these four aerosol species to
indicate the fraction of the fine aerosol mass that is not accounted for. Most notable among the
missing species is the contribution of nitrates, ammonium ion, and hydrogen ion.
There is mild seasonality in the nationally aggregated sulfate and organic carbon fractions.
Throughout the year, sulfate aerosol, including the ammonium cation, accounts for 30 to 40% of
the fine mass. Organic carbon also contribute 30 to 40% of the nationally averaged fine particle
mass. Thus, sulfates and organic carbon are the two dominant species, contributing about 70%
of the fine aerosol mass.
The contribution of soil dust to the fine mass ranges between 4% in the winter months to
12% during April through July. Elemental carbon is about 2% during the summer and 5%
during the winter.
The sum of the four measured fine mass components, sulfates, soil, organic carbon, and
elemental carbon add up to about 80% of the measured fine mass throughout the year. The
remaining, unaccounted fine mass may be contributed by nitrates, trace metals (e.g., Pb, Br, sea
salt [NaCl], etc.).
The seasonal pattern of concentration of primary emission tracers, selenium, Se and
vanadium, V is shown Figure 6-14d. Se is a known tracer for coal combustion, while V is a
trace constituent of fuel oil (Altshuller, 1980; Kleinman et al., 1980; Cass and McRae, 1983;
Tuncel et al. 1985). The Figure also shows the monthly average concentration of fine particle
sulfur as well as the S/Se ratio.
6-27
-------
The national average Se concentration is rather uniform over the seasons, ranging between
400 to 600 pg/m3. Since Se is a primary pollutant, the seasonal invariance means that the
combined effect of emissions and dilution is seasonally invariant over the year.
The concentration of V is between 500 to 700 pg/m3, with the higher concentrations
occurring in the winter season. Evidently, the contributions from V-bearing fuel oil are more
pronounced during the cold season. The monthly average sulfur in aerosols exhibits the highest
concentrations 1.5 //g/m3, during June, July, and August, and the lowest values 0.9 //g/m3,
during November, December, and January.
The S/Se mass ratio is about 700 during November to January and climbs to about
1,500 during April through September. The higher S/Se ratio during the warm season is an
indication of secondary sulfate production from SO2 in the plumes of coal fired power plants
(Chapter 3).
Eastern United States
The seasonal pattern of the eastern U.S. aerosol chemistry is shown in Figure 6-15. The
concentration of PM10, PM25, PMCoarse (Figure 6-15b) indicates a similar seasonality, highest
concentrations in the summer, and lowest in the winter. The PM10 levels range between 12 to
24 //g/m3, the PM2 5 ranges between 8 to 12 //g/m3, while PMCoarse ranges between 4 to 7
Mg/m3 over the year. The size segregated aerosol data for the nonurban East show that the fine
mass concentration (8 to 12 //g/m3) is higher than the national average (4 to 8 //g/m3), while the
coarse mass concentration is comparable to the national average. Eastern U.S. nonurban fine
particles contribute 60 to 70% of the fine mass throughout the year.
The apportionment of the fine particle mass into its chemical components (Figure 6-15c)
favors sulfates which amount to 40 to 50% of the fine mass throughout the year, compared to
about 30% of organic carbon. The contribution of soil dust is about 5% throughout the year,
while soot is more important in the winter (6%) than in the summer (3%). The above three
aerosol chemical components account for 85 to 90% of the measured fine particle mass, leaving
only relatively small contribution to nitrates, hydrogen ions, trace metals, and sea salt.
The coal tracer selenium (Figure 6-15d) exhibits a modest winter peaked seasonality
between 600 to 800 ng/m3. The fuel oil tracer vanadium on the other hand, is factor of two
6-28
-------
PM2.5 Concentration - Eastern U.S.
IMPROVE/NESCAUM Data
PM10, PM2.5 and PMC - Eastern U.S.
IMPROVE/NESCAUM Data
o
o
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
(b)
1989 Mar May Jul Sep Nov
+PM2.5 ^PM Coarse
Chemical Fine Mass Balance - Eastern U.S.
IMPROVE/NESCAUM Data
«
E
0)
U. D.4
O
1
o
U
1989 Mar May Jul Sep
Nov
Soil
Sulfate + OC + Soil + EC
Chemical Tracers - Eastern U.S.
IMPROVE/NESCAUM Data
4,000
3,500
3,000
2,500
2,000
1,500
1,000
(d)
1989 Mar May Jul Sep Nov
Sulfur-Max = 4000 -a-Selenium - Max = t
Vanadium -Max = 10 S/Se - Max = 4000
Figure 6-15. Seasonal pattern of nonurban aerosol concentrations for the eastern
United States: (a) monitoring locations; (b) PM10, PM2 5, and PMCoarse
(PMC); (c) sulfate, soil, organic carbon (OC), and elemental carbon (EC)
fractions; and (d) tracers.
6-29
-------
higher in the winter (1,500 ng/m3) compared to the summer (750 ng/m3). Evidently, the primary
contribution from fuel oil is winter peaked. The S/Se ratio is about 1,000 in the winter, and it is
over 2,000 in the summer months. This suggests the seasonality of secondary sulfate formation
during the summer months.
Western United States
The aggregated western U.S. aerosol seasonality is presented in Figure 6-16. The
non-urban aerosol concentrations for PM10, PM2 5, and PMCoarse are well below the
concentrations over the eastern United States (Figure 6-16b). The western United States differs
from the eastern United States, having lower fine mass concentrations, which range between 3 to
5 //g/m3. The coarse mass concentration (4 to 8 //g/m3) and seasonality is similar over the East
and the West. It is worth emphasizing, however, that these measurements are at remote national
parks and wilderness areas in both East and West. The examination of monitoring data in urban
areas and confined airsheds (Sections 6.4 and 6.5) reveals a highly textured pattern in space and
time.
The fine particle chemical mass balance (Figure 6-16c) for the aggregated western United
States shows the substantial contribution of organic carbon, which account for 30 to 45% of the
fine mass. The higher organic carbon fraction occurs in the November through January season.
Sulfates range between 20 to 25% throughout the year. Soil dust plays a prominent role in the
western fine mass balance, contributing 20% in April through May, but declining to 5% by
January. Elemental carbon ranges between 5% in the winter and 2 % during the summer. About
25% of the fine mass over the western United States is not accounted for by sulfates, soil,
organic carbon, and elemental carbon. It is known that nitrates are major contributors to the fine
particle mass in the South Coast Basin, as well as other western regions (White and Macias,
1987a; Chowetal., 1992a, 1993a, 1995a).
The concentration of the trace substances (Figure 6-16d) selenium and vanadium shows
both low concentrations and weak seasonality. The sulfur concentrations are also less than half
of the eastern U.S. values. The S/Se ratio is about 500 in the winter months and 1,000 during
the summer. The lower S/Se ratios compared to those in the eastern U.S. are the result
6-30
-------
PM2.5 Concentration -Western U.S.
IMPROVE/NESCAUM Data
O
'a
o
O
Chemical Fine Mass Balance -Western U.S.
IMPROVE/NESCAUM Data
3 0.7
U_ 0.4
(c)
o
o
.
1989 Mar
^Sulfate
May
Jul
Sep
Nov
Soil
Sulfate + OC + Soil + EC
PM10, PM2.5 and PMC -Western U.S
IMPROVE/NESCAUM Data
40,0001 • • • • • • • • • • •-
35,000
30,000
25,000
15,000
10,000
5,000
(b)
1989 Mar May Jul Sep Nov
-H- PM10 -+- PM2.5 -A- PM Coarse
Chemical Tracers - Western U.S.
IMPROVE/NESCAUM Data
4,000
3,500
3,000
2,500
2,000
1,500
1,000
(d)
1989 Mar May Jul
^^ Sulfur - Max = 4000
-+- Vanadium - Max = 10
Sep Nov
Selenium - Max = 4
S/Se - Max = 4000
Figure 6-16. Seasonal pattern of nonurban aerosol concentrations for the western
United States: (a) monitoring locations; (b) PM10, PM2 5, and PMCoarse
(PMC); (c) sulfate, soil, organic carbon (OC), and elemental carbon (EC)
fractions; and (d) tracers.
6-31
-------
of selenium emitting coal-fired power plants not being the only sources of western U.S. sulfur.
Smelters make a contribution to S but not Se in the atmosphere.
The above general discussion of national pattern of chemical and size dependence do not
provide the more detailed spatial and temporal texture of the U.S. aerosol pattern discussed in
the following sections. However, it provides the national scale gross features and serves as a
broader context for the more detailed examinations.
6.3.1.7 Background Concentrations of Particle Mass and Chemical Composition
The concentration and chemical composition of background particulate matter can very
with geographic location, from monitoring site to monitoring site; with season of the year; and
with meteorological conditions which affect the emissions and secondary production of biogenic
or geogenic species to the background.
A number of types of background can be considered. These backgrounds include the
following: (1) a "natural" background excludes all anthropogenic contributions. This
background includes any natural sources contributing to the background for chemical species in
North America or globally; (2) a background which excludes all anthropogenic sources within
North America, but not from anthropogenic sources contributing to background from outside of
North America; (3) a background which excludes the anthropogenic sources inside the United
States, but not from elsewhere in North America; (4) a background which excludes
anthropogenic sources from other regions into a specified region in the United States; (5) a
background which would exclude all sources of parti culate matter except those associated with a
particular urban area. The two backgrounds directly relevant to the Criteria Document are
backgrounds (1) and (2). The problems and limitations in obtaining reasonably accurate annual
average and seasonal values for these backgrounds are discussed below. Backgrounds (4) and
(5) can be more readily be obtained by measurements. These backgrounds are relevant to
subsequent stages in the implementation process. The averaging period over which background
levels are defined should also be stated. Annual and seasonal averages may be more appropriate
for risk assessments but daily peak values may be more relevant for control strategy
implementation.
More specifically, the term non-manmade is meant to encompass sources such as geogenic
dust plumes and sea salt as well as biogenic sources. Biogenic sources include (a) combustion
6-32
-------
products of biomass burning caused by lightning; (b) emissions of volatile sulfur compounds
from marshes, swamps or oceans; (c) organic particulate matter formed by the atmospheric
reactions of biogenic volatile organic compounds such as terpenes; and (d) particulate nitrates
formed by the atmospheric reactions of NOX emitted from soils. There is an intermediate class
of sources associated with agricultural activities. These include biomass burning caused by
human intervention and the addition of fertilizers to soils resulting in emissions of NH3 and NOX
(Section 5.2, 5.3).
Anthropogenic sources include vehicular and stationary sources which emit particles
directly or precursors such as sulfur dioxide, nitrogen oxides, or those volatile organic
compounds capable of reacting in the atmosphere to form organic particles. Stationary sources
of primary particulate matter as well as sulfur oxides and nitrogen oxide precursors include fossil
fuel power plants, while smelters are sources of primary particles and sulfur oxides. Vehicles
emit primary particulate matter as well as nitrogen oxides and volatile organic compounds.
Solvent usage, agricultural coatings, and many other industrial operations also may emit
precursors or particulate matter. Wood burning for heating of homes is a source of organic
carbon and elemental carbon (Section 5.2, 5.3).
The formation of sulfates from sulfur dioxide emitted by power plant plumes can occur
over distances exceeding 300 km and 12 h of transport (Section 3.4.2.1). Nitric acid also can be
formed in these plumes and it can be converted to ammonium nitrate, if sufficient ammonia is
available to first neutralize the sulfate in plumes. Similar transport can occur in urban plumes.
The transport distances in plumes depend on both formation rates of particles and their removal
by deposition processes. However, the residence times of fine particles can be long. For
example, if the dominant removal process is dry deposition, fine particles transported through a
1000 m deep mixed layer near the surface with deposition velocities of 1 to 0.1 cm/s have
atmospheric residence times ranging from 1 to 11 days (Section 3.5.1, 3.5.3). When particles
are trapped in a layer well aloft they may survive even longer periods. Therefore, transport
distances of several hundred to several thousand kilometers are possible.
Direct evidence of such transport aloft is available from satellite monitoring of back
scattered solar radiation. The most prominent plume is that of Sahara dust from West Africa
(Section 6.2). This plume has been observed to extend during the spring and summer months to
the east coast of the United States, especially over Florida (Figure 6-6). Ground level
6-33
-------
measurements in Bermuda indicate that southeasterly winds bring high concentrations of soil-
and crustal-related aerosols which appear to be from the Sahara desert (Wolff et al., 1986).
Other large plumes exist, such as the Asian plume. However, the satellite observations do not
indicate that it reaches the west coast of the United States (Figure 6-6).
Field measurements and modeling studies can be used as aids in the derivation of
background values for aerosol constituents. Either approach is subject to considerable
uncertainty and each has its own advantages and limitations. Field data would be the most
logical choice if it could be shown to be completely free of anthropogenic influences originating
within North America, i.e., background (2), (following the guidelines set out above for defining
background levels unaffected by pollution sources within North America). A number of
difficulties arise in interpreting field data for this purpose, namely: (a) there are very few tracers
(e.g., 14C) which can be used to distinguish between anthropogenic and biogenic source
categories of aerosol constituents; (b) multilayer trajectories should be used to identify source
regions since layer-average trajectories may underestimate the geographic area contributing
pollutants to the air mass sampled; (c) sampling must also be carried out for long enough periods
to obtain statistically representative values over seasonal time scales. Determining the history of
air parcels is difficult in locations subject to small scale circulations such as cumulus convection
and land-sea or mountain-valley breezes. In addition, all small localized anthropogenic sources
of particulate matter must be identified during sampling. Ideally, measurements should be
carried out long enough for the measurements to be shown to be generally representative of the
time period of interest e.g., seasonal average, annual average.
Alternatively, models which include only natural sources and anthropogenic sources
located outside North America could be used. Their utility is limited by inadequacies in model
formulation, such as grid spacing and knowledge of the strengths, locations, and variability of
various sources. Since a large fraction of parti culate matter is secondary, uncertainties in the
chemistry of precursor gases will play a large role in determining the uncertainty of the final
results. These uncertainties are especially large for the yield of aerosol produced by the
oxidation of biogenic hydrocarbons as pointed out in Chapters 3 and 5. Uncertainties in the
chemistry of NOX and SO2 are also important in that they affect estimates of the yield of aerosol
products versus the deposition of intermediate species.
6-34
-------
Trijonis (1982, 1991) has attempted to estimate PM25 and PM10 concentrations
corresponding to background (1), the "natural" background. His approach was to obtain
concentration values only from those biogenic and geogenic sources which are at or below those
possibly associated with preindustrial conditions over North America. Annual average
concentrations of the chemical species in particulate matter were estimated for the eastern United
States and for mountain/desert regions of the western United States. Seasonal "natural"
background concentration values were not estimated. The annual average concentrations of fine
particles were estimated separately for sulfates; as NH4HSO4, nitrates; as NH4NO3; organic
carbon; elemental carbon; soil dust and water (Trijonis, 1982, 1991). In the later work, coarse
particle concentration values were also estimated (Trijonis, 1991). In addition, in the later work,
it was emphasized that the concentration values proposed can have error factors ranging from
1.5 to 3 for individual chemical species in parti culate matter.
In the earlier work (Trijonis, 1982), a fine particle "natural" background for the eastern
United States is estimated at 5.5 ± 2.5 //g/m3. Excluding water, the background value would be
4 ± 2 //g/m3 with the largest contribution, 2 //g/m3, from organic carbon. In the later estimates
(Trijonis, 1991), a fine particle "natural" background for the eastern United States of 3.3 //g/m3
is estimated. Excluding water, this background would be 2.3 //g/m3 with 1.5 //g/m3 associated
with organic carbon. A separate estimate is given for the fine particle "natural" background over
the mountain/desert regions of the western United States of 1.2 //g/m3. Excluding water, this
background would be 1 //g/m3 with 0.5 //g/m3 associated with organic carbon. The coarse
particle "natural" background for both the eastern and western United States is estimated at
3 Mg/m3.
Fernam et al. (1981) also estimated "natural" background concentrations for PM25
constituents in the eastern United States during summer. They estimated natural contributions to
sulfate of 0.5-1.9 //g/m3, to organic carbon of 3.7 //g/m3, and to crustal material of 1.7 //g/m3.
To obtain these "natural" background estimates, a wide range of approaches are used
varying from natural SO2 and NOX emissions inventories to SO4, NO3 and elemental carbon
concentration measurements in remote locations in the northern and southern hemispheres.
Carbon isotope ratios and organic composition measurements for organic components are used
from several sites in the southwestern United States.
6-35
-------
Results of three-dimensional models that could be used to estimate each of the five
background levels for all the major categories of aerosol composition listed above are not
available. Liousse et al. (1996) have performed three-dimensional chemical tracer model
simulations of the global distribution of elemental and organic carbon. Background values
assuming only natural sources (background 1) were also calculated. Average organic carbon
concentrations calculated for the month of July were all less than 1 |ig/m3 in the United States.
These calculations were made assuming a 5% yield of secondary organic carbon from the
oxidation of terpenes (cf. Section 5-3).
Another approach is to use results from rural/remote sites in national parks, wilderness
areas and national monuments from the IMPROVE monitoring measurements. Results for the
period between March 1988 and February 1991 have been published (Malm et al., 1994). The
tabulations of results are given on an annual average basis for individual IMPROVE sites and on
a seasonal basis by IMPROVE subregion for fine mass; sulfate, as (NH4)2SO4; nitrate, as
NH4NO3; organic and elemental carbon; fine soil and coarse mass. These measurements do not
differentiate between anthropogenic and non-anthropogenic contributions and do not stratify
measurements values by wind direction or by use of trajectories representing various air masses
(Malm et al., 1994). However, a large set of measurements, including seasonal measurements,
are provided at a substantial number of rural/remote sites, especially in the western United
States.
In stratifying the IMPROVE results a problem arises because the Colorado plateau
"subregion" with seven sites straddles the boundary between the southwest and northwest used
subsequently (Figure 6-28). Four of the sites are north of the boundary in Utah and Colorado
and three of the sites are south of the boundary in Arizona and New Mexico. The authors place
the Colorado plateau in the southwest for purposes of a fine mass composition budget (Malm et
al., 1994). Since they assign only one other subregion, Sonora desert, with two sites to the
southwest, the method of assigning sites can significantly affect the resulting estimates of
regional fine mass concentrations. This problem can be avoided for the annual average values
which are shown by individual sites, but not for the seasonal values which are lumped by
subregion. This lumping also requires deciding whether a subregion with five sites, central
Rocky Mountains, should be given the same or five times the weight of the other subregions in
the northwest with only one or two sites each. For the annual average values given in Table 6-2
6-36
-------
the sites are assigned consistent with the division between the northwestern and southwestern
regions shown in Figure 6-28, excluding three sites in the northern California coastal mountains
considered separately. A transitional region between the western mountains and deserts and the
eastern United States has been considered consisting of five sites in three subregions from West
Texas (2), to South Dakota (1) up to the Boundary Waters subregion (2) near the Canadian
border. In addition, the result for particulate matter from the Appalachian subregion (2) are
given. Previous measurements of particulate matter at sites in the eastern mountains are
available (Stevens et al., 1980); Pierson et al., 1980b; Wolff et al., 1983). The measurements
listed in Table 6-2 include PM25 sulfate, as (NH4)2SO4, organic carbon, and PM(10.25).
The annual average PM2 5 increases substantially from west to east in Table 6-2 from a
value of 3.55 //g/m3 in the northwestern United States to 10.91 //g/m3 in the Appalachian
mountains. The annual average (NH4)2SO4 concentration increases even more substantially from
west to east from a value of 0.88 //g/m3 in the northwestern United States to 6.33 //g/m3 in the
Appalachian Mountains. The lowest annual average organic carbon concentration of 1.38 //g/m3
occurs in the southwestern United States and increases to 2.97 //g/m3 in the Appalachian
Mountains. A smaller range of concentrations occurs for organic carbon from west to east than
for PM2 5 and (NH4)2SO4. The (NH4)2SO4, as a percentage of PM2 5, increases into the
transitional region and the Appalachian Mountains from as low as 25% of the PM2 5 at sites in
the northwestern United States up to 58% at sites in the Appalachian Mountains. Conversely,
organic carbon decrease as a percentage of PM2 5 from 46% at sites in northwestern United
States down to 27% in the Appalachian Mountains. Within the
6-37
-------
TABLE 6-2. ANNUAL AVERAGE CONCENTRATIONS AND
CHEMICAL COMPOSITION FROM IMPROVE MONITORING SITES
Northwest3
Southwest11
California Coastal Mountains0
Transitional Region4
Appalachian Mountains0
No. of
15
5
3
5
2
Annual
3.55
3.91
4.99
5.15
10.91
Average Concentrations, ws/m3 and Composition
(NH4)2S04/%
0.88/25
1.28/33
1.41/28
1.97/38
6.33/58
Organics/
1.63/46
1.38/35
1.95/39
2.01/39
2.97/27
PM
4.46
5.62
8.85
6.54
6.24
8.0
9.5
13.8
11.7
17.2
"Cascades (1), central Rocky Mt. (5), Great Basin (1), N. Rocky Mt. (1), Sierra Nevada (1), Sierra Humboldt (2), and
Colorado Plateau (4)
bColorado Plateau (3), Sonora Desert (2)
°Same as subregion
dWestern Texas (2), northern Great Plains (1), Boundary Waters (2).
western United States there are somewhat higher percentages of (NH4)2SO4 and lower
percentages of organic particles in the southwestern United States than in the northwestern
United States. (NH4)2SO4 plus organic carbon account for from 67% to 85% of PM2 5, with the
higher percentages at IMPROVE sites east of the Rocky Mountains (Table 6-2).
Compared to the estimates discussed by Trijonis (1982, 1991) for "natural" background,
PM2 5 values in the western United States of 1 //g/m3, the average measured contractions of PM2 5
in the northwestern and southwestern United States of 3.55 //g/m3 and 3.91 //g/m3 suggest
anthropogenic contributions. The IMPROVE measurements are likely to include anthropogenic
contributions from sources within North America (background 3). Even the lowest annual
average PM2 5 value in the contiguous United States of 2.5 //g/m3 at Bridger Wilderness Area,
WY, is over twice the "natural" background. The Denali NP in Alaska has an average annual
PM25 of 2 //g/m3 (Malm et al., 1994). The organic carbon concentrations measured there are
somewhat closer to the estimated "natural" background in the western mountains/desert of
0.5 //g/m3 (Trijonis, 1991). However, average annual concentrations in the northwestern and
southwestern United States are higher with values of 1.63 //g/m3 and 1.38 //g/m3. The annual
average values at several IMPROVE monitoring sites in the Rocky Mountains are near 1 //g/m3,
while the Denali NP in Alaska has an average annual organic carbon concentration of
6-38
-------
0.85 //g/m3. These latter organic carbon concentration values ares at the two fold upper limit of
uncertainly in the estimate of "natural" background. On the other hand, the (NH4)2SO4
concentrations measured in the west are far above the "natural" background for (NH4)2SO4 of
0.1 //g/m3 (Trijonis, 1991). The lowest measured annual average (NH4)2SO4 at several sites are
near 0.5 //g/m3. For PM(10.25), the annual average concentrations in the northwestern and
southwestern United States of 4.46 //g/m3 and 5.62 //g/m3 are within the two fold upper limit of
uncertainty in the estimate of "natural" background. At a number of individual sites, annual
average PM(10_25) concentrations are 3 //g/m3 to 3.5 //g/m3, close to the estimated "natural"
background. Therefore, the largest deviations from the "natural" background estimates for a
major component occur for (NH4)2SO4.
Comparisons of the measured concentration values in the "transitional" area of the eastern
United States, using sites from west Texas to the Boundary Waters, find that the average annual
concentrations for PM25 of 5.15 //g/m3; (NH4)2SO4 of 1.97 //g/m3; organic carbon of 2.01 //g/m3
and PM(10_25) of 6.54 //g/m3 (Table 6-2) usually are well above the estimates of "natural"
background in the eastern United States (Trijonis, 1991) for PM25 of 2.3 Mg/m3; (NH4)2SO4 of
0.2 //g/m3; organics of 1.5 //g/m3; and PM(10.25) of 3 //g/m3. As in the western United States, the
measured (NH4)2SO4 concentration values are far above the "natural" background value, while
the measured concentrations of organics are well within the two fold uncertainty in the "natural"
background value.
Another source of lower PM10 concentrations are rural/remote AIRS monitoring sites.
Based on 1993 measurements, the lowest values of PM10 are as follows: Rosebud Co., MT
(maximum of 10 //g/m3, annual mean of 4.5 //g/m3); Campbell Co., WY (maximum of
15 //g/m3, annual mean of 7.0 //g/m3); and Washington Co., ME (maximum of 23 //g/m3, annual
mean of 8.8 //g/m3). These PM10 values agree within a factor of two with the estimated "natural"
background PM10 in the western United States of 4 //g/m3, and in the eastern United States of
5.3 Mg/m3 (Trijonis, 1991).
Seasonal variations in particulate matter are also important and have been considered. The
source used for these seasonal values in particulate matter is the IMPROVE monitoring network
(Malm et al., 1994). Because the seasonal values are reported only by IMPROVE subregions,
there is no good approach to averaging values from differing numbers of sites within the varying
geographical extent of IMPROVE subregions. Therefore, the values of annual average, summer
6-39
-------
and winter values for PM2 5, (NH4)2SO4, organic carbon, and PM(10_2 5) are listed for a number of
IMPROVE subregions (Table 6-3).
TABLE 6-3. ANNUAL SUMMER AND WINTER CONCENTRATIONS FROM
IMPROVE MONITORING SITES3
Subregion Region of U.S.
Central Rockies NW
Colorado Plateau NW-SW
Coastal Mountains NW
Sonora Desert SW
West Texas Transitional to
east
Northern Great Plains Transitional to
east
Boundary Waters Transitional to
east
Appalachian Eastern U.S.
Mountains
No of Seasons of
Sites the Year
5 annual
summer
winter
7 annual
summer
winter
3 annual
summer
winter
2 annual
summer
winter
2 annual
summer
winter
1 annual
summer
winter
2 annual
summer
winter
2 annual
summer
winter
PM,,
3.3
4.8
2.0
3.4
4.1
2.9
5.0
4.5
5.6
4.4
5.6
3.2
5.4
6.6
3.6
4.5
5.6
3.4
5.3
6.2
5.2
10.9
16.6
6.5
(MUSO,
0.8
1.0
0.5
1.1
1.3
0.9
1.4
1.9
0.9
1.5
2.1
1.2
2.1
2.5
1.5
1.5
1.8
1.2
2.0
2.2
2.0
6.3
10.5
3.0
Organics
1.5
2.4
0.9
1.2
1.6
1.1
1.9
1.4
2.3
1.5
1.8
1.1
1.5
1.7
1.1
1.5
2.2
1.1
2.1
3.1
1.4
3.0
4.4
2.0
PM
Coarse
4.8
7.5
3.0
4.7
6.4
3.2
8.9
10.7
7.7
6.0
7.6
3.3
7.5
7.4
5.1
6.3
9.7
3.9
5.7
8.2
3.2
6.2
11.2
3.1
a From Malm et al., 1994.
Annual average concentration almost always are intermediate between the summer and
winter concentration of particulate matter listed in Table 6-3. With a few exceptions, the
6-40
-------
summer concentrations are higher than winter concentrations. The exceptions are the higher
winter concentrations for PM2 5 and organics in the coastal mountains. Ratios of summer to
winter concentrations can equal or exceed two for all listed particulate components in both the
central rockies subregion and the Appalachian Mountains. The summer to winter concentration
ratios for PM2 5 are within the 1.5 to 2.5 range except for the coastal mountains and Boundary
Waters subregion. The summer to winter concentration ratios for PM coarse equal or exceeds
two except for the coastal mountains. Therefore, in most rural remote sites in IMPROVE
subregions summer concentrations of particulate matter substantially exceed winter
concentrations. However, it must be emphasized that it is not appropriate to extrapolate these
results obtained at IMPROVE sites in 1988 to 1991 to other sites or even to other years of
monitoring at IMPROVE sites.
Within the continental United States, there are measurements of particulate mass and
chemical composition under conditions identified as "clean" background conditions (Wolff et al.,
1983). These are based on 7 days of measurements during the summer of 1978 at a site 40 km
northwest of Pierre, South Dakota and 18 days during the summer of 1979 at a site 15 km north
of the Gulf Coast, near Abbeville, LA. At the South Dakota site the small variations in
anthropogenic pollutants observed was attributed to a lack of any major pollution sources along
the trajectories. In contrast, at the Louisiana site the days were stratified into "clean" days when
the air had passed over the Gulf of Mexico for several days and much more polluted episode
days when the maritime air was modified by air which had undergone transport from the
midwestern and northeastern United States.
Fine particle mass on "clean" days averaged 11 to 13 |ig/m3 and coarse mass between 9
and 19 |ig/m3 at the two sites. The total mass averaged between 21 and 32 |ig/m3. Organic
carbon at both sites was the most important fine particle species averaging 4 to 8 |ig/m3 (organic
mass multiplied by 1.2 to include H and O), while sulfate averaged 3 |ig/m3.
At the closest IMPROVE site, the Badlands National Monument, SD in the northern great
plains subregion (Table 6-3), for the summers of 1988 and 1989 (Malm et al., 1994) the
concentrations were PM25, 5.6 //g/m3; (NH4)2SO4, 1.8 //g/m3; organic carbon, 2.2 //g/m3 and
PM(10_2 5), 9.7 //g/m3. These concentration values are substantially lower than those obtained at
the site 40 km northwest of Pierre, SD in the summer of 1978 as follows: PM25, 13 //g/m3,
(NH4)2SO4, 3.2 //g/m3; organic carbon 3.8 //g/m3 and PM(10.25), 19 //g/m3.
6-41
-------
There are several reasons for the differences between the "clean" values and the IMPROVE
values, (a) The measured background varies from year to year and site to site, (b) Precipitation
periods were excluded by Kelly et al. (1982) and Wolff et al. (1983), but the IMPROVE
monitoring network measurements include all weather conditions. All other conditions being
the same, the inclusion of precipitation events in the IMPROVE measurements probably biases
the results low because of rain-out of particulate matter, (c) On the other hand, the presence of
material from anthropogenic sources probably biases the results high. Wolff et al. (1983) used
trajectory analyses to exclude periods with intrusions of polluted air from their analysis. This
was not done with the IMPROVE results. However, the layer-averaged trajectories used by
Wolff et al. (1983) may have underestimated the mixing of air parcels from surrounding
geographical areas leading to an underestimate of the potential for anthropogenic contributions.
The exact causes for the differences between these two types of "background" estimates cannot
be quantitated from available data.
For sulfate, it is possible to make a limited comparison with measurements at rural sites
outside of St. Louis with air flow from the northwest during the third quarters of 1975 and 1976
(Altshuller, 1987), background 5. The average third quarter sulfate concentrations at these sites
for these two years was 7 //g/m3, a substantially higher sulfate concentration than in South
Dakota (Wolff et al., 1983), but lower than measured in other wind directions. These
measurements outside of St. Louis also indicate substantially lower sulfate concentrations during
the first and fourth quarters of 1975, 1976, and 1977 averaging 3.4 //g/m3, comparable to the
third quarter sulfate concentrations in South Dakota.
It is important to emphasize that the "background" for particulate matter moving toward
cities along the east coast over the Great Smoky Mountains (Stevens et al., 1980); the Allegheny
Mountains (Pierson et al., 1980b) and the Blue Ridge Mountains (Wolff et al., 1983),
background 4, are much higher than for the "clean" air days in South Dakota and Louisiana. For
example, the fine particle matter at the Blue Ridge Mountain site in July and August 1980 with
trajectories from the midwest source areas and the Tennessee Valley source area averaged 27
and 24 //g/m3, approximately twice the values under "clean" air conditions in South Dakota and
Louisiana (Wolff et al, 1983). The sulfate concentrations for these two trajectory directions
averaged 14 and 9 //g/m3, with sulfate substantially exceeding organic carbon. This result is a
reversal in the chemical composition under the "clean" air conditions in South Dakota and
6-42
-------
Louisiana, but more consistent with the chemical composition under "episodic" conditions in
Louisiana when the sulfate concentration averaged 20 //g/m3 and the organic carbon averaged 15
Because of the repeated occurrence of (NH4)2SO4 concentrations far above "natural"
background even at rural/remote sites, this aspect justifies additional consideration.
A low contribution of natural sources of gaseous sulfur (both terrestrial and marine) occurs
in the eastern United States (Trijonis, 1991). However, a more detailed consideration of the
contribution of natural sources of gaseous sulfur indicates wide variations over the United States
(Placet, 1991). The following estimates for the ratios of total natural gaseous sulfur to total
anthropogenic gaseous sulfur by region (Placet, 1991) are as follows: northeast, 0.01; southeast,
0.03; west gulf, 0.03; southwest, 0.12; northwest, 0.19. The corresponding ratios for coastal
areas are higher with an estimate of 0.52 for the California coastal areas. If these ratios are
converted to ratios of total natural gaseous sulfur to total gaseous sulfur, the ratios would be 0. 1 1
in the southwest and 0.16 in the northwest. If the following assumptions are made (a) both
natural and anthropogenic sulfur are converted to (NH4)2SO4 to about the same extent; (b) the
concentrations of natural (NH4)2SO4 can be obtained by multiplying the above ratios by the
measured (NH4)2SO4 concentrations, the natural sulfur concentrations in the southeast would
range from 0.1 to 0.15 //g/m3 and in the northwest from 0.08 //g/m3 to 0.2 //g/m3.
A more detailed consideration of the contribution of natural gaseous sulfur at sites near the
Pacific coast is available (Kreidenweis, 1993). In particular, comparisons with measured
(NH4)2SO4 concentrations were made at the Crater Lake National Park in southwestern Oregon
with estimates of natural (NH4)2SO4 concentrations. The measured annual average concentration
at this site of (NH4)2SO4 was 0.5 //g/m3 and an average "low" concentration was approximately
0.13 //g/m3 (Kreidenweis, 1993). This latter value can be compared with several estimates of
natural (NH4)2SO4 concentration based on the following approaches (a) a natural source column
burden between 35 to 50° north of 0.05 to 0.15 //g/m3; (b) a Pacific natural source column
estimate between 35 to 50° N of 0.18 //g/m3 and (c) a 3 D model value of 0.14 to 0.28 //g/m3.
Other approaches gave higher possible values for natural (NH4)2SO4 (a) "clean" rainfall sulfate
concentrations of 0.1 to 0.5 //g/m3 and (b) another 3-D model value of 0.6 //g/m3. These
comparisons results in a wide range of annual average values of (NH4)2SO4 from less than 0.1
3 to less than 0.5 //g/m3 (Kreidenweis, 1993).
6-43
-------
Even an upper limit value for natural (NH4)4SO4 of 0.5 //g/m3 would be a third to a half of
the measured (NH4)2SO4 at IMPROVE sites near the Pacific Coast (Malm et al., 1994). Further
inland, at interior western sites, the marine sources of natural sulfur should make an even smaller
contribution to the measured concentrations of (NH4)2SO4. Comparison of these (NH4)2SO4 with
the estimates based on regional sulfur inventories (Placet, 1991) of 0.08 to 0.2 //g/m3 would
indicate a significant anthropogenic contribution even at relatively remote western IMPROVE
sites. This result suggests that background 3 may have a substantial contribution from
anthropogenic sulfur sources in North America.
As a summary to the discussion in Section 6.3.1.7, the estimated lower limit and upper
limit background concentrations for PM10 and PM2 5 are given on an annual average basis and for
winter and summer for the western and eastern United States in Table 6-4.
TABLE 6-4. SUMMARY OF ANNUAL AND SEASONAL AVERAGE RANGES OF
BACKGROUND CONCENTRATION LEVELS OF PM,n AND PM, s
PM
PM10
PM25
PM10
PM25
PM10
PM7,
Annual or Seasonal
Annual average
Annual average
Winter
Winter
Summer
Summer
Concentrations, Mg/m3
Western United States Eastern United States
4-
1 -
4-
1 -
4-
1 -
8
4
6
3
12
5
5-
2-
5-
2-
5-
2-
11
5
8
4
14
6
The lower limit concentrations are based on the "natural" background midrange
concentrations discussed (Trijonis 1991). There are error factors associated with the chemical
species used to obtain these concentrations range from 1.5 to 3.
The upper limit concentrations are based on measured concentrations from IMPROVE
monitoring sites (Malm et al., 1994). The PM25 concentrations are the sum of concentrations
measured for individual chemical species. As noted earlier in Section 6.3.1.7, these measured
concentrations can include some anthropogenic source contributions within North America.
6-44
-------
Therefore, the upper limit concentrations represent conservative values for the upper end of the
background concentration type.
To obtain the upper limit concentrations, the averages were obtained from the
concentrations for nine subregions in the western United States giving each region equal weight
and also weighing the contribution of each subregion by the number of sites in the subregion.
The median values were also obtained. For the eastern United States, the averages were
obtained from the concentrations for three subregions in the "transitional" region. For the annual
average from 23 individual sites in the western United States and 5 sites in the "transitional"
region (Table 6-3). The resulting values for upper limit concentrations were closely clustered
usually with a 1 //g/m3 range. Within these values, the lower whole value concentration was
listed in Table 6-4.
As a supplement to the data collected in the IMPROVE/NESCAUM networks, seasonal
and annual average PM10 concentrations were also taken from AIRS (1990 - 1995). Four
inhabited areas with the lowest annual average PM10 concentrations were chosen in areas without
nearby IMPROVE/NESCAUM sites. Annual, summer, and winter averages for Penobscot Co.,
ME (11.1, 13.8, and 10.0 |ig/m3); Marquette, MI (11.2, 15.5, and 7.0 |ig/m3); Mercer Co., ND
(11.7, 12.9, and 10.6 |ig/m3); and Lakeport, CA (11.6, 14.3, and 10.0 |ig/m3) all fall within the
upper limits set for PM10 shown in Table 6-4. All areas exhibit summertime maxima and
wintertime minima. The similarity of these results to the upper limits shown in Table 6-4
suggests an anthropogenic component to those upper limits, since the AIRS values were obtained
in inhabited areas.
Again, it should be mentioned that seasonal or annual average "background" values
presented above will likely underpredict 24-hour maximum "background" values. Ambient data
could be used to estimate 24-hour maximum values, but their use is subject to considerable
uncertainty because of possible anthropogenic inputs.
6.3.2 Urban National Aerosol Pattern—Aerometric Information Retrieval
System
The urban monitoring network is operated by state and local agencies as mandated by the
Clean Air Act. The data from this network are used to determine exceedences above the
particulate matter standards. Federal regulations also require that these monitoring data be
6-45
-------
submitted to the EPA Aerometric Information and Retrieval System (AIRS). In what follows,
AIRS PM10 refers to the PM10 mass concentration extracted from the AIRS database. The AIRS
database is a useful resource for analyzing trends and concentration patterns, and relationships
between the fine, coarse, and PM10 components of the atmospheric aerosol (Husar and Frank,
1991; Husar and Poirot, 1992).
The national average AIRS concentrations were calculated utilizing all of the available
data since the beginning of 1985, when less than 100 monitoring stations were operational
(Figure 6-17). Since that time, the number of monitoring stations has risen to more than
1,300 (Figure 6-17). The implications of the changing stations density to the above described
national PM10 trend is not well studied. The emergence of new stations appeared to be in rough
proportion to the final station density shown in Figure 6-17. In other words, in 1985, the
national coverage had a pattern similar to 1994, except less dense. Changes in sampling
equipment and monitoring protocols are also possible causes of systematic errors in the reported
spatial pattern and trends.
The AIRS PM10 database reports the concentrations every sixth day for a 24-h sampling
period, synchronously over the entire country. The sample duration is one day which, over the
long run, provides the concentration distribution function of daily samples. For determination of
the effects (human health, visibility, acid deposition) the concentration has to be known at the
specific location where the sensitive receptors reside. Also the concentrations have to be known
at a short (e.g., daily) time scale, as well as over the long term.
In order to characterize the one day-scale temporal variation over a given region, the entire
available data aggregated over the entire region for each monitoring day are plotted as
6-46
-------
lid f» IVI -I O Sta
US, A. II Stat'on
1986 1988 199O 1992
Figure 6-17. Trend of valid PM10 monitoring stations in the AIRS database.
time series. It is recognized that during the other five non-monitored days, the concentrations
may be different from the reported value. The six day sample increment ensures that both
weekday and weekend data are properly taken into account.
The AIRS PM10 stations are mostly in urban areas but some suburban and nonurban sites
are also reported. The analysis presented in this section is based on PM10 and PM2 5 data
retrieved from AIRS in October 1994.
The results of AIRS PM10 aerosol pattern analysis are presented in quarterly contour maps,
as well as seasonal time charts. For valid monthly and quarterly aggregation, it was required to
have at least two samples a month, and six samples per quarter. For the seasonal maps all the
available data between 1985 to 1994 were used.
The seasonal contour maps also show the location of the PM10 monitoring sites. The size
of the rectangle at each site is proportional to the quarterly average PM10 concentration
6-47
-------
using all available data between 1985 to 1994. Hence, sampling biases due to station density
that changed over time can not be excluded.
The quarterly concentration pattern of PM10 is shown in Figure 6-18. The high sampler
density allows the resolution of spatial texture on the scale of 100 km, particularly over major
metropolitan areas. However, remote regions in the central and western states have poor spatial
density. In the absence of rural monitoring data computerized contour plotting of PM10 is biased
toward extrapolating (spreading) high concentrations over large areas. This bias is particularly
evident in the maps for Quarters 1 and 4 in the western states, where the area of high
concentration hot spots is exaggerated.
The AIRS PM10 concentrations over the eastern United States are lowest during
Quarter 1, ranging between 20 to 30 //g/m3. The higher concentrations exceeding 30 //g/m3 are
confined to metropolitan areas.
6.3.2.1 National Pattern and Trend of Aerometric Information Retrieval System PM10
Two trend analysis approaches were used to obtain the 1988 to 1993 trends in PM10
shown in Figure 19b are subsequent figures providing AIRS concentration patterns. One of
these approaches uses all of the available stations operational each year between 1988 and 1994.
The second approach uses only those stations operational from 1988 to 1994, the long term
coverage, trend, stations.
During the 1988 to 1994 period there were decreases in the annual average PM10 for the
continental U.S. from 33 //g/m3 to 25 //g/m3, for all sites and from 35 //g/m3 to 28 //g/m3 for
trend sites resulting in 24% or 20% reductions in PM10.
The Figure 6-19b also shows the standard deviation among the yearly average PM10
concentrations for each year. On the national scale the standard deviation of yearly average
concentrations is about 40% of the mean.
The concentrations of PM25 and PM10 are compared in the scatter chart in Figure 6-19c.
Each point represents a pair of PM2 5-PM10 monthly average concentrations. The diagonal line is
the 1:1 line and shows the fine particle concentration ranges between 20 and 85% of PM10. The
heavy solid line is derived from a linear best fit regression. The detailed correlation statistics is
reproduced in the upper-left corner of the scatter charts. The
6-48
-------
O tarter 1
VO
, jj—_-. - ./.*::- ., - »;
i •--...., f--::::::::--:- *:.::::: -Si. M-..I
!•- iijfy^j
Mg/m3
• so
30
20
*y.
"' 4
/'
Figure 6-18. AIRS PM10 quarterly concentration maps using all available data.
-------
PM10 Average - Continental US
*";11t
'-? ••,- psHf" Qf-f-^'^f-f
- --, * .•!/> --Y-*, 'f-S'T 1 -
(a)
.. •>»--K?m --•--• - i-- *.» f--4* -: -»-- - ...
;!
- ii\ i .C4>.
>, - >':- ' ..«_.**? 'I- 3:ft -fc *••«??*• l-*-C<
PM10 Cone. Trend - Continental U.S.
EPA AIRS database
1988 1989 1990 1991 1992 1993 1994
-&- Avg for all sites -B- Avg for trend sites
-I- Avg + Std. Dev. -e- Avg - Std. Dev.
PM2.5 vs. PM10 - Conterminous U.S.
EPA AIRS - Monthly Averages
150
140
130
120
I 90
CORRELATION STATS
Avg X : 33.67
AvgY: 10.23
Avg Y/Avg X : 0.57
. CorrCoeff: 0.82
Slope : 0.56
Y offset : 0.24
Data Points : 2269
(c)
20 40 60 80 100 120 140
PM10 (
Seasonal PM Pattern - Continental U.S.
EPA AIRS Database
(d)
1986 Mar May Jul Sep Nov
-&-PM10 -B-PM2.5 -HPM Coarse
Figure 6-19. AIRS PM10 and PM2 5 concentration patterns for the conterminous
United States.
6-50
-------
ratio of overall average PM2 5 and overall average PM10 is also indicated. For the data when
both PM2.5 and PM10 data were available, nationally aggregated PM2 5 particles accounted for
57%ofthePM10mass.
The seasonal pattern of the national PM10 concentration is also depicted in Figure 6-19d,
utilizing all available data in AIRS. The national average PM10 seasonality ranges between
27 //g/m3 in March and April, and 33 //g/m3 in July and August, yielding a modest 16% seasonal
modulation. There is also evidence of slight bimodality with the December through January
peak.
The seasonal chart also shows the annual variation of PM25, and PM10-PM25 (i.e., coarse
particles). The national fine particle concentration shows clear evidence of bimodality with
peaks in July and December. It is shown below that the fine particle winter peak arises from
western sites, while the summer peak is due to eastern U.S. contributions. The national average
coarse particle concentration has a 50 % yearly modulation with a single peak in July.
Stratifying the national PM10 concentrations one can obtain results showing that the country has
several major aerosol regions, as discussed in more detail below. Each region has a discernible
geographic extent as well as seasonal pattern. Over the plains of the eastern United States the
spatial texture of PM10 is driven by the pattern of the emission fields, while the seasonality of
concentrations is likely to be determined by the chemical transformation and removal processes,
as well as by the regional dilution. In the mountainous western and Pacific states, pockets of
wintertime PM10 concentrations exist that well exceed the eastern U.S. values. It is believed that
haze and smoke in confined mountain valleys and air basins are strongly influenced by
topography which in turn influences the emission pattern, dilution, as well as the chemical
transformation and removal rate processes.
Given the regionality of the aerosol concentration pattern much of the discussion that
follows will be focused on the characteristics of these aerosol regions. The Rocky Mountains
produce a natural division between the eastern and western aerosol regimes which will be
discussed next.
6-51
-------
6.3.2.2 Eastern U.S. PM10 Pattern and Trend
During the 1988 to 1994 period there were decreases in the annual average PM10 for the
eastern U.S. from 31 //g/m3 to 26 //g/m3 for all sites and from 34 //g/m3 to 28 //g/m3 for trend
sites resulting in 16% or 18% reductions in PM10 (Figure 6-20b). The decline is rather steady
over time.
The highest eastern U.S. AIRS PM10 concentrations are recorded in Quarter 3
(Figure 6-20d). The peak concentrations are over the Ohio River Valley stretching from
Pittsburgh to West Virginia, southern Indiana and St. Louis. In this region, the PM10
concentration over the industrialized Midwest during the summer can exceed 40 //g/m3.
Additional hot-spots with > 40 //g/m3 are recorded in Birmingham, AL, Atlanta, GA, Nashville,
TN, Philadelphia, PA and Chicago. IL. The summertime PM10 concentrations in New England
and upstate Michigan are < 20 //g/m3.
The transition seasons Quarters 2 and 4 (Figure 6-20d) show PM10 concentrations ranging
from 25 //g/m3 to about 30 //g/m3 over much of the eastern U.S., with concentration hot-spots
over the industrial Midwest as well as in the Southeast, Atlanta, GA and Birmingham, AL. The
PM10 concentrations in urban-industrial "hot-spots" exceed their rural surrounding by less than a
factor of two.
The spatial variability of PM10 occurring over the eastern United States is driven
primarily by the varying primary aerosol emission density. This can be deduced from the
coincidence of higher concentrations within urban industrial areas. The atmospheric dilution
(i.e., horizontal and vertical dispersion) in these areas is not likely to be spatially variable. Also,
the chemical aerosol formation and removal processes are likely to have weak spatial gradients
when averaged over a calendrical quarter. Hence, the main factor that is believed to be
responsible for the spatial variability is the emission field of primary PM10 particles and the SO2,
NOX, and VOC precursors of secondary aerosols.
PM10 concentration in excess of 30 //g/m3 is recorded over the agricultural states of Iowa,
Kansas, Nebraska, and South Dakota. The elevated PM10 concentrations over this region tend to
persist over all four seasons. The eastern PM10 seasonality (Figure 6-20d) is rather
pronounced, with winter concentrations (December through March) of 24 //g/m3, and
6-52
-------
PM10 Average - Eastern US
(a)
J <•-*-. in Cr^/^rC^x•"^^^M^?
PM10 Cone. Trend - Eastern U.S.
EPA AIRS database
1988 1989 1990 1991 1992 1993 1994
-&- Avg for all sites -S- Avg for trend sites
-H Avg + Std. Dev. -©-Avg - Std. Dev.
PM2.5 vs. PM10 - Eastern U.S.
EPA AIRS - Monthly Averages
150
140
CORRELATION STATS
AvgX: 31.4
AvgY: 18.86
Avg Y/Avg X : 0.6
CorrCoBff: 0.83
Slope : 0.58
Y offset : 0.35
Data Pointc : 1651
(c)
20 40 60 80 100 120 140
PM10(|jg/m3)
Seasonal PM Pattern - Eastern U.
EPA AIRS Database
(d)
1986 Mar May Jul Sep Nov
-A-PM10 -B-PM2.S -H PM Coarse
Figure 6-20. AIRS concentration data for east of the Rockies: (a) monitoring locations;
(b) PM10 concentration trends; (c) PM10 and PM2 5 relationship; and
(d) PM10, PM2 5, and PMCoarse seasonal pattern.
6-53
-------
July through August peak of 35 //g/m3. The amplitude of the PM10 seasonal concentrations is
about 30%.
The scatter chart of the eastern AIRS PM2 5-PM10 relationship shows a significant amount
of scatter, with a slope of 0.58 (Figure 6-20b). The ratio of the overall average PM25 and PM10
concentration is 0.6 such that 60% of PM10 in the sub 2.5 //m size range. The seasonality of the
fine AIRS particle concentration over the East is bimodal with a major peak in July and a
smaller winter peak in January (Figure 6-20d). As shown in Figure 6-15b, the nonurban
IMPROVE/NESCAUM network results for the eastern U.S. for PM2 5 show a peak in summer
but does not show a winter peak. The coarse particle concentration shows a single broad peak
over the warm season, April through October (Figure 6-20d), but with a somwhat different
pattern than shown in Figure 6-15b for nonurban cities in the eastern U.S. It is therefore evident
that fine and coarse particles (from urban and nonurban measurements) have different seasonal
dynamics in the East.
6.3.2.3 Western U.S. PM10 Pattern and Trend
The mountainous states, west of the Rockies (Figure 6-21) have higher PM10
concentrations in Quarters 1 and 4 than in Quarters 2 and 4 and shown ever higher PM10
concentrations (>50 //g/m3) at localized hot-spots. These higher concentrations occur over both
metropolitan areas such as Salt Lake City, as well as in smaller towns in mountain valleys of
states west of the Rockies.
The main geographic feature regions considered in California are the Los Angeles basin
and the San Joaquin Valley. Both basins show monthly PM10 concentrations sometimes in
excess of 50 //g/m3. These basins are also confined by surrounding mountains that limit the
dilution, facilitate cloud formation, and have emissions that are confined to the basin floor.
Accordingly, they represent airsheds with characteristic spatial and temporal pattern. It is likely
that the actual local effects on the PM10 concentration field in the mountainous western states are
greater than depicted in Figure 6-2la.
It appears that the spatial pattern of these high concentration hot spots is driven by
emissions as well as by the restricted wintertime ventilation due to mountainous terrain. Over
the mountainous western states the atmospheric dilution by horizontal and vertical dispersion is
6-54
-------
PM10 Average - Western US
~f:''ff:y-^S"^-Ff-'"~!f^-^ *':!'.l'~-y-""- ......... - ..... """: ......... ==»»»»»»»»»»
:::a=
--'-'- ' - PTV» •*•=• \ - •;,, JJ__fe..* .% 1 .-.. vj
e *.. ^ ./ -.f**-^t" .-- --. ^=v-**BffA";(~«^-S- ' •j-"-tsss£:--:T~vT.-.- ^f-'
" "
-- -~ -
. _ ,4,.,™
(a)
PM10 Cone. Trend -Western U.S.
EPA AIRS database
°988 1989 1990 1991 1992 1993 1994
-A- Avg for all sites -B- Avg for trend sites
-I- Avg + Std. Dev. -S- Avg - Std. Dev.
PM2.5 vs. PM10 -Western U.S.
EPA AIRS - Monthly Averages
120
110
100
90
10 70
N
S
Q. 60
CORRELATION STATS:
Avg X : 39.75
Avg Y : 20.22
. Avg Y/Avg X : 0.5
Corr Coeff: 0.84
Slope : 0.57
Y offset : -2.81
Data Points : 618
20
10
(c)
Seasonal PM Pattern -Western U.S.
EPA AIRS Database
(d)
20 40 60 80 100 120 140
PM10 (Lig/m3)
1986 Mar May Jul Sep Nov
-A- PM10 -B- PM2.5 -I- PM Coarse
Figure 6-21. AIRS concentration data for west of the Rockies: (a) monitoring trends;
(b) PM10 concentration trends; PM10 and PM2 5 relationship; and (d) PM10,
PM2 5, and PMCoarse seasonal pattern.
6-55
-------
severely restricted by mountain barriers and atmospheric stratification due to strong and shallow
inversions. Radiative cooling also causes fog formation which enhances the production rate of
hydroscopic aerosols in the valleys. As a consequence, mountain tops are generally protruding
out of haze layers. Emissions arising from industrial, residential, agricultural, unpaved
roadways and other sources are generally confined to mountain valleys. In the wintertime the
mountain valleys are frequently filled with fog. All three major factors that determine the
ambient concentrations (i.e., emissions, dilution, and chemical rate processes) are strongly
influenced by the topography. For this reason, many of the maps depicting the regional pattern
use shaded topography as a backdrop.
In the western half of the U.S., west of and including the Rockies, there was a decrease in
the PM10 concentration of 1988 to 1994 from 36 //g/m3 to 25 //g/m3 for all sites and from 39
//g/m3 to 28 //g/m3 for trend sites (Figure 6-21b). The reductions were 31% for all sites and 28%
for trend sites. Standard deviation among the western stations of yearly average PM10
concentrations is about 40%.
The western AIRS PM2 5-PM10 relationship (Figure 6-2 Ic) shows that on the average about
50% of the PM10 is contributed by fine particles. The scatter of data points (Figure 6-2 Ic) also
shows that during high concentration PM10 episodes the fine fraction dominates.
The western PM10 seasonality (Figure 6-21d) is also rather pronounced, having about 30%
amplitude. However, the lowest concentrations (26 //g/m3) are reported in the late spring (April
through June), while the highest values occur in late fall (October through January).
The seasonality of PM25 west of the Rockies (Figure 6-2 Id) is strongly peaked in
November through January. In fact, the PM2 5 is several times higher than the summertime
values. On the other hand, the coarse fraction shows a broad peak during late summer, July
through October. It is to be noted that in Figures 6-20 and 6-21, the fine and coarse particle
concentrations do not add up to PM 10, because size resolved samples were only available for
tens of sites, while the PM10 concentrations were obtained from hundreds of monitoring stations.
In summary, there is a 20 to 24% reduction of PM10 concentrations for the continental U.S.
between 1988 and 1993. On the national average the PM10 seasonality is weak. Desegregation
of the national averages into east and west of the Rockies, shows that the downward trend west
of the Rockies is more pronounced than over the eastern half of the U.S. The east-west
desegregation also shows that the lack of national PM10 seasonality arises from two strong
6-56
-------
seasonal signals that are phase shifted, the eastern United States has a summer peak, the West a
fall and winter peak, and the sum of two signals is a weakly modulated seasonal pattern.
Nationally, PM2 5 mass accounts for about 57% of PM10 mass. The East and West show
comparable average fine particle fractions (60% in the East and 50% in the West). Fine particles
tend to dominate during the fall and winter season in the western U.S., except in the southwest.
It is evident that further examination discussed in the next sections will show that the East-
West division itself is rather crude and that dividing the conterminous United States into
additional subregions is beneficial in explaining the PM10 concentration pattern and trends.
6.3.2.4 Short-Term Variability of PM10 Concentrations
The previous aerosol concentration patterns were expressed as quarterly averages.
However, for health and other effects, the variance of the concentration, in particular the
occurrence of extreme high concentrations is of importance. The PM10 concentrations exhibit
marked differences in the shape of their distribution functions around the mean values. For
example in Figure 6-22, the day to day variations of PM10 concentrations in Knoxville, TN are
about 40% of the mean value of 35 //g/m3. On the other hand, the concentration time series for
Missoula, MT shows a coefficient of variation of 60% over the mean of 34 //g/m3. During the
winter season the coefficient of variation is even higher. It is therefore evident, that for
comparable mean concentrations the Missoula, MT site exhibits significantly higher short-term
variations. Also note the large variations from a high concentration day to the lower
concentrations on the day before and/or the day after (Figure 6-22).
The variability of concentration is examined spatially and seasonally by computing
logarithmic standard deviation (ratio of 84/50 concentration percentiles) for each monitoring
site. These deviations were then contoured for each season. The results are depicted in the
seasonal maps of the logarithmic standard deviation (Figure 6-23). The highest logarithmic
standard deviation is recorded over the northern and northwestern states during the cold
6-57
-------
140
'239
Mean : 34
CoVa : 60.64
Min : 1
Max : 239
Points: 1660
20
0 '•
1988
80
1989
1990
1991
1992
1993
Mean : 35
CoVa : 39.92
« 60 Min : 9
£ Max : 73
Points: 258
- 40
20
0L
1988
1989
1990
1991
1992
1993
Figure 6-22. Short-term PM10 concentration time series for Missoula, MT, and Knoxville,
TN.
season, Quarters 1 and 4. Regionally, the logarithmic standard deviation in the north-northwest
is about 2.0 with pockets of high winter variability such as Salt Lake City, UT, and Missoula,
MT. The lowest variability prevails over the warm season, Quarters 2 and 3, covering the
southeastern and southwestern states. Over multistate regions in the southern states the
summertime logarithmic standard deviation is below 1.5. This means that these areas are
covered more or less uniformly by summertime PM10, while the northern states are more
episodic.
6-58
-------
Quarter 1
2
fL,
r?...
VO
-4"
-^
.5'
Qy-arttH 4
Figure 6-23. Geographic variation of the standard deviation, og, of the lognormal distribution of PM10 concentrations from
the AIRS.
-------
6.3.2.5 Aerometric Information Retrieval System PM2 5 Concentrations
The mass concentration of fine particles in urban areas is not well known. Sampling and
analysis of PM25 is limited by small number of stations (<50), sampling period restricted to few
years, and different, non-standard sampling equipment was utilized for PM2 5
The yearly average AIRS PM2 5 concentrations are shown in Figure 6-24. Figure 6-24 also
shows the location and magnitude of PM25 concentrations from measurements of
IMPROVE/NESCAUM monitoring networks. The fine particle data from the
IMPROVE/NESCAUM show a pattern of high concentrations (> 15 //g/m3) occurring over the
eastern United States. This pattern of nonurban fine particle concentrations was discussed in
Section 6.3.1.
6.3.2.6 Other National Surveys
A summary of urban PM10, PM25, PMCoarse at eight urban areas, Birmingham, AL,
Buffalo, NY, Houston, TX, Philadelphia, PA, Phoenix, AZ, Pittsburgh, PA, Rubidoux, CA, and
Steubenville, OH was reported by Rodes and Evans (1985). The overall ratio of the PM10 to
Total Suspended Particulate (TSP) was 0.486. The relationships between PM10 and the 15 //m
fraction (IP) are linear for all sites. With exception of Phoenix, AZ, and Houston, TX, PM25
exceeded the PMCoarse mass concentration in all six urban areas.
Spengler and Thurston (1983) reported PM concentrations in six U.S. cities: Portage, WI,
Topeka, KS, Kingston, TN, Watertown, MA, St. Louis, MO, and Steubenville, OH, using
dichotomous virtual impactors in the two size ranges, PM25, having dp<2.5 //m, and coarse
particle mass with 2.5
-------
AIRS PM2.5 - IMPROVE PM2.5 Comparison
AIRS PM2.5 f IMPROVE/NESCAUM PM2.5
Figure 6-24. Annual PM2 5 concentration pattern obtained from IMPROVE/NESCAUM and AIRS networks.
-------
30
Portage, Wl
m
E
D)
20
10
• IP mass
• Fine mass
• Course mass
* Total sulfate mass
r>
60
50
40
o> 30
n.
20
10
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
Harriman, TN
°IP mass
•Fine mass
'Course mass
Total sulfate mass.
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
90
80
70
60
| 50
O)
a40
30
20
10
Topeka. KS
• IP mass
• Fine mass
• Course mass
'Total sulfate mass
60
50
40
rt
0)30
a.
20
10
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
Watertown, MA
, • IP mass
• Fine mass
h • Course mass
'Total sulfate mass
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
70
60
50
| 40
O)
a
30
20
10
St. Louis, MO
•IP mass
•Fine mass
Jcourse mass
Total sulfate mass
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
90
80
70
60
| 50
O)
11 40
30
20
10
Steubenville. OH
*IP mass
AFine mass
•Course mass
VTotal sulfate mas
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1979 1980 1981
Figure 6-25. Monthly mean concentrations in micrograms per cubic meter of PM15 (IP,
inhalable mass), PM2 5 (fine mass), coarse mass (PM15-PM2 5), and total sulfate
as (NH4)2SO4 in Portage, WI; Topeka, KS; Harriman, TN; Watertown, MA;
St. Louis, MO; and Steubenville, OH.
6-62
-------
The quartz content and elemental composition of aerosols, collected in dichotomous
samplers in selected sites in the EPA Inhalable Paniculate network, were reported by Davis et al.
(1984). For all network sites, an average of only 4.9 weight per cent of the coarse particle mass
and 0.4 weight per cent of the fine mass consisted of quartz. Continental interior sites show the
highest average quartz content as well as the greatest variability. The coastal regions and eastern
interior sites reveal the lowest quartz concentrations. The complete X-ray spectra from some
samples in Portland, OR, show that Si comes primarily from minerals such as feldspars, where
the Si in the Buffalo, NY aerosols comes from quartz.
6.3.3 Comparison of Urban and Nonurban Concentrations
Seasonal maps of the AIRS PM10-IMPROVE/NESCAUM PM10 spatial concentrations are
given in Figure 6-26. In evaluating the subsequent comparisons of the differences between
AIRS and IMPROVE/NESCAUM spatial concentrations possible sampling biases and
differences in sampling equipment and monitoring protocols may be significant. In addition, the
differences in geographical location between the stations for the two networks also can influence
the reliability of these comparisons. The AIRS PM2 5 concentrations everywhere exceed their
adjacent IMPROVE/ NESCAUM concentrations. The highest AIRS PM2 5 are reported over the
eastern urban industrial centers, such as Philadelphia and Pittsburgh, where the concentrations of
20 to 30 //g/m3 exceed the nonurban PM25 by a factor of 2 to 3. However, the excess urban
PM2 5 concentrations are evidently confined to the immediate vicinity of urban centers. This
indicates that over the eastern United States a regionally homogeneous background of PM2 5
concentration exists that has smooth spatial gradients. Superimposed on the smooth regional
pattern are local hot-spots with excess concentrations of factor of 2 to 3 that are confined within
a few miles of urban industrial centers. The regional homogeneity is an indication that the
eastern U.S. PM25 is composed of a secondary aerosol that is produced several days after the
emission of its gaseous precursors. Similar results have been discussed for SO42" since the 1970's
(Altshuller, 1980). The excess PM25 concentration in urban centers suggests that primary
emissions such as automobile exhaust and heating furnaces are responsible for much the urban
PM2 5 hot-spots.
6-63
-------
..-*
Quarter 2
Mg/m3
^•^ gQ
•30
20
10
0
PM10AIS* - PVia
/**===•
s-M * ? Ckf ^ 5 c-M 1 P t » PR & v'i
MS/m
J50
30
20
-10
Rs f V- <•. fi
Figure 6-26. Spatial maps of PM10 concentration difference between AIRS and IMPROVE/NESCAUM networks.
-------
The reported AIRS PM2 5 concentrations over the Pacific states are generally higher and
average at 20 to 50 //g/m3. This is 5 to 10 times higher than their companion IMPROVE PM2 5
concentrations. The dramatic difference is attributable to the pronounced concentration
differences between urban-industrial-agricultural centers that exist in mountainous air basins
and the concentrations monitored at remote national parks and wilderness areas that are
generally at higher elevations. However, it is fair to presume that the AIRS and IMPROVE
PM25 data sets represent the extreme of aerosol concentration ranges that exist over the western
U.S. The challenging task of filling in the details (i.e., spatially and temporally extrapolating the
aerosol concentrations over the rugged western United States) is discussed in further detail in
later regionally and locally focused sections below.
Comparisons have been made of the seasonality of the urban (AIRS) concentrations
relative to the nonurban (IMPROVE/NESCAUM) data. In Figure 6-27 the difference in PM10,
PM2 5, and PMCoarse between AIRS and IMPROVE/NESCAUM sites, using all available data,
is used to indicate the urban excess particle concentration compared to the rural concentration.
No attempt has been made to evaluate the possible uncertainties in these difference values.
Nationally, the urban excess fine particle concentration ranges between 18 //g/m3 in
December through February and 10 //g/m3 in April through June (Figure 6-27a). The urban
excess coarse mass concentration ranges between 10 to 7 //g/m3. The sum of the fine and coarse
national urban excess mass concentration is about 25 //g/m3 in the winter season, and 18 //g/m3
during the spring season. Hence, the nationally aggregated urban and nonurban data confirm
that urban areas may have excess concentrations on the order of 20 //g/m3, and well over half is
due to fine particles, particularly in the winter season.
The urban excess (AIRS-IMPROVE/NESCAUM difference) over the eastern United States
(Figure 6-27b) shows fine particles excess of 8 to 12 //g/m3, with higher value occurring during
both winter and summer. The urban excess coarse mass in the eastern United States is only 5 to
8 //g/m3, peaking during spring and summer. The sum of fine and coarse urban excess is 15 to
18 //g/m3 throughout the year.
6-65
-------
Urban Excess
Jan Mar May Jul Sep Nov Jan Jan Mar May Jul Sep Nov Jan Jan Mar May Ju I Sep Nov Jan
Fine + Coarse Mass
Fine — C— Coarse
Figure 6-27. Urban excess concentrations (AIRS minus IMPROVE) for (a) the United
States, (b) the eastern United States, and (c) the western United States.
The excess urban (AIRS-IMPROVE/NESCAUM) aggregated over the western United
States is much more pronounced in magnitude and seasonality. The urban excess fine mass is
about 30 //g/m3 in November through January and drops to 8 to 10 //g/m3 in April through
August. The urban excess coarse mass is less in magnitude and seasonality 15 to 18 //g/m3 in
July through December, and 10 to 12 //g/m3 in March through May. The sum of the urban
excess fine and coarse mass is 40 to 50 //g/m3 in November through January and about 20 //g/m3
in the spring March through June. The urban AIRS and nonurban IMPROVE) networks in the
western United States monitor aerosols differently because of different goals and mandates. The
urban nonurban difference is such that the western nonurban concentrations contribute little to
the much higher urban values, particularly in the winter season. On the other hand, the eastern
urban sites are greatly influenced by the nonurban, regionally representative concentrations,
particularly in the summer season.
6-66
-------
6.4 REGIONAL PATTERNS AND TRENDS
This section describes the spatial, temporal, size, and chemical characteristics of seven
aerosol regions of the conterminous U.S. The sizes and locations of these regions were chosen
mainly on the basis of the characteristics of their aerosol pattern. The main criteria for
delineating a region were (1) the region had to possess some uniqueness in aerosol trends,
seasonality, size distribution, or chemical composition; (2) each territory of the conterminous
United States had to belong to one of the regions; and (3) for reasons of computational
convenience the shape of the regions were selected to be rectangular on unprojected latitude
longitude maps. The resulting criteria yielded seven rectangular aerosol regions as shown in
Figure 6-28. It is recognized that this selection is arbitrary and for future analysis additional
regional definition criteria would be desirable. The limitations in the data bases of the two
different networks discussed previously also apply to the subsequent discussion.
For sake of consistency and intercomparisons each region is described using maps
delineating the spatial pattern and the sampling locations in the subsequent figures (Section a).
For the figures showing AIRS monitoring results, Section b shows trends in average PM10
concentrations and ± a. As discussed in Section 6.3.2.1 included in (b) are the results of two
trend analyses. One of these uses the annual concentrations from all available stations in
operation any time in the 1985 to 1994 period. The second approach uses the annual
concentrations from only those stations operated continuously from 1985 to 1994, the long term
coverage or trend stations. Section c show plots and correlations relating PM10 and PM2 5.
Monthly AIRS concentrations (Section d) for a given region were computed by averaging all the
available data for the specific month. In case of nonurban aerosol chemistry some regions only
had two to four monitoring stations. The monthly nonurban PM2 5, PMCoarse and PM10 shown
in the subsequent figures (Section b) over regions illustrate the relative seasonality of each
aerosol type. The nonurban regional average chemical composition is presented as seasonal
charts of chemical aerosol components as a fraction of the fine mass concentration (Section c).
The role of some primary sources, such as coal and fuel oil combustion is indicated through
seasonal charts of selenium (coal) and vanadium (fuel oil) trace metals (Section d). In addition,
for each region figures will be provided showing shorter term variability of PM10 concentrations
and PM10 urban excess concentrations.
6-67
-------
Industrial
Midwest
PM10 = 29
PM2.5= 17
Northeast
PM10 = 34
PM2.5 = 21
PM2.5/10 = 0.62
Upper
Midwest
PM10 = 31
PM2.5= 12
Northwest
PM10 = 28
PM2.5= 16
PM2.5/10 = 0.59
PM2.5/10 = 0.38 PM2.5/10 = 0.59
S.California
PM10=53
PM2.5=26
PM2.5/10=0.49
Southeast
PM 10=29
PM2.5=17
PM2.5/10=0.5a
Southwest
PM 10=34
PM2.5=12
PM2 5/10=0.37
Figure 6-28. Aerosol regions of the conterminous United States.
6.4.1 Regional Aerosol Pattern in Eastern New York, New Jersey, and the
Northeast
The Northeast aerosol region covers the New England states, including eastern
Pennsylvania and eastern Virginia to the south (Figure 6-29a). In the Northeast, terrain
features that significantly influence regional ventilation occur over the mountainous upstate
New York, Vermont and New Hampshire. Throughout the year, the Northeast is influenced
by Canadian as well as Gulf air masses. The region includes the Boston-New York
megalopolis, as well as other urban-industrial centers. It is known that the Northeast is
influenced by both local sources, as well as long range transport of fine particles and
transformations of precursors to particles from other regions, as well as transport
6-68
-------
PM2.5 Concentration - Northeast
IMPROVE/NESCAUM Data
PM10, PM2.5 and PMC - Northeast
IMPROVE/NESCAUM Data
40,000
35,000
™ 25,000
O
o
15,000
10,000
(b)
Chemical Fine Mass Balance - Northeast
IMPROVE/NESCAUM Data
1989 Mar May Jul Sep Nov
-+- PM2.5 -A- PM coarse
Chemical Tracers - Northeast
IMPROVE/NESCAUM Data
o
o
3,500
3,000
2.500
2,000
1,500
1,000
(d)
1989 Mar May Jul Sep Nov
OC +Soll
Sulfate + OC + Soil + EC
1989 Mar May Jul Sep Nov
Sulfur - Max = 4000 Selenium - Max = 4
Vanadium - Max = 10 S/Se - Max = 4000
Figure 6-29. IMPROVE/NESCAUM concentration data for the Northeast: (a) monitoring
locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate, soil, organic
carbon (OC), and elemental carbon (EC) fractions; and (d) tracers.
6-69
-------
and transformation of precursors in single and multiple urban plumes within the region
(Chapter 3).
6.4.1.1 Nonurban Size and Chemical Composition in the Northeast
The summary of the nonurban aerosol chemical composition in the Northeast is presented
in Figure 6-29c. The region has 14 monitoring sites, 8 of which are part of NESCAUM in upper
New England. The geographic locations with respect to nearby urban areas vary from those sites
within the northeast corridor to rural sites near the Canadian border.
The PM10 concentration exhibits a factor of two seasonal amplitude between 12 //g/m3 in
the winter, and 25 //g/m3 in June and July (Figure 6-29b). About 60% of PM10 is contributed by
fine particles throughout the year. The PM2 5 also contributes to the summer-peaked seasonality.
Data from a two year fine particle network in the Northeast (Bennett et al., 1994) yielded a
geometric mean concentration of PM25 of 12.9 //g/m3 and particulate sulfur (1.4 //g/m3,
equivalent to 4.2 //g/m3 of sulfate), which is somewhat lower than other comparable rural data.
Sulfates are the most important contributors of the fine particle mass in the Northeast,
particularly in the summer season when they account for half of the fine mass (Figure 6-29c).
The regionality of sulfate in the northeastern U.S. has been dicussed for many years (Altshuller,
1980). The organic carbon ranges from 30 to 40%, with the higher percentages occurring in the
fall and winter, September through January. In fact, during the late fall the contributions of
sulfate and organic carbon are comparable at 40%. Fine particle soil is unimportant throughout
the year (<5%). Elemental carbon, on the other hand, is somewhat more significant, particularly
during the fall when it contributes about 10% of the fine mass. The sum of the above four
nonurban fine particle aerosol components, account for over 90% of the measured fine particle
mass throughout the year. These results would appear to indicate ammonium ion, hydrogen ion,
nitrates, trace metals and sea salt are of minor importance in the northeastern U.S. fine particle
chemical mass balance.
6-70
-------
The seasonality of both selenium and vanadium indicates a winter peak (Figure 6-29d).
In particular, the vanadium concentration increases by a factor of two for December and
January compared to the summer values. Also, the V concentration is higher than over any
other region indicating the strongest contribution of fuel oil emissions. The S/Se ratio is
strongly seasonal with a winter value of 1,000 and a summer peak of 2,000 to 2,500 consistent
with a substantial secondary photochemical contribution of SO42" during the summer.
6.4.1.2 Urban Aerosols in the Northeast
In the northeastern U.S. there was a decrease in the annual average PM10 concentration
between 1988 and 1994 from 28 //g/m3 to 23 //g/m3 for all sites and from 31 //g/m3 to 25 //g/m3
for trend sites (Figure 6-30b). The reductions were 18% for all sites and 19% for trend sites.
The standard deviation among the monitoring stations for any given year is about 30%. The
map of the Northeast shows the magnitude of PM10 concentrations in proportion of circle radius.
The highest AIRS PM10 concentrations tend to occur in larger urban centers (Figure 6-3Oa).
The seasonality of the urban Northeast PM10 concentration (Figure 6-30d) is a modest
20%, ranging from 25 to 31 //g/m3, smaller than the seasonality of the nonurban northwest PM10
(Figure 6-29b). There is a summer peak in July, and a rather uniform concentration between
September and May showing only a slight winter peak. The PM2 5-PM10 relationship
(Figure 6-30c) shows that on the average 62% of PM10 is contributed by fine particles.
In general, the regional scale emissions are not expected to vary significantly from one day
to another. However, both meteorological transport (i.e., dilution), as well as aerosol formation
and removal processes, are important modulators of daily aerosol concentration. The daily
concentration of particulate matter exhibits strong fluctuation from one day to another, mainly
due to the role of the meteorological transport variability.
The regionally averaged daily concentration is associated with the regional scale
meteorological ventilation. High regionally averaged concentrations indicate poor ventilation
(i.e., a combination of low wind speeds and low mixing heights and the absence of fast aerosol
removal rates, such as cloud scavenging and precipitation). Low regional concentrations, on the
other hand, represent strong horizontal transport, deep mixing heights, or high regional
6-71
-------
PM10 Average - Northeast
(a)
=f*3L4"~ ^ ' "".i
£^ . '
™»H -"«""--S»i S-:.,i,!/-«? ":,«jftrf^
>-.r,-,:
:--r*f-*««&*:Jll-- '-^-iJr* •
2v?:...:i*.*:=::^:s. f- : jt^f*
r'p-r-cv-^fevjKM -•vc*J^> „
sr
•" "'" -'" "•»
LJ CT «!«<'
150
140
130
120
110
100
= 90
D)
PM2.5 vs. PM10 - Northeast
EPA AIRS - Monthly Averages
10
CORRELATION STATS:
Avg X : 34.28
Avg Y : 21.54
AvgY/AvgX: 0.62
Corr Coeff: 0.87
Slope: 0.63
Yoffset: -0.34
Data Points: 755
(c)
20 40 60 80 100 120 140
PM10([jg/m3)
PM10 Cone. Trend - Northeastern U.S.
EPA AIRS database
1989 1990 1991 1992 1993 1994
-A- Avg for all sites -B- Avg for trend sites
-+- Avg + Std. Dev. -e- Avg - Std. Dev.
Seasonal PM Pattern - Northeast
EPA AIRS Database
60
45
I35
= 30
25
15
(d)
1986 Mar May Jul
-A-PM10 -B-PM2.5 •
Sep Nov
• PM Coarse
Figure 6-30 AIRS concentration data for the Northeast: (a) monitoring locations;
(b) regional PM10 concentration trends; (c) PM10 and PM2 5 relationship;
and (d) PM10, PM2 5, and PMCoarse seasonal pattern.
6-72
-------
removal rates. Advection of high aerosol content air masses from neighboring regions may also
be a cause of elevated concentration in a given region.
The daily variation of the regional averaged urban PM10 concentration for the Northeast is
shown in Figure 6-31. The single day concentration data for every sixth day are connected by a
line between the data points, although five in-between days are not monitored. The lowest
regionally averaged daily urban PM10 is about 10 //g/m3, while the highest is about 55 //g/m3,
with a regional average in the early 1990s of 25 //g/m3. The highest concentrations (>40 //g/m3)
occur primarily in the summer season. The time series also indicate that the high concentration
episodes do not persist over consecutive six day periods. This is consistent with the notion that
the regional ventilation that is caused by synoptic scale air mass changes, which typically occur
every four to seven days over eastern U.S. The daily time series also convey the fact that day to
day variation in PM10 is higher than the seasonal amplitude. In fact, in Figure 6-31 the
concentration seasonality is barely discernible. It can be stated, therefore, that the PM10
concentration in the Northeast is highly episodic (i.e., the temporal concentration variation is
both substantial and irregular). The excess urban PM10 (AIRS-IMPROVE) is shown in
Figure 6-32. The excess urban PM10 concentration in the Northeast is a relatively small part of
the total urban PM10 concentration between May and October. The reliability of such estimates
of excess regional urban PM10 concentrations discussed earlier should be considered
(Section 6.3.3).
6.4.2 Regional Aerosol Pattern in the Southeast
The Southeast rectangle stretches from North Carolina to eastern Texas (Figure 6-33).
From the point of view of regional ventilation the Southeast terrain is flat, with the exception of
the mildly rolling southern Appalachian Mountains. The region is known for increasing
population over the past decades, high summertime humidity, and poor regional ventilation due
to stagnating high pressure systems.
6.4.2.1 Nonurban Size and Chemical Composition in the Southeast
Only six nonurban stations were available in the Southeast with the absence of monitoring
over the southern (Gulf Coast) part of the region, except for Florida. The
6-73
-------
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
Northeast Every Sixth Day
1991
1992
1993
Figure 6-31. Short-term variation of PM10 average for the Northeast. Data are reported
every sixth day.
Northeast urban excess
Jan Mar May Jul Sep Nov Ja
Figure 6-32. Urban excess concentration (AIRS minus IMPROVE) for the Northeast.
6-74
-------
PM2.5 Concentration - Southeast
IMPROVE/NESCAUM Data
Chemical Fine Mass Balance -Southeast
IMPROVE/NESCAUM Data
8
n
(c)
1989 Mar May Jul Sep Nov
OC ^Soil
Sulfate + OC + Soil + EC
PM10, PM2.5 and PMC - Southeast
IMPROVE/NESCAUM Data
40,000
o
15
o
o
35,000
30,000
25,000
20,000
15,000
10,000
5,000
(b)
Mar May Jul Sep Nov
-A- PM Coarse
Chemical Tracers - Southeast
IMPROVE/NESCAUM Data
o
o
4,000
3,500
3,000
2,500
2,000
1,500
1,000
(d)
1989 Mar May Jul Sep Nov
Sulfur-Max = 4000 Selenium - Max =
Vanadium - Max = 10 S/Se - Max = 4000
Figure 6-33. IMPROVE/NESCAUM concentration data for the Southeast: (a) monitoring
locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate, soil, organic
carbon (OC), and elemental carbon (EC) fractions; and (d) tracers.
6-75
-------
nonurban PM10 concentration in the Southeast (Figure 6-33b) is roughly comparable to the
Northeast, exhibiting about factor of two seasonal concentration amplitude between 12 //g/m3 in
the winter, and 25 //g/m3 in the summer. An anomalous high PM10 concentration is shown in
July which appears to be contributed by an excess coarse particle concentration of about
10 //g/m3. With exception of July, the fine particle mass accounts for about 70% of the
nonurban PM10, leaving the coarse mass of 30% or less throughout the year (Figure 6-33b).
The most prominent aerosol species in the Southeast are sulfates contributing 40 to 50% of
the fine mass (Figures 6-33c). The anomalously low sulfate fraction (35%) during July
coincides with the high (20%) soil contribution during July. For the other months, soil
contribution is <5% of the fine mass. The relative role of the organic carbon in the nonurban
Southeast is most pronounced during the winter (40%), but declines to 25% during the summer
months. The contribution of elemental carbon varies between 2% in the summer to 6% in the
winter months.
The trace element concentrations of selenium and vanadium (Figure 6-33d) are constant
throughout the year, implying that the combined role of emissions and dilution is seasonally
invariant. The concentration of sulfur, on other hand shows a definite summer peak, that is two
to three times higher than the winter concentrations. Consequently, the S/Se ratio is strongly
seasonal. In fact, the warm season S/Se ratio of 2,500 is higher than over any other region of the
country. If Se-bearing coal combustion is the dominant source of sulfur in the Southeast, then
the high S/Se ratio implies that the secondary photochemical sulfate production in the summer is
several times that in the winter.
6.4.2.2 Urban Aerosols in the Southeast
In the southeastern U.S. there was a decrease in the annual average PM10 concentrations
between 1988 and 1994 from 33 //g/m3 to 27 //g/m3 for all sites and from 35 //g/m3 to 29 //g/m3
for trend sites (Figure 6-34b). The reductions were 18% for all sites and 17% for trend sites.
The Southeast PM10 concentration trends and the PM10 seasonality resemble the industrial
Midwest described below. A unique feature of the Southeast is the uniformity of the aerosol
concentration among the monitoring stations. In fact the 17% station to station
6-76
-------
PM10 Average - Southeast
#aj™1 |""""^"J|-""""""""""-% jP""""«t -^ .. J^^^JP?^'"""""^ J
\. .: "Jf'"'":. "::vv*r:":.. £--?" ^ - "" ^^ t ^" ip'L-. I """""t ..-«-==.
PM2.5 vs. PM10 - Southeast
EPA AIRS - Monthly Averages
CORRELATION STATS
AvgX : 29.19
Avg Y : 16.32
Avg Y/Avg X : 0.55
Corr Coaff: 0.63
Slopo : 0.43
Y offset: 3.61
Data Points : 352
<"» 90
60
50
^4:;!f HoXoV
PM10 Cone. Trend - Southeastern U.S.
EPA AIRS database
1989 1990 1991 1992 1993 1994
-A- Avg for all sites -B- Avg for trend sites
-I-Avg + Std. Dev. -»-Avg -Std. Dev.
Seasonal PM Pattern -Southeast
EPA AIRS Database
(d)
20 40 60 80 100 120 140
PM10(|jg/m3)
1986 Mar May Jul Sep Nov
-A-PM10 -H-PM2.5 -I-PM Coarse
Figure 6-34. AIRS concentration data for the Southeast: (a) monitoring locations;
(b) regional PM10 concentration trends; (c) PM10 and PM2 5 relationship;
and (d) PM10, PM2 5, and PMCoarse seasonal pattern.
6-77
-------
standard deviation is by far the lowest among the aerosol regions (Figure 6-34b). This result
would appear to be associated with regional meteorological patterns.
The Southeast is also characterized by high seasonal amplitude of 37%, ranging between
22 //g/m3 in December through February and 35 //g/m3 in July through August (Figure 6-34d).
There is no evidence of a winter peak for the southeastern U.S.
The scattergram of PM25-PM10 for the Southeast (Figure 6-34c) shows an average of 58%
fine particle contribution, with considerable scatter. It should be noted, however, that size
segregated samples were available only briefly and these only for two monitoring sites which
may not be representative for the large southeastern region.
The regionally averaged daily PM10 concentrations over the Southeast (Figure 6-35) shows
a clearly discernible seasonality. The concentrations during the winter months are about factor
of two lower than during the summer. Overall, the lowest concentrations are about 12 //g/m3,
and the highest about 50 //g/m3, which is about factor of four. However, seasonality of the
temporal signal accounts for about half of the variation. Hence, within a given season the sixth
day to sixth day variation is only about 50%. It can be concluded that the PM10 concentration
over the southeastern United States region is quite uniform during shorter time intervals,
although it exhibits a substantial seasonality. The southeastern United States also exhibits the
highest spatial homogeneity (i.e., the smallest average deviations of average concentrations
between the stations). The PM10 urban excess (AIRS-IMPROVE) for the southeast region is
given in Figure 6-36. The range of monthly urban excess concentrations in the Southeast is
within approximately the same range, 5 //g/m3 to 10 //g/m3, as for the Northeast. The one
distinct feature is the sharp decrease in the urban excess in July which corresponds to the sharp
peak attributed to the nonurban coarse soil contribution in July for the Southeast (Figure 6-33).
6.4.3 Regional Aerosol Pattern in the Industrial Midwest
This aerosol region stretches between Illinois and western Pennsylvania, including
Kentucky on the south (Figure 6-37a). The industrial Midwest is covered by flat terrain west of
the Appalachian Mountains. In the winter the region is under the influence of cold Canadian air
masses, while during the summer moist air masses transported from the Gulf
6-78
-------
E
"5)
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
Southeast Every Sixth Day
J
1991
1992
1993
Figure 6-35. Short-term variation of PM10 average for the Southeast. Data are reported
every sixth day.
Southeast urban excess
40
Jan Mar May Jul Sep Nov
Figure 6-36. Urban excess concentration (AIRS minus IMPROVE) for the Southeast.
6-79
-------
PM2.5 Concentration - Industrial Midwest PM10
IMPROVE/NESCAUM Data
40,000
, PM2.5 and PMC - Industrial Midwest
IMPROVE/NESCAUM Data
35,000
30,000
I
o
o
15,000
(b)
1989 Mar May Jul Sep Nov
^PM10 +PM2.5 ^PM Coarse
Chemical Fine Mass Balance - Industrial Midwest Chemical Tracers - Industrial Midwest
IMPROVE/NESCAUM Data IMPROVE/NESCAUM Data
4,000 1 1 1 1 1 1 1 1 1-
a
£
o
tt
a
(C)
3,500
2,500
O
U
1,500
1,000
(d)
1989 Mar May Jul Sep Nov
T^Sulfate ^OC +Soil
-^EC ^Sulfate + OC + Soil + EC
1989 Mar May Jul Sep Nov
Sulfur-Max = 4000 Selenium - Max = 4
Vanadium-Max = 10 S/Se - Max = 4000
Figure 6-37. IMPROVE/NESCAUM concentration data for the industrial Midwest:
(a) monitoring locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate,
soil, organic carbon (OC), and elemental carbon (EC) fractions; and (d)
tracers.
6-80
-------
Coast prevail. However, the northern most portion of this region in Michigan and Wisconsin is
cooler and may be influenced by Canadian air flow at times during the summer. This region
includes the Ohio and Mississippi River Valleys that are known for high sulfur emission
densities. The region also includes major metropolitan areas.
6.4.3.1 Nonurban Size and Chemical Composition in the Industrial Midwest
The seasonal pattern of the nonurban aerosol in the Industrial Midwest is shown in
Figure 6-37b. Only five nonurban monitoring sites are available widely separately
geographically between those at the northern most sites and those in the southern portion of the
region with no sites over most of the region. Their representativeness is questionable. The PM10
concentrations range between 10 and 22 //g/m3, comparable to the nonurban levels in other
eastern U.S. regions. From 70 to 80% of PM10 is contributed by fine particles throughout the
year. The coarse particle concentrations are 4 to 5 //g/m3, which is lower than over any other
region of the U.S. Hence, the contribution of wind blown dust, fly ash, or other man-induced
dust entrainment is not a significant factor in the nonurban areas of the Industrial Midwest.
The chemical mass balance (Figure 6-37c) shows that sulfates are 45 to 55% of the fine
mass which is higher than the sulfate fractions in other regions. The concentration of vanadium,
which is a tracer for oil combustion, is low throughout the year. The concentration of fine
particle sulfur Organics exhibit a variable contribution that is high (40%) during the cold season
(October through February) and quite low (20%) in July and August. The strong winter peak for
the organic fraction differs markedly from the Northeast where the organics are seasonal.
Another unusual feature of the chemical mass balance is that the sum of sulfate, organic carbon,
soil, and elemental carbon is about 75% during the summer and 95% in the winter. It is not
known what is the composition of the missing 25% during the summer time, but the missing
fraction could be associated with nitrates, ammonium ion, hydrogen ion, and water.
Chemical tracer data are shown in Figure 6-3 7d. The chemical tracer for coal combustion,
selenium ranges between 1,000 and 1,500 pg/m3, which is higher than in any other region.
There is a sizeable month to month variation in Se concentration (partly due to a small number
of data points) and the seasonality is not appreciable. This means that the combined effects of
coal combustion source strength and meteorological dilution are seasonally invariant over the
industrial Midwest, exhibits random monthly variation but indicates a summer peak. The S/Se
6-81
-------
ratio is a rather smooth seasonal curve ranging between 1,000 in the winter and 2,000 during the
summer months. Hence, the sulfate yield is about twice as high during the summer as during
winter months. For comparison both the Northeast and Southeast exhibit somewhat higher
seasonality (factor of 2.5) in S/Se ratio. A possible explanation for this change in S/Se ratio is
that over the industrial Midwest the average age of the SO2 emissions traveling downwind may
be less than over the Northeast or Southeast.
6.4.3.2 Urban Aerosols in the Industrial Midwest
In the industrial midwester U.S. there was a decrease in the annual average PM10
concentrations between 1988 and 1994 from 33 //g/m3 to 29 //g/m3 for all sites and from
37 //g/m3 to 30 //g/m3 for trend sites (Figure 6-3 8b). The reductions were 12% for all sites and
19% for trend sites. There is also a 28% deviation among the stations within the region. As in
the Northeast, the higher concentrations occur within the larger urban-industrial areas
(Figure 6-38a). The PM10 seasonality (Figure 6-38d) is virtually identical (37% amplitude) to
the seasonality of the Southeast: the lowest concentrations (25 //g/m3) occur between November
and February, while the highest values are recorded in June through August (40 //g/m3). The
trends and the seasonality of the midwestern PM10 aerosols are comparable to those of the
Southeast.
Fine particles contribute 60% of the PM10 concentration on the average (Figure 6-3 8c), and
high PM10 can occur when either fine or coarse particles dominate.
Daily concentration over the industrial Midwest (Figure 6-39) varies between 14 and
75 //g/m3. The lowest regional concentrations occur during the winter months, while the highest
values (in excess of 40 //g/m3) occur during the summer. It is evident that seasonality is an
important component of the time series, accounting for about half of the variance. The elevated
concentrations occur only one sixth day observation at a time, consistent with the low frequency
of prolonged episodes. The industrial Midwest also shows substantial spatial variability. The
urban excess PM10 (AIRS-IMPROVE) for the industrial midwest is given in
6-82
-------
PM10 Average - Industrial Midwest
PM10 Cone. Trend - Industrial Midwest
EPA AIRS database
150
140
130
120
110
-,100
a
I 30
•» 80
CM
E
a. 70
60
50
40
30
20
10
PM2.5 vs. PM10 - Industrial Midwest
EPA AIRS - Monthly Averages
1989 1990 1991 1992 1993 1994
-A- Avg for all sites -B- Avg for trend sites
-+- Avg + Std. Dev. -9- Avg - Std. Dev.
Seasonal PM Pattern - Industrial Midwest
EPA AIRS Database
CORRELATION STATS:
Avg X : 29.02
AvgY: 17.62
• AvgY/AvgX: 0.6
Con Coeff: 0.86
Slope: 0.53
Y offset: 2.09
• Data Points : 465
(c)
40
=-30
0.
25
(d)
20 40 60 80 100 120 140
Mar May Jul Sep Nov
-A-PM10 -B-PM2.5 -i-PM Coarse
Figure 6-38. AIRS concentration data for the industrial Midwest: (a) monitoring
locations; (b) regional PM10 concentration trends; (c) PM10 and PM2 5
relationship; and (d) PM10, PM2 5, and PMCoarse seasonal pattern.
6-83
-------
Industrial Midwest Every Sixth Day
M.
80 -
75
70
65
60
55 ~
50
45
4 0 *
35 1 .
30 A'&J
25 -^4
20 *"g
15
10 ~
5
0
f
f-l
-"-!--
1
1-4,
* 1
*«.
(
f,
,^-,
tf]
i'"
*
If]
% ~ ::
^
J"
f
•*
1
J
-i- * ,{..
L - ' "^ f ' f
(-I' ,;. I - -- -'- 1 -
HI _,_ ^ ,;_ _ _.. J -;••
i' i°= ••'' ~\ ~-f i --. i -i '!•"* :- i
^' . 1 ™ .1? " ,"%. f c;J i
' ' fli^*"*! ' I '' ""*¥,!"- W1 1 1**"' '~'.-'W • r"
"•*• ~^ s '"V%,pl"*C^ "If *
1991
1992
1993
Figure 6-39. Short-term variation of PM10 average for the industrial Midwest. Data are
reported every sixth day.
Figure 6-40. The pattern for the urban excess PM10 differs seasonally from that in the northwest
(3-32) or southeast (6-34).
6.4.4 Regional Aerosol Pattern in the Upper Midwest
The upper Midwest covers the agricultural heartland of the country (Figure 6-41). The
region is void of any terrain features that would influence the regional ventilation. Industrial
emissions and the population density are comparatively low. However, the relatively high PM10
concentrations in this region warrant a more detailed examination. In the winter, the region is
covered by cold Canadian air masses, while in the summer moist Gulf air alternates and drier
Pacific air masses occur.
6.4.4.1 Nonurban Size and Chemical Composition in the Upper Midwest
There is a lack of nonurban monitoring sites in the upper midwest (Figure 6-41a).
Compared to the urban sites (Figure 6-42a), these nonurban sites are poorly representative of
6-84
-------
40
Industrial Midwest urban excess
Jan
Mar
Nov
May Jul Sep
Figure 6-40. Urban excess concentration (AIRS minus IMPROVE) for the industrial
Midwest.
the region. Based on these few sites in the upper Midwest, the PM10 concentration is about
8 //g/m3 during the November through April winter season, and increases to 15 //g/m3 during the
summer. Fine and coarse particles have a comparable contribution to the PM10 mass (Figure 6-
41b).
The chemical mass balance (Figure 6-4 Ic) indicates that during the March through May
spring season sulfates dominate, but during July through October season organics prevail. This
is a rather unusual pattern not observed over any other region. The contribution of fine particle
soil exceeds 10% in the spring as well as in the fall season.
Chemical tracers are shown in Figure 6-4Id. Selenium concentration is low throughout the
year (400 to 600 pg/m3), with the highest concentrations observed during the summer. This
suggests that either the Se sources from coal-fired power plants or the Se transport into the
Upper Midwest from other regions is stronger in the summer. The concentration of the fine
particle sulfur is <500 ng/m3 throughout the year, but somewhat higher during March and April.
The spring peak of fine particle sulfur has not been observed in any other region. It is also
worth noting that S/Se ratio is the highest during the spring and lowest in July
6-85
-------
PM2.5 Concentration - Upper Midwest
IMPROVE/NESCAUM Data
(a)
35,000
26,250
17,500
PM10, PM2.5 and PMC - Upper Midwest
IMPROVE/NESCAUM Data
40,000
30,000
25,000
O
O
15,000
10,000
(b)
1989 Mar May Jul Sep Nov
+PM2.5 T^PM Coarse
Chemical Fine Mass Balance - Upper Midwest Chemical Tracers - Upper Midwest
IMPROVE/NESCAUM Data IMPROVE/NESCAUM Data
4,0001
3,500
3,000
2,500
n
~ 2,000
O
o
1,500
(d)
1989 Mar May Jul Sep Nov
1989 Mar May Jul Sep Nov
Sulfur - Max = 4000 -a- Selenium -Max = 4
Sulfate + OC + Soil + EC -+- Vanadium - Max = 10 S/Se - Max = 4000
Figure 6-41. IMPROVE/NESCAUM concentration data for the upper Midwest:
(a) monitoring locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate,
soil, organic carbon (OC), and elemental carbon (EC) fractions; and (d)
tracers.
6-86
-------
PM10 Average - Upper Midwest
(a)
V
PM10 Cone. Trend - Upper Midwest
EPA AIRS database
1989 1990 1991 1992 1993 1994
-A- Avg for all sites -B- Avg for trend sites
-I- Avg + Std. Dev. -©- Avg - Std. Dev.
PM2.5 vs. PM10 - Upper Midwest
EPA AIRS - Monthly Averages
150
140
130
120
110
100
f 90
~ 80
in
ri
E 70
Q.
60
50
40
30
20
10
CORRELATION STATS
AvgX : 31.41
Avg Y : 12.18
Avg Y/Avg X : 0.38
CorrCooff: 0.54
Slope : 0.18
Y offset : 6.46
. Data Points : 34
(C)
20 40 60 80 100 120 140
PM10 (Lig/m3)
Seasonal PM Pattern - Upper Midwest
EPA AIRS Database
(d)
1986 Mar May Jul
-A- PM10 -B- PM2.5
Sep Nov
PM Coarse
Figure 6-42. Aerometric Information Retrieval System (AIRS) concentration data for the
upper Midwest: monitoring locations; regional PM10 monitoring trends;
PM10 and VM25 relationship; and PM10, PM2 5, and PMCoarse seasonal
trends.
6-87
-------
through September. It needs to be pointed out again that the above chemical patterns are based
on only two monitoring stations.
6.4.4.2 Urban Aerosols in the Upper Midwest
The agricultural upper Midwest (Figure 6-42b) shows the smallest decline in PM10
concentrations among the regions. In the upper midwestern U.S. there was a decrease in the
annual average PM10 concentration between 1988 and 1994 from 30 |ig/m3 to 25 |ig/m3 for all
sites and from 32 //g/m3 to 26 //g/m3 for trend sites (Figure 6-42b). The reductions were 17%
for all sites and 19% for trend sites. As over the eastern U.S., the highest concentrations occur
in the vicinity of urban areas. Some of the station-to-station concentration spread arises from
low concentrations over western North Dakota. On the average, the deviation among the
stations over the region is a moderate 30% (Figure 6-39). The upper Midwest is also unique in
that it shows the regionally lowest seasonal amplitude of 19%, with the slightly lower
concentrations occurring in December and January. The sparse size segregated data indicate that
only 38% of PM10 is contributed by fine particles. This is an indication that coarse wind blown
dust from natural or man-induced sources prevails. In this sense, the region is similar to the
Southwest (see below).
The daily regionally averaged PM10 concentrations in the upper Midwest (Figure 6-43)
range between 14 and 45 //g/m3. The highest values (>40 //g/m3) generally occur in the summer
season, while the low regional concentrations occur mainly in the cold season, but low values
also occur in the summer. It is interesting that the lowest PM10 concentrations over the upper
Midwest (15 //g/m3) are comparable to the Southeast and the industrial Midwest, but differ from
these regions by the absence of immediately subsequent high concentration events or episodes.
In fact, the PM10 "episodes" over the upper Midwest are all in the 40 to 45 //g/m3 concentration
range, compared to 50 to 75 //g/m3 in the Midwest. The seasonality is barely discernible from
the time series confirming that the day to day variation exceeds the seasonal modulation. The
urban excess PM10 (AIRS-IMPROVE) for the upper midwest is given in Figure 6-44, but its
reliability may be in question because of the very small number of nonurban sites.
-------
Upper Midwest Every Sixth Day
1991 1992 1993
Figure 6-43. Short-term variation of PM10 average for the Upper Midwest. Data
reported every sixth day.
are
40
35 --
30 --
25 --
20 --
4 »•—
15 --
10 --
5 --
Upper Midwest urban excess
Jan
Mar
May
Jul
Sep
Nov
Figure 6-44. Urban excess concentration (AIRS minus IMPROVE) for the Upper
Midwest.
6-89
-------
6.4.5 Regional Aerosol Pattern in the Southwest
The Southwest covers the arid states from western Texas to Arizona (Figure 6-45a). The
Southwest is characterized by mountainous terrain features between the southern Rockies and
the Colorado Plateau. The industrial activity and agriculture is minor compared to other regions.
Major population centers include El Paso, Phoenix, and Tucson. The meteorology of the region
is characterized by low annual precipitation, except during the periods when moist air penetrates
from the Gulf of Mexico toward these states, bringing moisture and precipitation.
6.4.5.1 Nonurban Size and Chemical Composition in the Southwest
The PM10 concentrations at nonurban southwestern sites show a double peak, one during
the late spring (April through July), and another in October. This bimodal seasonality is
imposed by the coarse particle mode. The PM2 5 mass concentration is unimodal with a summer
maximum. Overall, the nonurban PM10 concentrations are comparatively low (8 to 15 //g/m3)
and over 60% contributed by coarse particles (Figure 6-45b).
The chemical mass balance (Figure 6-45c) shows sulfates to be the larger contributor
during the winter (December through March) as well as in late summer (July through October).
However, sulfate and organic carbon contributions are comparable during March through June as
well as during November through December. Fine particle soil plays a prominent role in the
spring fine particle chemical mass balance reaching 25%, but the contribution of soil decreases
during the summer, and during December through February dwindles to below 10%.
The selenium and vanadium concentrations (Figure 6-45d) are very low and rather
invariant throughout the year. The fine particle sulfur concentration is low and exhibiting a
weak maximum during August. The S/Se ratio is comparatively low and bimodal, with peaks in
April through May as well as August through October.
6.4.5.2 Urban Aerosols in the Southwest
In the southwestern U.S. there was a decrease in PM10 concentrations between 1988 and
1994 from 38 //g/m3 to 24 //g/m3 for all sites and from 43 //g/m3 to 29 //g/m3 for trend sites
(Figure 6-46b). The reductions were 37% for all sites and 33% for trend sites. The
6-90
-------
PM2.5 Concentration -Southwest
IMPROVE/NESCAUM Data
O o
35,000
26,250
17,500
(a)
PM10, PM2.5 and PMC - Southwest
IMPROVE/NESCAUM Data
40.000 I •—
35,000
30,000
25,000
O
'n 20,000
10,000
5,000
(b)
Jan Mar May Jul Sep Nov
+ PM2.5 ^PM Coarse
Chemical Fine Mass Balance -Southwest
IMPROVE/NESCAUM Data
o
c
o
(c)
Jan Mar
^Sulfate
Chemical Tracers - Southwest
IMPROVE/NESCAUM Data
4,000
3,000
£ 2,500
O)
O
3= 2,000
O 1,500
O
1,000
May Jul Sep Nov
^Organics ^Soil
Org + Soil + Soot
(d)
Jan Mar May Jul Sep Nov
Sulfur-Max = 4000 Selenium - Max = 4
Vanadium - Max = 10 S/Se - Max = 4000
Figure 6-45. IMPROVE/NESCAUM concentration data for the Southwest:
(a) monitoring locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate,
soil, organic carbon (OC), and elemental carbon (EC) fractions; and (d)
tracers.
6-91
-------
PM10 Average - Southwest
PM10 Cone. Trend - Southwest
EPA AIRS database
1989 1990 1991 1992 1993 1994
-A-Avg for all sites -B-Avg for trend sites
-HAvg + Std. Dev. -©-Avg - Std. Dev.
140
130
120
110
1°°
10 80
CM
PM2.5 vs. PM10 - Southwest
EPA AIRS - Monthly Averages
CORRELATION STATS
Avg X : 37.58
Avg Y : 13.28
Avg Y/Avg X : 0.35
Corr Co0ff : 0.77
Slope : 0.15
Y offset : 7.40
Data Points : 107
(c)
Seasonal PM Pattern -Southwest
EPA AIRS Database
"E 35
- 30
£L
25
(d)
20
1986 Mar May Jul
-A-PM10 -B-PM2.5 -
Sep Nov
• PM Coarse
40 60 80 100 120 140
PM10(|jg/m3)
Figure 6-46. AIRS concentration data for the Southwest: (a) monitoring locations;
(b) regional PM10 monitoring trends; (c) PM10 and PM2 5 relationship; and
(d) PM10, PM2 5, and PMCoarse seasonal trends.
6-92
-------
downward trends in PM10 concentrations were not monotonic. In the Southwest is the large
concentration spread of 45% among the monitoring sites (Figure 6-46b). Sites with low
concentrations (<20 //g/m3) occur adjacent to high concentration sites (>50 //g/m3).
Seasonally, the Southwest PM10 concentration shows two peaks, one in late spring April
through June, and another during the fall October through November. The concentration dip in
August and September has not been observed for any other region. The late summer
concentration drop coincides with the occurrence of the moist air flows from the Gulf of
Mexico. The size segregated aerosol samples from the Southwest clearly show that coarse
particles make the major contribution to the PM10 concentration, the fine particles contributing
only 37% (Figure 6-46a). The scatter in Figure 6-46c indicates that high PM10 concentration
months can occur with low concentrations of fine particles. In the Southwest natural and man-
induced coarse particle dust is a major contributor to PM10 aerosols (Figure 6-45c).
The short term PM10 concentration over the Southwest (Figure 6-47) exhibits a highly
irregular pattern, that ranges between 11 to 52 //g/m3 regional average. Both the lowest (10 to
15 //g/m3) as well as the highest values are dispersed throughout the year.
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
Southwest Every Sixth Day
1991
1992
1993
Figure 6-47. Short-term variation of PM10 average for the Southwest. Data are reported
every sixth day.
6-93
-------
The urban excess PM10 (AIRS-IMPROVE) for the Southwest is given in Figure 6-48, and
the urban excess is substantially larger than in the regions discussed previously.
Southwest urban excess
Jan Mar May Jul Sep Nov
Figure 6-48. Urban excess concentration (AIRS minus IMPROVE) for the Southwest.
6.4.6 Regional Aerosol Pattern in the Northwest
The Northwest is defined to cover the bulk of the western United States north of the
Arizona border (Figure 6-49a). It is covered by mountainous terrain of the Rockies, as well as
the Sierra-Cascade mountain ranges. The Northwest is actually a collection of many aerosol
subregions. The meteorology is highly variable between the Pacific Northwest and the Rocky
Mountains with prevailing winds generally from the west. The main feature of the Northwest is
pronounced elevation ranges between mountain tops and valleys, and the resulting consequences
on emission pattern (confined to the valleys) and limited ventilation. The Northwest has also
industrial population centers, such as Seattle, Portland, Salt Lake City and Denver.
6-94
-------
PM2.5 Concentration - Northwest
IMPROVE/NESCAUM Data
Chemical Fine Mass Balance - Northwest
IMPROVE/NESCAUM Data
£ O.B
•5
I 0.5
(C)
1989 Mar
T^SU Ifate
May Jul Sep
Nov
Sulfate + OC + Soil + EC
PM10, PM2.5 and PMC - Northwest
IMPROVE/NESCAUM Data
30,000
"E
o,
25.000
20,000
O
O
15,000
(b)
1989 Mar May Jul Sep Nov
+PM2.5 T^PM Coarse
Chemical Tracers - Northwest
IMPROVE/NESCAUM Data
3,500
2,000
O
O
(d)
1989 Mar May Jul
Sulfur -Max = 4000
Vanadium - Max = 10
Sep Nov
Selenium - Max = 4
S/Se -Max = 4000
Figure 6-49. IMPROVE/NESCAUM concentration data for the Northwest:
(a) monitoring locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate,
soil, organic carbon (OC), and elemental carbon (EC) fractions; and (d)
tracers.
6-95
-------
6.4.6.1 Nonurban Size and Chemical Composition in the Northwest
The nonurban PM10 concentrations show low values ranging between 7 to 14 //g/m3 in the
northwestern U.S. The seasonality shows a peak in the summer which is contributed by both
fine and coarse particles. Coarse particles account for more than half of the PM10, particularly
during March through June spring season (Figure 6-49b).
The chemical mass balance (Figure 6-49c) shows roughly comparable contributions from
sulfates and organics, but their seasonality is phase shifted. Sulfates prevail during the spring
season while organics dominate during late fall (October through January). Fine particle soil
dust contributes 20% during April and May, but decline well below 10% during the winter
months (November through February). Overall, about 80% of the fine mass is accounted for by
the sulfates, organic carbon, soil, and elemental carbon.
Examining the carbonaceous particles and regional haze in the western and northwestern
U.S., White and Macias (1989) concluded that in the rural areas the concentrations of particulate
carbon are comparable to those of sulfate. Examining particulate nitrate, White and Macias
(1987) showed that the particulate nitrate concentration in the northern states (MT, ID, WY)
were substantially higher than sulfate concentrations. Aerosol particulate nitrates over rural
mountainous West were also episodic (i.e., few samples contributed a large fraction of the fine
particle integrated dosage).
Both selenium and vanadium concentrations (Figure 6-49d) are low in the Northwest, but
there is an indication of a summer peak of Se. The S/Se ratio is between 500 to 1,000, which is
the lowest among the regions. This ratio has both spring peak as well as fall peak, similar to the
pattern observed for the southwestern United States.
6.4.6.2 Urban Aerosols in the Northwest
In the northwestern U.S. there was a decrease in the annual average PM10 concentration
between 1988 and 1994 from 33 //g/m3 to 24 //g/m3 for all sites and from 35 //g/m3 to 27 //g/m3
for trend sites (Figure 6-50b). The reductions were 27% for all sites and 23% for trend sites.
However, the 1985 to 1994 reductions may be overestimates because of the low station density
in the early years. Once again, the average 1993 concentration is 25 //g/m3 which is comparable
to the 1993 concentrations of the eastern U.S. regions. The spread of
6-96
-------
PM10 Average - Northwest
:_ ,
'•, ,Jf
. •• >-}, " "-
(a) w- sW
PM10 Cone. Trend - Northwest
EPA AIRS database
1989 1990 1991 1992 1993 1994
Avg for all sites -B- Avg for trend sites
Avg + Std. Dev. -3- Avg - Std. Dev.
150
140
130
120
110
100
90
PM2.5 vs. PM10 - Northwest
EPA AIRS - Monthly Averages
"s
O> go
N 70
60
50
40
30
20
CORRELATION STATS
Avg X : 29.85
Avg Y : 17.29
Corr Co off : 0.9
Slope : 0.72
Y offset : -4.42
Data Points : 347
(c)
20 40 60 80 100 120 140
PM10 (|jg/m3)
Seasonal PM Pattern - Northwest
EPA AIRS Database
(d)
1986 Mar May Jul Sep Nov
-A- PM10 -B- PM2.S -I- PM Coarse
Figure 6-50. AIRS concentration data for the Northwest: (a) monitoring locations;
(b) regional PM10 monitoring; (c) PM10 and PM2 5 relationship; and (d) PM10,
PM2 5, and PMCoarse seasonal trend.
6-97
-------
concentration among the Northwest stations is large, with standard deviation of 45% (Figure 6-
50b). This spread in the concentration values is also evident from the various circle sizes of the
Northwest map (Figure 6-50a). The highest PM10 concentrations in the Northwest occur in more
remote mountainous valleys, rather than in the center of urban-industrial areas.
The seasonality of the northwestern United States has an amplitude of 36% which is
comparable to the strong seasonality of the eastern U.S. The peak PM10 concentrations occur in
the winter. The lowest PM10 concentration occurs during March through May and gradually
increases to a peak in December through January, falling sharply between January and March.
The limited PM2 5-PM10 data for the Northwest indicate that on the average 57% of PM10
particles are PM2 5. Figure 6-50c also indicates that the extreme PM10 concentrations are
contributed mainly by fine particles. Furthermore, the extreme PM10 concentrations also occur
in the winter season.
The daily concentration when averaged over the large and heterogeneous northwestern
region exhibits a remarkably small sixth day to sixth day variation (Figure 6-51). Furthermore,
there is clear seasonality with a strong winter peak. Within a given season, the regionally
averaged concentrations only vary by 20 to 40% from one sixth day to another. Examination of
the logarithmic standard deviation (Figure 6-50b) shows that the Northwest is spatially the most
heterogeneous and has the highest logarithmic standard deviation among all regions. Evidently,
in the Northwest high concentration PM10 pockets in topographically confined airsheds result in
strong spatial and temporal variations. However, large scale elevated PM10 concentrations that
cover the entire Northwestern region do not exist because high concentrations are not
"synchronized" among the different airsheds. In this sense, the Northwest differs markedly from
the eastern U.S., where large regional scale air masses with elevated PM10 determine the
regionally averaged values. The urban excess PM10 (AIRS-IMPROVE) for the Northwest is
given in Figure 6-52. The winter urban excesses are almost as large as in the Southwest
(Figure 6-48). However, if the region is a collection of aerosol subregions, the small number of
nonurban sites may not be representative of this collection of subregions.
6-98
-------
5
o_
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
Northwest Every Sixth Day
1991 1992 1993
Figure 6-51. Short-term variation of PM10 average for the Northwest. Data are reported
every sixth day.
Northwest urban excess
Jan Mar May Jul Sep Nov
Figure 6-52. Urban excess concentration (AIRS minus IMPROVE) for the Northwest.
6-99
-------
6.4.7 Regional Aerosol Pattern in Southern California
The region covers California south of San Francisco Bay (Figure 6-53a). It was considered
as a separate region primarily because of the known high aerosol concentrations in the Los
Angeles and San Joaquin basins. Meteorologically the region is exposed to the air flows from
the Pacific that provide the main regional ventilation toward the south and southeast. The
precipitation in the region occurs in the winter season, with the summer being hot and dry. The
regional ventilation of the San Joaquin Valley is severely restricted by the Sierra Nevada
Mountain range. Also, the San Gabriel Mountains constitute an air flow barrier east of the
Los Angeles basin. Both basins have high population, as well as industrial and agricultural
activities. Hence, human activities are believed to be the main aerosol sources of the region.
6.4.7.1 Nonurban Size and Chemical Composition in Southern California
The PM10 concentration at the few nonurban sites over southern California ranges between
10 //g/m3 during December through February, and 20 to 25 //g/m3 in April through October.
Coarse particles contribute more than 50% of the PM10 during the warm season May through
October. Both the fine and coarse aerosol fractions are lowest during the winter months
(December through March). The summer peak fine particle seasonality at nonurban southern
California sites is in marked contrast to the strongly fall peaked urban fine particle
concentrations (Figures 6-53b, 6-54d).
The chemical mass balance (Figure 6-53c) of nonurban southern California aerosol has a
substantial contribution by organics of 30 to 40% throughout the year. Sulfates account for only
10 to 15% of the fine mass in the winter, and about 20% in the summer months. The sulfate
fraction of the nonurban southern California fine mass is the lowest among the regions. Fine
particle soil dust is about 10% between April through November and drops to 5% during the
winter months. A notable feature of the southern California chemical mass balance is that 45%
of the winter, and 35% of the summer fine mass concentration is not accounted by sulfates, soils,
organic carbon,and elemental carbon. Nitrates are a major contributor to the southern California
aerosols (Solomon et al., 1989).
6-100
-------
PM2.5 Concentration -S. California
IMPROVE/NESCAUM Data
Chemical Fine Mass Balance - S. California
IMPROVE/NESCAUM Data
(C)
0.0
1989 Fab Mar Apr May Jun Jul Aug Sap Oct Nov Dae
-A-Sulfate -B-Organics —I—Soil
-©-Soot -0-Sulf + Org + Soil + Soot
PM10, PM2.5 and PMC - S. California
IMPROVE/NESCAUM Data
40,000
35,000
30.000
25,000
1
C
o
''S 20,000
15,000
5,000
(b)
1989 F0b Mar Apr May Jun Jul Aug Sap Oct Nov Dae
-Q-PM10 -I-PM2.5 -&-PM Coarse
Chemical Tracers - S. California
IMPROVE/NESCAUM Data
4,000
3,000
2,000
1,500
(d)
1989 Fab Mar Apr May Jun Jul Aug Sap Oct Nov Dae
-ft-Sultur -Max = 4000 -Q-Salanium - Max = 4
—I— Vanadium -Max = 10 -©-S/Se - Max = 4000
Figure 6-53. IMPROVE/NESCAUM concentration for Southern California:
(a) monitoring locations; (b) PM10, PM2 5, and PMCoarse (PMC); (c) sulfate,
soil, organic carbon (OC), and elemental carbon (EC) fractions; and
(d) tracers.
6-101
-------
Both selenium and vanadium (Figure 6-53d) show low values throughout the year without
significant seasonality. On the other hand the fine particle sulfur concentration shows a definite
summer peak at 500 ng/m3, compared to 200 ng/m3 during the winter. Consequently, the S/Se
ratio increase from 500 in the winter 1,000 to 1,500 in the summer.
6.4.7.2 Urban Aerosols in Southern California
In the southern California region there was a decrease in the annual average PM10
concentration between 1988 and 1994 from 41 //g/m3 to 30 //g/m3 for all sites and from
42 //g/m3 to 32 //g/m3 for trend sites (Figure 6-54b). The reductions were 27% for all sites and
241% for trend sites. There is a sizable concentration spread among the stations (40% standard
deviation). Inspection of the circle sizes in the map points (Figure 6-54a) to uniformly high
concentrations in the San Joaquin Valley as well as in the Los Angeles basin. The low
concentration sites are located either on the Pacific coast outside of the Los Angeles basin or in
the Sierra Nevada Mountains. Thus there are clear patterns of basin-wide elevated PM10
concentrations with lower values in more remote areas (Figure 6-54a).
The seasonality of the PM10 pattern in southern California is significant at 27%.
Furthermore, the seasonal pattern is unique that the highest concentrations occur in November
and the lowest in March. However, it is a see saw rather than a sinusoidal pattern.
On the average, about half of southern California PM10 is contributed by fine particles as
shown in the PM2 5-PM10 scattergram. Most of the high PM10 concentration months dominated
by fine particles tend to be in the fall.
The sixth day average time series for the southern California region (Figure 6-55) shows
remarkably high sixth daily variance, between 10 and 75 //g/m3. The lowest values tend to occur
between January and April, while the highest concentrations (>50 //g/m3) tend to occur during
October through December. Concentration excursions of a factor of two are common between
two consecutive six day time periods. However, visual inspection of the sixth daily signal also
reveals a substantial seasonality highest in the fall (September through December) and lowest in
the spring.
The urban excess PM10 (AIRS-IMPROVE) for Southern California is given in Figure 6-56.
The urban excesses are larger especially in winter, as are the urban excesses
6-102
-------
PM10 Average -Southern California
PM10 Cone. Trend - S. California
EPA AIRS database
PM2.5 vs. PM10 - S. California
EPA AIRS - Monthly Averages
1989 1990 1991 1992 1993 1994
-A-Avg for all sites -B-Avg for trend sites
-HAvg + Std. Dev. -©-Avg - Std. Dev.
Seasonal PM Pattern -S. California
EPA AIRS Database
140
130
120
110
100
fl~
£ 90
%
a.
— so
in
ni
S 70
O.
BO
50
40
30
20
CORRELATION STATS
Avg X : 54.1
Avg Y : 26.78
Avg Y/Avg X : 0.40
CorrCoaff: 0.87
Slope : 0.66
Y offset : -9
. Data Points : 209
(c)
I
(d)
20 40 60 80 100 120 140
PM10 (
Mar May Jul
-A-PM10 -B-PM2.5
Sep Nov
• PM Coarse
Figure 6-54. AIRS concentrations for Southern California: (a) monitoring locations;
(b) regional PM10 monitoring trends; (c) PM10 and PM2 5 relationship; and
(d) PM10, PM2 5, and PMCoarse seasonal trend.
6-103
-------
Southern California Every Sixth Day
1991
1992
1993
Figure 6-55. Short-term variation of PM10 average for Southern California. Data are
reported every sixth day.
Southern California urban excess
Jan
Mar
May
Jul
Sep
Nov
Figure 6-56. Urban excess concentration (AIRS minus IMPROVE) for Southern
California.
6-104
-------
in the Northwest. Again, these results depend on measurements from a small number of
nonurban sites.
6.5 SUBREGIONAL AEROSOL PATTERNS AND TRENDS
The health and other effects of aerosols are imposed on individuals, and the density of
population varies greatly in space. Consequently, the evaluation of effects requires the
knowledge of aerosol concentrations over specific locations where sensitive receptors reside.
The purpose of this section is to characterize the aerosol pattern at specific sites, small airsheds
or subregions. The discussions is organized by region and then by monitoring site within a
region. Most urban aerosol sampling is confined to PM10 or in some instances to PM2 5 and
PMCoarse . However, detailed chemical composition data are reviewed for several urban areas.
6.5.1 Subregional Aerosol Pattern in the Northeast
In the northeastern region, the Shenandoah National Park and Washington, DC constitute a
useful urban-nonurban set of size and chemically resolved aerosol data. New York City and
Philadelphia are also major metropolitan areas with substantial aerosol data bases. Whiteface
Mountain site distinguishes itself from its background by high elevation.
6.5.1.1 Shenandoah National Park
The PM10 concentration at the Shenandoah National Park IMPROVE site (Figure 6-57a)
exhibits a pronounced summer peak (27 //g/m3), which is a factor of three higher than the winter
value of 9 //g/m3. The strong seasonality is driven by the seasonal modulation of the fine mass
which accounts for 70 to 80% of the PM10 mass (Figure 6-57a). The coarse particle
concentration ranges between 3 and 6 //g/m3, which is small compared to the fine particle mass,
particularly in the summer season, when it accounts for < 25% of the PM10. It is clear that at this
nonurban site, in the vicinity of industrial source regions, fine particles determine the magnitude
ofPM10.
The chemical mass balance for the Shenandoah IMPROVE monitoring site (Figure 6-57b)
clearly documents the dominance of sulfate aerosols, which contribute about
6-105
-------
PM10, PM2.5, and PMC Monthly Average
Shenandoah NP
1989 Mar May Jill Sep Nov
^rPM Coarse
Chemical Fine Mass Balance
Shenandoah NP
0.9
0.8
ra
a, 0.6
c
I
I 0.5
c
0
E 0.4
o
0.2
X
x-
1989 Mar May Jul Sep Nov
T^T Sulfate ^Organics -F- Soil ^
~6" Sulfate + Organics + Soil + Soot
Chemical Tracers
Shenandoah NP
1989 Mar May Jul Sep Nov
^r Sulfur-Max = 4,000 -EH Selenium -Max =
-\- Vanadium-Max = 10 -&- S/Se-Max =4,000
Figure 6-57. IMPROVE/NESCAUM concentration for Shenandoah National Park: (a) PM10, PM2 5, and PMCoarse;
(b) chemical fraction of sulfate, soil, organic carbon (OC), and elemental carbon (EC); and (c) tracers.
-------
60% of the fine mass during April through September and about 50% during the winter months.
Organic carbon, on the other hand, range from 20% in summer to 30% in the winter months.
The contribution of fine particle soil and elemental carbon is well below 5%. Throughout the
year about 90% of the fine mass is accounted for by these measured substances. At the
Shenandoah site, the sulfate aerosols constitute a higher percentage of the chemical mass
balance, and lower percentages of organic carbon and elemental carbon than for the averaged
nonurban Northeastern sites (Figure 6-29).
Chemical tracer data are shown in Figure 6-57c. The concentration of coal-tracer selenium
shows two maximum, one during December through March, and another in June through
September. Vanadium is relatively constant throughout the year. The fine particle sulfur
concentration is almost a factor of five higher in August (3,300 ng/m3) than in December
(700 ng/m3). This extreme sulfur seasonality is stronger at the Shenandoah site relative to the
averages for sulfur seasonality at all nonurban Northeastern sites (Figure 6-29). The S/Se ratio
has a remarkably smooth but highly seasonal variation that varies by about factor of four
between the winter (700) and summer (2,600) values. If Se-bearing coal combustion is the
exclusive source of sulfur at the Shenandoah National Park , then the sulfate production from the
SO2 associated with coal-fired sources is 3 to 4 times higher in the summer than in the winter.
An examination of the nature and sources of haze in the Shenandoah Valley/Blue Ridge
Mountains area (Ferman et al., 1981) showed that sulfate aerosols were the most important
visibility reducing species. Averaging 55% of the fine particle mass, sulfates (and associated
water) accounted for 78% of the total light extinction. The second most abundant fine particles,
accounting for 29% of the fine mass, was organic carbon. The remaining particle mass and
extinction were due to crustal materials.
Using an in-situ rapid response measurement of H2SO4/(NH4)2SO4 aerosol in Shenandoah
National Park, VA, Weiss et al. (1982) found that the summer sulfate and ammonium ions
average 58% of particle mass smaller than 1 mm. The particle composition in terms of
NH4+/SO42" molar ratio ranged from 0.5 to 2.0 with strong diurnal variation. The particles were
most acidic at 1500 EDT and least acidic in the period 0600 to 0900 EDT. The water contained
in ambient aerosol particles was more strongly associated with sulfate and ammonium ions than
with the remainder of the fine particle mass.
6-107
-------
6.5.1.2 Washington, District of Columbia
The PM10 concentration at Washington DC (at the top of the National Park Service
Headquarters building) is virtually constant over the seasons at 25 to 30 //g/m3. Fine particles
contribute over 70% of PM10 throughout the year (Figure 6-58a). The weak seasonality in the
fine particle mass is in sharp contrast to the factor of three seasonal fine mass modulation at the
Shenandoah National Park. The coarse particle concentration in Washington, DC is 8 to
10 //g/m3 throughout the year, exhibiting virtually no seasonality.
PM2 5 at the urban Washington, DC site (figure 6-58b) is dominated by sulfates during the
summer months (over 50%), but declines to 30% in January. Organic carbon, on the other hand,
is 40% during October through January but only 30% during May through August. Thus, the
relative roles of organics and sulfates at the Washington, DC urban site is fully phase shifted by
half a year. Elemental carbon is a substantial contributes 9 to 12% during October through
December. Fine particle soil contributes a low 2 to 5% to PM2 5 at this urban site.
The chemical tracer species are shown in Figure 6-5 8c. The concentration of the coal
tracer selenium ranges between 1.5 to 2.0 pg/m3 without appreciable seasonality. The urban Se
in Washington, DC, is much higher than the Se at the northeastern nonurban sites. Vanadium,
the tracer for fuel oil, varies by factor of two between the high winter values (>8 pg/m3) and low
summer values (3 pg/m3). The pronounced V concentration seasonality is a clear indication of
that the emissions from fuel oil and other vanadium sources are seasonal. The fine particle sulfur
concentration varies by about factor of two between 1,400 ng/m3 winter concentration, and about
3,000 ng/m3 summer peak. The seasonal modulation of sulfur in Washington, DC is only factor
of two compared to the factor of four fine sulfur modulation at Shenandoah National Park. The
difference is primarily due to the elevated winter sulfur in Washington, DC. The S/Se ratio is
about 600 in the winter and about 1500 in the summer. It differs from Shenandoah by the lower
summer S/Se ratios. This result may be associated with differences in the air parcels involved in
long-range transport and transformation of SO2 to sulfate at Shenandoah compared to
Washington, DC.
6-108
-------
O
VO
60
55
50
45
40
! 35
01
u 30
u)
5
PM10, PM2.5, and PMC Monthly Average
Washington DC
20
15
10
5
1989 Mar May Jul Sep Nov
T^PM Coarse
a
5
0)
c
IL
0.9
0.8
0.7
0.6
0.5
0.3
0.2
0.1
0.0
Chemical Fine Mass Balance
Washington DC
*
A
NT
^
V-
1989 Mar May Jul Sep Nov
^rSulfate ^Organics ^Soil -c^
-fr Sulfate + Organics + Soil + Soot
Chemical Tracers
Washington, DC
1989 Mar May Jul Sep Nov
^r Sulfur - Max = 4,000 -Eh Selenium - Max = 4
-+ Vanadium-Max = 10 -& S/Se-Max =4,000
Figure 6-58. IMPROVE/NESCAUM concentration for Washington, DC: (a) PM10, PM25, and PMCoarse; (b) chemical fraction
of sulfate, soil, organic carbon (OC), and elemental carbon (EC); and (c) tracer concentrations.
-------
6.5.1.3 Comparison of Nonurban (Shenandoah) to Urban (Washington, District of
Columbia) Aerosols
The Washington, DC, urban site and the companion nonurban Shenandoah monitoring site
constitute a rare data pair that allows the quantification of urban-rural differences in fine and
coarse particle concentration, and chemical composition. Within Washington, DC, industrial
emissions are moderate compared to the industrial midwestern cities. However, both
automobile emission density and emissions from winter time heating are expected to be high. In
this section the excess aerosol concentrations in Washington, DC, over the Shenandoah site are
examined to elucidate the urban influence.
The Washington, DC, excess PM10 concentration (Figure 6-59a) ranges between
15-20 //g/m3 in the winter, and <3 //g/m3 in the summer. Hence, there is almost an order of
magnitude higher urban excess during the winter, compared to the summer. The seasonality of
the excess PM10 is driven by the winter peak excess fine particle concentration of 10-12 //g/m3.
The modest excess coarse particles is in the 3 to 6 //g/m3 range throughout the year. Thus, the
urban Washington, DC concentration exceeds its nonurban regional aerosol values during the
winter season, and the excess winter time urban aerosol is largely contributed by fine particle
mass. This indicates the smaller role of coarse particle fly ash, road dust resuspended by
automobiles or construction, road salt and all other sources of urban coarse particles in
Washington, DC, in winter.
The chemical composition of the excess fine particle concentration over the Shenandoah
nonurban background is also shown in Figure 6-59b. Fine organic carbon dominates the urban
excess ranging between 1 //g/m3 during the summer, and 5.5 //g/m3 during the winter. The
seasonality of excess organic carbon also drives the seasonality of excess fine mass. There is an
excess sulfate concentration of 1 to 2 //g/m3 in Washington, DC, except during July, August, and
September. In fact, in August in Washington, DC, sulfate concentration is about 0.3 //g/m3
below the Shenandoah values. The urban excess elemental carbon concentration is 1 to 2 //g/m3
throughout the year. The soil contribution to the fine particle mass is identical to the values of
the Shenandoah National Park, yielding virtually no excess fine soil contribution in the urban
area.
6-110
-------
PM10, PM2.5, and PMC Monthly Average
Washington DC -Shenandoah NP Difference
Chemicalpine Mass Balance
Washington DC -Shenandoah NP Difference
"E
O)
a.
u
u
n
5
20
18
161
14
12
10
8
6
i
2
0
19
I I I I I I I I I
(a)
?v
\ 1
\
vA /r
s \
'- \X ^A^
.A Y\ //A
-' V J4/T*
1 1 1 1 1 l^ 1 / 1 1 1 1
89 Mar May Jul Sep Nov
-B- PM10 -^ PM2.5 -*" PM Coarst
1989
Mar May Jul
Su Ifate -B- oc -
Sep Nov
Soil ^EC
^Sutate + OC + SoN+EC
Figure 6-59. Excess aerosol concentration at Washington, DC, compared to Shenandoah
National Park: (a) PM10, PM2 5, and PMCoarse (PMC); (b) concentration of
sulfate, soil, organic carbon (OC), and elemental carbon (EC).
The short-term fine mass concentration at Washington, DC and Shenandoah National Park
for the year 1992 is shown in Figure 6-60a. Although the sampling is conducted Wednesdays
and Saturdays for 24 h, the data points have been connected. The figure also compares the daily
(Wednesdays and Saturdays) fine particle sulfur concentrations at the two monitoring sites. The
fine mass concentration time series for Washington, DC, show elevated concentrations
(>30 //g/m3) that can occur throughout the year. On the other hand, high fine mass levels at
Shenandoah are recorded only during the summer season. Particulate sulfur concentrations at
the urban and nonurban site are often comparable during the summer (Figure 6-60b). This
indicates that particulate sulfur often is a large part of the regional air mass that at any given day
influences Washington, DC, and the Shenandoah National Park. Fine particle mass, on the other
hand, shows an excess concentration at Washington, DC, particularly during the winter months.
The fine mass daily time series clearly indicates that the concentration change from one daily
sample to another can be an order of magnitude different. Consequently, most of the
concentration variance is due to random synoptic air mass changes, and to a lesser degree due to
periodic seasonal variations.
6-111
-------
E
O)
c
en*
a
1992 Mar
T!S-Washington D.C.
-B- Shenandoah National Park
Nov
1992 Mar May Jul Sep Nov
-A- Washington D.C.
-£• Shenandoah National Park
Figure 6-60. Daily concentration of (a) fine mass and (b) fine sulfur at Washington, DC,
and Shenandoah National Park.
6.5.1.4 New York City, New York
The New York City metropolitan area is characterized by high population density,
moderate industrial activity, and relatively flat terrain. The PM10 concentration over the
metropolitan area is shown in Figure 6-6 la. The circles in the map show the locations of the
monitoring sites and the magnitude of each circle is proportional to the average PM10
concentration at that site using all available data. The observed average concentrations change by
about of factor of two to three from one location to another. Higher average concentrations tend
to occur near the center of the metropolitan area.
In the New York City metropolitan area there was a decrease in the annual PM10
concentration between 1988 and 1994 from 35 //g/m3 to 27 yUg/m3 for all sites and from
41 fj.g/m3 to 34 /ug/m3 for trend sites (Figure 6-6 Ib). The reductions were 23% for all sites and
17% for trend sites. There was unusually large difference between the two trends. The average
6-112
-------
(a)
\iglnf (25 C)
PM10 Cone. Trend - New York City Seasonal PM Pattern - New York City
EPA AIRS database EPA AIRS Database
£
"Si
45
PM10 Station Months : 1676
PM2.5 Station Months : 258
PMC Station Months : 258
(C)
1988 1989 1990 1991 1992 1993 1994 1986 Mar May Ju| Sep Nov
for all sites ^Avg for trend sites ^PM10 ^PM2.5 ^ PM Coarse
+ Std. Dev. ^Avg - Std. Dev.
Figure 6-61. New York City region: (a) aerosol concentration map, (b) trend, and
(c) seasonal pattern.
6-113
-------
seasonal pattern over the subregion (Figure 6-61c) is 25 to 30 //g/m3 throughout the year, but
rises to about 40 //g/m3 in July.
The seasonal pattern at three different individual monitoring sites in the New York City
metropolitan area is shown in Figure 6-62a. The three sites all show similar seasonality with a
summer peak, but with elevated concentrations closer to the city center.
Size segregated aerosol samples in New York City (Figure 6-62b,c) show that at both sites,
PM10 concentrations are contributed primarily by fine particles. Based on the discussion of the
more extensive Washington, DC (Section 6.5.1.2) measurements, it can be inferred that the
summer peak in the fine mass is mainly due to the regional formation of the fine aerosols, while
the winter peak is contributed by the local sources, confined to the inner metropolitan area.
As part of the New York Summer Aerosol study (Leaderer et al., 1978) continuous size
monitoring confirmed the expected bimodal volume distribution with modes between 0.1 to
1.0 (j,m and >3.0 //m. A number of interesting patterns were observed when the size distribution
was averaged by hour of day. The diurnal average total number concentration showed a pattern
which corresponded closely with the normalized diurnal traffic pattern. Particles <0.1 //m
showed the most marked diurnal variation, following the total number curve. Moreover,
particles in size ranges >0.1 //m showed little variation in the diurnal pattern. Analysis of
samples processed by the diffusion battery indicated that approximately 54%±18% of the sulfate
measured was in the suboptical range (approximately 0.04 |im to 0.3 jim) with the remainder
above 0.3 //m. Little sulfate mass was found in particles in the nuclei range (<0.04 //m).
Analysis of impactor samples for sulfates consistently showed that more than 85% of all water
soluble sulfates were <2.0 //m in size. Virtually no nitrate was present in the nuclei size range
while the suboptical size range accounted for approximately 30% of the total nitrate. 70% of the
total nitrate was found in the size range >0.3 //m. Analysis of large stages of Anderson impactor
showed that approximately 50% of particulate nitrate was greater than 5.5 //m in size.
Urban and rural particulate sulfur monitoring near New York City in the summer
(Leaderer et al., 1982) indicated that sulfate concentration distributions were regionally
homogeneous and increased with increasing ozone levels and covariant with several other
pollutant and meteorological parameters. Sulfate concentrations correlated strongly with
ammonium and strong acid at all sites. Strong acid concentrations were highest at the rural and
semi-rural sites,
6-114
-------
)
3.50
(a)
1985 Mar May Jul Sep Nov
-A- = PM10 AVG NEW YORK CITY
-B- = PM2.5 AVG NEW YORK CITY
-+- = PMC AVG NEW YORK CITY
(b)
1°985 Mar May Jul Sep Nov
-&- = PM10 AVG NEW YORK CITY
-B- = PM2.S AVG NEW YORK CITY
-+- = PMC AVG NEW YORK CITY
100
(c)
1985 Mar May Jul Sep Nov
T^= PM10 AVG NEW YORK CITY
-B-= PM2.5 AVG NEW YORK CITY
-H= PMC AVG NEW YORK CITY
Figure 6-62a,b,c. Fine, coarse, and PM10 particle concentrations at three New York City
sites.
6-115
-------
lowest at the urban sites, increased with increasing ozone levels and exhibited diurnal patterns
which matched the ozone diurnal patterns.
Size dependent, mass and composition of New York aerosol for low, medium, and high
visibility levels was reported by Patterson and Wagman (1977). At all levels of visibility,
bimodal or multimodal particle size distribution were observed for total mass and for individual
components. Decreased visibility corresponded to increased particle mass concentrations
especially in the fine particle fraction. Increases in the proportion of particulate sulfate and to a
lesser extent of nitrate, chloride, ammonium, and carbon were also associated with decreased
visibility.
Aerosol pattern analysis of a major wintertime (1983) pollution episode near New York
City in northern New Jersey (Lioy et al., 1985) revealed that the intensity of the episode was the
greatest in the area of the highest commercial, residential and industrial activities, and that the
atmospheric stagnation conditions resulted in the significant accumulation of aerosol mass. The
aerosol mass was primarily fine aerosols, and the extractable organic matter comprising about
50% of the particle mass.
6.5.1.5 Philadelphia, Pennsylvania
The metropolitan area of Philadelphia includes urban-industrial emissions over flat terrain.
Relatively uniform PM10 concentrations throughout the metropolitan area, with the exception of
a single site (AIRS #421010149) in the middle of the urban area (Figures 6-63 and 6-64).
The downward trends in PM10 concentrations between 1988 and 1994 were largely or
completely negated by the upward trends in 1993 and 1994 (Figure 6-63b). The decrease in
annual PM10 concentrations at trend sites between 1988 and 1994 for all sites was from 39 //g/m3
to 32 //g/m3, a decrease of 18%. The seasonal concentration of PM10 (Figure 6-63c) is about
30 to 35 //g/m3 throughout the year, except during the summer months when it rises above
40 Mg/m3.
The seasonal average PM10 concentrations for four sites near the center of Philadelphia is
shown in Figure 6-64. The high concentration site noted on the metropolitan map in
Figure 6-63a and two nearby sites in the industrial area long the riverfront are shown in
Figure 6-64a. The average PM10 concentration at that site ranges between 100 to 150 //g/m3
6-116
-------
(a)
PM10 Cone. Trend - Philadelphia
EPA AIRS database
1988 1989 1990 1991 1992 1993 1994
Avg for all sites Avg for trend sites
+ Std. Dev. ^Avg - Std. Dev.
jjg/m3 (25 C)
Seasonal PM Pattern - Philadephia
EPA AIRS Database
60
40
25
PM10 Station Months : 1263
PM2.5 Station Months : 59
PMC Station Months : 59
(C)
1986 Mar May Jul
Sep Nov
Coarse
Figure 6-63. Philadelphia region: (a) aerosol concentration map, (b) trend, and
(c) seasonal pattern.
6-117
-------
1985 Mar May Jul Sep Nov
PM10, Philadelphia, AIRS #42-101-xxxx
Sites, -A-=0149, -B-=0449, n-=0049
100
90
80
70
60
50
40
30
20
10
(b)
1985 Mar May Jul Sep Nov
Philadelphia, AIRS #42-101-0004
-&-=PM10, -B-=PM2.5, H-=PM Coarse
Figure 6-64a,b. Seasonal particle concentrations at four Philadelphia sites. (Note scale
for (a) is 150 u£/m3.)
which is a factor of 2 to 3 higher than the concentration at the neighboring sites. The daily
concentrations at these source monitoring sites correlate poorly with a nearby site (Figure 6-64b)
that is only 4 km away but outside the industrial area. This is an indication that the
concentrations at the industrial sites are under the influence of a strong local source of PM10. In
contrast, community sites in downtown and suburban Philadelphia that are as much as 30 km
apart show a strong correlation of daily measurements, indicating that a spatially uniform
regional aerosol influences the daily values in Philadelphia.
Size segregated aerosol samples (Figure 6-64b) show that fine particles contribute more
than coarse particles to the PM10 at this site. It is possible, however, that at other sampling sites,
e.g., the industrial sites (Figure 6-64a), coarse particles may prevail.
Outdoor summertime sulfate (SO4) concentrations were found to be uniform within
metropolitan Philadelphia (Suh et al., 1995). However, aerosol strong acidity (FT)
6-118
-------
concentrations were found to vary spatially. Also, the wintertime sulfate pattern was likely to be
more heterogeneous in space and time. This variation generally was independent of wind
direction, but was related to local factors, such as the NH3 concentration, population density, and
distance from the center of the city.
6.5.1.6 Whiteface Mountain, New York
The AIRS sampling location at the Whiteface Mountain in Upstate New York is a
high mountain top site elevated from the surrounding terrain. The monitoring site offers the
possibility of comparing mountain top concentrations to the surrounding lower elevation sites.
The seasonal pattern of PM10 concentration for Whiteface Mountain and the surrounding low
elevation sites, Saranac Lake and Saratoga Springs, is shown in Figure 6-65. The concentration
at the three sites is virtually identical during June through September. However, during the
winter the mountain top site at Whiteface has a PM10 concentration which is only one third of the
low elevation sites. This indicates that during the winter, the Whiteface mountain top is above
the surface-based aerosol layer, while during the summer the height of the well mixed aerosol
layer rises above the mountain top producing a reasonably uniform concentration at all sites.
6.5.2 Subregional Aerosol Pattern in the Southeast
6.5.2.1 Atlantic Coast States
The average yearly concentration in the southeast Atlantic coast states for all sites and
trend sites has decreased from 32 to 24 |ig/m3 and 25 |ig/m3 (Figure 6-66a,b). The reductions
were 25% and 22%. Seasonal concentrations show a summer peak largely due to PM2 5
(Figure 6-66c). Comparison of three AIRS PM10 monitoring sites in North Carolina's Piedmont,
Winston-Salem, Greensboro, and Raleigh-Durham (Figure 6-66d) shows virtually identical
concentrations (within 10%), both in absolute magnitudes and in the seasonality with summer
peaks in PM10. This is an indication that these sites in this subregion are exposed to essentially
the same air masses throughout the year. It also suggests that the excess PM10 concentrations
due to local urban sources probably are not signficant.
6-119
-------
g "
II »
I i
• F»IWI 1 O AVC5 WHITE FACE
= Pltfl 1 O AVQ SARANAG L
Figure 6-65. PM10 concentration seasonality at Whiteface Mountain and neighboring
low-elevation sites.
Size segregated monitoring data for Winston-Salem (Figure 6-66F) show that fine particles
contribute 70 to 80% of the PM10 mass of 25 to 35 //g/m3. Coarse particles are seasonally
invariant at about 10 //g/m3 which is typical for eastern U.S.
The PM10 concentration at monitoring sites in Florida (Orlando, Miami, Tampa) show
virtually identical concentrations ranging between 25 to 30 //g/m3 throughout the year, without
appreciable seasonality (Figure 6-66e).
6.5.2.2 Texas and Gulf States
The average yearly concentration between 1988 and 1994 in the Texas-Gulf states has
decreased for all sites and tend sites from 28 to 25 |ig/m3 (Figure 6-67b), a reduction of 11%.
Seasonal concentrations show a summer peak largely due to PM2.5 (Figure 6-68c). The
seasonal PM10 concentration at sites in Odessa, Amarillo, and Lubbock, TX, and in New
Orleans, LA, Mobile and Birmingham, AL show uniformity (20 to 40 //g/m3) with modest
seasonality
6-120
-------
Southeast Atlantic Coast States
PM10 Cone. Trend
EPA AIRS database
Seasonal PM Pattern
EPA AIRS Database
PNI.D Station Months: 5758
PM2.5 Station Months: 211
PMC Station Months: 210
(C)
I 1BBO 1999 1991 1992 1991 1994
^rAvgforallBlteB ^Avg for trend BlteB +Avg + std. Dev. ^Avg - std. Dev.
19GG Mar
May
Sep Nov
- PM Coarse
to
(e)
(I)
May Jul Sep
-*- - PM10 AVO WINSTON-SALEM
= PM10 AVG GREENSBORO
+ -PM10AVQ RALEIQH
May
Jul Sep
= PM10AVG ORLANDO
•PM10 AVO TAMPA
-PM10 AVO MIAMI
May Jul Sep
= PM10AVO WINSTON-SALEM
• PM2.5 AVG WINSTON-SALEM
+• PMC AVG WINSTON-SALEM
Figure 6-66. Aerosol concentration patterns for the Southeast Atlantic Coast states and sites in North Carolina and Florida:
(a) monitoring sites, (b) trends, (c) seasonal pattern, (d) North Carolina sites, (e) Florida sites, and (f) seasonal
pattern for Winston-Salem.
-------
PM10 Cone. Trend - S. Texas/Alabama
EPA AIRS database
60
55
50
45
40
35
30
25
20
15
10
5
(a)
Seasonal PM Pattern - Texas/Alabama
EPA AIRS Database
PM10 Station Months : 6774
PM2.5 Station Months: 185
PMC Station Months: 185
(c)
Mar May Jul Sep Nov
-A- PM10 -B- PM2.5 -+- PM Coarse
60
50
*£40
"oi
=•30
S
0- 20
10
1989 1990 1991
-Avg for all sites
Avg + Std. Dev.
1992 1993 1994
for trend sites
- Std. Dev.
(d)
ISM
Mar May Jul Sep Nov
-&-= PM10 AVG ODESSA
-B-= PM10 AVG AMARILLO
-+-= PM10 AVG LUBBOCK
50
40
«
£ 30
"S>
a 20
S
°- 10
1985 Mar May Jul Sep Nov
T^= PM10 AVG NEW ORLEANS
-B-= PM10 AVG MOBIL
-H= PM10AVG BIRMINGHAM
Figure 6-67a,b,c,d,e,f,g,h,i. Aerosol concentration patterns in Texas and Gulf states.
6-122
-------
100
90
BO
70
•3)
60
50
40
30
20
10
(f)
19B5 Mar May Jul Sep Nov
-A-= PM10 AVG HOUSTON
-B-= PM10 AVG AUSTIN
H-= PM10 AVG SAN ANTONIO
100
90
80
70
« 60
E
"3)
* 50
40
30
20
10
(h)
1985 Mar May Jul Sep Nov
-&-= PM10 AVG FORTWORTH
-B-= PM2.5 AVG FORTWORTH
-l-= PMC AVG FORTWORTH
100
90
80
70
60
50
40
30
20
10
(9)
1985 Mar May Jul Sep Nov
-&-= PM10 AVG CORPUS CHRISTI
~B~= PM2.5 AVG CORPUS CHRISTI
-l-= PMC AVG CORPUS CHRISTI
100
90
SO
70
60
50
40
30
20
10
1985 Mar May Jul Sep Nov
-&-= PM10 AVG NEW ORLEANS
~H-= PM2.5 AVG NEW ORLEANS
-*-= PMC AVG NEW ORLEANS
(i)
Figure 6-61 (cont'd). Aerosol concentration patterns in Texas and Gulf states.
6-123
-------
(Figure 6-67d,e). The sites in Houston, Austin, and San Antonio, TX show a wider range of
PM10 values with summer peaks (Figure 6-67f).
The size segregated aerosol samples collected in the cities of the Gulf states, Corpus
Christi, Forth Worth and New Orleans, LA (Figure 6-67g,h,i) all show fine particle
concentrations that are relatively low (10 to 20 //g/m3) compared to large eastern cities. Coarse
particle concentrations, on the other hand, can account for more than half of the PM10 mass. The
coarse particle contribution is most pronounced during the summer season.
In Houston, TX, Dzubay et al. (1982) found that in summertime fine particle mass
contained 58% sulfate and 18% of carbonaceous material. They also found that the coarse
fraction (2.5 to 15 //m) consisted of 69% crustal matter, 12% carbon, and 7% nitrate species.
6.5.2.3 Atlanta
Characterization of the Atlanta area aerosol (Marshall et al., 1986) show that elemental
carbon and particulate sulfur represent, respectively 3.1 to 9.9% and 1.9 to 9.4% of the total
suspended particulate mass. The concentrations of elemental carbon, sulfur, and TSP exhibit
strong seasonal variations, with elemental carbon decreasing from winter to summer, and sulfur
and TSP increasing. Elemental carbon appears to be statistically separate from sulfur, indicating
that the sources for elemental carbon and particulate sulfur are distinct.
6.5.2.4 Great Smoky Mountains
Size segregated fine and coarse aerosol concentrations were measured at the Great Smoky
Mountains National Park in September of 1980 (Stevens et al., 1980). Sulfate and its associated
ions contributed to 61% of the fine particle mass, followed by organics (10%) and elemental
carbon (5%).
6.5.3 Subregional Aerosol Pattern in the Industrial Midwest
Since the turn of the century, the major cities in the industrial midwestern states had air
pollution problems due to smoke and dust. Pittsburgh, St. Louis, Chicago, and Detroit were
among the formerly notorious air pollution hot spots. The recently acquired PM10 database now
allows the re-examination of these metropolitan areas in the industrial Midwest for their
concentration pattern in the 1990s.
6-124
-------
6.5.3.1 Pittsburgh, Pennsylvania
The average PM10 concentrations for sites in the extended metropolitan area is shown in
Figure 6-68. The Pittsburg, PA subregion includes the industrial cities, Steubenville, OH, and
Weirton, OH, located on the Ohio River. The average PM10 concentration at the 80 sites shown
on the map varies only by about 20% from site to site. Outstanding high concentration hot spots
are now absent. It is thus evident that during the 1985 to 1993 period, the average PM10
concentrations in the Pittsburgh subregion were spatially rather uniform.
In the Pittsburgh, PA metropolitan area there was a decrease in the annual average PM10
concentrations between 1988 and 1994 from 37 //g/m3 to 32 //g/m3 for all sites and from
41 //g/m3 to 33 //g/m3 for trend sites (Figure 6-68b). The reductions were 14% for all sites and
19% for trend sites. Figure 6-68b also marks the concentration standard deviation among the
monitoring sites for each year, which is about 15 to 20% and shrinking over time.
The seasonality of the PM10 pattern (Figure 6-68c) is dominated by a summer peak
(45 //g/m3), which is about 50% higher than the winter concentrations (30 //g/m3). The sites in
Pittsburgh, PA, Weirton, OH, and Steubenville, OH (Figure 6-69) show comparable seasonality
and values that are slightly above the subregional average. Hence, the particles at these formerly
highly polluted locations are now virtually identical to their subregional background.
Size segregated aerosol samples in Pittsburgh, PA and Steubenville, OH (Figure 6-69)
show that fine particles contribute 70 to 80% of the PM10 mass, and also dictate the summer-
peak seasonality of the PM10 concentrations. As in other urban monitoring sites in the eastern
U.S., the coarse particle concentration in Pittsburgh is about 10 //g/m3 and seasonally invariant.
The size segregated seasonal data for Steubenville, OH, exhibit more random fluctuations as
well as a discrepancy between the sum of fine and coarse on one hand, and PM10 on the other.
The discrepancy is attributed to the small number of size segregated aerosol samples.
Sulfate acidity measurements (Waldman et al., 1991) at Chestnut Ridge, PA (east of
Pittsburgh), suggest higher acidity occurred in the overnight period (0000-0800) in the late fall,
while sulfate had its highest levels in the morning to afternoon period.
6-125
-------
(a)
PM10 Cone. Trend - Pittsburgh
EPA AIRS database
1988 1989 1990 1991 1992 1993 1994
-&r Avg for all sites -B- Avg for trend sites
+ Std. Dev. -e- Avg - Std. Dev.
(25 C)
60
55
50
45
40
35
30
25
20
15
10
5
Seasonal PM Pattern - Pittsburgh
EPA AIRS Database
PM10 Station Months : 2937
PM2.5 Station Months : 159
PMC Station Months : 162
(C)
1986 Mar May Jul Sep Nov
^PM2.5 -+- PM Coarse
Figure 6-68. Pittsburgh subregion: (a) aerosol concentration map, (b) trends, and
(c) seasonal pattern.
6-126
-------
£
"3)
(a)
19B5 Mar May Jul Sep Nov
A = PM10 AVG PITTSBURGH
= PM10 AVG WEIRTON
- = PM10 AVG STEUBENVILLE
n 60
£
- 50
0.
40
(c)
1985 Mar May Jul Sep Nov
= PM10 AVG PITTSBURGH
= PM2.5 AVG PITTSBURGH
— = PMC AVG PITTSBURGH
(b)
1985 Mar May Jul Sep Nov
= PM10 AVG STEUBENVILLE
= PM2.5 AVG STEUBENVILLE
= PMC AVG STEUBENVILLE
(d)
1985 Mar May Jul Sep Nov
= PM10 AVG PITTSBURGH
= PM2.5 AVG PITTSBURGH
— = PMC AVG PITTSBURGH
Figure 6-69a,b,c,d. Fine, coarse, and PM10 concentration at sites in or near Pittsburgh.
6-127
-------
Pierson et al. (1980b, 1989) found no appreciable night/day difference in aerosol FT
(or NH4+ or SO42") - and almost no diurnal variation in O3 - at two elevated sites (Allegheny
Mountain and Laurel Hill, elevations 838 and 850 m) in southwest Pennsylvania. The contrast
with behavior at lower sites and particularly with the concurrent measurements at Deep Creek
Lake (Vossler et al., 1989) is attributable to isolation from ground-based processes at the
elevated sites at night.
The remarkable uniformity of fine particle mass and elemental composition from site to
site in the Ohio River Valley was also shown by Shaw and Paur (1983). Sulfur was the
predominant element in fine particles. Factor analysis of element concentrations indicated three
clusters throughout the year (1) coarse particle crustal elements (2) fine particle sulfur and
selenium (3) fine particle manganese, iron and zinc.
The chemical mass balance of Weirton-Steubenville aerosol was examined by Skidmore
et al. (1992). Primary aerosols from motor vehicles and secondary ammonium sulfate were
dominant contributors to the PM2 5 aerosol. Steel emissions were also significant contributors to
PM2 5. Wood burning and oil combustion were occasionally detected. Geological material was
the major contributor to the coarse aerosol fraction. Primary geological material, primary motor
vehicle exhaust, and secondary sulfate were the major contributors to PM10 at all five monitoring
sites.
The composition of size-fractionated summer aerosol in nearby Charleston, West Virginia
was reported by Lewis and Macias (1980). Ammonium sulfate was the largest single chemical
component (41%) of the fine aerosol mass. Carbon was also a large component of both fine and
coarse particle mass constituting 16% and 12% respectively. Factor analysis indicated that four
factors were sufficient to satisfactorily represent the variance of 26 measured parameters. The
factors were characteristic of crustal material, ammonium sulfate, automotive emissions, and
unidentified anthropogenic sources.
6.5.3.2 St. Louis, Missouri
Historically, the St. Louis metropolitan area has been known for high particulate
concentrations. The map of the metropolitan area (Figure 6-70a) shows about factor of 2 to
3 concentration differences among the PM10 monitoring stations.
6-128
-------
(a)
ug/ms (25 C)
PM10 Cone. Trend - St. Louis
EPA AIRS database
Seasonal PM Pattern - Pittsburgh
EPA AIRS Database
1988 1989 1990 1991 1992 1993 1994
-&- Avg for all sites -H- Avg for trend sites
-+- Avg + Std. Dev. -Q- Avg -Std. Dev.
60
50
PM10 Station Months : 2937
PM2.5 Station Months : 159
PMC Station Months : 162
(C)
1986 Mar May Jul Sep Nov
^PM2.5 ^PM Coarse
Figure 6-70. St. Louis subregion: (a) aerosol concentration map, (b) trends, and
(c) seasonal pattern.
6-129
-------
In the St. Louis metropolitan area there was a decrease in the annual average PM10
concentration between 1988 and 1994 from 37 //g/m3 to 30 //g/m3 for all sites and from
40 //g/m3 to 31 //g/m3 for trend sites (Figure 6-70b). The reductions were 23% for all sites and
22% for trend sites. This decline is comparable to the average reductions over the industrial
midwestern region. The seasonality of the sub-regionally averaged concentrations
(Figure 6-70c) shows the summer peak with 40 to 50 //g/m3 which is about 50% higher than the
winter averages.
Seasonal comparison of the individual monitoring sites in the area shows that Granite City,
IL and East St. Louis, IL have higher PM10 concentrations throughout the year compared to
western St. Louis, MO sites.
Size segregated aerosol samples at three sites west of the Mississippi River (Ferguson, MO,
Affton, MO, and Clayton, MO) show that fine particles are mostly responsible for PM10,
including the seasonality (Figure 6-71). Coarse particles contribute 10 //g/m3 or less throughout
the year, although corresponding size segregated aerosol data for the more polluted east side of
the Mississippi River are not available.
Monitoring the diurnal and seasonal patterns of particulate sulfur and sulfuric acid in
St. Louis, Cobourn and Husar (1982) noted an afternoon increase in particulate sulfur
concentration of about 20%. For the summertime, particulate sulfur concentration was higher
than the annual mean by 40%.
Measurements were made using dichotomous samplers of PM10 (expressed as PM20), PM25
and S at urban, suburban, semi-rural, and rural sites in and around St. Louis, MO, during 1975 to
1976 as part of the Regional Air Pollution Study (RAPS) (Altshuller, 1982). Unlike the
nonurban sites compared from the IMPROVE/NESCAUM network with urban sites from AIRS,
these rural sites were within 45 km of the center of St. Louis. The comparisons are between
three urban sites (103, 105, 106) and three rural sites (118, 112, 124).
The PM2 5 constituted 45 to 60% of the PM10 with the percentages at rural sites ranging
from 0 to 10% higher than at urban sites. The ratios of the concentrations of PM2 5 at urban sites
to PM25 at rural sites ranged from 1.4 to 1.5 for the six quarters between the third quarter of
1975 to the fourth quarter of 1976. The ratios of the concentrations of PMCoarse at urban sites
to PMCoarse at rural sites ranged from 1.5 to 1.8 for the same six quarters. For fine S, the ratios
6-130
-------
60
O)
50
40
(a)
100
1985 Mar May Jul Sep Nov
^ =PM10AVGST LOUIS
= PM10 AVG EAST ST LOUIS
100
80
« 60
•••.
O)
3.
- 50
20
(C)
1985 Mar May Jul Sep Nov
^ = PM10 AVG CLAYTON
~^~ = PM2.5 AVG CLAYTON
70
60
40
20
10
(b)
1985 Mar May Jul Sep Nov
A =PM10 AVG FERGUSON
~®~ = PM2.5 AVG FERGUSON
~*~ = PMC AVG FERGUSON
100
80
50
20
10
(d)
1985 Mar May Jul Sep Nov
^ = PM10 AVG AFFTON
~®~ = PM2.5 AVG AFFTON
~^~ = PMC AVG CLAYTON ~^~ = PMC AVG AFFTON
Figure 6-71a,b,c,d. Fine, coarse, and PM10 seasonal concentration patterns in or near
St. Louis.
6-131
-------
between urban and rural sites ranged from 1.1 to 1.2, while for coarse S, the ratios between
urban and rural sites ranged from 1.7 to 2.6 for the same six quarters.
These results indicate a very strong regional influence on fine S with a lesser regional
influence on PM25. The ratios of PMCoarse and coarse S indicate stronger local influences on
their concentrations than on fine S and PM2 5. The percentage of fine S expressed as (NH4)2 SO4
to the PM2 5 was consistently higher at rural sites than at urban sites in and around St. Louis
(Altshuller, 1982). In the third quarters of 1975 and 1976, these percentages averaged 70% at
rural sites and 55% at urban sites, while in the fourth quarters of 1975 and 1976, these
percentages averaged 45% at rural sites and 35% at urban sites.
As observed near New York City (Leaderer et al., 1982), the fine S in the St. Louis area
was regionally homogenous and, during episodic periods, the fine S concentrations followed the
variations in O3 concentrations reasonably closely (Altshuller, 1985). A linear relationship was
obtained for fine S and O3 flows into St. Louis. The fine S with increasing fine S concentration
constituted an increasingly large percentage of the PM25 at an urban site (Altshuller, 1985).
6.5.3.3 Chicago, Illinois
Historically, Chicago has been known for industrial dust, smoke, and haze, as in adjacent
East Chicago and Gary, IN. The average PM10 concentrations over the Chicago subregion
(Figure 6-72a) vary by a factor of two or less throughout the subregion. In the Chicago
subregion, there was a decrease in the annual average PM10 concentrations between 1988 and
1994 from 32 //g/m3 to 29 //g/m3 for all sites and from 39 //g/m3 to 31 //g/m3 for trend sites
(Figure 72b). The reductions were 9% for all sites and 20% for trend sites. The seasonality of
PM10 is also typical with the summer peak of 40 //g/m3 and winter values of 20 to 30 //g/m3.
Superposition of seasonal PM10 data at Chicago, IL, East Chicago, IL, and Gary, IN,
demonstrates significant spatial uniformity, as well as indicating in more recent years
comparatively low PM10 concentrations in this area that has historically been a smoky and dusty
industrial subregion.
In the Chicago subregion there was a decrease in the annual average PM10 concentration
between 1985 and 1994 from 40 //g/m3 to 29 //g/m3 for all sites and from 40 //g/m3 to 31 //g/m3
for trend sites (Figure 6-72b). The reductions were 28% for all sites and 23% for trend sites.
6-132
-------
PM10 Cone. Trend - Chicago
EPA AIRS database
(a)
Hg/m3(25C)
1°988 1989 1990 1991 1992 1993 1994
-A- Avg for all sites -a- Avg for trend sites
Seasonal PM Pattern - Chicago
EPA AIRS Database
Avg + Std. Dev.
Avg - Std. Dev.
PM10 Station Months : 3245
PM2.5 Station Months : 0
PMC Station Months : 0
(C)
1°986 Mar May Jul Sep Nov
-fr- PM10 -B- PM2.S -+- PM Coarse
150
140
130
120
110
100
90
fl
£ .0
Dl
=• 70
E~
Q_ BO
50
40
30
20
10
(d)
1°985 Mar May Jul Sep
= PM10 AVG CHICAGO
Nov
= PM10 AVG EAST CHICAGO
-H= PM10 AVG GARY
Figure 6-72. Chicago subregion: (a) aerosol concentration map, (b) trends, (c) and
(d) seasonal patterns.
6-133
-------
Chemical composition measurements in Chicago (Lee et al., 1993) showed that mean
concentrations for SO42' (5.55 //g/m3), NH4+ (2.74 //g/m3), NH3 (1.63 //g/m3), HNO3
(0.81 Mg/m3), HNO2 (0.99 //g/m3), for SO2 (21.2 //g/m3), NO3' (4.21 //g/m3), and H+
(7.7 nmol/m3). The highest values occurred in the summer, except for HNO2 and NO3" which
had the highest values in the winter.
Comparison of atmospheric coarse particles at an urban and nonurban site near Chicago,
IL, show that the concentration were 50% higher during mid-day than at night. Dry ground
samples were 30 % higher than wet ground and 90% higher than frozen ground samples. (Noll
etal., 1985).
The analysis of coarse particles in Chicago, IL (Noll et al., 1990) show that the coarse
particle mass could be divided into two categories: material that was primarily of crustal origin
(Al, Ca, Fe, and Si) and material that was primarily of anthropogenic origin (Cd, Cu, Mn, Ni,
Pb,
and Zn). The mass of crustal material varied between 14 and 24% of the total coarse mass. The
mass of Cd, Cu, Mn, Ni, Pb, and Zn totaled less than 1%.
The composition of atmospheric coarse particles at urban (Chicago, IL) and nonurban
(Argonne, IL) were reported by Noll et al. (1987). Limestone and silicates were the main source
of material at the non urban site. Anthropogenic sources, represented by flyash and coal, were
present in the industrial sector sample and rubber tire was present in the commercial sector
sample.
6.5.3.4 Detroit, Michigan
In Detroit, in July, 1981 (Wolff and Korsog, 1985) the average fine mass was found to be
42.4 //g/m3. The chemical composition of the fine particles (Wolff et al., 1982) was 52%
sulfates, 27% organic carbon, 4% elemental carbon, 8% soil dust. Nitrate was found to be
absent from fine mass. Fine particles themselves contributed about 64% of the aerosol mass.
The sulfate associated with coal combustion contributed to 50% of the fine particles. The coarse
fraction, which averaged as 25.8 //g/m3, was dominated by crustal material which accounted for
about two-thirds of the coarse material. Significant contributions were also identified from
motor vehicles (mostly due to re-entrained road dust) and iron and steel industry emissions.
6-134
-------
The seasonal variations in nitric acid, nitrate, strong aerosol acidity, and ammonia in
Warren, MI, was examined by Cadle (1985). The greatest variations was for ammonia, which
was 8.5 times higher in summer than winter. The least variation was for particulate nitrate
which had a summer maximum only 1.8 times higher than in spring minimum. It was noted that
ammonium nitrate volatilization from filters and impactors can cause large errors in summertime
measurements, but the errors are not significant during the winter.
The influence of local and regional sources on the concentration of parti culate matter in
urban and rural sites near Detroit, MI was investigated by Wolff et al. (1985). Analysis of
spatial variations of the various particulate components revealed: (1) at all four sites the PM2 5
was dominated by regional influences rather than local sources. The site in industrial sector had
the largest impact of local sources, but even at his site the local influences appears to be smaller
than
the regional ones. (2) The regional influences were most pronounced on the sulfate levels which
accounted for 40 to 50% of the PM25. (3) Organic carbon compounds were the second most
abundant PM2 5 species accounting for 20 to 40% of the mass. Organic carbon seems to be
controlled by both local and regional organic carbon influences. Vehicular emissions and
possibly secondary reactions appear to affect the organic carbon concentrations. (4) Elemental
carbon appears to be dominated by local emission. (5) PMCoarse was dominated by local
sources, but at the industrial site unknown non-crustal elements were significant components of
coarse mass.
6.5.5 Subregional Aerosol Pattern in the Southwest
The arid southwestern U.S. includes metropolitan areas (El Paso, TX, Phoenix-Tucson,
AZ) with modest industry and national parks (Grand Canyon) where the prevention of visibility
degradation has been stated as a national goal. The southwest is a dusty region and much of the
discussion below pertains to coarse particles and soil dust.
6.5.5.1 El Paso, Texas
The PM10 concentration in the El Paso, TX, subregion shows that the high and low
concentration sites occur near each other (Figure 6-73a). This is an indication that local sources
of PM10 with limited range of impact are important. In the El Paso, TX, subregion
6-135
-------
PM10 Cone. Trend - El Paso
EPA AIRS database
1989 1990 1991 1992 1993 1994
Avg for all sites Avg for trend sites
+ Std. Dev. ^Avg - Std. Dev.
(a)
(25 C)
Seasonal PM Pattern - El Paso
EPA AIRS Database
55
50
35
30
15
PM10 Station Months : 1108
PM2.5 Station Months: 32
PMC Station Months: 32
(C)
1986 Mar May Jul Sep Nov
-A- PM10 -B- PM2.5 -+- PM Coarse
Figure 6-73. El Paso subregion: (a) aerosol concentration map, (b) trends, and
(c) seasonal pattern.
6-136
-------
there is a decrease in the annual average PM10 concentration between 1988 and 1994 from
46 //g/m3 to 25 //g/m3 for all sites and from 57 //g/m3 to 34 //g/m3 for trend sites (Figure 6-73b).
The reductions were 46% for all sites and 40% for trend sites. This substantial reduction
exceeds the PM10 decline over the entire southwestern region (Figure 6-46b).
The seasonality of PM10 over the El Paso, TX subregion (Figure 6-73c) is bimodal with
peaks in the spring time, March through July, as well as another stronger peak, October through
November. This double peak seasonality at El Paso, TX, also parallels the seasonality of the
entire region. The concentration reduction in August which coincides with the arrival of moist
flow from the Gulf of Mexico into states in the southwest (Figure 6-46d). Size segregated
aerosol samples for El Paso, TX (AIRS #481410037) show that coarse particles dominate the
PM10 concentrations, accounting for about 70% of the PM10 mass (Figure 6-74a). This is
consistent with the important role of coarse particles over the arid Southwest. In comparison,
size segregated data for San Antonio, TX (Figure 6-74b) located closer to the Gulf Coast in
Texas, show that fine and coarse mass have comparable contributions, similar to Houston, TX.
6.5.5.2 Phoenix and Tucson, Arizona
The Phoenix-Tucson subregion (Figure 6-75a) shows a substantial PM10 concentration
range. Samplers within the Phoenix or Tucson area indicate 2 to 3 times higher concentrations
than the more remote sites, particularly the ones in the mountains. For the Phoenix-Tucson
subregion there was a decrease in the annual average PM10 concentration between 1988 and 1993
from 39 //g/m3 to 28 //g/m3 for all sites and from 49 //g/m3 to 32 //g/m3 for trend sites
(Figure 6-75b). The reductions were 28% for all sites and 35% for trend sites. The decrease in
PM concentration were not monotonic. The average PM10 seasonality of the Phoenix-Tucson
subregion (Figure 6-75c) shows the bimodal spring and fall peak pattern which is characteristic
for the entire Southwest region.
During the Phoenix Urban Haze Pilot Study during the winter 1988 to 1989 (Frazier, 1989)
a definite diurnal cycle in PM2 5 concentrations was observed. The maximum, generally but not
6-137
-------
100
90
70
60
50
40
30
20
10
(a)
1985 Mar May Jul Sep Nov
A = PM10 AVG EL PASO
~ = PM2.5 AVG EL PASO
-I- = PMC AVG EL PASO
100
90
80
70
60
50
40
20
10
(b)
1985 Mar May Jil Sep Nov
A - PM10 AVG SAN ANTON. 0
-B-. PM2.5 AVG SAN ANTON. O
-H- PMC AVG SAN ANTON O
Figure 6-74a,b. Fine, coarse, and PM10 concentration patterns in El Paso and San
Antonio.
always, occurred at night, which is consistent with the meteorological observations of poor
dispersion and dilution.
The wintertime aerosol chemical pattern in Phoenix was reported by Chow et al. (1990)
and Solomon and Moyers (1986). These investigators found fine particle crustal species,
sulfates, nitrates, and organic and elemental carbon to be at least five times higher in
concentration when comparing samples during a period of limited visibility to samples taken
during good visibility.
A chemical characterization of wintertime fine particles in Phoenix, AZ (Solomon and
Moyers, 1986) showed a dominance of organic carbon and nitrate aerosols. The composition in
Phoenix is most like that of Denver, CO, a city which also experiences wintertime inversions
6-138
-------
C)
PM10 Cone. Trend - Phoenix/Tucson
EPA AIRS database
(a)
Seasonal PM Pattern - Phoenix/Tucsoi
EPA AIRS Database
60
55
50
35
30
25
20
15
10
PM 10 Station Months: 1630
PM2.5 Station Months : 0
PMC Station Months : 0
(C)
1986 Mar May Jul
-&- PM10 -H- PM2.5
Sep Nov
PM Coarse
1988 1989 1990 1991 1992 1993 1994
~&~ Avg for all sites ~H- Avg for trend sites
~+~ Avg +• Std. Dev. ^ Avg - Std. Dev.
Figure 6-75. Phoenix-Tucson subregion: (a) aerosol concentration map, (b) trends, and
(c) seasonal pattern.
6-139
-------
(Pierson and Russell, 1979; Countess et al., 1980; Groblicki et al., 1981). In both cities, the
average measured NO3" concentrations were about 1 to 2 times that of the average SCT4
concentration. In addition, the average SO4 concentration measured in Phoenix was much lower
than those observed at other locations throughout the U.S., but similar to the regional values
observed in the Southwest (Moyers, 1982).
Wintertime PM10 and PM2 5 chemical compositions and source contributions in Tucson, AZ
(Chow et al., 1992a) show that the major contributors to the highest PM10 concentrations were
geological material (>50%) and primary motor vehicle exhaust (> 30%) at three urban sampling
sites. Secondary ammonium sulfate, secondary ammonium nitrate, and copper smelter aerosols
were found to contribute less than 5% to elevated PM10 concentrations.
The OC/EC ratio was one to one at Phoenix sites. The average arsenic concentrations in
Phoenix was four times higher than observed in other cities, which indicates the potential
influence of Arizona smelters located within 100 miles of Phoenix. Average sulfate levels in
Phoenix were higher than they were in Denver, which has less local emissions of SO2.
6.5.5.3 Grand Canyon National Park
McMurry and Zhang (1989) reported the size distribution of ambient organic and
elemental carbon near the Grand Canyon and in the Los Angeles basin. Virtually all of the
carbon was found in the submicron range, some below 0.1 //m. However, positive sampling
artifacts for sub 0.1//m organics were considered significant.
At the Grand Canyon National Park, Zhang et al. (1994) showed that sulfates and
carbonaceous particles were the major contributor to PM2 5 particle scattering during the three
winter months and that their contributions were comparable. Scattering by nitrates and soil dust
was typically a factor of five to ten smaller. The low pressure impactor measurements also
showed that sulfur size distributions vary considerably (0.07 to 0.66 //m).
6.5.6 Subregional Aerosol Pattern in the Northwest
The mountainous northwestern United States has many aerosol regions with different
characteristics. The discussion below will examine South Lake Tahoe, as a case study for
mountain-valley difference, Salt Lake City, UT, Denver, CO, Idaho-Montana sites, and several
Washington-Oregon sites.
6-140
-------
6.5.6.1 South Lake Tahoe
South Lake Tahoe IMPROVE monitoring site is located in a in a populated area on the
south shore of Lake Tahoe. The Bliss State Park IMPROVE monitoring site is to the northwest,
elevated (700ft) and removed from the populated areas. The pair of sites illustrates the
populated area-remote difference in aerosol pattern. The aerosol and visibility at the two lake
Tahoe sites were also examined (Molenar et al., 1994).
The concentration of all aerosol components is substantially higher on the south lake shore
compared to the more remote site. The seasonality and chemical composition is also
substantially different. The excess PM10 concentration at the S. Lake Tahoe site compared to
Bliss State Park (Figure 6-76) is about 5 //g/m3 during the warm season, May through
September, and it climbs to 28 //g/m3 excess in January. The factor of five seasonal modulation
for valley excess PM10 is likely contributed by winter time emission sources, poor dispersion
compared to the summer, as well as fog, all of which tend to enhance the aerosol formation.
Fine and coarse particles contribute roughly equally to excess PM10 mass concentrations.
However, fine particles contribute about 60% during the fall season and coarse particles prevail
(>60%) during the spring. Both fine and coarse particles show a winter peak concentration.
The chemical composition of the valley excess fine particle mass concentration also shows
a strong seasonality for organic carbon and elemental carbon. In fact, the excess organic carbon
concentration in the winter (13 //g/m3) is almost an order of magnitude higher than the summer
values. The seasonal concentration of excess elemental carbon is similar to that of the organic
carbon. However, the relative magnitude of organic carbon compared to elemental carbon is
higher in the winter (factor of five) than in the summer (factor of two). The concentration of
fine particle sulfate is virtually identical for South Lake Tahoe and Bliss State Park. This
implies that the South Lake Tahoe aerosol sources do not contain sulfur. It is also worth noting
that the excess fine particle soil at South Lake Tahoe is below 1 Mg/m3, which is a small fraction
of the coarse mass. Thus, the crustal component of the South Lake
6-141
-------
to
PM10, PM2.5, and PMC Monthly Average
S. Lake Tahoe - Bliss State Pk. Difference
Nov
Chemical Fine Mass Balance
S. Lake Tahoe - Bliss State Pk. Diffferenc
PM10
PM Coarse
Sool
Fine Mas;
Figure 6-76. Excess aerosol concentration (a) and composition (b) at South Lake Tahoe compared to Bliss State Park.
Mar May Jul
Sulfate -B-Organics
Sulfate + Organics + Soil + Soot
Sep
Soil
Nov
-------
Tahoe aerosol contributes to the coarse mass but not appreciably to the fine mass concentration.
In summary, there is a significant excess PM10 aerosol concentration at S. Lake Tahoe
compared to the adjacent Bliss State Park remote site, particularly during the winter season
(28 //g/m3). The excess mass is about equally distributed between fine and coarse particles. The
fine mass is largely composed of organics.
6.5.6.2 Salt Lake City, Utah, Subregion
Salt Lake City, Ogden, and Provo, UT, are part of an airshed that is confined by tall
mountains to the East, limiting the dispersion by westerly winds.
The seasonal average PM10 concentration at three AIRS sites in Salt Lake City, Ogden, and
Provo, UT, is shown in Figure 6-77b. All three sites show virtually identical seasonality, having
peak concentrations during December through January. This confirms that the three sites belong
to the same airshed with similar source pattern, meteorological dispersion and chemical
transformation and removal processes.
During the 1988 to 1994 period there were overall decreases in the annual average PM10
for the Salt Lake City, UT subregion from 49 //g/m3 to 29 //g/m3 for all sites and from 54 //g/m3
to 30 //g/m3 for trend sites (Figure 6-77b). The reductions were 41% for all sites and 48% for
trend sites. The trends were not monotonic, but showed substantial shifts upwards and
downwards during the 1988 to 1994 period.
The size segregated fine and coarse concentration data exhibit a dynamic seasonal pattern.
Fine particles clearly dominate the high winter concentrations reaching 40 to 50 //g/m3,
compared to summer concentrations of 10 //g/m3. This magnitude of fine mass concentration is
among the highest recorded in the AIRS data system. Coarse particles are less seasonal and
they are more important during the dry summer season. The formation of sulfate and nitrate
during winter inversion fogs near Salt Lake City, UT were studied by Mangelson et al. (1994).
6-143
-------
PM10 Cone. Trend -Salt Lake City
EPA AIRS database
Seasonal PM Pattern - Salt Lake City
EPA AIRS Database
Uly Jul S.p
* • PM10 AVO SALT LAKE CITY
= PM10AVOOODEN
+ -PH10AVOPROVO
1988 1989 1980 1991 1992 1993 19ft
^Avp, for all sites ^Avgfortrend sites H-Avg + Std. Dev. -«-Avg-Std.Dev.
1°986
I Jul S.p
-PM10AVO NOTINACITY
• PM2.5 AVG NOTINA CITY
• PMCAVO NOTINACITY
PM10Station Months: 91
(c)
Mar May Jul
•*- PM1D -fr PM2.5
Sep
- PM Coarse
(f)
May Jul S>p
- PII10 AVO SALT LAKE CITY
= PII2.5 AV6 SALT LAKE CITY
• PMC AVO SALT LAKE CITY
Figure 6-77. Salt Lake City region: (a) aerosol concentration map, (b) trends, (c) seasonal pattern, and (d,e,f) seasonal
patterns at sites in or near Salt Lake City.
-------
6.5.6.3 Denver, Colorado
The Denver brown cloud is a manifestation of high wintertime concentration of particles
and gases. Several recent studies have focused on the characterization of the Denver brown
cloud aerosols.
Size distribution measurements of winter Denver aerosol (Countess et al., 1981) show that
on high pollution days that the mass median aerodynamic diameter of the accumulation mode
aerosol was about 0.31 //m with og±2.0. Wolff et al. (1981) found that on the average motor
vehicles were responsible for 27% of the elemental carbon while wood burning was responsible
for 39% of the elemental carbon.
The chemical composition of wintertime Denver fine aerosol mass (16.4 //g/m3) (Sloane et
al., 1991) shows the dominance of total carbon consisting of organic carbon (8.1 //g/m3) and
elemental carbon (2.6 //g/m3) over sulfate (1.2 //g/m3) and nitrate (3.4 //g/m3). The fine particle
size distribution of sulfate and nitrates were bimodal.
6.5.6.4 Northern Idaho-Western Montana Subregion
The mountainous northern Idaho and western Montana subregion is characterized by deep
valleys and the absence of major industrial sources or large urban-metropolitan areas.
Nevertheless, PM10 monitoring sites in northern Idaho and western Montana report
concentrations that are among the highest in the nation, as illustrated in Figure 6-78a, while
neaby sites are among the lowest. The large spatial concentration variability is evidently related
to the rugged terrain. Most of the monitoring sites are located in the flat valleys.
In the northern Idaho-western Montana subregion there was a decrease in the annual
average PM10 concentrations between 1988 and 1993 from 41 //g/m3 to 30 //g/m3 for all sites and
from 40 //g/m3 to 31 //g/m3 for trend sites. The reductions were 27% for all sites and 23% for
trend sites (Figure 6-78b). The average seasonality of the subregion is strongly winter peaked
(Figure 6-78c) with a factor of two modulation between 25 and 45 //g/m3.
The high spatial variability is illustrated in an example from northern Idaho (Figure 6-79a).
Three sites in Missoula, MT, show winter monthly averaged peak concentrations from less than
40 to more than 100 //g/m3. This is higher than the monthly average PM10 concentration
anywhere in the eastern U.S. The site closest to the city center
6-145
-------
(a)
C)
PM10 Cone. Trend - N. Idaho/NW Montana Seasonal PM Pattern - Idaho/Montana
EPA AIRS database
EPA AIRS Database
55
50
45
40
35
25
20
10
PM10 Station Months : 1985
PM2.5 Station Months :0
PMC Station Months : 0
(C)
1988 1989 1990 1991 1992 1993 1994
Avg for all sites Avg for trend sites
+ Std. Dev. ^Avg - Std. Dev.
1986 Mar May Jul
-&- PM10 -B- PM2.5
Sep Nov
PM Coarse
Figure 6-78. Northern Idaho-Northwestern Montana subregion: (a) aerosol
concentration map, (b) trends, and (c) seasonal pattern.
6-146
-------
£
'5)
1°985 Mar May Jul Sep Nov
= PM10 AVG MISSOULA
= PM10 AVG MISSOULA
= PM10 AVG MISSOULA
100
(b)
fees
(c)
Mar May Jul Sep Nov
^= PM10 AVG BOISE CITY
-B-
= PM10AVGSALMON
= PM10 AVG IDAHO FALLS
-B-
1985 Mar May Jul Sep Nov
= PM10 AVG ANACONDA- DEER LODGE COUNTY
= PM10 AVG ANACONDA DEER LODGE COUNTY
= PM10 AVG ANACONDA- DEER LODGE COUNTY
Figure 6-79a,b,c. PM10 concentration patterns at sites in the Northern Idaho-
Northwestern Montana subregion.
6-147
-------
shows the highest winter peak (>100 //g/m3), but has summer values that are comparable to the
other two sites. It is evident that in Missoula, MT, high concentration gradients exist between
the populated areas and remote sites. Boise and Salmon, ID (Figure 6-79b) also show elevated
PM10 concentrations during the cold season. Idaho Falls, ID, on the other hand, is seasonally
uniform at about 30 //g/m3, which is comparable to the lowest Missoula, MT, site.
Unusually low PM10 concentrations of 10 //g/m3 are reported at three PM10 monitoring sites
near Anaconda-Deer, ID (Figure 6-79c). This result is unexpected because the sites are in a
valley. The characteristic winter peak is completely absent. This suggests that pristine, low,
PM10 sites can exist in the northwestern valleys, and hence the region is not uniformly covered
by wintertime haze or smoke.
6.5.6.5 Washington-Oregon Subregion
The Pacific Northwest is also a mountainous subregion that exhibits unique aerosol
characteristics. During 1988 to 1994, there were decreases in the annual average PM10
concentrations for the Washington-Oregon subregion from 36 |ig/m3 to 26 |ig/m3 for all sites and
from 39 |ig/m3 to 28 |ig/m3 for trend sites. The reductions were 28% for both all sites and trend
sites. The subregion shows a strong seasonality with a winter peak due to PM2 5 (Figure 6-80b).
PM10 monitoring sites in Seattle, Bellevue, and Tacoma, WA (Figure 6-80d), show relatively
low concentrations and a lower seasonality although higher values occur in the winter. A much
more pronounced seasonality of PM10 concentrations is recorded in southern Oregon. Medford,
Grants Pass, and Klamath Falls, OR (Figure 6-80e) evidently belong to an airshed in which
emissions, dispersion, and aerosol formation mechanisms are conducive to the formation of
winter time aerosol (60 to 80 //g/m3).
Fine and coarse particle data collected over a limited period in 1987 show that the winter
peak of PM10 is entirely due to the strong winter peak of fine particle mass (50 to 100 //g/m3).
Coarse mass, on the other hand, is seasonally invariant at about 10 to 20 //g/m3. Fine particles
clearly are responsible for the winter peak. This is somewhat different from the observations at
South Lake Tahoe, where the winter peak was attributed to both fine and coarse particles.
6-148
-------
(a)
Seasonal PM Pattern -Washington/Oregon
EPA AIRS Database
M 35
£
D)
=L 30
PM10 Station Month! : 5142
PM2.5 Station Months : 99
PMC Station Months : 97
(c)
1986 Mar May Jul
ay Jul S»p Nov
rPM10 -B-PM2.5 -l-pM Colr«.
PM10 Cone. Trend - Washington/Oregon
EPA AIRS databasa
1989 1990 1991 1992 1993 1994
^Avg for all Bites "&Avg for trend sites
f-Avg * Std. Dav. •©"Avo - std. Dev.
E
D)
(d)
19BS Mar
May Jul Sep
= PM10 AVG SEATTLE
~ = PM10 AVG BELLEVUE
~ = PM10 AVG TACOMA
Figure 6-80a,b,c,d,e,f,g,h. Aerosol concentration patterns in Washington State and
Oregon.
6-149
-------
(e)
1985 Mar May Jul Sep Nov
A = PM10 AVG MEDFORD
~B- = PM10 AVG GRANTS PASS
~+~ = PM10 AVG KLAMATH FALLS
- 50
E
L
(g)
1985 Mar May Jul Sep Nov
~^~ = PM10 AVG BEND
~H- = PM2.5 AVG BEND
~+~ = PMC AVG BEND
(f)
1985 Mar May Jul Sep Nov
~&~ = PM10 AVG MEDFORD
~B- = PM2.5 AVG MEDFORD
~+~ = PMC AVG MEDFORD
(h)
1985 Mar May Jul Sap Nov
^^ = PM10 AVG CENTRAL POINT
~B~ = PM2.5 AVG CENTRAL POINT
~+~ = PMC AVG CENTRAL POINT
Figure 6-80 (cont'd). Aerosol concentration patterns in Washington State and Oregon.
6-150
-------
The size segregated aerosol data for Bend and Central Point, OR (Figure 6-80g,h), show
diminishing concentrations compared to Medford (Figure 6-80f), where the reduction of PM10 is
mainly due to the decrease of the fine particle mass during the winter season.
In Portland, OR, carbonaceous aerosol was found to account for about 50% of fine aerosol
mass (Shah et al., 1984).
6.5.6.6 Other Northwestern Locations
Dresser (1988) investigated the winter PM10 concentrations in a small ski resort town,
Telluride, CO, and found that the street dirt and sand are major contributors, particularly during
the dry post snow period. Wintertime source apportionment attributed to 45% of the PM10 mass
to residential wood combustion in San Jose, CA (Chow et al., 1995a).
6.5.7 Subregional Aerosol Pattern in Southern California
The southern California region has two subregions, the San Joaquin Valley and the
Los Angeles-South Coast Air Basin, discussed separately in sections below.
6.5.7.1 San Joaquin Basin
The wide air basin between the coastal mountain ranges of California to the west and the
Sierra Nevada Mountains to the east shows reasonably uniform PM10 concentrations as indicated
on the map (Figure 6-8la). There is evidence of PM10 concentration reduction but the trend is
not conclusive (Figure 6-8 Ib). The seasonal modulation amplitude over the San Joaquin Valley
(Figure 6-8 Ic) is about factor of 2.5 between the low spring concentration 30 to 35 //g/m3, and
high fall concentration (60 to 70 //g/m3). The unique feature of this seasonality is the fall peak
which differs from the summer peak in the eastern United States and winter peak over the
mountainous northwestern states.
The AIRS database contains valuable size segregated fine and coarse particle concentration
data within the San Joaquin Valley, as shown in Figure 6-82 for Fresno, Madera, Visalia, and
Bakersfield, CA. These monitoring sites show virtually identical concentration patterns for fine
and coarse mass. Both coarse and fine particles are important contributors to the San Joaquin
Valley PM10 aerosol. However, their respective prevalence is phase shifted. Fine particles are
6-151
-------
(a)
100
90
80
70
I!60
c 5
g I 50
u «
o
o a
1- O
E 40
30
20
10
PM 10 Cone. Trend - San Joaquin Valley
EPA AIRS database
(b)
Seasonal PM Pattern - San Joaquin Valley
EPA AIRS Database
100
90
PM10 Station Months : 1335
. PM2.5 Station Months : 123
PMC Station Months : 123
1989 1990 1991 1992 1993 1994
f Avg for all sites "Q" Avg for trend sites
- Avg + Std. Dev. -O- Avg - Std. Dev.
1986 Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
-A- PM 10 -B- PM2.5 -H PM Coarse
Figure 6-81. San Joaquin Valley: aerosol concentration map, trends, and seasonal
pattern.
6-152
-------
90
«0
E
D)
3.
5
Q.
SO
30
20
10
(a)
1985 Mar May Jul Sep Nov
~^= PM10AVG FRESNO
-B- - PM2.5 AVG FRESNO
•+- = PMC AVG FRESNO
100
10
1985 Mar May Jul Sep Nov
-£- » PM10 AVG VISALIA
-B- « PM2.S AVG VISALIA
~~*-- PMC AVG VISALIA
1985 Mar May Jul Sep Nov
A = PM10 AVG MADERA
-B-- PM2.5AVG MADERA
-+-= PMC AVG MADERA
100
1985 Mar May Jul Sep Nov
-A- = PM10 AVG BAKERSFIELD
-B- = PM2.5 AVG BAKERSFIELD
-+- = PMC AVG BAKERSFIELD
Figure 6-82. Fine, coarse, and PM10 seasonal patterns in the San Joaquin Valley.
6-153
-------
most important during the November through February winter season, while coarse particles
prevail during April through September. In November, both coarse and fine particles contribute
to the seasonal peak of PM10. During March through May, neither fine or coarse particles are
abundant and the PM10 concentration is lowest during the spring season.
The temporal dynamics of the emissions, ventilation and aerosol formation in the San
Joaquin Valley has been the subject of detailed aerosol monitoring, and source apportionment
studies.
The aerosol composition at nonurban sites (Chow et al., 1995b) provides further
characteristics of the central California aerosol pattern (Figure 6-82). A PM10 aerosol study was
carried out at six sites in California's San Joaquin Valley from 14 June 1988 to 9 June 1989, as
part of the 1988 to 1989 Valley Air Quality Study (VAQS). Concentrations of PM10 and PM2 5
mass, organic and elemental carbon, nitrate, sulfate, ammonium, and elements were determined
in 24-h aerosol samples collected at three urban (Stockton, Fresno, Bakersfield) and three
non-urban (Crows Landing, Fellows, Kern Wildlife Refuge) locations (Chow et al., 1993a). The
VAQS data indicate the federal 24-h PM10 standard of 150 //g/m3 was exceeded at four out of
the six sites and for reasons which differ by season and by spatial region of influence. The
annual average source contributions to PM10 at Bakersfield, the site with the highest annual
average, were 54% from primary geological material, 15% from secondary ammonium nitrate,
10 % from primary motor vehicle exhaust, 8% from primary construction, the remaining 4% is
unexplained. The results of the source apportionment at all sites show that geological
contributions dominate in summer and fall months, while secondary ammonium nitrate
contributions derived from direct emissions of ammonia and oxides of nitrogen from agricultural
activities and engine exhaust are largest during winter months. (Chow et al., 1992b).
6.5.7.2 Los Angeles-South Coast Air Basin-Southeastern Desert Air Basin
The Los Angeles basin is confined by the San Gabriel Mountains which limit the
ventilation during westerly winds. Intensive emissions from automotive and industrial sources
produce the Los Angeles smog with numerous secondary photochemical reaction products from
primary emissions. The map of the Los Angeles subregion shows (Figure 6-83a) the
6-154
-------
1988 1989 1990 1991 1992 1993
1985 Mar May Jul Sep Nov
Figure 6-83. Los Angeles: (a) aerosol concentration map, (b) trends, and (c) seasonal
pattern.
magnitude of PM10 concentrations for individual monitoring stations. Isopleths of PM10
concentration for 1992 are consistent with these results showing the highest PM10 concentrations
are measured in the center of the LA basin with the lower concentration of PM,n near the ocean
10
and out in the desert and the mountains (Hoggan et al., 1993).
There has been a substantial reduction of subregion average PM10 concentration from 1988
to 1993 from 54 //g/m3 down to 38 //g/m3 (Figure 6-83b), a reduction of 30%. The seasonality
of the basin averaged PM10 concentration shows a 50% amplitude, with the peak concentration
(60 //g/m3) during October and the lowest values (40 //g/m3) during January through March
(Figure 6-83c). Hence, this fall peaked seasonality is similar to the fall peak over the San
Joaquin Valley.
6-155
-------
The PM10 air quality in the California South Coast Air Basin (CSCAB) and to a limited
extent in the Southeastern Desert Air Basin have been analyzed for the 1985 to 1992 period
(Hoggan et al., 1993). Although a larger number of monitoring stations exist in more recent
years, the analysis involved only the monitoring stations with complete data in Long Beach,
Burbank, El Toro, Ontario, Rubidoux, Banning, and Indio. Measurements in downtown Los
Angeles also are used in parts of the analysis (Hoggan et al., 1993). The annual average PM10
trend line for 1985 to 1992 showed a statistical significant trend downwards with the decrease
averaging 3% per year. The sulfate and nitrate also were measured and they accounted for about
one-third of the decrease in PM10. The decreases between 1989 and 1993 for this set of stations
were smaller than for the larger group of stations (Figure 6-83b). There was a statistically
significant decrease (0.05 level of significance) at Burbank, Long Beach, Rubidoux, and
Banning. Use of both a decision tree analysis and a multiple linear regression analysis showed
that the temperature at 850 mb, a measure of mass stability, was an important variable associated
with PM10 in the CSCAB. Use of this variable suggests that the observed decreases in annual
average PM10 concentrations between 1987 and 1992 are not an artifact of meteorology. A more
detailed discussion of these analyses as related to various aspects of meteorology is given
(Hoggan et al., 1993).
The diurnal patterns of PM10 also are discussed (Hoggan et al., 1993). The Rubidoux
monitoring station showed peaks in PM10 at about the time of peak commuter traffic. The Los
Angeles monitoring station showed higher PM10 concentrations in the morning and evening than
at midday. Azusa and Long Beach monitoring stations showed broad daytime peaks. The Indio
monitoring station showed an evening peak.
The weekday to weekend mean PM10 concentrations at all monitoring stations showed
significantly lower concentrations on weekends (Hoggan et al., 1993). At the two SEDAB
stations, Indio and Banning, Saturday PM10 concentrations were slightly lower than weekdays,
but Sunday PM10 concentrations fell within the range of weekday means.
Some seasonal characteristics of the Los Angeles basin are depicted in Figure 6-84. The
monitoring sites at different parts of the basin have markedly different seasonal concentration
patterns. Hawthorne and Long Beach near the Pacific Coast and Burbank in an inland valley
have the higher PM10 concentration in late fall and early winter (Figure 6-84b,c). On the
6-156
-------
O)
150
140
130
120
110
100
90
80
70
BO
50
40
30
20
10
« 60
"3)
- 50
O.
40
(a)
19B5 Mar May Jul Sep Nov
~*~ = PM10 AVG HAWTHORNE
~^~ = PM10 AVG RUBIDOUX
+ = PM10 AVG BURBANK
(c)
1965 Mar May Jul Sap
^^ = PM10 AVG AZUSA
PM2.5 AVG AZUSA
PMC AVG AZUSA
Nov
-B-
80
1985
(b)
1985 Mar May Jul Sep Nov
^ = PM10 AVG LONG BEACH
~^~ = PM2.5 AVG LONG BEACH
+ = PMC AVG LONG BEACH
(d)
Mar M ay
-&-,
-B-.
Jul Sep
PH10 AVG RUBIDOUX
PH2.5 AVG RUBIDOUX
PMC AVG RUBIDOUX
Nov
Figure 6-84a,b,c,d. Fine, coarse, and PM10 seasonal patterns near Los Angeles. (Note
scale for (a) is 150 ug/m3.)
6-157
-------
other hand, Azuza and Rubidoux in the eastern part of the basin exhibits the higher PM10
concentration during the May to October 'smog season' (Hoggan et al., 1993) (Figure 6-84b,d,e).
The main causes of different seasonalities are likely to be associated with seasonally varying
meteorological, transport, and chemical transformation patterns. The role of coarse and fine
particles in the Los Angeles basin is also illustrated in Figure 6-84. At Long Beach, near the
coast (adjacent to Hawthorne), the fine particles dominate the PM10 during the November
through February winter season (40 to 50 //g/m3). Coarse particles at Long Beach are constant
throughout the year at about (20 //g/m3). At Azuza and Rubidoux fine and coarse particles
contribute roughly equally to the high PM10 concentrations. Thus, the PM10 aerosols over the
smoggiest parts of the Los Angeles basin are not dominated by fine secondary aerosols but
contributed by both fine and coarse particles.
The Rubidoux site in 1985 to 1988 showed violations of the 24-h PM10 standard
approximately 12% of the time with a large contribution from ammonium nitrate (Chow et al.,
1992c). A large group of dairies and animal husbandry operations in the Chino area
approximately 13 km west of the Rubidoux site were identified as major ammonia emitters
(Russell and Cass, 1986). To better evaluate the immediate area, measurements were made at
the Rubidoux, Riverside-Magnolia, and Riverside sites. The results indicated that the Rubidoux
site did represent urban-scale contributions of primary motor vehicle exhaust, secondary sulfate,
and secondary nitrate. However, there also were significant neighborhood-scale and urban-scale
contributions of primary geological sources and lime/gypsum sources contributing to the PM10
concentration (Chow et al., 1992c).
The Los Angeles smog has been the subject of extensive spatial, temporal, size and
chemical composition studies since the 1960s (Appel et al., 1976, 1978, 1979; Hidy et al., 1980).
A number of individual studies are discussed below.
The chemical characteristics of the PM10 aerosols were measured throughout 1986
(Solomon et al., 1989). Five major aerosol components (carbonaceous material, elemental
carbon and organic carbon [measured value multiplied by 1.4 to account for O and H associated
with C], nitrate, sulfate, ammonium, and soil-related materials, as measured) accounted for over
80% of the 1986 annual average PM10 mass. In all, measured chemical components were
included from 80 to 94% of the PM10 mass was chemically identified. The nitrate and
ammonium concentrations were substantially higher at the Rubidoux and Upland sites than at
6-158
-------
other sites. Measurements made off the coast at San Nicolas Island suggest that from 28 to 44%
of the annual average concentration at individual on-land sites can be associated with a regional
background (Solomon et al., 1989).
More recently the LA aerosol characteristics during 11 summer days and 6 fall days in
1987 have been further elucidated by Southern California Air Quality Study (SCAQS) (Lawson,
1990). Several of the SCAQS studies reported are discussed below. The SCAQS study is also
discussed in Chapter 3, Section 3.4.2.3.
Nitrate, sulfate, ammonium, and organic and elemental carbon were the most abundant
species in the PM2 5 fraction during SCAQS (Chow et al., 1994a). The coarse particle fraction
was composed largely of soil-related elements (e.g., aluminum, silicon, calcium, iron) at the
inland sites and with marine-related elements (e.g., sodium, chloride) at the coastal sites.
Average concentrations for most chemical compounds were higher during the fall than during
the summer, except for sulfate which was more abundant in summer. The PM2 5 constituted one-
half to two-thirds of PM10 at all sampling sites. PM25 nitrate and ammonium concentrations
were negatively biased for daytime samples compared to nighttime samples, consistent with
diurnal changes in temperature and the effect of these changes on the equilibrium between
particulate ammonium nitrate and gaseous ammonia and nitric acid. (Chow et al., 1994a; Watson
etal., 1994a).
Wolff et al. (1991) measured the smog aerosol pattern during SCAQS at Claremont, CA,
and Long Beach, CA, in the eastern and western Los Angles basin, respectively. Claremont's air
quality during the summer was characterized by high concentrations of photochemically
produced pollutants including ozone, nitric acid, particulate nitrate, and particulate organic
carbon (OC). The highest concentrations of these species were experienced during the daytime
sampling period (0600 to 1800) and were associated with transport from the western part of the
basin. Long Beach's air quality during the fall was characterized by frequent periods of air
stagnation that resulted in high concentrations of primary pollutants including PM10, OC and
elemental carbon (EC) as well as particulate nitrate. Night -time levels of most constituents
exceeded daytime levels due to poorer night-time dispersion conditions. At Claremont, OC and
nitrate compounds accounted for 52% of PM10, while at Long Beach they accounted for 67% of
PM10. On the average, there appears to be sufficient particulate ammonium to completely
neutralize the nitrate and acidic sulfates.
6-159
-------
In situ, time resolved analysis for aerosol organic and elemental carbon in Glendora, CA
(Turpin et al., 1990), showed strong diurnal variations with peaks occurring in the daylight
hours. Comparison of the diurnal profile of organic carbon with those of elemental carbon
provided evidence for the secondary formation of organic aerosol in the atmosphere. Turpin et
al. (1991) observed that secondary organic aerosol appears to have contributed roughly half of
the organic aerosol in Pasadena during midday summer conditions.
Turpin and Huntzicker (1991) also found that the organic and elemental carbon
concentrations exhibit strong diurnal variations. Peak concentrations occur during the daylight
hours in the summer and at night in the fall. The maximum concentrations observed in the fall
(maximum total carbon, 88 //g/m3) were two to three times higher than the summer maxima
(maximum total carbon, 36 //g/m3). Measurements of elemental and organic carbon have been
carried out by Gray et al. (1986). Extensive efforts have been made by Cass and coworkers (e.g.
Rogget et al., 1993; Hildemann et al., 1991) to identify the molecular composition of the organic
component. While some tracers have been identified, only a fraction of the organic PM has been
characterized in terms of its molecular composition.
Gaseous nitric acid and fine paniculate nitrate at Claremont, CA (Pierson and Brachaczek,
1988) both showed pronounced (~10-fold) diurnal variations; however, coarse particles showed
little diurnal variation. The average concentrations over the September 11 to 19 study period
were for HNO3, 7.1 //g/m3; fine NOS 7.29 //g/m3; and coarse NOj, 7.1 //g/m3. Fine NOj may
have been underestimated due to volatilization during or after sampling. This problem is
discussed in Chapter 4, Section 4.2.10.1.
Careful size distribution measurements in the Los Angeles basin (John et al., 1990) shed
light on the size spectrum dynamics for ammonium, sulfate and nitrate. Three modes, two
submicron and one coarse, were sufficient to fit all of the size distributions. The smallest mode,
at 0.2±0.1 //m aerodynamic diameter, is probably a condensation mode containing gas phase
reaction products. A larger mode at 0.7±0.2 //m is defined as a droplet mode. Most of the
inorganic particle mass was found in the droplet mode. The observed condensation and droplet
modes characterize the overall size distribution in the 0.1 to 1.0 //m range, previously described
by Whitby and coworkers as a single accumulation mode (Whitby et al., 1972; Whitby, 1978).
Wall et al. (1988) also found that in September 1985 at Claremont, CA fine particle nitrate was
6-160
-------
associated with ammonium, while coarse mode nitrate was associated with both ammonium and
sodium. Sulfate was primarily in two submicrometer modes.
A clear demonstration of the effect of relative humidity and aerosol loading on
atmospheric sulfate size distributions is given by Hering and Friedlander (1982). Days of high
relative humidity and aerosol loading correspond to high mass median diameters (0.54±0.07 //m)
for the sulfate while low relative humidity and low aerosol loadings correspond to small mass
median diameters (0.2±0.02 //m). According to their interpretation, the larger (0.54 //m) sulfate
particles resulted from aqueous phase reactions of SO2. The finer (0.2 //m) sulfate resulted from
homogeneous gas phase reactions leading to the nucleation of sulfuric acid particles.
McMurry and Stolzenburg (1989) provide evidence that Los Angeles smog aerosols are
externally mixed. Monodisperse ambient aerosols were often found to split into nonhygroscopic
(no water uptake) and hygroscopic portions when humidified. An average of 30% of the
particles in the 0.2 to 0.5 //m range were nonhygroscopic. However, the proportion of particles
that were nonhygroscopic varied considerably from day to day and on occasion was 70 to 80%
of the particles. The data show that for the hydrophilic aerosol, the larger particles (0.4 to
0.5 //m) grew more when humidified than did smaller particles (0.05 to 0.2 //m).
Size distributions of aerosol phase aliphatic and carbonyl groups at Claremont, CA (Pickle
et al., 1990) showed maxima in the 0.12 to 0.26 //m and the 0.5 to 1.0 //m size functions. From
the aliphatic carbon absorbency, the ambient samples generally showed maxima in the 0.076 to
0.12 //m size fraction. The authors attribute the carbonyl absorbance almost entirely attributed
to products of atmospheric reactions and the aliphatic absorbencies in particles smaller than 0.12
//m to automotive emissions.
Cahill et al. (1990) found that the sulfate aerosol size at Glendora, CA, is smaller, 0.33 //m
(MMD) during clear days compared to 0.5 //m on smoggy days.
The size distributions of organic nitrate groups in ambient Los Angeles aerosol were
typically bimodal (Mylonas et al., 1991). During periods of high photochemical activity, the
maxima in the mass loadings were in the 0.05 to 0.075 //m and the 0.12 to 0.26 //m size
fractions. During periods of low-moderate ozone concentrations, the distributions were shifted
to slightly larger sizes, with maxima appearing in the 0.075 to 012 //m and the 0.5 to 1.0 //m size
fractions. A principal component analysis of the organonitrate loadings revealed strong
correlations with ozone concentrations and with aerosol phase carbonyl loadings.
6-161
-------
The analysis of coarse particles in Claremont, CA (Noll et al., 1990) show that the coarse
particle mass could be divided into two categories: material that was primarily of crustal origin
(Al, Ca, Fe, and Si) and material that was primarily of anthropogenic origin (Cd, Cu, Mn, Ni,
Pb, and Zn). The mass of crustal material varied between 33 and 49% of the total coarse mass,
while the mass of anthropogenic elements listed above were <1%.
The daily frequency distribution of the chemical components of the Los Angeles aerosol
measured over a 1-year period were approximately lognormal (Kao and Friedlander, 1994). For
nonreactive aerosol components, the geometric standard deviation (GSD) is nearly constant at
1.85±0.14 even for components from different source types. An apparent bimodal frequency
distribution for sulfates probably corresponds to the two differing reaction pathways by which
gas-to-particle conversion occurs. However, the bimodal sulfate distribution function was not
found at other Los Angeles sites (Kao and Friedlander, 1995). The authors suspect a
relationship between GSD and the level of complexity of the stochastic physical and chemical
processes affecting the distributions of the individual species. They also point out that the
chemical concentration of the Los Angeles aerosol that corresponded to the peak in the (nearly)
lognormal frequency distribution of the total mass is lower than he simple average chemical
concentration.
A long term data base for organic and elemental carbon has been constructed (Cass et al.,
1984; Gray et al., 1984). The average elemental carbon concentrations at seven monitoring sites
in the Los Angeles area, for the 24-year period (1958 to 1982), were estimated to range from
6.4 //g/m3 at downtown Los Angeles to 4.5 //g/m3 at West Los Angeles. At most monitoring
sites studied, elemental carbon concentration were lower in recent years than during the late
1950s and early 1960s.
6-162
-------
6.6 CHEMICAL COMPOSITION OF PARTICULATE MATTER
AEROSOLS AT URBAN AND NONURBAN SITES
This section summarizes selected data from a number of studies for the composition of
atmospheric particles in suburban, urban, and a few rural areas for comparison purposes.
Emphasis has been placed on the Harvard six-city study and the inhalable particulate network
(1980-1981). Data for fine particle mass and elemental composition were available from these
studies. Data for sulfate, nitrate, and elemental and organic carbon content are included from
other studies to provide an overview of the chemical composition of the atmospheric aerosol in
the United States. Tables presented in Appendix 6A provide relatively detailed representations
of the properties of atmospheric particles to which U.S. populations are exposed. Unfortunately,
data this complete are generally collected only during intensive studies. The tables are meant to
provide examples of the types of information that could be collected as part of future monitoring
efforts in support of human exposure investigations.
A summary of all the aerosol sampling studies included in this compilation is given in
Tables 6A-la, 6A-lb, and 6A-lc. Sampling studies have been grouped by geographical region
roughly corresponding to the eastern, central, and western United States. Data are tabulated for
the PM25 (d < 2.5 |im), the coarse fraction of PM10 (2.5 |im < d < 10 |im) and PM-10 (d < 10
//m) size fractions of the ambient aerosol in Tables 6A-2a, 6A-2b, and 6A-2c. Compositional
data for all size fractions were broken down into the following major components: sulfate, as
SOJ; carbon, as organic carbon (OC), which as been multiplied by a factor of 1.4 to account for
the presence of oxidized species, and elemental carbon (EC); nitrate as NO3"; and remaining
trace elements. The NH4+, that would be required to neutralize all acidic species in the samples,
is shown as (NH4+)*. Representing sulfate as ammonium sulfate and using a factor of 1.4 to
account for the mass of organic carbon present in oxidized forms allows a firm lower limit to be
placed on the fractional mass that is not chemically identified in filter samples. Acidity is given
in units of nmoles/M3 in Tables 6A-2a and 6A-2c. The masses of the trace elements from
sodium through lead have been calculated by assuming they are in their most stable forms for
conditions at the earth's surface. Reconstructed masses calculated in this way are shown by the
entry, Sum, along with measured masses, and the ratio of the two are shown at the bottom of the
individual summaries for each size fraction. Not all compositional categories were measured in
the studies for inclusion in the tables. For instance, data for characterizing the carbon or nitrate
6-163
-------
content of the ambient aerosol are not available for many of the studies listed. Average data are
shown in graphical form in Figures 6-85a, 6-85b, and 6-85c for studies in the eastern, central,
and western United States.
As can be seen from inspection of Figure 6-85a, sulfate is the major identified component
of mass for fine particles (34.1%), followed by elemental and organic carbon (24.8%), minerals
(4.3%), and nitrate (1.1%) for studies in the eastern United States. However, this last inference
is based on only a few studies in which nitrate was measured. Pierson et al. (1980a,b, 1989)
measured nitrate as constituting only 0.8% to 1.4% of aerosol mass at Allegheny Mountain and
Laurel Hill in southwest Pennsylvania in the summers of 1977 and 1983. Presumably, the low
nitrate in these and other studies in the eastern United States is related to aerosol acidity. Coarse
particles are seen to consist mainly of mineral forming elements (51.8%) and sulfate (4.9%).
Not enough data were available to determine abundances of carbon species and nitrate in the
coarse fraction. A sizable fraction of both the fine (22.8%) and coarse (41.5%) particle mass is
shown as unknown. This unknown mass is assumed to be mainly water, either bound as water
of hydration or associated with hygroscopic particles. A small fraction of the mass, especially in
the coarse fraction, may be present as carbonates. Carbonates are difficult to quantify, in part
because of artifact forming reactions with atmospheric CO2 and acids on filters. Stable
carbonates could be identified by SEM in regions where they are known to represent a
substantial fraction of soil composition.
Fine particles sampled in the studies shown in Table 6A-1 in the central United States
(Figure 6-85b) are seen to consist mainly of sulfate (22.3%), minerals (7.6%), and elemental and
organic carbon (53.6%). The reconstructed mass percentages sum to 124.8%. This could be due
to an overestimation of the carbon content which was estimated from only a few samples
collected during winter in woodsmoke impacted areas. Coarse particles were found to consist
mainly of minerals (62.8%), sulfate (3.1%) and an unknown fraction (33.0%). No nitrate or
carbon data were available for the coarse fraction from the studies in the central United States.
While gross fine particle composition appears to be broadly similar between the eastern
and central United States on the basis of the studies shown in Tables 6A-la, 6A-lb, and 6A-lc,
the fine particle composition is seen to be distinctly different in the western United States
6-164
-------
PM2.5 Mass Apportionment
1 Minerals 4.3%
Unknown 22.8% ^~~~~~
EC 3.9%
OCx 1.4 20.9%
SO4 34.1%
(NHJ )* 13.0%
Nitrate based on 3 studies
NO; 1.1%
Coarse Mass Apportionment
Unknown 41.5%
Minerals 51.8%
4.9%
Insufficient Nitrate, OC, and EC data available
PM10 Mass Apportionment
Minerals 19.6%
Unknown 28.9%
EC 3.3% —
OCx 1.4 8.5%
NO" 1.2%
SO4 27.8%
)* 10.7%
Nitrate based on 2 studies
Figure 6-85a. Major constituents of particles measured at sites in the eastern United States,
as shown in Tables 6A-2a, 6A-2b, and 6A-2c. (NH4+)* represents the
concentration of NH4+ that would be required if all $£J were present as
(NH4)2SO4 and all NO3 as NUjNOj. Therefore, (NH+)* represents an upper
limit to the true concentration of NH4+.
6-165
-------
PM2.5 Mass Apportionment
EC 9.0% v / Minerals 7.6%
OCx1.4 44.6%
S04 22.3%
(NHJ)* 10.2%
NOj 8.1%
Reconstructed sum = 124.8%
Coarse Mass Apportionment
Unknown 33.0%
(NHj)* 11% * 7 Minerals 62.8%
SOJ 3.1%
Insufficient Nitrate, OC, and EC data available
PM10 Mass Apportionment
EC 29.6%
OCx 1.4 5.0%
Minerals 35.8%
SO4 3.3%
(Nh£)* 6.5%
Nitrate based on 2 studies; OC and EC based on 4 studies
Reconstructed sum = 103.9%
NOj 23.7%
Figure 6-85b. Major constituents of particles measured at sites in the central United States,
as shown in Tables 6A-2a, 6A-2b, and 6A-2c. (NH4+)* represents the
concentration of NH4+ that would be required if all SO4 were present as
(NH4)2SO4 and all NO3 as NE^NO*,. Therefore, (NH+)* represents an upper
limit to the true concentration of NH4+.
6-166
-------
PM2.5 Mass Apportionment
EC 14.7% ~~*- —— l~~ Minerals 14.6%
SOT 10.8%
)* 7.5%
OCx 1.4 38.9%
NO; 15.7%
Reconstructed sum = 102.2%
Coarse Mass Apportionment
Unknown 27.0%
Minerals 69.9%
Insufficient Nitrate, OC, and EC data available
PM10 Mass Apportionment
EC 5.1%
OCx 1.4 30.0%
Minerals 36.3%
SO4 4.6%
24.0% ' (NHt )* 6.7%
Reconstructed sum = 111.4%
Figure 6-85c. Major constituents of particles measured at sites in the western United States,
as shown in Tables 6A-2a, 6A-2b, and 6A-2c. (NH4+)* represents the
concentration of NH4+ that would be required if all 840 were present as
(NH4)2SO4 and all NO3 as NUNO,. Therefore, (N5,+)" represents an upper
limit to the true concentration of NH4+.
6-167
-------
(Figure 6-85c). Elemental plus organic carbon species (53.6%) are the major identified
component of mass, instead of sulfate (10.8%), and minerals and nitrate account for a larger
fraction of total mass. While minerals are seen to account for most of the coarse particle mass
(69.9%), available data were insufficient to estimate the contributions of elemental and organic
carbon species to the coarse mass. Table 6A-3 shows a comparison of selected ratios of mass
components for studies conducted in each of the three broad regions of the United States.
Many of the studies listed in Table 6A-3 involved data collected at more than one site
within an airshed. Information about the variability of particle mass within an airshed can yield
information about the nature of sources of the particles. The variability of mean concentrations
measured at multiple sites within a study area is used as a measure of the intersite variability in
fine particle composition and is shown in Tables 6A-4a and 6A-4b.
6.7 ACID AEROSOLS
6.7.1 Introduction
Acid aerosols are secondary pollutants formed primarily through oxidation of sulfur
dioxide (SO2), a gas emitted by the combustion of fossil fuels. Oxidation of SO2 forms sulfate
(SO4), the major component of acid aerosols. Sulfate is formed to a lesser extent through the
oxidation of sulfur species (H2S and CH3SCH3) from natural sources. The oxidation of SO2
occurs through a series of heterogeneous (gas-particle) or homogeneous (gas or aqueous) phase
oxidation reactions that convert SO2 to sulfuric acid (H2SO4) particles. The sulfate species are
typically expressed in terms of total SO4, with the acidic fraction expressed in terms of titratable
H+ ([H+] + [HSO4]) and referred to as aerosol strong acidity. The chemical aspects of oxidation
of SO2 and formation of aerosol strong acidity are discussed in Chapter 3, Section 3.3.1. H+ is
usually found in the fine particle size fraction (aerodynamic diameter (Dp) < 1.0 jam) (Koutrakis
and Kelly, 1993; Pierson et al., 1980a, 1989). However, acidity may be found in larger particles
during periods of fog or very high relative humidity. Keeler et al. (1988) and Pierson et al.
(1989) report finding acidity in the > 2.5 //m size range when the relative humidity was close to
100%. Although recent research has shown a high correlation between SO4 and acidity, data
from summertime sampling have shown that SO4 is not always a reliable predictor of FT for
individual events at a given site (Lipfert and Wyzga, 1993).
6-168
-------
A major determinant of the lifetime of H+ in the atmosphere is the rate of neutralization by
ammonia (NH3). Ammonia reacts with H2SO4 to form ammonium sulfate [(NH4)2SO4] and
ammonium bisulfate (NH4HSO4). The major sources of ammonia in the environment are
animals and humans (Fekete and Gyenes, 1993). The then current state-of-knowledge regarding
acid aerosols was reviewed by EPA in 1989 (U.S. Environmental Protection Agency, 1989) and
by Spengler et al., 1990. A more recent summary is given by Waldman et al. (1995).
6.7.2 Geographical Distribution
In North America, ambient concentrations of H+ tend to be regional in nature with the
highest concentrations found in the northeastern United States and southwestern Canada.
Spengler et al. (1990) have collected information on maximum values of SO4 and H+ found
across the U.S. and southern Canada. This information is shown in Table 6-5.
6.7.3 Spatial Variation (Regional-Scale)
Recent evidence has shown that meteorology and regional transport are extremely
important to acid sulfate concentrations. Elevated levels of ambient H+ were measured
simultaneously during a regional episode at multiple sites located from Tennessee to Connecticut
(Keeler et al., 1991). Lamborg et al. (1992) measured H+ concentrations to investigate the
behavior of regional and urban plumes advecting across Lake Michigan. Results suggested that
aerosol acidity is maintained over long distances (up to 100 km or more) in air masses moving
over large bodies of water. Lee et al. (1993) reported that H+ and SO4 concentrations measured
in Chicago over a year were similar to levels measured in St. Louis. In an analysis of acid
sulfate concentrations measured at Pittsburgh, State College, and Uniontown, PA, Liu et al.
(1996) reported high correlations for H+ between all three locations. The three locations are
separated by large distances (approximately 60 to 240 km) and have vastly different population
densities. It is commonly believed that the source region for most of the H+ precursors (primary
inorganic pollutant gases —SO2 and NOX) is the Ohio River Valley (Lioy et al., 1980). The
conversion of the primary gases to secondary pollutants takes place as the prevailing winds carry
the precursors
6-169
-------
TABLE 6-5. MAXIMUM SCT4 AND H+ CONCENTRATIONS
MEASURED AT NORTH AMERICAN SITES
(H+ concentrations expressed as sulfuric acid (H2SO4)equivalents;
"SC" indicates semi-continuous measurements.)
Location
Lennox, CA
Smoky Mountains
High Point, NJ
Brookhaven, NY
Tuxedo, NY
St. Louis, MO
St. Louis, MO
Los Angeles, CA
Harriman, TN
Watertown, MA
Fairview Lake, NJ
Warren, MI
Whiteface Mt, NY
Toronto, ON, Canada
Allegheny Mt., PA
Laurel Hill, PA
Harriman, TN
St. Louis, MO
Topeka, KS
Watertown, MA
Steubenville, OH
Portage, WI
Kanawha Valley, WV
Dunville, ON, Canada
Hendersonville, TN
Livermore, CA
Morehead, KY
Monroeville, PA
Pembroke, ON, Canada
Springdale, AR
Newtown, CT
Allegheny Mt., PA
Uniontown, PA
State College, PA
Philadelphia, PA
Pittsburgh, PA
Sample Duration (h)
2-8
12
6
3
1-12
SC
SC
12
SC
SC
SC,4
24
24
8,16
7,10
7,10
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
12
12,24
12
24
6,24
Maximum
so;(//g-m-3)
18
17
37
24
41
25
43
10
47
31
27
37
59
75
45
56
28
40
14
23
56
33
46
31
23
9
23
42
29
11
26
33
52
47
39
27
Concentration
H2S04 (//g-m-3)
0.1
10
18
10
9
7
34
3
18
14
12
9
14
19
31
42
14
6
3
9
18
4
22
15
11
2
14
18
14
2
8
20
39
25
9
15
Source: Spengler et al. (1990).
6-170
-------
from the source region, northeastward to the northeastern United States and southwestern
Canada. This type of northeasterly wind flow occurs on the backside (western side) of
mid-latitude anti-cyclones (high pressure systems).
Pierson et al. (1980a,b, 1989) conducted studies of atmospheric acidity on Allegheny
Mountain and Laurel Hill in southwest Pennsylvania, 80 and 100 km southeast of Pittsburgh, in
the summers of 1977 and 1983. The aerosol H+ appeared to represent the net after H2SO4
reaction with NH3(g). The resulting HVSO^ ratio depended on SO^ concentration, approaching
that of H2SO4 at the highest SO^ concentrations. The atmospheric was acidic; the average
concentrations of HNO3 (78 nmole/m3) and aerosol H4" (205 nmole/m3), NH4+ (172 nmole/m3),
and SCT4 (201 nmole/m3), and the dearth of NH3 (<15 nmole/m3), show that the proton acidity of
the air exceeded the acid-neutralizing capacity of air by a factor of >2, with one 10-hour period
averaging 263 nmole/m3 for HNO3 and 844 nmole/m3 for H+. SO2 added another 900 nmole/m3
(average) of potential H+ acidity. HNO3 and aerosol H+ episodes were concurrent, on 7-8 day
cycles, unrelated to SO2 which existed more in short-lived bursts of apparently more local origin.
NOX was sporadic like SO2. Laurel and Allegheny, separated by 35.5 km, were essentially
identical in aerosol SO^ , and in aerosol H+, less so in HNO3; apparently, chemistry involving
HNO3 and aerosol H+ or SO^ was slow compared to inter-site transport times (1-2 hours). From
growth of bscat and decline of SO2, daytime rate coefficients for SO2 oxidation and SO2 dry
deposition were inferred to have been, respectively, -0.05 and <0.1 hr"L
HNO3 declined at night. Aerosol H+ and SO4 showed no significant diurnal variation, and
O3 showed very little; these observations, together with high PAN/NOX ratios, indicate that
regional transport rather than local chemistry is governing. The O3 concentration (average
56 ppb or 2178 nmole/m3) connotes an oxidizing atmosphere conducive to acid formation.
Highest atmospheric acidity was associated with (1) slow westerly winds traversing
westward SO2 source areas, (2) local stagnation, or (3) regional transport around to the back side
of a high pressure system. Low acidity was associated with fast-moving air masses and with
winds from the northerly directions; upwind precipitation also played a moderating role in air
parcel acidity. Much of the SO2 and NOX, and ultimately of the HNO3 and aerosol H+, appeared
to originate from coal-fired power plants. An automotive contribution to the NOX and HNO3
could not be discerned.
6-171
-------
Size distributions of aerosol FT and SC^ were alike, with MMED -0.7 //m, in the optimum
range for efficient light scattering and inefficient wet/dry removal. Thus, light scattering and
visual range degradation were attributable to the acidic SO^ aerosol. With inefficient removal of
aerosol FT, and inefficient nighttime removal of HNO3, strong acids may be capable of long-
distance transport in the lower troposphere. Water associated with the acidic aerosol was shown
to account for much of the light scattering.
6.7.4 Spatial Variation (City-Scale)
A study of acid aerosols and ammonia (Suh et al., 1992) found no significant spatial
variation of H+ at Uniontown, Pennsylvania, a suburb of Pittsburgh. Measurements at the central
monitoring site accounted for 92% of the variability in outdoor concentrations measured at
various homes throughout the town. There was no statistical difference (p > 0.01) between
concentrations of outdoor H+ among five sites (a central site and four satellite sites) in Newtown,
Connecticut (Thompson et al., 1991). However, there were differences in peak values which
were probably related to the proximity of the sampling sites to ammonia sources. These studies
suggest that long-term averages should not substantially differ across a suburban community,
although peak values may differ significantly.
In small suburban communities outdoor concentrations of H+ are fairly uniform, suggesting
that minor differences in population density do not significantly affect outdoor FT or NH3
concentrations (Suh et al., 1992). In urban areas, however both H+ and NH3 exhibit significant
spatial variation. Waldman et al. (1990) measured ambient concentrations of H+, NH3, and SO4
at three locations in metropolitan Toronto. The sites, located up to 33 km apart, had significant
differences in outdoor concentrations of H+. Waldman and co-workers reported that the sites
with high NH3 measured low H+ concentrations. However, the limited number of sampling sites
did not allow for a conclusive determination of the relationship between population density,
ammonia concentrations, and concentrations of acid aerosols.
An intensive monitoring study has been conducted during the summers of 1992 and 1993
in Philadelphia (Suh et al., 1995). Twenty-four hour measurements of aerosol acidity (H+)
sulfate and NH3 were collected simultaneously at 7 sites in metropolitan Philadelphia and at
Valley Forge, 30 km northeast of the city center. The researchers reported that 864 was evenly
6-172
-------
distributed throughout the measurement area but H+ concentrations varied spatially within
metropolitan Philadelphia. This variation was related to local NH3 concentrations and the local
population density (Figure 6-86). The amount of NH3 available to neutralize H+ increased with
population density, resulting in lower H+ concentrations in more densely populated areas. The
extent of the spatial variation in H+ concentrations did not appear to depend on the overall H+
concentration. It did, however, show a strong inverse association with local NH3 concentrations.
Q
I-
<
0.5
0.4
0.3
0.2
0.1
0.0
120
_ 90
Si
o
E
.£ 60
30
SO,
NH,
H /S04
0 5,000 10,000 15,000 20,000
POPULATION DENSITY (persons/sq.mMe)
Figure 6-86. Mean air pollutant concentrations for days when winds were from the southerly
direction, plotted versus population density. The solid line represents H+
concentrations; the long dashed line represents SO^" concentrations; the dashed
and dotted line represents the ratio of H+ to SO^" levels; and the dotted line
represents NH3 concentrations. All data collected in Philadelphia, PA, during
the summers of 1992 and 1993.
Source: Adapted from Suh et al. (1995).
6.7.5 Seasonal Variation
An analysis of results from Harvard's 24-City Study (Thompson et al., 1991), which
measured acid aerosols concentrations at 8 different small cities across North America each year
during a three year period, revealed that the summer H+ mean concentrations were significantly
higher than the annual means at all sites. The results showed that at the sites with high H+
concentrations, approximately two-thirds of the aerosol acidity occurred from May through
6-173
-------
September (Figure 6-87). Little or no seasonal variation was observed at sites with low acidity.
These findings were supported by those of Thurston et al. (1992) in which H+ concentrations
measured at Buffalo, Albany, and White Plains, NY, were found to be highest during the
summertime. Thurston and co-workers also reported that moderate concentrations of H+ could
occur during non-summer months
— «|^u
2
0
E
S 90-
_>.
jo
'o
>. 60-
£
C
o
S-
g.30-
k
«
D Hendersonville, TN • Morehead, KY
DDunnville, Ontario, Canada
D Pembroke, Ontario, Canada
DLivermore, CA
n
ikl
-
r
Lit
1
i
1
-
r
-
-i
1
DPenn Hills, PA
DNewtown, CT
DSpringdale, AR
-
.
'l
n
IttliiJ
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Month
Figure 6-87. Average monthly aerosol strong acidity for Year 1 sites of the Harvard 24-City
Study.
Source: Thompson et al. (1991).
6.7.6 Diurnal Variation
Evidence exists of a distinct diurnal pattern in outdoor FT concentrations. Wilson et al.
(1991) examined concentration data for FT, NH3, and SO4 from the Harvard 24-City Study for
evidence of diurnal variability (Figure 6-88). This investigation found a distinct diurnal pattern
for H+ concentrations and the FF7SO4 ratio, with daytime concentrations being substantially
6-174
-------
60 80 100 120 140 160 180 200
(0
TJ
c
10
(0
3
O
10
c>
O
E
c
3.2
3.0
2.8
2.6
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Sulfate
Hydrogen Ion
•A A A
0 2 4 6 8 10 12 14 16 18 20 22 24
Hour
Figure 6-88. Diurnal pattern of sulfate and hydrogen ion at Harriman, TN, weekly pattern
and daily average.
Source: Wilson et al. (1991).
6-175
-------
higher than nighttime levels. Both H+ and SO4 concentrations peaked between noon and
6:00 pm. No such diurnal variation was found for NH3. Wilson and co-workers concluded that
the diurnal variation in H+ and SO4 was probably due to atmospheric mixing. Air containing
high concentrations of H+ and SO4 mixes downward during daylight hours when the atmosphere
is unstable and well-mixed. During the night, ammonia emitted from ground-based sources
neutralizes the acid in nocturnal boundary layer, the very stable lower part of the atmosphere,
but a nocturnal inversion prevents the ammonia from reacting with the acid aerosols aloft. Then
in the morning as the nocturnal inversion dissipates, the acid aerosols mix downward again as
the process begins anew. Spengler et al. (1986a) also noted diurnal variations in sulfate and
sulfuric acid concentrations and suggested atmospheric dynamics as the cause. The diurnal
variation in SO4 has been observed by other workers and discussed in terms of atmospheric
dynamics by Wolff et al. (1979) and Wilson and Stockberger (1990).
This diurnal variation in mixing heights and concentrations does not seem to hold at
elevated sites. For example, Pierson et al. (1980a,b, 1989) found no appreciable night/day
difference in aerosol FT (or NH4+ or SO4 ), and almost no diurnal variation in O3, at two
elevated sites (Allegheny Mountain and Laurel Hill, elevations 838 and 850 m) in southwest
Pennsylvania. They contrasted this behavior with that at lower sites, and particularly with the
concurrent measurements at Deep Creek Lake (Vossler et al., 1989). The differences were
attributed to isolation from ground-based processes at the elevated sites at night.
6.7.7 Indoor and Personal Concentrations
Several studies have examined indoor concentrations of acid aerosols and personal
monitoring. Brauer et al. (1989) monitored personal exposures to particles (including acidic
sulfates) and gases in metropolitan Boston in the summer of 1988, and compared these to
measurements collected at a centrally located ambient monitor. They found that personal
concentrations of acidic aerosols and gases differed significantly from those measured at the
centrally located site. Summer and winter concentrations of acid aerosols and gaseous pollutants
also collected in Boston (Brauer et al., 1991) showed indoor/outdoor ratios of FT to be 40-50%
of the indoor/outdoor SO4 ratio indicating neutralization of the acid by the higher indoor NH3
levels, which were reported.
6-176
-------
Indoor, outdoor, and personal acid aerosol monitoring was performed for children living in
Uniontown, Pennsylvania, during the summer of 1990 (Suh et al., 1992). The indoor, outdoor,
and personal measurements were compared to outdoor measurements collected from a centrally
located ambient monitor. Personal concentrations were lower than corresponding outdoor levels
but higher than indoor levels. Air conditioning was found to be an important predictor of indoor
H+, while NH3 was found to influence indoor and personal H+ concentrations. Similar results
were obtained in a study of the relationships between indoor/outdoor concentrations of H+ and
NH3 conducted in State College, PA, in 1991 (Suh et al., 1994).
In a study characterizing H+ concentrations at child and elderly care facilities, Liang and
Waldman (1992) measured indoor and outdoor acid aerosol concentrations. Results from this
study showed that indoor/outdoor H+ and SO4 ratios were comparable to those measured inside
residential buildings. Air conditioner use and indoor NH3 concentrations were again identified
as important determinants of indoor FT concentrations.
6.8 NUMBER CONCENTRATION OF ULTRAFINE PARTICLES
6.8.1 Introduction
Recent work has suggested that ultrafme particles may be responsible for some of the
health effects associated with exposure to particulate matter (Chapter 11, Section 11.4). The
hypothesis for explaining a biological effect of ultrafme particles is based on the number,
composition and size of particles rather than their mass (Seaton et al., 1995). This has led to an
interest in the number concentration of ambient particles. This section examines data on particle
number concentration and the relationship between particle number and particle mass or volume.
6.8.2 Ultrafine Particle Number-Size Distribution
In the context of ambient particles, the term ultrafme particles refers to those particles with
diameters below 0.1 //m. Ultrafine aerosol size distributions from an urban site at Long Beach,
California (Karch et al., 1987), and from a background site in the Rocky Mountains, Colorado
(Kreidenwies and Brechtel, 1995) are shown in Figures 6-89 and 6-90. Both of these sets of data
were obtained by electrical mobility measurements. For the urban aerosols of Long Beach, the
6-177
-------
120,0001
100,000"
'E 80,000
u
a
a 60,000
O)
o
40,000"
20,000"
(a)
12T
(b)
0.00
Long Beach, CA
•1200-2400
•1200-1300
•1400-1500
•2100-2200
0.01
Particle Diameter (pm)
Long Beach, CA
•J 1U
u
*£ 8-
^
a
a e-
O)
0
S
> 4-
TJ
-•-1200-2400
-6-1200-1300
-D- 1400-1500
-0-2100-2200
-0—B D—0—0—D-
0.01
Particle Diameter (um)
0.10
0.10
Figure 6-89. Aerosol number (a) and volume (b) size distributions from an urban site at
Long Beach, CA.
6-178
-------
1,200
Rocky Mountains, CO
z 400"
•o
0.1
Particle Diameter (urn)
11/23/941304
11/23/941804
11/24/941205
*^
<£
o
•«.
1
Q.
Q
O)
0
TJ
^
>
0.6'
0.5-
0.4-
0.3"
0.2-
(b)
V1*/
-D- 11/23/94 1304
-•-11/23/941804
-0-11/24/941205
Rocky Mountains, CO
0.1"
Q •lUUHUUHiUIUUM
0.01
0.1
Particle Diameter, Dp (urn)
Figure 6-90. Aerosol number (a) and volume (b) size distributions from a background site
in the Rocky Mountains, CO.
6-179
-------
number geometric mean diameter can vary from 0.012 //m to 0.043 //m. Some of the ultrafine
distributions, such as that shown for the 1,200 to 1,300 PST time period, are bimodal. The
number concentrations were higher in the early afternoon, 1400-1500 PST, as shown in
Figure 6-91. For the background aerosols from Rocky Mountains the number geometric mean
diameter of the ultrafine aerosols was somewhat larger than for Long Beach, with geometric
mean diameters ranging from 0.047 to 0.075 //m for periods without urban influence. A
bimodal character for the ultrafine distribution was also observed for some measurements, as
seen in Figure 6-90.
12
14
16
20
22
24
18
Time of Day
Figure 6-91. Number concentrations as a function of time of day at Long Beach, CA.
The contrast between urban and background ultrafine aerosol size distribtution is
demonstrated in Figure 6-92, where a change in the wind direction brought transport from an
urban area to the background site at Rocky Mountains. Within a 2-h period, the number
6-180
-------
60,000
Rocky Mountains, CO
12/25/941524
12/25/941550
12/25/941648
0.1
Particle Diameter, Dp (urn)
Rocky Mountains, CO
(b)
E
O
rt
E
3.
Q.
Q
O)
O
>
TJ
2.5
2'
1.5'
1 '
0.5'
-D- 12/25/94 1453
-•-12/25/941546
-0-12/25/941653
infrtin
0.1 1
Particle Diameter, Dp (urn)
Figure 6-92. Number (a) and volume (b) size distributions at the Rocky Mountain site
showing an intrusion of urban air.
6-181
-------
concentration increased from 850 cm"3 to 19,000 cm"3, an increase of more than a factor of 20.
In contrast, the volume distribution increased by less than a factor of 5. The number geometric
mean diameter decreased from 0.052 //m for the background aerosol to 0.024 //m for the urban
influenced aerosol. For the urban influenced size distributions, over 96% of the particle number
was measured in particles below 0.1 //m, while 80% of the particle volume was associated with
particles above that size.
6.8.3 Relation of Particle Number to Particle Mass
In general, the majority of airborne particle volume and mass is associated with particles
above 0.1 //m, while the highest number concentration of particles is found in particles below
0.1 //m. This was shown for volume in Figures 6-89 to 6-92 and can be seen for mass in the
recent data collected in the Los Angeles, CA shown in Figure 6-93. As with the data of Whitby
and Sverdrup (1980), the size distributions of Figure 6-93 show data collected by several
instruments. Physical size distributions were measured with an electrical aerosol analyzer for
particles between 0.01 and 0.4 //m, and with a laser optical particle counter for particles between
0.14 and 3 //m. Additionally, Berner (John et al., 1989, 1990) and MOUDI (Marple et al., 1991)
impactors were used to measure the mass size distribution of inorganic ion species and
carbonacous species. These data have been combined (Hering et al., 1996) to give a total mass
distribution from which the number distribution has been calculated assuming an effective
aerosol density of 1.6 g/cm3 and assuming that the water associated with the aerosol is 15% of
the measured dry particle mass (see McMurry and Stolzenburg, 1989). The optical particle
counter was calibrated with ambient particles, size classified by a differential mobility analyzer.
The ambient aerosol has a lower effective refractive index than the polystyrene latex usually
used for calibration (Hering and McMurry, 1991). No fitting was applied to match the different
size distributions in the region of overlap.
Figure 6-93 shows the average of distributions collected over six different days in the fall
of 1987 in downtown Los Angeles, as part of the Southern California Air Quality Study. Particle
number distributions emphasize the ultrafme particles, or "nuclei" mode. Volume distributions
place importance on 0.1 to 1 //m particles which are associated with the "accumulation" mode.
For this average distribution 88% of the particle number is associated with particles below 0.1
6-182
-------
125,000
E 100,000
u
0.1 1
Particle Diameter, Dp(|jm)
10
Impactor ~°—OPC
EAA
160
140
^^
CO
'E -o
•Q
80
0.1
1
10
Particle Diameter, Dp(um)
Figure 6-93. Number (a), and volume and mass (b) size distributions from Los Angeles, CA,
showing comparison of three measurement techniques.
6-183
-------
(j.m, but 99% of the particle volume is from particles above that size. Both the impactor and
optical counter data indicate a weakly bimodal character for the accumulation mode aerosol.
For unimodal, log normal size distributions, the particle volume Fis simply related to the
particle number Nby the relation:
where Dgn is the number geometric mean diameter, and ogis the geometric standard deviation.
However, because of the multimodal character of ambient aerosol size distributions, one does
not expect this simple relationship to hold in the atmosphere. The relationship between particle
number and particle volume was examined for data from the Southern California Air Quality
Study collected at Riverside, CA over 11 days in the summer of 1987, and at downtown Los
Angeles in the fall of 1987 using the methods described above. As shown in Figure 6-94,
particle number concentrations are correlated with the volume associated with particles below
0.1 //m, but are not correlated with the total fine particle volume. Similar results are found for
the data reported from Rocky Mountains, CO and for the data reported by Whitby and Sverdrup
(1980).
6.8.4 Conclusion
The size distribution measurements of aerosols in urban and continental background
regions indicate number geometric mean diameters which vary from 0.01 to 0.08, with the larger
values found in background regions. Particle number concentrations may vary from less than
1,000/cm3 at clean, background sites to over 100,000/cm3 in polluted urban areas. Particle
number concentrations are dominated by the ultrafine or nuclei mode aerosols. In contrast, the
volume (or mass) of fine particles is associated with particles above 0.1 //m, which are
associated with the accumulation mode identified by Whitby and coworkers (Willeke and
Whitby, 1975; Whitby and Sverdrup, 1980). Particle number concentrations are correlated with
the volume of particles below 0.1 //m. The number concentration of ultrafine particles results
from a balance between formation and removal. The rate of removal by coagulation with
accumulation mode
6-184
-------
160,000'
140,000'
•T
g 120,000-
q; 100,000-
o
1 80,000-
| 60,000-
*• 40,000-
20,000-
n.
\fi)
O
• •
•
•V?
• i^^^n
c
JgTD
* Los Angeles
n Riverside
A Whitby Background
0 W/i/tfiy l//*an
A Rocky Mountains
- ^fj§^
A
fib \ \ \ \
0.00 2.00 4.00 6.00
Volume < 0.1um (um 3/cm3)
8.00
160,000-r
W
140,000
\ 120,000
q) 100,000
•0
1 80,000-
^ 80,000-
*• 40,000
20,000
.
0
•
• •
" • '
• •
1 D DD" '•" *"
• mm
•
3 • D D a
^A \ \ \ \
50 100 150
Volume <2.5um (ums/cm3)
I Los Angeles
D Riverside
A Whitby Backgrouni
O Whitby Urban
^ Rocky Mountains
200
Figure 6-94. Relationship between particle number and particle volume ([a] volume <0.1
and [b] <2.5
6-185
-------
particles will increase as the number (and mass and volume) of accumulation mode particles
increases. Therefore, a correlation between number and accumulation mode volume or mass on
a short term (e.g., hourly basis), would not be anticipated. However, as suggested by the
differences in particle number concentrations from 850 cm"3 at a remote site in the Rocky
Mountains, to 19,000 cm"3 in air transported from an urban area, to in excess of 105 cm"3 in
polluted urban areas, a correlation, between the total number concentration and the total fine
article mass or volume, might be expected if comparisons were made over longer periods, e.g.
days. However, no such studies have been done.
6.9 AMBIENT CONCENTRATIONS OF ULTRAFINE METALS
6.9.1 Introduction
Nucleation theory (Seinfeld, 1986) indicates that ultrafine particles will consist of materials
that have very low vapor pressure but which will, at some time, exist in significant vapor
concentrations. This could be the result of rapid formation of a condensible vapor from
chemical conversion of a gas or the formation of a vapor at relatively high concentrations during
combustion. Very small particles, because of their high curvature, have a higher vapor pressure
than larger particles. This is known as the Kelvin effect and becomes increasingly important as
the particle size decreases below 0.1 //m in diameter. The critical size, at which a particle will
grow instead of evaporating, depends on the saturation ratio, the ratio of the vapor pressure of
the particle, pA, to the vapor pressure over a flat surface, p^) (S = pA\p^); the surface tension; and
the molar volume of the condensed phase. Thus, materials such a elemental carbon, formed in
flames, or metal (or metal compound) vapor, formed during combustion, are likely candidates
for ultrafine particles. Sulfuric acid can also form ultrafine particles (Weber et al., 1995) but
whether it nucleates into ultrafine particles or condenses on existing particles depends on the
balance between the formation rate of sulfuric acid and the surface area of preexisting particles
(Seinfeld, 1986).
Thus, ultrafine aerosols may be primary, formed from vapor generated during combustion,
or secondary, formed from vapor generated by chemical reactions in the atmosphere. Because of
their small size, ultrafine particles diffuse rapidly and are lost by deposition to surfaces or by
6-186
-------
growth into larger particles by coagulation. Ultrafme particles also serve as nuclei for
condensation of vapors. Thus, ultrafine particles grow rapidly by coagulation and condensation,
into the accumulation mode. For these reasons, the mass of ultrafine particles in the ambient
atmosphere is generally much smaller than that of the accumulation mode, where removal rates
of particles reach a minimum in non-cloud conditions. The result is that in ambient conditions,
the ultrafine mode is generally indistinct or absent from mass or volume profiles of aerosol
particles versus size. However, a distinct ultrafine mode below 0.1 //m diameter has been
observed in quasi-ambient samples taken close to combustion sources. In these cases, the
distinct ultrafine particle mode is referred to as the nuclei mode (Whitby, 1978).
While there is consensus that ultrafine metal particles are produced and emitted into the
atmosphere, there is little information on ambient concentrations of ultrafine metals. The few
direct measurements available can be extended with some confidence using indirect methods;
i.e., from particle counting techniques that have size information but no chemical information, or
from filter collection methods that have limited size information but detailed compositional
information. Nevertheless, it is clear that more data on ultrafine metals are urgently needed to
gain confidence in the spatial and temporal concentration profiles of this key atmospheric
component.
6.9.2 Formation of Ultrafine Particles
Nucleation theory establishes that high temperature processes are generally required to
form ultrafine metallic aerosols. Such processes are usually anthropogenic, although natural
fires, volcanic eruptions, and other such events can contribute to ultrafine transition and heavy
metals in some circumstances. Table 6-6, taken from Seeker (1990), gives the vaporization
temperature of EPA-regulated metals (Federal Register, 1986) as a function of temperature, with
and without chlorine available in the combustion process.
Note the dramatic shift in temperature for several elements, including lead, for the
chlorine-rich combustion scenario. A similar process has been used to prevent lead from coating
surfaces in internal combustion engines using leaded gasoline. The process used
6-187
-------
TABLE 6-6. REGULATED METALS AND THE VOLATILITY TEMPERATURE
With No Chlorine
Metal
Chromium
Nickel
Beryllium
Silver
Barium
Thallium
Antimony
Lead
Selenium
Cadmium
Osmium
Arsenic
Mercury
Volatility
Temp. (°F)
2935
2210
1930
1660
1560
1330
1220
1160
605
417
105
90
57
Principal
Species
CrO2/CrO3
Ni(OH)2
Be(OH)2
Ag
Ba(OH)2
T1203
Sb2O3
Pb
SeO2
Cd
OsO4
As2O3
Hg
With 10% Chlorine in Waste
Volatility
Temp. (°F)
2930
1280
1930
1160
1660
280
1220
5
605
417
105
90
57
Principal
Species
CrO2/CrO3
MC12
Be(OH)2
AgCl
BaCl2
T1OH
Sb2O3
PbCl4
SeO2
Cd
OsO4
As2O3
Hg
Source: Seeker (1990).
chlorine and bromine-containing additives to form compounds such as PbBrCl which are
gaseous at combustion temperatures but form ultrafine particles after leaving the vehicle.
Numerous theoretical and laboratory studies have shown that the typical size of metals
derived from combustion is ultrafine (Friedlander, 1977; Senior and Flagan, 1982; Seeker,
1990). Analysis of particles from coal combustion by Natusch and Wallace, 1974 and Natusch
et al., 1974 showed an additional aspect. There is a tendency for the condensing metal vapors to
form relatively uniform thickness surface coatings on more refractory particles present in the
combustion effluent stream. If the particles upon which the metals coat themselves are crustal,
as in coal fly ash, this results in a final particle whose enrichment factor compared to crustal
averages depends upon the initial size of the refractory particle—minor for large particles,
extreme for ultrafine particles (Davison et al., 1974). This result also places the (potentially)
toxic metals on the biologically-accessible surface.
Thus, the presence of metals in a combustion process such as incineration of biological and
chemical wastes or treatment of contaminated soils poses a problem. Raising the temperature of
6-188
-------
combustion high enough to completely (> 99.99%) destroy the biological and chemical species
will also enhance the volatilization of metallic components in the feed stock, requiring more
efficient removal methods for ultrafine and accumulation mode metals. Figure 6-95 shows the
enhanced volatilization of metals as the combustion temperature is raised from 1000 °F (540 °C)
to 1800 °F (980 °C) (Seeker, 1990).
As
~ 40
30
o
'Z
c
11J
20
(0
+j
-------
The combustion effluent can be partitioned into three components (Seeker, 1990; Barton et
al., 1990); emitted (as fly ash), captured (assuming there is an attempt to capture fine particles),
and collected in the bottom ash. Assuming no particle removal equipment is in place on the
combustion process, emitted particles will include both the "emitted" component and most of the
"captured" component. In an uncontrolled incineration facility, 96% of mercury, 88% of
cadmium, 58% of lead, and 11% of copper might by emitted into the atmosphere. If control is
attempted, the capture efficiency is only 25% for mercury, but is better for most other metals,
ranging from 86% for cadmium to 91% for copper (Barton et al., 1990). In addition, the
chemical state of the metals in the ultrafine mode can vary from the more toxic phases (for
example, arsenite versus arsenate) as a function of combustion conditions (Chesworth et al.
1994). Thus, we must expect that ultrafine metallic components will be emitted from high
temperature processes in both toxic and less toxic forms.
6.9.3 Techniques for Collecting and Analyzing Ultrafine Metals
Relatively little information exists on concentrations of ultrafine metal particles in ambient
air samples away from combustion sources. There are many reasons. The ultrafine mode falls
off rapidly away from the combustion source, due to the rapid migration of some types of
ultrafine particles into the accumulation mode, and increased dispersion as one moves away
from the source. Many sources of ultrafine metals use tall exhaust stacks, which enhances
dispersion. The largest of the ultrafine particles can overlap the smallest particles of the much
more abundant accumulation mode, roughly 0.2 to 0.7 //m aerodynamic diameter. Particles
must be size-separated using a device with a sharp cut point, ususally a multistage physical
impactor, that entails problems in particle collection and analysis. Since ultrafine particles may
be hard and dry, adhesive coatings are essential in order to avoid particle bounce in the
impactors. Particle bounce typically translates coarser particles onto finer stages, contaminating
the ultrafine particles with the enormously more abundant coarser particles. Finally, one can
collect only a few monolayers of particles (at most) on the adhesive stages before particle
bounce becomes important, assuming the particles themselves are not "sticky". A few
monolayers of particles of 0.1 //m diameter amounts to only about 50 //g/cm2 of total deposit. If
one then desires to perform minor or trace elemental analysis of the deposit, one is then faced
6-190
-------
with analytical requirements that reach picogram (10"12gm) sensitivities. This clearly limits
analytical options.
For these reasons, much of the data available on ultrafine particles does not depend on
compositional analysis. Most information on the presence of ultrafine particles is derived from
particle counting techniques such as the Electrical Mobility Analyzer (EMA), in situations in
which the source is well known (source-enriched). This was the method pioneered in the 1972
ACHEX studies of Los Angeles (Whitby, 1978). Particle counting devices do not normally
result in collection of ultrafine particles in a manner suitable for compositional analysis,
although some of the devices ("particle classifiers") could be modified to provide samples for
subsequent compositional analysis, if desired. The same can be argued for devices such as
diffusion batteries, but to date little has been done along this line in ambient conditions.
Integrated samples of fine particles can be collected on substrates suitable for analysis.
While some optical information is available as one approaches the ultrafine size, most optical
techniques do not work in the ultrafine size range, which is well below the wavelength of light.
A Scanning Electron Microscope (SEM) beam can still resolve ultrafine particles although some
details are lost. The ultrafine particle distribution can then be derived by particle counting
techniques, either manual or automated, and metal composition can be found by X-ray analysis
of the single particles. The enormous gain in signal to noise ratio by selecting individual
particles offsets the loss of X-ray sensitivity (typically parts per thousand) caused by use of the
electron beams to induce the X rays. SEM and electron microprobe analyses rarely achieve any
better than one part per thousand sensitivity. However, for single particles, this is often enough
to classify them by source. Proton microprobes are, at present, not quite able to operate in the
0.1 //m diameter region, but can perform Proton Induced X-ray Emission (PIXE) analysis to one
part per million by mass on single particles as small as 0.3 //m (Cahill, 1980).
Impactors are designed to separate particles by aerodynamic size in such a way as to allow
compositional analysis. Yet here, too, ultrafine particles pose problems. First, most impactors
can not operate effectively below 0.1 //m. The Stokes number for separation of a 0.1 //m
diameter particle from an air stream requires either extremely high jet velocities, extremely low
pressures in the gas stream, or both. While such performance can be achieved in a physical
impactor, most impactors used for ambient particle collection in the 1970's and early 1980's did
not possess this capability. For example, the very popular cyclones and virtual impactors are
6-191
-------
ineffective below about 0.5 //m diameter. The Lundgren-type impactors widely used in
California studies (Lundgren,1967; Flocchini et al., 1976; Barone et al. 1978) used 0.5 //m as the
lowest cut point. Everything smaller was collected on a filter. The Battelle-type samplers
(Mercer, 1964) favored by other groups (Van Grieken et al., 1975) used a lowest cut point of
0.25 //m diameter. Thus, while both these units generated copious information on aerosol
composition, they could not separate ultrafine aerosols from accumulation mode aerosols.
In the mid-1980's four new impactors were developed capable of providing information on
the composition of particles near 0.1 //m diameter: the Low Pressure Impactor, (LPI) (Hering et
al., 1978), the Berner Low Pressure Impactor (BLPI) (Berner and Liirzer, 1980; Wang and John,
1988), the Davis Rotating-drum Unit for Monitoring impactor, (DRUM) (Cahill et al., 1985;
Raabe et al., 1988), and the Multiple Orifice Uniform Deposit Impactor (MOUDI) (Marple et
al., 1986; Marple et al., 1991). Battelle-type impactors were also modified to add two size cuts
below 0.25 //m diameter. However, unlike the other four units, no certification of performance
has been published to date on its performance in the ultrafine region. The development of
reliable, clean adhesive coatings such as Apiezon™-L grease was also a major advance in the
field (Wesolowski et al., 1977; Cahill, 1979), allowing separation of abundant soils from
ultrafine size ranges even in dry, dusty conditions. For nominally PM-10 soils, for example, a
ratio of coarse to ultrafine soils was measured at 6,600:1 at a temperatures above 30 °C and
relative humidity below 20% (Cahill et al., 1985). Performances and specifications of all these
units is included in a recent review paper (Cahill and Wakabayashi, 1993)
It is important to mention, however, that the motivation for development of this ultrafine
capability was not for extensive studies of ultrafine metals, but rather to get a more complete
picture of the accumulation mode behavior of sulfates, nitrates, organics, and other major
components of the fine aerosol mix. Thus, compositional analysis was often limited to these
species even when suitable samples had been collected. For example, many LPI samples were
collected on stainless steel substrates, ideal for combustion analysis of sulfur, but unsuitable for
analysis of transition metals by X-ray techniques.
6.9.4 Observations of Very Fine Metals
Few techniques exist for collecting particles below 0.1 //m diameter for chemical analysis.
No compositional data was found for particles below 0.1 //m diameter. However, since ultrafine
6-192
-------
particles rapidly grow into the accumulation mode, it may be assumed that measurements of the
small-size tail of the accumulation mode provide some insight into the composition of the
ultrafine particles. Thus, the concentration of metals in the smallest available size-cut will be
examined. In order to avoid problems with definitions, particles in the smallest size-cut, which
may extend to diameters above 0.1 //m, will be called "very fine" and ultrafine will be reserved
for particle distributions with a mass mean diameter below 0.1 //m.
6.9.4.1 Stack and Source-Enriched Aerosols
Observation of very fine metals in source or source-enriched situations lessens problems
with dilution of the sample and identification of the source. This eases both particle collection
and analysis. Figure 6-96 shows the results of such a study on a coal fired power plant
(Maenhaut et al., 1993) using the Berner Low Pressure Impactor (BLPI). The extreme
volatilization of selenium is clearly seen, which is also confirmed in aircraft sampling of power
plant stacks. Note, however, that the enrichment factor, as a function of particle size, for both
sulfur and its chemical analog selenium. More refractory elements, on the other hand, are
strongly enhanced in the very fine particles as compared to coarser modes.
The BLPI cuts are as follows: Stage number 1-0.011 //m diameter, 2-0.021, 3-0.032,
4-0.07, 5-0.17, 6-0.30, 7-0.64, 8-1.4, 9-2.6, 10-5.5, 11-10.7 //m. All are for particle density 2.45
g/cm3 and a temperature 120 °C, the conditions of stack sampling in the coal fired power plant.
Both these figures were normalized to Earth crustal averages. Thus, even a two order of
magnitude rise in the normalized concentration may not result in a visible "combustion mode"
since the mass of soil falls very rapidly as one moves towards very fine particles. This is exactly
what is predicted by the results of Natusch et al. (1974). Thus, source testing
6-193
-------
1,000q
100
c
.2
c
o
u
•a
I
n
E
o
10
1:
0.1
0.1 xSe
-«-
S
-e-
Ca
-K-
Al
-t-
Si
1,000q
23456789 10
Stage number
34567
Stage number
10
Figure 6-96. Average normalized concentrations as a function of stage number, for
selenium (Se), sulfur (S), calcium (Ca), aluminum (Al), silicon (Si),
potassium (K), molybdenum (Mo), tungsten (W), nickel (Ni), and chromium
(Cr) for five BLPI samples from a coal fired power plant. The smallest size
mode is to the left, Stage number 1, 0.011 to Stage number 11,10.7 //m
diameter. Normalization is to average crustal composition.
Source: Maenhaut et al. (1993).
confirms nucleation theory and the laboratory studies and predicts emissions of metals in the
very fine particle size range from many types of high temperature combustion sources.
6.9.4.2 Ambient Aerosols
Direct Observations
Because of the difficulties in sampling and analysis, there is relatively little information on
the concentrations of very fine metal particles in ambient air. Some quantitative determinations
of ambient concentrations have become available in the past 15 years, however, generally as a
result of a number of short but intensive aerosol studies. Examples include the extensive studies
6-194
-------
near the Grand Canyon National Park (NP) in 1979 (Macias et al., 1981) to the Mohave Studies
near the Grand Canyon NP in 1993 and the Southern is California Air Quality Study (SCAQS)
in 1985-1987 (Hering et al., 1990; Cahill et al., 1990; Cahill et al., 1992a); studies at
Shenandoah NP in 1991 (Cahill and Wakabayashi, 1993) and Mt. Rainier NP in 1992 (Malm et
al., 1994a; Cahill and Wakabayashi, 1993), and others. While almost all of these studies used
several different types of impactors with ultrafine capabilities, relatively few were analyzed for
trace metal content.
An example of very fine particles persisting in ambient air is shown in Figure 6-97 using
data collected at Grand Canyon NP 1984 (Cahill et al., 1987). The very fine particles behave
independently from the accumulation mode, in fact often showing a net anti-correlation in
concentrations of sulfur as well as dramatic differences in metals (Table 6-7). The very fine
particles in Table 6-8 can be attributed to non-ferrous metal smelting activities in the region
(Eldred et al., 1983; Small et al., 1981), which puts the nearest important sources a hundred
miles away from the sampling site. The completely different behaviors of the accumulation and
very fine particles in this arid site also show that mis-sizing by particle bounce is not significant.
Table 6-8 presents a summary of more recent data for major EPA-regulated metals (lead,
nickel) and other metals, at Long Beach, CA, December in 1987 (Cahill et al., 1992a) and at
Shenandoah NP in 1991 (Cahill and Wakabayashi, 1993). The elements span the range from
refractory metals like nickel and vanadium to metals with low melting temperatures such as zinc
and lead. These data were all taken with the same unit, the Davis Rotating-drum Unit for
Monitoring (DRUM) using greased stages and a single orifice impactor (Cahill et al., 1985). The
last two stages were modified form the Gand Canyon configuration as a result of theoretical and
laboratory studies (Raabe et al., 1988), yielding 0.069 to 0.24 //m for Stage 8, and 0.24 to
0.34 //m diameter for Stage 7.
The DRUM data were used for several reasons: the DRUM's slowly rotating greased stages
have a documented ability to handle large amounts of coarse, dry soils without contaminating
the very fine stages (Cahill et al., 1985; Cahill and Wakabayashi, 1993), the elemental data are
of unprecedented sensitivity for ambient very fine trace metals (PIXE and
6-195
-------
l_
«
E
o
3
U
^
Q.
nograms
n
z
400
200
400
200
400
200
400
200
400
200
Stage 8
-
Stage 7
-
Stage 6
?Lji/u
Stage 5
~^n „
Stage 4
0.088 - 0.15 um diameter (very fine)
. S\. 1
0.15 - 0.24 um diameter
0.24 - 0.34 um diameter
__T L
^/uijiju M,- "vrnAT1
0.34 - 0.56 um diameter
t^^^^J-H-^^ j-w^,^
0.56 - 1.15 um diameter
^^mr\^f\j—^jj~~\j**—*L.
n 1 1 ^"S_i n~^~ ">—
n
K
^ilV^JL
rAl-rlJ1-Ln jO.
10
20
30
August 1984
Figure 6-97. Fine and very fine sulfur at Grand Canyon National Park, summer 1984.
The sulfur peaks on August 15 and August 16 were used for the
compositional analysis in Table 3. The first three cut points are somewhat
uncertain due to altitude and flow rate corrections. Final stage
configurations are given in Raabe et al. (1988), which were used for all later
studies using the DRUM.
Source: Cahill et al. (1987).
synchrotron-XRF), there is a consistency of sampler type and protocols at very different
locations, and there are more trace element data from the DRUM than from any other type of
unit. These advantages outweigh its disadvantages; the DRUM does not have the very fine
sizing detail of either the LPI or BLPI impactor, or the ability to measure mass, ions and organic
matter of the MOUDI or BLPI. The analyses were done both by PIXE and by synchrotron-XRF
(Cahill et al., 1992a), with most of the trace metal data from the latter
6-196
-------
TABLE 6-7. COMPOSITION OF THE AEROSOLS PRESENT AT GRAND CANYON
NATIONAL PARK IN THE SUMMER OF 1984 FOR THE SULFATE EPISODES OF
AUGUST 15 (ACCUMULATION MODE, STAGE 6) AND AUGUST 16
(VERY FINE PARTICLES, STAGE 8)
Elements
Sodium
Silicon and Aluminum
Sulfur
Chlorine
Potassium
Calcium
Titanium
Vanadium
Iron and Nickel
Copper
Zinc
Arsenic
Bromine
Lead
Stage 8,
0.088-0.15 //m
(ng/m3)
420
8
204
208
59
150
2
2
2
100
931
13
2
63
Stage 6,
0.24-0.34 //m
(ng/m3)
10
6
392
5
3
5
4
3
2
1
2
2
2
4
Source: Cahill et al. (1987).
source. In order to obtain sulfate, multiply sulfur by 3.0. These average values, however, obscure
a great deal of structure as a function of time.
The variability as a function of size and time is shown in Figure 6-98 for nickel, selenium,
and lead in Long Beach, CA as part of the SCAQS studies of 1987. By 1987, much of the lead
was no longer automotive, and there are significant changes in the very fine fraction over
periods of four to twelve hours. Note the behavior of very fine metals; almost total absence for
selenium, partial absence for nickel, and constant presence for lead. Almost all elements at
almost every site show similarly complex behavior. Thus, the summary of Table 6-8 can include
only the most basic types of information on fine and very fine metals in the atmosphere.
6-197
-------
TABLE 6-8. MEASUREMENTS OF FINE AND VERY FINE METALS
Site Name
Duration
Frequency
Dates
Particle
Aerodynamic
Diameters
Very Fine Particles
Accumulation Mode
Long Beach, CA
6 days
6 samples/day
(11, 12/87)
Mean detectable
limit - 0.3 ng/m3
Shenandoah NP
21 days
6 samples/day
(9/91)
Mean detectable
limit - 0.15 ng/m3
( ae=-"m)
Element
Vanadium
Nickel
Zinc
Selenium
Lead
Sulfur3
Vanadium
Nickel
Zinc
Selenium
Lead
Sulfur3
From
To
0.069
0.24
Maximum
Values
(ng/m3)
6.6
3.4
51
MDL
199
1.2
1.2
3.8
2.7
50
From
To
0.069
0.24
2.5
1.3
17.6
MDL
71.4
200
0.24
0.58
1.42
0.14
5.38
334
From
To
0.24
0.34
Mean
6.1
4.4
46.3
0.32
47.6
250
0.67
0.48
2.16
0.11
5.49
929
From
To
0.34
0.56
From
To
0.56
1.15
From
To
1.15
2.5
Values (ng/m3)
10.5
7.7
140.4
3.0
59.9
350
0.52
0.13
2.60
0.52
3.01
1235
12.2
4.5
189.4
1.4
69.9
500
0.30
0.03
1.92
0.35
10.87
1727
8.6
0.5
39
0.65
25.4
250
0.80
0.01
1.66
0.14
16.06
101
Estimated from graphs.
Source: Cahill et al. (1992a, 1996a).
In addition to the limited US data, comparison data have also become available from foreign
sources such as from the Kuwaiti oil fires (Reid et al., 1994) and a study in Santiago, Chile
(Cahill et al., 1996). While the former is a unique situation, the Santiago data are
6-198
-------
20.0.
1 0.0.
20.0.
0.0-
20.0-
ij 20.0.
a
S
" o.oi
-------
especially useful since leaded gasoline is still routinely used in Chile and other countries,
generating data impossible to obtain in the United States. Table 6-9 summarizes some of these
data for a refractory element, nickel, and a volatile metal, lead. However, the full data set
includes 450 samples of four to six hours duration, each analyzed in five fine size fractions,
generally with about 20 elements found in each fraction, or approximately 40,000 individual
elemental values.
Some general observations can be made from the data; first, there is an enormous
variation in the concentration of fine and very fine metals, sometimes spanning 4 or 5 orders of
magnitude in a few days. Such behavior can be modeled by plumes of particles that sweep over
the site episodically, as opposed to area or regional sources. Second, one often finds a mixture
of very fine particle or nuclei mode behavior as well as accumulation mode behavior. However,
these modes may be physically separated in time.
Lead in the United States follows a variety of very different patterns. In the rural
samples, lead tends to be bimodal, with a coarse component above 1.0 //m diameter and a very
fine component below 0.34 //m diameter. This can be modeled by a very fresh fine particle
mode and a coarser mode associated with resuspended soil. Urban sites, however, both in the
U.S. and in Santiago, show lead in very fine particles as well as in the accumulation mode. Lead
in resuspended soil is found in the coarse particle mode.
Other metals at Long Beach, however, lack a distinct concentration of very fine
particles all the time (selenium) or part of the time (nickel), merely possessing an accumulation
mode that closely mimics sulfates and other secondary species (Cahill et al., 1990). It is well
known that nickel and vanadium were derived from high temperature combustion sources, and
since each is highly refractory, they will occur primarily as very fine particles near the source.
Thus, the similarity between the distributions of these elements and less refractory elements such
as zinc can be understood through a rapid condensation and coagulation of the abundant
secondary species around these metals, leading to an accumulation mode distribution as the
secondary acidic species hydrate. Clearly, such processes are weaker at dry sites such as the arid
west in summer (Table 6-8). On the other hand, Shenandoah NP has a mixture of urban and
rural behavior, with occasional sharp peaks of very fine metals (nickel) superimposed on an
accumulation mode behavior (sulfur, selenium) with some coarse contribution (lead,
6-200
-------
TABLE 6-9. MEASUREMENTS OF FINE AND VERY FINE METALS
(LEAD AND NICKEL)
Site
Duration
Frequency
Dates Element
Long Beach Lead
6 days
4 samples/
day (11/87) Nickel
Shenandoah NP Lead
21 days
6 samples/
day (9/91) Nlckel
Mt. Rainier NP Lead
28 days
6 samples/
day (7, 8/92) Nlckel
Santiago, Chile Lead
14 days
6 samples/
day (9/93)
Kuwait Lead
14 days
4 samples/
day (6/91) Nickel
Particle
Aerodynamic
Diameters
(Dae, ,um
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Mean
Maximum
Very
Fine
Particles
From
To
0.069
ng/m3
71.4
199
1.3
3.4
5.4
50
0.58
1.2
2.3
6
Always less
MDL
101
920
429.9
2580
1.5
5
Accumulation Mode
From
To
0.24
0.34
ng/m3
47.6
95
4.4
11.4
5.5
20
0.48
1.6
6.5
From
To
0.34
0.56
ng/m3
59.9
129
7.7
15.0
3.0
16
0.13
0.8
2.0
15 21
than MDL
0.4
53
340
154.2
580
2.5
18
0.8
38
320
84.7
128
4.3
11
From
To
0.56
1.15
ng/m3
69.9
164
4.5
13.4
10.9
70
0.03
1.0
3.4
14
0.4
108
640
44.7
86
3.7
8
From
To
1.15
2.5
ng/m3
25.4
58
0.5
3.7
16.1
130
0.01
0.14
6.7
29
0.7
41
270
38.1
70
6.0
9
MDLa
0.45
0.22
0.2
0.09
0.5
0.07
8
0.35
0.22
aMDL = minimum detectable limit at 95% confidence level, in nanograms per cubic meter
Source: Cahill et al. (1992a,b, 1996a), Malm et al. (1994a), Reid et al. (1994), Cahill and Wakabayashi (1993).
6-201
-------
vanadium). Only through a detailed study of meteorology and knowledge of emission sources
can such ambient behavior be understood.
Indirect Methods
Lacking a large body of direct data on very fine metallic aerosols, there are indirect ways
to increase our knowledge of such aerosols;
1. Combustion studies have established the formation mechanism of very fine metallic
aerosols, and,
2. Considerable ambient data exist that, when combined with known combustion
processes, yield estimates for the concentration of very fine metallic aerosols by time
and locations.
3. In conditions of low ambient concentrations of particles and low humidity, very fine
particles have been shown to persist for many hours. (Cahill et al., 1985).
Thus, the numerous observations of fine (Dp < 2.5 //m) metallic aerosols in low humidity
conditions can yield estimates of the presence of such metals in the very fine particles and set
upper limits on their concentrations. The relatively small number of actual measurements can
then serve as tests or as confirmation of our level of understanding of these biologically
important aerosols. As an example, Figure 6-99 shows concentration profiles of sulfur,
selenium, zinc, and arsenic, all of which can occur as very fine particles in the western United
States. Arsenic and zinc are annual averages, March, 1993 to February, 1994, while the sulfur
(for sulfate, times 3.0) and selenium are for summer, 1993. This was done to exhibit the
correlation of these elements, which are chemically akin, during the eastern U.S. sulfate
maximum each summer. The regional nature of the elements is very evident, as are certain
strong sub-regional sources such as the copper smelter region of Arizona and New Mexico
(arsenic).
The non-urban values shown in Figure 6-99, which are derived from the cleanest areas of
the United States, are surprisingly relevant to urban areas in the same region for some of the
species. Table 6-10 compares major and minor fine elements at Shenandoah NP, where there
are detailed measurements of particle size, and Washington, DC, where such size information is
lacking. Summer 1993 is the comparison period. Finally, two western sites are compared, both
downwind of Los Angeles; San Gorgonio Wilderness, and Grand Canyon NP.
6-202
-------
Figure 6-99. Patterns of zinc, arsenic, sulfur, and selenium in the United States.
-------
TABLE 6-10. COMPARISON OF SELECTED SPECIES AT SHENANDOAH
NATIONAL PARK; WASHINGTON, DISTRICT OF COLUMBIA;
SAN GORGONIO WILDERNESS, CALIFORNIA; AND
GRAND CANYON NATIONAL PARK DURING SUMMER 1993
Shenandoah Washington, San Gorgonio Grand Canyon
Concentration (//g/m3) National Park DC Wilderness National Park
Mass-PM10 31.00 34.90 21.70 9.37
Mass-PM25 22.50 26.50 10.30 4.50
Composition - PM2 5
Ammonium sulfate 11.80 14.60 2.55 1.09
Ammonium nitrate 0.40 1.47 4.44 0.25
Organic matter 2.84 5.42 3.88 1.22
Soil 1.41 1.55 0.86 0.63
Trace compositon (ng/m3)
Nickel 0.24 0.97 0.18 0.09
Copper 1.06 3.37 0.76 0.30
Zinc 7.93 13.90 3.72 0.63
Arsenic 0.22 0.56 0.16 0.18
Selenium 1.58 2.48 0.44 0.18
Bromine 2.14 4.18 3.67 2.11
Lead 2.17 4.48 1.36 0.51
Bio-smoke tracer 8.33 < 2.00 10.00 32.30
(non-soil fine potassium)
Optical Absortion 19.60 41.90 13.90 5.40
(b(abs), 10'6m'')
Source: Malm et al. (1994b).
Inhalation of Very Fine Metals
An extensive literature exists on the deposition of fine metals in the human lung, much of
which was derived from laboratory studies, some using radioactive tracer isotopes. But an
example of one of the few direct measurements of lung capture of ambient very fine metals is
found in Desaedeleer et al. (1977) and shown in Figure 6-100. The lower cut point is only
0.25 //m, but even so, the increased capture efficiency of the lung for very fine and very fine
particles is clearly shown.
6-204
-------
£.
1
^
O
!a
E
< 0.5
^
To
£
X
UJ
B B
.2 0.2
+J
(0
0£
^
o
'1 °-1
c
0
O
o 0.05
o
0.02
i i i i i i
-r TT ~
— ti •! T ~r
TT * i\ -
- S ^fi \
^^^ \
~ -|- ,- \ -
mr ^JT \ T
— -H^JB^ Ti
"^E, I ^~
P
i ^
\ -
^P ^L
• ^k
^ \
%
J\
^
, * ~
— *
- A A Pb -
• n Br V
_ » Cl * .
— —
i i i i i i
0
20
40
60
80
90
92
94
96
98
65432 1
c
o
j?
4, 4
to 2, 2 to 1,1 to 0.5, 0.5 to 0.25, and < 0.25 fj,m particles of size classes 1
through 6, respectively. Extension of the curve to particles of diameter
>2 fj.m (classes 2 and 1) is supported by separateexperiments using chalk
dust aerosol.
Source: Desaedeleer et al. (1977).
6.9.5 Conclusions
There are few data on ambient concentrations of ultrafine metals. The few direct
measurements can be extended with some confidence using indirect methods; i.e., particle
counting techniques that have size information but no chemical information, or filter collection
methods that have limited size information but detailed compositional information.
6-205
-------
Nevertheless, it is clear that more information is needed on the size and concentration and the
spatial and temporal concentration profiles of ultrafine metal particles.
Ultrafine metals are produced by a wide variety of anthropogenic activities and emitted
into the ambient air. Ambient concentrations of such metals have been seen not only in urban
settings but also at the cleanest sites in the United States. Concentrations are highly variable as a
function of site and time. While ultrafine metals have been seen to persist for many hours, or
more, in the clean, dry environment of the arid west, they appear to be rapidly transformed into
the accumulation mode in polluted urban or humid rural sites.
6.10 FINE AND COARSE PARTICULATE MATTER TRENDS
AND PATTERNS
Data for characterizing PM10 are available from a number of AIRS sites across the country.
However, data for characterizing PM2 5 and PM(10_2 5) as well as PM10 are not readily available.
As discussed in 6.3.1.7, data for PM2 5 and PM(10_2 5) have been obtained at sites in the
IMPROVE/NESCAUM networks. However, these sites are located in uninhibited areas.
Measurements suitable for determining trends and patterns of PM2 5 and PM(10_2 5) in populated
areas are available from only a few sites.
Most such data have been obtained with dichotomous samplers which measure PM2 5 (an
indicator of fine mode particles) and PM(10.25) (an indicator of the coarse fraction of PM10).
These two fractions may be added together to give PM10. PM2 5 is sometimes referred to as fine
and PM(10.25) as coarse although it is understood that PM2 5 will contain that fraction of the coarse
mode PM below 2.5 //m diameter and neither PM10 nor PM(10.25) will contain that portion of the
coarse mode above 10//m diameter. Sources of PM25 (fine) and PM(10.25) (coarse) data include
EPAs Aerometric Information Retrieval System (AIRS) (AIRS, 1995), IMPROVE (Eldred and
Cahill, 1994; Cahill, 1996), The California Air Resources Board (CARB) (CARB, 1995), the
Harvard Six-Cities Data Base (Spengler et al., 1986b; Neas, 1996), and the Harvard Philadelphia
Data Base (Koutrakis, 1995). The Inhalable Paniculate Network (IPN) (IPN, 1985; Rodes and
Evans, 1982) provides TSP, PM15 and PM25 data with only a small amount of PM10 data.
Data suitable for characterizing the daily variability in PM2 5 and PM10 are available from
only one site in southwestern Philadelphia. The National Weather Service provides daily
6-206
-------
observations of visual range, which when suitably treated, can provide an indication of fine
mode particle concentration. The Harvard Six Cities study obtained data for PM2 5 and PM15
every other day for several years. The California Air Resources Board operates about twenty
sites that collect PM25 and PM(10_25) data with a sampling frequency of every sixth day. Every
sixth day data for a few sites may be found in AIRS. Because of the small number of data sets
for PM2 5 and either PM(10.25) or PM10 levels detailed intercomparisons of the behavior of these
aerosol size fractions in different regions of the United States cannot yet be made. Data for
characterizing the daily and seasonal variability of PM2 5, PM(10_2 5), and PM10 will be discussed
in 6.10.1, the longer term variability (i.e., trends) of PM25, PM10_25, and PM10 will be discussed
in 6.10.2, and the interrelations and correlations among the various PM components and
parameters will be discussed in 6.10.3.
The results presented in this section were derived from data bases available to the public.
Except for the visibility and National Park trend data, the results presented in this section were
prepared for this Criteria Document and have not yet been published elsewhere.
6.10.1 Daily and Seasonal Variability in PM25 and PM10
In addition to considering patterns of seasonal variations over broad geographical areas, a
great deal of information, useful for relating ambient concentrations to health effects, can be
obtained by analyzing long time series of concentration data at a single site. Collocated 24-hour
PM2 5 and PM10 filter samples were collected at a site in southwestern Philadelphia from
May 1992 through April 1995 (Koutrakis, 1995). This unique data set was collected on a nearly
daily basis, thereby allowing an assessment of day-to-day variability in aerosol properties.
The data are presented as box plots showing the lowest, lowest tenth percentile, lowest
quartile, median, highest quartile, highest tenth percentile, and highest PM2 5 values in
Figure 6-101. The four three-month averaging periods shown (March-May, June-August,
September-November, December-February) correspond to the so-called climatological or
meteorological seasons. Highest median (20.8 |ig/m3) and extreme (72.6 |ig/m3)
6-207
-------
ou -
70 -
-^ 60 -
S
2 50 -
c
c
o
o 30 -
0
o
20 -
10 -
n -
\
<
L
1
»
J
i
•
f
<
L
•
•
1
»
J
•
i
Philadelphia - PBY site
PM2.5
(n = 1024)
•
[
«
L
1 1
» <
T 1
i
•
I
»
I
Mar-May
J u n - A u g
Sep-Nov
Dec-Fob
Figure 6-101. Concentrations of PM25 measured at the PBY site in southwestern
Philadelphia. The data show the lowest, lowest tenth percentile, lowest
quartile, median (black circles), highest quartile, highest tenth percentile,
and highest PM2 5 values.
PM25 concentrations were found during summer, with a difference of 50 |ig/m3 between them.
Median PM2 5 concentrations are 14.6, 14.2, and 13.4 |ig/m3 for the three quarterly periods from
September through May, while maximum concentrations ranged from 41 to 55 |ig/m3.
Corresponding PM10 data are shown in Figure 6-102. PM10 concentrations exhibit strong
maxima during both the summer (82.4 |ig/m3) and winter (77.5 |ig/m3). Maximum PM10
concentrations during spring and fall are 54.7 and 58.5 |ig/m3. The difference between median
and maximum values was 54.4 |ig/m3 during summer and 58.3 |ig/m3 during winter. The median
PM10 concentration was 28.0 |ig/m3 in summer, and ranged between 19.2 and 20.9 |ig/m3 during
the other seasons.
PM25 and PM10 concentrations were highly correlated (r=0.92). PM10 and PM(10_25)
concentrations were less highly correlated (r=0.63) and PM2 5 and PM(10_2 5) concentrations were
even less well correlated (r=0.30). The day-to-day difference in PM25 concentrations was 6.8 ±
6.5 |ig/m3 and the maximum difference was 54.7 |ig/m3, while the day-to-day
6-208
-------
a>
o
O
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
Philadelphia - PBY site
•
- [
<
1
•
[
1
' [
PM10
(n = 1024)
•
I
I
I '
m
•
1 1
»
J \
m
m
m
1
»
I
Mar-May
Jun-Aug
Sep-Nov
Dec-Feb
Figure 6-102. Concentrations of PM10 measured at the PBY site in southwestern
Philadelphia. The data show the lowest, lowest tenth percentile, lowest
quartile, median (black circles), highest quartile, highest tenth percentile,
and highest PM2 5 values.
difference in PM10 concentrations was 8.6 ± 7.5 |ig/m3 with a maximum difference of
50.4 |ig/m3. The day-to-day difference in PM(10.25) concentrations was 3.7 ± 3.5 |ig/m3 with a
maximum difference of 3 5.1 |ig/m3. The ratio of PM2 5 to PM10 throughout the measurement
period was 0.71 ± 0.13. The high correlation coefficient between PM25 and PM10 along with
the high ratio of PM2 5 to PM10 suggests that variability in PM2 5 was driving the variability in
PM10 levels.
Frequency distributions for the entire three-year PM2 5, PM(10.2 5), and PM10 data sets are
shown in Figures 6-103, 6-104, and 6-105, respectively. Concentrations predicted from the log-
normal distribution, using mean values and geometric standard deviation derived from the data,
are also shown. The small number of apparently negative PM(10.25) values reflects measurement
error at low concentration levels. Frequency distributions of aerosol concentrations at several
sites in the South Coast Air Basin (Kao and Friedlander, 1995) have also been shown to be
reasonably approximated by log-normal distributions.
6-209
-------
35O
PM2.5
geometric mean = 15.2 pg/nrT
og= 1 .69
1 O
2O
3O
4O
SO
6O
7O
8O
Concentration (|jg/m )
Figure 6-103. Frequency distribution of PM25 concentrations measured at the PBY site in
southwestern Philadelphia. Log-normal distribution fit to the data shown
as solid line.
45O
4OO
35O --
o> 3OO -t-
ta
tn
2SO --
200 --
1 SO --
1 OO
SO +
-I-
O 1O 2O 3O 4O SO 6O 7O
Concentration (|jg/m3 )
Figure 6-104. Frequency distribution of coarse mode mass derived by difference between
PM10 and PM2 5. Log-normal distribution not shown because of derivative
nature of entries.
6-210
-------
25O
ZOO --
PM
1O
geometric mean = 21.4 pg/m
og= 1.66
(A
O>
a. 150 --
CO
"o
1OO --
SO --
10
20
30
40
50
60
70
80
Concentration (|jg/m )
Figure 6-105. Frequency distribution of PM10 concentrations measured at the PBY site in
southwestern Philadelphia. Log-normal distribution fit to the data shown
as solid line.
In general, the highest PM2 5 values are observed when winds are from the southwest
during sunny but hazy high presure conditions. In contrast, the lowest values are found after
significant rainstorms during all seasons of the year. The highest PM2 5 values were observed
during episodes driven by high sulfate abundances and are due, at least partly, to higher sulfate
concentrations. Correlation coefficients between SO=4 and PM2 5 were 0.97 during the summer
of 1993. Similar correlations between SO4 and PM2 5 were found at a site in northeastern
Philadelphia (24 km distant from the site under discussion) during the summer of 1993.
In addition, PM25 was found to be stongly correlated (r > 0.9) between seven urban sites and one
background site (Valley Forge, PA) during the summer of 1993 (Suh et al., 1995). The same
relations were also found during the summer of 1994 at four monitoring sites as part of a
separate study (Pinto et al., 1995). The results from these studies strongly suggest that PM2 5 and
SO4 concentrations are spatially uniform throughout the Philadelphia area, and that variability
in PM10 levels is caused largely by variability in PM2 5 (Wilson and Suh, 1996). However, not
enough data are available from regional sites to define the total areal extent of the spatial
6-211
-------
homogeneity observed in the urban concentrations.
Different conclusions could be drawn about data collected elsewhere in the United
States. PM2 5 and PM(10_2 5) data were obtained at a number of sites in California on a sampling
schedule of every six days with dichotomous samplers (California Air Resources Board, 1995).
As an example, frequency distributions of PM25, PM(10_25), and PM10 concentrations (calculated
as the sum of PM2 5 and PM(10_2 5) obtained at Riverside-Rubidoux from 1989 to 1994 are shown
in Figures 6-106, 6-107, and 6-108, respectively. It can be seen that the data cannot be
satisfactorily fit by a single function, mainly as the result of the complexity of the concentration
distribution of the coarse size mode shown in Figure 6-107.
80
70-
60-
2 50-
Q.
E
5 40-
30-
20-
1O-
O 2O 4O 6O 80 1OO 12O 14O 1 6O 1 8O
Concentration (|jg/nri3)
Figure 6-106. Frequency distribution of PM25 concentrations measured at the Riverside-
Rubidoux site.
The data are also presented as box plots showing the lowest, lowest tenth percentile,
lowest quartile, median, highest quartile, highest tenth percentile, and highest PM2 5 values in
Figure 6-109 for four three-month averaging periods (January-March, April-June,
6-212
-------
80
70-
6O-
2 50H
S 40H
30-
20-
10-
0 20 40 60 8O 10O 12O 140 160 180
Concentration (pg/m3)
Figure 6-107. Frequency distribution of PM(10_25) concentrations measured at the
Riverside-Rubidoux site.
ou
40-
01
H. 30-
E
a
U)
"o
o 20-
z
10-
n —
n
-
i — ,
-
-
i — ,
-
n n n n n N l~l n
0 20 40 60 80 100 120 140 160 180
Concentration (\iglm3)
Figure 108. Frequency distribution of PM10 concentrations calculated as the sum of PM25
and PM(10_2 5) masses measured at the Riverside-Rubidoux site.
6-213
-------
160
140-
100-
so-
60-
40 -
20 -
Riverside-Rub id oux
Fine
(n = 382)
n.
T
Y
Jan - Mar
1st Qtr
Apr -Jun
2nd Qtr
Jul - Sept
3rd Qtr
Oct - Dec
4th Qtr
Figure 6-109. Concentrations of PM25 measured at the Riverside-Rubidoux site. The data
show the lowest, lowest tenth percentile, lowest quartile, median (black
squares), highest quartile, highest tenth percentile, and highest PM2 5 values.
July-September, October-December). Data for PM(10_2 5) and reconstructed PM10 are similarly
plotted in Figures 6-110 and 6-111. As can seen from these figures, variability in concentrations
within an averaging period is high. Differences between median and maximum PM2 5 levels
range from 40 |ig/m3 during the spring to 123 |ig/m3 during the winter, while differences
between median and maximum PM(10_2 5) levels range from 23 |ig/m3 during winter to 83 |ig/m3
during summer. Variations in both size fractions combine to yield differences between median
and maximum PM10 levels ranging between 83 |ig/m3 and 136 |ig/m3. Median PM25 levels do
not show a clear seasonal cycle. However, PM(10_2 5) concentrations show a maximum during
the summer which causes a weak maximum in PM10 levels. In fact, median PM2 5 (30 |ig/m3)
and PM(10_2 5) (34 |ig/m3) levels are identical during the spring and fall quarters. The ratio of
PM2 5 to PM10 mass throughout the measurement period was 0.48 ±0.13 and PM2 5 and PM10
levels were moderately correlated (r = 0.47).
An examination of the data from Philadelphia, PA and Riverside, CA indicates that
substantial differences exist in aerosol properties between widely separated geographic
6-214
-------
Riverside-Rub id oux
nto -
120 -
£" 100 -
E
T 80 -
0
2
1 60 -
u
o
O
40 -
20
Coarse
(n = 382)
n
•
^
r
•
]
i
•
^
0 I I
Jan - Mar Apr - Jun Jul
1st Qtr 2nd
T
n f\
IL
T
I y
i i
- Sept Oct - Dec
Qtr 3rd Qtr 4th Qtr
Figure 6-110. Concentrations of PM(10_25) measured at the Riverside-Rubidoux site. The
data show the lowest, lowest tenth percentile, lowest quartile, median (black
squares), highest quartile, highest tenth percentile, and highest PMcoarse
values.
200
Riverside-Rubidoux
150-
o
a 100
o
O
50 -
PM10
(n = 382)
n
n
A
Jan - Mar
1st Qtr
Apr -Jun
2nd Qtr
Jul - Sept
3rd Qtr
Oct - Dec
4th Qtr
Figure 6-111. Concentrations of PM10 measured at the Riverside-Rubidoux site. The data
show the lowest, lowest tenth percentile, lowest quartile, median (black
squares), highest quartile, highest tenth percentile, and highest PM10 values.
6-215
-------
regions. Fine mode particles make up most of the PM10 mass observed in Philadelphia and
appear to drive the daily and seasonal variability in PM10 concentrations there. Coarse mode
particles are a larger fraction of PM10 mass in Riverside and drive the seasonal variability in
PM10 seen there. The range in the seasonal variation of the ratio of PM2 5 to PM10 mass is much
smaller in Philadelphia (0.70 to 0.75) than in Riverside (0.41 to 0.57) for the four averaging
periods used. Differences between median and maximum concentrations in any size fraction are
much larger at the Riverside site than at the Philadelphia site. Many of these differences could
reflect the more sporadic nature of dust suspension at Riverside. These considerations
demonstrate the hazards in extrapolating conclusions about the nature of variability in aerosol
characteristics inferred at one location to another.
6.10.2 Fine and Coarse Particulate Matter Trends and Relationships
6.10.2.1 Visual Range/Haziness
Observations of visual range, obtained by the National Weather Service and available
through the National Climatic Data Center of the National Oceanic and Atmospheric
Administration, provide one of the few truly long-term, daily records of any parameter related to
air pollution. After some manipulation, the visual range data may be used as an indicator of fine
mode particle pollution. The data reduction process and analyses of resulting trends have been
reported by Husar et al. (1994), Husar and Wilson (1993), and Husar et al. (1981).
Visual range i.e., the maximum distance at which an observer can discern the outline of an
object, is an understandable and for many purposes an apporpriate measure of the optical
environment. It has the disadvantage, however, of being inversely related to aerosol
concentration. It is usual, therefore, to convert visual range to a direct indicator of fine mode.
particle concentration. The quantitative measure of haziness is the extinction coefficient, Bext,
defined as Bext=K/visual range, where K is the Koschmieder constant. The value of K is
determined both by the threshold sensitivity of the human eye and the initial contrast of the
visible object against the horizon sky. Husar et at. (1994) use K=1.9 in accordance with the data
by Griffing (1980). The extinction coefficient is in units of km"1 and is proportional to the
concentration of light scattering and absorbing aerosols and gases. The radiative transfer
characteristics which determine the visual range depend on time of day. Only local noon
observations are used.
6-216
-------
Haze Trend Summary
The U.S. haze patterns and trends since 1960 are presented in 16 haze maps that represent
four time periods and four seasons (Figure 6-112). The selected time periods are 5 year averages
centered at 1960, 1970, 1980, and 1990. The quarters are calendrical, i.e., winter is January,
February, and March. View horizonally for secular trends by quarter. View vertically for
seasonal variation by decade.
The overall national view shows two large contiguous haze regions, one over the eastern
U.S. and another over the western Pacific states. The two haze regions are divided by a
low-haze territory between the Rocky Mountains and the Sierra-Cascade mountain ranges. This
general pattern is preserved over the past 30-year period. However, notable trends have
occurred over both the western and eastern haze regions.
The haziness in the western Pacific states covers all of the coastal states, with California
having the highest values. In the 1960s a large fraction of western California was very hazy,
particularly during Quarters 1 and 4. By the 1990s the magnitude of the Pacific Coast haziness
has declined markedly for all seasons.
The eastern haze region extends from the East Coast to the Rocky Mountains. The western
boundary of the eastern haze region has been markedly constant over both the seasons and the
years. In fact, haze in the mid-section of the U.S., extending from the Rocky Mountains to the
Mississippi River, has changed little over the 30-year history.
The most dynamic pattern can be observed over the eastern U.S., extending from the
Mississippi River to the East Coast. The eastern U.S. shows a significant seasonal variation.
There is also a significant trend over the past 30 years. Furthermore, these seasonal and secular
(long-term) trends are different for sub-regions within the eastern U.S., such as the Northeast,
Mid-Atlantic and Gulf States regions.
In the 1960s, the highest extinction values were recorded for the cold season (Ql, Q4),
with significantly lower values for the warm quarters (Q2, Q3). The remarkable reduction in
haziness during the cold season and the strong increase during the warm season has shifted the
6-217
-------
figure 6-112. United States trend maps for the 75th percentile extinction coefficient, Bext for winter (Ql), spring (Q2),
summer (Q3), and fall (Q4). Bext [km'1] is derived from visual range, VR, data by Bext=1.9/VR. Data
obtained during natural obstructions to vision (i.e., rain, snow, fog) were eliminated.
-------
haze peak from winter to summer. This seasonal change has been accompanied by a regional
shift in highest haze pattern. In the 1960s, the worst haziness occurred around Lake Erie and the
New York-Washington megalopolis, during the cold season. By the 1990s the area with the
worst haze had shifted southward toward Tennessee and Carolinas and occurred in the summer
season.
The decade of the 1980s shows less change than the earlier decades. However, there has
been a continued haze reduction in the Northeast, north of the Ohio and east of the Mississippi
Rivers. The southeastern U.S. as well as the Pacific states remained virtually unchanged in the
1980s.
Regional Pattern
Trends for specific regions in the eastern U.S., and the number and location of visual range
reporting stations for each region, are shown in Figure 6-113. The trend graphs represent the
75th percentile of Bext for the stations located within the designated region. The trends are
presented for Quarters 1 (winter) and 3 (summer) separately. The northwestern U.S. exhibits an
increase of Quarter 3 haze between 1960 and 1970, and a steady decline between 1973 (0.22)
and 1992 (0.12). In the winter quarter the haziness has steadily declined from 0.15 to 0.10 in the
30-year period. The Mid-Atlantic region that includes the Virginias and Carolinas shows a
strong summer increase between 1960 and 1973, followed by a decline. The winter haze was
virtually unchanged over the 30-year period. The haziness over the Gulf states increased
between 1960 and 1970, and remained virtually unchanged since then. The central Midwest
including Missouri and Arkansas exhibit virtually no change during the winter season and a
slight increased in the summer (1960-1970). The upper Midwest (Figure 14) shows an opposing
trend for summer and winter. While summer haze has increased, mostly 1960-1973, the winter
haze has declined.
6.10.2.2 IMPROVE
The National Park Service-EPA monitoring network for Class I areas is designed to
monitor visibility in national parks and other designated areas. Most of these are remote.
However, data from two southeastern sites, Shenandoah National Park and the Great Smoky
6-219
-------
Si
Upper Midwest
Midwest
^ 0.28
ZL °-24
^ 0.20
S? 0.16
•R o.os
CD 0.04
West Gulf
Southeast
Mid Atlantic Northeast
Quarter 3
1M01M01M01I701M01MO 1»401«S018eOU701Be01WO 1M0116019601*70 ttMIMO 10401(601MO WTO IMG 1WO 1M01K01M01*n>1M01tM 1MO1M01«801»701«W 1«80
s*
to
^
o
0.28
0.24
0.20
0.16
0.08
0.04
-
Quarter 1
1MOigS01M01(701M01»0 1«4019601M0187019801WO 1M01«6019S01WBIBS01WO 1MO»M1M01«ni1M01WO 1M01tS01M01«7l)lgai>1WO 1M0196019M187D1M019M
Figure 6-113. Secular haze trends (1960 to 1992) for six eastern U.S. regions, summer (Ql) and winter (Q3)
-------
Mountains National Park, provide useful information on the regional background of sulfate
(Eldred and Cahill, 1994; Cahill et al., 1996b). As shown in Figure 6-114, there is a distinct
increase in sulfate. This increase can be correlated with increases in SO2 emissions in the
summer from power plants in the Tennessee Valley (Cahill et al., 1996b). The increased
emissions may be related to an increase in demand for power for air conditioning. The increase
in regional background will impact urban centers along the eastern U.S. Visibility
measurements over the northeastern U.S. show an increase in haze from 1960 to 1970 in both
winter and summer. Between 1970 and 1983, there was a decrease in haze in the winter but
little change during the summer (Husar and Wilson, 1993; Husar et al., 1994). Concern has been
expressed that the indicated trends may have been impacted, or even produced, by changes in
monitoring protocols (White, 1996a,b). However, these issues have been addressed by Cahill
etal. (1996b).
6.10.2.3 Philadelphia
Philadelphia is of special interest because of the extensive monitoring conducted there
and the use of Philadelphia data in epidemiological studies. Extensive measurements of TSP
have been conducted in Philadelphia. Several data sets have been combined to give an
indication of long-term trends in Philadelphia (Figure 6-115). The TSP data set was construed
from the AIRS database (Wyzga and Lipfert, 1996; Li and Roth, 1995). There was a steady
decrease in TSP from 1973 to 1983 with variable but slightly increasing TSP levels between
1983 and 1990.
Fine PM was estimated from the Inhalable Particle Network (Rodes and Evans, 1985)
from 1980 to 1983, from AIRS (AIRS, 1995), from 1987 to 1990, and from the Harvard Data
Base (Koutrakis, 1996) for 1993 and 1994. During the period 3/79 to 12/83, the Inhalable
Particulate Network conducted measurements in Philadelphia with dichotomous samplers.
These used 15 //m upper cut points except for a period at the end of the study (3/82 to 12/83)
when two co-located PM10 samplers were run at one site. The IPN data set allows construction
of four annual averages for 1980 through 1983 by averaging PM25 data from PM15/PM2 5
dichotomous samplers from the several IPN sites across Philadelphia. These are shown in
Figure 6-115, along with the one year of PM2 5 data from PM10/PM2 5 dichotomous samplers at
the South Broad St. site.
6-221
-------
11
10
8
CO
£ 7
O)
tf 6
4-1
I 5
CO
Sulfate Concentration Trends
1982
1984
1986
1988
1990
1983 1985 1987 1989 1991
Shenandoah Smoky Mountains
1992
1993
1994
Figure 6-114. Eastern U. S. regional background trend of sulfate indicated by seasonal
trend data from Shenadoah and Great Smoky Mountains National Parks.
A PM10/PM2 5 dichotomous sampler, run in the Philadelphia area from 1987 through
1990 allows annual averages of PM2 5 for those years to be added to Figure 6-115. Harvard
University measured PM10 and PM2 5 at the Presbyterian Home site from 5/92 to 5/92 allowing
annual averages for 93 and 94 to be added to the graph. Since PM2 5 is expected to be relatively
uniform across Philadelphia (Wilson and Suh, 1996), this data can be used to estimate a PM2 5
trend from 1979 to 1994. A downward trend is indicated.
The samplers were not at the same sites during the different time periods. Since PM(10_2 5)
does not seem to be uniform across Philadelphia (Wilson and Suh, 1996), no PM10 or PM(10.25)
trend could be constructed. Comparisons of PM10 and PM(10_2 5) and PM2 5/PM10 (Figure 6-116)
for 1983 and 1993 are shown. Differences in PM(10.25) and the ratio of
6-222
-------
D.
tn
95
90
85
80
75
70
65
60
55
TSP and PM2 5 Trends
IPN, AIRS, and Harvard Databases
\
30
25
20
15 i
i/
(
10 Q.
0)
19731 19751 19771 19791 1981 I 19831 19851 19871 19891 1991 I 19931
1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994
Year
D TSP + PM25, IPN Avg o PM25, IPN, SBROAD A PM25,AIRS x PM25, PBY
Figure 6-115. TSP and PM2 5 trend data for the city of Philadlphia from AIRS, IPN, and
Harvard database.
PM2 5/PM10 may represent geographical differences in the coarse fraction of PM10 as well as
relative changes in PM2 5 and PM(10_2 5).
6.10.2.4 Harvard Six-Cities Study
During 1979 to 1986, the Harvard School of Public Health measured particulate matter in
6 cities in eastern and central United States (Spengler et al., 1986b; Neas, 1996). Means and 90th
percentiles for fine, coarse, PM15, and TSP are shown in Figures 6-117 to 6-119. (Measurements
were made with dichotomous samplers with a 15 //m diameter cut point from 1979 to 1984 and
with a 10//m diameter cut point from 1984 to 1986. The coarse fractions of PM10 and PM15 were
not significantly different during the overlapping year.) In the dirtier cities, Steubenville, St.
Louis, and Harrison, there were decreases in all PM indicators, especially in the earlier years.
6-223
-------
to
to
V20
io
S. Broad, 1983
PM2.8 and PMnnj,,,
PHILADELPHIA
1
0.9
0.8
0.7
°-6
0.5
0.4
0.3
0.2
24-Jan-B3 I 25-Mar-e3 I 24-May-B3 I 23nJul-B3 I 21-Bep-83 I 2D-NOV-83 I
23-F>b-B3 24-Apr-83 23Jun-83 22^Aug-83 21-Oct-83 20-Dec-83
Date
PM2.t as a Fraction ol PM,0
PBY,1983
PM2.6 and
(C)
24-Jan-63 I 25-Mar-B3
24-May-S3 23-JUI-B3 I 21-Sap-E3 I 2D-NDV-B3
23-Feb-B3 24-Apr-B3 23^Jun-B3 22-Aug-B3 21-Oct-B3 20-Dec-83
Date
01JAN93 I 05MAR93 I D8MAY93 I D6JUL93 I DSBEP93 I 07NOV93
01FEB93 05APR93 06JUN93 06AUQ93 07OCT93 OBDEC93
Date
° PMa.B + PM,,,^..,
PM2.< as a Fraction ol PM,0
01JAN93 OSMAR93 I OSMAY93 I 06JUL93 06BEP93
07NOV93
01FEB93 05APR93 OSJUN93 OSAUC93 07OCT93 08DEC93
Data
Figure 6-116. Comparison of fine and coarse particle parameters in Philadelphia in 1983 and 1993: (a) PM25 and PM(10_25) at
South Broad St. site, 1983; (b) PM2/PM10at South Broad St. site, 1983; (c) PM25 and PM(10_25) at
Presbyterian Home site, 1993; (d) PM2 5/PM10 at Presbyterian Home Site, 1993.
-------
Stubenville
Harvard Six Cities Data
80
70
St. Louis
, 30
20
10
1979 1980
a PM2.5
181 1982 1983 1984
Year
* PM-C « PM15 » TSP
1980 1981 1982 1983
Year
D PM2.5 * PM-C • PM15
to
to
150
140
130
120
a 110
^ 100
4)
U 90
I BO
S 70
g 60
« " 50
75, *°
a 30
20
10
0
Stubenville
150
140
130
120
01 110
e 100
41
O 90
I BO
£ 70
S so
30
20
10
St. Louis
(d)
1980 1981 1982 1983 1984
Year
O PM2.5 + PM-C « PM15 » TSP
O PM2.5
>81 1982 1983
Year
* PM-C • PM15
Figure 6-117. Trend data from the Harvard Six-Cities Study: (a) Steubenville, fine, coarse, PM15, and TSP means;
(b) Steubenville, fine, coarse, PM15, and TSP 90th percentiles; (c) St. Louis, fine, coarse, PM15, and TSP
means; (d) St. Louis, fine, coarse, PM10, and TSP 90th percentiles.
-------
Harvard Six Cities Data
1979 1980
OPM2.5
181 1982 1983
Year
•PM-C «PM15
1984
A TSP
1980
1981
1982 1983
Year
OPM2.5 * PM-C «PM15
1984
A TSP
Oi
to
to
Oi
150
140
130
12°
g 100
a 90
,,-so
40
30
20
10
0
Harrlman
(b)
1979 1980
OPM2.5
>81 1982 1983 1984
Year
•PM-C «PM15 A TSP
150
140
130
120
= 110
U 100
u 90
°- 80
fj 70
". 60
"E so
5 40
30
20
10
0
Watertown
(d)
1980
1981
1982 1983
Year
DPM2.5 * PM-C «PM15
1984
» TSP
Figure 6-118. Trend data from the Harvard Six-Cities Study: (a) Harriman, fine, coarse, PM15, and TSP means;
(b) Harriman, fine, coarse, PM15, and TSP 90th percentiles; (c) Watertown, fine, coarse, PM15, and
TSP means; (d) Watertown, fine, coarse, PM1S, and TSP 90th percentiles.
-------
80
70
60
2 so
a
0)
* 40
"E
D> 30
20
10
0
Portage
Harvard Six Cities Data
801
(a)
1980
1981
1982 1983
Year
O PM2.5 * PM-C • PM1S
1984
• TSP
70
EO
i 50
0)
» 40
"E
"Si 30
20
10
0
Topeka
(c)
1980 1981
1982 1983
Year
O PM2.5 * PM-C « PM15 » TSP
1984 1985
to
to
01
a
0)
150
140
130
120
110
100
90
^ 80
O 70
O)
„" 60
•i so
3. 40
30
20
10
Portage
(b)
D PM2.5
181 1982 1983
Year
* PM-C » PM15
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
Topeka
(d)
PM2.5
1 1982 1983
Year
' PM-C »PM15
' TSP
Figure 6-119. Trend data from the Harvard Six-City Study: (a) Portage, fine, coarse, PM15, and TSP means; (b) Portage, fine,
coarse, PM1S, and total TSP 90th percentiles; (c) Topeka, fine, coarse, PM15, and TSP means; (d) Topeka, fine,
coarse, PM15, and TSP 90th percentiles.
-------
There was also an apparent decrease in Topeka, one of the cleaner cities. No trend can be
discerned in Watertown or Portage. It is difficult to determine whether there was a greater trend
in fine or coarse particles.
6.10.2.5 AIRS
The AIRS data base was searched for sites with 4 or more years of fine and coarse data
(AIRS, 1995). Five such sites were found. Values for the mean and the 90th percentile are
shown in Figures 6-120 to 6-123. No significant trends are evident in PM2 5 or PM(10_2 5) either in
the means or the 90th percentile values. PM10 and PM(10_2 5) at the dirtier site in New York City
do appear to have decreased from 1988 to 1992 but to have increased between 1992 and 1994.
6.10.2.6 California Sites
The California Air Resources Board conducted dichotomous sample measurements, every
sixth day, beginning in 1989 at a number of California sites (CARB, 1995). Some results from 8
sites are shown in Figures 6-124 to 6-130. The means (Panel a) and 90th percentile values
(Panel b) are given for PM2 5, PM(10_2 5), and PM10. Most of the sites show slight downward
trends for PM10 and both PM2 5 and PM(10_2 5).
The California sites are of special interest because of the substantial seasonal and daily
variability. The individual every-sixth-day values are plotted for 1991 (plus 1 day in the
preceeding and following years)(Panel c). Strong seasonal and daily variation are evident.
Based on the every-sixth-day measurements, it would appear that the day-to-day variability at
the California sites is higher than in Philadelphia. Also shown is the PM2 5 fraction of PM10
(Panel d). These ratios are also show a strong seasonal variation.
6.10.3 Interrelations and Correlations
The availability of data on four PM size fractions at several sites for a number of years
makes it possible to examine relationships and correlations among PM2 5, PM(10_2 5), PM10, and
TSP. It is also possible to examine the distribution of values in the upper range and the
relationship of the fine fraction to other PM parameters. Sufficient data for these purposes are
6-228
-------
New York, NY
Site 69
Annual Arithmetic Mean (ug/m*)
Site 71
Annual Arithmetic Mean (\iglm3)
70
60
50
40
30
20
10
°8
6 8
7 8
PM
— - -
8 8
10
^^
""" "" -
9 9
0 9
Coan
1 9
se
2 9
(a)
70
60
50
40
30
20
10
3 94 8
PM2.5
NAACJ
6 8
S
7 8
PM1(
X
it*
X*
8 8
\
9 9
> c
r^^^
0 9
oarse
•v^— -
^^^
1 9
-~~~~~
^^ ^^
2 9
PP
(c)
3 9
i/l 2. 5
to
to
VO
90th Percentile (\iglm )
90th Percentile (ug/m3)
1 00
80
60
40
20
°8
^
6 8
X
X
-"'
7 8
' PM1
\s
^v
"^^
8 8
0
\
•—
9 9
0 9
Coars
1 9
e
2 9
F
(b)
3 9
>M2.5
1 00
80
60
40
20
4
6 8
7 8
PM10
^
8 8
X
^•x
9 g
' C
x
**.
0 9
oarse
- — -
1 9
-,-•
2 9
pi
(d)
y
S
3 9
i/l 2. 5
Figure 6-120. Trend data from AIRS: (a) New York City, Site 69, fine, coarse, and PM10 means; (b) New York City,
Site 69, fine, coarse, and PM10 90th percentiles; (c) New York City, Site 71, fine, coarse, and PM10 means;
(d) New York City, Site 71, fine, coarse, and PM10 90th percentiles.
-------
Detroit, Ml
Annual Arithmetic Mean (ug/m3)
St.Louis,MO-IL
Annual Arithmetic Mean (ug/m3)
70
60
50
40
30
20
1 0
°8
NAAC
6 8
S
7 8
PM1 0
- — -
8 8
— — >
9 9
' C
^^
0 9
oarse
\
"W
1 9
2 9
pf
(a)
3 9
i/l 2. 5
70
60
50
40
30
20
10
4
NAAC
6 8
S
7 8
PM1
8 8
0
• ,
-~ ^
9 9
1 (
0 9
2oars
.
1 9
e
~
2 9
P
(c)
^^
3 9
M2.5
Detroit, Ml
90th Percentile (ug/m3)
St.Louis, MO-IL
90th Percentile (ug/m3)
100
80
60
40
20
8
6 8
7 8
PM10
--—^
8 8
9 9
C
•
.---^
•->'**
0 9
oarse
'x-.
V
1 9
2 9
Pi
(b)
3 9
1/I2.5
100
80
60
40
20
4
6 8
7 8
PM1C
8 8
)
—
9 9
C
^
0 9
;oars<
1 9
—- — •
2 9
* p
(d)
— '
3 9
M2.5
Figure 6-121. Trend data from AIRS: (a) Detroit, fine, coarse, and PM10 means; (b) Detroit, fine, coarse, and PM10 90th
percentiles; (c) St. Louis, fine, coarse, and PM10 means; (d) St. Louis, fine, coarse, and PM10 90th percentiles.
-------
Philadelphia, PA - NJ
Annual Arithmetic Mean (ug/m3)
7O
6O
SO
4O
30
2O
1O
B
NAAQS
6 8
.
7 8
^ KJI ••• r
•
8 8
k —
"--^_
^^
9 9
,^~*'
^^
O 9
^^
^\
1 9
2 9
(a)
^~*-
^^
3 9
*> fS.
9Oth Percentile (ug/m )
1OO
8O
6O
4O
2O
86
87 88
- PM1 O
89 9O 91
— — C o a rs e
92 93 94
- - - - PM2.5
Figure 6-122. Trend data from AIRS: (a) Philadelphia, fine, coarse, and PM10 means;
(b) Philadelphia, fine, coarse, and PM10 90th percentiles.
1
2
3
4
5
6
1
available from several sites in California (CARB, 1995) and from Philadelphia (IPN, 1985;
AIRS, 1995; Harvard 1995). However, only the Philadelphia data allows examination of the
relationship of PM25 and PM10 with TSP.
6.10.3.1 Upper Range of Concentration for Various PM Size Fractions
Some information on the upper range of concentrations and relationships among the four
PM size fractions are shown in Tables 6-11 and 6-12. The maximum value; the 2nd, 3rd, 4th,
6-231
-------
to
OJ
to
100
80
60
40
20
0 i
89
180
160
140
120
100
80
60
40
20
0
89
Annual Arithmetic Mean (ug/m )
90 91 92 93
•Total Coarse
NAAQS
94 95
•••• Fine
01/06/91 I 03/01/91 I 05/06/91 I 07/05/91 I 09/03/91 I 11/02/91 I 12/20/91
02/05/91 04/06/91 06/05/91 08/04/91 10/03/91 12/02/91
Date
° Coarse * Fine
90th Percentlle (ug/m )
(b)
1
0.9
0.8
0.7
0.6
0.4
0.3
0.2
0.1
PM 2.5 as a Fraction of PN(o
90
-Total
91
92 93
Coarse
94 95
• • • • Fine
01/17/90 I 01/06/91 I 01/01/92 I 01/01/93 I 01/02/94 I 01/04/95 109/30/95
O7/04/9O 07/O5/91 07/O5/92 07/06/93 07/01/94 07/O2/95
Date
Figure 6-123. Trend data from San Jose from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM,n.
-------
Stockton-Hazelton, CA
100
80
60
40
20
0
89
180
160
140
120
100
80
60
40
20
0
Annual Arithmetic Mean (ug/rA )
(a)
NAAQS
90 91 92 93
•Total Coarse
90th Percentile (pg/rfl )
94 95
Fine
(b)
89
90 91 92 93 94 95
•Total Coarse Fine
100
90
80
70
60
I so
40
30
20
10
0
1
0.9
0.8
0.7
r0'8
a.
•5 0.5
o
a °'4
* 0.3
0.2
0.1
0
Every Sixth Day, 1991
01/06/91 Io3/01/91 I 05/06/91 I 07/05/91 I 09/03/91 I 11/02/91 112/26/91
02/11/91 04/06/91 06/05/91 08/10/91 10/03/91 12/08/91
Date
D COARSE * FINE
PM2 5 as a Fraction of PMJ0
01/OB/89I 01/05/90 I 01/06/911 01/01/921 01/01/93 I 01/02/941 01/04/95 lo8S31/95
07/03/89 07/04/90 07/05/91 07/05/92 07/08/93 07/01/94 07/02/95
Figure 6-124. Trend data from Stockton-Hazelton from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and
total 90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM,n.
-------
Annual Arithmetic Mean
100
80
60
40
20
Visalia, CA
100
Every Sixth Day, 1991
(a)
NAAQS
89
90 91
Total
92 93
Coarse
94 95
Fine
90th Percentile (|jg/m )
to
180
160
140
120
100
80
60
40
20
0
(b)
90
80
70
60
50
40
30
20
10
0
1
0.9
0.8
01/06/91 I 03/01/91 I 05/06/91 I 07/05/91 I 09/15/9* 11/02/91 I 12/26/91
02/05/91 04/06/91 06/05/91 08/10/91 10/03/91 12102/91
Date
0 Coarse * Fine
£ 0.3
0.2
0.1
PM2,5 as a Fraction
89
90 91
-Total
92 93
— Coarse
94
95
Fine
01/04/89 I 01/05/90 I 01/06/91 I 01/01/92 I 01/02/93 I 01/05/94 I
07/03/89 07/04/90 07/05/91 07/05/92 07/06/93 07/01/94
Date
Figure 6-125. Trend data from Visalia from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM10.
-------
to
Annual Arithmetic Mean (tig/rft )
100
80
60
40
20
Bakersfield.CA
(a)
NAAQS
89
90
-Total
91
92 93
Coarse
94
95
Fine
90th Percentile
180
160
140
120
100
80
60
40
20
0
89
90
-Total
91
92 93
-- Coarse
94
95
110
100
90
80
70
60
>
L 50
40
30
20
10
0
Every Sixth Day, 1991
01/06/91 I 03/07/91 I 05/12/91 I 07/05/91 I 09/04/91 I 11/02/91 I 12/26/91
02/06/91 04/06/91 06/05/91 06/04/91 10/03/91 12/02/91
Date
D Coarse + Fine
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
PM2.6 as a Fraction of PM,0
Fine
01/04/89 I 01/05/90 I 01/06/91 I 01/01/92 I 01/01/93 lo 1/06/84!
07/03/89 07/04/90 07/05/91 07/07/92 07/06/93 04/08/94
Figure 6-126. Trend data from Bakersfield from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM10.
-------
Oi
to
Annual Arithmetic Mean (\iglflt )
Azusa, CA
100H
80
60
40
20
(a)
NAAQS
89
90 91 92 93
•Total Coarse
90th Percentile (ug/rr? )
94 95
•••• Fine
180
160
140
120
100
80
60
40
20
0
(b)
89
90
•Total
91
92 93
Coarse
94 95
Fine
110
100
90
80
70
6o
>
L so
40
30
20
10
0
1
0.9
0.8
0.7
S" 0.6
"o
§ °-5
s
u
S 0.4
u.
0.3
0.2
0.1
Every Sixth Day, 1991
01/12/91 03/13/91 05/06/91 07/05/91 09/03/91 11/02/91 01/01/92
02/05/91 04/06/91 06/05/91 08/04/91 10/03/91 12102/91
Date
n Coarse ' Fine
PM2 6 as a Fraction of PM|a
(d)
01/04/8908/26/891 07/04/90 107/05/91 I 07/05/92 I 07/06/93 lo7/01/94 07/02/95I
05/28/89 01/05/90 01/12/91 01/01/92 01/01/93 01/08/94 01/03/95 09/30/95
Date
Figure 6-127. Trend data from Azusa from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM,n.
-------
Annual Arithmetic Mean (ug/m )
to
100
80
60
40
20
Riverside-Rubidoux, CA
(a)
; P _NAAgs_
89
90 91 92 93
•Total Coarse
90th Percentile (Mg/rr? )
94 95
•••• Fine
180
160
140
120
100
80
60
40
20
0
(b)
89
90
•Total
91
92 93
Coarse
94 95
Fine
Every 6th Day, 1991
1
0.9
0.8
0.7
Q- 0.6
0.3
0.2
01/06/91 I 03/01191 lo5J12/91 I D7f05f9ll 09/03/91 I 11/02191 loi/01/92
02/05/91 04/06/91 06/05/91 08/10/91 10/03/91 12/02/91
Date
D Fine + Coarse
PM2 5 as a Fraction of PM,0
(d)
01/22/89 I 02/17/90 I 03/13/91 I 04/30/92 I 05/25/93 I 06/01/94 I 05/27/95
08/02/89 08/21/90 10/15/91 11/26/92 11/27/93 11/28/94
Figure 6-128. Trend data from Riverside-Rubidoux from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and
total 90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM10.
-------
Annual Arithmetic Mean (ug/m3)
100
80
60
40
20
; '_ JNIAAQS
89
90 91 92 93 94 95
•Total Coarse Fine
to
OJ
oo
90th Percentile (ug/m3)
180
160
140
120
100
80
60
40
20
0
(b)
89
90 91
•Total
92 93
Coarse
94 95
Fine
Every Sixth Day, 1991
01/06/91 I 03/01/91 I 05/06/91 17/05/911 09/03/91 I 11/02/91 H 2/14/91
02/05191 04/06/91 OS/29/91 08/04/91 10/03/91 12/02/91
Date
Coarse Fine
PM,
as a Fraction of PM,
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
01/16/89 I 01/05/90 I 01/06/91 I 01/31/92 I 01/01/93 lo 1/02/94 101/09/95
07/03/89 07/04/90 07/05/91 07/05/92 07/06/93 07/01/94
Date
Figure 6-129. Trend data from Lone Pine from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM,n.
-------
to
OJ
VO
Annual Arithmetic Mean (ug/fti )
El Centro, CA
100
80
60
40
20
0
(a)
89
90 91 92 93
Total Coarse
94
Fine
90th Percentile (ug/itf )
180
160
1401
120
1001
80
601
40
20
0
(b)
89 90 91 92 93 94 95
Total Coarse Fine
120
110
100
90
80
70
"
SO
40
30
20
10
0
0.9
0.8
0.7
£ °-«
a.
'o 0.5
c
_O
3 0.4
a
u_
0.3
0.2
0.1
Every Sixth Day, 1991
01/06/91 I D3fo'lf91 I 05/06/91 I 07ll'll91 I 09/0*6/91I 11102/91 11*2114191
02/05191 04/06/91 06/05/91 08/04/91 10/03/91 12/02/91
Date
D Coarse ' Fine
PM2 5 as a Fraction of PM,C
(d)
01/04/89 J09/19/89 105/23/90 101/30/91 110/03/91 100/23/92103/02/93110/28/93107/07/94 103/22/85
05/22/89 01/17/90 09/14/90 08/05/91 02/18/92 11/02/92 08/30/93 03/03/94 11/16/94
Figure 6-130. Trend data from El Centro from CARB: (a) Fine, coarse and total means; (b) Fine, coarse and total
90th percentiles; (c) Every sixth day fine and coarse mass for 1991; (d) Fine and coarse mass as a
fraction of PM10.
-------
TABLE 6-11. MAXIMUM VALUE; 2ND, 3RD, 4TH, AND 5TH HIGHEST VALUES;
98TH AND 95TH PERCENTILE VALUES; 50TH PERCENTILE VALUE (MEDIAN); A,
THE DIFFERENCE BETWEEN THE MEDIAN AND THE MAXIMUM VALUES AND
#, THE NUMBER OF MEASUREMENTS AVAILABLE FROM EIGHT CALIFORNIA
AIR RESOURCES BOARD SITES:
(a) PM, s (b) PMnn., „, and (c) PM,n
PM25
SITE
Riverside
Azusa
Bakersfield
Visalia
Stockton
San Jose
El Centre
Lone Pine
Max
142
98
447
140
94
105
73
29
2nd
130
95
147
121
92
88
62
23
3rd
129
88
119
105
91
86
52
22
4th
122
88
100
91
75
69
49
19
5th
121
87
98
91
75
66
47
18
98%
114
84
93
82
70
59
39
17
95%
77
60
77
69
55
44
26
13
50%
29
23
16
15
11
9
11
6
A
113
75
431
125
83
96
62
23
#
368
371
296
389
381
341
392
322
PM(10-2.5)
Riverside
Azusa
Bakersfield
Visalia
Stockton
San Jose
El Centra
Lone Pine
123
108
320
86
66
55
324
107
114
98
104
75
57
45
176
105
87
71
99
74
57
41
160
84
86
62
98
73
56
39
150
71
86
61
90
70
56
32
132
67
76
57
76
64
54
64
108
42
68
50
61
51
41
51
63
26
34
24
27
21
16
11
27
10
89
84
293
65
50
44
297
97
368
371
296
389
381
341
392
322
PM10
Riverside
Azusa
Bakersfield
Visalia
Stockton
San Jose
El Centra
Lone Pine
194
203
766
187
126
151
347
122
189
152
218
164
119
109
228
120
189
139
183
138
112
102
222
101
182
139
163
137
110
87
167
93
182
135
144
130
102
85
158
76
178
127
135
109
98
76
130
54
130
99
120
98
82
61
90
36
68
50
48
43
30
22
39
16
126
153
718
144
96
129
308
106
368
371
296
389
381
341
392
322
6-240
-------
TABLE 6-12. MAXIMUM VALUE; 2ND, 3RD, 4TH, AND 5TH HIGHEST
VALUES; 98TH AND 95TH PERCENTILE VALUES; 50TH PERCENTILE
VALUE (MEDIAN); A, THE DIFFERENCE BETWEEN THE MEDIAN AND
THE MAXIMUM VALUES AND #, THE NUMBER OF MEASUREMENTS
AVAILABLE FOR STIES IN PHILADELPHIA FROM 1979 TO 1995:
(a) PM,s (b) PMnn.^, and (c) PM,n,AND (d) TSP
Philadelphia
Site
IPN
Average
IPN
S. Broad
AIRS
Harvard
PBY
Dates
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/59
Max
98
54
55
73
PM,,
2nd
94
54
55
72
3rd
74
52
47
56
4th
65
50
46
53
5th
65
50
45
53
98%
61
53
46
43
95%
50
50
43
36
50%
21
22
18
15
A
74
32
37
58
#
366
91
219
1014
PM,,n,«
IPN
Average
IPN
S. Broad
AIRS
Harvard
PBY
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/59
NA
28
39
40
NA
25
39
28
NA
20
38
27
NA
19
37
25
NA
17
30
24
NA
25
37
18
NA
18
25
15
NA
9
12
6
NA
19
27
34
0
91
219
970
PM,n
IPN
Average
IPN
S. Broad
AIRS
Harvard
PBY
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/59
NA
71
86
82
NA
66
83
78
NA
66
82
72
NA
65
79
64
NA
64
73
64
NA
67
79
54
NA
64
60
48
NA
30
31
22
NA
41
55
60
0
91
219
1025
TSP
IPN
Average
IPN
S. Broad
AIRS
Harvard
PBY
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/59
196
116
131
NA
150
107
124
NA
148
105
116
NA
140
101
116
NA
138
99
112
NA
129
109
116
NA
114
100
104
NA
64
61
56
NA
132
55
75
NA
366
91
219
0
6-241
-------
and 5th highest values; the 98th and 95th percentile values; the 50th percentile (median value)
and the difference between the median and the maximum value are given for the measurement
period available at each site. The maximum PM2 5, PM(10_2 5), and PM10 levels were substantially
higher at all the California sites, including the site at Lone Pine (estimated 1980 population,
1800), than at the Philadelphia sites. Differences between maximum and median levels are also
larger at the California sites. The causes for the extremely high values observed at the
Bakersfield site are not known. Data on the upper ranges of TSP are shown for Philadelphia
sites as available.
6.10.3.2 Relationships Between PM25; PM(1025), PM10, and TSP in Philadelphia
Epidemiologists have made extensive use of a long-term TSP data set from Philadelphia
(Chapter 12; Wyzga and Lipfert, 1996; Li and Roth, 1995) to investigate the statistical
relationships between TSP and mortality. It is possible, however, that PM2 5 or PM10, instead of
TSP, may be the causal agent and that TSP may serve as an indicator for PM2 5 or PM10. PM
indicators for Philadelphia, other than TSP, have not been available until recently. Therefore, an
examination of relationships between TSP, PM25, and PM10 in the Philadelphia area may provide
data that will be useful in interpreting the epidemiological results obtained in Philadelphia with
TSP. Such relationships are displayed in a series of Figures (6-131 to 6-135) that show:
(Panel a) TSP plotted versus PMX (where PMX is either PM2 5 or PM10) (Panel b) the distribution
of values of PM/TSP, (Panel c) PM/TSP plotted versus PMX, and (Panel d) PM/TSP plotted
versus TSP.
It would appear from Figures 6-131 to 6-135 that there is some relationship between PIV^
and TSP and that the relationship improves at higher values of TSP. The PM/TSP ratio does
not appear to vary significantly with PMX. However, the ratio does appear to increase with TSP
until a certain level of TSP is reached and then levels off. These visual observations are
quantified by comparison of the PM/TSP ratios at various levels and statistical regressions of
with various TSP fractions shown in Table 6-13.
6-242
-------
PHILADELPHIA, IPN, 3/79 to 12/83
Comparison of TSP and PM2
ON
to
OJ
no
100
90
80
"E 70
a
^ 60
w
S 50
a.
40
30
20
10
a
n
-
-
n
-
n n
B
a a a
n n n
Da n <& g
[_^, n B~^ n [jj|p na n
OHSSi §33 D not] aa
a rJ^^^K^^^^** ra
° U
- J2gjpSPipiP° ^ D
I I I I I I I I I I I I I I I I I I I
4D
24
22
20
n 18
3
™ 16
= 14
E
i 12
0
£ 10
B
6
4
2
A
-
-
-
~
-
-
-
~
-
-
-
_
I 1 ' I
20 40 60 80 100 120 140 160 180 200 " ' I 0.15 1
—
• 1
1 ]
' '•' 1 1
r~i i — i
1 v ii ,| i , , ,
0.25 1 0.35 0.45 0.55 0.65 0.75 1
3 0.1 0.2
TSP, MB/m
0.8
0.7
0.6
0.0.5
09
t
"0.4
S
a.
0.3
0.2
0.1
t\
Comparison of PM>5 and PM>5 /TSP
n
-
n
m n n
n D n
n n n D Dn n
n nD n n n
n ° IQDqi,n D B n D
D nft^Ja ^J, * ° ° ° D
n nji^^^^& n D jj] an
-
ngrifiJiBppiB1 a^
c^H^^Si o
~ ^,1^ °
^°ff n n
n
n
i i i i i i i i i i
0.8
0.7
0.6
JL 0.5
t
S 0.4
S
a.
0.3
0.2
0.1
n
0.3 0.4 0.5 0.6 0.7 0.8
PM^/TSP
2.5
Comparison of TSP and PM2.s /TSP
n
-
n
-
D n^ n
n D :
D D <^^<:
n DDfc 'ft1
^^^S
n n^nnrffi^igHj
n p3™P fflip
ft n 3=U
fln „ a
-
i i i i
:
:n
^S
ipy
^
fci
p
SPC
3^ r
|P
D c
n
n
n n
i n a
n a
an D
ji^ npB n aa n
? ^n 03° v® ° a
J«, mP n Dn
&3& Ofa D n
jaflgr^ib DcP
:DD D g
4:1 n
a
^ n on
, , , ,
TSP, \iglm
Figure 6-131. PM25 and TSP Relationships in Philadelphia, IPN Average, 3/79 to 12/83: (a) comparison of PM25 with
TSP, (b) frequency distribution of PM2 5/TSP, (c) comparison of PM25 /TSP with PM2 5 , (d) comparison
ofPM25/TSPwithTSP.
-------
PHILADELPHIA, IPN, S. BROAD, 3/82-12183; PM
ON
K)
DU
Kfl
OU
40
E
01
a.
S 30
/I
20
10
n o
n
n
n
D
— a
n
° n D n n n
— cm D n
n ag a Dn
n n dP D nn3 ° °
_ n n n n
Dn °^ff^ ° %°nDn
_ n n a n n n
i i i i i i i i i i
21
20
19
1B
17
16
2 15
3 14
- 13
5 12
S 11
^ 10
o 9
7
6
5
4
3
2
1
_
-
-
-
-
-
—
-
—
~
_
-
-
_
_
-
r~i
~~~l
: ' n
: \
•
20 40 60 80 100 120 " D.15 0.2 0.2S 0.3 0.35 0.4 0.45 D.5 0.55 0.8 O.S5 O.T
TSP, |JQ/m
0.7
0.6
0.5
0.
to
t
a 0.4
a.
0.3
0.2
Comparison of PM,5 lAvg TSP and PM^
n
~~ n
a n
aa a a
Q
~~ D D E r-1
n n D n
D DnnDCBiD
D D § n D
n D
nn °D0 nDDDn D°
D n^n °n n DD ° ™
n n n D D
D •? ^*
^ n n
a
n
n
i i i i i i i i i i
0.7
0.6
0.5
a.
to
t
2 0.4
a.
0.3
0.2
Comparison
PM
25 /Average TSP
of PMj5 /TSP and Average TSP
D
-
—
D
n
n
a
n
n n Dr
D
n
cc
n
D
n
6
D
D D i)
n n
_ n
n
n
a n qb
p
CD
n
D D
D
D
n
Q
^i °
D
n n
D
n ,-,
Dn
n
n n
n
n
u° nn n ° n
n n
a
•a
n
u
nn
n
DnnnD
~ n
i
°'1 5 15 25 35 45 55 "' ' 20 40
PM , |jg/m
n
a
n
60
TSP, |jg/m
n
80 100 120
Figure 6-132. PM25 and TSP Relationships in Philadelphia, IPN, South Broad Site, 3/82 to 12/83: (a) comparison of PM25
with TSP, (b) frequency distribution of PM25/TSP, (c) comparison of PM25/TSP with PM25 ,
(d) comparison of PM2 /TSP with TSP.
-------
Comparison of TSP and PM
PHILADELPHIA, AIRS, 1987-1990; PM 2.5
Distribution of PM^ /Average TSP
ON
ou
50
40
30
20
10
n n
-
n
no on
n n
n
no D D
n n n n
n n n_,
n n TD
n
^ nn m™ nn D
a n aim a, m
_ a o c^W/^ BDDc D
D MM TI U nrn n rn\ n
a ""BUB m a~^ K
cfl-i rPPHn EF n D
jj DDJETD D
r-pn n cr Q n n
a
TH
18
17
16
15
« 14
3 13
« 12
iq 1 1
s" 10
0- 9
o 8
* 7
6
5
4
3
2
1
/\
_
-
-
-
-
-
-
~
_
-
-
-
~
-
-
I — I
° 20 40 60 80 100 120 140 I 0.15
a °-1
TSP, |jg/m
Comparison of TSP and PM,,
11 ~ ~
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
f.
a
-
n
n
n
n
a a
n D
D r-i Q
DDDBD B>° *"° S
r- 1 2 '— ' i— ' ,— , n n n
aniing nnD^D§BDniin| n°° ° n ° ° °
an ni B| Dg n0^ °Dn nD a
'""''-' DQD BpHRurP ^^ na ^
Dn,— iH n^ nBn @ ^
nBn|| B a D
. D.*Bo
1.1
1
0.9
0.8
1 0.7
?
aT °-6
£ 0.5
•q
«
> 0.4
0.3
0.2
0.1
"
-
a
-
- na
n
n
n D
n ^h
0.
i— i
0.25
I — I
; '•
n. — |
1' Rlrnrnm
0.35 0.45 0.55 1 0.65 0.75 0.85 1
2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
PM 2.5 /TSP
Comparison of TSP and PM^
n
n
n
n
1 1 1 Q
n n D
a
ifl n i-' ^
QTQI O '-"-' D p.
DQ u D JW LJ-1 n-,H a u " a
n g! n Q a Di dftp n^ an n^1^1 n a a
an
D n
n a
c
[S
n c
bui [J n^ 5 Cd a
H.JJ ^QD r? rrrD ^ 9 HP ^
3nin rr '""'HH^TI ^ ^ n D ^^
D fln D ° °
Dan n ^n an
0 20 40 60 " 20
40
60 80 100 120 140
TSP, |jg/m
Figure 6-133. PM25 and TSP Relationships in Philadelphia, AIRS, 1987 to 1990: (a) comparison of PM25 with TSP,
(b) frequency distribution of PM2 5/TSP, (c) comparison of PM2 5 /TSP with PM2 5 , (d) comparison of
PM25//TSP with TSP.
-------
PHILADELPHIA, IPN, S. BROAD, 3/82-12/83; PM ,„
Distribution of Pty), /Average TSP
ON
K>
ON
BO
70
60
n
1 5°
a.
° 40
=
30
20
10
_ n
n nn
n
~~ n
n
-
n n
Dn ° D
n D Q
D n HO n
n n T-i n n
— D [SB nn Q
n n n_, i-i n D n
^T3 a nn n
J n™ ^ ° H
n ^b nn D D
—
24
22
20
18
at
S 16
o
S 12
Q ...
,B 1 U
eT"
B
6
4
2
-
-
-
-
_
~
-
| 1
" n
20 40 60 80 100 120 " 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 Of 0.85
TSP, (JQ/m3 PM ,„ /Average TSP
0.9
0.8
0.7
0.6
a.
g 0.5
O
»'o.4
0.3
0.2
0.1
Comparison of PM,n lAvg TSP and Avg TSP
_
n
n
° n D
D n n n
n
— |-| '-'nrjrj D'""''""'
rf33 ffli m p. n m n Q
n i-i |= ._. n n n
n Bn§n a
n n
n n
a
— n n
u a
n
-
-
0.9
0.8
0.7
0.6
a.
(? O.S
a
= 0.4
0.3
0.2
0.1
Comparison
_
n
n
ofPM,, /Average TSP and Pl/l
— n
n n
n n
— n
D nn
n n n
n
n *
-
-
I I
°0 20 40 60 80 100 120 " 10
TSP, |jg/m
n
c
c
[ffl c
n
n
D n ft,"
H
Hn c
n
n
]
n
s
^
c
u
n
3 OL
n q_
DC
n
30
PM
n
n
n n
n n a n
n *
m D n n
3D D
i i i i
50 70
„ , ug/m"
Figure 6-134. PM10 and TSP Relationships in Philadelphia, IPN, South Broad Site, 3/82 to 12/83: (a) comparison of PM10
with TSP, (b) frequency distribution of PM10/TSP, (c) comparison of PM10/TSP with PM10, (d) comparison
ofPM10/TSPwithTSP.
-------
to
PHILADELPHIA, AIRS, 19B7-1990; PM1Q
90
80
70
"E 60
^ 50
°
? 40
30
20
10
Comparison of TSP and PM10
n
n D
n
D
n
n
on a an ^
° D mVoD D\ ° °D
a a a a i§ Q gOD a DDD D D
n al a D npT-, ann D
tn5p rjiP D^) Dn D
j% a ^ D
1=11=1 D g^nnD @ ^ D D
D OlP ro|-| Q [Jll |Efi D Q rj
~~ n SjS r&fin-' n n Q
n cH3-1 n tgjrj n n n
n^™ D
17
16
15
14
13
2 12
1 11
>=10
s' 9
^ B
SS 7
6
4
3
2
1
Distribution of PMn /Average TSP
-
-
-
-
-
-
:
_
~
-
-
-
° 20 40 60 BO 100 120 140 "
0.35
-,-.. . a 0.3
TSP, |JQ/m
Comparison of PM,0 lAvg TSP and Avg TSP
1 .6
1.5
1.4
1.3
1.2
E 1.1
° 1
2) 0.9
•=,'••
SQT
a.
0.6
0.5
0.4
0.3
n
-
n
-
n
- n
n
c^
n n
n n ^
- °D B?P BD i D e V Dt
cpnnD ^ Bg^a^cm S1 ^Q n n Q
CTJj'-T: n n'fen QnEnnnDD °a n
rBjrfr? en re O] i-i cP n ^ n
- D Dn g™ n R^nan nn^ D
^n cP ^ n D ^
n an QD rati n D
D, D i i °
1 .D
1.5
1.4
1.3
"E 1-2
= 1.1
a." 1
t 0.9
a
=- O.B
*• 0.7
0.6
0.5
0.4
0.3
-
_
-
~
-
-
0.
, 1
0.45
0.55
I
- 1 1
' ' 1 1
'. -m
0.65 1 0.75 0.85 1 0.95
4 0.5 0.6 0.7 O.B 0.9 1.0
PM10 ITSP
Comparison of PM,n /Averaae TSP and PM,n
n
-
"
'-E
a
n
I
i
20 40 60 60 100 120 140 "'' 0
3
TSP, |jg/m
D°
^tj
n S
3 dp
B*
q
n
Bfit
jfT
nBf
_p nu
ffan
R n
n
3
20
n
n
n
n
n
B
n °
n a cfa ° D ° *
" "n^0 D^*DJ*
^fef^_|[^dP n n n
ft^ nq^S nDn
i 4t~i D cnu § ^ n
ha n n n n
i n
n
n
40 60 80
3
PM1D , ug/m
Figure 6-135. PM10 and TSP Relationships in Philadelphia, AIRS, 1987 to 1990: (a) comparison of PM10 with TSP,
(b) frequency distribution of PM10/TSP, (c) comparison of PM10/TSP with PM10, (d) comparison
ofPM10/TSPwithTSP.
-------
a\
K>
oo
TABLE 6-13. RELATIONSHIPS BETWEEN PMX (PM2 5 OR PM10) AND TSP AS A FUNCTION OF TSP
CONCENTRATION LEVELS FOR SEVERAL SITES IN PHILADELPHIA: (a) RATIO OF PMX TO TSP,
(b) COEFFICIENT OF DETERMINATION (R2)
(a)RatioofPM/TSP
Philadelphia
Site
IPN
Average
IPN
S. Board
AIRS
Dates
3/79
12/83
3/82
12/83
1/87
12/90
TSP
All
0.335 ±0.108
0.371 ±0.105
0.345 ±0.137
PM25/TSP
TSP
<80
0.325 ±0.107
0.361 ±0.106
0.350±0.114
TSP
>80
0.363 ±0.107
0.416 ±0.090
0.317 ±0.083
TSP
All
NA
0.525 ±0.105
0.573 ±0.187
PM10/TSP
TSP
<80
NA
0.516±0.107
0.581 ±0.194
TSP
>80
NA
0.573 ±0.079
0.528±0.131
(b) Coefficients of Determination, R2
Philadelphia
Site
IPN
Average
IPN
S. Board
AIRS
Dates
3/79
12/83
3/82
12/83
1/87
12/90
TSP
All
0.64
0.57
0.45
PM, , with
TSP
<80
0.36
0.38
0.29
TSP
>80
0.50
0.48
0.34
TSP
All
NA
0.78
0.55
PM,nwith
TSP
<80
NA
0.57
0.42
TSP
>80
NA
0.61
0.24
-------
6.10.3.3 Correlations Between PM2 5, PM(10 2 5), and PM10
The analysis of epidemiological results suggest that the smaller size fraction of particulate
matter may have a stronger association with health outcomes than fractions that contain larger
size particles (Chapter 12). It is of interest, therefore, to examine the correlations between PM25,
PM(10.25), and PM10. The means of these fractions and the coefficient of determinination, R2, for
their relationships are shown in Table 6-14 for eight sites in California (CARB, 1995) and in
Table 6-15 for several sites and times for Philadelphia (IPN, 1985; AIRS, 1995; Harvard, 1995).
If correlation between PM25 and PM10 is high but the correlation of PM(10_2 5) with both
PM2 5 and PM10 is low, it is possible that PM10 is serving as an indicator of PM2 5 and that any
health effects of PM(10_2 5) would be masked by the larger PM2 5 (Wilson and Suh, 1996). This
may be the case in Philadelphia since PM2 5 to PM10. In general, PM(10_2 5) is a larger fraction of
PM10 at the California sites than at the Philadelphia site. However, there is still substantial
variability (-40% from minimum to maximum) in this ratio in the data sets from California.
Correlations between PM2 5 and PM(10_2 5) are highly variable at the sites in California and
encompass the Philadelphia value. The large correlations seen between PM2 5 and PM(10_2 5) at
several California sites suggest a significant contribution from crustal material to PM2 5. In
contrast, at the Philadelphia site, only PM2 5 and not PM(10_2 5) was highly correlated with PM10.
These data support the desirability of having independent data on fine mode particles and
coarse mode particles for epidemiological investigations.
6.10.3.4 Fine Fraction
The fine fractions of PM10 (PM2 5/PM10) were shown for Philadelphia in Figure 6-116
(Panels c and d) and for California sites in Figures 6-123 to 6-130. A strong seasonal variation
is evident at the California sites but not in Philadelphia. Numerical values of the PM2 5 fractional
contribution to PM10 are given for Philadelphia and for several California sites in Table 6-16.
These variations in PM2.5/PM10 demonstrate the difficulty of inferring PM25 from PM10
measurements unless some information is available on PM2 5/PM10 on a seasonal and geographic
basis.
6-249
-------
TABLE 6-14. MEANS AND STANDARD DEVIATIONS FOR PM2 s, PM(10 2 5),
AND PM10 AND COEFFICIENTS OF DETERMINATION (R2) BETWEEN PAIRS
FOR EIGHT CALIFORNIA AIR RESOURCES BOARD SITES DURING
THE PERIOD 1989 TO 1990
Mean ± Standard Deviation
Site
Riverside
Azusa
Bakersfield
Visalia
Stockton
San Jose
El Centre
Lone Pine
PM25
34.1 ±24.3
25. 9 ± 17.2
24. 2 ±24. 2
23.0 ±20.5
17.4 ± 16.7
13. 9± 14.1
12.3 ±8.2
6.5 ±3.7
PM(10.2.5)
34.5 ±19.5
25. 5 ± 14.5
33. 7 ±33.6
23. 3 ± 15.9
17.8 ± 10.8
11.9 ±6.7
31. 5 ±25.4
12.1 ± 11.7
PM10
68.6 ±37.6
51. 3 ±27. 7
57.0 ±27.7
46.3 ±26.7
35.6±21.8
25. 8 ± 17.9
43. 8 ±30.5
18.6± 13.8
Coefficient of Determination, R2
Site
Riverside
Azusa
Bakersfield
Visalia
Stockton
San Jose
El Centre
Lone Pine
PM25toPM(10.25)
0.21
0.27
0.36
0.36
0.05
0.16
0.27
0.19
PM2 5 to PM10
0.79
0.79
0.86
0.66
0.77
0.88
0.50
0.42
PM(10.25)toPM10
0.67
0.71
0.74
0.41
0.44
0.48
0.94
0.94
6-250
-------
TABLE 6-15. MEANS AND STANDARD DEVIATIONS FOR PM2 s, PM(10 2 5), PM10,
and TSP AND COEFFICIENTS OF DETERMINATION (R2) BETWEEN PAIRS
FOR SEVERAL SITES IN PHILADELPHIA DURING
PERIODS FROM 1979 TO 1995
Philadephia
Site
IPN Average
IPN S. Board
AIRS
Harvard PBY
Dates
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/95
Mean ± Standard Deviation
PM25 PM(10.2.5) PM10 TSP
23. 3 ±13. 3 NA NA 68.2 ±24.7
22.6 ±11.0 9.7±4.7 32.1 ± 13.5 61.1 ±20.5
19.9 ±10.0 13.1 ±6.7 33. 0± 14.9 58.4 ±21. 9
17.4 ±9.4 7.0 ±4.3 24.3 ±11. 5 NA
Coefficient of Determination, R2
Site
IPN Average
IPN S. Board
AIRS
Harvard PBY
Dates
3/79
12/83
3/82
12/83
1/87
12/90
5/92
5/95
PM25with PM25 PM(10.25)with PM25
PMdo.2.5) withPM10 PM10 with TSP
NA NA NA 0.64
0.14 0.90 0.42 0.57
0.32 0.86 0.69 0.45
0.11 0.88 0.41 NA
6.11 SUMMARY AND CONCLUSIONS
This chapter presents ambient concentration measurements of particulate mass, PM10,
PM2 5, and PM(10_2 5), and of the chemical composition of parti culate matter. For PM10
measurements the number of urban monitoring stations in the AIRS network increased rapidly in
the years immediately after 1985, but the increase slowed substantially in the early 1990s. The
measurements of PM10 at most of these stations were made every 6th day. Measurements
6-251
-------
TABLE 6-16. PM, s/PM,n (FRACTION OF PM,n CONTRIBUTED BY PM, s)
Philadelphia
Mar-May
Jun-Aug
Sept-No v
Dec-Feb
Azusa
Visalia
San Jose
Riverside
Stockton
Bakersfield
Lone Pine
El Centre
Riverside
Winter
Spring
Summer
Fall
Mean
0.71
0.73
0.73
0.72
0.75
0.50
0.49
0.49
0.49
0.46
0.44
0.38
0.29
0.57
0.48
0.41
0.48
Standard Deviation
0.13
0.14
0.16
0.17
0.15
0.13
0.22
0.15
0.14
0.18
0.19
0.14
0.10
0.14
0.13
0.09
0.15
Coefficient of Variation
(%)
18
19
22
24
20
26
45
31
29
39
43
37
34
25
27
22
15
Range
0.09-1.09
0.30-1.56
0.17-1.81
0.03-1.55
0.22-0.99
0.22-0.76
0.23-0.69
0.16-0.74
of chemical species in urban areas usually are obtained in special studies of limited duration.
Data for chemical species in urban areas are discussed as appropriate in the text.
The mass concentration measurements in urban areas have been used to obtain (a) annual
trends in PM10, (b) ratios and correlations of PM2 5 to PM(10_2 5) and PM10 and (c) seasonal
variations in PM10, PM2 5, and PM(10_2 5).
The measurements at non-urban sites were collected at a much smaller number of locations
relative to the number of urban stations by region. The geographical location of the sites in the
IMPROVE/NESCAUM networks were not selected to optimize their locations relative to AIRS
stations in the same region. As a result, not only are there small numbers of non-urban sites by
region, but most of these sites are geographically well displaced from urban areas.
6-252
-------
The non-urban concentration measurements include both mass and chemical composition
so they were used to obtain (a) the variations in PM10, PM2 5, and PM(10_2 5) with month of the
year, (b) the chemical balances for sulfates, organic carbon, elemental carbon, and soil with
month of the year and (c) the variations in the concentrations of S, Se, and V and the S to Se
ratio with month of the year.
From the urban and non-urban PM10 concentration measurements, an "urban excess" was
obtained from the monthly differences in AIRS and IMPROVE/NESCAUM PM10 values.
Because of the limitations mentioned above and the lack of tests of statistical significance, these
"urban excess" values should be viewed as preliminary and used very cautiously with respect to
quantitative results.
Additional sections of Chapter 6 include the following discussions: (1) the mass
apportionment of chemical species obtained from a group of selected research studies of the
chemical composition at locations in the eastern, central and western U.S.; (2) acid sulfate study
results by (a) their geographical distribution in the U.S. and southern Canada, (b) spatial
variations on a city and urban scale, (c) seasonal variations, (d) diurnal variations, and (e) indoor
and personal monitoring relative to outdoor hydrogen ion concentration measurements; (3)
particle number concentrations with emphasis on ultrafme particles; (4) some information on
metals potentially present in ultrafme particles; and (5) information on fine and coarse PM
trends and patterns for sites where both fine and coarse PM measurements were available.
Based on these various concentration measurements a considerable number of conclusions
may be obtained. Many of these conclusions are limited by (a) the number of monitoring sites
available, (b) their geographical location, (c) the frequency of measurement and (d) differences
in methodology used between networks or stations as well as between individual studies of
chemical composition.
Trends in PM10 mass concentration, averaged over regions or by city, usually indicate a
substantial decrease in PM10 concentrations by year from 1988 to 1994. There are exceptions to
this significant downward trend in Philadelphia and at some locations within the Southern
California Basin. The trend plots shown in Chapter 6 have not been tested for statistical
significance. The trend plots can also be influenced by the approach taken in the selection of
stations. Since the number of stations increased rapidly between 1985 and 1990, the trends that
might be obtained using early data could be biased by the added stations being influenced by
6-253
-------
location towards higher or lower PM10 concentration measurements. For this document, the set
of stations in operation from 1988 to 1994 was used to obtain PM10 concentration trends during
this period. It should also be noted that meteorological influences which are known to be
important for deducing trends of O3 concentrations also may affect PM10 concentrations on a
year-to-year basis.
Keeping the limitations mentioned above in mind, urban trend analyses for PM10 are
presented using all stations in operation in a given year and the smaller set of trend stations in
operation over the entire 1988 to 1994 time period. The range for the averaged decrease in PM10
between 1988 and 1994 at urban stations was: for the contiguous U.S., all sites, 24%, trend
sites, 20%; for the eastern U.S., all sites, 16%, trend sites, 18%; and for the western U.S., all
sites, 31%, trend sites 28%. There were appreciable differences between regions in the range of
averaged decreases in PM10 between 1988 and 1994 with the decrease for urban stations in the
northeast ranging from 18% (all) to 19% (trend) while in the industrial midwest the decreases
ranged from 12% (all) to 19% (trend). The ranges of averaged decreases for the three western
regions were from 27% to 37% (all) and 23% to 33% (trend). These decreases in PM10
concentrations resulted in 1994 annual average regional AIRS concentrations in the range of
25 //g/m3 to 32 Mg/m3.
For individual cities, both between and within cities, the decreases in PM10 for individual
stations could show substantial variability. In the Los Angeles Basin, 3 of 6 stations showed
statistically significant downward trends in PM10 while other stations showed no significant
trends. In the western U.S. several large cities showed larger downward trends in PM10 than the
regional averages. PM2.5 and PM(10-2.5) or PM10 data, suitable for determining trends of both
fine and coarse components of PM10, are available from only a few sites in the eastern United
States and a few sites in California. While a general decrease is evident in both fine and coarse
components of PM10 at most sites where data is available, it is not possible to ascertain
differential trends in the two components.
A few attempts to infer various types of background levels of PM2 5 and PM10 have been
made. The backgrounds most relevant to the Criteria Document include a "natural" background
which excludes all anthropogenic sources anywhere in the world, and a background which
excludes anthropogenic sources in North America, but not elsewhere. Annual average natural
background levels of PM10 have been estimated to range from 4 to 8 |ig/m3 in the western United
6-254
-------
States and 5 to 11 |ig/m3 in the eastern United States. Corresponding PM2 5 levels have been
estimated to range from 1 to 4 |ig/m3 in the western United States and 2 to 5 |ig/m3 in the eastern
United States. Twenty-four hour average concentrations may be substantially higher than the
annual or seasonal average background concentrations presented in Chapter 6. The 24-hour
averages are usually considered for control strategies while the annual and seasonal averages are
suitable for risk analyses.
Based either on the correlation of individual values or on the average PM2 5 to PM10 values,
the annual ratios of PM2 5 to PM10 from urban stations fell within a relatively narrow range of
0.55 to 0.6, for both the entire eastern and western U.S. However, for two regions, the upper
midwest and southwest, the correlations yielded ratios of less than 0.2 while the average PM25 to
PM10 values yielded ratios between 0.3 and 0.4.
Ratios of PM25 to PM(10.25) from urban stations can vary with season as well as between
regions. In the northeast, southeast, and industrial midwest regions, there is appreciable
uniformity with PM2 5 exceeding PM(10_2 5) during all seasons of the year. In contrast, in the
southwest, the PM2 5 is less than the PM coarse during all seasons of the year. In the northwest
and in southern California, PM2 5 exceeds PM10 in the fall and winter with the reverse occurring
in the spring and summer.
Measurements of the day to day variability in PM2 5 and PM10 are available from only one
site located in Philadelphia, PA. The data show day to day variations of 8.6±7.5 |ig/m3 for PM10,
6.8±6.5 |ig/m3 for PM25, and 3.7±3.4 |ig/m3 for PM10.25 from May 1992 to April 1995.
Maximum day to day differences were 50 |ig/m3 for PM10, 55 |ig/m3 for PM2 5, and 35 |ig/m3 for
PM(10_2 5). The ratio of PM2 5 to PM10 was 0.72±0.16 over the measurement period and the
correlation between PM2 5 and PM10 was 0.86 (R2) suggesting that variability in PM2 5 was
forcing the variability in PM10. Data collected by dichotomous samplers at several sites in
California showed that PM(10_2 5) accounted for roughly half of PM10 and that both PM2 5 and
PM(10_2 5) were highly correlated with PM10. Differences among the Philadelphia data set and the
California data sets illustrate the dangers in extrapolating relations among different size fractions
from one region of the country to other regions.
Comparisons of seasonal profiles of PM10 show summer peaks for both urban and
nonurban sites in the northeast, southeast, and industrial midwest. These summer peaks usually,
but not exclusively, are associated with the summer peaks in PM2 5. The PM2 5 concentrations at
6-255
-------
non-urban sites in the northeast, southeast, and industrial midwest exceed the PM(10_2 5)
concentrations in all seasons of the year, as is the case for urban stations. The northwest urban
PM10 and PM2 5 concentrations show a spring and early summer minimum with the highest
values in fall and winter, while the non-urban PM10 and PM2 5 concentrations show a summer
peak similar to the seasonal profiles in the eastern U.S. In southern California, the urban PM10
and PM2 5 seasonal profiles show fall peaks, while the non-urban seasonal profiles have a
relatively flat maximum from spring into early fall. Again it must be emphasized that with so
few nonurban sites in most regions any conclusions drawn from the comparisons above are very
tentative for most regions of the U.S.
The every-sixth-day urban PM10 averaged concentrations for most regions of the
United States ranged during 1990 to 1994 from 10 to 15 //g/m3 up to 40 to 60 //g/m3. The
southern California region had PM10 values averaging up to 70 to 75 //g/m3. Day-to-day
variations in PM10 concentrations in Knoxville, TN, ranged from 10 to 20 //g/m3, while in
Missoula, MT, PM10 concentrations ranged from <10 to 120 to 140 //g/m3 with one value over
200 Mg/m3.
A quantity termed an urban excess has been discussed extensively in the text of Chapter 6.
In view of the distinctions discussed above between the number and geographical distribution of
urban and non-urban sites, the quantitative results probably should be interpreted with
considerable caution. While it is reasonable that additional sources within cities should increase
PM10 concentrations significantly above those at non-urban sites, the quantitative differences can
be sensitive to the location of the non-urban sites with respect to individual cities. The most
striking feature of the urban excess is its large increase in the fall and winter in the western
United States compared to the eastern United States.
The chemical compositions at the nonurban IMPROVE/NESCAUM sites are discussed
within the earlier sections of Chapter 6. Later in Chapter 6 an independent evaluation of
chemical composition is given based on a mixture of intensive studies at both urban and
nonurban sites. The results from both approaches appear reasonably consistent in showing
geographical variations in chemical composition.
Both approaches indicate that sulfate, presumably present either as (NH4)HSO4 or as
(NH4)2SO4, is the largest contributor to the chemical species measured in the eastern
United States. Other results indicate that a large regional background of sulfate is superimposed
6-256
-------
on a smaller urban contribution. Results also indicate that sulfate is relatively uniform in
concentration throughout much of the eastern United States. These results are less pronounced
in the late fall and winter months. The contribution of sulfate to PM10 is somewhat smaller than
sulfate is to PM2 5. Comparisons of the eastern United States with the central United States and
western United States show a decreasing contribution of sulfate to the chemical composition.
Conversely, the soil and/or mineral concentrations become an increasingly important contributor
to PM10 and PM2 5 going from the eastern to the western United States. The nitrates, as NH4NO3,
also appear to be a much more important contributor to the composition in areas of the western
United States than in the eastern United States. Organic compounds also appear to increase in
importance relative to sulfate going from the eastern to the western United States. For PM(10_2 5),
sulfates are relatively unimportant. Soil or mineral components dominate the PM(10_2 5), but there
is a substantial unknown fraction of PM(10.25).
Particle strong acidity, defined as H2SO4 plus HSO^ is a regional pollutant fairly evenly
distributed across large areas of the central portion of the eastern United States. It is relatively
evenly distributed across small cities, but in the one large urban area from which results have
been reported, the higher concentrations of ammonia in the central city apparently neutralize a
significant portion of the acidity. Thus, higher concentrations of acidity are found in rural areas,
small towns, and suburban areas than in the centers of larger urban areas. The concentration of
acidity is higher in the summer and peaks during the early afternoon in urban areas. Indoor,
outdoor, and personal monitoring indicates that indoor and personal concentrations of acidity are
lower than outdoor concentrations, presumably due to neutralization by indoor ammonia.
Particle strong acidity is normally found exclusively in the fine particle mode. Coarse particles
tend to be basic. Exceptions may occur during periods of fog or very high relative humidity.
The number concentration of particles is generally dominated by particles below 0.1 //m or
100 nm in diameter, termed ultrafme particles. When a distinct mode is present, it is called the
nuclei mode. Number geometric mean diameter ranged from 12 to 43 nm in Long Beach, CA
and 47 to 75 nm in clean air in the Rocky Mountains. Particle number concentrations varied
from less than 1,000/cm3 at clean background sites to over 100,000/cm3 in polluted urban areas
and were correlated with the volume of particles below 0.1 //m. Particle number concentrations
were not found to be correlated with accumulation mode volume on an hourly basis.
6-257
-------
Correlations of particle number and accumulation mode volume might be expected if compared
over longer time intervals (e.g., days), but such studies have not yet been done.
An examination of the size distribution of metals suggests that metals that may be
volatilized during combustion may appear as ultrafine particles. Such metals include copper,
zinc, and lead and possibly nickel and vanadium, as well as nonmetals selenium and sulfur.
Ultrafine particles appear to exist longer under conditions of low concentrations and high
relative humidity.
6-258
-------
REFERENCES
AIRS, Aerometric Information Retrieval System [database]. (1995) [PM10 and PM25 data]. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards.
Altshuller, A. P. (1980) Seasonal and episodic trends in sulfate concentrations (1963-1978) in the eastern United
States. Environ. Sci. Technol. 14: 1337-1349.
Altshuller, A. P. (1982) Relationships involving particle mass and sulfur content at sites in and around St. Louis,
Missouri. Atmos. Environ. 16: 837-843.
Altshuller, A. P. (1985) Relationships involving fine particle mass, fine particle sulfur and ozone during episodic
periods at sites in and around St. Louis, MO. Atmos. Environ. 19: 265-276.
Altshuller, A. P. (1987) Relationships between direction of wind flow and fine particle sulfur concentrations within
and upwind of St. Louis, MO. Atmos. Environ. 21: 1023-1032.
Appel, B. R.; Colodny, P.; Wesolowski, J. J. (1976) Analysis of carbonaceous materials in southern California
atmospheric aerosols. Environ. Sci. Technol. 10: 359-363.
Appel, B. R.; Kothny, E. L.; Hoffer, E. M.; Hidy, G. M.; Wesolowski, J. J. (1978) Sulfate and nitrate data from the
California Aerosol Characterization Experiment (ACHEX). Environ. Sci. Technol. 12: 418-425.
Appel, B. R.; Hoffer, E. M.; Kothny, E. L.; Wall, S. M.; Haik, M.; Knights, R. L. (1979) Analysis of carbonaceous
material in southern California atmospheric aerosols. 2. Environ. Sci. Technol. 13: 98-104.
Ashbaugh, L. L.; Watson, J. G.; Chow, J. (1989) Estimating fluxes from California's dry deposition monitoring data.
Presented at: 82nd annual meeting of the Air & Waste Management Association; June; Anaheim, CA.
Pittsburgh, PA: Air & Waste Management Association; paper no. 89-65.3.
Barone, J. B.; Cahill, T. A.; Eldred, R. A.; Flocchmi, R. G.; Shadoan, D. J.; Dietz, T. M. (1978) A multivanate
statistical analysis of visibility degradation at four California cities. Atmos. Environ. 12: 2213-2221.
Barton, R. G.; Clark, W. D.; Seeker, W. R. (1990) In: Proceedings of the 1st international conference on toxic
combustion byproducts.
Bennett, R. L.; Stockburger, L.; Barnes, H. M. (1994) Comparison of sulfur measurements from a regional fine
particle network with concurrent acid modes network results. Atmos. Environ. 28: 409-419.
Bemer, A.; Lilrzer, C. (1980) Mass size distributions of traffic aerosols at Vienna. J. Phys. Chem. 84: 2079-2083.
Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
Technol. 23: 1408-1412.
Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1991) Indoor and outdoor concentrations of inorganic acidic
aerosols and gases. J. Air Waste Manage. Assoc. 41: 171-181.
Burton, R. M.; Liu, L. J. S.; Wilson, W. E.; Koutrakis, P. (1995) Comparison of aerosol acidity in urban and
semi-rural environments. Atmos. Environ.: submitted.
Cadle, S. H. (1985) Seasonal variations in nitric acid, nitrate, strong aerosol acidity, and ammonia in an urban area.
Atmos. Environ. 19: 181-188.
6-259
-------
Cahill, T. A. (1979) Comments on surface coating for Lundgren-type impactors. In: Lundgren, D. A.; Lippmann, M.;
Harris, F. S., Jr.; Clark, W. E.; Marlow, W. H.; Durham, M. D., eds. Aerosol measurement: [papers from a
workshop]; March 1976; Gainesville, FL. Gainesville, FL: University Presses of Florida; pp. 131-134.
Cahill, T. A. (1980) Proton microprobes and particle-induced X-ray analytical systems. Annu. Rev. Nucl. Part. Sci.
30:211-252.
Cahill, T. A. (1996) [Data from IMPROVE data base.] Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards.
Cahill, T. A.; Wakabayashi, P. (1993) Compositional analysis of size-segregated aerosol samples. In: Newman, L., ed.
Measurement challenges in atmospheric chemistry. Washington, DC: American Chemical Society; pp.
211-228. (Advances in chemistry series no. 232).
Cahill, T. A.; Goodart, C.; Nelson, J. W.; Eldred, R. A.; Nasstrom, J. S.; Feeney, P. J. (1985) Design and evaluation of
the drum impactor. In: Ariman, T.; Veziroglu, T. N, eds. Proceedings of the international symposium on
particulate and multi-phase processes. Washington, DC: Hemisphere Publishing Corporation; pp. 319-325.
Cahill, T. A.; Feeney, P. J.; Eldred, R. A. (1987) Size-time composition profile of aerosols using the drum sampler. In:
Proceedings of the 4th international PIXE conference; June 1986; Tallahassee, FL. Nucl. Instrum. Methods
Phys. Res. B22: 344-348.
Cahill, T. A.; Surovik, M.; Wittmeyer, I. (1990) Visibility and aerosols during the 1986 Carbonaceous Species
Methods Comparison Study. Aerosol Sci. Technol. 12: 149-160.
Cahill, T. A.; Wakabayashi, P.; Raabe, O.; Teague, S. (1992a) A new inertial impactor for visibility-reducing aerosols.
Presented at: 85th annual meeting and exhibition of the Air & Waste Management Association; June; Kansas
City, MO. Pittsburgh, PA: Air & Waste Management Association; paper no. 92-60.08.
Cahill, T. A.; Wilkinson, K.; Schnell, R. (1992b) Composition analyses of size-resolved aerosol samples taken from
aircraft downwind of Kuwait, spring 1991. J. Geophys. Res. [Atmos.] 97: 14,513-14,520.
Cahill, T. A.; Morales, R.; Miranda, J. (1996a) Comparative aerosol studies of Pacific rim cities—Santiago, Chile
(1987); Mexico City, Mexico (1987-1990); and Los Angeles, U.S.A. (1973 and 1987). Atmos. Environ. 30:
747-749.
Cahill, T. A.; Eldred, R. A.; Wakabayashi, P. H. (1996b) Trends in fine particle concentrations at Great Smoky
Mountains National Park. For presentation at: Annual meeting of the Air & Waste Management Association;
June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association; submitted paper no.
96-MP1A.05.
California Air Resources Board. (1995) [Data base provided by John R. Holmes, Chief, Research Division.]
Sacramento, CA: California Air Resources Board.
Cass, G. R.; McRae, G. J. (1983) Source-receptor reconciliation of routine air monitoring data for trace metals: an
emission inventory assisted approach. Environ. Sci. Technol. 17: 129-139.
Cass, G. R.; Conklin, M. H.; Shah, J. J.; Huntzicker, J. J.; Macias, E. S. (1984) Elemental carbon concentrations:
estimation of an historical data base. Atmos. Environ. 18: 153-162.
Chesworth, S.; Yang, G.; Chang, D. P. Y.; Jones, A. D.; Kelly, P. B.; Kennedy, I. M. (1994) The fate of arsenic in a
laminar diffusion flame. Combust. Flame 98: 259-266.
Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal source
contributions to PM10 and PM25 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM-10:
6-260
-------
implementation of standards, an APCA/EPA international specialty conference; February; San Francisco,
CA. Pittsburgh, PA: Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
Chow, J. C.; Watson, J. G.; Pntchett, L.; Lowenthal, D.; Frazier, C.; Neuroth, G.; Evans, K.; Moon, D. (1990)
Wintertime visibility in Phoenix, Arizona. Presented at: the 83rd annual meeting & exhibition; June;
Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association; paper no. 90-66.6.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hmsvark, B. A.; Pnchett, L. C.; Neuroth, G. R. (1992a)
Wintertime PM10 and PM^ chemical compositions and source contributions in Tucson, Arizona. In: Chow, J.
C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an A&WMA/EPA
international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management
Association; pp. 231-243. (A&WMA transactions series no. 22).
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
(1992b) PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A
26: 3335-3354.
Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
PM10 source contributions in Rubidoux, California. Atmos. Environ. Part A 26: 693-706.
Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
(1993a) PM10 andPM25 compositions in California's San Joaquin Valley. Aerosol Sci. Technol. 18: 105-128.
Chow, J. C.; Watson, J. G.; Ono, D. M.; Mathai, C. V. (1993b) PM10 standards and nontraditional particulate source
controls: a summary of the A&WMA/EPA international specially conference. J. Air Waste Manage. Assoc.
43: 74-84.
Chow, J. C.; Watson, J. G.; Fujita, E. M.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994a) Temporal and spatial
variations of PM25 and PM10 aerosol in the Southern California Air Quality Study. Atmos. Environ. 28:
2061-2080.
Chow, J. C.; Watson, J. G.; Solomon, P. A.; Thmllier, R. H.; Magliano, K. L.; Ziman, S. D.; Blumenthal, D. L.;
Richards, L. W. (1994b) Planning for SJVAQS/AUSPEX particulate matter and visibility sampling and
analysis. In: Solomon, P. A.; Silver, T. A., eds. Planning and managing regional air quality: modeling and
measurement studies: a perspective through the San Joaquin Valley air quality study and AUSPEX. Boca
Raton, FL: Lewis Publishers; pp. 171-216.
Chow, J. C.; Fairley, D.; Watson, J. G.; DeMandel, R.; Fujita, E. M.; Lowenthal, D. H.; Lu, Z.; Frazier, C. A.; Long,
G.; Cordova, J. (1995a) Source apportionment of wintertime PM10 at San Jose, Calif. J. Environ. Eng. (NY)
121:378-387.
Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. H.; Frazier, C. A.; Solomon, P. A.; Thuiller, R. H.; Magliano, K.
(1995b) Descriptive analysis of PM25 and PM10 regionally representative locations during
SJVAQS/AUSPEX. Atmos. Environ.: in press.
Coboum, W. G.; Husar, R. B. (1982) Diurnal and seasonal patterns of particulate sulfur and sulfuric acid in St. Louis,
July 1977-June 1978. Atmos. Environ. 16: 1441-1450.
Conner, T. L.; Miller, J. L.; Willis, R. D.; Kellogg, R. B.; Dann, T. F. (1993) Source apportionment of fine and coarse
particles in southern Ontario, Canada. Presented at: 86th annual meeting & exhibition of the Air & Waste
Management Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper
no. 93-TP-58.05.
Countess, R. J.; Wolff, G. T.; Cadle, S. H. (1980) The Denver winter aerosol: a comprehensive chemical
characterization. J. AirPollut. Control Assoc. 30: 1194-1200.
6-261
-------
Countess, R. J.; Cadle, S. H.; Groblicki, P. J.; Wolff, G. T. (1981) Chemical analysis of size-segregated samples of
Denver's ambient particulate. J. AirPollut. Control Assoc. 31: 247-252.
Davis, B. L.; Johnson, L. R.; Stevens, R. K.; Courtney, W. I; Safriet, D. W. (1984) The quartz content and elemental
composition of aerosols from selected sites of the EPA inhalable particulate network. Atmos. Environ. 18:
771-782.
Davison, R. L.; Natusch, D. F. S.; Wallace, J. R.; Evans, C. A., Jr. (1974) Trace elements in fly ash: dependence of
concentration on particle size. Environ. Sci. Technol. 8: 1107-1113.
Desaedeleer, G. G.; Winchester, J. W.; Akselsson, K. R. (1977) Monitoring aerosol elemental composition in particle
size fractions for predicting human respiratory uptake. Nucl. Instrum. Methods 142: 97-99.
Desert Research Institute. (1995) [Unpublished data]. Reno, NV: Energy and Environmental Engineering Center.
Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
and acid aerosols. Environ. Res. 59: 362-373.
Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
(1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329:
1753-1759.
Dresser, A. L. (1988) A dispersion and receptor model analysis of a western community's PM-10 problem. JAPCA
38: 1419-1421.
Dzubay, T. G. (1980) Chemical element balance method applied to dichotomous sampler data. Ann. N. Y. Acad. Sci.
338: 126-144.
Dzubay, T. G.; Stevens, R. K.; Lewis, C. W.; Hem, D. H.; Courtney, W. J.; Tesch, J. W.; Mason, M. A. (1982)
Visibility and aerosol composition in Houston, Texas. Environ. Sci. Technol. 16: 514-525.
Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, L; Sheffield, A. E.; Courtney, W. J. (1988) A composite
receptor method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
Eldering, A.; Cass, G. R. (1994) A source-oriented model for air pollutant effects on visibility. J. Geophys. Res.
[Atmos.]: submitted.
Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the United
States of America. Atmos. Environ. 28: 1009-1019.
Eldred, R. A.; Ashbaugh, L. L.; Cahill, T. A.; Flocchini, R. G.; Pitchford, M. L. (1983) Sulfate levels in the southwest
during the 1980 copper smelter strike. J. Air Pollut. Control Assoc. 33: 110-113.
6-262
-------
Eldred, R. A.; Cahill, T. A.; Feeney, P. J.; Malm, W. C. (1987) Regional patterns in particulate matter from the
National Park Service network, June 1982 to May 1986. In: Bhardwaja, P. S., ed. Visibility protection:
research and policy aspects, an APCA international specialty conference; September 1986; Grand Teton
National Park, WY. Pittsburgh, PA: Air Pollution Control Association; pp. 386-396. (APCA transactions
series no. TR-10).
Eldred, R. A.; Cahill, T. A.; Pitchford, M.; Malm, W. C. (1988) IMPROVE-a new remote area particulate monitoring
system for visibility studies. Presented at: 81st annual meeting of the Air Pollution Control Association;
June; Dallas, TX. Pittsburgh, PA: Air Pollution Control Association; paper no. 88-54.3.
Eldred, R. A.; Cahill, T. A.; Wilkinson, L. K.; Feeney, P. J.; Chow, J. C.; Malm, W. C. (1990) Measurement of fine
particles and their chemical components in the IMPROVE/NPS networks. In: Mathai, C. V., ed. Visibility
and fine particles: an A&WMA/EPA international specialty conference; October 1989; Estes Park, CO.
Pittsburgh, PA: Air & Waste Management Association; pp. 187-196. (A&WMA transactions series no.
TR-17).
Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1994a) Regional patterns of selenium and other trace elements in the
IMPROVE network. In: Proceedings of the international specially conference on aerosol and atmospheric
optics: radiative balance and visual air quality, volume A. Pittsburgh, PA: Air and Waste Management
Association.
Eldred, R. A.; Cahill, T. A.; Flocchmi, R. G. (1994b) Comparison of PM10 and PM25 aerosols in the IMPROVE
network. In: Proceedings of an international specialty conference on aerosol and atmospheric optics:
radiative balance and visual air quality, volume A. Pittsburgh, PA: Air and Waste Management Association.
Air Waste: submitted.
Ellenson, W. D.; Schwab, M.; Egler, K. A.; Shadwick, D.; Willis, R. D. (1994) Draft technical report for the pilot
project of lower Rio Grande valley environmental study. ManTech Environmental Technology. Submitted
to U.S. EPA.
Fairley, D. (1990) The relationship of daily mortality to suspended particulates in Santa Clara county, 1980-86.
Environ. Health Perspect. 89: 159-168.
Federal Register. (1986) Hazardous waste management system; land disposal restrictions. F. R. (November 7)
51:40572-40654.
Fekete, K. E.; Gyenes, L. (1993) Regional scale transport model for ammonia and ammonium. Atmos. Environ. Part
A 27: 1099-1104.
Ferman, M. A.; Wolff, G. T.; Kelly, N. A. (1981) The nature and sources of haze in the Shenandoah Valley/Blue
Ridge Mountains area. J. AirPollut. Control Assoc. 31: 1074-1082.
Flocchmi, R. G.; Cahill, T. A.; Shadoan, D. J.; Lange, S. J.; Eldred, R. A.; Feeney, P. J.; Wolfe, G. W.; Simmeroth, D.
C.; Suder, J. K. (1976) Monitoring California's aerosols by size and elemental composition. Environ. Sci.
Technol. 10: 76-82.
Flocchini, R. G.; Cahill, T. A.; Eldred, R. A.; Feeney, P. J. (1990) Particulate sampling in the northeast: a description
of the Northeast States for Coordinated Air Use Management (NESCAUM) network. In: Mathai, C. V., ed.
Visibility and fine particles: an A&WMA/EPA international specially conference; October 1989; Estes Park,
CO. Pittsburgh, PA: Air & Waste Management Association; pp. 197-206. (A&WMA transactions series no.
TR-17).
Frazier, C. A. (1989) Letter [report on mass and chemical concentrations and babs measurements for Phoenix pilot
study samples]. Reno, NV: Desert Research Institute, Energy and Environmental Engineering Center.
6-263
-------
Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals of aerosol behavior. New York, NY: John Wiley &
Sons.
Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1984) Elemental and organic carbon particle
concentrations: a long-term perspective. Sci. Total Environ. 36: 17-25.
Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1986) Characteristics of atmospheric organic
and elemental carbon particle concentrations in Los Angeles. Environ. Sci. Technol. 20: 580-589.
Griffing, G. W. (1980) Relations between the prevailing visibility, nephelometer scattering coefficient and
sunphotometer turbidity coefficient. Atmos. Environ. 14: 577-584.
Groblicki, P. J.; Wolff, G. T.; Countess, R. J. (1981) Visibility-reducing species in the Denver "brown cloud" -1.
relationships between extinction and chemical composition. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr.,
eds. Plumes and visibility: measurements and model components, proceedings of the symposium; November
1980; Grand Canyon National Park, AZ. Atmos. Environ. 15: 2473-2484.
Hasan, H.; Dzubay, T. G. (1987) Size distributions of species in fine particles in Denver using a microorifice
impactor. Aerosol Sci. Technol. 6: 29-39.
Hering, S. V.; Friedlander, S. K. (1982) Origins of aerosol sulfur size distributions in the Los Angeles basin. Atmos.
Environ. 16: 2647-2656.
Hering, S. V.; McMurry, P. H. (1991) Optical counter response to monodisperse atmospheric aerosols. Atmos.
Environ. Part A 25: 463-468.
Hering, S. V.; Flagan, R. C.; Friedlander, S. K. (1978) Design and evaluation of new low-pressure impactor.
I. Environ. Sci. Technol. 12: 667-673.
Henng, S. V.; Appel, B. R.; Cheng, W.; Salaymeh, F.; Cadle, S. H.; Mulawa, P. A.; Cahill, T. A.; Eldred, R. A.;
Surovik, M.; Fitz, D.; Howes, J. E.; Knapp, K. T.; Stockburger, L.; Turpin, B. J.; Huntzicker, J. J.; Zhang,
X.-Q.; McMurry, P. H. (1990) Comparison of sampling methods for carbonaceous aerosols in ambient air.
Aerosol Sci. Technol. 12: 200-213.
Hering, S.; Eldering, A.; Seinfeld, J. H. (1996) Bimodal character of accumulation mode aerosols in southern
California. Atmos. Environ.: submitted.
Hidy, G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. (1980) The character and origins of
smog aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX). New
York, NY: John Wiley & Sons.
Hildemann, L. M.; Markowski, G. R.; Cass, G. R. (1991) Chemical composition of emissions from urban sources of
fine organic aerosol. Environ. Sci. Technol. 25: 744-759.
Hoggan, M. C.; Banayan, S. C.; Sin, R. W.; Davidson, A. (1993) An analysis of PM10 air quality in California's South
Coast Air Basin, 1985-1992. In: Criteria air pollutants: volume 3B, papers from the 86th annual meeting &
exhibition of the Air & Waste Management Association; June ; Denver, CO. Pittsburgh, PA: Air & Waste
Management Association; paper no. 93-WA-75.04. (A&WMA publication no. AM93-3).
6-264
-------
Houck, J. E.; Rau, J. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM10 nonattainment
area. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an
A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
Waste Management Association; pp. 219-230. (A&WMA transactions series no. 22).
Husar, R. B.; Frank, N. H. (1991) Interactive exploration and analysis of EPA's Aerometric Information and Retrieval
System (AIRS) data sets. Presented at: 84th annual meeting and exhibition of the Air & Waste Management
Association; June; Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
no. 91-119.4.
Husar, R. B.; Poirot, R. (1992) Exploration of the AIRS database: weekly cycle of ozone and PM 10 aerosols.
Presented at: 85th annual meeting of the Air & Waste Management Association; June; Kansas City, MO.
Pittsburgh, PA: Air & Waste Management Association; paper no. 92-77.06.
Husar, R. B.; Wilson, W. E. (1993) Haze and sulfur emission trends in the eastern United States. Environ. Sci.
Technol. 27: 12-16.
Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
haziness: a summary. Atmos. Environ. 15: 1919-1928.
Husar, R. B.; Elkins, J. B.; Wilson, W. E. (1994) U.S. visibility trends, 1960-1992, regional and national. Presented at:
87th annual meeting of the Air & Waste Management Association; June; Cincinnati, OH. Pittsburgh, PA: Air
& Waste Management Association; paper no. 94-MP3.05.
IPN. (1985) [Inhalable Particulate Network Database.] Research Triangle Park, NC: U.S. Environmental Protection
Agency, National Exposure Research Laboratory; Jose M. Sune, contact.
Jaenicke, R. (1980) Natural aerosols. In: Kneip, T. J.; Lioy, P. J., eds. Aerosols: anthropogenic and natural, sources
and transport. Ann. N. Y. Acad. Sci. 338: 317-329.
Japar, S. M.; Brachaczek, W. W.; Gorse, R. A., Jr.; Norbeck, J. M.; Pierson, W. R. (1986) The contribution of
elemental carbon to the optical properties of rural atmospheric aerosols. Atmos. Environ. 20: 1281-1289.
John, W.; Wall, S. M.; Ondo, J. L.; Winklmayr, W. (1989) Acidic aerosol size distributions during SCAQS.
Sacramento, CA: California Air Resources Board; report no. CA/DOH/AIHL/SP-51.
John, W.; Wall, S. M.; Ondo, J. L.; Winklmayr, W. (1990) Modes in the size distributions of atmospheric inorganic
aerosol. Atmos. Environ. Part A 24: 2349-2359.
Johnson, R. L.; Shah, J. J.; Gary, R. A.; Huntzicker, J. J. (1981) An automated thermal-optical method for the analysis
of carbonaceous aerosol. In: Macias, E. S.; Hopke, P. K., eds. Atmospheric aerosol: source/air quality
relationships; based on a symposium jointly sponsored by the Divisions of Nuclear Chemistry and
Technology at the 180th national meeting of the American Chemical Society; August 1980; Las Vegas, NV.
Washington, DC: American Chemical Society; pp. 223-233. (Comstock, M. J., ed. ACS symposium series:
167).
Johnson, D. L.; Davis, B. L.; Dzubay, T. G.; Hasan, H.; Crutcher, E. R.; Courtney W. J.; Jaklevic, J. M.; Thompson,
A. C.; Hopke, P. K. (1984) Chemical and physical analyses of Houston aerosol for interlaboratory
comparison of source apportionment procedures. Atmos. Environ. 18: 1539-1553.
6-265
-------
Joseph, D. B.; Metsa, J.; Malm, W. C.; Pitchford, M. (1987) Plans for IMPROVE: a federal program to monitor
visibility in class I areas. In: Bhardwaja, P. S., ed. Visibility protection: research and policy aspects, an
APCA international specialty conference; September 1986; Grand Teton National Park, WY. Pittsburgh, PA:
Air Pollution Control Association; pp. 113-125. (APCA transactions series no. TR-10).
Kao, A. S.; Friedlander, S. K. (1994) Chemical signatures of the Los Angeles aerosol (dp<3.5 //m). Aerosol Sci.
Technol. 21:283-293.
Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
Environ. Sci. Technol. 29: 19-28.
Karch, R.; Kruisz, C.; Reischl, G. P.; Winklmayr, W. (1987) Continuous measurement of ultrafine particle size
distributions in the range from 2 to 200 nm from the southern California air quality study, June-July 1987
[preliminary report]. Vienna, Austria: Institut fur Experimentalphysik der Universitat Wien.
Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity on
dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
Keeler, G. I; Japar, S. M.; Brachaczek, W. W.; Gorse, R. A., Jr.; Norbeck, J. M.; Pierson, W. R. (1990) The sources
of aerosol elemental carbon at Allegheny Mountain. Atmos. Environ. Part A 24: 2795-2805.
Keeler, G. J.; Spengler, J. D.; Castillo, R. A. (1991) Acid aerosol measurements at a suburban Connecticut site.
Atmos. Environ. Part A 25: 681-690.
Kelly, N. A.; Wolff, G. T.; Ferman, M. A. (1982) Background pollutant measurements in air masses affecting the
eastern half of the United States - 1. air masses arriving from the northwest. Atmos. Environ. 16: 1077-1088.
Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan development
in the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate
source controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ.
Pittsburgh, PA: Air & Waste Management Association; pp. 979-991. (A&WMA transactions series no. 22).
Kleinman, M. T.; Pasternack, B. S.; Eisenbud, M.; Kneip, T. J. (1980) Identifying and estimating the relative
importance of sources of airborne particulates. Environ. Sci. Technol. 14: 62-65.
Koutrakis, P. (1995) [PM25 and PM10 data for Philadelphia, 1992-1995]. Boston, MA: Harvard School of Public
Health.
Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity and
ambient relative humidity. J. Geophys. Res. [Atmos.] 98: 7141-7147.
Koutrakis, P.; Spengler, J. D. (1987) Source apportionment of ambient particles in Steubenville, OH using specific
rotation factor analysis. Atmos. Environ. 21: 1511-1519.
Kreidenweis, S. M. (1993) Towards development of a northern hemisphere natural sulfate aerosol budget. Presented
at: 86th annual meeting of the Air & Waste Management Association; June; Denver, CO. Pittsburgh, PA: Air
& Waste Management Association; paper no. 93-MP-4.08.
Kreidenweis, S.; Brechtel, F. (1995) [Particle number and size-distribution database]. Fort Collins, CO: Colorado
State University, Foothill Campus, Atmospheric Sciences Department.
Lamborg, C.; Keeler, G. J.; Evans, G. (1992) Atmospheric acidity measurements during the Lake Michigan Urban Air
Toxics Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research
and Exposure Assessment Laboratory; EPA report no. EPA/600/A-92/250. Available from: NTIS,
Springfield, VA; PB93-121069.
6-266
-------
Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
Leaderer, B. P.; Bernstein, D. M.; Daisey, J. M.; Kleinman, M. T.; Kneip, T. J.; Knutson, E. O.; Lippmann, M.; Lioy,
P. J.; Rahn, K. A.; Sinclair, D.; Tanner, R. L.; Wolff, G. T. (1978) Summary of the New York summer
aerosol study (NYSAS). J. Air Pollut. Control Assoc. 28: 321-327.
Leaderer, B. P.; Tanner, R. L.; Holford, T. R. (1982) Diurnal variations, chemical composition and relation to
meteorological variables of the summer aerosol in the New York subregion. Atmos. Environ. 16: 2075-2087.
Lee, H. S.; Wadden, R. A.; Scheff, P. A. (1993) Measurement and evaluation of acid air pollutants in Chicago using
an annular denuder system. Atmos. Environ. Part A 27: 543-553.
Lewis, C. W.; Dzubay, T. G. (1986) Measurement of light absorption extinction in Denver. Aerosol Sci. Technol. 5:
325-336.
Lewis, C. W.; Macias, E. S. (1980) Composition of size-fractionated aerosol in Charleston, West Virginia. Atmos.
Environ. 14: 185-194.
Lewis, C. W.; Baumgardner, R. E.; Stevens, R. K.; Russwurm, G. M. (1986) Receptor modeling study of Denver
winter haze. Environ. Sci. Technol. 20: 1126-1136.
Li, Y.; Roth, H. D. (1995) Daily mortality analysis by using different regression models in Philadelphia County,
1973-1990. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution
and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 45-58.
Liang, C. S. K.; Waldman, J. M. (1992) Indoor exposures to acidic aerosols at child and elderly care facilities. Indoor
Air 2: 196-207.
Liousse, C.; Penner, J. E.; Chuang, C.; Walton, J. J.; Eddleman, H.; Cachier, H. (1996) A global three-dimensional
model study of carbonaceous aerosols. J. Geophys. Res. [Atmos.] 101: in press.
Lioy, P. J.; Samson, P. J.; Tanner, R. L.; Leaderer, B. P.; Minnich, T.; Lyons, W. (1980) The distribution and transport
of sulfate "species" in the New York metropolitan area during the 1977 summer aerosol study. Atmos.
Environ. 14: 1391-1407.
Lioy, P. J.; Daisey, J. M.; Greenberg, A.; Harkov, R. (1985) A major wintertime (1983) pollution episode in northern
New Jersey: analyses of the accumulation and spatial distribution of inhalable particulate matter, extractable
organic matter and other species. Atmos. Environ. 19: 429-436.
Lipfert, F. W.; Wyzga, R. E. (1993) On the spatial and temporal variability of aerosol acidity and sulfate
concentration. Air Waste 43: 489-491.
Liu, L.-J. S.; Burton, R.; Wilson, W. E.; Koutrakis, P. (1996) Comparison of aerosol acidity in urban and semi-rural
environments. Atmos. Environ. 30: 1237-1245.
Lundgren, D. A. (1967) An aerosol sampler for determination of particle concentration as a function of size and time.
J. Air Pollut. Control Assoc. 17: 225-229.
Macias, E. S.; Zwicker, J. O.; Ouimette, J. R.; Hering, S. V.; Friedlander, S. K.; Cahill, T. A.; Kuhlmey, G. A.;
Richards, L. W. (1981) Regional haze case studies in the southwestern U.S. -1. aerosol chemical
composition. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and visibility: measurements and
model components, proceedings of the symposium; November 1980; Grand Canyon National Park, AZ.
Atmos. Environ. 15: 1971-1986.
6-267
-------
Maenhaut, W.; Roy set, O.; Vadset, M.; Kauppinen, E. I; Lind, T. M. (1993) Analysis of size-fractionated coal
combustion aerosols by PIXE and other analytical techniques. Nucl. Instrum. Methods Phys. Res. Sect. B 75:
266-272.
Malm, W. C.; Gebhart, K. A.; Molenar, I; Cahill, T.; Eldred, R.; Huffman, D. (1994a) Examining the relationship
between atmospheric aerosols and light extinction at Mount Rainier and North Cascades National Parks.
Atmos. Environ. 28: 347-360.
Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994b) Spatial and seasonal trends in particle
concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
Mangelson, N F.; Rees, L. B.; Lewis, L.; Joseph, J. M.; Cui, W.; Machir, J.; Eatough, D. J.; Wilkerson, T. D.; Jensen,
D. T. (1994) The formation of sulfate and nitrate during winter inversion fogs in Cache Valley, Utah. In:
Proceedings of an international specialty conference on aerosol and atmospheric optics: radiative balance and
visual air quality, volume A. Pittsburgh, PA: Air and Waste Management Association.
Marple, V.; Rubow, K.; Ananth, G.; Fissan, H. J. (1986) Micro-orifice uniform deposit impactor. J. Aerosol Sci. 17:
489-494.
Marple, V. A.; Rubow, K. L.; Behm, S. M. (1991) A microorifice uniform deposit impactor (MOUDI): description,
calibration, and use. Aerosol Sci. Technol. 14: 434-446.
Marshall, B. T.; Patterson, E. M.; Grams, G. W. (1986) Characterization of the Atlanta area aerosol, elemental
composition and possible sources. Atmos. Environ. 20: 1291-1300.
McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
aerosols. Atmos. Environ. 23: 497-507.
McMurry, P. H.; Zhang, X. Q. (1989) Size distributions of ambient organic and elemental carbon. Aerosol Sci.
Technol. 10: 430-437.
Mercer, T. T. (1964) The stage constants of cascade impactors. Ann. Occup. Hyg. 7: 115-124.
Molenar, J. V.; Dietrich, D. L.; Cismoski, D. S.; Cahill, T. A.; Wakabayashi, P. H. (1994) In: Proceedings of an
international specialty conference on aerosol and atmospheric optics: radiative balance and visual air quality,
volume A. Pittsburgh, PA: Air & Waste Management Association.
Moyers, J. L. (1982) 1981-1982 aerosol studies for nine southwest national parks and monuments. Washington, DC:
U.S. Department of the Interior, National Park Service, Southwest Region.
Mukerjee, S.; Fox, D. L.; Stevens, R. K.; Shy, C. M.; Vescio, N; Lumpkm, T. A.; Kellogg, R. B. (1993)
A methodology to apportion ambient air measurements to investigate potential effects on air quality near
waste incinerators. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric
Research and Exposure Assessment Laboratory; report no. EPA/600/A-93/104. Available from: NTIS,
Springfield, VA; PB93-191393.
Mylonas, D. T.; Allen, D. T.; Ehrman, S. H.; Pratsinis, S. E. (1991) The sources and size distributions of
organonitrates in Los Angeles aerosol. Atmos. Environ. Part A 25: 2855-2861.
Natusch, D. F. S.; Wallace, J. R. (1974) Urban aerosol toxicity: the influence of particle size. Particle size, adsorption,
and respiratory deposition profiles combine to determine aerosol toxicity. Science (Washington, DC) 186:
695-699.
Natusch, D. F. S.; Wallace, J. R.; Evans, C. A., Jr. (1974) Toxic trace elements: preferential concentration in respirable
particles. Science (Washington, DC) 183: 202-204.
6-268
-------
Neas, L. M. (1996) [Data provided for 1986 from Harvard six cities data base.] Boston, MA: Harvard School of
Public Health.
Noll, K. E.; Pontius, A.; Frey, R.; Gould, M. (1985) Comparison of atmospheric coarse particles at an urban and
non-urban site. Atmos. Environ. 19: 1931-1943.
Noll, K. E.; Draftz, R.; Fang, K. Y. P. (1987) The composition of atmospheric coarse particles at an urban and
non-urban site. Atmos. Environ. 21: 2717-2721.
Noll, K. E.; Yuen, P.-F.; Fang, K. Y.-P. (1990) Atmospheric coarse particulate concentrations and dry deposition
fluxes for ten metals in two urban environments. Atmos. Environ. Part A 24: 903-908.
Patterson, R. K.; Wagman, J. (1977) Mass and composition of an urban aerosol as a function of particle size for
several visibility levels. J. Aerosol Sci. 8: 269-279.
Pickle, T.; Allen, D. T.; Pratsinis, S. E. (1990) The sources and size distributions of aliphatic and carbonyl carbon in
Los Angeles aerosol. Atmos. Environ. Part A 24: 2221-2228.
Pierson, W. R.; Brachaczek, W. W. (1988) Coarse- and fine-particle atmospheric nitrate and HNO3(g) in Claremont,
California, during the 1985 Nitrogen Species Methods Comparison Study. Atmos. Environ. 22: 1665-1668.
Pierson, W. R.; Russell, P. A. (1979) Aerosol carbon in the Denver area in November 1973. Atmos. Environ. 13:
1623-1628.
Pierson, W. R.; Brachaczek, W. W.; Korniski, T. J.; Truex, T. J.; Butler, J. W. (1980a) Artifact formation of sulfate,
nitrate, and hydrogen ion on backup filters: Allegheny Mountain experiment. J. Air Pollut. Control Assoc.
30: 30-34.
Pierson, W. R.; Brachaczek, W. W.; Truex, T. J.; Butler, J. W.; Komiski, T. J. (1980b) Ambient sulfate measurements
on Allegheny Mountain and the question of atmospheric sulfate in the northeastern United States. Ann. N. Y.
Acad. Sci. 338: 145-173.
Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Norbeck, J. M. (1986) On the acidity of dew. J.
Geophys. Res. [Atmos]. 91: 4083-4096.
Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Norbeck, J. M.; Keeler, G. J. (1987) Acid ram and
atmospheric chemistry at Allegheny Mountain. Environ. Sci. Technol. 21: 679-691.
Pierson, W. R.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Norbeck, J. M.; Keeler, G. J. (1989) Atmospheric
acidity measurements on Allegheny Mountain and the origins of ambient acidity in the northeastern United
States. Atmos. Environ. 23: 431-459.
Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, Sz.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995) Chemical
characterization of the ambient aerosol in Philadelphia. Presented at: Particulate matter: health and regulatory
issues: an A&WMA international specially conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
Management Association; submitted abstract.
Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition processes.
In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume 1: emissions, atmospheric
processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment Program.
(State of science and technology report no. 1).
Poirot, R. L.; Flocchini, R. G.; Husar, R. B. (1990) Winter fine particle composition in the northeast: preliminary
results from the NESCAUM monitoring network. Presented at: 83rd annual meeting of the Air & Waste
6-269
-------
Management Association; June; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
paper no. 90-84.5.
Poirot, R. L.; Galvin, P. J.; Gordon, N.; Quan, S.; Van Arsdale, A.; Flocchini, R. G. (1991) Annual and seasonal fine
particle composition in the northeast: second year results from the NESCAUM monitoring network.
Presented at: 84th annual meeting and exhibition of the Air & Waste Management Association; June;
Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-49.1.
Pope, C. A., Ill; Schwartz, J.; Ransom, M. R. (1992) Daily mortality and PM10 pollution in Utah valley. Arch.
Environ. Health 47: 211-217.
Raabe, O. G.; Braaten, D. A.; Axelbaum, R. L.; league, S. V.; Cahill, T. A. (1988) Calibration studies of the drum
impactor. J. Aerosol Sci. 19: 183-195.
Reid, J. S.; Cahill, T. A.; Wakabayashi, P. H.; Dunlap, M. R. (1994) Geometric/aerodynamic equivalent diameter
ratios of ash aggregate aerosols collected in burning Kuwaiti well fields. Atmos. Environ. 28: 2227-2234.
Rodes, C. E.; Evans, E. G. (1985) Preliminary assessment of 10 ,um particulate sampling at eight locations in the
United States. Atmos. Environ. 19: 293-303.
Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993) Quantification of urban
organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part
A 27: 1309-1330.
Russell, A. G.; Cass, G. R. (1986) Verification of a mathematical model for aerosol nitrate and nitric acid formation
and its use for control measure evaluation. Atmos. Environ. 20: 2011-2025.
Schichtel, B. A.; Husar, R. B. (1991) Apportionment of light extinction by aerosol types over the U.S. Hawscom Air
Force Base, MA: Air Force Systems Command, Phillips Laboratory; scientific report no. 4.
Schwartz, J. (1994) Air pollution and hospital admissions for the elderly in Birmingham, Alabama. Am. J. Epidemiol.
139: 589-598.
Schwartz, J.; Dockery, D. W. (1992) Increased mortality in Philadelphia associated with daily air pollution
concentrations. Am. Rev. Respir. Dis. 145: 600-604.
Seaton, A.; MacNee, W.; Donaldson, K.; Godden, D. (1995) Particulate air pollution and acute health effects. Lancet
(8943): 176-178.
Seeker, W. R. (1990) Waste combustion. In: Proceedings of the 23rd international symposium on combustion;
pp. 867-885.
Seinfeld, J. H. (1986) Atmospheric chemistry and physics of air pollution. New York, NY: John Wiley and Sons.
Senior, C. L.; Flagan, R. C. (1982) Ash vaporization and condensation during combustion of a suspended coal
particle. Aerosol Sci. Technol. 1: 371-383.
Shah, J. J.; Watson, J. G., Jr.; Cooper, J. A.; Huntzicker, J. J. (1984) Aerosol chemical composition and light
scattering in Portland, Oregon: the role of carbon. Atmos. Environ. 18: 235-240.
Shaw, R. W., Jr.; Paur, R. J. (1983) Composition of aerosol particles collected at rural sites in the Ohio River Valley.
Atmos. Environ. 17: 2031-2044.
Sisler, J. F.; Malm, W. C. (1994) The relative importance of soluble aerosols to spatial and seasonal trends of impaired
visibility in the United States. Atmos. Environ. 28: 851-862.
6-270
-------
Sisler, J. F.; Huffman, D.; Lattimer, D. A.; Malm, W. C.; Pitchford, M. L. (1993) Spatial and temporal patterns and the
chemical composition of the haze in the United States: an anlysis of data from the IMPROVE network,
1988-1991. Fort Collins, CO: Colorado State University, Cooperative Institute for Research in the
Atmosphere (CIRA).
Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PM10 air quality assessment for the Jefferson County, Ohio air
quality control region. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source
controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
Air and Waste Management Association; pp. 1016-1031. (A&WMA transactions series no. 22).
Sloane, C. S.; Watson, J.; Chow, J.; Pritchett, L.; Richards, L. W. (1991) Size-segregated fine particle measurements
by chemical species and their impact on visibility impairment in Denver. Atmos. Environ. Part A 25:
1013-1024.
Small, M.; Germani, M. S.; Small, A. M.; Zoller, W. H.; Moyers, J. L. (1981) Airborne plume study of emissions from
the processing of copper ores in southeastern Arizona. Environ. Sci. Technol. 15: 293-299.
Solomon, P. A.; Moyers, J. L. (1986) A chemical characterization of wintertime haze in Phoenix, Arizona. Atmos.
Environ. 20: 207-213.
Solomon, P. A.; Fall, T.; Salmon, L.; Cass, G. R.; Gray, H. A.; Davidson, A. (1989) Chemical characteristics of PM10
aerosols collected in the Los Angeles area. JAPCA 39: 154-163.
Spengler, J. D.; Thurston, G. D. (1983) Mass and elemental composition of fine and coarse particles in six U.S. cities.
J. AirPollut. Control Assoc. 33: 1162-1171.
Spengler, J. D.; Allen, G. A.; Foster, S.; Severance, P.; Ferris, B., Jr. (1986a) Sulfuric acid and sulfate aerosol events
in two U. S. cities. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk
assessment and control strategies - proceedings of the second U.S.-Dutch international symposium; May
1985; Wilhamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 107-120.
Spengler, J. D.; Briggs, S. L. K.; Ozkaynak, H. (1986b) Relationship between TSP measurements and
size-fractionated particle mass measurements in six cities participating in the Harvard Air Pollution Health
Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards; EPA cooperative agreement no. CR-812667-01-0; December 5.
Spengler, J. D.; Brauer, M.; Koutrakis, P. (1990) Acid air and health. Environ. Sci. Technol. 24: 946-956.
Stevens, R. K. (1985) Sampling and analysis methods for use in source apportionment studies to determine impact of
wood burning on fine particle mass. Environ. Int. 11: 271-283.
Stevens, R. K. (1995) [Unpublished data].
Stevens, R. K.; Dzubay, T. G.; Shaw, R. W., Jr.; McClenny, W. A.; Lewis, C. W.; Wilson, W. E. (1980)
Characterization of the aerosol in the Great Smoky Mountains. Environ. Sci. Technol. 14: 1491-1498.
Stevens, R. K.; Dzubay, T. G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods applied to the
determination of the origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ. 18:
261-272.
Stevens, R. K.; Hoffman, A. J.; Baugh, J. D.; Lewis, C. W.; Zweidinger, R. B.; Cupitt, L. T.; Kellogg, R. B.;
Simonson, J. H. (1993) A comparison of air quality measurements in Roanoke, VA, and other integrated air
cancer project monitoring locations. In: Measurement of toxic and related air pollutants: proceedings of the
1993 U.S. EPA/A&WMA international symposium; May; Durham, NC. Pittsburgh, PA: Air & Waste
Management Association; pp. 185-189. (A&WMA specially conference VIP-34).
6-271
-------
Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
Technol. 26: 2507-2517.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993) Validation of personal exposure models for sulfate and aerosol
strong acidity. J. Air Waste Manage. Assoc. 43: 845-850.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international specially
conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association; pp.
244-261. (A&WMA transactions series no. 22).
Thompson, K. M.; Koutrakis, P.; Brauer, M.; Spengler, J. D.; Wilson, W. E.; Burton, R. M. (1991) Measurements of
aerosol acidity: sampling frequency, seasonal variability, and spatial variation. Presented at: 84th annual
meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC, Canada.
Pittsburgh, PA: Air & Waste Management Association; paper no. 91-89.5.
Thurston, G. D.; Gorczynski, J. E., Jr.; Jaques, P.; Currie, J.; He, D. (1992) An automated sequential sampling system
for particulate acid aerosols: description, characterization, and field sampling results. J. Exposure Anal.
Environ. Epidemiol. 2: 415-428.
Trijonis, J. (1982) Existing and natural background levels of visibility and fine particles in the rural East. Atmos.
Environ. 16: 2431-2445.
6-272
-------
Trijonis, J. C. (1991) Visibility: existing and historical conditions—causes and effects. Appendix A: characterization of
natural background aerosol concentrations. In: Irving, P. M., ed. Acidic deposition: state of science and
technology, volume III: terrestrial, materials, health and visibility effects. Washington, DC: The U.S.
National Acid Precipitation Assessment Program; pp. 24-A1 - 24-AT. (State of science and technology
report no. 24).
Trijonis, J.; McGown, M.; Pitchford, M.; Blumenthal, D.; Roberts, P.; White, W.; Macias, E.; Weiss, R.; Waggoner,
A.; Watson, J.; Chow, J.; Flocchini, R. (1988) Visibility conditions and causes of visibility degradation in the
Mojave Desert of California: executive summary, RESOLVE project final report. China Lake, CA:
Department of the Navy, Naval Weapons Center; document no. NWC TP 6869. Available from: NTIS,
Springfield, VA; AD-A206 322.
Tuncel, S. G.; Olmez, I.; Parrington, J. R.; Gordon, G. E.; Stevens, R. K. (1985) Composition of fine particle regional
sulfate component in Shenandoah Valley. Environ. Sci. Technol. 19: 529-537.
Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and elemental
carbon. Aerosol Sci. Technol. 12: 161-171.
Turpin, B. J.; Huntzicker, J. J.; Larson, S. M.; Cass, G. R. (1991) Los Angeles summer midday particulate carbon:
primary and secondary aerosol. Environ. Sci. Technol. 25: 1788-1793.
U.S. Environmental Protection Agency. (1989) An acid aerosols issue paper: health effects and aerometrics. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; EPA report no. EPA-600/8-88-005F. Available from: NTIS, Springfield, VA; PB91-125864.
Van Grieken, R. E.; Johansson, T. B.; Winchester, J. W.; Odom, L. A. (1975) Micro-determination of
zirconium-hafnium ratios in zircons by proton induced X-ray emission. Z. Anal. Chem. 275: 343-348.
Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM10 source apportionment using local surface dust profiles:
examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate
source controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ.
Pittsburgh, PA: Air & Waste Management Association; pp. 262-278. (A&WMA transactions series no. 22).
Vossler, T. L.; Lewis, C. W.; Stevens, R. K.; Dzubay, T. G.; Gordon, G. E.; Tuncel, S. G.; Russwurm, G. M.; Keeler,
G. J. (1989) Composition and origin of summertime air pollutants at Deep Creek Lake, Maryland. Atmos.
Environ. 23: 1535-1547.
Waldman, J. M.; Lioy, P. J.; Thurston, G. D.; Lippmann, M. (1990) Spatial and temporal patterns in summertime
sulfate aerosol acidity and neutralization within a metropolitan area. Atmos. Environ. Part B 24: 115-126.
Waldman, J. M.; Liang, S.-K. C.; Lioy, P. J.; Thurston, G. D.; Lippmann, M. (1991) Measurements of sulfate aerosol
and its acidity in the SO2 source region of Chestnut Ridge, PA. Atmos. Environ. Part A 25: 1327-1333.
6-273
-------
Waldman, J. M; Koutrakis, P.; Allen, G. A.; Thurston, G. D.; Burton, R. M.; Wilson, W. E. (1995) Human exposures
to particle strong acidity. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air
pollution and human mortality and morbidity, part II; January 1994; Irvine, CA. Inhalation Toxicol. 7:
657-669.
Wall, S. M.; John, W.; Ondo, J. L. (1988) Measurement of aerosol size distributions for nitrate and major ionic
species. Atmos. Environ. 22: 1649-1656.
Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactor for sampling inorganic ions. Aerosol Sci.
Technol. 8: 157-172.
Watson, J. G.; Chow, J. C.; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olnez, I.
(1988) The 1987-88 metro Denver brown cloud study: v. I, II, III. Reno, NV: Desert Research Institute; final
report no. 8810.1F(l-3).
Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994a) Chemical
mass balance source apportionment of PM10 during the Southern California Air Quality Study. Aerosol Sci.
Technol. 21: 1-36.
Watson, J. G.; Roth, P. M.; Blumenthal, D. L.; Ranzien, A. J.; Solomon, P. A.; Thuilher, R. H. (1994b)
Recommendations on planning a large-scale combined monitoring and modeling effort based on the
SJVAQS/AUSPEX air quality study. In: Solomon, P. A.; Silver, T. A., eds. Planning and managing regional
air quality modeling and measurement studies: a perspective through the San Joaquin Valley air quality study
and AUSPEX. Boca Raton, FL: Lewis Publishers; pp. 369-382.
Weber, R. J.; McMurry, P. H.; Eisele, F. L.; Tanner, D. J. (1995) Measurement of expected nucleation precursor
species and 5 to 500 run diameter particles at Mauna Loa, Hawaii. J. Atmos. Sci. 52: 2242-2257.
Weiss, R. E.; Larson, T. V.; Waggoner, A. P. (1982) In situ rapid-response measurement of H2SO4/(NH4)2SO4
aerosols in rural Virginia. Environ. Sci. Technol. 16: 525-532.
Wesolowski, J. J.; John, W.; Devor, W.; Cahill, T. A.; Feeney, P. J.; Wolfe, G.; Flocchmi, R. (1977) Collection
surfaces of cascade impactors. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental samples.
Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 121-131.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy, G.
M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of smog
aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX). New York,
NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and technology: v. 9).
Whitby, K. T.; Husar, R. B.; Liu, B. Y. H. (1972) The aerosol size distribution of Los Angeles smog.
J. Colloid Interface Sci. 39: 177-204.
White, W. H. (1996a) Alternative perspectives on a trend in eastern fine-particle sulfur concentrations. Environ.
Manager: in press.
White, W. H. (1996b) Deteriorating air or improving measurements?—on interpreting concatenate time series.
J. Geophys. Res. [Atmos.]: submitted.
White, W. H.; Macias, E. S. (1987) Particulate nitrate measurements in rural areas of the western United States.
Atmos. Environ. 21: 2563-2571.
6-274
-------
White, W. H.; Macias, E. S. (1989) Carbonaceous particles and regional haze in the western United States. Aerosol
Sci. Technol. 10: 111-117.
Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. Air Pollut. Control Assoc.
25: 529-534.
Wilson, W. E.; Stockburger, L. (1990) Diurnal variations in aerosol composition and concentration. In: Masuda, S.;
Takahashi, K., eds. Aerosols: science, industry, health and environment: proceedings of the third
international aerosol conference, v. 2; September; Kyoto, Japan. Oxford, United Kingdom: Pergamon Press
pic; pp. 962-965.
Wilson, W. E.; Suh, H. H. (1996) Fine and coarse particles: concentration relationships relevant to epidemiological
studies. J. Air Waste Manage. Assoc.: accepted.
Wilson, W. E.; Koutrakis, P.; Spengler, J. D. (1991) Diurnal variations of aerosol acidity, sulfate, and ammonia in the
atmosphere. Presented at: 84th annual meeting and exhibition of the Air & Waste Management Association;
June; Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-89.9.
Wolff, G. T.; Korsog, P. E. (1985) Estimates of the contributions of sources to inhalable particulate concentrations in
Detroit. Atmos. Environ. 19: 1399-1409.
Wolff, G. T.; Monson, P. R.; Ferman, M. A. (1979) On the nature of the diurnal variation of sulfates at rural sites in
the Eastern United States. Environ. Sci. Technol. 13: 1271-1276.
Wolff, G. T.; Countess, R. J.; Grobhcki, P. J.; Ferman, M. A.; Cadle, S. H.; Muhlbaier, J. L. (1981) Visibility-reducing
species in the Denver "brown cloud"—II. sources and temporal patterns. Atmos. Environ. 15: 2485-2502.
Wolff, G. T.; Ferman, M. A.; Kelly, N. A.; Stroup, D. P.; Ruthkosky, M. S. (1982) The relationships between the
chemical composition of fine particles and visibility in the Detroit metropolitan area. J. Air Pollut. Control
Assoc. 32: 1216-1220.
Wolff, G. T.; Kelly, N. A.; Ferman, M. A.; Morrissey, M. L. (1983) Rural measurements of the chemical composition
of airborne particles in the eastern United States. J. Geophys. Res. C: Oceans Atmos. 88: 10,769-10,775.
Wolff, G. T.; Korsog, P. E.; Stroup, D. P.; Ruthkosky, M. S.; Morrissey, M. L. (1985) The influence of local and
regional sources on the concentration of inhalable particulate matter in southeastern Michigan. Atmos.
Environ. 19: 305-313.
Wolff, G. T.; Ruthkosky, M. S.; Stroup, D. P.; Korsog, P. E.; Ferman, M. A.; Wendel, G. J.; Stedman, D. H. (1986)
Measurements of SOX, NOX and aerosol species on Bermuda. Atmos. Environ. 20: 1229-1239.
Wolff, G. T.; Ruthkosky, M. S.; Stroup, D. P.; Korsog, P. E. (1991) A characterization of the pnncipal PM-10 species
in Claremont (summer) and Long Beach (fall) during SCAQS. Atmos. Environ. Part A 25: 2173-2186.
Wyzga, R. E.; Lipfert, F. W. (1996) [Philadelphia TSP data.] Palo Alto, CA: Electric Power Research Institute.
Zhang, X.; Turpin, B. J.; McMurry, P. J.; Herring, S. V.; Stolzenburg, M. R. (1994) Mie theory evaluation of species
contributions to the 1990 wintertime visibility reduction in the Grand Canyon. J. Air Waste Manage. Assoc.
44: 153-162.
6-275
-------
APPENDIX 6A:
TABLES OF CHEMICAL COMPOSITION OF
PARTICULATE MATTER
6A-1
-------
TABLE 6A-la. SUMMARY OF PM2. STUDIES
EAST
Smoky Mtn.
Shenandoah
Camden
Philadelphia
Deep Creek
Roanoke
Raleigh
Watertown
Hartford
Boston
Res.Tr. Pk.
Charlotte
Allegheny Mtn.
Allegheny Mtn.
Laurel Hill
REF
1
1
2
3
4
5
5
6,7
8
8
8
20
44
45-50
45-50
NOTE WEST
Boise
T arrant CA
b Five Points CA
Riverside CA
c San Jose
d Honolulu
d Winnemucca NV
Portland
a Seattle
a Southern California
a San Joaquin Valley
e Phoenix
Nevada
REF
5
8
8
8
8
8
8
8
8
9,31
10
11
12
NOTE CENTRAL
d Albuquerque
a St. Louis
a Steubenville
a Harriman
a Portage
a Topeka
a Inglenook AL
a Braidwood IL
a Kansas City KS
g,h Minneapolis
i St. Louis
j Kansas City MO
f Akron
Cincinnati
Buffalo
Dallas
El Paso
Denver
Urban Denver
Non-urban Denver
Chicago
Houston
St.Louis
Harriman
St. Louis
Steubenville
Brownsville
Ontario
REF
5
6,7
6,7
6,7
6,7
6,7
8
8
8
8
8
8
8
8
8
8
8
13
14
14
15
16
17
17
18
21
24
37
NOTE
d
a
a
a
a
a
a
a
a
a
a
a
m
aa
k
n
1
6A-2
-------
TABLE 6A-lb. SUMARY OF COARSE FRACTION STUDIES
EAST
Smoky Mtn.
Shenandoah
Camden
Philadelphia
Watertown
Hartford
Boston
Res.Tr. Pk.
Allegheny Mtn.
Allegheny Mtn.
Laurel Hill
REF NOTE
1 0
1 0
2 b
3 ab
6,7 o,p
8 a,o
8 a,o
8 a,o
44
45-50
45-50
WEST
Tarrant CA
Five Points CA
Riverside CA
San Jose
Honolulu
Winnemucca NV
Portland
Seattle
Southern California
San Joaquin Valley
Phoenix
REF
8
8
8
8
8
8
8
8
9,31
10
11
NOTE CENTRAL
a,o St. Louis
a,o Steubenville
a,o Harriman
a,o Portage
a,o Topeka
a,o Inglenook AL
a,o Braidwood IL
a,o Kansas City KS
g Akron
i Cincinnati
j Buffalo
Dallas
El Paso
Denver
Chicago
Houston
St. Louis
Harriman
St. Louis
Brownsville
Ontario
REF
6,7
6,7
6,7
6,7
6,7
8
8
8
8
8
8
8
8
13
15
16
17
17
18
24
37
NOTE
o,p
o,p
o,p
o,p
o,p
a,o
a,o
a,o
a,o
a,o
a,o
a,o
a,o
o
s
o
k,r
n
1
6A-2
-------
TABLE 6A-lc. SUMMARY OF PM,,, STUDIES
EAST
Smoky Mtn.
Shenandoah
Camden
Philadelphia
Kingston
Watertown
Hartford
Boston
Res.Tr. Pk.
Allegheny Mtn.
Allegheny Mtn.
Laurel Hill
REF NOTE
1 o,q
1 o,q
2 b
3 ab
6,7 p,q
6,7 p,q
8 a,q
8 a,q
8 a,q
44
45-50
45-50
WEST
Tarrant CA
Five Points CA
Riverside CA
San Jose CA
Honolulu HI
Winnemucca NV
Portland OR
Seattle
Southern California
San Joaquin Valley
Phoenix
San Fran. Bay
San Jose
Palm Springs
Pocatello, ID
Tuscon
Rillito, AZ
REF
8
8
8
8
8
8
8
8
9,31
10
11
29
29
38
39
40
42
NOTE
a,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
&h
i
J
V
w
t
u
CENTRAL
St. Louis
Harriman
Steubenville
Portage
Topeka
Inglenook AL
Braidwood IL
Kansas City KS
Minneapolis
St. Louis
Kansas City MO
Akron
Cincinnati
Buffalo
Dallas
El Paso
Denver
Chicago
Houston
St.Louis
Harriman
St. Louis
Brownsville
Utah Valley
Ontario
SE Chicago, IL
Ohio
REF
6,7
6,7
6,7
6,7
6,7
8
8
8
8
8
8
8
8
8
8
8
13
15
16
17
17
18
24
26
37
41
43
NOTE
p,q
p,q
p=q
p,q
p,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
a,q
q
s
q
X
1
y
6A-4
-------
FOOTNOTES FOR TABLES 6A-la THROUGH 6A-2c
a. Inhalable Particle Network (IPN) Data. Only represents days of elevated concentrations—dichot filter
loadings >50 (ig/cm2.
b. Data from Site 28 only.
c. Average of all 6-h samples.
d. Avg over all day/nite samples.
e. Average of all 12-h samples at 2 incin. sites and 2 background sites. Only XRF values which exceeded
associated uncertainties more than half the time at all four sites were included.
f. Average from Sparks site and Reno site.
g. Sampling only during intensive episodes.
h. Averages based on 12-h day/nite samples. There were 59 sampling days at Claremont and 23 sampling days
at Long Beach.
i. Avg over all sites: Stockton, Crow's Landing, Fresno, Kern, Fellows, and Bakersfield.
j. Average of Central Phoenix, West Phoenix, and Scottsdale sites.
k. Avg of RAPS site 106.
1. Average from Walpole, Windsor 1, and Windsor 2 sites.
m. Avg of 3 urban sites: Auraria, Federal, and Welby.
n. Median VAPS values from Central site.
o. 2.5-15 (jm.
p. Coarse concentrations may be 30% or more underestimated due to losses from handling filters.
q. PM15.
r. 2.4-20 ^m.
s. No upper size cutoff on VAPS inlet.
t. Average of Palm Springs and Indio, C A.
u. Avg. of Downtown Tuscon, Orange Grove, Cray croft, and Corona de Tuscon sites.
v. Mean of annual avgs (1988-1992) from ~9 sites in Alameda, San Francisco, and Santa Clara counties.
w. 24-h average of day/nite concentrations at two sites in San Jose.
x. PM20. Average from RAPS site 106.
y. Avg. of Follansbee, Mingo, Sewage Plant, Steubenville, and WTOV Tower sites in Ohio.
z. Average of urban sites: Fresno, Bakersfield and Stockton.
aa. Average of nonurban sites: Brighton and Tower.
ab. Castor Avenue site only.
6A-5
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c. BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
1) Smoky Mtn.
2) Shenandoah Valley
3) Abastumani Mtn.
Philadelphia - 3 sites
Philadelphia
Deep Creek Lake
1) Albuquerque
2) Raleigh
3) Boise
4) Roanoke
Portage,
Topeka,Harriman,
Kingston, St. Louis,
Steubenville, Watertown
1) Sept 1978
2) Jul-Aug 1980
3)Mar24-Jull979
Jul 14-Aug 13 1982
Jul25-Augl4 1994
August, 1983
l)Dec 1984-Marl985
2)Janl985-Marl985
3) Dec 1986-Marl987
4)Oct 1988-Feb 1989
1979-1981
Multi-season
F+C(2.5-15),EC, OC, SO;
Nitrate. 12-h samples.
F+C(2.5-10), EC, OC, SO=
N0'3 12-h (0600-1800) and
(1800-0600).
Fine mass, elements, OC, EC,
SD, uncert, from 4 sites
Day/nite sampling (1000-
2200, 2200-1000).
Dichots. FM, CM, OC, EC,
Gases, FP nitrate
F&C (2.5-10) +Carbon,
EOM, VOCs. 12-h samples,
Day/night:
0700-1900,1900-0700.
FP&IP(2.5-15). 24-h
(midnite-midnite), every other
day. No Carbon data.
1) Comparison of avg F&C
composition for 3 sites.
1) F+C composition at site 28.
2) 9-source CMB source app. for site
28.
3) Mass Balance for 3 sites.
1) Measured PM25 mass, OC, EC,
elements, SD, unc. at each site.
1) Mean FP mass, OC, EC, nitrate,
elements stratified by day/nite/all.
1) Mean comp. of F mass, EC, OC
EOM, at 4 sites. 2) daytime/
nightime/24-havgs for key species
at 4 sites.
1) Mean+-SE by city for F+C mass,
metals.
2) Box-line plots by city showing
means and percentiles for F+C mass,
sulfate, Cl.
3) Time-series plots of F+C mass &
tot Sulfate.
4) Data summaries only—no raw data.
No CP data
presented;
Sampling only in
winter; focus on
woodstove impact
Source of info on
geographical and
temporal PM
composition
variability.
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
Harvard 6-cities
IPN study -25 sites.
Los Angeles (SCAQS)
40 locations
Aerosol composition
1) 1977-1985 TSP
2) 1979-1985 PM10
&PM25
3) 1979-1984
Sulfate
Throughout 1980.
Summer (11 episode
days) and fall (7
episode days) 1987
F+C(2.5-15), 24-h sample every
6th day. Only moderately or
highly-loaded samples were
included. No Carbon.
Sequential 4-, 5-, and 7-h PM25
and PM10 on summer episode
days, and 4- and 6-h samples in
fall.
Mass, elements, ions, sulfate,
nitrate, Carbon, ammonium.
1) Table of Mean Air pollution
values for 6 cities: TSP, Inhalable,
Fine, Sulfate. No comp.
1) F+C mass for ~25 sites.
2) F+C mass, composition for 22
sites (No carbon)
1) Avg & Max PM10 and PM25
mass, ions, comp, Cv, Ce stratified
into summer and fall.
2) Plots of temp and spatial
variations of PM2 5 and PM10, PM2 5
nitrate.
3) Cto/EC for some sites
Temp and spatial
variations of PM2 5 and
PM,n
10 San Joaquin Valley
6 sites
Aerosol Composition
Jun 1988-Jun 1989 24-h PM10 & PM, 5 every 6 days.
Mass, elements, ions (K+, SCQ,
NH4+,Na+), EC, OC
1) Summary of annual geometric
avg, arith. avg, max 24-h PM10 and
PM2 5 mass by site.
2) Ann. Avg Mass and comp. for
PM10andPM25by site.
PM10 highest in winter
and dominated by F
mass;C>50%ofPM10
in summer and fall.
Data show spatial and
temporal variations of
PM,nandPM7,
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
11
12
13
Sites
Phoenix PM Study
Phoenix
4 sites
Also comparison
aerosol data from
Denver, Reno, and
Sparks
Denver
Dates
Oct. 1989 -Jan.
1990
Sept. 1989 -Jan.
1990
Jan. 11-30, 1982
Types of Samples
F&C mass, elements, uncertainties from
6 sites
6-h samples, 2x/day, (0600-1200, 1300-
1900)
PM10 & PM25: mass, elements, HNO3,
SO2, NH3, FP NCr3and SO;, ionic
species, OC, EC.
Dichotomous sampler, OC, EC, nitrate,
sulfate
Data
1) temporal variation of PM25 mass
at 4 sites.
2) Mean, SD, & Max: PM25, EC,
OC, NOj SO;, NFfJ and elements for
3 Phoenix sites
3) Same for Denver (1 1/87-1/88)
4) Same for Reno (1 1/86-1/87)
5) Same for Sparks (1 1/86-1-87)
1) Measured PM25 and Coarse,
elements, OC, EC, nitrate, day /night
samples; light extinction.
Comments
Moudi size-
resolved (0- 5.6 (jm
in 9 bins) mass,
NOj so; oc, EC.
Source
apportionment for
F&C particles and
14
oo
15
16
Denver (SCENIC)
Nov. 1987-Jan.
1988
2x daily (0900-1600, 1600-0900). PM25
mass, comp, sulfate, nitrate, OC, EC,
ionic species, gases
Chicago
Houston
July, 1994
Sept. 10-19, 1980
VAPS & Dichot. FM, CM, OC, EC,
elements, SO2, HONO, HNO3.
Dichotomous sampler: 0.1-2.5, 2.5-15. 4
sites. Consecutive 12 h samples.
1) Avg, SD, Min, Max PM25 mass
for 6 sites.
2) Avg, SD, Mm, Max, for PM2 5
mass, ionic species, EC, OC,
elements for 3 sites.
3) Source profiles
4) SCE for 4 sites by day and night
1) Avg VAPS mass, SD, uncert. for
F&C, OC, EC.
1) Average F&C mass, elements,
Carbon, NH^, NO3, Sulfate
extinction.
Source
Apportionment
study
Source
apportionment.
17
St. Louis & Harriman
Sept. 1985-Aug.
1986
Daily F&C (2.5-10pm). Also SO2,
NO,, and O3.
1) Mean, SD, range for PM10, PM25,
SO;, H+, SO,, NO,, O3for both sites.
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
18
19
St. Louis
1) Albuquerque
2) Denver
F(<2.4)&C (2.4-20) 6-12 hr. No
Carbon.
Jul 1976-Aug 1976
(St.Louis)
RAPS data for St.
Louis exist for May
1975-Mar 1977 but
were not in this
article
1) Jan 3-4, 1983 F & C (2.5-10) + Carbon, Nitrate &
2) Jan 19-20, 1982 Sulfate (1C) 12-h samples,
Day/Night:
0700-1900,1900-0700.
1) 2-mo avg of F+C mass, metals, sulfate, for
one site.
2) F+C composition of selected samples
(different sites) during events.
3) CMB apportionment of F+C fractions to 6
components (crustal shale, crustal limestone,
ammonium sulfate, motor vehicles, steel,
paint).
4) Plots ofintercity variations in source
component concentrations
1) Mean daytime and nightime comp. of F&C,
EC, OC, nitrate, sulfate, for each site.
2) Source app. of Denver winter FP
composition.
1) Crude CMB
source
apportionment of
FP with 6
sources.
More complete
source app
results in Lewis
& Enfield paper.
20
21
22
23
24
Charlotte (2 incin sites Apr 30-Jun 4, 1992 VAPS F&C + Acid gases.
and 2 control sites). & Sept 21-28, 1992. no carbon. 12-h samples
Steubenville
Jan-Dec 1984
Review of PM10 studies 1984-1990
Phoenix
Jan 5-27, 1983
Brownsville — Spring+Summer
residential and central 1993
sites.
24-h, F+C. No Carbon
PM,,
F(<2.8)+C(>2.8). 1800-0800 12 h
samples.
1)FP MES indoor/outdoor
2) VAPS central site
3) Dichot central site
1) Mean ambient FP cone. + XRF unc. at 4
sites
2) CMB results for FP.
1) avg F mass + comp.
2) avg source contributions by SRFA
3) SRFA-derived source profiles
1) SCE's for PM10 mass for -15 studies
l)avg F+C nightime comp, mass, Cv,Ce, gases.
2) CMB of FP
1) min, med, max for fine MES comp+mass
2) min, med, max F+C comp, mass for VAPS
and dichot at central site
ambient PM10
data sources are
cited but no data
is presented
No avg values,
only median .
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
25 Sparks, Reno, Verdi, NV Apr 1986-Mar 1987
(SNAPS)
26 Utah Valley (Linden Apr 1985-Dec 1989
site)
27 Santa Clara County 1980-1986: only
Nov, Dec, Jan data
used.
28 San Joaquin Valley Jun 1988-Jun 1989
6 sites
Source apportionment
1) PM25 & PM10 every 6th day. 24-h
samples. Also diurnal sampling.
1) PM10 for 1736 days. Also, SO2,
NO2, O3, acidity data.
"COH" -coefficient of haze.
[COH/PM10=1.87 or 1.64 (1985 and
1986)].
24-h PM10 & PM25 every 6 days.
Mass, elements, ionic species,
Carbon,
1) Seasonal avg SCE for PM10 at 3 sites. No raw data
(geological, motor veh, construction,
vegetative, sulfate, nitrate, OC, EC)
sd=38, (min,max)=(l,365 ,ug/m3).
2) freq distribution of PM10 mass.
1) Plots of COH vs daily mortality for
2-yr periods.
1) Table of ann. avg. SCE to PM10 and
PM2.5 for data above, by site
no comp. data.
Highest pm 10
during winter.
Examines relation
between mortality
and COH
For PM 10 Mass,
Sulfate, and Nitrate
data, see ref 27.
>
o
29
SF Bay Area
2 sites
Los Angeles
(SCAQS)
40 locations
CMB Source Apport.
Dec 16, 1991-Feb
24, 1992
12-h daily day & nite (0600-1800,
1800-0600) PM10 samples.
Mass, elements, ions (K+,C1, SO^ ,
NHj, , Na+) Carbon, ammonium.
Summer (11 episode Sequential 4-, 5-, and 7-h PM25 and
l)Table of ann. avg. PM10 mass, sulfate,
nitrate statistics at 3 sites for 1988-1992
2) Avg. & Max day & nite PM10 mass,
ions, comp, EC, OC, for both sites
3) Source profiles
4) SCE pie charts for each site.
1) Source profiles
days) and fall (7
episode days) 1987
PM10 on summer episode days, and 4- 2) PM10 SCE for summer and fall.
and 6-h samples in fall.
Mass, elements, ions, sulfate, nitrate,
Carbon, ammonium.
3) Diurnal SCE to PM10 at each site.
l.HighestPM10
mass during Nov,
Dec, Jan.
2. Wood combust.
contributes -45%
ofPM10.
Data show diurnal
changes in SCE for
PM10 mass.
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
31
32
>
34
1) Claremont (SCAQS) 1) Summer 1987
(59d)
2) Long Beach (SCAQS) 2) Fall 1987
(23d)
CADMP - 8 sites:
Gasquet, Fremont,
Bakersfield, Yosemite,
Sequoia, Long Beach, Los
Angeles, Azusa
Central California -53
sites in SF Bay area,
Sacramento Valley, San
Joaquin Valley, North and
South Central Coast,
Mountain Counties
Birmingham
Philadelphia
State College, PA
Summer 1988
1) 1989
2) July & August,
1988
1986-1989
1973-1980
summer 1990
Continuous 12-h
PM10 and PM2 5.
Mass, elements, ionic species, EC,
OC
2 samples every 6th day.
0600-01800, 1800-0600.
PM2 5, PM10. Mass, ionic species,
PM10 every 6th day. Sulfate and
nitrate measured on a subset of
these samples.
Daily 24-h PM10 mass. Also Ozone
data.
No composition data.
24-h (midnite-midnite) TSP.
No composition data.
Indoor, outdoor, personal SOJ , H+,
andNLL,
1) Mean, SD, & Max: PM10, FPM,
CPM, EC, OC, NO,, SO; , NH+4 .
2) Mean values of above species
during intensive and non-intensive
periods.
3) Day/nite values of above
4) PM10 and PM2 5 mass balances
5) Summary of EC, OC data.
1) Graph of avg PM10 & PM2 5 mass
and ratio at 8 sites
2) Graphs of PM10 & PM25 ionic
concentrations.
1) 1989 Max and Avg PM10 mass,
Sulfate, and Nitrate for ~53 sites.
2) Summertime 1988 Avg, SD, and
Max PM10 and PM2 5 Mass, comp,
OC,EC, Ionic species, for 3 SJVAQS
sites. [Annual data summary is in ref
20].
1) Table of percentile points of the
distribution of PM10, O3, T, DewPoint,
Pneumonia, Chronic obstructive
pulmonary disease.
2) Avg PM10 and O3 by season
1) Table of percentile points of the
distribution of TSP, SO2, T,
DewPoint, Mortality.
Ask Chow/Watson for
raw data.
Aside: Indoor/Outdoor
ratios of 0.63 for PM10
were reported in Tuscon.
Validation of personal
exposure models
-------
APPENDIX FOR TABLES 6A-la THROUGH 6A-2c (cont'd). BIBLIOGRAPHY FOR PM STUDIES
RefNo.
Sites
Dates
Types of Samples
Data
Comments
37
38-43
44
45-50
Southern Ontario
3 sites
Miscellaneous sites
14 sites
Jan. -Nov., 1991
1984-1990
Allegheny Mm. SW PA July 24-Aug. 10
elev. 838m 1977
Allegheny Mtn. and
Laurel Hill, SW PA
separation 35.5 km
Aug. 5-Aug. 28,
1983
24-h, midnite-midnite, every 6th day.
PM10 dichot sampler.
PM10 concentrations.
Filters, impactors, gas samplers,
day/night
Filters, dichotomous samplers,
impactors, denuders, gas analyzers,
day/night
l)Avg mass, elements, for F&C fractions,
for 3 sites. No OC, EC.
1) Measured PM10 mass and avg source
contributions (up to 10 source
categories).
Aerosol mass, elements, H+, NFTj , SO; ,
NO3, total C, size distributions, bscat, Lv,
gases
Fine, coarse, and PM10 mass, elements,
EC, H+, NH+4 , SO; ,
NO3, size distributions, CN counts, bscat,
babs, Lv, HNO3 and other gases, rain, dew,
2-site correlation
Primary reference
isRef 10.
Strong aerosol H+
found, associated
with so;
Coordinated with
Deep Creek Lake
experiment, Ref
4,^60 km to SSW
References:
to
1. Stevens et al. (1984)
2. Dzubayetal. (1988)
3. Pinto etal. (1995)
4. Vossleretal. (1989)
5. Stevens etal. (1993)
6. Spengler and Thurston (1983)
7. Dockeryetal. (1993)
8. Davis etal. (1984)
9. Chow etal. (1994a)
10. Chow etal. (1993a)
11. Desert Research Institute (1995)
12. Chow etal (1990)
13. Lewis etal. (1986);
Lewis and Dzubay (1986)
14. Watson et al. (1988)
15. Stevens, R. K. (1995) [Unpublished
data].
16. Johnson etal. (1984)
17. Dockery et al. (1992)
18. Dzubay (1980)
19. Stevens (1985)
20. Mukerjee et al. (1993)
21. Koutrakis and Spengler (1987)
22. Chow etal. (1993b)
23. Solomon and Moyers (1986)
24. Ellenson et al. (1994)
25. Chow etal. (1988)
26. Pope etal. (1992)
27. Fan-ley (1990)
28. Chow etal. (1992b)
29. Chow etal. (1995a)
30. Watson et al. (1994a)
31. Wolff etal. (1991)
32. Ashbaugh et al. (1989)
33. Chow etal. (1994b);
Watson etal. (1994b)
34. Schwartz (1994)
35. Schwartz and Dockery
(1992)
36. Suhetal. (1993)
37. Conner etal. (1993)
38. Kim etal. (1992)
39. Houck et al. (1992)
40. Chow etal. (1992a)
41. Vermette et al. (1992)
42. Thanukos et al. (1992)
43. Skidmore etal. (1992)
44. Pierson et al. (1980b)
45. Pierson et al. (1986)
46. Japar et al. (1986)
47. Pierson et al. (1987)
48. Keeler etal. (1988)
49. Pierson et al. (1989)
50. Keeler etal. (1990)
-------
TABLE 6A-2a. PM2, COMPOSITION FOR THE EASTERN UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
1
Smoky Mtn.
9/20-26/78
0-12-24
12
12
24.00
2.22
1.10
0.30
12.00
<0.054
<0.003
0.018
0.016
<0.010
0.003
0.028
0.040
1
Shenandoah
7/23-5/08/80
0-12-24
12
28
27.00
0.44
1.12
13.60
<0.105
<0.003
0.008
0.035
0.010
0.005
0.054
0.061
2(b)
Camden
7/14-8/13, 1982
6-18-6
12
50
28.70
2.05
1.87
<0.48
11.20
0.053
0.001
0.029
0.040
0.002
0.003
0.002
0.091
0.101
0.006
0.001
0.146
0.011
3
Philadelphia
7/25-8/14/94
9-9
24
21
32.18
4.51
0.76
0.114
0.009
0.058
0.026
0.007
0.127
0.060
0.023
0.003
0.070
0.007
4(c)
Deep Creek
8/83
4x daily
6
98
40.00
1.45
0.18
0.57
0.001
0.005
0.048
0.058
0.044
0.003
0.034
46, 49, 50
Allegheny Mtn.
8/5-28/83
day/night
-10
44
49
2
1.2
0.5
17
9
0.058
0.0005
0.0048
0.004
0.027
0.0004
0.061
0.0016
0.0012
0.046
0.041
0.011
0.0032
0.0037
0.036
0.0009
46, 49, 50
Laurel Hill
8/6-27/83
day/night
-10
39
46
2
1.4
0.6
18
10
0.048
0.0006
0.0033
0.004
0.023
0.0004
0.038
0.0011
0.0020
0.062
0.040
0.009
0.0038
0.0031
0.034
0.0011
5(d)
Raleigh
1/85-3/85
7-19-7
12
NR
30.30
10.00
0.50
0.009
0.001
0.028
0.018
0.007
0.020
0.044
0.159
0.003
0.001
5(d)
Roanoke
10/88-2/89
7-19-7
12
NR
19.90
7.30
1.50
0.176
0.002
0.005
0.047
0.053
0.001
0.007
0.114
0.177
0.012
0.001
6,7
Watertown
5/79-6/81
00-24
24
354
14.90
5.85
20.300
0.088
0.041
0.084
0.074
0.004
0.009
8(a)
Hartford
1980
NR
24
2
26.75
0.035
0.036
0.070
0.003
0.043
0.125
0.171
0.007
0.010
8(a)
Boston
1980
NR
24
1
34.80
0.002
0.020
0.070
0.004
0.035
0.121
0.096
0.001
0.012
8(a)
Res.Tr.Pk
1980
NR
24
3
28.77
0.073
0.002
0.007
0.035
0.016
0.120
0.148
0.003
0.001
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE EASTERN UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
1
Smoky Mtn.
9/20-26/78
00-12-24
12
12
0.097
3.744
0.001
0.038
<0.006
<0.004
0.009
1
Shenandoah
7/23-5/08/80
00-12-24
12
28
0.052
4.539
0.001
0.116
<0.010
<0.010
0.011
2(b)
Camden
7/14-8/13 '82
6-18-6
12
50
0.249
4.200
0.079
0.002
0.103
<0.012
<0.002
<0.027
0.013
0.082
3
Philadelphia
7/25-8/14/94
9-9
24
21
0.015
0.019
3.251
<0.002
0.165
<0.042
<0.013
0.041
4(c)
Deep Creek
8/83
4x daily
6
98
0.048
6.700
0.001
0.003
0.150
0.001
0.013
44, 45-50
Allegheny Mtn.
8/5-28/83
day/night
-10
44
0.013
0.035
0.0005
5.9
0.0006
0.0018
0.23
0.0026
0.0041
0.0019
0.010
45-50
Laurel Hill
8/6-27/83
day/night
-10
39
0.019
0.039
0.0002
5.5
0.0006
0.0020
0.21
0.0027
0.0047
0.0017
0.012
5(d)
Raleigh
1/85-3/85
7-19-7
12
NR
0.096
1.729
0.002
0.076
0.003
0.015
5(d)
Roanoke
10/88-2/89
7-19-7
12
NR
0.027
1.177
0.002
0.077
0.004
0.083
6,7
Watertown
5/79-6/81
00-24
24
354
0.329
1.800
0.001
0.100
0.022
8(a)
Hartford
1980
NR
24
2
0.510
2.219
0.001
0.177
0.002
0.017
0.079
8(a)
Boston
1980
NR
24
1
0.009
0.285
3.869
0.001
0.144
0.020
0.046
8(a)
Res.Tr.Pk
1980
NR
24
3
0.042
0.106
2.835
0.002
0.350
0.018
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE WESTERN UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
9(g)
Los Angeles
Summer 1987
NR
4,5 and 7
1 1 days
41.10
8.27
2.37
4.34
9.41
0.035
0.022
0.015
0.013
0.022
0.093
0.022
0.063
0.099
0.041
0.024
0.016
0.202
0.005
9(g)
Los Angeles
Fall 1987
NR
4 and 6
6 days
90.20
18.46
7.28
22.64
4.38
0.250
0.015
0.043
0.065
0.335
0.453
0.025
0.273
0.557
0.217
0.075
0.043
0.466
0.007
10(i)
San Joaquin
Valley
6/88-6/89
NR
24
-35
29.89
4.87
3.24
8.17
3.00
0.152
0.012
0.010
0.096
<0.007
0.094
0.003
0.096
0.180
0.188
0.006
0.016
n(i)
Phoenix
10/13/89-
1/17/90
NR
6 h, 2x/day
-100 days
29.37
10.10
7.47
3.60
1.33
0.130
<0.020
<0.106
0.011
0.170
<0.018
0.365
0.003
0.015
0.216
0.207
0.023
<0.006
0.003
5(d)
Boise
12/86-3/87
7-19-7
12
NR
35.70
12.70
1.70
0.102
0.002
0.014
0.026
0.122
0.001
0.011
0.022
0.145
0.002
0.002
12(f)
Nevada
11/86-1/87
00-24
24
24
56.92
19.97
15.17
2.43
1.67
0.275
0.001
0.013
0.033
0.215
0.145
0.002
0.010
0.310
0.280
0.015
0.006
8(a)
Tarrant CA
1980
NR
24
6
57.05
0.177
0.102
0.455
0.002
0.047
0.316
0.186
0.032
0.003
8(a)
Five Points
CA
1980
NR
24
3
31.80
0.239
0.015
0.150
0.004
0.001
0.024
0.216
0.244
0.005
0.025
8(a)
Riverside
CA
1980
NR
24
4
35.18
0.036
0.037
0.301
0.009
0.040
0.127
0.120
0.007
0.007
8(a)
San Jose
CA
1980
NR
24
6
36.28
0.123
0.001
0.188
0.089
0.050
0.003
0.043
0.148
0.248
0.006
0.006
8(a)
Honolulu
1980
NR
24
1
21.10
1.127
0.017
1.024
0.518
0.004
0.018
0.726
0.371
0.020
0.002
8(a) 8(a)
Winnemucca Portland
1980 1980
NR NR
24 24
5 4
9.68 37.18
0.361 0.581
0.012
0.006 0.093
0.243 0.154
0.021
0.009
0.026 0.072
0.231 0.270
0.149 0.218
0.003 0.052
0.001 0.027
8(a)
Seattle
1980
NR
24
1
10.70
0.002
0.006
0.019
0.037
0.002
0.024
0.098
0.080
0.004
0.006
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE WESTERN UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
9(g)
Los Angeles
Summer 1987
NR
4, 5 and 7
1 1 days
0.060
0.038
2.832
0.013
0.052
0.019
0.005
0.006
0.090
9(g)
Los Angeles
Fall 1987
NR
4 and 6
6 days
0.046
0.185
1.998
0.011
0.520
0.028
0.060
0.007
0.298
10(i)
San Joaquin
Valley
06/88-06/89
NR
24
-35
0.007
0.029
0.001
1.242
<0.002
0.001
0.460
<0.015
0.002
0.017
0.015
0.078
n(i)
Phoenix
10/13/89-
1/17/90
NR
6 h, 2x/day
-100 days
<0.051
0.039
<0.0025
0.437
<0.033
<0.002
0.430
<0.028
<0.030
<0.016
0.056
5(d)
Boise
12/86-3/87
7-19-7
12
NR
0.045
0.603
0.001
0.069
0.001
0.019
12(f)
Nevada
11/86-1/87
00-24
24
24
0.041
0.115
0.001
0.765
0.860
0.004
0.043
0.009
0.033
8(a)
Tarrant CA
1980
NR
24
6
0.619
2.578
0.583
0.010
0.095
8(a)
Five Points
CA
1980
NR
24
3
0.007
0.087
1.129
0.001
0.656
0.005
0.006
0.016
8(a)
Riverside
CA
1980
NR
24
4
0.376
1.653
0.001
0.234
0.003
0.029
8(a)
San Jose
CA
1980
NR
24
6
0.013
0.891
0.852
0.292
0.002
0.061
8(a)
Honolulu
1980
NR
24
1
0.002
0.071
0.313
2.363
0.063
0.001
0.011
8(a)
Winnemucca
1980
NR
24
5
0.042
0.358
0.914
0.009
0.011
8(a)
Portland
1980
NR
24
4
0.017
0.422
1.944
0.001
0.377
0.005
0.014
0.081
8(a)
Seattle
1980
NR
24
1
0.006
0.215
0.831
0.001
0.092
0.059
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE CENTRAL UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
5(d)
Albuquerque
12/84-3/85
7-19-7
12
NR
20.60
13.20
2.10
0.077
0.085
0.059
0.036
0.045
0.074
13
Denver
1/11-30/82
6-18-6
12
-26
20.73
7.11
2.15
2.22
2.06
0.394
<0.002
0.031
0.103
0.047
0.006
0.052
<0.009
0.010
0.079
0.079
0.011
0.003
14(m) 14(aa)
Urban Denver Non-urban Denver
11/87-1/88 11/87-1/88
9-16-9 9-16-9
7&17 7&17
-136 -150
19.67 10.35
7.25
4.41
3.96
1.55
0.037
0.018
0.058
0.005
0.141
0.003
0.017
0.111
0.077
0.012
0.002
15
Chicago
7/94
8-8
24
16
13.57
5.39
1.31
0.046
<0.003
<0.091
0.004
0.045
<0.029
0.011
<0.005
0.011
0.089
0.061
0.012
0.005
<0.002
0.022
<0.001
16
Houston
9/10-19/80
NR
12
20
38.60
5.68
1.42
0.59
14.61
0.123
<0.005
0.048
0.055
0.155
<0.003
0.032
<0.005
0.028
0.162
0.119
0.014
<0.38
0.004
6,7
Harriman
5/80-5/81
00-24
24
256
20.80
8.10
36.1
0.038
0.150
0.021
0.120
0.017
BQL
17 6,7 6,7
Harriman Kingston Portage
9/85-8/86 5/80-6/81 3/79-5/81
NR 00-24 00-24
24 24 24
330 169 271
21.00 24.60 11.00
8.70 4.95
36.1 10.5
0.044 0.011
0.120 0.045
BQL 0.027
0.097 0.049
0.010 0.003
BQL BQL
6,7
Topeka
8/79-5/81
00-24
24
286
12.50
4.40
11.6
0.045
0.250
0.031
0.090
0.004
BQL
8(a)
El Paso
1980
NR
24
10
27.16
0.155
0.025
0.070
0.332
0.001
0.036
0.134
0.127
0.004
0.001
8(a)
Inglenook
1980
NR
24
8
32.03
0.082
0.001
0.040
0.326
0.003
0.002
0.032
0.281
0.408
0.037
0.001
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE CENTRAL UNITED STATES
a\
oo
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
5(d)
Albuquerque
12/84-3/85
7-19-7
12
NR
0.237
0.507
0.076
0.007
13
Denver
1/11-30/82
6-18-6
12
-26
0.043
0.326
<0.003
0.709
0.277
<0.003
<0.027
0.046
14(m) 14(aa)
Urban Denver Non-urban Denver
11/87-1/88 11/87-1/88
9-16-9 9-16-9
7&17 7&17
-136 -150
0.075
0.642
0.004
0.001
0.272
0.006
0.001
0.009
0.031
15
Chicago
7/94
8-8
24
16
0.008
0.027
1.321
<0.042
<0.001
0.074
<0.049
<0.029
<0.009
0.052
16
Houston
9/10-19/80
NR
12
20
0.028
0.465
<0.002
4.834
0.006
<0.002
0.210
<0.005
<0.002
<0.014
<0.008
0.084
6,7
Harriman
5/80-5/81
00-24
24
256
0.180
2.500
0.002
0.120
BQL
17 6,7
Harriman Kingston
9/85-8/86 5/80-6/81
NR 00-24
24 24
330 169
0.194
2.400
0.002
0.200
BQL
6,7
Portage
3/79-5/81
00-24
24
271
0.061
1.400
0.001
0.075
BQL
6,7
Topeka
8/79-5/81
00-24
24
286
0.163
1.100
0.190
BQL
8(a)
El Paso
1980
NR
24
10
0.481
0.823
0.002
0.436
0.003
0.055
8(a)
Inglenook
1980
NR
24
8
0.008
0.309
2.655
0.001
0.685
0.133
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE CENTRAL UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
8(a) 8(a)
Braidwood Kansas City KS
1980 1980
NR NR
24 24
1 8
28.20 25.66
0.089 0.091
0.003
0.003 0.027
0.084 0.519
0.004
0.024 0.032
0.071 0.189
0.052 0.311
0.001 0.006
0.001 0.002
8(a) 8(a)
Minneapolis Kansas City MO
1980 1980
NR NR
24 24
6 3
15.50 16.77
0.004 0.007
0.047 0.064
0.103 0.213
0.001 0.002
0.035 0.021
0.087 0.140
0.092 0.142
0.005 0.006
0.001 0.001
8(a)
Akron
1980
NR
24
7
36.09
0.046
0.012
0.039
0.110
0.010
0.037
0.609
0.268
0.085
0.006
8(a)
Cincinnati
1980
NR
24
2
29.80
0.062
0.013
0.024
0.062
0.003
0.024
0.174
0.136
0.011
0.004
8(a)
Buffalo
1980
NR
24
14
38.75
0.192
0.009
0.003
0.218
0.002
0.026
0.671
0.310
0.033
0.008
8(a)
Dallas
1980
NR
24
4
28.93
0.111
0.033
0.223
0.691
0.005
0.043
0.248
0.125
0.015
0.002
8(a)
St. Louis
1980
NR
24
5
23.06
0.119
0.003
0.025
0.090
0.018
0.076
0.126
0.002
0.002
18(k)
St. Louis
8-9/76
NR
6-12
NR
34.00
0.203
0.002
0.020
0.132
0.132
0.004
0.087
0.006
0.029
0.275
0.261
0.036
0.004
6,7
St. Louis
9/79-6/81
00-24
24
306
19.00
7.40
10.3
0.078
0.101
0.052
0.190
0.021
0.003
17 6,7
St. Louis Steubenville
9/85-8/86 4/79-4/81
NR 00-24
24 24
311 499
17.70 29.60
8.00 10.94
9.7 25.2
0.042
0.097
0.092
0.590
0.029
0.005
-------
TABLE 6A-2a (cont'd). PM2, COMPOSITION FOR THE CENTRAL UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
8(a)
Braidwood
1980
NR
24
1
0.041
2.060
0.001
0.220
0.011
8(a)
Kansas City KS
1980
NR
24
8
0.013
0.180
1.816
0.001
0.434
0.004
0.034
8(a)
Minneapolis
1980
NR
24
6
0.308
0.907
0.001
0.169
0.045
8(a)
Kansas City MO
1980
NR
24
3
0.369
0.763
0.177
0.046
8(a)
Akron
1980
NR
24
7
0.059
0.412
3.419
0.008
0.522
0.009
0.150
8(a)
Cincinnati
1980
NR
24
2
0.043
0.343
2.876
0.005
0.328
0.003
0.053
8(a)
Buffalo
1980
NR
24
14
0.060
0.359
3.706
0.005
0.241
0.001
0.078
8(a)
Dallas
1980
NR
24
4
0.018
1.066
1.514
0.442
0.007
0.002
0.054
8(a)
St. Louis
1980
NR
24
5
0.020
0.277
2.333
0.002
0.170
0.023
18(k)
St. Louis
8-9/76
NR
6-12
NR
0.001
0.688
4.655
0.006
0.004
0.458
0.009
0.002
0.112
0.002
0.101
6,7
St. Louis
9/79-6/81
00-24
24
306
0.327
2.100
0.002
0.160
BQL
17 6,7
St. Louis Steubenville
9/85-8/86 4/79-4/81
NR 00-24
24 24
311 499
0.216
4.700
0.005
0.290
0.011
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2b. COARSE PARTICLE COMPOSITION FOR THE EASTERN UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
1(0)
Smoky Mtn.
9/20-26/78
NR
12
12
5.60
<0.300
<0.001
0.005
0.322
<0.012
<0.005
0.118
0.108
<0.002
1(0)
Shenandoah
7/23-5/08/80
NR
12
28
7.40
0.78
0.311
<0.002
0.003
0.304
0.179
0.006
0.158
0.129
<0.006
<0.003
2(b)
Camden
7/14-8/13 '82
6-18-6
12
50
11.40
<3.00
0.42
0.57
<0.90
0.550
0.015
0.360
<0.006
0.069
<0.009
0.490
0.151
0.011
0.004
3(ab)
Philadelphia
7/25-8/14/94
NR
24
21
8.42
0.325
0.003
0.421
0.047
0.014
0.352
0.100
0.104
0.006
0.136
0.002
4(c) 46,49,50
Deep Creek Allegheny Mtn.
8/83 8/5-28/83
4x daily day/night
6 -10
98 44
15
0.39
0.0002
0.007
0.0011
0.27
0.0004
0.044
0.0014
0.0016
0.24
0.11
0.060
0.0063
0.0026
0.054
0.0008
46,49,50 5(d)
Laurel Hill Raleigh
8/6-27/83 1/85-3/85
day/night 7-19-7
-10 12
39 NR
13
0.39
0.0002
0.006
0.0011
0.28
0.0003
0.039
0.0015
0.0025
0.24
0.10
0.061
0.0068
0.0021
0.044
0.0009
5(d) 6,7(o,p)* 8(a,o)
Roanoke Watertown Hartford
10/88-2/89 5/79-6/81 1980
7-19-7 00-24 NR
12 24 24
NR 354 2
9.30 27.85
0.65
1.875
0.022 0.046
0.209 0.864
0.305 0.302
0.008
0.026
0.276 1.070
0.310
0.006 0.021
0.005
8(a,o)
Boston
1980
NR
24
1
105.60
3.458
0.001
0.025
1.069
0.301
0.004
0.023
1.612
0.533
0.029
0.022
8(a,o)
Res.Tr.Pk
1980
NR
24
3
8.17
0.606
0.003
0.086
0.002
0.010
0.182
0.068
0.003
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE EASTERN UNITED STATES
Oi
to
to
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
1(0)
Smoky Mtn.
9/20-26/78
NR
12
12
0.014
<0.560
<0.0006
0.580
0.018
<0.004
1(0)
Shenandoah
7/23-5/08/80
NR
12
28
0.009
<0.711
<0.001
0.813
0.017
0.006
2(b)
Camden
7/14-8/13 '82
6-18-6
12
50
0.054
0.230
0.181
<0.0015
1.610
<0.009
0.002
0.065
0.007
0.030
3(ab)
Philadelphia
7/25-8/14/94
NR
24
21
0.027
0.013
BQL
BQL
0.933
0.030
BQL
0.052
4(c) 46,49,50
Deep Creek Allegheny Mtn.
8/83 8/5-28/83
4x daily day/night
6 -10
98 44
0.006
0.007
0.0004
0.59
0.0002
0.0003
1.48
0.0029
0.029
0.0011
0.010
46,49,50 5(d)
Laurel Hill Raleigh
8/6-27/83 1/85-3/85
day/night 7-19-7
-10 12
39 NR
0.007
0.007
0.0005
0.56
0.0002
0.0003
1.41
0.0025
0.027
0.0010
0.011
5(d) 6,7(o,p)* 8(a,o)
Roanoke Watertown Hartford
10/88-2/89 5/79-6/81 1980
7-19-7 00-24 NR
12 24 24
NR 354 2
0.033
0.076 0.171
0.200 0.428
1.000 4.517
0.094
0.008
0.054
8(a,o)
Boston
1980
NR
24
1
0.016
0.177
0.502
6.760
0.154
0.008
0.054
8(a,o)
Res.Tr.Pk
1980
NR
24
3
0.013
0.223
1.387
0.021
0.007
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE WESTERN UNITED STATES
a\
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
9(g)'
Los Angeles
Summer
1987
NR
4,5 and 7
1 1 days
26.30
3.34
0.82
5.13
1.87
0.723
BQL
0.055
0.003
0.563
1.026
0.002
BQL
0.737
0.196
0.311
0.017
1.431
BQL
9(g)'
Los Angeles
Fall 1987
NR
4 and 6
6 days
8.50
4.89
1.21
4.86
1.01
0.597
0.004
0.084
0.006
0.854
0.426
0.017
BQL
1.635
0.243
0.212
0.021
0.052
BQL
10(i)'
San Joaquin
Valley
6/88-6/89
NR
24
-35
44.17
5.71
2.38
2.38
0.62
3.418
0.040
0.006
0.961
0.393
0.007
BQL
1.453
0.632
0.031
BQL
H(i)
Phoenix
10/13/89-
1/17/90
NR
6 h, 2x/day
-100 days
33.09
4.46
0.84
0.86
0.37
2.539
<0.002
<0.077
0.002
1.929
<0.016
0.194
0.008
0.021
1.259
0.669
0.032
<0.005
0.003
5(d) 12(f) 8(a,o)
Tarrant
Boise Nevada CA
12/86-3/87 11/86-1/87 1980
7-19-7 00-24 NR
12 24 24
NR 24 6
43.85
2.230
0.047
4.088
0.005
0.030
0.941
0.255
0.035
0.003
8(a,o)
Five
Points CA
1980
NR
24
3
92.57
7.078
0.004
1.636
0.022
0.006
0.013
3.059
1.193
0.050
0.012
8(a,o)
Riverside
CA
1980
NR
24
4
71.03
3.513
0.028
4.781
0.164
0.005
0.021
1.888
0.961
0.042
0.006
8(a,o)
San Jose
CA
1980
NR
24
6
30.40
1.930
0.062
0.682
0.430
0.006
0.028
1.066
0.260
0.021
0.008
8(a,o)
Honolulu
1980
NR
24
1
25.80
1.865
0.006
0.957
0.938
0.005
0.007
0.658
0.294
0.014
0.003
8(a,o)
Winnemucca
1980
NR
24
5
55.74
6.564
0.004
1.934
0.176
0.006
0.017
1.764
1.051
0.041
0.002
8(3,0)
Portlsnd
1980
NR
24
4
80.38
6.351
0.002
0.028
1.305
0.176
0.010
0.037
1.789
0.587
0.056
0.009
8(3,0)
Se3ttle
1980
NR
24
1
25.30
2.294
0.002
0.014
0.548
0.228
0.003
0.017
0.903
0.151
0.018
0.001
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE WESTERN UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
ON si
*f" Sn
to
•J^ Sr
Ti
V
Zn
9(g)*
Los Angeles
Summer
1987
NR
4,5 and 7
1 1 days
0.127
0.046
0.520
BQL
1.988
BQL
0.072
BQL
0.024
9(g)'
Los Angeles
Fall 1987
NR
4 and 6
6 days
0.053
0.066
0.264
BQL
1.642
BQL
0.106
0.003
BQL
10(i)'
San Joaquin
Valley
6/88-6/89
NR
24
-35
0.052
0.032
0.222
7.577
0.012
0.130
BQL
0.016
H(i)
Phoenix
10/13/89-
1/17/90
NR
6 h, 2x/day
-100 days
0.038
0.022
0.003
0.178
<0.030
<0.002
7.013
<0.026
0.014
0.121
<0.014
0.034
5(d) 12(f) 8(a,o)
Tarrant
Boise Nevada CA
12/86-3/87 11/86-1/87 1980
7-19-7 00-24 NR
12 24 24
NR 24 6
0.002
0.167
0.310
5.208
0.083
0.052
8(a,o)
Five Points
CA
1980
NR
24
3
0.148
0.018
0.293
16.001
0.272
0.007
0.016
8(a,o)
Riverside
CA
1980
NR
24
4
0.144
0.113
0.720
7.544
0.182
0.030
8(a,o)
San Jose
CA
1980
NR
24
6
0.032
0.228
0.257
5.214
0.086
0.044
8(a,o)
Honolulu
1980
NR
24
1
0.022
0.258
3.766
0.067
0.008
8(a,o)
Winnemucca
1980
NR
24
5
0.021
0.215
11.903
0.164
0.015
8(a,o)
Portland
1980
NR
24
4
0.011
0.115
0.427
12.128
0.186
0.004
0.038
8(a,o)
Seattle
1980
NR
24
1
0.077
0.121
4.332
0.091
0.034
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE CENTRAL UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Oi
j> Ba
^ Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
5(d) 13(o)
Albuquerque Denver
12/84-3/85 1/11-30/82
7-19-7 6-18-6
12 12
NR -26
35.73
0.39
2.900
0.058
0.024
0.658
0.012
1.235
<0.009
0.008
0.954
0.648
0.021
0.005
14(m) 14(ab) 15(s)
Urban Denver Non-urban Denver Chicago
11/87-1/88 11/87-1/88 7/94
9-16-9 9-16-9 8-8
7&17 7&17 24
-136 -150 16
14.97
0.223
<0.0013
<0.038
0.007
0.716
<0.012
0.036
<0.0024
0.006
0.344
0.101
0.106
0.008
<0.0017
<0.017
<0.0007
16(o) 6,7(o,p)* 17 6,7(o,p)*
Houston Harriman Harriman Kingston
9/10-19/80 5/80-5/81 9/85-8/86 5/80-6/81
NR 00-24 NR 00-24
12 24 24 24
20 256 330 169
24.80 11.70 9.00 10.80
3.10
1.63
0.91
1.093
<0.006
0.091
0.036 0.014 0.012
2.780 1.650 0.840
<0.006
0.366 0.029 0.018
0.007
0.018
0.604 0.570 0.263
0.170
0.021 0.021 0.018
<0.74
0.004 0.001 BQL
6,7(o,p)* 6,7(o,p)* 8(a,o)
Portage Topeka El Paso
3/79-5/81 8/79-5/81 1980
00-24 00-24 NR
24 24 24
271 286 10
7.20 13.90 49.05
0.35 0.40
2.748
0.012
0.003 0.010 0.033
0.335 2.150 3.632
0.056 0.043
0.003
0.047
0.181 0.490 0.812
0.496
0.006 0.016 0.023
0.001 0.001 0.001
8(3,0)
Inglenook
1980
NR
24
8
40.43
2.426
0.021
2.598
0.004
0.027
1.193
0.309
0.041
0.002
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE CENTRAL UNITED STATES
ON
to
ON
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
5(d) 13(o)
Albuquerque Denver
12/84-3/85 1/11-30/82
7-19-7 6-18-6
12 12
NR -26
0.113
0.099
0.005
<0.48
7.460
0.009
0.090
0.039
14(m) 14(ab) 15(s)
Urban Denver Non-urban Denver Chicago
11/87-1/88 11/87-1/88 7/94
9-16-9 9-16-9 8-8
7&17 7&17 24
-136 -150 16
0.027
0.005
0.043
<0.017
<0.0006
0.739
<0.021
0.019
<0.004
0.038
16(o)
Houston
9/10-19/80
NR
12
20
<0.1
0.124
<0.003
<1.29
<0.009
2.990
<0.009
<0.008
0.036
<0.03
0.058
6,7(o,p)* 17 6,7(o,p)*
Harriman Harriman Kingston
5/80-5/81 9/85-8/86 5/80-6/81
00-24 NR 00-24
24 24 24
256 330 169
0.057 0.040
BQL BQL
1.880 1.700
6,7(o,p)*
Portage
3/79-5/81
00-24
24
271
0.013
BQL
0.905
6,7(o,p)*
Topeka
8/79-5/81
00-24
24
286
0.040
BQL
2.310
8(a,o)
El Paso
1980
NR
24
10
0.191
0.249
0.001
5.377
0.077
0.057
8(3,0)
Inglenook
1980
NR
24
8
0.022
0.079
0.314
6.312
0.116
0.055
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE CENTRAL UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Oi
j> Ba
^> Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
8(a,o)
Braidwood
1980
NR
24
1
28.70
1.931
0.002
0.003
1.406
0.002
0.020
0.656
0.303
0.017
0.001
8(a,o)
Kansas City KS
1980
NR
24
8
41.67
2.284
0.003
0.029
3.754
0.530
0.004
0.015
0.979
0.361
0.025
0.002
8(a,o)
Minneapolis
1980
NR
24
6
30.85
2.191
0.001
0.022
1.571
0.293
0.002
0.022
0.744
0.310
0.026
0.001
8(a,o)
Kansas City MO
1980
NR
24
3
41.67
2.284
0.003
0.029
3.754
0.530
0.004
0.015
0.979
0.361
0.025
0.002
8(a,o)
Akron
1980
NR
24
7
34.81
2.509
0.003
0.025
1.431
0.572
0.014
0.018
1.640
0.324
0.044
0.005
8(a,o)
Cincinnati
1980
NR
24
2
33.15
2.910
0.017
1.312
0.103
0.002
0.014
0.883
0.363
0.021
0.003
8(a,o)
Buffalo
1980
NR
24
14
44.57
2.808
0.012
2.550
0.728
0.015
0.022
2.040
0.206
0.078
0.009
8(a,o)
Dallas
1980
NR
24
4
32.63
1.294
0.006
0.051
3.436
0.029
0.005
0.023
0.720
0.210
0.020
0.002
8(a,o)
St. Louis
1980
NR
24
5
33.76
3.837
0.001
0.021
1.784
0.053
0.001
0.014
0.587
0.291
0.017
0.002
18(k,r) 6,7(o,p)* 17 6,7(o,p)*
St. Louis St. Louis St. Louis Steubenville
8-9/76 9/79-6/81 9/85-8/86 4/79-4/81
NR 00-24 NR 00-24
6-12 24 24 24
306 311 499
28.00 12.40 9.90 16.90
0.70 1.86
1.209
0.001
0.034
0.047 0.021 0.010
2.817 1.499 1.023
0.001
0.257 0.093 0.211
0.009
0.014
1.218 0.580 1.610
0.392
0.035 0.019 0.039
0.005 0.002 0.004
-------
TABLE 6A-2b (cont'd). COARSE PARTICLE COMPOSITION FOR THE CENTRAL UNITED STATES
Oi
K>
oo
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
8(a,o)
Braidwood
1980
NR
24
1
0.014
0.013
0.572
0.001
5.767
0.083
0.012
8(a,o)
Kansas City KS
1980
NR
24
8
0.109
0.280
4.809
0.074
0.040
8(a,o)
Minneapolis
1980
NR
24
6
0.098
0.224
4.679
0.062
0.027
8(a,o)
Kansas City MO
1980
NR
24
3
0.109
0.280
4.809
0.074
0.040
8(a,o)
Akron
1980
NR
24
7
0.097
0.451
5.009
0.107
0.069
8(a,o)
Cincinnati
1980
NR
24
2
0.037
0.099
0.389
6.633
0.096
0.148
8(a,o)
Buffalo
1980
NR
24
14
0.108
0.765
2.675
0.051
0.043
8(a,o)
Dallas
1980
NR
24
4
0.252
0.240
3.210
0.051
0.030
8(a,o)
St. Louis
1980
NR
24
5
0.095
0.279
4.468
0.058
0.021
18(k,r)
St. Louis
8-9/76
NR
6-12
0.098
0.189
0.002
0.533
0.001
0.001
4.470
0.001
0.007
0.475
0.004
0.074
6,7(o,p)*
St. Louis
9/79-6/81
00-24
24
306
0.088
0.200
1.940
BQL
17 6,7(o,p)*
St. Louis Steubenville
9/85-8/86 4/79-4/81
NR 00-24
24 24
311 499
0.043
0.800
2.010
0.002
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2c. PM,n COMPOSITION FOR THE EASTERN UNITED STATES
,n
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Oi
j> Ba
fe Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
l(o,q)*
Smoky Mtn.
9/20-26/78
NR
12
12
29.60
2.22
1.10
0.30
12.00
BQL
BQL
0.023
0.338
BQL
0.003
0.146
0.148
BQL
l(o,q)'
Shenandoah
7/23-5/08/80
NR
12
28
34.40
0.44
1.12
14.38
0.311
0.011
0.339
0.189
0.011
0.212
0.190
BQL
BQL
2(b)*
Camden
7/14-8/13 '82
6-18-6
12
50
40.10
2.05
2.29
0.57
11.20
0.603
0.001
0.044
0.400
0.002
0.072
0.002
0.581
0.252
0.017
0.001
0.146
0.015
3(ab)* 4(c)
Philadelphia Deep Creek
7/25-8/14/94 8/83
NR 4x daily
24 6
21 98
40.60
4.51
0.76
0.439
0.012
0.479
0.073
0.021
0.479
0.160
0.126
0.010
0.206
0.009
5(d) 5(d) 6,7(p,q) 8(a,q)'
Raleigh Roanoke Watertown Hartford
1/85-3/85 10/88-2/89 5/79-6/81 1980
7-19-7 7-19-7 00-24 NR
12 12 24 24
NR NR 354 2
24.20 54.60
6.50
1.910
0.110 0.082
0.250 0.934
0.389 0.302
0.011
0.069
0.350 1.195
0.481
0.009 0.028
0.011 0.015
8(a,q)*
Boston
1980
NR
24
1
140.40
3.458
0.003
0.045
1.139
0.301
0.008
0.058
1.733
0.629
0.030
0.034
8(a,q)*
Res.Tr.Pk
1980
NR
24
3
36.93
0.679
0.002
0.010
0.121
0.002
0.026
0.302
0.216
0.006
0.001
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE EASTERN UNITED STATES
Oi
OJ
o
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
l(o,q)*
Smoky Mtn.
9/20-26/78
NR
12
12
0.111
3.744
0.001
0.618
0.018
BQL
0.009
l(o,q)*
Shenandoah
7/23-5/08/80
NR
12
28
0.061
4.539
0.001
0.929
0.017
BQL
0.017
2(b)*
Camden
7/14-8/13/82
6-18-6
12
50
0.303
4.430
0.260
0.002
1.713
BQL
0.002
0.065
0.020
0.112
3(ab)* 4(c)
Philadelphia Deep Creek
7/25-8/14/94 8/83
NR 4x daily
24 6
21 98
0.042
0.032
3.251
1.098
0.030
0.092
5(d) 5(d) 6,7(p,q)
Raleigh Roanoke Watertown
1/85-3/85 10/88-2/89 5/79-6/81
7-19-7 7-19-7 00-24
12 12 24
NR NR 354
0.405
2.000
0.001
1.100
0.022
8(a,q)*
Hartford
1980
NR
24
2
0.033
0.681
2.647
0.001
4.694
0.096
0.025
0.133
(a,q)*
Boston
1980
NR
24
1
0.025
0.462
4.371
0.001
6.904
0.154
0.028
0.100
8(a,q)*
Res.Tr.Pk
1980
NR
24
3
0.042
0.119
3.058
0.002
1.737
0.021
0.025
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE WESTERN UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
9(g)
Los Angeles
Summer 1987
NR
4,5 and 7
1 1 days
67.40
11.61
3.19
9.47
11.28
0.758
0.007
0.070
0.016
0.585
1.119
0.023
0.022
0.836
0.237
0.335
0.033
1.632
0.005
9(g)
Los Angeles
Fall 1987
NR
4 and 6
6 days
98.70
23.35
8.49
27.50
5.39
0.847
0.019
0.127
0.072
1.190
0.880
0.042
0.178
2.192
0.460
0.287
0.063
0.518
0.005
10(i)
San Joaquin Valley
Jun. 1998- Jun. 1989
NR
24
-35
74.05
10.59
5.62
10.55
3.62
3.570
0.051
0.015
1.057
0.487
0.010
0.087
1.633
0.820
0.037
0.010
n(i)
Phoenix
10/13/89-1/17/90
NR
6 h, 2x/day
-100 days
62.45
14.56
8.30
4.46
2.34
2.669
BQL
0.013
0.014
2.099
BQL
0.559
0.011
0.036
1.475
0.876
BQL
0.054
BQL
BQL
0.006
5(d)* 12(f) 8(a,q)*
Tarrant
Boise Nevada CA
12/86-3/87 11/86-1/87 1980
7-19-7 00-24 NR
12 24 24
NR 24 6
100.90
2.407
0.149
4.543
0.007
0.077
1.257
0.441
0.067
0.006
8(a,q)*
Five Points
CA
1980
NR
24
3
124.37
7.317
0.019
1.786
0.026
0.007
0.037
3.275
1.437
0.055
0.037
8(a,q)*
Riverside
CA
1980
NR
24
4
106.20
3.549
0.065
5.082
0.173
0.005
0.061
2.015
1.081
0.049
0.013
8(a,q)*
San Jose
CA
1980
NR
24
6
66.68
2.053
0.001
0.250
0.771
0.480
0.009
0.071
1.214
0.508
0.027
0.014
8(a,q)*
Honolulu
1980
NR
24
1
46.90
2.992
0.023
1.981
1.456
0.009
0.025
1.384
0.665
0.034
0.005
8(a,q)*
Winnemucca
1980
NR
24
5
65.42
6.925
0.010
2.177
0.176
0.006
0.043
1.995
1.200
0.044
0.003
8(a,q)*
Portland
1980
NR
24
4
117.55
6.932
0.014
0.121
1.459
0.197
0.019
0.109
2.059
0.805
0.108
0.036
8(a,q)*
Seattle
1980
NR
24
1
36.00
2.296
0.008
0.033
0.585
0.228
0.005
0.041
1.001
0.231
0.022
0.007
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE WESTERN UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
5 Sn
U> Sr
" Ti
V
Zn
9(g)
Los Angeles
Summer 1987
NR
4,5 and 7
1 1 days
0.187
0.084
3.353
0.008
2.040
0.018
0.077
0.005
0.114
9(g)
Los Angeles
Fall 1987
NR
4 and 6
6 days
0.099
0.251
2.262
0.010
2.162
0.024
0.165
0.009
0.293
10(i)
San Joaquin Valley
Jun. 1988 -Jun. 1989
NR
24
-35
0.059
0.061
0.004
1.463
0.001
8.037
0.014
0.147
0.014
0.094
n(i)
Phoenix
10/13/89-1/17/90
NR
6 h, 2x/day
-100 days
0.054
0.062
BQL
0.615
BQL
BQL
7.443
BQL
0.014
0.136
BQL
0.090
5(d)* 12(f) 8(a,q)*
Tarrant
Boise Nevada CA
12/86-3/87 11/86-1/87 1980
7-19-7 00-24 NR
12 24 24
NR 24 6
0.002
0.786
2.888
5.791
0.093
0.147
8(a,q)*
Five Points
CA
1980
NR
24
3
0.155
0.105
1.422
0.001
16.657
0.277
0.013
0.032
8(a,q)*
Riverside
CA
1980
NR
24
4
0.144
0.489
2.373
0.001
7.778
0.182
0.003
0.059
8(a,q)*
San Jose
CA
1980
NR
24
6
0.045
1.119
1.109
5.506
0.086
0.002
0.105
8(a,q)*
Honolulu
1980
NR
24
1
0.002
0.093
0.571
6.129
0.130
0.001
0.019
8(a,q)*
Winnemucca
1980
NR
24
5
0.063
0.573
12.817
0.173
0.026
8(a,q)*
Portland
1980
NR
24
4
0.028
0.537
2.371
0.001
12.505
0.191
0.018
0.119
8(a,q)*
Seattle
1980
NR
24
1
0.006
0.292
0.952
0.001
4.424
0.091
0.093
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE CENTRAL UNITED STATES
Ref 8(a,q)*
Site Albuquerque
Dates 12/84-3/85
Time 7-19-7
Duration (h) 12
Number NR
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Ba
ON Br
> Ca
% Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
P
Pb
13(q)*
Denver
1/11-30/82
6-18-6
12
-26
56.46
7.11
2.15
2.22
2.45
3.294
<0.004
0.089
0.127
0.705
0.018
1.287
<0.018
0.018
1.033
0.727
0.031
0.008
0.155
0.424
14(m) 14(aa) 15(s)*
Urban Denver Non-urban Denver Chicago
11/87-1/88 11/87-1/88 7/94
9-16-9 9-16-9 8-8
7&17 7&17 24
-136 -150 16
28.54
5.39
1.31
5.46
0.269
<0.0043
<0.130
0.011
0.761
<0.041
0.047
<0.0073
0.017
0.432
0.161
0.118
0.013
<0.0041
0.022
<0.0018
0.035
0.032
16(q)*
Houston
9/10-19/80
NR
12
20
63.40
8.78
1.42
2.22
15.52
1.216
<0.015
0.139
0.091
2.935
<0.012
0.398
0.007
0.046
0.766
0.289
0.035
<1.49
0.008
0.128
0.589
6,7(p,q) 17*
Harriman Harriman
5/80-5/81 9/85-8/86
00-24 NR
24 24
256 330
32.50 30.00
8.10 8.70
36.1
0.052
1.800
0.050
0.690
0.038
0.001
0.237
6,7(p,q)
Kingston
5/80-6/81
00-24
24
169
35.40
0.056
0.960
0.018
0.360
0.027
ND
0.234
6,7(p,q)
Portage
3/79-5/81
00-24
24
271
18.20
5.30
0.014
0.380
0.083
0.230
0.009
0.001
0.074
6,7(p,q)
Topeka
8/79-5/81
00-24
24
286
26.40
4.80
0.055
2.400
0.031
0.580
0.020
0.001
0.203
8(a,q)*
El Paso
1980
NR
24
10
76.21
2.903
0.037
0.103
3.964
0.043
0.004
0.083
0.946
0.623
0.027
0.002
0.672
8(a,q)*
Inglenook
1980
NR
24
8
72.45
2.508
0.001
0.061
2.924
0.003
0.006
0.059
1.474
0.717
0.078
0.003
0.030
0.388
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE CENTRAL UNITED STATES
ON
>
OJ
rv
Ref 8(a,q)*
Site Albuquerque
Dates 12/84-3/85
Time 7-19-7
Duration (h) 12
Number NR
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
13(q)*
Denver
1/11-30/82
6-18-6
12
-26
0.005
0.709
<0.004
<0.004
7.737
<0.004
0.009
0.09
<0.004
0.085
14(m) 14(aa) 15(s)*
Urban Denver Non-urban Denver Chicago
11/87-1/88 11/87-1/88 7/94
9-16-9 9-16-9 8-8
7&17 7&17 24
-136 -150 16
1.363
<0.059
<0.0017
0.813
<0.070
0.019
<0.013
0.090
16(q)*
Houston
9/10-19/80
NR
12
20
<0.006
4.83
0.006
<0.003
3.200
0.036
<0.045
0.142
6,7(p,q) 17*
Harriman Harriman
5/80-5/81 9/85-8/86
00-24 NR
24 24
256 330
2.500
0.002
2.000
ND ND
6,7(p,q)
Kingston
5/80-6/81
00-24
24
169
2.400
0.002
1.900
ERR
6,7(p,q)
Portage
3/79-5/81
00-24
24
271
1.500
0.001
0.980
ND
6,7(p,q)
Topeka
8/79-5/81
00-24
24
286
1.200
2.500
ND
8(a,q)*
El Paso
1980
NR
24
10
1.072
0.003
5.813
0.080
0.112
8(a,q)*
Inglenook
1980
NR
24
8
2.969
0.001
6.997
0.116
0.188
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE CENTRAL UNITED STATES
Ref
Site
Dates
Time
Duration (h)
Number
Mass
OC
EC
Nitrate
Sulfate
Acidity5
Al
As
Oi
j> Ba
^ Br
Ca
Cd
Cl
Cr
Cu
Fe
K
Mg
Mn
Mo
Na
Ni
8(a,q)*
Braidwood
1980
NR
24
1
56.90
2.020
0.002
0.006
1.490
0.002
0.044
0.727
0.355
0.018
0.002
8(a,q)*
Kansas City KS
1980
NR
24
8
70.33
2.144
0.003
0.036
4.371
0.010
0.048
0.989
0.660
0.026
0.005
8(a,q)*
Minneapolis
1980
NR
24
6
46.35
2.191
0.005
0.069
1.674
0.293
0.003
0.057
0.831
0.402
0.031
0.002
8(a,q)*
Kansas City MO
1980
NR
24
3
58.43
2.284
0.010
0.093
3.967
0.530
0.006
0.036
1.119
0.503
0.031
0.003
8(a,q)*
Akron
1980
NR
24
7
70.90
2.555
0.015
0.064
1.541
0.572
0.024
0.055
2.249
0.592
0.129
0.011
8(a,q)*
Cincinnati
1980
NR
24
2
62.95
2.972
0.013
0.041
1.374
0.103
0.005
0.038
1.057
0.499
0.032
0.007
8(a,q)*
Buffalo
1980
NR
24
14
83.32
3.000
0.009
0.015
2.768
0.728
0.017
0.048
2.711
0.516
0.111
0.017
8(a,q)*
Dallas
1980
NR
24
4
61.55
1.405
0.039
0.274
4.127
0.029
0.010
0.066
0.968
0.335
0.035
0.004
8(a,q)*
St. Louis
1980
NR
24
5
56.82
3.956
0.004
0.046
1.874
0.053
0.001
0.032
0.663
0.417
0.019
0.004
18(x)* 6,7(p,q)
St. Louis St. Louis
8-9/76 9/79-6/81
NR 00-24
6-12 24
306
62.00 31.40
8.10
1.412
0.003
0.054
0.179 0.099
2.949 1.600
0.005
0.344 0.145
0.015
0.043
1.493 0.770
0.653
0.071 0.040
0.009 0.005
17* 6,7(p,q)
St. Louis Steubenville
9/85-8/86 4/79-4/81
NR 00-24
24 24
311 499
27.60 46.50
8.00 12.80
9.7
0.052
1.120
0.303
2.200
0.068
0.008
-------
TABLE 6A-2c (cont'd). PM,n COMPOSITION FOR THE CENTRAL UNITED STATES
ON
OJ
ON
Ref
Site
Dates
Time
Duration (h)
Number
P
Pb
Rb
S
Sb
Se
Si
Sn
Sr
Ti
V
Zn
8(a,q)*
Braidwood
1980
NR
24
1
0.014
0.054
2.632
0.002
5.987
0.083
0.023
8(a,q)*
Kansas City KS
1980
NR
24
8
0.013
0.237
2.031
0.001
4.976
0.076
0.060
8(a,q)*
Minneapolis
1980
NR
24
6
0.406
1.131
0.001
4.848
0.062
0.072
8(a,q)*
Kansas City MO
1980
NR
24
3
0.478
1.043
4.986
0.074
0.086
8(a,q)*
Akron
1980
NR
24
7
0.059
0.509
3.870
0.008
5.531
0.116
0.219
8(a,q)*
Cincinnati
1980
NR
24
2
0.080
0.442
3.265
0.005
6.961
0.099
0.201
8(a,q)*
Buffalo
1980
NR
24
14
0.060
0.467
4.471
0.005
2.916
0.051
0.001
0.121
8(a,q)*
Dallas
1980
NR
24
4
0.018
1.318
1.754
3.652
0.058
0.002
0.084
8(a,q)*
St. Louis
1980
NR
24
5
0.020
0.372
2.612
0.002
4.638
0.058
0.044
18(x)*
St. Louis
8-9/76
NR
6-12
0.099
0.877
0.002
5.188
0.007
0.005
4.928
0.010
0.009
0.587
0.006
0.175
6,7(p,q)
St. Louis
9/79-6/81
00-24
24
306
0.415
2.300
0.002
2.100
ND
17* 6,7(p,q)
St. Louis Steubenville
9/85-8/86 4/79-4/81
NR 00-24
24 24
311 499
0.259
5.500
0.005
2.300
0.013
References are listed in Table 1 Appendix. Associated notes are explained in Table 1.
'Values for this size fraction are calculated from the average measured values reported for the other two size fractions.
'Units for acidity are nmoles/m3.
NR = not reported.
-------
TABLE 6A-3. SELECTED RATIOS OF PM COMPOSITION BY
GEOGRAPHIC REGION
EAST
FM/CM
FM/PM10
Tot Carbon/FM
SOJ/FM
Mean
2.59
0.65
0.25
0.34
N
8
8
7
12
WEST
Mean
0.89
0.41
0.54
0.11
N
11
11
5
13
CENTRAL
Mean
1.06
0.51
0.64
0.28
N
25
25
5
28
N = number of studies contributing to the calculated ratios.
FM, CM, PM10 = Mass concentrations of PM25, Coarse fraction, and PM10 respectively.
Total Carbon = (OC x 1.4 + EC).
6A-37
-------
TABLE 6A-4a. SITE-TO-SITE VARIABILITY OF PM2. CONCENTRATIONS
Study Area
No. of Sites
Study Dates
Reference
Fine Mass
OC
EC
Nitrate
Sulfate
Al
Br
Ca
Cl
Cr
£ Cu
w Fe
°° K
Mn
Ni
Pb
S
Si
Ti
V
Zn
Denver Metropolitan
3,a
11/2/87- 1/31/88
14
Mean
19.672
7.245
4.409
3.956
1.547
0.037
0.018
0.058
0.141
0.003
0.017
0.111
0.077
0.012
0.002
0.075
0.642
0.272
0.009
0.031
Spread
2.889
0.789
0.780
0.931
0.162
0.006
0.006
0.001
0.013
0.002
0.008
0.023
0.009
0.003
0.002
0.017
0.077
0.009
0.001
0.008
Phoenix
3,b
10/13/89- 1/17/90
11
Mean
29.379
10.089
7.490
3.597
1.329
0.131
0.011
0.167
0.366
0.003
0.015
0.216
0.209
0.023
0.003
0.039
0.436
0.430
0.056
Spread
3.493
2.690
1.710
0.370
0.240
0.015
0.003
0.034
0.356
0.001
0.003
0.035
0.020
0.010
0.001
0.009
0.038
0.066
0.030
Philadelphia
4,c
7/25/94 - 8/14/94
3
Mean
32.183
4.164
0.685
13.426
0.114
0.009
0.058
0.026
0.007
0.127
0.060
0.003
0.007
0.019
3.251
0.165
0.019
0.041
Spread
2.172
0.935
0.215
0.333
0.009
0.005
0.014
0.007
0.001
0.037
0.008
0.000
0.002
0.010
0.081
0.022
0.003
0.018
San Joaquin Valley
6,d
6/14/88-6/9/89
10
Mean
29.888
4.873
3.242
8.165
3.003
0.152
0.010
0.096
0.094
0.003
0.096
0.180
0.188
0.006
0.016
0.029
1.242
0.460
0.017
0.015
0.078
Spread
10.020
2.695
2.580
2.270
1.325
0.055
0.006
0.050
0.070
0.002
0.036
0.060
0.080
0.003
0.030
0.021
0.565
0.245
0.004
0.028
0.027
Mean = Mean over all sites of the average concentrations determined at each site for the sampling period.
Spread = ABS ({Highest Mean Cone. - Lowest Mean Conc.}/2) for all the sites.
a. Federal, Auraria, and Welby sites in urban Denver.
b. Central Phoenix, Scottsdale, and Western Phoenix sites.
c. Broad Street, Castor Avenue, Roxboro, and Northeast Airport sites.
d. Stockton, Crow's Landing, Fresno, Kem, Fellows, and Bakersfield sites.
-------
TABLE 6A-4b. SITE-TO-SITE VARIABILITY OF PM,n CONCENTRATIONS
10
Study Area
No. of Sites
Study Dates
Reference
Fine Mass
OC
EC
Nitrate
Sulfate
Al
Br
Ca
Cl
Cr
Cu
Fe
K
Mn
Ni
Pb
S
Si
Ti
V
Zn
San Jose
2,a
12/16/91 -
29
Mean
64.950
19.390
9.015
10.900
2.240
0.845
0.012
0.670
0.728
0.003
0.029
0.834
0.823
0.014
0.003
0.035
1.147
2.905
0.088
0.007
0.065
2/24/92
Spread
1.650
0.150
0.415
0.600
0.090
0.035
0.001
0.049
0.032
0.001
0.002
0.027
0.021
0.001
0.000
0.004
0.091
0.045
0.024
0.003
0.005
Phoenix
3,b
10/13/89-
11
Mean
62.465
14.549
8.327
4.459
1.704
2.670
0.014
2.096
0.559
0.011
0.036
1.475
0.878
0.054
0.006
0.062
0.615
7.442
0.121
0.090
1/17/90
Spread
7.064
3.481
1.777
0.452
0.287
0.273
0.003
0.317
0.349
0.002
0.009
0.170
0.083
0.014
0.002
0.013
0.041
0.862
0.024
0.034
San Joaquin Valley
6,c
6/14/88 -
10
Mean
62.920
7.870
3.505
9.437
3.565
2.993
0.012
0.950
0.388
0.009
0.084
1.413
0.720
0.030
0.019
0.039
1.472
7.517
0.128
0.022
0.085
6/9/89
Spread
17.280
4.150
2.760
3.015
1.460
1.570
0.005
0.390
0.225
0.003
0.046
0.445
0.220
0.011
0.032
0.027
0.605
1.765
0.033
0.031
0.029
Mean = Mean over all sites of the average concentrations determined at each site for the sampling period.
Spread = ABS ({Highest Mean Cone. - Lowest Mean Cone. }/2) for all the sites.
a. San Carlos St. and Fourth St. sites.
b. Central Phoenix, Scottsdale, and Western Phoenix Sites.
c. Stockton, Crow's Landing, Fresno, Kern, Fellows, and Bakersfield sites.
6A-39
-------
7. HUMAN EXPOSURE TO PARTICULATE MATTER:
RELATIONS TO AMBIENT AND INDOOR
CONCENTRATIONS
7.1 INTRODUCTION
The 1982 Air Quality Criteria Document for Particulate Matter and Sulfur Oxides
(U.S. Environmental Protection Agency, 1982) thoroughly reviewed the PM exposure literature
through 1981. The later "Second Addendum to Air Quality Criteria for Parti culate Matter and
Sulfur Oxides (1982)" (U.S. Environmental Protection Agency, 1986a) added coverage of newly
available health effects information up to 1986. This chapter first summarizes key points from
the 1982 Criteria Document, and then thoroughly reviews the PM exposure literature from 1982
through 1995 and includes some literature published and in press through February, 1996.
The U.S. Environmental Protection Agency (U.S. EPA) regulatory authority for PM only
extends to the ambient air, defined in 40 CFR 50.1(e) as that portion of the atmosphere, external
to buildings, to which the general public has access (Code of Federal Regulations, 1994). By the
operative definition of ambient air, polluted air inside a building, or on private property owned
or controlled by the source of pollution, is not regulated by the National Ambient Air Quality
Standards (Costle, 1980; Bennett, 1983). However, it is necessary to consider total personal
exposure to PM, both from the regulated ambient air and non-regulated indoor air. This is
because ambient (outdoor) particles penetrate into non-ambient environments (indoors) where
people spend approximately 85% of their time (U.S. Environmental Protection Agency, 1989).
Therefore, when people are indoors, they are exposed to a mixture of ambient PM and particles
generated indoors from non-regulated sources, such as PM from cigarette smoke and personal
activities.
Personal exposure to total PM is important in itself, because the body may react differently
to ambient and non-ambient particles of identical size but different chemical composition.
Comparison of personal exposures to indoor and outdoor concentrations may provide clues as to
whether or not these two types of PM have similar toxicity on a unit size and mass basis.
Personal exposure may also act as a confounder in epidemiological studies which use an inferred
community exposure to ambient PM as a parameter to correlate with community health
7-1
-------
parameters, and an individual's personal exposure to total PM is a critical parameter for analysis
if that person is a member of a cohort whose health outcomes are being tracked individually.
Therefore, this chapter examines not only indoor air quality in regard to PM, but also
community and individual exposures to PM, which include that portion of ambient PM which
penetrates into indoor microenvironments (|iEs). This is to aid in interpretation of acute and
chronic epidemiology studies assessed in Chapter 12, in which ambient PM concentrations are
assumed to be an indicator or a surrogate for mean community exposure to ambient PM or an
individual exposure to ambient PM. Thus, this chapter has three objectives: (a) to provide a
review of pertinent studies of indoor and personal exposures to PM; (b) to evaluate linkages
between monitored personal exposures and exposures estimated from a fixed-site monitor
located at some central monitoring site; and (c) to quantify the contribution of ambient air to
personal PM exposure.
In this chapter, Sections 7.1.1 - 7.1.3 discuss the concept of ambient PM as a surrogate for
a personal exposure and the relationship of a measured personal PM exposure to the ambient and
nonambient concentrations of PM that may influence it.
Section 7.2 next reviews PM concentrations found indoors where people spend about 85%
of their time (U.S. Environmental Protection Agency, 1989). This subject is discussed in detail
because of the importance of indoor conditions for understanding total exposure to PM. Indoor
air particles from indoor sources may be an important factor in the analysis and interpretation of
epidemiology studies, because they may influence both the personal PM exposure and personal
health of the exposed people.
Section 7.2.5 reviews the literature covering biological aerosols, which may produce direct
health effects or act as a source of antigens capable of sensitizing people to the effects of other
PM exposures.
Section 7.3 reviews the fundamental principles of personal PM monitoring and factors that
influence the personal PM measurement.
Section 7.4 covers the literature on direct measurements of personal exposures to PM and
PM constituents such as sulfates.
Section 7.5 reviews the literature on indirect exposure estimation procedures that predict
exposures from time-weighted averages of concentrations measured indoors and outdoors.
7-2
-------
Section 7.6 discusses the relationship of individual PM exposures to ambient PM
concentrations and establishes a linkage between average personal PM exposures in a
community to the ambient PM concentrations.
Section 7.7 discusses implications of PM exposure relationships for mortality and
morbidity analyses.
Section 7.8 provides a Summary of Conclusions for Chapter 7.
7.1.1 Ambient Particulate Matter Concentration as a Surrogate for
Particulate Matter Dosage
The health effects of PM experienced by an individual depend upon the mass, size and
composition of those particles deposited within various regions of the respiratory tract during the
time interval of interest. The amount of this potential dose will depend on the concentration
inhaled (e.g., the instantaneous personal exposure); the ventilation rate (a function of physical
activity and basal metabolism); and the fractional deposition, which is a function of ventilation
rate, mode of breathing (e.g., oral or nasal), and any alterations due to lung dysfunction. If all
people had identical ventilation rates and deposition patterns, then the potential-dosage
distribution could be linearly scaled to the personal exposure distribution which would serve as a
suitable primary surrogate. The usage of ambient PM concentration in health studies as a
surrogate for personal PM exposure, and thereby a secondary surrogate for the PM dosage,
would be suitable if ambient concentration was also linearly related to the personal exposure
(Mage, 1983).
Adult ventilation rates are lowest (mean ~ 6 L/min) during the night while asleep, at a
maximum (mean ~ 12 L/min; peak ~ 60 L/min) during the day while awake (Adams, 1993), and
in phase with PM exposure, which is also lower at night than during the day (Clayton et al.,
1993). Consequently, the product of the 24-h average PM exposure, the 24-h average ventilation
rate, and the average deposition parameter for the average ventilation would seriously
under-predict the amount of PM deposited in the respiratory tract (Mage, 1980).
In practice, when relating human health to PM pollution variables (as in Chapter 12) one is
forced to use time-weighted-average (TWA) ambient PM concentration as a surrogate for PM
exposure and PM dosage because only fragmentary data are typically available on personal
exposures to PM in populations. Data are also limited on ventilation rates as a function of basal
-------
metabolism and physical activities (Adams, 1993), as are data on pulmonary deposition rates of
particles people are inhaling, since the size distribution is unknown and deposition is affected by
unmeasured individual physiological parameters. According to Hodges and Moore (1977),
"even when an explanatory variable (ambient PM concentration) can be measured with
negligible error it may often be standing as a proxy for some other variable (dosage) which
cannot be measured directly, and so it (dosage) is subject to measurement error". Pickles (1982)
shows "that (such) uncertainties in air pollution levels lead to two kinds of error in the air
pollution/mortality regression coefficient - a systematic underestimate and a random scatter". In
addition, measurement error can also bias a threshold in the dose-response function towards zero
(Yoshimura, 1990).
In the sections that follow, the relationships between ambient PM concentration, indoor
PM concentrations and personal exposures to PM are discussed in detail. The following five
caveats should be kept in mind while reading this chapter:
1. Ambient PM concentrations are often measured as a 24-h time-weighted-average
(TWA) expressed as |ig/m3. This quantity, by necessity, is assumed to be a surrogate
for the mass of ambient PM deposited in people's respiratory tracts per unit body
weight, expressed as |ig/kg-day.
2. This daily quantity of ambient PM deposited per unit body weight is in turn a surrogate
for the amount of the true (but unknown) species and/or size fraction of the total PM
that is the specific etiologic toxic agent(s) that act by a presently unknown mechanism.
This latter quantity should be the independent variable for delineating underlying
relationships between ambient PM TWA concentrations to the health indices used as
the dependent variables.
3. Virtually all analyses and discussions of exposure presented here are based on personal
exposure to PM of non-smokers. Only Dockery and Spengler (1981b) included
6 smokers out of 37 subjects. Smokers are often excluded from these studies because a
personal exposure monitor (PEM) on a smoker will not capture the main-stream
tobacco smoke that is directly inhaled. In Section 7.2 on indoor air pollution, it is
shown that side-stream environmental tobacco smoke (ETS) is the largest identifiable
indoor source of PM where smoking occurs. For the average smoker, the amount of
direct inhalation (several milligrams of PM per cigarette) can be two-to-three orders of
magnitude greater than the microgram amounts of ETS which the PEM captures
(Federal Trade Commission, 1994). The relationships presented below, of ambient PM
concentration to individual total PM exposure, therefore only apply to non-smokers.
4. A total TWA personal exposure to PM (ambient PM plus indoor PM) will be a poor
surrogate for the personal exposure to PM of ambient origin for those people whose
7-4
-------
personal exposures are dominated by indoor (residential and occupational) sources,
such as ETS.
5. All studies of indoor concentrations and personal exposures described below evaluated
subjects recruited either in a nonrandom manner or in a scientific probability sampling
scheme. In the former case, the results cannot be extrapolated with confidence beyond
the subjects themselves. In the latter case, the results can be extrapolated with a known
confidence to the target population from which the sample was drawn. However, in
both cases, there is a cohort of people who are nonresponders. If the reason for their
refusal to participate in the survey is directly or indirectly related to their PM exposure,
then the study results represent a sample with a bias of unknown sign and magnitude.
7.1.2 General Concepts for Understanding Particulate Matter Exposure and
Microenvironments
Particulate matter represents a generic class of pollutants which requires a different
interpretation of exposure in contrast to that for the other specific criteria gaseous pollutants,
such as CO (Mage, 1985). Whereas a molecule of CO emitted from a motor vehicle is
indistinguishable from a molecule of CO emitted from a fireplace, a l-|im aerodynamic diameter
(AD) particle emitted from a motor vehicle and a l-|im particle emitted from a fireplace can
have a different shape, mass, chemical composition, and/or toxicity. Thus, a "particle" can be a
single entity, or an agglomeration of smaller particles, such as a small Pb particle bound to a
larger crustal particle. Furthermore, indoor sources of particles produce a wide variety of
particles of varying size and composition that people are exposed to, as shown in Figure 7-1
(Owen et al., 1992). Given that the health effects of inhalation of any particle can depend upon
its mass and chemical composition, it would be of use to measure PM exposure in terms of mass
and chemical composition as a function of size distribution (Mage, 1985).
The total PM exposure of an individual during a period of time is composed of exposure to
many different particles from various sources in different microenvironments (|iE). A //E was
defined by Duan (1982) as "a chunk of air space with homogeneous
7-5
-------
Particle Diameter (um)
0.
Plant
Animal
Mineral
Combustion
Home/
Personal
Care
Radioactive
01 0.1 1 10 100 1,000
^ Soores ^ -* sPanish "°« Pollei^
j Mold ^
:> ~«4 Starches ^
JPM^? JUH^ ___
Carbon Black (£=
^ Pudding M lv -4 Corn Cob Chaffe. ^
•* SnuTT ^.^ Sawdust ^
Bfct.rlophjg. ^ -« Dropl.t Mu.l.l ». •„,„,. ^
^ Epithelial Cells (human)
^ ^^ *" ^ pray Dried Milk ^ ^ Duct Miti*
^ i-0CDS Bone Dust ^>
Asbestos
<4 Coal Dust ^
•4 Clay ^
• Cnocial Ut. Insulatiprti Fib.rrrtaBS Glass WocH
%jrning Wood>.
Coal Flue 6as ^~ ^" Smok» Fly Ash
>• ^^ hunting Aid
^ ^ I>^ --*!•• -y P«»il Spray Paint Dust
Paint Pigments ^ ^_^^.
Alkali Fume ^ Insecticide B«JG.t& ^ -*^^ Emollient^
. Pigment Binrlar^ ^ ^
Cop.er Toner: ^ ^ Artificial Tertile Fibers
Lint
4 ^
Man-made mineral fibers
Figure 7-1. Sizes of various types of indoor particles.
Source: Owen etal. (1992).
-------
pollutant concentration"; it has also been defined (Mage, 1985) as a volume in space, during a
specific time interval, during which the variance of concentration within the volume is
significantly less than the variance between that |iE and its surrounding jiEs. For example, a
kitchen with a wood stove can constitute a single jiE for total PM when the stove is off, and all
people in the kitchen would have similar PM exposures. When the stove is in operation, the
kitchen could have a significant vertical PM concentration gradient and a child on the floor in a
far corner and an adult standing at the stove could be exposed to significantly different PM
concentrations.
In a given jiE, such as one in the kitchen example, the particles may come from a wide
variety of sources. PM may be generated from within (e.g. the stove, deep frying, burning
toast), from without (ambient PM entering through an open window), from another indoor jiE
(cigarette smoke from the living room), or from a personal activity that generates a
heterogeneous mix of PM (sweeping the kitchen floor and resuspending a mixture of PM from
indoor and outdoor sources that had settled out).
In general, as people move through space and time, they pass through a series of jiEs and
their average total exposure (X |ig/m3) to PM for the day can be expressed by the following
equation,
X = SXiti/St; (7-1)
where X; is the total exposure to PM in the ith |iE, visited in sequence by the person for a time
interval t; (Mage, 1985).
With appropriate averaging over sets of 4 classes of jiEs (e.g., indoors, ambient-outdoors.
occupational, and in-traffic) Equation 7-1 can be simplified as follows (Mage, 1985):
X = (Xfc tin + Xout tout + Xocc tocc + X^ ttra) / T (7-2)
where each value of X is the mean value of total PM concentration in the jiE class while the
subject is in it, time (t) is the total time the subject is in that |iE during the day, and T is equal to
the sum of all times (usually one day). Similar equations may be written for personal exposures
to particles from specific sources (e.g., diesel soot), for specific chemicals (e.g., Pb), or for
specific size intervals (PM < 2.5 jim AD).
Many excellent studies have reported data on air quality concentrations in jiE settings that
do not meet a rigorous definition of an exposure, which requires actual occupancy by a person
7-7
-------
(Ott, 1982). Section 7.2, on Indoor Concentrations and Sources of PM, cites Thatcher and
Layton (1995) who report that "merely walking into a room increased the particle concentration
by 100%". Consequently, an integrated measurement of air quality in an enclosed space that
includes time when it is unoccupied may not be a valid measure that can be used to estimate an
exposure while occupied. If this measure includes periods of time when the space is unoccupied,
it will tend to be biased low as a measure of the exposure within it during periods of occupancy.
For example, it is incorrect to associate an average PM exposure to a person while cooking at a
stove in a kitchen with a kitchen concentration measurement that is influenced by periods when
the stove was off (Smith et al., 1994).
The literature on 24-h average PM concentrations in indoor jiEs, such as residential
settings, is treated separately in Section 7.2, as is done for 24-h average ambient PM
concentrations in Chapter 6. In the exposure portion of this chapter, specific reference is made
to some studies where simultaneous personal PM exposures and indoor PM measurements have
been made, so that the relationship between indoor concentration and personal exposure can be
examined.
In practice, a cascade sampler can collect ambient PM samples by size fractionation for
separate chemical analyses, but such a complete definition of personal exposure to PM by
chemistry and size is difficult to obtain. Although some personal monitors can be equipped with
a cyclone or impactor separator and several filters to capture several PM sizes (e.g., <2.0 jim, 2.0
to 10 |im, and >10 //m; Tamura et al., 1996), most published studies of PM exposure used a
PEM with a single integrated measurement of particle mass collected (e.g., <2.5 jim or <10 jim).
Consequently, health studies on individuals are usually only able to develop associations
between their observed health effects and their observed exposure expressed as an integral mass
of PM collected and its average chemical composition.
Health studies on populations can make multiple measurements of ambient and indoor PM
concentrations simultaneously (e.g., PM25, PM10, TSP) along with components of PM, such as
polycyclic aromatic hydrocarbons (PAHs), to help understand the size distribution and chemistry
of the particles in the ambient and indoor atmospheres.
-------
7.1.3 Summary of State-of-Knowledge in the 1982 Criteria Document
In 1982 it was known, from personal monitoring and indoor monitoring, that SO2 is almost
always lower indoors than outdoors because of the virtual absence of indoor sources and the
presence of sinks for SO2 in indoor settings (exceptions can occur if high sulfur coal or kerosene
are used as fuel in a poorly vented stove or space heater). However, this relationship does not
hold for PM, as the indoor and personal monitoring data show both higher- and lower-than
ambient PM concentrations in indoor settings as a function of particle size and human activity
patterns.
The largest coarse mode particles (>10 |im), which are generally of nonanthropogenic
origin (wind blown dust, etc.), require turbulence to provide vertical velocity components
greater than their settling velocity to allow them to remain suspended in the air (Figure 7-1).
Outdoor particles enter into an indoor setting either by bulk flow, as through an open window, in
which all particles can enter at the inlet condition, or by pressure driven drafts and diffusional
flows through cracks and fissures in the barriers of the building envelope when all windows are
closed. In the latter mode of entry, velocities are relatively lower, thereby settling out the largest
coarse particles (>25 //m AD) in the passage through the barriers (Thatcher and Layton, 1995).
Indoor settings are usually quiescent (Matthews et al., 1989), and ambient particles that
enter indoors quickly settle out by gravity or electrostatic forces, leading to familiar dust layers
on horizontal surfaces and vertical TV screens that require constant cleaning (Raunemaa et al.,
1989). However, human activity in indoor settings, such as smoking, dusting, vacuuming and
cooking, does generate fine particles (<2.5 jim) and coarser particles (>2.5 jim) and resuspends
coarse particles (>10 //m) that previously had settled out (Thatcher and Layton, 1995; Litzistorf
etal., 1985).
Only three studies of personal PM exposures, compared to ambient PM concentrations,
were referenced in the 1982 Criteria Document (U.S. Environmental Protection Agency, 1982).
Binder et al. (1976) reported that "outdoor air measurements do not accurately reflect the air
pollution load experienced by individuals who live in the area of sampling", in a study in
Ansonia, CT, where personal exposures to PM5 were double the outdoor PM concentrations
measured as TSP (115 versus 58 //g/m3). Spengler et al. (1980) was cited as reporting that
"there was no correlation [R2 = 0.04] between the outdoor level [of respirable particles] and the
personal exposure of individuals" in a study in Topeka, KS. Figure 7-2, from Repace et al.
7-9
-------
(1980), was cited as an example of the variability of PM exposures which show very little
influence of ambient concentration. Thus, at the time of the 1982 Criteria Document, two major
factors were known to influence ambient PM relationships to indoor PM air quality: (1) the
variability of indoor levels of PM compared to outdoor concentrations as a function of particle
size (e.g., fine indoor > fine outdoor, and coarse indoor < coarse outdoor); and (2) the variation
of exposures of individuals related to different activities involved in local generation of particles
in their immediate surroundings (smoking, traffic, dusting and vacuuming at home, etc.). This
understanding was summarized on pg. 5-136 of the 1982 Criteria Document, as follows:
• long term personal exposures to fine fraction PM (<2.5 jim) of outdoor origin, may be
estimated by ambient measurements of the <2.5 jim PM fraction.
• Personal activities and indoor concentrations cause personal exposures to PM to vary
substantially. Ambient measurements appear to be a poor predictor of personal
exposure to PM.
• Tobacco smoke is an important contributor to indoor concentrations and personal
exposures where smoking takes place (U.S. Environmental Protection Agency, 1982).
7.2 INDOOR CONCENTRATIONS AND SOURCES OF PARTICULATE
MATTER
7.2.1 Introduction
Although EPA regulates particles in ambient air, which excludes the air internal to
buildings, it is still important to consider indoor air. Most people spend most of their time
indoors. A U.S. Environmental Protection Agency (1989) report indicates that U.S. residents
spend 85.2% of their time indoors, 7.4% in or near a vehicle, and only 7.4% outdoors. Also, it
is important to understand how outdoor particles are affected as they cross building envelopes.
For a home with no indoor sources, how much protection is offered against particles of various
size ranges? How do parameters such as volume of the house, air exchange rate, cleaning
frequency and methods, and materials in the home affect
7-10
-------
280
260
240
220
"E
I200
^ 180
c
~ 160
I 140
S 120
o
o
o
100
80
60
40
20
0
• Indoors
• In Transit
O Outdoors
Well-Ventilated Kitche
Cafeteria, Smoking Sectio
Behind Smoky Diesel True
Outside Cigar
Smoker's Office
uburbs
Vehicle
In City
Commutin
Bedroom
^m ^m ^m ^m ^
Street Suburbs, Outdo
Cafeteria,
Nonsmoking
Section
Sidewalk
BusExhaust
**~ Livin
ommuting Room ""
Suburbs _
U°9ging
I I I I I
City, Outdoor
Library, Unoccupied Cafeteria Livina Room
12 1 2
Midnight
34 567 89 10 11 12 1 234 567 89101112
A.M. Noon P.M.
Time of Day
Figure 7-2. An example of personal exposure to respirable particles.
Source: Repace et al. (1980).
-------
concentrations of particles of outdoor origin? This section has several parts that address these
questions.
The first part (7.2.2; 7.2.3; and 7.2.4) deals with field studies of particles indoors and
outdoors, focussing mainly on large-scale surveys of many homes and buildings. Besides
presenting observed indoor and outdoor particle concentrations, information on important
parameters such as air exchange rates, source emission rates, and deposition rates is also
reported. This section also discusses a few studies dealing with inorganic and organic
constituents of particles, as well as other considerations such as the role of house dust in
exposure to metals. Section 7.2.3 provides a brief introduction to indoor air quality models.
Finally, Section 7.2.4 summarizes the main findings.
The second part (7.2.5) is a discussion of bioaerosols from plants, molds, insects, etc.
Although these sources of PM are uncontrolled by EPA, they affect measured PM indoors and
can potentiate the effects of PM from other sources through allergenic properties.
In keeping with EPA's regulatory responsibilities, the many studies in industrial
workplaces and the "dusty trades" are omitted, as are studies of lead (Pb) in indoor locations,
since lead is a separate criteria pollutant and such studies are reviewed in a separate lead criteria
document (U.S. Environmental Protection Agency, 1986b).
7.2.2 Concentrations of Particles in Homes and Buildings
At least seven major reviews of field studies of indoor particles have been published since
1980 (Sterling et al., 1982; National Research Council, 1986; Repace, 1987; Guerin et al., 1992;
U.S. Environmental Protection Agency, 1992; Holcomb, 1993; Wallace, 1996). The last of
these reviews reports on several recently completed important studies, including EPA's major
probability-based PTEAM Study. Since the two microenvironments where people spend the
most time are (a) home and (b) work or school, studies of these environments are summarized in
turn, with emphasis on the former.
7.2.2.1 Particle Concentrations in Homes: Large-Scale Studies in the United States
There have been three large-scale studies (greater than 150 homes) of airborne particles
inside U.S. homes. In chronological order, these are:
1. The Harvard Six-City study, carried out by the Harvard School of Public Health from
1979 through 1988, with measurements taken in 1,273 homes;
7-12
-------
2. The New York State ERDA study, carried out by Research Triangle Institute (RTI) in
433 homes in two New York State counties during 1986;
3. The EPA Particle TEAM (PTEAM) study, carried out by RTI and Harvard School of
Public Health in 178 homes in Riverside, CA in 1990.
The findings of each are discussed in detail, since these studies present the most complete
investigations to date of indoor and outdoor concentrations of particles.
7.2.2.1.1 The Harvard Six-City Study
The Harvard Six-City Study is a prospective epidemiological study of health effects of
particles and sulfur oxides. Focused mainly on children, it has included pulmonary function
measurements on more than 20,000 persons in the six cities, chosen to represent low (Portage,
WI and Topeka, KS), medium (Watertown, MA and Kingston-Harriman, TN), and high
(St. Louis, MO and Steubenville, OH) outdoor particle and sulfate concentrations.
The study took place in two measurement phases. The first involved monitoring of about
10 homes in each city for respirable particles (PM3 5), with measurements made every sixth day
(24-h samples) for one to two years. In the second phase, a larger sample of 200 to 300 homes
was selected from each city, with week-long PM2 5 samples collected both indoors and outdoors
during two weeks of sampling in summer and winter. Ultimately, more than 1,200 homes were
monitored in this way.
Spengler et al. (1981) described the first five years of the study. During the Phase I period,
pulmonary function measurements were made for 9,000 adults, and 11,000 children in grades 1
through 6. In each home, a 24-h sample (beginning at midnight) was collected every sixth day,
using a cyclone sampler with a cut point of -3.5 jim at a flow rate of 1.7 Lpm. About 10 sites in
each city were kept in operation for two years. The annual mean indoor and outdoor PM3 5
concentrations are shown in Figure 7-3. The indoor concentrations exceeded the outdoor levels
in all cities except Steubenville, OH, where the outdoor levels of about 46 |ig/m3 slightly
exceeded the indoor mean of about 43 |ig/m3. The authors noted that the major source of indoor
particles was cigarette smoke, and categorized their data by number of smokers in the home
(Table 7-1).
7-13
-------
1 1 1 VI W
140j-
100
1
E 80
"5)
u>
0.
40
20
^™
^
(376)....
•
X
^
(•
u
X
^
'I ^OU^/
a
(342)
)
X
-
3.
X
^
)
^
^
X
^
olid
highes
(183)
X
•
**S VILVI W W 1
( ) No. of samples
B+ — Highest site mean
•* — Composite overall mean
* — Lowest site mean
C
(274)
(355) n
8
±J
K
)
If'B'I'
P T K W SL S
P T K W SL S
Figure 7-3. The annual mean concentration of respirable particles (PM3 5) for the highest
and lowest site from the network of indoor and outdoor monitors in each city
(P-Portage, T-Topeka, K-Kingston/Harriman, W-Watertown, SL-St. Louis,
S-Steubenville) in the Harvard Six-City study. Overall composite mean and
the number of samples are also shown.
Source: Spengler et al. (1981).
TABLE 7-1. CONCENTRATIONS OF PARTICLES (PM3 5) IN HOMES OF
CHILDREN PARTICIPATING IN THE HARVARD SIX-CITY STUDY
Location
Indoors
No smokers
One smoker
Two or more smokers
Outdoors
No. of Homes
35
15
5
55
No. of Samples
1,186
494
153
1,676
Mean (SD) (//g/m3)
24.4(11.6)
36.5(14.5)
70.4 (42.9)
21.1 (11.9)
Source: Spengler et al. (1981).
7-14
-------
Dockery and Spengler (198la) provided additional data analyses drawn from the same
study but including data from 68 homes compared to the 55 reported on in Spengler et al.
(1981). Annual (every sixth day) mean indoor PM3 5 concentrations (in |ig/m3) were 20 and 23
in the two "clean" locations (Portage and Topeka); 31 and 36 in the two "medium" locations
(Watertown and Kingston-Harriman); and 39 and 47 in the two "dirty" locations (Steubenville
and St. Louis). Outdoor PM2 5 concentrations measured by dichotomous samplers every other
day ranged from 13 //g/m3 in Portage and Topeka to 20 //g/m3 in St. Louis, 24 //g/m3 in
Kingston-Harriman, and 36 //g/m3 in Steubenville (Spengler and Thurston, 1983). A mass
balance model allowed estimation of the impact of cigarette smoking on indoor particles. Long-
term mean infiltration of outdoor PM3 5 was estimated to be 70% for homes without air
conditioners, but only 30% for homes with air conditioners. A contribution of 0.88 |ig/m3 per
cigarette (24-h average) was estimated for homes without air conditioning; for homes with air
conditioning, it increased to 1.23 |ig/m3 per cigarette. A residual amount of 15 |ig/m3 not
explained by the model was attributed to indoor sources such as cooking, vacuuming and
dusting.
From the one to two years of indoor-outdoor data on 57 homes in the six cities, Letz et al.
(1984) developed an equation relating indoor to outdoor particle concentrations:
Cin = 0.385 Cout + 29.4 (Smoking) +13.8.
Thus, homes with smokers had a PM3 5 ETS component of 29.4 |ig/m3. The residual of
13.8 |ig/m3 was assumed to be due to other household activities.
Neas et al. (1994) presented summary results for the entire Phase 2 of the Six-City Study
(1983 to 1988). In Phase 2, for 1,237 homes containing white, never-smoking children, 7 to 11
years old at enrollment, three questionnaires were completed and two weeks of summer and
winter monitoring indoors and outdoors for PM2 5 was done, using the Harvard PM2 5 impactor.
At the start of the indoor monitoring study, 55% of the children were exposed to ETS in the
home, and 32% were exposed to two or more smokers. Household smoking status changed for
173 children, (13% of smoking households ceased to smoke, and 15% of the nonsmoking
households became smoking ones). The annual (winter and summer) household PM2 5 mean
concentration for the 580 children living in consistently smoking households was 48.5 ±1.4
(SE) |ig/m3 compared to 17.3 ±0.5 |ig/m3 for the 470 children in consistently nonsmoking
7-15
-------
households. Among the 614 exposed children for whom complete information on smoking
consumption was available, 36% were exposed to < 1/2 pack daily, 40% to 1/2 to 1 pack daily,
and 25% to >1 pack daily. The distribution of household concentrations for children in these
smoking categories is shown in Figure 7-4.
Spengler et al. (1985) reported on the Kingston-Harriman, TN data from the Six-City
Study. Of 101 participants, 28 had cigarette smoke exposure at home, and each had an indoor
and personal monitor (cutpoints of 3.5 |im). Each town had a centrally located outdoor
dichotomous sampler providing two size fractions (2.5 jim and 15 jim). Both towns had similar
outdoor PM2 5 concentrations of 18 |ig/m3, so the values were pooled for subsequent analyses.
Indoor concentrations averaged 42 ± 2.6 (SE) |ig/m3. Indoor values in homes with smoking
averaged 74 ± 6.6 |ig/m3, compared to 28 ± 1.1 |ig/m3 in homes without smoking (p < 0.0001).
No significant correlations between indoor and outdoor concentrations were observed.
Lebret et al. (1987) reported on the Watertown, MA portion of the Six-City Study where
265 homes were monitored for two one-week periods. Homes with smoking averaged 54 |ig/m3
(N = 147 and 152 during weeks 1 and 2), while homes without smoking averaged 21.6 |ig/m3 (N
= 70 and 74). The effect of smoking one cigarette/day was estimated at 0.8 |ig/m3 of PM25.
Spengler et al. (1987) reported on a new round of measurements in three Six-City Study
communities: Watertown, MA; St. Louis, MO; and Kingston-Harriman, TN. In each
community, about 300 children were selected to take part in a year-long diary and indoor air
quality study. PM2 5 measurements were taken indoors at home for two consecutive weeks in
winter and in summer, using the automated Harvard sampler which collected an integrated
sample for the week except for 8 a.m. to 4 p.m. weekday periods when the child was at school.
During this 40-h period, samples were taken in one classroom in each of the elementary schools
involved. Results were presented for smoking and non-smoking homes in each city by season
(Figure 7-5); the authors noted that mean concentrations in homes with smokers were about 30
|ig/m3 greater than homes without smokers, the difference being greater in winter than in
summer for all cities.
Santanam et al. (1990) reported on a more recent and larger-scale monitoring effort in
Steubenville and Portage as part of the Six-City Study; 140 homes in each city, equally
7-16
-------
120 -i
0)100-
£ 80-
o
4-1
73
5
o
4-1
re
= 60-
ra
Q.
ra
'o.
(0
0)
a:
40-
20 H
90th %tile
75th %tile
50th %tile
25th %tile
10th%tile
Never Changed
and Status
Former
Consistently Smoking
Pack
1/2-1
Pack
Packs
Figure 7-4. Distribution percentiles for annual average concentrations of indoor
respirable particulate matter (PM2 5) by household smoking status and
estimated number of cigarette packs smoked in the home during Phase 2
Harvard Six-City study.
Source: Neas et al. (1994).
7-17
-------
iou -
150-
140-
130-
120-
| 100-
ra 90-
~ BO-
S' 70-
o- 60-
50-
40-
30-
20-
10-
0J
S = smoking
N = non-smokim
--- = outdoors
\
ii
ts. —
i
ii
i
J r
i
— ^T\ — E
1 h
i L
n
i- H~t:
I
|i_ _
i 90th%tile
r1] 75th %tile
• Mean
- 50th%tile
U 25th %tile
1 10th%tile
i
H
f
h
]- -
SN SN SN SN SN
I
1
S N
Winter Summer Winter Summer Winter Summer
Watertown St. Louis Kingston
Figure 7-5. PM2 5 (jUg/m3) in smoking (S) and nonsmoking (N) homes in three of the
Harvard Six-City Study sites.
Source: Spengler et al. (1987).
distributed among households with and without smoking were monitored for one week in
summer and in winter. The Harvard impactor sampler was used to collect PM2 5 samples
between 4 p.m. and 8 a.m. on weekdays and all day on weekends, corresponding to likely times
of occupancy for school-age children. Outdoor samples were collected from one site in each
city. Target elements were determined by XRF. A source apportionment using principal
components analysis (PCA) and linear regressions on the elemental data were carried out
(Table 7-2a,b). Cigarette smoking was the single largest source in smokers' homes, accounting
for 20 to 27 |ig/m3 indoor PM2 5 in Steubenville (Table 7-2a) and 10 to 25 |ig/m3 in Portage
(Table 7-2b). Wood smoke was estimated to account for about 4 |ig/m3 indoors and outdoors in
Steubenville in winter, but only for about 1 |ig/m3 indoors and outdoors in Portage. Sulfur-
related sources accounted for 8 to 9 |ig/m3 indoors and 16 |ig/m3 outdoors in Steubenville in the
summer, but were apparently not important in winter. Auto-related sources accounted for 2 to 5
|ig/m3 in the two cities. Soil sources
7-18
-------
TABLE 7-2a. RECONSTRUCTED SOURCE CONTRIBUTIONS
TO INDOORPM,, MASS FOR STEUBENVILLE, OH1
Source
Soil
Wood smoke
O.C.-I
Tobacco Smoke
Sulfur-related
Auto-related
O.C.-II
Indoor dust
Unexplained
Total
Smokers'
Homes
7.9(3.45)
9.5(4.15)
10.3 (4.47)
45.6(19.9)
NS
NS
NS
NS
26.7(11.6)
100 (43.57)
WINTER
Non- Smokers'
Homes
17.6(3.45)
21.2(4.15)
22.9 (4.47)
NA
NS
NS
NS
NS
38.3 (7.47)
100(19.54)
Outdoor
Site
9.6(1.79)
23.0(4.31)
24.8 (4.65)
NA
NS
NS
NS
NA
42.6 (7.95)
100(18.7)
Smokers'
Homes
NS
NS
NS
53.7 (26.8)
17.8(8.90)
7.3 (3.65)
8.8 (4.40)
7.4 (3.70)
5.0 (2.4)
100 (49.85)
SUMMER
Non-Smokers'
Homes
NS
NS
NS
NA
33.3 (8.23)
14.8 (3.65)
16.5 (4.07)
15.0(3.70)
20.4 (5.05)
100 (24.7)
Outdoor
Site
NS
NS
NS
NA
52.5(15.5)
5.3(1.55)
26.0 (7.67)
NA
16.2 (4.78)
100 (29.5)
'All entries in % (ug/m3)
NS = not significant.
NA = not applicable.
O.C.-I: Iron and steel, and auto-related sources.
O.C.-II: Iron and steel, and soil sources.
Source: Santanam et al. (1990).
TABLE 7-2b. RECONSTRUCTED SOURCE CONTRIBUTIONS
TO INDOOR PM,, MASS FOR PORTAGE, WI1
Source
Sulfur-related
Auto-related
Soil
Tobacco Smoke
Wood smoke
Unexplained
Total
Smokers'
Homes
13.2(4.56)
5.1 (1.78)
3.8(1.31)
71.0(24.6)
2.7 (0.94)
4.2(1.38)
100 (34.6)
WINTER
Non- Smokers'
Homes
30.7 (4.56)
12.0(1.78)
8.8(1.31)
NA
6.3 (0.94)
42.2 (6.23)
100(14.8)
Outdoor
Site
39.2 (4.04)
17.3(1.78)
13.4(1.38)
NA
13.0(1.34)
17.1 (1.80)
100(10.3)
Smokers'
Homes
23.3 (5.80)
18.1 (4.50)
7.5(1.86)
40.1 (9.99)
NA
11.0(2.75)
100 (24.9)
SUMMER
Non-Smokers'
Homes
38.1 (5.30)
29.6(4.12)
13.4(1.86)
NA
NA
18.9 (2.62)
100(13.9)
Outdoor
Site
45.8 (6.23)
35.6 (4.84)
16.5 (2.25)
NA
NA
2.10(0.28)
100(13.6)
'All entries in % C"g/m3)
NA = not applicable.
Source: Santanam et al. (1990).
7-19
-------
accounted for only about 1 to 3 |ig/m3 of indoor and outdoor PM2 5 concentrations. Nonsmoking
homes in both cities had indoor mean PM2 5 concentrations very close to the outdoor mean
concentrations. Quite large percentages of particle concentrations were due to unexplained
sources.
7.2.2.1.2 The New York State ERDA Study
Sheldon et al. (1989) studied PM2 5 and other pollutants in 433 homes in two New York
State counties. One goal of the study was to determine the effect of kerosene heaters, gas stoves,
wood stoves or fireplaces, and cigarette smoking on indoor concentrations of combustion
products. A stratified design included all 16 combinations of the four combustion sources and
required about 22,000 telephone calls to fill all cells. The sampler was a portable dual-nozzle
impactor developed at Harvard University. Two oiled impactor plates in series were used to
reduce the probability that some particles larger than 2.5 jam would reach the filter. Samples
were collected in the main living area and in one other room (containing a combustion source if
possible) using a solenoid switch to collect alternate 15-min samples over a 7-day period.
Outdoor samples were collected at a subset of 57 homes. All samples were collected during the
winter (January to April) of 1986.
PM2 5 mean concentrations indoors for all homes, with and without any combustion
sources, were approximately double those outdoors in both counties (Table 7-3). However, in
homes without combustion sources, PM2 5 concentrations were approximately equal (Leaderer et
al., 1990). Of the four combustion sources, only smoking created significantly higher indoor
PM25 concentrations in both counties (Table 7-4). Use of kerosene heaters was associated with
significantly higher concentrations in Suffolk (N = 22) but not in Onondaga (N = 13). Use of
wood stoves/fireplaces and gas stoves did not significantly elevate indoor concentrations in
either county.
Leaderer et al. (1990) extended the analysis of these data by collapsing the gas stove
category, reducing the number of categories from 16 to 8 (Table 7-5). By inspection of Table 7-
5, it is clear that smoking was the single strongest source of indoor fine particles, with geometric
means of indoor PM ranging from 28.5 to 61.4 |ig/m3, whereas the four nonsmoking categories
ranged from 14.1 to 22.0 |ig/m3.
7-20
-------
TABLE 7-3. WEIGHTED SUMMARY STATISTICS BY NEW YORK COUNTY FOR
RESPIRABLE SUSPENDED PARTICULATE (PM2,) CONCENTRATIONS
Percent Detected
Sample Size
Population Estimate
Arithmetic Mean (//g/m3)
Arithmetic Standard Error
(Mg/m3)
Geometric Mean (//g/m3)
Geometric Standard Error
Minimum (//g/m3)
Maximum (//g/m3)
Percentiles
10th
16th
25th
50th (median)
75th
84th
90th
95th
99th
Main Living
Onondaga
98.9
224
Area
Suffolk
99.6
209
Outdoors
Onondaga
100
37
Suffolk
100
20
94,654 286,580
36.7a
2.14
25.7a
1.07
0.72
172
9.93
11.2
13.5
23.9
48.4
68.0
85.2
112
136
46.4
2.77
35.9
1.06
2.18
284
13.8
16.8
18.9
33.6
62.8
76.6
89.4
112
155
16.8
1.00
15.8
1.06
6.32
28.4
12.8
15.1
20.5
21.8
4.54
18.6
1.11
12.0
106
13.6
16.7
22.3
a Significantly different between counties at 0.05 level.
Source: Sheldon et al. (1989).
Leaderer and Hammond (1991) continued analysis of the New York State data by selecting
a subset of 96 homes for which both nicotine and PM2 5 data were obtained. In the 47 homes
where nicotine was detected (detection limit = 0.1 |ig/m3), the mean concentration of RSP was
44.1 (± 25.9 SD) |ig/m3 compared to 15.2 (± 7.4) |ig/m3 in the 49 homes without detected
nicotine. Thus, homes with smoking had an increased weekly mean PM2 5 concentration of
about 29 |ig/m3. Imperfect agreement with reported smoking was observed, with nicotine being
measured in 13% of the residences that reported no smoking, while nicotine was not detected in
28% of the residences that reported smoking. A regression on
7-21
-------
TABLE 7-4. WEIGHTED ANALYSIS OF VARIANCE OF RESPIRABLE
SUSPENDED PARTICULATE (PM2 5) CONCENTRATIONS (^g/m3) IN THE MAIN
LIVING AREA OF HOMES VERSUS SOURCE CLASSIFICATION
Onondaga(R2 = 0.17)
Model
Independent variables:
Intercept
Gas stove
Kerosene heater
Tobacco smoking
Wood stove/fireplace
Suffolk (R2 = 0.21)
Model
Independent variables:
Intercept
Gas stove
Kerosene heater
Tobacco smoking
Wood stove/fireplace
F Value
20.5
1.87
1.06
81.6
2.42
36.9
0.13
12.0
114
0.71
Probability
0.00
0.17
0.30
0.00
0.12
0.00
0.72
0.00
0.00
0.40
Coefficient
20.3
5.25
5.05
45.1
7.81
26.1
-1.52
30.1
46.8
9.88
Source: Sheldon et al. (1989).
TABLE 7-5. RESPIRABLE SUSPENDED PARTICULATE (PM2 5)
CONCENTRATION Qag/m3) IN HOMES BY SOURCE CATEGORY
Suffolk _ _ Onondaga
Source
None
W
K
S
KW
SW
SK
SKW
Outdoor
N
30
15
7
61
29
23
6
19
Mean
17.3
18.1
22.0
49.3
38.0
61.4
30.3
16.9
Standard
1.7
1.6
1.6
1.8
1.8
2.0
1.4
1.3
N
45
16
4
80
4
31
4
4
36
Mean
14.1
19.1
21.2
36.5
19.7
33.9
35.3
28.5
15.8
Standard
1.7
1.7
1.0
2.4
1.5
2.2
1.5
1.6
1.5
Abbreviations: W = woodstove; K = kerosene heater; S = tobacco smoking.
Source: Leaderer et al. (1990).
7-22
-------
all (smoking and nonsmoking) homes of PM2 5 on total number of cigarettes smoked during the
week (T) gave the result:
PM25 = 17.7 + 0.322T (N = 96; R2 = 0.55).
For the subset of 47 homes with measured nicotine, the regression gave the result:
PM25 = 24.8 + 0.272T (N = 47; R2 = 0.40).
Thus each cigarette produces about a 0.3 (±0.03) |ig/m3 increase in the weekly mean PM2 5
concentration, equivalent to a 2.1 (±0.2) |ig/m3 increase in the daily concentration.
Koutrakis et al. (1992) also analyzed the New York State data, using a mass-balance model
to estimate PM2 5 and elemental source strengths for cigarettes, wood burning stoves, and
kerosene heaters. Homes with cigar or pipe smoking and fireplace use were eliminated,
resulting in 178 indoor air samples. PM25 source strength for smoking was estimated at 12.7 ±
0.8 (SE) mg/cigarette; but PM2 5 source strengths could not be estimated for wood burning or
kerosene heater usage (only seven homes in each category were available for analysis). For a
residual category of all other indoor sources, a source strength of 1.16 mg/h was calculated. For
nonsource homes (N = 49), the authors estimated that 60% (9 |ig/m3) of the total PM2 5 mass was
from outdoor sources and 40% (6 |ig/m3) from unidentified indoor sources. However, indoor
concentrations were not significantly correlated with outdoor levels. For smoking homes, they
estimated that 54% (26 |ig/m3) of the PM25 mass was from smoking, 30% (15 |ig/m3) from
outdoor sources, and 16% (8 |ig/m3) from unidentified sources. The elemental emissions profile
for cigarettes included potassium (160 |ig/cig), chlorine (69 |ig/cig), and sulfur (65 |ig/cig), as
well as smaller amounts of bromine, cadmium, vanadium, and zinc. The woodburning profile
included three elements: potassium (92 |ig/h), silicon (44 |ig/h) and calcium (38 |ig/h). The
kerosene heater profile included a major contribution from sulfur (1500 |ig/h) and fairly large
inputs of silicon (195 |ig/h) and potassium (164 |ig/h). A drawback of the mass-balance model
was an inability to separately estimate the value of the penetration coefficient P and the decay
rate k for particles and elements; Koutrakis et al. (1992) assumed a constant rate of 0.36 h"1 for &,
and then solved for P.
7-23
-------
7.2.2.1.3 The U. S. Environmental Protection Agency Particle Total Exposure Assessment
Methodology Study
EPA designed a study of exposure to particles and associated elements in the late 1980s.
Personal exposure and indoor and outdoor PM2 5 and PM10 concentrations were measured. The
personal exposure portion of the study is discussed in 7.4.1.1.1. The study was carried out under
the Total Exposure Assessment Methodology (TEAM) program, and is known as the Particle
TEAM, or PTEAM Study.
A pilot study was undertaken in nine homes in Azusa, CA in March of 1989 to test the
sampling equipment. The first five households were monitored concurrently for seven days
(March 6-13, 1989; Wiener, 1988, 1989; Wiener et al., 1990; Spengler et al., 1989); the last four
households were then monitored for four consecutive days (March 16-20, 1989). Indoor and
outdoor particle concentrations were monitored using impactors with a 10 Lpm pump (Marple et
al., 1987). Indoor monitors, capable of sampling both fine and inhalable particles
simultaneously, were placed in different rooms in each house to determine the magnitude of
room-to-room variation.
Room-to-room variation of 12-h integrated particle levels was generally less than 10%.
Therefore the several indoor values in a particular house were averaged to provide a single mean
indoor value to compare to the corresponding outdoor value. The mean (SE) 24-h indoor PM10
concentration was 58.7 (3.4) |ig/m3 compared to the outdoor mean of 62.6 (3.5) |ig/m3.
Corresponding PM25 concentrations were 36.3 (2.6) |ig/m3 indoors and 42.6 (3.0) |ig/m3
outdoors.
Regressions of indoor on outdoor concentrations (N = 26 for each size fraction and time
period) resulted in the following equations for PM10:
Cin (day) = 36 (11) + 0.44 (0.14) Cout (R2 = 0.17)
Ctn (night) = 44 (11) + 0.14 (0.19) Cout (R2 = 0.01)
andforPM25:
Cin (day) = 18 (5) + 0.47 (0.10) Cout (R2 = 0.30)
Ctn (night) = 24 (6) + 0.23 (0.15) Cout (R2 = 0.05)
7-24
-------
where the values in parentheses are the standard errors of the parameter estimates. (In most
epidemiology studies, PM exposures are related to PM concentrations at a community ambient
monitoring station, rather than to these PM concentrations measured outside indivdual homes).
The R2 values improved considerably when the regressions for individual homes were
calculated (Wallace, 1996; see also Table 7-6). For the five homes with seven days of
monitoring (14 12-h periods) all slopes were significant, and R2 values ranged from 0.34 to 0.79
for PM10 and from 0.49 to 0.85 for PM2 5. For the four homes having only four days of
monitoring, only home 8 had significant slopes and R2 values above 0.5.
TABLE 7-6. REGRESSIONS OF INDOOR ON OUTDOOR PM
10
AND PM2 5 CONCENTRATIONS
PARTICLE TOTAL EXPOSURE
ASSESSMENT METHODOLOGY PREPILOT STUDY
PM,n (MR/m3)
House
1
2
3
4
5
6
7
8
9
N
13
13
14
13
14
8
8
8
7
Intercept
23
-25
13
16
14
175
30
-2.7
48
SE
9
17
7
9
13
38
34
23
42
P
0.026
NS
NS
NS
NS
0.004
NS
NS
NS
Slope
0.27
1.14
0.64
0.52
0.67
-1.52
0.34
1.38
0.94
SE
0.12
0.23
0.1
0.14
0.16
0.78
0.62
0.5
0.87
P
0.038
0.0003
0.00002
0.004
0.001
NS
NS
0.03
NS
R2
0.34
0.7
0.79
0.54
0.59
0.39
0.05
0.56
0.19
PM, , Cug/m3)
House
1
2
3
4
5
6
7
8
9
N
14
14
14
13
14
8
8
8
8
Intercept
14
-12
7.3
6
11
65
10
-0.34
37
SE
3.4
9
4.5
5
6
26
8
13
47
P
0.001
NS
NS
NS
NS
0.046
NS
NS
NS
Slope
0.19
0.96
0.72
0.52
0.58
-0.32
0.35
0.99
0.78
SE
0.06
0.16
0.09
0.13
0.1
1.01
0.22
0.39
1.3
P
0.005
0.00007
0.00001
0.002
0.0001
NS
NS
0.045
NS
R2
0.49
0.74
0.85
0.6
0.72
0.02
0.3
0.51
0.05
Source: Data from PTEAM Prepilot Study upon which R2 values were generated as reported by
Wallace (1996).
7-25
-------
After the pilot study in Azusa, CA, the EPA sponsored a study of personal, indoor, and
outdoor concentrations of PM10, and indoor and outdoor concentrations of PM25 in Riverside,
CA (Pellizzari et al., 1992, 1993; Perritt et al., 1991; Sheldon et al., 1992; Clayton et al., 1993;
Thomas et al., 1993; Ozkaynak et al., 1993a,b, 1996). Personal exposure results of this study are
discussed in Section 7.4.1.1.2. The main goal was to estimate the frequency distribution of
exposures to PM10 for all nonsmoking Riverside residents aged 10 and above; and 178
households were selected, using probability sampling to represent about 61,000 households
throughout most of the city of Riverside. Homes were sampled between September 22 and
November 9, 1990, and each home had two 12-h samples for both size fractions. A central site
operated throughout the 48 days of the study, producing 96 12-h samples collected by side-by-
side reference samplers (dichotomous samplers and modified hi-volume samplers) along with
the low-flow (4 Lpm) impactors with nominal cutpoints at 2.5 and 10 jim designed for this
study. (Laboratory tests [Thomas et al., 1993] revealed that the actual cutpoints were 2.5 //m
and 11.0 //m, but this section shall refer to PM10 in keeping with the investigators [Clayton
et al., 1993] who reported their data as PM10). A subset of the homes was monitored for PAHs
(Sheldon et al., 1992); 125 were monitored indoors and 65 of those were monitored outdoors for
two consecutive 12-h periods.
The precision of the three types of particle samplers at the central site was excellent, with
median RSDs of about 4 to 5% (Wallace, et al., 1991a). The low-flow sampler produced
estimates about 12% greater than the dichotomous sampler, which was about 7% greater than the
modified hi-vol sampler (Wallace, et al., 1991b). Part of the difference may be due to the
different cutpoints (estimated to be 11 |im for the new sampler, 9.5 for the dichot, and 9.0 for the
modified hi-vol), and part due to particle bounce (large particles bouncing off the impactor and
being re-entrained in the flow to the filter), such that the PM2 5 and PM10 fractions in the
low-flow sampler may be contaminated with a small number of larger-size particles. However,
particle bounce was found in laboratory tests to account for less than 7% of the total mass.
The population-weighted distributions of personal (PEM), indoor (SIM), and outdoor
(SAM) particle concentrations are provided in Table 7-7. PM10 mean concentrations
(150 |ig/m3) were more than 50% higher than either indoor or outdoor levels (95 |ig/m3).
7-26
-------
TABLE 7-7. WEIGHTED DISTRIBUTIONS OF PERSONAL, INDOOR, AND
OUTDOOR3 PARTICLE CONCENTRATIONS
to
PN
SAM
Sample size
Minimum
Maximum
Mean
(Std. error)
Geometric Mean
(Std. error)
Std. deviation
Geometric std. deviation11
Percentiles
10th
25th
50th (median)
75th
90th
Std. errors of percentiles
10th
25th
50th
75th
90th
167
7.
187.
48.
(3.
37.
(2.
37
2.
14.
23.
35.
60.
102.
1.
2
4.
3.
4.
4
8
9
5)
7
5)
.6
07
9
4
5
1
2
6
1
0
9
6
SIM
173
2.8
238.3
48.2
(4.1)
35.0
(3.3)
41.2
2.25
11.5
19.3
33.5
61.5
101.0
3.4
1.4
4.5
3.3
6.7
DAYTIME
SAM
165
16.2
506.6
94.9
(5.5)
82.7
(4.1)
57.2
1.68
42.8
56.9
84.1
110.8
157.2
2.3
4.5
4.7
4.0
7.2
NIGHTTIME
PM,n
SIM
169
16.6
512.8
94.7
(5.7)
78.2
(5.0)
61.4
1.88
30.9
49.5
81.7
127.2
180.7
3.4
4.3
8.3
9.4
11.0
PEM
171
35.1
454.8
149.8
(9.2)
128.7
(8.5)
84.3
1.75
59.9
86.1
129.7
189.1
263.1
4.0
9.4
7.5
10.8
12.0
PM,
SAM
161
3.4
164.2
50.5
(3.7)
37.2
(3.1)
40.3
2.23
14.5
23.0
35.0
64.9
120.7
2.1
2.7
2.4
4.6
5.8
;
SIM
166
2.9
133.3
36.2
(2.2)
26.7
(1.9)
29.5
2.21
10.0
14.8
25.9
48.9
82.7
0.9
1.3
2.4
5.3
5.8
SAM
162
13.6
222.9
86.3
(4.4)
74.5
(4.0)
47.7
1.74
39.3
53.6
74.1
103.7
167.8
7.4
3.4
4.8
5.1
4.3
PM,n
SIM
163
14.1
180.3
62.7
(3.2)
53.1
(3.1)
37.4
1.78
25.2
33.5
51.6
84.8
116.9
1.5
2.4
3.5
4.7
5.3
PEM
168
19.1
278.3
76.8
(3.5)
67.9
(3.1)
39.7
1.64
36.6
48.1
66.2
98.8
135.0
1.5
3.1
4.3
8.2
10.1
^Statistics other than the sample size, minimum, and maximum are calculated using weighted data; they provide estimates for the target population of person-days (PEM)
or of household-days (SIM, SAM).
bln contrast to the other statistics, the gsd is a unitless quantity.
Source: Pellizzan et al. (1992).
-------
Overnight mean personal PM10 concentrations (77 |ig/m3) were similar to the indoor (63 |ig/m3)
and outdoor (86 |ig/m3) levels. The reason for the higher daytime personal exposures (PEM)
than daytime SIM or SAM is not completely understood: it may be due to persons often being
close to sources of particles (e.g., cooking, dusting, or vacuuming) or to re-entrainment of
household dust (Thatcher and Layton, 1995). It appears not to be due to skin flakes or clothing
fibers; many skin flakes were found on filters but their mass does not account for more than 10%
of the excess personal exposure (Mamane, 1992).
Mean PM2 5 daytime concentrations were similar indoors (48 |ig/m3) and outdoors
(49 |ig/m3), but indoor concentrations fell off during the sleeping period (36 |ig/m3) compared to
50 |ig/m3 outdoors. Thus the fine particle contribution to PM10 concentrations averaged about
51% during the day and 58% at night, both indoors and outdoors. The distributions of these
ratios are provided in Table 7-8.
TABLE 7-8. WEIGHTED DISTRIBUTIONS3 OF
PM2.s/PMin CONCENTRATION RATIO
Daytime
Sample Size
Mean
(Std. error)
Geometric Mean
(Std. error)
Percentiles
10th
25th
50th (median)
75th
90th
Std. errors of percentiles
10th
25th
50th
75th
90th
Outdoor
160
0.470
(0.016)
0.444
(0.017)
0.274
0.371
0.469
0.571
0.671
0.018
0.018
0.015
0.019
0.012
Indoor
167
0.492
(0.021)
0.455
(0.022)
0.250
0.347
0.498
0.607
0.735
0.030
0.046
0.020
0.024
0.028
Nighttime
Outdoor
154
0.522
(0.017)
0.497
(0.019)
0.308
0.406
0.515
0.646
0.731
0.023
0.028
0.022
0.027
0.016
Indoor
160
0.550
(0.014)
0.517
(0.016)
0.301
0.440
0.556
0.694
0.771
0.023
0.017
0.015
0.023
0.012
aStatistics other than sample size are calculated using weighted data; they provide estimates for the target
population of household-days.
Source: Pellizzan et al. (1992).
7-28
-------
Unweighted distributions are displayed in Figures 7-6 and 7-7 for 24-h average PM10 and
PM2 5 personal, indoor, and outdoor concentrations. For 24-h data, the indoor PM is less than
the outdoor PM at all percentiles. Most of the distributions were not significantly different from
log-normal distributions, as determined by a chi-square test. About 25% of the nonsmoking
population of Riverside was estimated to have 24-h personal PM10 exposures exceeding the 150
|ig/m3 24-h NAAQS for ambient air. Since participants were monitored for only one day, the
percentage of persons with exposures exceeding the outdoor 24-h standard more than once per
year would be greater than 25%.
300
270
240
210
180
150
I 120
o>
a.
0
S 90
Q.
60
30
^* Personal
-fr Indoor
••&•• Outdoor
300
270
240
210
180
150
120
90
60
25 50 75 90 95 98 99
Cumulative Frequency (%)
30
Figure 7-6. Cumulative frequency distribution of 24-h personal, indoor, and outdoor
PM10 concentrations in Riverside, CA.
Source: Adapted from PTEAM study data (Pellizzari et al., 1992).
The 48-day sequence of outdoor PM10 and PM25 concentrations is shown in Figure 7-8
(Wallace et al., 1991a). At least two extended episodes of high fine-particle concentrations
occurred, and four days of high Santa Ana winds, with correspondingly high coarse-particle
concentrations from desert sand, were observed.
7-29
-------
50 75 90 95 98 99
Cumulative Frequency (%)
20
Figure 7-7. Cumulative frequency distribution of 24-h indoor and outdoor PM2 5
concentrations in Riverside, CA.
Source: Adapted from PTEAM study data (Pellizzari et al., 1992).
200
20 40 60 80
12-Hour Periods Beginning Sept. 22,1990
100
Figure 7-8. Forty-eight day sequence of PM10 and coarse PM (PM10 - PM2 5) in Riverside,
CA, PTEAM study. Santa Ana wind conditions are noted by an asterisk.
Source: Wallace et al. (1991 a).
7-30
-------
Central-site PM2 5 and PM10 concentrations agreed well with back yard concentrations.
Pearson correlations of the log-transformed data were 0.96 and 0.92 for overnight and daytime
PM2 5 and 0.93 for overnight PM10 values (Ozkaynak et al., 1993a), but dropped to 0.64 for
daytime PM10 values. However, two homes in one Riverside area showed very high outdoor
concentrations of 380 and 500 |ig/m3 on one day, while two homes in another Riverside area and
the central-site monitor showed more typical concentrations. A local event likely produced the
higher concentrations at the former two homes. If they are removed from the data set, the
correlation improves from 0.64 to 0.90, suggesting that a single central-site monitor can
represent well PM2 5 and PM10 concentrations throughout a wider area such as a town or small
city (at least in the Riverside area) except for unusual local conditions.
Daytime indoor PM10 and PM2 5 concentrations showed low-to-moderate Pearson
correlations of 0.46 and 0.55, respectively, with outdoor concentrations (N = 158 to 173). At
night, the correlations improved somewhat to 0.65 and 0.61, respectively (N = 50 to 168).
Outdoor PM10 concentrations explained about 27% of the variance of indoor levels (Figure 7-9)
with the two outliers included.
Simple regressions of indoor on outdoor PM10 and PM25 resulted in the following
equations (standard errors in parentheses):
Indoor PM10 = 48 (9) + 0.51 (0.08) x Outdoor PM10 (day) N=159 R2 = 0.22
Indoor PM10 = 20 (5) + 0.52 (0.05) x Outdoor PM10 (night) N=151 R2 = 0.42
Indoor PM2 5 = 14 (4) + 0.70 (0.07) x Outdoor PM2 5 (day) N=162 R2 = 0.42
Indoor PM25 = 9 (3) + 0.56 (0.04) x Outdoor PM25 (night) N=153 R2 = 0.54
Simple regressions of personal PM10 on outdoor and indoor PM10 resulted in the following
equations:
Personal PM10 = 71 (9) + 0.78 (0.08) x Indoor PM10 (day) N=163 R2 = 0.40
Personal PM10 = 21 (4) + 0.90 (0.05) x Indoor PM10 (night) N=158 R2 = 0.65
Personal PM10= 100 (12)+ 0.48 (0.10) x Outdoor PM10 (day) N=158 R2 = 0.12
Personal PM10 = 31 (6) + 0.53 (0.06) x Outdoor PM10 (night) N=155 R2 = 0.38
7-31
-------
4600
o 500
m
+J
C
ffi
u
C
o
o
400
- 300
o
o
•o
CM
200
S, 100
2
o
Indoor = 0.54*Outdoor + 32
R2 = 27% (n = 309)
100 200 300 400
Average 12-h outdoor concentration
500
600
Figure 7-9. Average indoor and outdoor 12-h concentrations of PM10 during the PTEAM
study in Riverside, CA.
Source: Ozkaynak et al. (1993b).
Correlation analyses and regressions relating personal to indoor, indoor to outdoor, and
personal to outdoor concentrations of the 14 prevalent elements were carried out for the
appropriate size fractions and both 12-h monitoring periods. For most of the elements, as with
particle mass, moderate correlations were noted for personal-indoor and indoor-outdoor
concentrations but low correlations for personal-outdoor concentrations. One element was a
strong exception to this rule: sulfur. Unlike any of the other elements, sulfur was not elevated
in the PEM relative to the SIM, and, thus, personal concentrations were much more closely
related to indoor concentrations (rs = 0.91 during the day and 0.95 at night). Moreover, because
few sources of sulfur are found indoors, the indoor-outdoor correlations were high (rs varied
between 0.90 and 0.95 for both size fractions), and even the personal-outdoor correlations
showed little degradation (the Spearman correlation rs = 0.85 during the day and 0.92 at night).
Regressions of outdoor sulfur on indoor levels gave the following results for PM10 sulfur
(Mg/m3):
7-32
-------
Sin (day) = 0.26 (0.06 SE) + 0.80 (0.02) Soui N=164 R2 = 0.88
Sin (night) = 0.20 (0.06) + 0.71 (0.03) Soui N = 155 R2 = 0.84
and for fine (PM2 5) sulfur:
Sin (day) = 0.046 (0.04 SE) + 0.85 (0.02) Soui N=164 R2 = 0.92
Sin (night) = 0.061 (0.04) + 0.80 (0.02) Sout N = 154 R2 = 0.89
Stepwise regressions resulted in smoking, cooking, and either air exchange rates or house
volumes being added to outdoor concentrations as significant variables (Table 7-9). Homes with
smoking added about 27 to 32 |ig/m3 to the total PM2 5 concentrations and about 29 to 37 //g/m3
to the PM10 values. Cooking added 12 to 26 |ig/m3 to the daytime PM10 concentration and
about 13 //g/m3 to the daytime PM2 5 concentration, but was not significant during the overnight
period.
A model developed by Koutrakis et al. (1992) was solved using nonlinear least squares to
estimate penetration factors, decay rates, and source strengths for particles and elements from
both size fractions in the PTEAM study. In this model, which assumes perfect instantaneous
mixing and steady-state conditions throughout each 12-h monitoring period, the indoor
concentration of particles or elements is given by
=
d + k
where
Cin = indoor concentration (ng/m3 for elements, |ig/m3 for particles)
P = penetration coefficient
a = air exchange rate (h"1)
Cout = outdoor concentration (ng/m3 or |ig/m3)
Qis = mass flux generated by indoor sources (ng/h or |ig/h)
V = volume of room or house (m3)
k = decay rate due to diffusion or sedimentation (h"1)
From initial multivariate analyses, the most important indoor sources appeared to be
smoking and cooking. Therefore the indoor source term Qis was replaced by the following
expression:
7-33
-------
TABLE 7-9. STEPWISE REGRESSION RESULTS FOR INDOOR AIR
CONCENTRATIONS OF PM10 AND PM2 5 (^g/m3)
COEFFICIENTS (STANDARD ERRORS OF ESTIMATES)
PM,n
Variable
N
R2
Intercept
Outdoor air
Smoking3
No. cigarettes'1
Cooking0
Air exchange
House volume*1
All
310
41%
0.52
(0.05)
37
(6)
3.2
(0.7)
20
(5)
5.2
(2.0)
-0.08
(0.02)
Day
158
39%
57
(21)
0.66
(0.09)
29
(8)
3.0
(1.0)
26
(9)
-2.7
(1)
Night
147
58%
0.45
(0.05)
38
(11)
3.9
(0.9)
12
(5)
12
(5)
All
324
55%
0.64
(0.04)
28
(3.5)
2.5
(0.4)
9.4
(2.9)
Day
156
53%
21
(7.8)
0.71
(0.07)
27
(7)
2.4
(0.6)
13
(5)
-2.0
(0.6)
Night
149
71%
0.53
(0.04)
32
(10)
4.0
(0.6)
4.5
(2)
All listed coefficients significantly different from zero at p < 0.05.
aBinary variable: 1 = at least one cigarette smoked in home during monitoring period.
bThis variable was interchanged with the smoking variable in alternate regressions to avoid colinearity problems.
"Binary variable: 1 = cooking reported for at least one min in home during monitoring period.
"Volume in thousands of cubic feet.
Source: Ozkaynak et al. (1996).
Qls = (HclgSdg + TcookScook)/T + Bother (7-4)
where
T = duration of the monitoring period (h)
Ndg = number of cigarettes smoked during monitoring period
^cig = mass of elements or particles generated per cigarette smoked (ng/cig or |ig/cig)
^cook = time spent cooking (min) during monitoring period
^cook = mass of elements or particles generated per min of cooking (ng/min or jig/min)
= mass flux of elements or particles from all other indoor sources (ng/h or |ig/h)
7-34
-------
With these changes, the equation for the indoor concentration due to these indoor sources
becomes
r> _ out + ^cig^cig + 'cook^cook + Bother (^ ,,
1n d+k (a + k)V T (a + k)V
The indoor and outdoor concentrations, number of cigarettes smoked, monitoring duration,
time spent cooking, house volumes, and air exchange rates were all measured or recorded. The
penetration factor, decay rates, and source strengths for smoking, cooking, and all other indoor
sources (2other) were estimated using a nonlinear model (NLIN in SAS software). The Gauss-
Newton approximation technique was used to regress the residuals onto the partial derivatives of
the model with respect to the unknown parameters until the estimates converge. On the first run,
the penetration coefficients were allowed to "float" (no requirement was made that they be < 1).
Since nearly all coefficients came out close to 1, a second run was made bounding them from
above by 1. The NLIN program provides statistical uncertainties (upper and lower 95%
confidence intervals) for all parameter estimates. However, it should be noted that these
uncertainties assume perfect measurements and are therefore underestimates of the true
uncertainties.
Results are presented in Table 7-10 for the combined day and night samples. The
penetration factors were very close to unity for nearly all particles and elements. The calculated
average decay rate (lower and upper 95% confidence levels) for PM2 5 was 0.39 (0.22; 0.55) h"1,
and for PM10 was 0.65 (0.36; 0.93) h"1. Since PM10 contains the PM2 5 fraction, a separate
calculation was made for the coarse particles (PM10 - PM25) with a resulting decay rate of 1.01
(0.6; 1.4) h"1. Each cigarette emitted 22 (14; 30) mg of PM10 on average, about two-thirds of
which 14 (10; 17) mg is in the fine fraction. Cooking emitted 4.1 (2.6; 5.7) mg/min of inhalable
particles, of which about 40% or 1.7 (1.0; 2.3) mg/min, was in the fine fraction. All target
elements emitted by cooking were limited almost completely to the coarse fraction. Sources
other than cooking and smoking emitted about 5.6 (2.6; 8.7) mg/h of PM10, of which only about
1.1 mg/h (0.0; 2.1) (20%) was in the fine fraction.
Decay rates for elements associated with the fine fraction were generally lower than for
elements associated with the coarse fraction, as would be expected. For example, sulfur,
7-35
-------
TABLE 7-10. PENETRATION FACTORS, DECAY RATES, AND SOURCE STRENGTHS: NONLINEAR ESTIMATES
VAR
PM25'
Al
Mn
Br
Pb
Ti
Cu
Sr
Si
Ca
Fe
K
S
Zn
Cl
PM10'
Al
Mn
Br
Pb
Ti
Cu
Sr
Si
Ca
Fe
K
S
Zn
Cl
Mean
1.00
1.00
0.87
0.90
Penetration
195
0.89
0.95
0.78
0.81
Decay Rate (1/h)
u95
1.11
1.05
0.95
0.99
Mean 195
0.39 0.22
0.03 -0.03
0.23 0.07
0.28 0.15
u95
0.55
0.09
0.38
0.41
S
Mean
1.7
0.9
0.1
0.1
cook (//g/min)
195
1.0
-1.4
-0.1
0.0
u95
2.3
3.1
0.2
0.2
Mean
13.8
9.0
0.2
1.9
S smoke (//g/cig)
b 195
10.2
-2.5
-0.4
1.3
u95
17.3
20.5
0.8
2.5
Other Sources (,ug/h)
Mean
1.1
3.0
0.5
0.6
b 195
0.0
-3.7
0.2
0.3
u95
2.1
9.8
0.9
0.9
Fail to converge
1.00
0.97
0.98
1.00
1.00
1.00
1.00
0.71
0.50
1.00
1.00
1.00
1.00
1.00
1.00
0.83
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.94
0.56
0.93
0.75
0.65
0.76
0.81
0.97
0.57
0.28
0.85
0.80
0.80
0.90
0.89
0.80
0.62
0.83
0.81
0.68
0.80
0.83
0.96
0.81
0.44
Fail to conver
1.44
1.01
1.20
1.35
1.24
1.19
1.03
0.86
0.72
1.15
1.20
1.20
1.10
1.11
1.20
1.05
1.16
1.19
1.32
1.20
1.17
1.04
1.19
1.43
ge
1.63 0.38
0.07 0.01
0.54 0.04
0.61 -0.02
0.70 0.11
0.16 -0.04
0.16 0.12
0.78 0.31
0.64 0.05
0.65 0.36
0.80 0.38
0.69 0.30
0.21 0.11
0.14 0.01
0.60 0.22
0.77 0.18
0.62 0.28
0.62 0.26
0.63 0.06
0.66 0.26
0.46 0.17
0.21 0.17
0.37 0.10
2.36 0.48
2.88
0.12
1.05
1.25
1.29
0.37
0.19
1.25
1.24
0.93
1.21
1.07
0.32
0.26
0.98
1.36
0.97
0.97
1.20
1.06
0.75
0.26
0.64
4.24
0.6
0.0
6.1
11.9
4.5
0.0
1.0
0.4
5.9
4.1
69.5
0.9
0.1
0.0
4.0
0.5
0.3
149.3
118.7
46.7
17.6
6.8
1.2
45.7
0.0
0.0
-8.6
-0.6
-3.3
-4.4
-3.9
-0.5
0.1
2.6
16.6
0.1
0.0
-0.3
0.3
0.0
0.0
26.9
37.3
8.5
0.1
-0.7
-0.2
17.6
1.2
0.0
20.9
24.4
12.3
4.4
5.9
1.2
11.6
5.7
122.4
1.7
0.3
0.3
7.8
1.1
0.5
271.8
200.1
84.8
35.2
14.3
2.5
73.9
3.7
0.1
14.4
165.6
23.8
121.3
27.1
2.9
102.6
21.9
97.6
1.1
1.8
2.1
10.0
3.5
2.6
296.4
800.0
73.0
215.7
68.0
4.0
320.2
0.2
-0.1
-58.3
72.0
-16.3
85.7
2.4
-1.5
54.0
13.6
-159.0
-2.7
1.2
0.4
-8.4
0.4
1.2
-293.9
329.0
-109.8
116.9
29.3
-3.0
107.0
7.2
0.2
87.2
259.1
63.9
156.9
51.7
7.4
151.2
30.2
354.2
4.9
2.5
3.9
28.4
6.5
3.9
886.6
1271.0
255.9
314.5
106.7
11.0
533.4
3.8
0.1
57.3
34.1
23.8
8.9
4.0
7.5
20.6
5.6
154.5
1.2
0.4
0.0
10.3
3.2
0.9
237.8
107.6
51.5
43.6
22.7
7.4
148.4
1.4
0.0
12.5
3.4
1.8
-0.5
-3.7
4.2
7.2
2.6
52.0
-0.2
0.1
-0.6
2.6
1.3
0.3
16.1
-27.0
-15.5
8.6
10.4
3.4
49.4
6.3
0.2
102.0
64.8
45.7
18.3
11.7
10.9
34.0
8.7
257.0
2.6
0.6
0.6
18.1
5.1
1.5
459.6
242.3
118.5
78.5
34.9
11.4
247.4
TVIass units in mg for PM2 5 and PM10 only.
bA negative lower confidence interval implies a nonzero mean is not statistically significant.
Source: Ozkaynak et al. (1993a).
-------
which has the lowest mass median diameter of all the elements, had calculated decay rates of
0.16 (0.12; 0.19) h'1 and 0.21 (0.17; 0.26) h'1 for PM25 and PM10 fractions, respectively. The
crustal elements (Ca, Al, Mn, Fe) had decay rates ranging from 0.6 to 0.8 h"1.
Based on the mass-balance model, outdoor air was the major source of indoor particles,
providing about 3/4 of fine particles and 2/3 of thoracic particles in the average home. It was
also the major source for most of the target elements, providing 70 to 100% of the observed
indoor concentrations for 12 of the 15 elements. It should be noted that these conclusions are
applicable only to Riverside, CA. In five of the six cities studied by Harvard and in both New
York counties, outdoor air could not have provided as much as half of the indoor air particle
mass for the average home, because the observed indoor-outdoor ratios of the mean
concentrations were > 2. However, for homes without smoking or combustion sources
(Santanam et al., 1990; Leaderer et al., 1990; Table 7-5), indoor-outdoor ratios were ~ 1. In
general, homes in areas with colder winters (such as New York) would be expected to have
tighter construction than homes in warmer areas (such as Riverside) and, therefore, more
protection against outdoor air particles.
Unidentified indoor sources accounted for most of the remaining particle and elemental
mass collected on the indoor monitors. The nature of these sources is not yet completely
understood (Thatcher and Layton, 1995). They apparently do not include smoking, other
combustion sources, cooking, dusting, vacuuming, spraying, or cleaning, since all these sources
together account for less than the unidentified sources. For example, the unidentified sources
accounted for 26% of the average indoor PM10 particles, whereas smoking accounted for 4% and
cooking for 5% (Figure 7-10).
Of the identified indoor sources, the two most important were smoking and cooking
(Figures 7-11 and 7-12). Smoking was estimated to increase 12-h average indoor
concentrations of PM10 and PM25 by 3.2 and 2.5 |ig/m3 per cigarette, respectively. Homes with
smokers averaged about 30 |ig/m3 higher levels of PM10 than homes without smokers, most of
this increase being in the fine fraction. Cooking increased indoor concentrations of PM10 by
about 0.6 |ig/m3 per minute of cooking, most of the increase being in coarse particles.
Emission profiles for target elements were obtained for smoking and for cooking. Major
elements emitted by cigarettes were K, Cl, and Ca; those from cooking included Al,
7-37
-------
Cooking
4%
Other Indoor
14%
Smoking
5%
Outdoor ^-^__ ^^^
76%
N = 352 Samples from 178 homes
Outdoor
66%
Cooking
5%
Other Indoor
26%
Smoking
4%
N = 350 Samples from 178 homes
Figure 7-10. Sources of fine particles (PM25) (top) and thoracic particles (PM10) (bottom)
in all homes (Riverside, CA).
Source: Ozkaynak et al. (1993a).
7-38
-------
Cooking
Other Indoor
Outdoor
60%
Smoking
30%
N = 61 Samples from 31 homes
Outdoor
56%
Cooking
3%
Other Indoor
16%
Smoking
24%
N = 61 Samples from 31 homes
Figure 7-11. Sources of fine particles (PM25) (top) and thoracic particles (PM10) (bottom)
in homes with smokers (Riverside, CA).
Source: Ozkaynak et al. (1993a).
7-39
-------
Cooking
Outdoor \
62%
Other Indoor
8%
Smoking
5%
N = 62 Samples from 33 homes
Outdoor
56% \
Cooking
25%
Other Indoor
16%
N = 62 Samples from 33 homes
Smoking
4%
Figure 7-12. Sources of fine particles (PM25) and thoracic particles (PM10), top and
bottom panels, respectively, for homes with cooking during data collection
(Riverside, CA).
Source: Ozykaynak et al. (1993a).
7-40
-------
Fe, Ca, and Cl. Other household activities such as vacuuming and dusting appeared to make
smaller contributions to indoor particle levels. Commuting and working outside the home
resulted in lower particle exposures than for persons staying at home. As with the particle mass,
daytime personal exposures to 14 of 15 elements were consistently higher than either indoor or
outdoor concentrations. At night, levels of the elements were similar in all three types of
samples.
7.2.2.1.4 Comparison of the Three Large-Scale Studies
The three studies had somewhat different aims and therefore different study designs. The
Harvard Six-City study selected homes based on various criteria, especially a requirement that a
school-age child be in the home, but did not employ a probability-based sample. Therefore the
results strictly apply only to the homes in the sample and not to a wider population; however, the
very large number of homes suggests that the results should be broadly applicable to homes with
school-age children in the six cities. The New York State study used a probability-based
sample, but stratified on the basis of combustion sources. Hence, there are likely to be a higher
fraction of homes with kerosene heaters, wood stoves, and fireplaces in the sample than in the
general population. The PTEAM study used a fully probability-based procedure, and its results
are likely the most broadly applicable to the entire population of Riverside households.
However, the participants were limited to nonsmokers, so homes with only smokers were
excluded; as a consequence, maximum indoor concentrations were likely underestimated. Also,
the three studies used different monitors, with different cutpoints precluding exact comparisons.
However, large differences between the PM3 5 and PM2 5 cutpoints and the PMn and PM10
cutpoints are not likely (Willeke and Baron, 1993); thus, these results can be more readily
compared. In what follows, the term "fine particles" refers to the PM3 5 and PM2 5 size fractions
collected in the three studies.
Indoor-Outdoor Relationships. Outdoor concentrations of fine particles in five of the
Harvard six cities and the two New York counties were relatively low, typically in the range of
10 to 20 |ig/m3 (Table 7-11). Only Steubenville, with an annual mean of 45 |ig/m3 (but a range
among the outdoor sites of 20 to 60 |ig/m3) approached the mean outdoor level of 50 |ig/m3
observed in Riverside. It is interesting to note that average indoor concentrations
7-41
-------
TABLE 7-11. INDOOR-OUTDOOR MEAN CONCENTRATIONS
OF FINE PARTICLES IN THREE LARGE-SCALE STUDIES
Study Name
Harvard Six-City Study
Portage, WI
Topeka, KS
Kingston-Harriman, TN
Watertown, MA
St. Louis, MO
Steubenville, OH
New York State ERDA Study
Onondaga County
Suffolk County
EPA Particle TEAM Study
Riverside, CA
Homes
11
10
8
8
10
8
224
209
178
Out
10
10
18
15
18
45
17
22
50
In
20
22
44
29
42
42
37
46
43
In/Out
2.0
2.2
2.4
1.9
2.3
0.9
2.2
2.1
0.9
Harvard: PM3 5 measured using cyclone sampler. Samples collected every sixth day for one year (May 1986 to April
1987).
NYS: PM2 5 measured using impactor developed at Harvard. Samples collected for one week at each household
between January and April 1986.
PTEAM: PM25 measured using Marple-Harvard-EPA sampler. Samples collected for two 12-h periods at each
home between September and November 1990.
Source: Harvard data—Spengler et al. (1981); NYS data—Sheldon et al. (1989); PTEAM data—Pellizzan et al.
(1992).
exceeded outdoor concentrations in the seven sites with low outdoor levels, (indoor/outdoor
ratios were contained in a small range between 1.9 and 2.4), but were slightly less than outdoor
concentrations in the two sites with high outdoor levels (ratios of 0.9).
Effect of Smoking. All three studies found cigarette smoking to be a major source of
indoor fine particles. Multivariate calculations in all three studies result in rather similar
estimates of the effect of smoking on fine particle concentrations. Spengler et al. (1981)
estimated an increase of about 20 |ig/m3 per smoker based on 55 homes from all six cities. Since
the 20 homes with at least one smoker averaged at least 1.25 smokers per home, this corresponds
to about 25 |ig/m3 per smoking home. Spengler et al. (1985) found a smoking effect of about 32
|ig/m3 for smoking homes in multivariate models based on the Kingston-Harriman data.
Santanam et al. (1990) found a smoking-related increase of 20-27 //g/m3 in Steubenville and
Portage (winter only) but only 10 //g/m3 in Portage in summer. Sheldon et al. (1989) found an
increase of 45 (Onondaga) and 47 (Suffolk) |ig/m3 per smoking home in a multivariate model of
the New York State data. Ozkaynak et al. (1993b) found an increase of about 27 to 32 |ig/m3 in
7-42
-------
homes with smokers in a multivariate regression model of the PTEAM PM2 5 data. Thus, the
effect of a home with smokers on indoor fine particle concentrations was estimated to be about
20 to 30 Mg/m3 in the Six-City and PTEAM studies, but about 45 //g/m3 in the New York State
study.
Dockery and Spengler (1981a) found an effect of 0.88 |ig/m3 per cigarette for homes
without air conditioning, and 1.23 |ig/m3 per cigarette for homes with air conditioning, based on
68 homes from all six cities. Lebret et al. (1987) found an effect of 0.8 |ig/m3 per cigarette for
homes in the Watertown, MA, area. Leaderer and Hammond (1991) found an effect ranging
between 1.9 and 2.3 |ig/m3 per cigarette contribution to a 24-h average. In a series of stepwise
regressions on the PTEAM data, Ozkaynak et al. (1993b) found an effect ranging between 1.2
and 2.4 |ig/m3 per cigarette smoked during a 24-h period. Taking the midpoint of these ranges
leads to estimates for the Harvard Six-City, New York State and PTEAM studies of about 1.1,
2.1, and 1.8 |ig/m3 increases in fine particle concentrations per cigarette smoked in the home
over a 24-h period.
Both the New York State study and the PTEAM study were able to estimate source
strengths for different variables using a mass-balance model. The estimates for PM2 5 emissions
from cigarettes were very comparable, with Koutrakis et al. (1992) estimating 12.7 mg/cig
compared to the PTEAM estimate of 13.8 mg/cig (Ozkaynak et al., 1993a). Both studies also
found similar elemental profiles for smoking, with potassium and chlorine being emitted in
substantial amounts.
Effect of Other Variables. In the PTEAM Study, the second most powerful indoor source
of PM10, and possibly PM25 particles, was cooking. Quite large emission strengths of several
mg/minute of cooking were determined from the mass-balance model, while multiple
regressions indicated that cooking could contribute between 10 and 20 |ig/m3 PM10, and
somewhat smaller amounts of PM25, to the 12-h concentration.
Both the New York State and PTEAM studies also measured air exchange in every home,
and both studies found that air exchange significantly affected indoor particle concentrations. In
the PTEAM study, increased air exchange led to increased indoor air concentrations for both
PM2 5 and PM10 at night only, perhaps because outdoor concentrations were larger than indoor
levels at night. In the New York State study, increased air exchange led to decreased RSP
concentrations in Onondaga (p < 0.02) but no effect was noted in Suffolk (p < 0.90). In both of
7-43
-------
these counties, indoor levels generally exceeded outdoor levels, so increased air exchange would
generally reduce indoor concentrations.
7.2.2.2 Other Studies of PM Indoors
Several other large-scale studies of indoor PM in homes have taken place in other
countries, and a number of smaller U.S. studies have been conducted. These are discussed
below in order of the number of homes included in the study.
Lebret et al. (1990) carried out week-long RSP measurements (cutpoint not described) in
260 homes in Ede and Rotterdam, The Netherlands, during the winters of 1981 to 1982 and 1982
to 1983, respectively; 60% of the Ede homes and 66% of the Rotterdam homes included
smokers. Diary information collected during the measurement period indicated that, on average,
one to two cigarettes were smoked during the week, presumably by guests, even in the
nonsmoking homes. Homes with one smoker averaged seven cigarettes smoked per day at home
in Ede (N = 53) and 11 per day in Rotterdam (N = 35). Homes with two smokers averaged 21
cigarettes per day in Ede (N = 23) and 25 per day in Rotterdam (N = 15).
Geometric means for the combined smoking and nonsmoking homes were similar in the
two cities (61 and 56 |ig/m3, respectively), with maxima of 560 and 362 |ig/m3. Outdoor
concentrations averaged about 45 |ig/m3 (N not given). Indoor concentrations in the homes with
smokers averaged about 70 |ig/m3 (calculated from data in the paper), compared to levels in the
nonsmoking homes of about 30 |ig/m3. Multiple regression analysis indicated that the number of
smoking occupants explained about 40% of the variation in the log-transformed RSP
concentrations—family size, frequency of vacuuming, volume of the living room, type of space
heating, and city (Ede versus Rotterdam) had no significant effect on RSP concentrations. In a
second regression, the number of smoking occupants was replaced by the number of cigarettes
and cigars smoked during the week. The regression equation was
log(RSP) = 1.4 + 0.37 log(# cigarettes) + 0.53 log(# cigars)
+ 0.03 log(family size)
R2 = 0.49; d.f. = 250 F = 83.7 p < 0.0001
From this equation, the authors estimated that one cigarette smoked per day would increase
weekly average indoor RSP concentrations by 2 to 5 |ig/m3, whereas one cigar smoked per day
7-44
-------
would increase indoor levels by 10 |ig/m3. Instantaneous RSP concentrations were measured
using a TSI Piezobalance on the day the technicians were setting up the equipment. Table 7-12
shows the influence of smoking on these measurements.
TABLE 7-12. INFLUENCE OF RECENT CIGARETTE SMOKING
ON INDOOR CONCENTRATIONS OF PARTICIPATE MATTER1
Time Since Smoking
No smoking
More than 1 h ago
Between 1/2 and 1 h ago
Less than 1/2 an hour ago
During the measurements
N
98
18
7
27
54
RSP (geom. mean)
41
52
76
141
191
(Mg/m3)
'Size cuts for measured particles not specified.
Source: Lebret et al. (1990).
Heavner et al. (1995) studied PM3 5 at home and at work for 104 New Jersey and
Pennsylvania females. The personal sampler used consisted of a cyclone sampling head attached
to a 37-mm Fluoropore filter, connected by Tygon tubing to a 1.7 Lpm pump. The sampling
head was worn on a lapel, collar, or pocket in the breathing zone of the participant until she went
to bed, when the sampler was placed on the bedside table. The "home" pumps were turned on at
6 p.m. and sampled until about 8 a.m. the next morning (an average of 14 h); the "work" pumps
were turned on at work and sampled for an average of 7 h. Participants were selected to include
those with exposure to smoking at home or at work or both or neither. The 14-h evening and
overnight concentrations in the homes averaged 86.7 ± 145.4 (SD) //g/m3 for 30 homes with
smokers and 27.6 ± 19.9 //g/m3 for 58 homes without smokers. Corresponding values for
workplaces were 67.0 ± 44.3 //g/m3 for those 28 allowing smoking and 30.3 ± 17.6 //g/m3 for
the 52 without smoking, the differences being significant at p < 0.0001 (Wilcoxon rank sum) for
both comparisons.
Diemel et al. (1981) measured particles in 101 residences in an epidemiological study
related to a lead smelter in Arnhem, the Netherlands. The indoor sampler collected samples at a
flowrate of 1 to 1.5 Lpm. The authors stated that particles < 3 to 4 jim diameter should have
7-45
-------
been sampled efficiently, but presented no data on measured outpoint size. The outdoor
samplers (number not given) were high-volume samplers. The 28-day average levels indoors
ranged from 20 to 570 |ig/m3, with an arithmetic mean of 140 |ig/m3 (SD not presented) and a
geometric mean of 120 |ig/m3; corresponding outdoor concentrations (2-mo averages of 24-h
daily samples) ranged from 53.7 to 73.3 |ig/m3 (N not given), with nearly identical arithmetic
and geometric means of 64 |ig/m3.
Kulmala et al. (1987) measured indoor and outdoor air in approximately 100 dwellings
(including some office buildings) in Helsinki, Finland between 1983 and 1986. Samples were
collected on Nuclepore filters using a stacked foil technique. The geometric mean for the
combined fine particle (<1 //m) samples indoors was 16 |ig/m3, with a 95% range of 4 to
67 |ig/m3. The corresponding value for the indoor coarser particles (>1 //m) was 13 |ig/m3 with a
range of 3 to 63 |ig/m3. Outdoors, the fine particles had a geometric mean of 20 |ig/m3 with a
95% range of 5 to 82 |ig/m3, and the coarser particles had a geometric mean of 16 |ig/m3 with a
range of 3 to 91 |ig/m3.
Quackenboss et al. (1989) reported PM10 and PM25 results from 98 homes in the Tucson,
AZ area selected as part of a nested design for an epidemiological study. The Harvard-designed
dual-nozzle indoor air sampler (Marple et al., 1987) was used for indoor air measurements.
Outdoor air was measured within each geographic cluster by the same instrument;
supplementary data were obtained from the Pima County Air Quality Control District, but these
data did not include PM2 5 measurements and some data were apparently PM15. Homes were
classified by (a) tobacco smoking and (b) use of evaporative ("swamp") coolers, which
apparently act as a removal mechanism for particles (Table 7-13). Homes without smoking
averaged about 15 |ig/m3 PM2 5, compared to 27 |ig/m3 for homes reporting one or less pack a
day, and 61 |ig/m3 for homes reporting more than one pack a day. PM2 5 particles accounted for
about half of the PM10 fraction in nonsmoking homes, increasing with the amount of smoking to
about 80% in those homes with heavy smoking. Outdoor PM10 particles were not strongly
correlated with indoor levels (R2 = 0.18; N ~ 90).
7-46
-------
TABLE 7-13. INDOOR AVERAGE PM25 AND PM10
BY REPORTED SMOKING IN THE HOME AND EVAPORATIVE
COOLER USE DURING SAMPLING WEEK FOR TUCSON, AZ STUDY
Smoking
Cigarettes/Day
None
1-20
>20
Evaporative Cooler
Yes
No
Total
Yes
No
Total
Yes
No
Total
Mean
8.8
20.3
15.2
19.3
32.3
27.3
36.2
82.7
60.8
PM25
S.D.
5.0
19.0
15.5
8.8
28.5
23.6
32.9
55.4
50.8
Homes
(20)
(25)
(45)
(10)
(16)
(26)
(8)
(9)
(17)
Mean
21.0
38.4
30.3
33.9
53.4
46.2
47.4
102.5
75.0
PM10
S.D.
9.7
22.9
19.9
12.0
33.9
29.1
39.6
60.6
57.2
Homes
(20)
(23)
(43)
(10)
(17)
(27)
(9)
(9)
(18)
PM25: Significant (p < 0.01) main effects for smoking and evaporative cooler use; two-way interaction nearly
significant (p = 0.06).
PM10: Significant (p < 0.01) main effects for evaporative cooler and smoking.
Source: Quackenboss et al. (1989).
Quackenboss et al. (1991) extended the analysis of the Tucson homes over three seasons.
Median indoor PM2 5 levels in homes with smokers were about 20 |ig/m3 in the summer and
spring/fall seasons compared to about 10 |ig/m3 in homes without smokers in those seasons. In
winter, however, the difference was considerably increased, with the median level in 24 homes
with smokers at about 36 |ig/m3 compared to 13 |ig/m3 in 26 homes without smokers.
Sexton et al. (1984) reported on a study in Waterbury, VT. This study included 24 homes,
19 with wood-burning appliances, and none with smokers. 24-h samples were collected in each
home every other day for two weeks, providing 163 valid indoor samples. Indoor RSP levels
ranged from 6 to 69 |ig/m3 with a mean value of 25 |ig/m3. Outdoor levels ranged from 6 to 30
|ig/m3 with a mean value of 19 |ig/m3. Indoor concentrations were not significantly correlated
with outdoor concentrations (r = 0.11, p > 0.16.)
Kim and Stock (1986) reported results for 11 homes in the Houston, TX area. (Year and
the season not reported in the paper.) For most homes, two 12-h PM2 5 samples (day and night)
were collected for approximately one week. Sampling methods were not fully discussed, but
apparently they involved samples collected using a mobile van near each home. The mean
weekly concentrations in the five smoking homes averaged 33.0 ± 4.7 (SD) |ig/m3, versus mean
outdoor concentrations averaging 24.7 ± 7.4 |ig/m3 (calculated from data presented in paper).
7-47
-------
Indoor concentrations in the six nonsmoking homes averaged 10.8 ± 4.9 |ig/m3 compared to
outdoor levels of 12.0 ± 5.9 |ig/m3.
Morandi et al. (1986) reported on 13 Houston, TX, homes monitored during 1981 as part
of a larger personal monitoring study of 30 nonsmoking participants. The TSI Piezobalance
(cutpoint at about PM3 5) was employed for personal monitoring, with technicians "shadowing"
the participants and taking consecutive 5-min readings. At the homes, dichotomous samplers
(cutpoints at PM2 5 and PM10) were used for two 12-h daytime samples (7 a.m. to 7 p.m.) both
inside and outside the homes for seven consecutive days. Little difference was noted in the
indoor concentrations at homes (25 ± 30 (SD) |ig/m3) and at work or school (29 ± 25 |ig/m3).
The highest overall respirable suspended particle (RSP) concentrations occurred in the presence
of active smoking (89 |ig/m3), significantly different from mean RSP values measured in the
absence of smokers (19 |ig/m3; p < 0.0001). Among homes with smokers, those homes with
central air conditioning were significantly (p<0.0001) higher (114 versus 52 |ig/m3) than those
with no air conditioning. Cooking was associated with significantly higher RSP concentrations
(27 |ig/m3 compared to 20 |ig/m3, p < 0.01). The single highest RSP concentration (202 |ig/m3)
was found in a home with no smokers and no air conditioning but with active cooking. The
authors concluded that cooking was a more important source of indoor RSP than smoking, at
least in the few homes they studied.
Coultas et al. (1990) measured PM25 in 10 homes containing at least one smoker, using the
Harvard aerosol impactor. Samples were collected for 24 h every other day for 10 days and then
for 24 h every other week for 10 weeks, resulting in 10 samples per household. The mean
concentrations of PM25 ranged from 32.4 ± 13.1 (SD) to 76.9 ± 32.9 |ig/m3. Outdoor particle
concentrations were not reported; thus it is difficult to calculate the portion of the observed PM2 5
that might be due to ETS.
Kamens et al. (1991) measured indoor particles in three homes without smokers in North
Carolina in November and December 1987 (no measurements of outdoor particles were taken).
Two dichotomous samplers (PM25 and PM10), several prototype personal samplers (also PM25
and PM10), three particle sizing instruments including a TSI electrical aerosol mobility analyzer
(EAA) with 10 size intervals between 0.01 and 1.0 jam, and two optical scattering devices
covering the range of 0.09 to 3.0 and 2.6 to 19.4 |im were employed. Air exchange
measurements were made using SF6 decay over the course of the seven 8-h (daytime) sampling
7-48
-------
periods. There were also three 13-h (evening and overnight) sampling periods. For the entire
study, 37% of the estimated total mass collected was in the fine fraction, and another 37% was >
10 |im. The remainder (26%) was in the inhalable coarse (PM10 - PM2 5) fraction. However,
considerable variation was noted in these size distributions. For example, on one day with
extensive vacuuming, cooking, and vigorous exercising of household pets, 52% of the total mass
appeared in the fraction >10 |im, but only 18% in the fine fraction. The peak in particle mass on
that day coincided with vacuuming and sweeping of carpets and floors. On another day, cooking
of stir-fried vegetables and rice produced a large number of small particles, with those <0.1 jim
accounting for 30% of the total EAA particle volume, much more than the normal amount. The
cooking contribution of that one meal to total 8-h daytime particle volume exposure was
calculated to be in the range of 5 to 18%. The authors concluded that the most significant indoor
source of small particles (<2.5 jim) in all three of these nonsmoking homes was cooking, while
the most significant source of large particles (>10 jim) was vacuum sweeping. Inhalable coarse
particles (PM10 - PM2 5) appeared to be of largely biological (human dander and insect parts) and
mineral (clay, salt, chalk, etc.) origin.
In a test of a new sampling device (a portable nephelometer), Anuszewski et al. (1992)
reported results from indoor and outdoor sampling at nine Seattle, WA, homes sampled for an
average of 18 days each during the winter of 1991 to 1992. The nephelometer is a light-
scattering device with rapid (1-min) response to various household activities such as sweeping,
cigarette smoking, frying, barbecuing, and operating a fireplace. Homes with fewer activities
showed high correlations of indoor and outdoor light-scattering coefficients, both between
hourly averages and 12-h averages. However, homes with electrostatic precipitators, with
weather-stripped windows or doors, and with gas cooking or heating devices showed weak 12-h
indoor-outdoor correlations.
Chan et al. (1995) studied particles and nicotine in seven homes with one smoker each in
Taiwan. Sampling was carried out in summer and winter of 1991. Each home had one indoor
PM5 sampler in the living room and another in the yard. In the winter study, two homes had
PM10 samplers added inside and outside and at two central sites. Indoor mean PM5
concentrations averaged 44 ± 32 (SD) |ig/m3 in summer compared to outdoor levels of
27 ± 15 |ig/m3. Corresponding winter values were 107 ± 44 |ig/m3 and 92 ± 40 |ig/m3.
7-49
-------
Daisey et al. (1987) measured RSP, PAH, and extractable organic matter (EOM) in seven
Wisconsin homes with wood stoves; one 48h (1,000 m3) sample was collected during
woodburning and a second sample was collected when no woodburning occurred. Five of seven
homes had somewhat higher RSP levels during woodburning, but the mean difference was not
significant.
Highsmith et al. (1991) reported on two 20-home studies in Boise, ID and Roanoke, VA.
The Boise study assessed the effects of wood burning on ambient and indoor concentrations in
the area. Ten homes with wood burning stoves were matched with 10 homes without such
stoves. One matched pair of homes was monitored from Saturday through Tuesday for eight
consecutive 12-h periods. Ambient PM2 5 levels increased by about 50% at night, suggesting an
influence of woodburning. Indoor PM2 5 concentrations also were increased (by about 45%) in
the homes with the wood burning stoves compared to those without (26.3 versus 18.2 //g/m3),
although coarse particles showed no increase (10.2 versus 9.7 //g/m3). The Roanoke study,
designed to assess the effects of residential oil heating, showed no effects on indoor levels of
fine or coarse particles.
Lofroth et al. (1991) measured particle emissions from cigarettes, incense sticks, "mosquito
coils," and frying of various foods. Emissions were 27 and 37 mg/g for two brands of Swedish
cigarettes, 51 and 52 mg/g for incense sticks and cones, and 61 mg/g for the mosquito coil.
Emissions from frying pork, hamburgers, herring, pudding, and Swedish pancakes ranged from
0.07 to 3.5 mg/g.
Mumford et al. (1991) measured PM10, PAH, and mutagenicity in eight mobile homes with
kerosene heaters. Each home was monitored for 2.6 to 9.5 h/day (mean of 6.5 h) for three days
a week for two weeks with the kerosene heaters off and for two weeks with them on (average
on-time of 4.5 h). Mean PM10 levels were not significantly increased when the heaters were on
(73.7 ± 7.3 (SE) |ig/m3 versus 56.1 ± 5.7 |ig/m3), but in two homes levels increased to 112 and
113 |ig/m3 when the heaters were on. Outdoor concentrations averaged 18.0 ± 2.1 |ig/m3.
Colome et al. (1990) measured particles using PM10 and PM5 (cyclone) samplers inside and
outside homes of 10 nonsmokers, including eight asthmatics, living in Orange County, CA.
Indoor PM10 samples were well below outdoor levels for all homes (mean of 42.5 ±3.7 (SE)
|ig/m3 indoors versus 60.8 ± 4.7 |ig/m3 outdoors). No pets, wood stoves, fireplaces, or kerosene
heaters were present in any of these homes.
7-50
-------
Lioy et al. (1990) measured PM10 at eight homes (no smokers) for 14 days in the winter of
1988 in Phillipsburg, NJ, which has a major point source consisting of a grey-iron pipe
manufacturing company. The Harvard impactor was used indoors to collect 14 24-h samples
beginning at 4:30 p.m. each day; Wedding hi-vol PM10 samples were deployed at three outdoor
sites. A fourth outdoor site was located on a porch of one of the homes, directly across the street
from the pipe manufacturer. The first three sites showed little difference from one another,
whereas on day 4 and day 6 of the study, the outdoor sampler on the porch had readings that
were considerably (about 40 |ig/m3) higher than the other outdoor samplers, suggesting an
influence of the nearby point source. The geometric mean outdoor PM10 concentration was 48
|ig/m3 (GSD not provided) compared to 42 |ig/m3 indoors. A simple regression equation for all
homes (N = 101 samples) explained 45% of the cross-sectional variance in indoor PM10:
Indoor PM10 = 0.496 Outdoor PM10 + 21.5 (R2 = 0.45)
However, individual regressions by home showed much better R2 values in most cases, ranging
from 0.36 to 0.96 (Table 7-14). All slopes were significant.
Thatcher and Layton (1995) measured optical particle size distributions inside and outside
a residence in the summer. Measured deposition velocities for particles between 1 and 5 jim
closely matched the calculated gravitational settling velocities; however, for particles >5 jim,
the deposition velocity was less than the calculated settling velocity, perhaps due to the non-
spherical nature of these particles. The deposition velocities determined by the authors
corresponded to a particle deposition rate k of 0.46 h"1 for particles of size range 1 to 5 jam and
1.36 h"1 for particles of size range 5 to 10 jim. These values are very comparable with the values
of 0.39 h"1 for particles less than 2.5 jim and 1.01 h"1 for particles between 2.5 and 10 jim found
by the PTEAM Study. The authors measured the penetration factor/1 by the following method:
They first carried out vigorous house cleaning activities to raise the level of resuspended dust
well above outdoor levels.
7-51
-------
TABLE 7-14. REGRESSION OF INDOOR ON OUTDOOR PM10
CONCENTRATIONS: THEES STUDY, PHILLIPSBURG, NJ
House
1
2
O
4
5
6
7
8
N
14
14
9
14
13
13
12
14
Intercept
19
16
9
20
6
-1
24
27
SE
9
14
5
21
10
18
25
8
P
NS
NS
NS
NS
NS
NS
NS
S
Slope
0.44
0.40
0.55
0.73
0.43
0.89
0.70
0.54
SE
0.06
0.08
0.04
0.15
0.07
0.13
0.29
0.05
P
S
S
S
S
S
S
S
S
R2
0.79
0.68
0.96
0.66
0.75
0.81
0.36
0.91
S = Significant
NS = Non-significant
Source: Data from THEES Study (Lioy et al, 1990).
They then left the house, while automated instruments measured the deposition rate k for the
different particle sizes and the air exchange rate a for SF6 tracer gas. With these values of a and
k in hand, they solved the equation for/1, using the steady-state values for Cin and C^ observed
long after the dust had settled:
(7-6)
For all size ranges tested, including the largest (10 to 25 jim), the experimentally determined
value for P was not significantly different from 1 (Figure 7-13). This result is in agreement with
the PTEAM conclusion that P is 1 for both fine and coarse particles, although the latter
conclusion was derived from a nonlinear (statistical) approach whereas the present result was
experimentally obtained.
The resuspension results of Thatcher and Layton (1995) (Figure 7-14) show the effect of a
vigorous housecleaning activity. The authors concluded "Although particles larger than 5 jim
show significant resuspension in these experiments, particles smaller than 5 |im are not readily
resuspended, and particles less than 1 jim show almost no resuspension even with vigorous
activity." Figure 7-15 shows that just one person walking in and out of a carpeted
7-52
-------
2
a.
1
5
§7
S7
£
25/jm
10
20 30 40
Time (minutes)
50
60
Figure 7-14. The change in suspended particle mass concentration versus time, as
measured by optical particle counter, assuming spherical particles of unit
density. All resuspension activities are stopped at t = 0.
Source: Thatcher and Layton (1995).
7-53
-------
Cleaning
I I
2 Mm
Walk/Sit
4 People
5 minutes
4 People
30 minutes
Walk In
0.5 to 1 um
1 to 5 um 5 to 10 um
Particle Diameter (um)
10 to 25 um
Figure 7-15. The ratio of the suspended particle concentration after a resuspension
activity to the indoor concentration before that activity, by particle size. The
activities tested are (1) vigorous vacuuming and housecleaning, (2) 2 min of
continuous walking and sitting in the living area by one person, (3) 5 min of
normal activity by four people, (4) 30 min of normal activity and (5) one
person walking into and out of the living area.
Source: Thatcher and Layton (1995).
living area can increase indoor particle concentrations in the ranges 5 to 10 jim by 100% and 10
to 25 |im by 200%. The absolute increase in indoor concentrations by this activity is a function
of the surface dust loading in those size ranges. Surface dust loadings (//g/m2) increase with the
time since last cleaning (Raunemaa et al., 1989; Wilmoth et al., 1991).
Because fluffy house dust can be resuspended, it will contribute to total airborne exposure
to particles and constituents such as metals and pesticides. Roberts et al. (1990) studied
42 homes in Washington State. Geometric mean lead concentration in 6 homes where shoes
were removed on entry was 240 |ig/m2 on carpets, compared to 2,900 |ig/m2 on carpets in 36
homes where shoes were kept on. In Japan, where shoes are removed on entry and straw mats
(tatami) are usually used instead of carpets, Tamura et al. (1996) found evidence of negligible
7-54
-------
PM10 resuspension. These findings suggest that most of the carpet dust in a home enters via
track-in on shoes rather than by infiltration of ambient air.
7.2.2.3 Personal Exposures to Environmental Tobacco Smoke.
Jenkins et al. (1995a) reported on the first 12 cities of a 16-city sampling survey comparing
ETS exposures at home and at work. About 100 nonsmoking persons in each city were recruited
to wear a personal monitor at work and another personal monitor away from work. The
monitors collected PM3 5 particles, which were then analyzed for tobacco smoke markers
(UVPM, FPM and solanesol). Nicotine and other gas-phase markers were also collected.
Subjects provided saliva samples, which were used to screen out smokers reporting themselves
as nonsmokers. (Using different cutoff points of 10, 30, or 100 |ig/L, between 1.82 and 5.2% of
the 1073 subjects would have been misclassified as nonsmokers). Four cells were defined:
persons exposed to smoke at home and at work (N = 119); persons exposed at home but not at
work (110); persons exposed at work but not at home (163); and persons exposed neither at
home nor at work (504). All four particle markers agreed in ranking the four cells for total ETS
exposure in the order listed—that is, nonwork (including home) ETS exposures were greater
than work exposures as shown in Table 7-15. The authors identified several problems with the
selection of the sample. First, the sample was 68% female. Secondly, the socioeconomic level
was biased high, with about twice as many persons having some college or being college
graduates as the population as a whole. It is well known that smoking rates decrease as
education and income rise, and this study confirmed that observation—when broken out by
income, ETS markers decreased by factors of 2 to 5 as annual income rose from $10,000 to
$100,000. The authors compared ETS levels in offices with no smoking (N = 629), restricted
smoking (N = 297) and unrestricted smoking (N = 113). Median (mean) levels of RSP increased
from 13 (18) to 16 (28) to 33 (58) |ig/m3 in the three categories, with corresponding nicotine
medians (means) of 0.025 (0.11), 0.09 (0.87), and 0.44 (2.7) |ig/m3
Jenkins et al. (1995b) updated the results to the full 16 cities. The final number of
participants in the four cells were 157, 234, 281, and 808, respectively. The median RSP
(PM3 5) values changed only slightly, increasing to 33.6 from 32 //g/m3 in Cell 1 and decreasing
to 23.3 //g/m3 in Cell 2, with no changes in the remaining two cells.
7-55
-------
TABLE 7-15. MEDIAN VALUES (jig/m3) FOR
ENVIRONMENTAL TOBACCO SMOKE MARKERS
Cell (N)
1(119)
2(110)
3 (163)
4 (504)
Nonwork
S
S
NS
NS
Work
S
NS
S
NS
RSP
32
24
20
15
UVPM
12
7.6
2.3
1.1
FPM
7.7
5.9
1.2
0.6
Solanesol
0.113
0.058
0.003
ND
Nicotine
1.46
0.56
0.11
0.02
S = smoker; NS = nonsmoker; ND = not detectable.
Source: Jenkins et al. (1995a).
ETS Exposures in Restaurants and Buildings. Oldaker et al. (1993) reported results of
analyzing ETS markers in four office buildings. Median RSP levels were 30 and 34 |ig/m3 in
two buildings allowing smoking, compared to 5 and 7 |ig/m3 in two buildings without smoking.
Crouse et al. (1989) reported on measurements of RSP (PM3 5) in 42 North Carolina restaurants.
Geometric mean (arithmetic mean) values were 5.3 (8.6), 26.1 (34.1) and 62.0 (80.8) |ig/m3,
respectively. Oldaker et al. (1990) measured PM3 5 in 33 restaurants in the Winston-Salem, NC,
area during the summer of 1986 and the winter of 1988 to 1989; in the winter, the cutpoint was
changed to PM25. A wide range of particle concentrations was noted, from 18 to 1,374 |ig/m3 in
the summer, and <25 to 281 |ig/m3 in winter.
7.2.2.4 The Fraction of Outdoor Air Particles Penetrating Indoors
Having reviewed the literature on particles in homes, it is useful to return to one of the
questions we asked at the outset: For a home with no indoor sources or resuspension of settled
dust of ambient origin, how much protection is offered against outdoor particles of various size
ranges?
The governing equation in this case is
Cm Pa
(7-6)
7-56
-------
Thus, there are three parameters affecting the fraction of outdoor air particles to be found
indoors: the penetration factor/1, the air exchange rate a, and the particle deposition rate k.
Penetration Factor P. The penetration factor Pisa measure of the ability of a gas or
particle to penetrate the building envelope; 0 < P < 1. For a nonreactive gas, such as CO, the
factor is expected to be 1. For large particles, the factor would be expected to go to zero with
increasing particle size and decreasing air exchange rate. The question is at what combinations
of size range and air exchange rate does the factor P begin to decrease significantly from unity
for PM?
Two recent studies have attempted to determine the value of P for different particle size
ranges. The PTEAM study (Ozkaynak et al., 1996) found a value of P ~ 1 for both PM2 5 and
PM10 particles. The value was determined statistically by a nonlinear solution of Equation 7-5
(including all indoor sources) for 178 homes. Thatcher and Layton (1995) also found a value of
P ~ 1 for all size ranges tested, including the ranges 1 to 3 jim, 3 to 6 jim, 1 to 5 jim, 5 to 10 jim,
and 10 to 25 jim. The authors determined their values experimentally by direct measurement on
one instrumented house. The results for the first two size ranges were obtained in five replicate
experiments; for the last three size ranges, in only one experiment (Figure 7-13). Thus the two
studies used different methods but arrived at the same conclusion: particles less than 10 pm in
aerodynamic diameter penetrate building envelopes with an efficiency approaching that of
nonreactive gases. Clearly, more work needs to be done to test this finding at lower air
exchange rates.
Air Exchange Rate a. Air exchange rates in residences depend on three major factors:
building construction, ambient conditions, and resident activities.
The building construction determines the lower bound of the air exchange rate. That is,
rates cannot be reduced below the rate allowed by diffusion through the building cracks, holes,
and other uncontrolled means of particle ingress in the absence of wind and buoyancy
differences. Tests by building pressurization (e.g., using "blower doors") are able to determine a
parameter ("crack length") that quantifies this lower bound. Buildings that are extremely tightly
constructed for energy efficiency are able to reduce the lower bound of the air exchange rate to
the order of 0.1 air change per hour (ach, or h"1).
7-57
-------
Ambient conditions, particularly temperature and wind velocity, can also drive air
exchange rates. Strictly speaking, it is the difference between indoor and outdoor temperatures
that creates either a pressure difference (closed windows) or a convective behavior (open
windows) leading to higher air exchange rates as the temperature difference increases. As wind
velocity rises, pressure differences also increase and therefore the air exchange rate rises.
Besides these immediate ambient conditions we also have climatic conditions. A region that can
expect a daily sea breeze is more likely to use open windows than air conditioning for
ventilation. Northern areas are more likely to have tightly constructed buildings than southern
areas in the USA.
In most cases, by far the most important factor affecting air exchange rates is the behavior
of the resident(s). This includes such considerations as the number of residents, the number and
age of children, the number of pets that spend time outdoors, whether or not air conditioning is
used, and how much time doors and windows are open. Since residents are more active during
the day, and doors are opened and closed more often, air exchange rates during the day typically
exceed those at night, both in winter and in summer. In the PTEAM Study, the median daytime
air exchange rate was 1.02 h"1 compared to an overnight median of 0.80 h"1 (Wallace et al.,
1993). In the Parkville community of Baltimore, MD, in the spring, the daytime median was
0.40 h"1 and the overnight median was 0.28 h"1. In Los Angeles coastal communities in the
summer, the daytime median was 2.2 h"1 and the overnight median was 1.2 h"1. (All values
derived from U.S. Environmental Protection Agency, 1995)
Fortunately, a large number of surveys have been carried out in which air exchange rates
of homes have been measured. These include the three major particle studies already mentioned,
and some studies of other pollutants. A paper collecting results from many surveys found a
geometric mean for 2844 U.S. residences of 0.53 h"1 with a geometric standard deviation of 2.3
(Murray and Burmaster, 1995). The mean value for all 2844 homes was 0.76 h"1, which
corresponded to the 70th percentile. However, the geometric means varied by season (a low of
0.31 h"1 in fall and a high of 1.00 h"1 in summer) and by region (a low of 0.31 h"1 in the North
and a high of 0.69 h"1 in the South—mainly southern California). The geometric standard
deviations for individual seasons and regions were generally very close to 2, ranging from 1.9 to
2.5. (It should be noted here that the homes were not selected to represent the nation, and that
there are very great disparities in the number of homes sampled in any one region.)
7-58
-------
A second paper (Koontz and Rector, 1995) used a nearly identical data set, but weighted
the 2889 measured homes by the state populations to estimate more closely the national
distribution. Their estimates are similar to those of Murray and Burmaster (1995) with an
arithmetic mean of 0.63 h"1, a geometric mean of 0.46 h"1 and a GSD of 2.25.
However, certain smaller areas with pronounced climatic conditions could have very much
higher air exchange rates. In a region such as the South Bay of Los Angeles, Wallace et al.
(1991c) found that 49 of 50 homes had no air conditioning and depended on the daily land-sea
breeze for ventilation. In this area, winter air exchange rates had a geometric mean of 0.75 h"1
and summer air exchange rates were much higher, with a geometric mean of 2.16 h"1. Both these
ranges are much higher than the typical values reported above. Thus, it is important to consider
the individual geographic region of study and its local climatic characteristics before selecting a
range of air exchange rates to characterize the region.
With that caveat, the empirical distribution for a large number of U.S. homes across all
seasons, but with disparate representation among the various regions of the country, appears to
have a median value of about 0.5 h"1, with a one geometric standard deviation (± a) range of 0.2
to 1.1 h"1, and a ±2o range of 0.1 to 2.2 h"1 (Murray and Burmaster, 1995; Koontz and Rector,
1995).
Deposition Rate k. In a residence, the deposition rate k depends on many factors, such as
scale of turbulence, and the size, shape, electrostatic charge, and density of the particle. For
larger particles, the deposition rate is determined largely by gravitational settling; for smaller
particles, deposition on vertical surfaces by diffusion may also be important (Nazaroff et al.,
1993). Unfortunately, fine particle deposition rates are not well characterized. Typically, one
must measure over very long periods of time (weeks to months) to collect enough particles for
analysis by sophisticated techniques. A series of studies in nearly unoccupied buildings
containing telephone-switching electrical equipment resulted in average values for the deposition
velocity of sulfate particles ranging from 0.003 to 0.005 cm/s (Sinclair et al., 1988, 1990, 1992;
Weschler et al., 1989); these values correspond to values of k (using a surface to volume ratio of
3 m"1) of 0.3 to 0.5 h"1. However, another series of studies in museums resulted in values an
order of magnitude smaller (Ligocki et al., 1990; Nazaroff et al., 1990a,b). Results for the sulfur
(PM25) deposition rate in the PTEAM studies were 0.16 h"1, lying between the values found by
7-59
-------
these two groups. Nazaroff et al. (1993) concluded that deposition rates could vary as a result
of different surfaces or near-surface air flows, amount of thermal isolation of the surfaces from
building walls, turbulence, and many other factors. Thus it is not likely that theoretical
calculations of deposition rates will provide trustworthy estimates. Nor is it likely that chamber
studies, with their limited ability to reproduce the variety of floor coverings and air flows found
in residences, can provide much information relevant to real-world residences.
In the absence of precise theory or widely applicable chamber study estimates, the largest
study of residences including a calculation of empirical deposition rates is the PTEAM study.
The estimate for PM2 5 was 0.39 h"1, for PM10 it was 0.65 h"1, while for the coarse fraction (the
difference between PM10 and PM2 5) it was 1.01 h"1.
What Is the Fraction of Outdoor Air Particles Found Indoors at Equilibrium?
Based on the values of/1, a, and k discussed above, an answer can be provided to this
question. Figure 7-16 shows the fraction of outdoor fine and coarse particles found in homes
under equilibrium conditions for a range of air exchange rates. This fraction is calculated using
the value of P = 1 determined in the PTEAM and the Thatcher and Layton (1995) studies, and
the values of & for fine and coarse particles calculated in the PTEAM study. The fractions are
displayed over the 95% range of observed air exchange rates (0.1 to 2.2 h"1) in studies reported
on by Murray and Burmaster (1995). It can be seen that at the mean air exchange rate of 0.76 h"1
reported in Murray and Burmaster (1995), the fractions of outdoor fine (<2.5 //m) and coarse
particles (>2.5 and <10 //m) that will be found indoors under equilibrium conditions are 66%
and 43%, respectively. The fraction of PM10 found indoors will lie between these two curves,
with the exact placement dependent on the relative proportions of fine and coarse particles
constituting the PM10.
The actual distribution of values of a/(a+k) observed in the PTEAM Study is provided in
Table 7-16 for PM10 and for its fine and coarse fractions. As can be seen, the average values
across day and night were about 67% for fine particles and 47% for coarse particles, with PM10
exactly between the two size fractions at 57%.
7-60
-------
0.5
1 1.5 2
Air exchange rate (air changes per hour)
2.5
Deposition rate = 0.39/h for fine particles, 1.01/h for coarse
Figure 7-16. Fraction of indoor particulate matter (PM) from outdoor airborne PM,
under equilibrium conditions, as a function of air-exchange rate, for two
different size fractions.
Source: Calculated from PTEAM database (Ozkaynak et al., 1993a; Wallace, 1996).
These results suggest that if persons at risk of health effects from outdoor particle pollution
are able to significantly decrease the air exchange rates in their homes (by weatherization,
installation of air conditioning to reduce use of windows, etc.) they could decrease the fraction
of outdoor air particle concentration in their homes. A decrease in the air exchange rate from the
mean level of 0.76 h"1 reported above to an achievable (16th percentile) value of 0.25 h"1 would
decrease the indoor air level of outdoor-generated fine PM2 5 particles from 66% to 39% of the
outdoor level, and of PM10 from 54% to 28%.
7.2.2.5 Studies of PM in Buildings
The single largest study of particles in buildings was carried out by the Lawrence Berkeley
Laboratory (LBL) for the Bonneville Power Administration (BPA) (Turk et al., 1987, 1989).
Thirty-eight buildings were chosen from two climatic regions in the Pacific Northwest: Portland-
Salem, OR (representing mild coastal conditions), and Spokane-Cheney,
7-61
-------
TABLE 7-16. FRACTION OF CONCENTRATION OF
OUTDOOR PARTICLES ESTIMATED TO BE FOUND INDOORS AT EQUILIBRIUM:
RESULTS FROM THE PARTICULATE TOTAL EXPOSURE ASSESSMENT
METHODOLOGY STUDY
Daytime (N=174)
Statistic
Mean
Standard deviation
Standard error
Geometric mean
Minimum
25th percentile
Median
75th percentile
Maximum
Fine
0.68
0.17
0.013
0.66
0.28
0.55
0.70
0.83
0.95
PM,n
0.58
0.19
0.015
0.55
0.19
0.42
0.58
0.75
0.93
Coarse
0.49
0.20
0.015
0.45
0.13
0.32
0.47
0.65
0.89
Overnight (N= 175)
Fine
0.66
0.15
0.012
0.64
0.28
0.55
0.66
0.79
0.94
PM,n
0.55
0.17
0.013
0.53
0.19
0.43
0.54
0.69
0.90
Coarse
0.46
0.17
0.013
0.42
0.13
0.32
0.43
0.59
0.85
Fractions calculated from the formula Pal(a+k), where
P=l;
A: = 0.39 h'1 for fine particles, PM2 5;
£= 0.65 h'1 for PM10; and
k= 1.01 h"1 for coarse particles 2.5 ,wm < AD < 10 /j,m.
Values for a measured in 175 homes during the PTEAM Study.
Source of data: Values calculated from PTEAM database (Wallace, 1996).
WA (representing extreme inland conditions). The buildings were studied for a variety of
pollutants to determine how ventilation rates affect indoor air quality. Buildings were measured
in winter (21 buildings in both regions), spring (10 buildings in both regions) and summer (nine
buildings in the inland region only). All but four buildings were government or public
properties, and therefore the 38 buildings cannot be considered to represent the full mix of
building types.
Each building was monitored for 10 working days over a two-week period. From four to
eight particle sampling sites were chosen in each building according to size. The sampler was an
LBL-developed flow controlled device with a 3 jim cutpoint. The pumps sampled only during
hours the building was occupied. If filters had to be changed due to excessive loading, the
combined weight of all filters from one site was determined—thus all values are approximately
7-62
-------
10 working-day (80-h) averages. Buildings had varied types of smoking policies, from
relatively unrestricted to very tightly controlled, as in one elementary school. In most buildings,
an attempt was made to site at least one monitor in an area where smoking was allowed. Data
were obtained from smoking areas in about 30 of the 38 buildings.
Results comparing smoking and non-smoking areas are provided in Table 7-17 and
Figure 7-17. Mean RSP concentrations in the smoking areas were more than three times higher
than in the non-smoking areas (70 versus 19 |ig/m3). Since these arithmetic means showed
evidence of being driven by one or two high values, the geometric mean (averaged across all
sites in a building) may be a better comparison. Here the ratio is very close to 3 to 1 (44 versus
15 |ig/m3). Outdoor results at 30 sites had the identical arithmetic mean as the indoor non-
smoking sites: 18.9 |ig/m3.
Repace and Lowrey (1980) sampled 19 establishments allowing smoking (seven
restaurants, three bars, church bingo games, etc.) and 14 where no smoking occurred (including
five residences and four restaurants) between March and early May of 1978. Sampling occurred
for short periods of time (2 to 50 min) using a TSI Piezobalance to measure PM3 5. Indoor
concentrations ranged from 24 to 55 |ig/m3 in the areas without smoking, and from 86 to 697
|ig/m3 in places with active smoking.
Miesner et al. (1989) sampled particles and nicotine in 57 locations within 21 indoor sites
in Metropolitan Boston, MA, between July 1987 and February 1988. PM2 5 was sampled using
Harvard aerosol impactors. Sampling times ranged from about 3 h in a bus station to 16 h in a
library, depending partly on how "clean" the environment was perceived to be. PM2 5
concentrations ranged from 6 |ig/m3 (in the library) to 521 |ig/m3 in a smoking room in an office
building. For 42 measurements in non-smoking areas, the mean PM2 5 concentration was 25 ±
30 (SD) |ig/m3. Six of these measurements included a classroom with visible levels of chalk dust
on the impactor, four measurements in subways, and the bus station. The remaining 36
nonsmoking areas had a mean PM25 concentration of 15 ± 7 |ig/m3. The 15 smoking areas
ranged from 20 to 520 |ig/m3 with a mean of 110 ± 120 |ig/m3.
Sheldon et al. (1988a,b) reported on the EPA 10-building study of hospitals, homes for the
elderly, schools, and office buildings. Particle measurements were taken in six buildings using a
National Bureau of Standards portable particle sampler (McKenzie et al., 1982) to
7-63
-------
TABLE 7-17. SMOKING, NONSMOKING, AND
OUTDOOR RSP CONCENTRATIONS AND RATIOS
Building No.
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
AM
ASD
GM
GSD
Outdoor —
C"g/m3)
ND
ND
ND
8
BD
35
35
8
8
9
8
ND
10
6
BD
10
7
7
7
18
17
20
11
11
68
32
52
65
29
33
13
ND
ND
16
18
20
19
14
11
11
19
16
14
2.2
Indoor
(//g/m3) Arithmetic Mean (Range)
Nonsmoking
25(19-36)
19(18-21)
ND
7(6-8)
13(13)
12(11-13)
38(32-44)
7(7-8)
11(11)
65(53-74)
23(9-49)
10(10)
5(5-6)
ND
11(7-14)
9(8-11)
11(10-13)
ND
ND
11(10-11)
11(9-12)
18(18)
9(BD-20)
44(10-77)
35(32-38)
45(20-70)
36(33-38)
36(29-43)
10(8-12)
24(20-30)
12(8-18)
13(10-17)
ND
13(10-16)
20(6-35)
14(9-18)
21(12-32)
7(BD-9)
8(8-9)
10(8-12)
19
14
15
1.9
Smoking0
ND
ND
20(16-25)
ND
14(14)
35(23-59)
39(39)
ND
16(13-20)
95(67-127)
209(209)
63(63)
ND
30(26-34)
12(12)
73(73)
105(105)
19(19)
20(11-29)
ND
ND
57(22-165)
ND
24(24)
109(109)
82(55-123)
61(33-89)
BD
144(144)
113(113
268(268)
36(21-52)
29(12-74)
54(13-117)
50(50)
72(17-127)
27(11-62)
308(308)
13(11-14)
26(11-40)
70
73
44
2.7
"Repeated test of building #11.
'Repeated test of building #17.
"Smoking within
10 m radius of site.
Mean"
25(19-36)
19(18-21)
20(16-25)
7(6-8)
13(13-14)
28(11-59)
38(32-44)
7(7-8)
15(11-20)
86(53-127)
63(9-209)
36(10-63)
5(5-6)
30(26-34)
11(7-14)
31(8-73)
40(10-105)
19(19)
20(11-29)
11(10-11)
11(9-12)
50(18-165)
9(BD-20)
37(10-77)
60(32-109)
67(20-123)
48(33-89)
24(BD-43)
32(8-144)
37(20-113)
64(8-268)
21(10-52)
29(12-74)
28(10-117)
23(6-50)
28(9-127)
25(11-62)
46(BD-308)
11(8-14)
15(8-40)
30
19
24
2.0
NA =
ND =
BD =
Indoor
Nonsmoking +
Outdoor
NA
NA
NA
0.9
NA
0.3
1.1
0.9
1.3
7.0
2.9
NA
0.5
NA
NA
0.9
1.6
NA
NA
0.6
0.7
0.9
0.8
4.0
0.5
1.4
0.7
0.6
0.3
0.7
0.9
NA
NA
0.8
1.1
0.7
1.1
0.5
0.7
0.9
1.2
1.3
0.9
2.0
Not applicable.
No data collected.
Below detection limit.
Ratios
Indoor
Smoking +
Outdoor
NA
NA
NA
NA
NA
1.0
1.1
NA
2.0
11.0
26.1
NA
NA
5.0
NA
7.3
15.0
2.7
2.9
NA
NA
2.9
NA
2.2
1.6
2.6
1.2
NA
5.0
3.4
20.6
NA
NA
3.4
2.8
3.6
1.4
22.0
1.3
2.4
6.0
7.2
3.6
2.6
Indoor
Mean +
Outdoor
NA
NA
NA
0.9
NA
0.8
1.1
0.9
1.9
9.6
7.9
NA
0.5
5.0
NA
3.1
6.1
2.7
2.9
0.6
0.7
2.5
0.8
3.4
0.9
2.1
0.9
0.4
1.1
1.1
4.9
NA
NA
1.8
1.3
1.4
1.3
3.3
1.0
1.4
2.3
2.2
1.7
2.3
11 Arithmetic average of all sites in building.
Source: Turk et al. (1987).
7-64
-------
80
70
-v 60
*>
I 50
§»
^ 40
jg 30
20
10
0
Mean Concentrations
Smoking areas Nonsmoking areas Outdoors
Geometric Means
50
40
30
20
10
Smoking areas Nonsmoking areas Outdoors
Figure 7-17. Comparison of respirable particles in smoking and nonsmoking areas of
38 buildings in the Pacific Northwest.
Source: Turk et al. (1987).
7-65
-------
collect two size fractions: PM3 and a coarse fraction between PM3 and PM15. The sampler
employed two filters in series: an 8.0 jim Nuclepore filter for PM15 and a 3 jim Ghia Zefluor
Teflon filter for fine particles. The flow rate was 6 Lpm for a 24-h sample. Three consecutive
24-h samples were collected at each building. Additional particle monitoring was provided at
certain locations (e.g., smoking lounge, cafeteria) using a Piezobalance (PM3 5) and a
dichotomous sampler (PM2 5 and PM10).
In areas without smoking, indoor concentrations of both size fractions were generally
lower than outdoor levels; for example, the coarse fraction ranged from 0.2 to 0.66 of the
outdoor level (13 to 17 |ig/m3) in the three buildings with no smoking. The fine fraction was
present at higher indoor-outdoor ratios, ranging from 0.56 to 0.99 in the same three buildings
(outdoor fine fraction ranged from 16 to 33 |ig/m3). The fine fraction was elevated in the
regions of smoking (range of 14 to 56 |ig/m3). Piezobalance results for several buildings showed
uniformly low (7 to 29 |ig/m3) for 800 min of monitoring in nonsmoking areas.
Concentrations in the areas allowing smoking were more often in the 40 to 60 |ig/m3 range,
with short-term peaks as high as 345 |ig/m3. It was possible to use the observed declines in
PM3 5 following cessation of smoking to calculate an effective air exchange rate and thus a
source strength for PM3 5 emissions from cigarettes. Four estimates gave an average value of
about 6 mg/cigarette, somewhat below the chamber study estimates of 10 to 15 mg/cig. An
estimate due to Repace and Lowrey (1980) of concentrations of respirable parti culates due to
smoking was also tested, with good agreement. The Repace and Lowrey equation is
(7-7)
where Pa is smoking occupancy in persons per 100 square meters and a is the air exchange rate
h"1. Equation 7-7 was developed assuming one of every three occupants are smokers who smoke
two cigarettes per hour. Assuming a background concentration of 15 |ig/m3, the measured
values for the smoking lounge for zero, three, and nine smokers were 10, 78, and 284 |ig/m3,
respectively. Equation 7-7 predicts 0, 99, and 296 |ig/m3, respectively. In two of the homes for
the elderly, apartments with smokers and nonsmokers were measured for three consecutive days
using the NBS samplers. In one building, the smoker's apartment had a 2-day PM3 average of 39
|ig/m3, compared to 9.4 |ig/m3 in the nonsmoker's apartment; in the other home for the elderly,
7-66
-------
where two smokers shared one apartment, the average 2-day PM3 concentration was 88 |ig/m3
compared to 8.6 |ig/m3 in the nonsmoking apartment. The simultaneous ambient values were not
measured at Home 1. At Home 2, the ambient value was 11 //g/m3.
Owen et al. (1990) studied particle size distributions in an office under varying conditions
of ventilation and occupancy. The unoccupied office using minimum outdoor air had
concentrations at least as low as the occupied office using maximum outdoor air. PM3 5
concentrations (measured using the TSI Piezobalance) were about twice as high (75 versus
39 //g/m3) in the occupied office when the dampers were closed as when they were open. The
main source of particle generation appeared to be the hallway, suggesting that resuspension of
tracked-in dust was an important indoor source of particles as reported by Roberts et al. (1990)
for residences.
7.2.3 Indoor Air Quality Models and Supporting Experiments
Indoor concentrations of particles are a function of penetration of outdoor particles and
generation of particles indoors. The concentrations are modified by air exchange rates and
deposition rates of the particles onto indoor surfaces.
7.2.3.1 Mass Balance Models
Mass balance models have been used for more than a century in various branches of
science. All such models depend on the law of the conservation of mass. They simply state that
the change in mass of a substance in a given volume is equal to the amount of mass entering that
volume minus the amount leaving the volume. Usually they are written in the form of first-order
linear differential equations. That is, consider a volume V filled with a gas of mass m. The
change in mass Aw over a small time A^ will simply be the difference between the mass entering
the volume (min) and the mass leaving the volume (Vwout):
A/77
At At
Taking the limit as At approaches zero, we have the differential equation for the rate of change
of the mass:
7-67
-------
(7-9)
If we require that the mass be uniformly distributed throughout the volume at all times, we
have a condition that the physical chemists call "well-mixed". We assume that any mass gained
or lost in the volume Fis instantaneously distributed evenly throughout the volume. We may
then replace the mass term (rri) by the concentration C = m/V, so that dm/dt = V dC/dt.
The above equations are the basis for all such mass-balance models. Equation 7-9 takes on
many forms depending on the type of processes involved in transporting mass into or out of the
volume being considered. A large class of models assume that the volume Fis a single perfectly
mixed compartment. More complex models assume multiple compartments to allow for
incomplete mixing in the total volume F(Mage and Ott, 1996). A detailed mass-balance model
that includes changes in particle size, chemical composition, and turbulence is described in
Nazaroff and Cass (1989).
7.2.4 Summary of Indoor Particulate Matter Studies
At low outdoor levels of fine (PM3 5 or PM2 5) particles (as in most of the cities in the
Harvard Six-City and New York State studies), mean indoor concentrations have been found to
be twice as high as outdoor levels. However, for homes without smokers or combustion sources,
indoor levels are often roughly equal to outdoor levels (Santanam et al., 1990; Leaderer et al.,
1994; Neas et al., 1994). At high outdoor levels, mean indoor concentrations have been about
10% lower than the mean outdoor concentrations in the two areas studied (Steubenville, OH, and
Riverside, CA). Indoor concentrations are considerably higher during the day, when people are
active, than at night. Based on a mass-balance model, outdoor air was the major source of
indoor particles in the PTEAM study, providing about 3/4 of fine particles (PM25) and 2/3 of
inhalable particles (PM10) in the average home. However, outdoor air contributed less than half
of the indoor particle concentrations at seven out of eight other sites with extensive indoor-
outdoor measurements. Indoor concentrations are much higher during the day, when people are
active, than at night.
In the PTEAM study (with very high outdoor particle concentrations), indoor levels were
significantly influenced by outdoor levels, but with relatively low R2 values ranging between
7-68
-------
0.22 and 0.54. In the other two major studies, no significant indoor-outdoor relation was
observed. Regressions of indoor on outdoor particles seldom explained more than half the
variance of any study (R2 < 50%). However, in those studies with repeated measures on the
same house, (e.g., the PTEAM prepilot [Table 7-6], the Phillipsburg, NJ, study [Table 7-15] and
Tamura et al. [1996] in Section 7.4.2.1), longitudinal regressions of indoor on outdoor particles
often had much higher R2 values of 0.6 to 0.9 for each individual house. Since the
epidemiological studies of health effects of particles have been studies of variation overtime, the
longitudinal regressions by individual home are expected to be more relevant to the
epidemiology studies than cross-sectional regressions across all homes in the study. The better
relationship showed by these regressions suggests that whatever structural or behavioral
characteristics affect indoor particle concentrations in the home tend to persist or be repeated
over time. This gives better support to the epidemiological findings than would be inferred from
the typically low R2 values reported for the cross-sectional regressions performed in most
studies.
Deposition rates k ranged from 0.16 h"1 for sulfur to 0.4 h"1 for fine (PM2 5) particles to 1 h"1
for coarse particles (PM10 - PM25), with an intermediate estimate of 0.65 h"1 for PM10. The
penetration factor P for both fine and coarse fractions was estimated to be unity. For a home
with no indoor sources whatever and a typical air exchange rate of about 0.75 h"1, these values
for k and P would imply that sulfur indoors would be about 0.757(0.16 + 0.75) = 82% of the
outdoor value at equilibrium, fine particles indoors would be about 0.757(0.4+0.75) = 65% of the
outdoor value at equilibrium, indoor PM10 would be about 54% of outdoor levels, and indoor
coarse particles would be about 43% of outdoor levels. Since very few homes were observed to
have concentrations this low, it can be inferred that very few homes are free of important indoor
sources of particles.
A crucial question is the impact of outdoor particles on indoor particle concentrations. It
was found that the governing equation is a function of only two parameters: air exchange rate a
and particle deposition rate k: a/(a+k). Air exchange rates measured in the United States appear
to follow a roughly log-normal distribution with a geometric mean of 0.5 and a geometric
standard deviation about 2. With the values for the deposition rates provided above, one can
calculate the impact of outdoor particles on indoor concentrations for any given value of the air
exchange rate. At a low air exchange rate of, say, 0.4 h"1, sulfates indoors will be 71% of their
7-69
-------
outdoor values, fine particles indoors will be 50% of their outdoor values, while coarse particles
will be 0.4/1.4 or 28% of their outdoor values. At a higher air exchange rate of 1 h"1, sulfates
will be 86% of their outdoor concentration, fine particles will be 1/1.4 or 71% of their outdoor
concentration, whereas coarse particles will be 50% the outdoor concentration. The difference
in both cases between the two size fractions is about 0.2; that is, for the entire range of realistic
air exchange rates (from 0.2 h"1 to 2 h"1), if the fraction of outdoor coarse (PM10 - PM25)
particles found indoors is/ then the fraction of fine particles found indoors will be
approximately/+ 0.2. It can be seen that a reduction in air exchange rate would reduce the
impact of outdoor air on indoor air particle concentrations.
7.2.5 Bioaerosols
Biologically-derived particles are frequently ignored components of both ambient and
indoor aerosols. This lack of attention is, in part, due to the fact that the bioaerosols are
considered "natural" and not amenable to control. Methods for their analysis are, in many cases,
highly variable, and very little exposure or exposure/response information is available.
Measurement methods for bioaerosols are discussed in Chapter 4 (Section 4.4). Various health
effects associated with bioaerosols are discussed in Chapter 11. A few reference works that
focus on bioaerosols include Gregory (1973), Edmonds (1979), Cox (1987), Lighthart and Mohr
(1994), and Cox and Wathes (1995).
For bioaerosols, there is considerable confusion among the terms reservoir, source,
particle, and agent. For the purposes of this chapter, the following definitions apply:
• Reservoir: the environmental niche in which source organisms are living
• Source: the organism that produced the particle
Particle: the particle shed from the organism
• Agent: the part(s) of the particle that actually mediate the disease process.
Examples of bioaerosol sources, particles and agents are presented in Table 7-18.
7-70
-------
TABLE 7-18. AN OVERVIEW OF ORGANISMS, AEROSOLS,
AND DISEASE AGENTS
Sources
Aerosol Particles
Disease Agents
Plants
Animals
Fungi
Bacteria
Viruses
Pollen and pollen fragments, fragments of other
plant parts, spores (ferns, mosses), algal cells
Skin scales, secretions (saliva, skin secretions),
excreta, body parts (arthropods)
Spores, hyphae, yeast cells, metabolites (toxins,
digested substrate material)
Cells, fragments, metabolites (toxins, digested
substrate material)
Viral particles
Glycoprotein allergens
Glycoprotein allergens
Glycoprotein
allergens, infectious
units, glucans,
mycotoxins
Infectious units,
allergens, endotoxin,
exotoxins
Infectious units
7.2.5.1 Plant Aerosols
Pollen
Pollen is produced by vascular flowering plants: trees (pines, cedars, birch, elm, maple,
oak, hickory, walnut, etc.), grasses, and weeds (ragweed, sage, Russian thistle, lambs quarters,
etc.). Within these large groupings, specific types are regionally common. For example,
ragweed is most common in the eastern United States. Birch pollen dominants the spring pollen
season in New England, while mountain cedar pollen is abundant early in the year in the
southwest (Lewis et al., 1983).
Pollen levels outdoors are controlled by the number of plants available for pollen release,
the amount of pollen produced by each plant, factors that control pollen release and dispersion
from the plant, and factors that directly affect the aerosols (Edmonds, 1979). The number of
plants available depends on the many environmental factors that control plant prevalence, some
of which are human factors. As an example, the abundance of the ragweed plant in a particular
year depends on the number of plants that produced seed in the previous year, disturbed ground
available for seed germination and growth, and meteorological factors during the growing
season. Once a crop of ragweed has been produced, pollen production depends on temperature,
rainfall, and day length.
7-71
-------
Pollen grains are relatively large complex particles that consist of cellular material
surrounded by a cell membrane and a complex wall. Pollen grain structure has been well
studied. Pollen shed is controlled by temperature, humidity, wind, and rain. Pollen levels in air
depend on all of these factors as well as wind and rain conditions after release, and on surfaces
available for impaction. Figure 7-18 represents day to day ragweed pollen prevalence in
Kalamazoo, MI, for 1994.
1,000
1994 Kalamazoo, Michigan
100-
10-
II 111 .1. . .1. Jill Jllllllll
111. Ill IIII ill 11 Ijllllllll „
IIIIIIIIllll.llllllll JlllllllIIIIIJll 1
•••I iiiiiiiiiiiiiii i .ill liiiiiiiiiiiiiiiiiiiiiiiiiiiiii" i"
1IIIIII1II1«^^^^ I' i
•in iiiiiiir ill iiiiiii ill
11 IIIIIIIIIIIIIII Jill III'
.Illllllllllllllllll ,1 ,1, 1 JlLlLlllllIlllllllIllllllllllllllllllI
' uniform
iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii iiiiiiiiiiiiiiiii iiiiiiiiiii 1111111111111" i
llllllllllllllllllllllllllllllllllllllllllll
Illlllllllllllllllllllllllllllllllllllllllllllllllllllllllll IIIIIIIIIII
Jan-1 Jan-31 Mar-2
Apr-1 May-1 May-31 Jun-30 Jul-30 Aug-29 Sep-28 Oct-28 Nov-27 Dec-27
24-hour Total Pollen Counts
Figure 7-18. Chart of ragwood pollen prevalence. Sampling was not conducted before
April and during the first few days of October.
Source:
Pollen allergens are (apparently) water-soluble glycoproteins that rapidly diffuse from the
grain when it contacts a wet surface. The glycoproteins are (generally) specific to the type of
pollen, although large groups may be represented by a single allergen. For example, many
different kinds of grasses carry similar allergens in their pollen grains. A number of pollen
allergens have been characterized: Amb a I (ragweed), Bet p I (birch), Par j I (parietaria), etc.
7-72
-------
Other Natural Plant Aerosols
Other plant-derived particles that are a natural part of outdoor air include algal cells; spores
of mosses, liverworts, club mosses, and ferns; and fragments of all kinds of plants. Very little
has been reported about the prevalence or human impact of any of these aerosol particles,
although they are presumed to carry allergens.
Man-Made Plant Aerosols (Soy, Latex, Occupational)
Man-made accumulations of plant material that are subsequently handled inevitably
produce bioaerosols. The most common practices that involve such accumulations are storage,
handling, and transport of farm products (hay, straw, grain), composting, and manufacturing
processes that involve the use of plant material. In addition, the use of some plant products can
result in disease-causing aerosols (Alberts and Brooks, 1992). The aerosols produced from most
of these processes are complex, and few have been accurately characterized.
Grain Dust. It is well-recognized that grain dusts include respirable-size particles
(< 10 //m) although the exact nature of the particles and the agents of disease remain speculative.
Soybean dust aerosols released from freighters unloading the beans in port have been blamed for
epidemics of asthma.
Wood Dust. Wood trimmer's disease (from particles released from wood during high-
speed cutting). Sewage composting involves the use of wood chips that can release allergenic
aerosols.
Latex. Latex-containing powder aerosols are produced when surgical gloves are used.
Latex particles also may be released from automobile tires.
7-73
-------
7.2.5.2 Animal Aerosols
Mammalian Aerosols
All mammals produce aerosols, from humans to the smallest mouse. Human aerosols (skin
scales, respiratory secretions) do not cause disease except, of course, for agents of infection (see
below). Other mammals release aerosols that cause hypersensitivity diseases. The most
common of these are cats, dogs, farm animals, laboratory animals, and house mice, although all
animals release aerosols that could be sensitizing under appropriate conditions (Burge, 1995).
Mammals only cause human disease when appropriate exposure conditions occur. For cats,
simply having a cat in a house will create such conditions, as will handling any animal regardless
of the environment. Cat allergens apparently become aerosolized on very small particles (<1
//m) shed from skin and saliva. There is some indication that dog, mouse, and other rodent
allergens are borne on dried urine particles, and particle sizes are similar to those of cat allergen.
Little is known about other mammalian aerosols. Cat and dog allergens have been characterized
(Pel d I, Can f I) and other mammalian allergens are under active study.
Avian Aerosols
Wild and domesticated birds associated with disease-causing aerosols include for example:
starlings (histoplasmosis); pigeons (histoplasmosis, pigeon-breeders disease); parrots
(psittacosis); poultry (poultry-handlers disease); etc. Of these diseases, only the hypersensitivity
diseases (pigeon breeders and poultry handlers disease) are caused by "bird" aerosols. The
others are infections caused by agents inhabiting the birds (see below). The birds that release
antigens that have caused human disease are those that are confined or congregate close to
people. The avian aerosol-hypersensitivity diseases are almost exclusively confined to sites
where birds are bred and handled extensively, especially in indoor environments. Relatively
little is known about avian aerosols. Probably skin scales, feather particles, and fecal material
are all released as antigen-containing aerosols. The antigens (allergens) responsible for avian-
related diseases have not been characterized.
7-74
-------
Insect Aerosols
Dust Mites. Dust mites are arthropods belonging to the family Pyrogliphidae. There are
two common species in temperate climates: Dermatophagoides farinae (which proliferates
under relatively dry conditions) and D. pteronyssinus which dominates populations in more
humid environments (Arlian, 1989). Dust mites thrive in environments where relative humidity
consistently exceeds 60 % and where skin scales and fungal spores are available as a food
source. Primary reservoirs for exposure are bedding and carpet dust. The mite itself is about
100 //m long, but it excretes 20 //m membrane-bound fecal particles that contain the allergens.
Exposure to dust mite allergens apparently occurs only when reservoirs are disturbed. Dust
mites produce allergens that are a major cause of sensitization in children. The allergens are
digestive enzymes that gradually diffuse from fecal particles after deposition on mucous
membranes. Several dust mite allergens have been characterized and monoclonal antibodies
against each have been raised and cloned. These include Der f I and II, and Der p I and II
(Platts-Mills and Chapman, 1987).
Cockroaches. Cockroaches are insects belonging to the Orthoptera (Mathews, 1989). The
most common cockroach infesting temperate climate buildings is Blatella germanica, the
German cockroach. Cockroaches are nocturnal, and inhabit dark environments where food and
water are available. Common food sources include stored animal or human food, and discarded
food (garbage). Cockroaches are extremely prolific, given appropriate environmental
conditions. Population pressure will eventually drive the roaches into the daylight in search of
food. Cockroaches shed body parts, egg cases, and fecal particles, all of which probably carry
allergens. Little is known about the particles that actually carry the allergens. Two German
cockroach allergens have been characterized: Bla g I, and Bla g II. The function within the
cockroach of these allergens is unknown. Cockroach allergens are probably a major cause of
asthma for some populations of children.
Other Insects. Fragments of gypsy moths and other insects that undergo massive
migrations can become abundant in ambient air. Sizes, nature, and allergen content of such
particles have not been studied. Cases of occupational asthma from exposure to insects (e.g.,
sewer flies) have been reported.
7-75
-------
Other Animal Allergens
It is likely that proteinaceous particles shed from any animal could cause sensitization if
exposure conditions are appropriate. For example, exposure to proteins aerosolized during
seafood processing have caused epidemics of asthma.
7.2.5.3 Fungal Aerosols
Fungi are primarily filamentous microorganisms that reproduce and colonize new
environments by means of airborne spores. Most use complex non-living organic material for
food, require oxygen, and have temperature optima within the human comfort range. The major
structural component of the cell wall is acetyl-glucosamine polymers (chitin). Cell walls also
may contain B-glucans, waxes, mucopolysaccharides, and a wide variety of other substances. In
the process of degrading organic material, the fungi produce CO2, ethanol, many other volatile
organic compounds, water, organic acids, ergosterol, and a broad spectrum of secondary
metabolites including many antibiotics and mycotoxins.
The fungi colonize dead organic materials in both outdoor and indoor environments. Some
fungi are able to invade living plant tissue and cause many important plant diseases. A few
fungi will invade living animal hosts, including people. Fungi are also universally present in
indoor environments unless specific efforts are made for their exclusion (i.e., as in clean rooms).
The kinds of fungi that are able to colonize indoor materials are generally those with broad
nutritional requirements (e.g., Cladosporium sphaerospermum), those that are able to colonize
very dry environments (e.g., members of the Aspergillus glaucus group), or organisms that
readily degrade the cellulose and lignin present in many indoor materials (e.g., Chaetomium
globosum, Stachybotrys atra, Merulius lacrymans). Yeasts (which are unicellular fungi) and
other hydrophilic taxa (e.g., Fusarium, Phialophora) are able to colonize air/water interfaces.
Water, in fact, is the most important factor controlling indoor fungal growth, since food sources
are ubiquitous (Kendrick, 1992).
Particles that become airborne from fungal growth include spores (the unit of most fungal
exposure), fragments of the filamentous body of the fungus, and fragments of decomposed
substrate material. Fungal spores range from about 1.5 //m to >100 //m in size and come in
many different shapes. The simplest are smooth spheres; the most complex are large
multicellular branching structures. Most fungal spores are near unit density or less. Some
7-76
-------
include large air-filled vacuoles. Fungal spores form the largest and most consistently present
component of the outdoor bioaerosols. Levels vary seasonally, with lowest levels occurring
during periods of snow. While rain may initially wash large dry spores from the air, these are
immediately replaced by wet (hydrophilic) spores that are released in response to the rain.
Some kinds of spores are cosmopolitan in outdoor air (e.g., Cladosporium herbamm,
Alternaria tenuissima). Others produced by fungi with more fastidious nutritional requirements
are only locally abundant. A typical indoor fungal aerosol is composed of particles penetrating
from outdoors, particles released from active growth on indoor substrates, and reaerosolized
particles that have settled into dust reservoirs. Indoor fungal aerosols are produced by active
forcible discharge of spores, by mechanisms intrinsic to the fungus that "shake" spores from the
growth surface, and (most commonly) by mechanical disturbance (e.g., air movement,
vibration).
Allergic rhinitis and asthma are the only commonly reported diseases resulting from fungal
exposures outdoors, and which also commonly occur indoors. The allergens of fungi are
probably digestive enzymes that are released as the spore germinates. Other spore components
(of unknown function) may also be allergenic. Only very few fungal allergens (out of possibly
hundreds of thousands) have been characterized: (e.g., Alt a I, Cla h I, and AspfT).
Allergic fungal sinusitis and allergic bronchopulmonary mycoses occur when fungi
colonize thick mucous in the sinuses or lungs of allergic people. The patterns of incidence of
allergic fungal sinusitis may be explained in part by geographic variability in ambient fungal
exposures. Figure 7-19 shows total fungal spore counts in Kalamazoo, MI, for 1994. This
disease is most commonly caused by Bispora, Curvularia, and other dark-spored fungi.
Exposure patterns required for allergic bronchopulmonary mycoses are unknown. This disease
is usually caused by Aspergillus fumigatus. Histoplasmosis and Coccidioidomycoses are fungal
infectious diseases that result from outdoor exposures to Histoplasma capsulatum (a fungus that
contaminates damp soil enriched with bird droppings) and Coccidioides inmitis (a fungus that
growth in desert soils. Indoor aerosol-acquired fungal infections are rare, and restricted to
immunocompromised people (Rippon, 1988).
7-77
-------
1994 Kalamazoo, Michigan
10,000
« 1,000
"«
u
a
u
S 100'
10-
"i.<:»; •<
liii"!!!!
;,;;; NI ;;; i, INN imimiiN,. i i,;
I Mill MM IIIIMI Illllllimilllllllll lllll.ll
I .illlllillll llllllll!.lllllllllllllllllllll.lllinill
IIIIIIIIILilllllllllllll
MM
llli..lllllllll II .lii Illllli.lllllll II
I. Illllllllllllll 1.1, Illl Illlllllllllllllll
,i. iiiiiiiiiiiuiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii iiiiiiiiiiiiiiiiiii miii 1111 illinium
Illllli
.I Illilll I
1.IIIIIII Illl
mi MI niii 11 inmiiMiiii iiiiiiMiiMiiiiiiiMiiiiiiiiiiiiMiiiiiiiMiiMiiiiiiiiniii
Illl ." .1.11 Hill II. Illl IIIIIIIIIIIIII..IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
I . ,,1111II. Illl I IllllllllJIIIlllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllir
I!!.!!!!!!!!!!!!!!.!.!!!!!I!!!!!!!!!!!!!!!!!)!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!)!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!! !!!!!!!!!n! !!!!!!!!!!!!{!!!!! ;I!!!!!!!!!!!!!!!!!!!!!
iiiiiiiiiiiiiiiiiii!! iiiiii iniiSiiiSiiiiiiiiiiiiinliiilra
Illllllllllllllllllll IIIIII Illlllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll
•!•••
iiiiiiiiiiiimimil iiiiii iliiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiliiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
iiiniiiiiiiiiiiiiiii iiiiii iniiiiiiiiiiiiiiiiiiiiiiiiiiiiiiniiiniiiiiiiiiiiiiiiiiiiiniiiiiiiiiiiiiiiiiiiiiiiiiniiiiiiii
~ I IIIIII Illl - - - ~ ^—~~ -.--^.-.
iiilliiiiliiiiiiiii IIIIMHII iiiiiiiiiiilliiii!
Illllllllllllll Illlllllllll Illlllllllllllllll,
1111111111111111111111
iiiiiiiiiiliiiniiiiii
Illlllllllllllllllllll.
1-
Jan-1
Jan-31
Mar-2
Apr-1
May-1 May-31 Jun-30 Jul-30
24-hour Total Spore Counts
Aug-29 Sep-28 Oct-28 Nov-27 Dec-27
Figure 7-19. Chart of fungal spore prevalence in Kalamazoo, MI, for 1994.
Source:
Toxic agents produced by fungi include antibiotics, mycotoxins, and some cell-wall
components that have toxic or irritant properties. The antibiotics and mycotoxins are secondary
metabolites that are produced during fungal digestion of substrate materials, and their presence
depends, in part, on the nature of the substrate. The locations of the toxins in spores or other
mycelial fragments are unknown, as are the dynamics of release in the respiratory tract. Aerosol
exposure to fungal antibiotics in levels sufficient to cause disease is unlikely. Mycotoxicoses
have been reported as case studies from exposure to spores of Stachybotrys atra (Croft et al.,
1986), and epidemiologically for Aspergillus flavus (Baxter et al., 1981).
7.2.5.4 Bacterial Aerosols
Bacteria, in contrast to plants, animals and fungi, contain neither nuclei or mitochondria.
Most are unicellular, although some form "pseudo" filaments when cells remain attached
following cell division. The actinomycetes are bacteria that do form filaments and (in some
cases) dry spores designed for aerosol dispersal. The bacteria can be broadly categorized into
two groups based on a response to the Gram stain procedure. The cell walls of Gram positive
bacteria are able to absorb a purple stain; the walls of Gram negative bacteria resist staining.
The Gram negative cell wall contains endotoxin (see above).
7-78
-------
Most infectious agents are maintained in diseased hosts. A few, including Legionella
pneumophila, reside in water-filled environmental reservoirs such as water delivery systems,
cooling towers, air conditioners, and (outdoors) oceans, lakes, streams, etc.
Infectious agents are often released from hosts in droplets released from the respiratory
tract. Each droplet contains one or more of the infectious agent, probably one or more other
organisms, and respiratory secretions. Most droplets are very large and fall quickly. Smaller
droplets dry quick to droplet nuclei, which range in size from the size of the individual organism
(<1 //m for the smallest bacteria) to clumps of larger organisms (>10 //m for larger bacteria).
Environmental-source aerosols are produced by mechanical disturbances that include wind, rain
splash, wave action, and by mechanical disturbance such as occurs in recirculation and sprays of
washes and coolants, and in humidifiers. Particle sizes from all of these activity cover a wide
range from well below 1 //m to >50 //m. The thermophilic actinomycetes produce dry aerial
spores that require only slight air movements to stimulate release. Each spore is about 1 //m in
diameter.
Whole living bacteria are agents of infectious disease (e.g., Tuberculosis, Legionnaires'
disease). For tuberculosis, a single virulent bacterial cell deposited in the appropriate part of the
lung is likely to cause disease in a host without specific immunity. For Legionnaires' disease,
the number of organisms required to make disease development likely depends on how well the
host's general protective immune system is operating. Some bacteria release antigens that cause
hypersensitivity pneumonitis. The antigens may be enzymes (e.g., Bacillus subtilis enzymes
used in the detergent industry) or may be cell wall components as in the thermophilic
actinomycetes. Bacteria also produce toxins of which endotoxin is the most important from an
aerosol exposure point of view.
7.2.5.5 Viral Aerosols
The viruses are units of either RNA or DNA surrounded by a protein coat. They have no
intrinsic mechanism for reproduction, and require living cells whose enzyme systems they utilize
to make new particles. They can be crystallized and remain able to reproduce, and are often
considered intermediates between non-life and life. Because viruses require living cells to
reproduce, reservoirs for them are almost exclusively living organisms. Rarely, viruses survive
(but do not reproduce) in environmental reservoirs from which they are re-aerosolized to cause
7-79
-------
disease. The Hanta virus that causes severe respiratory disease in people exposed to intense
aerosols of infected mouse urine is an example of this phenomenon. Viral aerosols are produced
when the infected organism coughs, sneezes, or otherwise forces respiratory or other secretions
into the air. The viral particles are coated with secretions from the host, and, as for the bacteria,
there may be one to many in a single droplet. The size of a single viral particle is very small (a
small fraction of a //m). However, infectious droplets are probably within a much larger size
range (1 to 10 //m). Each kind of virus produces a specific disease, although some of the
diseases present with similar symptoms. Thus, the measles virus produces measles, the chicken
pox viruses produces chicken pox and shingles. Influenza and common colds are produced by a
range of viruses all of which produce symptoms that are similar (but not necessarily identical).
7.2.5.6 Ambient and Indoor Air Concentrations of Bioaerosols
A general rough estimate of the contribution of bioaerosols to collected PM mass can be
made as follows: for an "average" 3 jim spherical spore of 0.9 density, each spore would weigh
~ 13 x 10"6 jig; for a clean indoor environment with ~ 103 spores/m3 the mass would be on the
order of 0.01 |ig/m3; for a typical outdoor condition, with ~ 50 x 103 spores/m3, the contribution
would be on the order of 0.5 |ig/m3. In contaminated indoor environments, where spore levels
above 106 spores/m3 are possible, the spore weight could be on the order of 10 |ig/m3 or more.
In summary, the minor mass concentrations of bioaerosols in ambient and indoor air are
independent of the concentrations of the non-bioaerosol constituents in ambient and indoor air.
However, the deposition of bioaerosols at the same respiratory tract loci as the other PM can
cause irritation and infection foci that may make the affected host more susceptible to the effects
of other deposited PM.
7-80
-------
7.3 DIRECT METHODS OF MEASUREMENT OF HUMAN PM
EXPOSURE BY PERSONAL MONITORING
7.3.1 Personal Monitoring Artifacts
Human exposure to air pollution can be measured by placing a personal exposure monitor
(PEM) close to the breathing zone of an individual. However, the very act of studying the
subjects can alter their behavior, which influences the measured values of their exposures and
creates an erroneous reading. This influence, known as the "Hawthorne Effect" (Mayo, 1960;
Last, 1988), arises because the subjects are aware of the study objectives, and the presence of the
PEM on their body is a constant reminder.
The physical location of the monitor inlet, as worn by the subject, can also influence the
subject's PM exposure and the recorded PM (Cohen et al., 1982, 1984). The movements of the
subject's body and the PEM sampling flow rate can alter the air currents in the subject's
breathing zone. "The presence of the body and its movement affect what a personal sampler
collects" (Ogden et al., 1993). When in close proximity to a source actively emitting PM (within
a meter) a small change in PEM position (e.g. from left side to right side) can vary the PM
measurement. The vertical position of the personal monitor sampling inlet (e.g., at the waist or
at the lapel near the breathing zone), can influence the captured amount of PM that is generated
from the floor and stuffed furniture (Aso et al., 1993).
In performance of a personal monitoring study, people often refuse to participate. The
refusal rate increases with the burden on the respondents due to the time required to complete
questionnaires, diaries and the need to carry the personal monitor with them throughout the
study. If the cohort of people who refuse to participate have significantly different personal PM
exposures than the participants, then the study will produce a biased estimate of the exposures of
the total population.
Two other important errors that influence the personal exposure measurements are:
(1) "the monitor effect", by which the monitor reduces PM concentration in the breathing zone
by "self dilution" (Cohen et al., 1984), the alteration of stream lines in the area of the nose and
mouth, or by electrostatic charge on a plastic cassette filter holder collecting charged particles
(Cohen et al., 1982); and (2) "the subject effect", by which the subject
7-81
-------
contaminates the data set by a purposeful action, such as blowing smoke into the inlet, or
forgetting to wear the monitor and not admitting that error in the log of daily activity.
These unquantifiable "errors" in a PM PEM measurement study may be greater than the
filter weighing errors and flow rate measurement errors that can be quality controlled through
calibration procedures. This may be important for interpretation of published PM PEM data
because these errors likely inflate the variance of the measurements.
7.3.2 Characterization of Participate Matter Collected by Personal
Monitors
The amount of PM collected by different types of personal monitors with the identical
nominal cut-point can be variable. The difference between two PM measurements, made by two
nominally identical monitors of different design, can be a function of the wind speed and the size
distribution of the PM in the air mass being sampled. A recent field comparison by Groves et al.
(1994) of different types of respirable dust samplers used in occupational settings where coarse
mode PM predominates shows that there is considerable difference between the mass collected
by sets of paired cyclones and paired impactors sampling in a concentration range of 500 to 6600
Mg/m3. The cyclones collected from 53 to 165% of the mass collected by the impactors. This
type of comparison study has not been done for personal monitors used in nonoccupational
studies at ambient and indoor respirable PM concentrations on the order of 10 to 100 //g/m3,
where the fine mode can be more important.
7.3.3 Microscale Variation and the Personal Cloud Effect
The study of Thatcher and Layton (1995) described in Section 7.2.2.2 reports the increase
of indoor PM of various size ranges from household activities, such as walking into and out of a
room. The tendency for such human activity in the home or at work to generate a "personal
activity cloud" of particles from clothing and other items (stuffed furniture, carpet, etc.), that
will be intense in the breathing zone and diluted near an area monitor located several meters
away, has also been cited as a contributing factor to the discrepancy between personal measures
of exposure and time-weighted-average (TWA) exposures using microenvironmental
measurements (Martinelli et al., 1983; Cohen et al., 1984; Rodes et al., 1991). Fletcher and
Johnson (1988) also measured metal concentrations (measurement method and size unspecified)
7-82
-------
in an occupational exposure situation (metal spraying of spindles on a lathe) and found 50%
higher concentrations measured from the left lapel compared to the right lapel, which reflected
the orientation of the operator to the lathe.
7.4 NEW LITERATURE ON PARTICLE EXPOSURES SINCE 1981
The following sections review studies that measured PEM PM in the general non-smoking
population. In these studies, the subjects spent time at home and in other indoor environments
that include time at work. In the USA, recent data indicate that on a daily basis, an average US
resident spends approximately 21 h indoors (85.6%), 100 minutes in (or near) a vehicle (7.2%),
and 100 minutes outdoors (7.2%) (U.S. Environmental Protection Agency, 1989).
Almost all the studies of PM exposure in the general public have been conducted on urban
and suburban residents. These subjects are often working in occupations that do not require PM
monitoring to assure that occupational standards are being met (e.g. in an office). However, PM
monitoring in an industrial workplace by a subject - independently of an official corporate
industrial hygiene program - can have legal or security implications for an employer. A further
complication arises from the fact that industrial exposures tend to be dominated by a specific
type of particle. Coal miners are exposed to coal dust, textile workers are exposed to cotton
dust, etc.
7.4.1 Personal Exposures in U.S. Studies
Dockery and Spengler (198 Ib) compared personal PM3 5 exposures and ambient PM3 5
concentrations in Watertown, MA, and in Steubenville, OH. In Watertown, 24-h personal
samples were collected on a 1-in 6-day schedule, and in Steubenville, 12-h personal samples (8
a.m. to 8 p.m.) were collected on a Monday-Wednesday-Friday schedule. A correlation
coefficient of 0.692 between the mean personal and the mean ambient concentration for
37 subjects, 18 in Watertown and 19 in Steubenville, was reported for the pooled data.
However, this appears to be an artifact of two separate clusters formed by these data, each with
considerably lower correlation. When these data are analyzed separately, the regression
coefficient between personal and ambient for Watertown is R2 = 0.00 and for Steubenville it is
R2 = 0.18.
7-83
-------
Sexton et al. (1984) studied personal exposures to respirable particles (PM3 5) for
48 nonsmokers during a winter period in Waterbury, VT, where firewood was either the primary
or secondary heating source for the subject. Their results showed that personal exposures were
45% higher than indoor averages (36 //g/m3 versus 25 //g/m3) and indoor averages were 45%
higher than outdoor averages (25 //g/m3 versus 17 //g/m3). Ambient air pollution, measured by
an identical stationary ambient monitor (SAM) outside each residence (a pump contained in a
heated box was connected to an external cyclone and filter), had no correlation with the
residents' personal exposures (R2 = 0.00) and 95% of the subjects had personal exposures greater
than the median outdoor concentration.
Spengler et al. (1985) reported a study of PM3 5 exposures in the non-industrial cities of
Kingston and Harriman, TN, during the winter months of February through March, 1981. In
this study, two Harvard/EPRI PM3 5 monitors were used for each person. One stationary indoor
monitor (SIM) remained indoors in the home, and the second monitor (PEM) was carried for 24-
h to obtain the personal exposure. In each community, identical Harvard/EPRI samplers (SAM)
were placed at a central site to represent ambient PM3 5 concentrations. The results of the study
are shown in Table 7-19. In both communities, 95% of the subjects had personal exposures to
PM3 5 greater than the average ambient concentrations. The mean personal exposure and indoor
concentrations (44 ± 3 //g/m3 and 42 ± 3 //g/m3) were more than 100% greater than the mean
ambient average of 18 ± 2 //g/m3 sampled on the same days.
For the complete cohort, the correlation between PM PEM and PM SAM was r = 0.07 (p =
0.30), and between PM PEM and PM SIM was r = 0.70 (p = 0.0001). The correlation between
simultaneous PM PEM and PM SAM was r = 0.15 for 162 nonsmoke exposed individual
observations (p = 0.06). For 63 observations on smoke exposed individuals, the correlation r =
0.16 was not significant (p = 0.16) between PM PEM and PM SAM. An important finding was
that in nonsmoking households, the PM PEM is always higher than SIM and SAM. "This
implies that individuals encounter elevated concentrations away from home and/or that home
concentrations are elevated while they are at home and reduced while they are away". This
observation is supported by the findings of Thatcher and
7-84
-------
TABLE 7-19. QUANTILE DESCRIPTION OF PERSONAL, INDOOR,
AND OUTDOOR PM3 5 CONCENTRATIONS (^g/m3),
BY LOCATION IN TWO TENNESSEE COMMUNITIES
City
Kingston
Harriman
Total3
Group
Personal
Indoor
Outdoor
Personal
Indoor
Outdoor
Personal
Indoor
Outdoor
N
133
138
40
93
106
21
249
266
71
95%
99
110
28
122
129
34
113
119
33
75%
47
47
22
54
45
23
48
46
23
50%
34
31
16
35
27
15
34
29
17
25%
26
20
12
24
18
13
26
20
13
5%
19
10
6
15
10
9
17
10
7
Mean
42
42
17
47
42
18
44
42
18
S.E.
2.5
3.5
2.7
4.8
4.1
4.0
2.8
2.6
2.1
includes samples from 13 subjects living outside Kingston and Harriman town limits and from four field
personnel residing in these communities.
N = number of samples.
S.E. = Standard error.
Source: Spengler et al. (1985).
Layton (1995), reported in Section 7.2.2.2: merely walking into a room can raise the
concentrations of PM by 100%. This study is relevant to the analyses by Dockery et al. (1992)
of PM mortality in St. Louis, MO, and in Eastern Tennessee counties surrounding Kingston and
Harriman as discussed in Chapter 12. Although the Spengler et al. (1985) and Dockery et al.
(1992) studies are not directly comparable, because different years of data were used (1981
versus 1985/1986), the authors' assumption in Dockery et al. (1992) that the Harriman, TN, data
represent exposures to PM in all of eastern Tennessee is called into question.
Morandi et al. (1988) investigated the relationship between personal exposures to PM and
indoor and outdoor PM concentrations, using a TSI Model 3500 piezobalance that measures
respirable particles in the range <3.5 jim. For the group of 30 asthmatics in Houston, TX, that
were studied, outdoor concentrations averaged 22 //g/m3, indoor concentrations averaged 22%
higher than outdoor (27 //g/m3) and, in motor vehicles, the average concentration of particles
was 60% higher than the average outdoors (35//g/m3). Personal 12-h (7 a.m. to 7 p.m.) daytime
exposures to PM were not predicted as well by fixed site dichotomous sampler ambient monitors
7-85
-------
(R2 = 0.34) as by the indoor exposures (R2 = 0.57). However, for 1-h exposures, they found no
correlation (R2 = 0.00) between the personal exposures to PM5 and the indoor exposures
measured with a TSI model 5000 stationary continuous piezobalance located in the "den" area of
the home. The authors noted that use of home air conditioning and recirculation tended to
increase the PM exposures.
Lioy et al. (1990) reported a study done during the winter (January 1988) in the industrial
community of Phillipsburg, NJ, where personal PM10 was monitored along with indoor and
outdoor PM10. They collected PM10 (fine plus coarse particles on a single filter). In this study of
eight residences of 14 nonsmoking individuals not smoke exposed at home, geometric mean 24-
h concentrations were 68, 48 and 42 //g/m3 for personal, outdoor and indoor sites, respectively.
The arithmetic mean personal PM exposure of 86 //g/m3 was 45% higher than the mean ambient
concentration of 60 //g/m3. The higher ambient than indoor concentrations in this study, a
reversal of the relationships found in the Sexton et al. (1984), Spengler et al. (1985) and
Morandi et al. (1988) studies, may be caused by the local industrial source of coarse particles in
that community and the absence of cigarette smokers in the residences sampled. This difference
also may be partially explained by the 10 jim particle sizes sampled in the NJ study and the 3.5
|im particle sizes in the other studies. The regression coefficient between personal and ambient
PM10 for all 14 people on the 14 days of the study (n = 191 valid personal values) was 0.19 (R2 =
0.037, p = 0.008). With three personal exposure extreme values removed (n = 188 personal
values) and without correction for missing data, the coefficient was 0.50 (R2 = 0.25, p = 0.007).
Lioy et al. (1990) report individual regression equations of PEM and SAM for the six of 14
subjects with significant relationships (p < 0.01). These data are shown in Table 7-20. For
individuals with constant daily activities in the same microenvironments, the increment of PM
exposure due to nonambient sources is repeatable with lower variability than that of the ambient
PM. Therefore their variation of personal exposure from day-to-day is highly driven by the
variation of the ambient PM. For subjects with intermittent exposures to nonambient PM,
through non-repetitive activity patterns or intermittent source operation, the regression of PEM
on SAM can become non-significant. This improvement in
7-86
-------
TABLE 7-20. REGRESSION EQUATION OF THOSE INDIVIDUALS
HAVING STATISTICALLY SIGNIFICANT RELATIONSHIPS OF
EXPOSURE (PEM) WITH OUTDOOR AIR CONCENTRATIONS (SAM)
Participant
01
31
52
62
81
91
Equation
y = 0.62 (0.12) X + 26.5 (17.3)
y = 0.55 (0.07) X + 7.3 (9.9)
y = 0.63 (0.11)X+ 15.3 (14.7)
y= 1.29 (0.27) X + 33.0 (37.1)
y= 1.07 (0.24) X + 39.0 (32.6)
y = 0.59 (0.12) X + 42.0 (19.9)
R2
0.66
0.83
0.74
0.67
0.63
0.63
N
14
14
14
13
14
13
P
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
y = Personal air PM-10.
X = Outdoor air PM-10.
() = Confidence interval.
Source: Lioy et al. (1990).
correlation was also shown for their indoor versus outdoor relationships, between cross-sectional
and individual comparisons, as described in Section 7.4.2.3.
In all these studies, the personal PM was measured to be higher than either the indoor or
the outdoor PM measurements. This relationship of PEM > SIM and PEM > SAM has also been
found in the PTEAM study (Clayton et al., 1993) described in detail in Section 7.2.2.1.3 and
later in Section 7.4.1.1. For the PTEAM study during the day (7 a.m. to 7 p.m.) average
personal PM10 exposure data (150 //g/m3) were 57% higher than the average indoor and outdoor
concentrations, which were virtually equal (95 //g/m3). Consequently, a time-weighted-average
(TWA) of the daytime indoor and outdoor PM concentrations appears to always underestimate
the personal exposures to PM because the daytime PEM data are higher than either the SIM or
SAM data. At night (7 p.m. to 7 a.m.) average PM10 personal exposures (77 //g/m3) were higher
than the average indoor concentrations (63 //g/m3) but lower than the average outdoor
concentration (86 //g/m3).
It has been proposed (World Health Organization, 1982a; Spengler et al., 1985; Mage,
1985) that such a discrepancy between the TWA and the personal monitoring measurements may
be caused by two factors described as follows: (1) human exposure to PM at work and in traffic
are only partially accounted for in a TWA of indoor and outdoor ambient PM values; and (2)
7-87
-------
indoor and outdoor averages reflect periods of low concentration during which the subject is not
present. The PM pollution generating activities in a home usually occur only when a person is at
home, as discussed in Section 7.1.2 concerning Equation 7-2. Therefore, the PM in a home will
be higher when a person is present than when the home is unoccupied. A 24-h average of the
indoor concentration thereby underestimates the average exposure of a person while in that
home.
Ambient PM is also higher during the day (when industry and traffic are active, and wind
speeds are high) than at night when PM generating activities are at a minimum and the air is still
(Miller and Thompson, 1970). Consequently, a 24-h average ambient PM value generally
underpredicts the concentrations during the daylight hours and the exposures of people going
outdoors during that period.
7.4.1.1 The Particle Total Exposure Assessment Methodology Study
In 1986, the U. S. Congress mandated that EPA's Office of Research and Development
"carry out a TEAM Study of human exposure to particles." The main goal of the study was to
estimate the frequency distribution of exposures to particles for nonsmoking Riverside, CA,
residents. Another goal was to determine particle concentrations in the participants' homes and
immediately outside the homes. The detailed analyses of the indoor PM and outdoor PM data
were described in Section 7.2.2.1.3.
7.4.1.1.1 Pilot Study
Study Design
A prepilot study, described in Section 7.2.2.1.3, was undertaken in nine homes in Azusa,
CA in March of 1989 to test the sampling equipment (Ozkaynak et al., 1990). Newly-designed
personal exposure monitors (PEMs) were equipped with thoracic (PM10) and fine (PM2 5) particle
inlets. The PEMs were impactors with 4-Lpm Casella pumps (Wiener, 1988). Two persons in
each household wore the PEMs for two consecutive 12-h periods (night and day). Each day they
alternated inlet nozzles. A central site with a PEM, a microenvironmental monitor (MEM), and
two EPA reference methods (dichotomous and high-volume samplers) with a 10 jim size-
selective inlet was also operated throughout the 11 days (22 12-h periods) of the study.
-------
Results
The personal exposure levels were about twice as great as the indoor or outdoor
concentrations for both PM10 (Table 7-2la) and PM2 5 (Table 7-2Ib). Considerable effort was
expended to demonstrate that this was not a sampling artifact, due for example to the constant
motion of the sampler; however, no evidence could be found for an artifactual effect.
Nonetheless, to reduce chances for an artifactual finding in the main study, it was decided to use
identical PEMs for both the personal and fixed (indoor and outdoor) samples in the main study.
Cross-sectional personal exposures were essentially uncorrelated (slightly negatively) with
outdoor concentrations (R2 = 0 to 2%) (Ozkaynak et al., 1993a). However, a serial correlation
analysis of these pilot PTEAM data were performed for the six or eight 12-h averages that
comprised the three or four 24-h averages reported for the residents of the first five homes in
Table 7-21a,b. The residents of four homes only carried the PEM for two days, so the four 12-h
individual measurements were too few for development of a meaningful serial relationship. The
results for the ten people in homes 1 to 5 are shown in Table 7-22. The medians of R2 equal
0.12 for PEM PM2 5 vs SAM PM2 5 and 0.07 for PEM PM10 vs SAM PM10, neither of which is
significant. More importantly, the serial slopes were positive for 15 of the 20 cases which is the
expected behavior, as opposed to the counter-intuitive negative correlation found for the pooled
PEM vs SAM data for all residents of the nine homes.
In Azusa, the excess PM2 5 and PM10 generated by personal activities increased the personal
exposures by approximately 100% above the average of the indoor and outdoor values. These
results are in marked contrast to the data of Tamura and Ando (1994) and Tamura et al. (1996)
in which seven Japanese elderly housewives and male retirees had PM10 PEM exposures less
than the time weighted average of SIM and SAM PM10 concentrations.
7.4.1.1.2 Main Study
Study Design
Ultimately 178 residents of Riverside, CA took part in the study in the fall of 1990.
Respondents represented 139,000 ± 16,000 (S.E.) nonsmoking Riverside residents aged 10 and
above. Their homes represented about 60,000 Riverside homes. Each participant wore the PEM
for two consecutive 12-h periods. Concurrent PM10 and PM2 5 samples were
7-89
-------
TABLE 7-21a. PARTICLE TOTAL EXPOSURE ASSESSMENT METHODOLOGY
PREPILOT STUDY: 24-HOUR PM,n CONCENTRATIONS
,n
House
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
5
5
5
6
6
7
7
8
8
9
9
Mean
SD
SE
Day
1
O
5
7
1
O
5
7
1
O
5
7
2
4
6
2
4
6
8
10
9
11
9
11
8
10
Person 1
102
142
158
92
109
99
131
62
98
100
143
76
109
90
99
80
70
80
130
150
209
80
135
97
136
273
117.2
44.9
8.8
Person 2
86
125
150
127
158
140
87
56
107
141
132
103
92
77
122
104
77
78
152
102
126
71
178
151
102
91
112.9
30.8
6.0
Indoors
54
38
49
34
122
37
41
32
86
39
71
36
77
34
36
76
62
54
114
106
46
29
73
38
63
121
60.3
28.5
5.6
Outdoors
132
49
70
49
112
48
70
46
115
45
79
44
102
47
37
99
65
50
39
51
72
39
59
28
43
48
63.0
27.1
5.3
Source: Data from PTEAM Prepilot Study used to calculate R2 values as shown in Table 7-22 and published by
Wallace (1996).
7-90
-------
TABLE 7-21b. PARTICLE TOTAL EXPOSURE ASSESSMENT METHODOLOGY
PREPILOT STUDY: 24-H PM?. CONCENTRATIONS
2,
House
1
1
1
2
2
2
3
3
3
4
4
4
4
5
5
5
5
6
6
7
7
8
8
9
9
Mean
SD
SE
Day
2
4
6
2
4
6
2
4
6
1
O
5
7
1
O
5
7
9
11
8
10
8
10
9
11
Person 1
44
55
55
58
46
51
53
62
109
75
46
118
40
65
59
40
34
71
77
64
111
53
110
178
105
71.2
32.7
6.5
Person 2
96
88
382
53
100
50
66
94
88
61
43
94
40
69
70
56
53
81
75
135
67
100
1453*
48
58
140.8*
275.5
55.1
Indoors
22
25
21
31
27
28
48
30
39
33
19
31
17
62
35
42
25
56
53
17
32
27
35
70
42
34.7
13.7
2.7
Outdoors
67
39
33
52
43
40
58
35
39
71
29
46
26
96
38
55
28
33
18
27
35
27
35
40
28
41.6
16.8
3.4
* Horseback riding at an indoor ring. If this point is deleted, mean = 86.1.
Source: Data from PTEAM Prepilot Study used to calculate R2 values as shown in Table 7-22 and published by
Wallace (1996).
7-91
-------
TABLE 7-22. REGRESSIONS OF PERSONAL EXPOSURE ON INDOOR AND
OUTDOOR PM10 AND PM2 5 CONCENTRATIONS: PARTICULE TOTAL EXPOSURE
ASSESSMENT METHODOLOGY PREPILOT STUDY
House
PM10:
1
2
3
4
5
PM?V
1
2
3
4
5
Person
Personal vs.
1
2
1
2
1
2
1
2
1
2
Personal vs.
1
2
1
2
1
2
1
2
1
2
N Intercept
Outdoor
8
8
8
8
8
8
6
6
6
6
Outdoor
6
6
6
6
6
6
8
8
8
8
124
134
47
26
83
116
87
106
47
22
41
274
8.8
47
87
40
40
45
27
46
SE
42
60
44
52
47
54
20
28
31
26
20
266
20
34
58
54
24
22
15
16
P
0.03
NS
NS
NS
NS
NS
0.01
0.02
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
0.03
Slope
-0.0004
-0.16
0.77
1.22
0.3
0.07
0.2
-0.15
0.42
0.9
0.22
-1.8
0.96
0.47
-0.29
0.97
0.7
0.34
0.42
0.3
SE
0.51
0.73
0.58
0.68
0.61
0.7
0.29
0.4
0.41
0.35
0.4
5.3
0.41
0.7
1.25
1.2
0.48
0.45
0.24
0.27
P
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
R2
0
0.01
0.23
0.35
0.04
0.002
0.1
0.03
0.2
0.63
0.07
0.03
0.58
0.1
0.01
0.15
0.26
0.09
0.34
0.17
NS = not significant (p > 0.05).
N = Number of 12-h observations.
Source: Wallace (1996).
collected by the stationary indoor monitor (SIM) and stationary ambient monitor (SAM) at each
home. A total often particle samples were collected for each household (day and night samples
from the PEM10, SIM10, SIM25, SAM10, and SAM25). Air exchange rates were also determined
for each 12-h period. Participants were asked to note activities that might involve exposures to
increased particle levels. Following each of the two 12-h monitoring periods, they answered an
7-92
-------
interviewer-administered questionnaire concerning their activities and locations during that time.
A central outdoor site was maintained over the entire period (September 22, 1990 through
November 9, 1990). The site had two high-volume samplers (Wedding & Assoc.) with 10-|im
inlets (actual cutpoint about 9.0 |im), two dichotomous PM10 and PM2 5 samplers (Sierra-
Andersen) (actual cutpoint about 9.5 |im), one PEM, one PM10 SAM, and one PM2 5 SAM.
Results
Of 632 permanent residences contacted, 443 (70%) completed the screening interview. Of
these, 257 were asked to participate and 178 (69%) agreed.
Quality of the Data
More than 2,750 particle samples were collected, about 96% of those attempted. All filters
were analyzed by X-ray fluorescence (XRF) for a suite of 40 metals. More than 1,000 12-h
average air exchange rate measurements were made. A complete discussion of the quality of the
data is found in Pellizzari et al. (1993) and in Thomas et al. (1993).
Concentrations
Concentrations of particles and target elements have been reported (Clayton et al., 1993;
Ozkaynak et al., 1993a; Pellizzari et al., 1993; Wallace et al., 1993). Population-weighted
daytime personal PM10 concentrations averaged about 150 //g/m3, compared to concurrent indoor
and outdoor mean concentrations of about 95 //g/m3 (Table 7-23). The overnight personal PM10
mean was much lower (77 //g/m3) and more similar to the indoor (63 //g/m3) and outdoor
(86 //g/m3) means. About 25% of the population was estimated to have exceeded the 24-h
National Ambient Air Quality Standard for PM10 of 150 //g/m3. Over 90% of the population
exceeded the 24-h California Ambient Air Quality Standard of 50 //g/m3.
Correlations
The central site appeared to be a moderately good estimator of outdoor particle
concentrations throughout the city. Spearman correlations of the central-site concentrations
7-93
-------
TABLE 7-23. POPULATION-WEIGHTED3 CONCENTRATIONS AND
STANDARD ERRORS (uz/m3) PTEAM STUDY
Sample type
Daytime PM10
Personal
Indoor
Outdoor
Overnight PM10
Personal
Indoor
Outdoor
Daytime PM2 5
Indoor
Outdoor
Overnight PM2 5
Indoor
Outdoor
N
171
169
165
168
163
162
173
167
166
161
Geom.
Mean
129
78
83
68
53
74
35
38
27
37
GSD
1.75
1.88
1.68
1.64
1.78
1.74
2.25
2.07
2.21
2.23
Arith.
Mean ± SE
150±9
95 ±6
94 ±6
77 ±4
63 ±3
87 ±4
48 ±4
49 ±3
36 ±2
51±4
Percentile
90% ± SE
260 ± 12
180± 11
160 ± 7
140 ± 10
120 ± 5
170 ± 5
100 ±7
100 ± 6
83 ±6
120 ± 5
98%
380
240
240
190
160
210
170
170
120
160
Tersonal samples weighted to represent nonsmoking population of 139,000 Riverside residents aged 10 or
above. Indoor-outdoor samples weighted to represent 61,500 homes with at least one nonsmoker aged 10 or
above.
Source: Pellizzari et al. (1993).
measured by all three methods (PEM-SAM, dichot, Wedding) with outdoor near-home
concentrations as measured by the SAMs ranged from 0.8 to 0.85 (p<0.00001). Linear
regressions indicated that the central-site 12-h readings could explain 57% of the variance
observed in the near-home 12-h outdoor concentrations (Figure 7-20).
Outdoor 12-h concentrations of PM10 could explain about 25 to 30% of the variance
observed in indoor concentrations of PM10, but only about 16% of the variance in 12-h personal
exposures to PM10 (Figure 7-21). This is understandable in view of the importance of indoor
activities such as smoking, cooking, dusting, and vacuuming on exposures to
7-94
-------
600
™ 500
2
2 400
o
E
•o
re
o
re
.a
<
CO
300
200
100
Backyard = 1.03*Central + 17.6
FT = 0.57 N = 323
50 100 150 200
Central site reference monitor mean
250
Figure 7-20. Residential outdoor monitors versus central-site mean of two dichotomous
samplers in Riverside, CA. R2 = 57%.
Source of Data: Pellizzari et al. (1993).
500
^ 400
(0
300
o
Q.
X
± 200
re
c
o
2
a>
a.
100
Pers = 0.54*Out + 62
R2= 16% N = 312
0 100 200 300 400 500
Backyard concentrations (\igftn )
Figure 7-21. Personal exposures versus residential (back yard) outdoor PM10
concentrations in Riverside, CA. R2 = 16%.
Source of Data: Pellizzari et al. (1993).
600
7-95
-------
particles. The higher daytime exposures were even less well represented by the outdoor
concentrations.
Indoor concentrations accounted for about half of the variance in personal exposures.
However, neither the indoor concentrations alone, nor the outdoor concentrations alone, nor
time-weighted averages of indoor and outdoor concentrations could do more than explain about
two-thirds of the observed variance in personal exposures. The remaining portion of personal
exposure is assumed to arise from personal activities or unmeasured microenvironments that are
not well represented by fixed indoor or outdoor monitors.
Discussion
The more than 50% increase in daytime personal exposures compared to concurrent
indoor or outdoor concentrations suggested that personal activities were important determinants
of exposure. However, the nature of this "personal cloud" of particles has not yet been
determined. An approach to the composition of the personal cloud is elemental analysis, using
X-ray fluorescence. Analysis of all personal and indoor filters showed that 14 of 15 elements
were elevated by values of 50 to 100% in the personal filters compared to the indoor filters
(Figure 7-22). This observation suggests that a component of the personal cloud is an aerosol of
the same general composition as the indoor aerosol. This could be particles created by activities
(e.g., cooking) or re-entrained household dust from motion (walking across carpets or sitting on
upholstered furniture; Thatcher and Layton, 1995). House dust is a mixture of airborne outdoor
PM (primarily coarse mode), tracked-in soil and road dust, and PM produced by indoor sources.
As such, it should contain crustal elements from soil, lead and bromine from automobiles, and
other elements from combustion sources. This would be consistent with the observation that
nearly all elements were elevated in personal samples. The lack of elevated values for sulfur
may be due to the fact that submicron particles are not resuspended by human activity (Thatcher
and Layton, 1995). The personal overnight samples that showed smaller mass increases than the
personal daytime samples are also consistent with the fact that the participants were sleeping for
much of the 12-h overnight monitoring period and were thus not engaging in these particle-
generating or reentraining activities.
7-96
-------
20
40 60 80 100
Percent increase in personal cloud
120
Figure 7-22. Increased concentrations of elements in the personal versus the indoor
samples.
Source: Ozkaynak et al. (1996).
A source apportionment of the personal PM10 mass during the daytime period is shown on
Figure 7-23 (Ozkaynak et al., 1996). This chart is derived by subtracting the average SIM and
SAM (95 |ig/m3) from the mean PEM (150 |ig/m3) given on Table 7-23. The 55 |ig/m3
difference is shown as the 37% fraction of the total of 150 |ig/m3 labelled Personal 37%. The
source of this "personal cloud" is indeterminable from the SIM, SAM and PEM data. As
discussed previously, it is likely to consist primarily of resuspended dust that would have a
composition of a mixture of all the other sources. The 15% other-indoor PM represents the
indoor mass that could not be assigned to ETS, cooking or ambient PM. It is likely that the 52%
of other-indoor plus personal-cloud categories contains an appreciable amount of ambient PM
that came indoors over a long period of time and is resuspended by activity. If so, then the PEM
would be about 50% of ambient origin.
7.4.2 Personal Exposures in International Studies
As part of World Health Organization/United Nations Environment Programme
(WHO/UNEP) Global Environment Monitoring System (GEMS) activities, four pilot studies
7-97
-------
Outdoor
42%
Personal
37%
Smoking
3%
N = 166 Samples
Cooking
3%
Other Indoor
15%
Figure 7-23. Source apportionment of PTEAM PM10 Personal Monitoring (PEM) Data.
"Other Indoor" represents PM found by the indoor monitor (SIM), for
which the source is unknown. "Personal" represents the excess PM
captured by the PEM that cannot be attributed to either indoor (SIM) or
outdoor (SAM).
Source: Clayton et al. (1993).
of personal exposure to PM were conducted in: Zagreb (World Health Organization, 1982a);
Toronto (World Health Organization, 1982b); Bombay (World Health Organization, 1984); and
Beijing (World Health Organization, 1985). In these studies, people who worked in the
participating scientific institutes were recruited to carry a PM sampler, and their exposures were
matched to the ambient concentrations measured outside their home or at a central station in
their communities. The results of these studies, expressed as mean personal exposure (PEM)
and mean ambient (SAM) concentration, and the cross-sectional regression R2 between them are
presented in Table 7-24.
The net result of these four international studies is that they appear to confirm the lack of a
consistent cross-sectional relationship between individual personal PM exposures and ambient
concentrations as found in the U.S. studies described in Section 7.4.1.
7-98
-------
TABLE 7-24. SUMMARY OF WHO/UNEP GLOBAL ENVIRONMENT MONITORING
SYSTEM/PERSONAL EXPOSURE PILOT STUDY RESULTS
Location
Season
Toronto
Winter
Summer
Zagreb
Summer
Winter
Bombay
Winter
Summer
monsoon
Beijing
Winter
Summer
PM Size
Cut (urn)
25*
5
3.5
3.5
N
13
12
15
20
m
72
78
12
12
105
102
101
71
40
Time
8-h
1-wk
24-h
24-h
1-wk
PEM
Mean ± SE
122±9
124±4
114±?
187±?
127±6
67±3
58±3
177±?
66±?
SAM
Mean ± SE
68±9
78±4
55±?
193±?
117±5
65±3
51±2
421±?
192±?
R2 PEM vs.
SAM
0.15
0.10
0.00
0.50
0.26
0.20
0.02
0.07
0.03
P
NS
NS
NS
NR
NR
NR
NS
NS
NS
N = number of subjects carrying personal exposure monitor (PEM).
m = total number of observations.
PEM = mean ± SD of PM concentrations (in ,wg/m3) from personal exposure monitors.
SAM = mean ± SD of PM concentrations (in ,ug/m3) from stationary ambient monitors.
NR = Not Reported, but listed as significant.
NS = Not significantly different from 0.
? = Not Reported.
*25 //m AD computed from flow rate and open filter design.
Source: World Health Organization (1982a,b, 1984, 1985).
7.4.2.1 Personal Exposures in Tokyo (Itabashi Ward), Japan
Tamura and Ando (1994), National Institute for Environmental Studies (1994) and Tamura
et al. (1996) report results of a PM personal monitoring study conducted during 1992 in Tokyo.
Seven elderly non-smoke exposed individuals who lived in traditional Japanese homes with
"tatami" reed mat or carpeting on tatami or wooden flooring, and cooked with city gas, carried a
PEM cascade impactor with cut-points of 2 jim and 10 jim (Sibata Science Technology, Ltd.).
The seven individuals lived near the Itabashi monitoring station close to a main road. Indoor
PM (SIM) and outdoor PM (SAM) were measured simultaneously for 11 48-h periods
distributed in all four seasons of the year. The dataset was screened to remove observations that
included indoor combustion source exposures, such as ETS from visitors, and burning of incense
or mosquito coils. The reported findings were as follows:
7-99
-------
1. The cross sectional correlation coefficient of SIM vs SAM was "relatively high" (r2 =
0.72), but the individual coefficients for each house were higher as shown in Figure 7-
24.
2. The cross sectional correlation coefficient of PEM vs SAM (measured under the eaves
of the subject's house) was "relatively high" (r2 = 0.70), but the individual coefficients
for most of the subjects were higher as shown in Table 7-25.
3. The cross sectional correlation coefficient of PEM vs PM measured at the Itabashi
monitoring station was slightly lower than that for the outside air (r2 = 0.68), as shown
in Figure 7-25, and the individual coefficients for most of the subjects were higher as
shown in Table 7-25.
4. The individual SAM values were all linearly related with the central monitor at the
Itabashi station with the coefficient of regression (R2) in the range between 0.70 and
0.94.
5. The individual PEM values varied from 30% to 50% of the SAM values. These {PEM
< SAM} data are quite different from the US data sets, such as PTEAM, where {PEM
> SAM}, because they were designed to measure the influence of the outdoors on
personal exposures. The difference may be due to the exclusion of ETS exposure and
incense/mosquito coil burning and the Japanese customs of using reed mat (tatami)
flooring and taking shoes off when entering a home. These factors would all tend to
reduce the generation and resuspension of PM in the home.
Tamura and Ando (1994) and Tamura et al. (1996) confirm the findings of Thatcher
and Layton (1995) that PM < 5 jim AD has negligible resuspension in homes. Their
SIM PM2 and SIM (PM10 - PM2) were highly correlated with the SAM of identical size
(r = 0.879 and 0.839 respectively) but there was a negative correlation between the
SIM and SAM (TSP - PM10) fraction (r = - 0.036).
The importance of this study is that it demonstrates that there are very strong correlations
between PEM and SAM (0.747 < r < 0.964) when the masking influences of indoor combustion
sources are removed and resuspension of PM is minimized. This provides strong support to the
use of an ambient monitoring station to represent the exposure of people in the community to
PM of ambient origin.
7.4.2.2 Personal Exposures in the Netherlands
Janssen et al. (1995) preliminarily reported in an abstract results of personal PM
monitoring conducted during 1994 in Amsterdam and Wageningen, NL as part of a doctoral
study. Participants were 13 non-smoking adults (age 50 to 70) in Amsterdam (urban) with
7-100
-------
110-
100-
— 90-
"£ 80-
270-
tTeo-
o
o 50-
1 40-
~ 30-
20-
10-
E
B m
D ffl °
D
D ^
m n
ra D m H
m m
ffl D
tyj!™ ° r = 0.922
™ ffl Winter r = 0.920
Summer r = 0.961
i i i i i i i i i
40
80
120
160
izu —
110-
100-
"90-
£ 80-
370-
S 60-
550-
B 40-
30-
20-
10-
F
n n
it
• n
BE • B
Q Q B B
JlE ^ r = 0.897
ff Winter r = 0.702
Summer r = 0.970
i i i i i i i i i
0 40 80 120 160 2C
izu —
110-
100-
-90-
z™~-
S60-
o 50-
= 40-
30-
20-
10-
G
B
D B
a » n
!!b I^
c^Sp0 r=0.898
Q D1 ffl Winter r=0.879
Summer r=0.919
i i i i i i i i i
40
80 120
Outdoor (ug/m )
160
200
110-
100-
_ 90-
-~ 80-
o 70-
3 60-
o 50-
o
•o 40-
- 30-
20-
10-
A
9 m
m
a °
iB°
IB
_jjff r = 0.983
H,H Winter r = 0.980
Summer r = 0.982
i i i i i i i i i
40
80
120
160
200 -
40
80
120
160
40
80
120
160
40
80 120
Outdoor (ug/rrl ;
160
200
110-
100-
-90-
-E 80-
ra 70-
3 60-
o 50-
•o 40-
- 30-
20-
10-
B
H a
" -W ™ Winter r = 0.966
Summer r = 0.877
200
izu —
110-
100-
_ 90-
-£ 80-
o 70-
r eo-
§ 50-
1 40-
~ 30-
20-
10-
C
B
m o
Dn«Dffi
OB
nlffl m
a iftf" r = 0.970
n Winter r = 0.968
Summer r = 0.964
i i i i i i i i i
200
110-
100-
_ 90-
•E 80-
o. 70-
~ 60-
o 50-
•o 40-
-30-
20-
10-
o —
D
B B
ffl ffl
B
B B D
a<^n *
8JB* r = 0.838
° D Winter r = 0.775
Summer r = 0.978
i i i i i i i i i
200
Figure 7-24. The relationship between PM10 in outdoor air and indoor air at each house
in the study. A, B, C, D, E, F, and G, refer to the individuals identified later
in Tables 7-29 and 7-30.
Source: Tamura and Ando (1994); Tamura et al, (1996).
7-101
-------
TABLE 7-25. SUMMARY OF CORRELATIONS BETWEEN PM,n PERSONAL
10
EXPOSURES OF 7 TOKYO RESIDENTS AND THE PM,n MEASURED OUTDOORS
10
UNDER THE EAVES OF THEIR HOMES, AND THE PM MEASURED AT THE
ITABASHI MONITORING STATION
Correlation between Correlation between Personal
Number of Samples Personal and Outdoor at and Itabashi Station (r)
Subject ID 48-hPM10 home (r)
A
B
C
D
E
F
G
A-
9 0.958
9 0.874
11 0.846
9 0.922
10 0.960
7 0.776
9 0.961
G 64 0.834
0.876
0.747
0.848
0.964
0.925
0.801
0.952
0.830
Source: Tamura et al. (1996).
~
«E
~a>
_3.
C
Q
M
(^
•o
n
^
o
0
•o
+J
O
130-
120-
110-
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
—
y = 1.07 x- 0.4 (R = 0.901)
DD ....--
_ n n.Q-'"
n P\'
D DD ..,---"ffl |
I l | I. . >*i*f~i ~"~ —
nn •••*' ^ H + —•— """
»n i^i--"$'""'"
.••'
D .-••''
a ...---"
R--%' '
i~~i >*F"*i
--••'i
__,^.
• I —
*---'"'"'"'
T^, — • "^
•—""j. T
"*"
"'j + * y = 0.46 x + 11.4 (R = 0.825)
^^••^'^ ' '
— * ,.***
,.•''
I I I I I I I I
0 20 40 60 80
1 1 1 1 1 1
100 120 14
Itabashi Monitoring Station
Figure 7-25. Correlations between PM10 at the Itabashi monitoring station and PM10 in
outdoor and personal exposure (D=outdoor; +=personal).
Source: Tamura and Ando (1994); Tamura et al. (1996).
7-102
-------
no occupational exposure to PM, and 15 children (age 10 to 12) in Wageningen (rural) who are
presumably non-smokers. Four to eight measurements were obtained for each subject which
allowed for correlating PEM and SAM within individuals (longitudinally). Only the median
individual regressions were reported, as follows: adults, PEM = 26 + 0.70 SAM, R = 0.57,
R2 = 0.32; and children, PEM = 78 + 0.43 SAM, R = 0.67, R2 = 0.44. For the children, parental
smoking explained 35% of the variance between PEM and SAM. For the adults, "living near a
busy road", time spent in traffic, and exposure to ETS explained 75% of the variance between
PEM and SAM. The authors interpreted their preliminary results to "suggest a reasonably high
correlation between personal and ambient PM10 within individuals". Janssen et al. (1995) also
note that the low correlations observed in most of the other studies reported in the literature were
cross-sectional (calculated on a group level), and were therefore mostly determined by the
variation between subjects (e.g., ETS exposed and non-ETS exposed subjects combined in the
same regression).
7.4.2.3 Reanalysis of Phillipsburg, NJ Data
With insight from the Jansen work, Wallace (1996) reanalyzed the complete Lioy et al.
(1990) data from Phillipsburg, NJ, as shown partially in Table 7-20 (see also Table 7-37).
Wallace (1996) compared the cross-sectional regressions of PEM on SAM for all the 14 subjects
on each of the 14 days sampled, to the longitudinal regressions of each of the 14 subjects on all
14 days sampled. He found that the median R2 (range) of the 14 individual (longitudinal)
regressions was 0.46 (0.02 to 0.82); and that for the 14 daily (cross-sectional) regressions was
0.06 (0.00 to 0.39). The difference appears to indicate that, although one household may have a
smoker and another not, the relationship of the indoor air in each home to the outdoor air may be
the same from day to day (i.e., consistently higher than ambient in the first case, but may be
consistently similar in the second). Because it provides a linkage between PEM and SAM, it
bears reiteration to make certain that it is clearly understood. This PEM vs SAM relationship
can be visually demonstrated with the following hypothetical example as shown on Figure 7-
26a,b.
• Let two people live next door to each other at a location where the ambient PM for 5
consecutive days has a sequence (1, 2, 3, 4, 5}.
• Let person A live without ETS exposure and have a corresponding PEM series (1,2,
3,4, 5},(R2=1).
7-103
-------
UJ
Q.
012345
SAM
15
10
UJ
Q.
(b)
0123
SAM
Figure 7-26. Example of difference between serial correlation (a) and cross-sectional
correlation (b) of PEM and SAM, showing how pooling of individuals
together can mask an underlying relationship of PEM and SAM.
• Let neighbor B live with ETS exposure and have a corresponding PEM series {11,
12, 13, 14, 15}, (R2= 1).
• When their PEM values are pooled so that they are analyzed together
(cross-sectionally) {(1,11), (2,12), (3,13), (4,14), (5,15)} vs the SAM set {1, 2, 3, 4,
5}, then R2 = 0.074.
• However, had the two PEM series been averaged each day, the sequence of averages
{6, 7, 8, 9, 10} would have a correlation of R2 = 1 with the same SAM sequence.
This averaging process is described later in more detail in Section 7.6.2.
The explanation by Janssen et al. (1995) for the low cross-sectional correlations of PM PEM
with PM SAM found in the literature and the new analyses reported by Tamura et al. (1996),
Jansen et al. (1995), and Wallace (1996) represent a major advance in our understanding of
contributions of ambient PM to personal exposures.
7-104
-------
7.4.2.4 Overview of Comparison of Personal Exposure to Ambient PM Concentrations
The PTEAM Study and the other key PEM studies discussed in this chapter so far are
summarized in Table 7-26. This table shows that many of the early studies reported no
statistically significant correlation between PEM and SAM. However, these early studies were
all characterized by a non-probability sample and a relatively small sample size. The PTEAM
study in Riverside which was a probability sample (Clayton et al., 1993) and the Lioy et al.
(1990) study in Phillipsburg, which was not a probability sample, have large sample sizes and
achieved significance. The other studies, such as World Health Organization (1982a,b) or
Morandi et al. (1988) are equivocal. In the following sections, PEM/SAM comparisons for
some PM constituents and two means of visualizing the complex relationships of PM measured
by a SAM and a PEM are developed.
7.4.3 Personal Exposures to Constituents of Particulate Matter
Suh et al. (1993) measured personal exposures to sulfate (SO4=) and acidity (H+), and
ambient and indoor concentrations in State College, PA, summer 1991. The correlations
between personal and ambient values of sulfate and acidity were R2 = 0.92 and 0.38 respectively,
which is in marked contrast to the R2 ~ 0 between earlier reported ambient PM and personal PM
studies (Table 7-26). This relationship is supported by Figure 7-22, indicating that personal
activities in the PTEAM study do not generate or resuspend sulfates less than 10 //m.
Figure 7-27 shows the consistent relation between ambient and personal sulfate
measurements (slope = 0.78 ± 0.02), and Figure 7-28 shows the improvement in prediction by
using the TWA with a correction factor (estimated personal sulfate = 0.885*TWA, R2 = 0.95
with slope = 0.96 ± 0.02). Personal acidity was also computed by the same equation with a
correction for personal ammonia (NH3) exposure that gave an R2 = 0.63. As opposed to PM
which has both indoor and outdoor sources, the sulfate and acidity are virtually all of outdoor
origin. Consequently, only the characteristics of the indoor environment, such as air
conditioning and ammonia sources, modify the personal exposures indoors.
7-105
-------
o
Oi
TABLE 7-26. COMPARISON OF PERSONAL EXPOSURE MONITOR (PEM) EXPOSURE OF
INDIVIDUALS TO THE SIMULTANEOUS AMBIENT PARTICULATE MATTER (SAM)
CONCENTRATION IN SEVERAL U.S. AND FOREIGN CITIES Qug/m3)
Reference
Binder et al.
Dockery and Spengler
Dockery and Spengler
Spengler et al.
World Health
Organization
Spengler et al.
World Health
Organization
Sexton et al.
World Health
Organization
World Health
Organization
Morandi et al.
Lioy et al.
Perritt et al.
Clayton et al.
Tamura et al.
Year
1976
1981b
1981b
1980
1982a
Winter
Summer
Winter
Summer
1985
1982b
Summer
Winter
1984
1984
Winter
Summer
Monsoon
1985
Winter
Summer
1988
1990
1991
1993
1996
Eocation
Ansonia
Watertown
Steubenville
Topeka
Toronto
Non-asthmatic
Non-asthmatic
Asthmatic
Asthmatic
Kingston/
Harriman
Zagreb
Waterbury
Bombay
Beijing
Houston
Phillipsburg
Azusa
Riverside
Tokyo
PMpm
5
3.5
3.5
3.5
25
3.5
5
3.5
3.5
3.5
3.5
10
2.5
10
10
10
N
20
18
19
46
13
13
13
13
97
12
48
15
20
30
14
14°
9
9
141
7
Time
24-h
24-h
12-h
12-h
8-h
8-h
8-h
8-h
24-h
1-wk
24-h
24-h
24-h
1-wk
12-h
24-h
24-h
24-h
24-h
24-h
48-h
Mean PEM
115
35
57
30
122
124
91
124
44
114
187
36
127
67
58
177
66
27
86
76
79
115
113
37
Mean SAM
59
17
64
13
68
78
54
80
18
55
193
17
117
65
51
421
192
16
60
60
43
62
84
56
R2 PEM vs SAM
NS
0.00
0.19
0.04
0.15
0.10
0.00
0.07
0.00
0.00
0.50
0.00
0.26
0.20
0.02
0.07
0.03
0.34
0.04
0.25
0.01
0.01
0.23
0.68
P
NS
NS
NR
NS
NS
NS
NS
NS
NS
NS
NR
NS
NR
NR
NS
0.09
NS
0.05
0.008
0.001
NS
NS
NR
0.000
N = Number of individuals carrying personal monitors.
NS = Not statistically significant from 0.
NR = p Value not reported, but mentioned as significant.
" = Year of publication.
b = 14 Subjects carried PEMS for 14 days for 191 valid measurements.
0 = Three outliers are removed and regression is for 188 measurements.
-------
600
500
400
E
| 300
0)
15
c
2 200
100
0 100 200 300 400 500 600
Outdoor Sulfate (nmoles/iVi )
Figure 1-21. Personal versus outdoor SO4=. Open circles represent children living in air
conditioned homes; the solid line is the 1:1 line.
Source: Suh et al. (1993).
500
100 200 300 400
Measured Sulfate (nmoles/iVi )
500
Figure 7-28. Estimated ("best fit" model) versus measured personal SO4 . Model
includes indoor and outdoor concentration and activity data. Open circles
are air conditioned homes; the solid line is the 1:1 line.
Source: Suh et al. (1993).
7-107
-------
Similar high correlations for total sulfur were found by Ozkaynak et al. (1996) in the
PTEAM study. Regressions of personal exposures in the PM10 fraction on outdoor sulfur gave
the following results (//g/m3):
Spets (day) = 0.62 (0.07 SE) + 0.69 (0.03) Sout N = 168 R2 = 0.78
Spets (night) = 0.27 (0.06) + 0.68 (0.03) Sout N = 162 R2 = 0.81
Another important consideration in evaluating personal exposures, from the indoor and
outdoor environmental measurements, is that the chemical composition of the excess in personal
exposure compared to the TWA exposure calculation may be significantly different than that
predicted from the indoor and ambient data alone.
In addition to the two factors cited just above, a microscale "personal cloud" can be
generated by the person's activities which complicates the exposure measurement process. This
effect is most important in occupational settings where personal exposures are not readily
comparable to weighted area sampling measurements. For example, Lehmann et al. (1990)
measured workers exposure to diesel engine exhaust by personal monitoring of PM10 with a
range of 0.13 to 1.2 mg/m3, compared to an area estimate range of 0.02 to 0.80 mg/m3. The
U.S. Centers for Disease Control (1988) reports the exposures of nurses and respiratory
therapists to the aerosols of ribavirin during treatment of patients by ribavirin aerosols
administered inside an oxygen tent. Bedside area monitors averaged 317 //g/m3 while personal
exposures ranged from 69 to 316 //g/m3 with an average of 161 //g/m3.
Environmental Tobacco Smoke (ETS) is a category of PM found in many indoor settings
where smoking is taking place or recently occurred. As stated in Section 7.2, ETS is the major
indoor source of PM where smoking occurs. Because of the depth of discussion of ETS in
Section 7.2.2.2, no further discussion is made here other than to note that ETS adds on the order
of 25 to 30 |ig/m3 to 24-h average personal exposures and residential indoor environments where
smoking takes place (Holcomb, 1993; Spengler et al., 1985).
The random ETS increment will tend to reduce the correlation between PEM and SAM. If
one were able to subtract out the ETS from the PEM PM data, the correlation of SAM with the
non-ETS PEM PM might be improved (Dockery and Spengler, 1981b). As stated as a caveat in
the introductory section 7.1, the inhalation of main-stream tobacco smoke will be a major
additive exposure to PM for the smokers, which dwarfs the nonsmoker's PEM PM. Therefore
the results presented so far apply only to nonsmokers, and a major proportion of the US
7-108
-------
population (e.g., smokers) has a total exposure to PM that is at least one order of magnitude
greater than that of the nonsmokers.
7.5 INDIRECT MEASURES OF EXPOSURE
7.5.1 Time-Weighted Averages of Exposure
The early air pollution literature related health to ambient particulate matter (TSP)
concentrations as a surrogate for personal exposures to PM. Although this relationship has been
shown to be highly questionable for specific individuals, it still is used in studies such as
Pengelly et al. (1987) who estimated TSP exposures of school children in Hamilton, Ontario, by
interpolation of ambient TSP concentrations to the school locations.
The first usage of a time-weighted-average (TWA) of environmental exposures to estimate
total human personal exposure to an air pollutant (Pb) was by Fugas et al. (1973). In theory, a
human exposure to PM could be estimated by use of Equation 7-2 and knowledge of the average
PM concentration while in each microenvironment (|iE) that a person experiences and the
duration of the exposure in each such jiE (Duan, 1982; Mage, 1985). For a room with no source
in operation, the whole room could be treated as a single jiE. However, when a PM source is in
operation and gradients exist, that very same room may need to be described by multiple jiEs.
These jiEs could have dimensions of an order of a few centimeters close to the source and of
several meters farther from the source.
Ogden et al. (1993) compared exposures from personal sampling and static area sampling
data for cotton dust exposures. The British cotton dust standard specifies static sampling,
because the 1960 dose-response study used to set the standard used static sampling data to
compute worker exposure and dosage. Ogden et al. (1993) found median personal exposures of
2.2 mg/m3 corresponding to a mean static background concentration of 0.5 mg/m3. They
concluded that "The presence of the body and its movement affect what a personal sampler
collects, so static comparisons cannot be used to infer anything about the relationship of the
(static) method with personal sampling." Ingham and Yan (1994) confirmed this finding by
modelling the human body as a cylinder and showing that unless the personal monitor
length/diameter ratio was greater than four, the aspiration efficiency (the fraction of particles
sampled that would be sampled in the absence of the body) could be greatly affected.
7-109
-------
Rodes et al. (1991) compared the literature relationships of personal exposure monitoring
(PEM) to |iE area monitoring (MEM) for PM, as shown in Figure 7-29, to which Ogden et al.
(1993) is added as a single point. The authors found that PEM/MEM ratios ranged from 3 to 10
in occupational settings, and from 1.2 to 3.3 in residential settings. These combined data show
that approximately 50% of all measured PEM PM values are more than 100% greater than the
estimated simultaneous MEM values using the TWA approach. Their explanation points to this
excess PM as due to the spatial gradient about indoor sources of PM which are usually well
away from area monitors which thus fail to capture the high exposures individuals may get when
in close proximity to a source. They suggest that clothing lint and skin dander could only add, at
most, a few percent to the total PM mass collected by a personal exposure monitor.
The Tokyo PM10 data of Tamura et al. (1996), added on Figure 7-29, show that for their
cohort of five elderly housewives and two male retirees that there is no evidence of a large
personal cloud effect as seen in the other studies listed. Japanese people customarily take shoes
off before entering a home and do not use wall-to-wall carpets, which would reduce track-in of
soil and eliminate a major reservoir for resuspension of dust. However, this same cohort does
display a "personal cloud" effect for the PM greater than PM10, with a maximum PEM/MEM
value of 3.3 for PEM = 55 |ig/m3 vs MEM 17 |ig/m3. This is consistent with the findings of
Thatcher and Layton (1995) showing, on Figure 7-15, an indoor increase due to human activity,
primarily for the PM greater than 10 |im in size, and Sheldon et al. (1988a,b) showing two U.S.
homes for the elderly with less than 10 |ig/m3 PM3 over a 72-h period in a nonsmoker's room.
7.5.2 Personal Exposure Models Using Time-Weighted Averages of Indoor
and Outdoor Concentrations of Particulate Matter
Several studies have used the relationship of Equation 7-2 to compute the time- weighted-
average (TWA) PM exposure of subjects. The procedure calls for a time-activity diary to be
kept so that the time at-home, outdoors, at-work, in-traffic, etc., can be defined. By use of jiE
monitoring data from the study itself (or literature values of PM concentrations
7-110
-------
100
_ ill Stevens (1969)
A Fletcher and Johnson (1988)
O Parker et al. (1990)
o Lioy et al. (1990)
A EPA PTEAM data
^Ogden et al. (1993)
Tamura et al. (1996)
10
.Q
4-i
«J
o:
0.1
10
—i—
30
—I—
50
—W—
70
-I—
90
-I—
95
—I—
98
+
Data median
10
30 50 70
Cumulative % less than
90 95
98
Figure 7-29. Personal activity cloud (PEM) and time-weighted average exposure (MEM).
Source: Rodes et al. (1991), Ogden et al. (1993), Tamura et al. (1996).
in similar jiEs) and concurrent ambient monitoring, one can predict the concentration that would
be measured if the subject had carried a PEM.
Because people in the United States spend, on average, 21 h indoors each day (U.S.
Environmental Protection Agency, 1989), the concentration in indoor jiEs is a most important
quantity for usage within a TWA PM model. The important articles on indoor air quality for
7-111
-------
PM have been reviewed extensively by Wallace (1996) and are covered in Section 7.2. The
articles that are discussed here predict PM exposures of non-smokers that include ETS, and most
provide PEM data for comparison. As opposed to the gaseous pollutants for which continuous
hour-to-hour time series of SAM data are available, PM SAM monitoring data have been often
only available as a time series of 24-h SAM measurements. Consequently, in much of the early
PM TWA literature, the modelers assumed, by necessity, the same ambient PM in the morning
and evening, which might not be accurate (Dockery and Spengler, 1981b).
Spengler et al. (1980) in a study of PEM, SAM and SIM in Topeka, Kansas, found the
averages of PEM = 30 |ig/m3, SIM = 24 |ig/m3 and SAM = 13 |ig/m3. They note "It suggests that
somewhere in an individual's daily activities, they are being exposed to PM at concentrations
higher than what is measured either indoors or outdoors". This relationship has been found in
almost all other studies, such as PTEAM (Clayton et al., 1993) where daytime PEM averaged
150 |ig/m3 and SIM and SAM averaged just under 100 |ig/m3. Spengler et al. (1985) measured
24-h PEM, SIM and SAM. The resulting relationship based on Equation 7-1 was: PEM = 17.7
|ig/m3 + 0.9 TWA. The authors noted, in addition to the previous suggestion, that the excess of
PEM over TWA may be due to an incorrect assumption that the indoor and outdoor are constant
during the 24-h sampling period.
Koutrakis et al. (1992), in a study discussed in Section 7.2 on Indoor Air, report that their
source-apportionment mass-balance model predicts penetration from outdoors to indoors on the
order of 85-90% for Pb and sulfur compounds. The authors claim that:
"We can satisfactorily predict indoor fine aerosol mass and elemental concentrations using
the respective outdoor concentrations, source type and usage, house volume and air
exchange rate."
The authors further note that this may be a cost-effective approach to estimating peoples'
exposure while indoors, since the necessary ambient data may be available and the housing
profile may be collected with a simple interview.
Colome et al. (1992) measured indoor and outdoor PM-10 at homes of asthmatics in
California. Their personal monitoring data, limited to three individuals, confirmed the relation
in Figure 7-16 that "some protection from higher outdoor concentration is afforded by shelter if
smokers and other particulate sources are not present". This observation may be important for
estimating the exposure of elderly and infirm people who are assumed to be the susceptible
cohort (Sheldon et al., 1988a,b).
7-112
-------
Klepeis et al. (1994) present an up-to-date TWA PM Model that uses, as an input, real-
time hourly PM SAM data and a mass balance equation to predict exposures of nonsmokers in
various indoor settings based on ambient PM data, presence of PM sources such as smokers, and
other variables relating to air exchange rates. The inclusion of the additive terms that allow for
sources, such as cooking and presence of smokers adds to the TWA of Equation 7-2, which in
effect is a correction for the underprediction of the |iE concentration.
In summary, as described by several authors, the PM PEM exposure of individuals who are
not smoke exposed has been shown to be higher than their corresponding TWA of SIM and
SAM in U.S. studies. The exact reason for this excess in PM, sometimes called a "personal
cloud", is not known (Rodes et al., 1991). It has been thought to reflect the fact that the person's
presence itself can stir up loosely settled-dust by induced air motion and vibration
(Ogden et al., 1993; Aso et al., 1993). Thatcher and Layton (1995) gave an example where
merely walking into and out of a room raised the total suspended dust (PM10) by 100%. A study
by Litzistorf et al. (1985) of asbestos type fibers in a classroom showed how fibers (f) were
stirred up when it was occupied. The levels rose from below the detectable level of 10000 f/m3
to 80000 f/m3 when occupied, and they returned to below detectable levels within 1 h after the
end of the class. Millette and Hays (1994) present a detailed discussion of the general topic of
resuspended dust in their text on settled asbestos dust.
It may not be a proper procedure to use a 24-h average concentration in a physical setting,
such as a kitchen, to estimate a person's exposure while in the kitchen. As described previously
in the discussion of the definition of a microenvironment in Section 7.1.2, the same kitchen can
constitute one or more jiEs depending on the source operation pattern. In many studies, such as
Spengler et al. (1985), the SIM sampled the indoor residential setting for 24-h in phase with the
PEM. The resulting average SIM will often underestimate the person's exposure while they are
at home and may contribute to the difference between a TWA exposure and the PEM.
In a similar manner, a person's workplace exposure may be more or less than that in their
home. In the PTEAM study (Clayton et al., 1993), there was a general decrease in exposure for
those employed outside their home. However, employment in a "dusty trade", such as welding,
may increase their PM PEM. Lioy et al. (1990) give an example of a subject with a hobby
involving welding having a 24-h PEM reading of 971 |ig/m3.
7-113
-------
Indirect estimation of a person's time-weighted-average (TWA) PM exposure may be a
cost-effective alternative to direct PEM PM measurement. Mage (1991) compared the
advantages and disadvantages of the TWA indirect method compared to the direct PEM method.
The primary advantages of the indirect method are the lower cost and lower burden on the
subject, because it uses only a time-activity diary and no PM PEM is required; the disadvantage
is the lower accuracy. The primary advantage of the PEM PM method is that it is a higher
accuracy direct measurement; the main disadvantages are the higher cost and higher burden on
the subject (see Section 7.3.1). Mage (1991) proposed a combined study design in which direct
measurements on a subset of subjects can be used to calibrate the TWA estimates of other
subjects. Duan and Mage (1996) present an expression for the optimum fraction of subjects to
carry the PEM as a function of the relative cost of the PM PEM to the TWA PM estimate and
the correlation coefficient between the PM PEM data and the PM TWA estimates.
7.6 DISCUSSION
7.6.1 Relation of Individual Exposures to Ambient Concentration
The previous sections discussed the individual PM PEM vs PM SAM relationships of the
studies listed in Table 7-26. In many of the cross-sectional PM studies, no statistically
significant linear relationship was found between PEM and SAM, but in some other studies the
relationship is positive and statistically significant. However, as shown by Lioy et al. (1990),
Janssen et al. (1995), and Tamura et al. (1996), the serial correlations between PEM and SAM
within an individual's time series are often highly positive and significant. This section discusses
these data in terms of understanding the complex relationship between the SAM concentrations
and the individual PEM exposures. In the following section, the relationship of the SAM to the
mean PEM in the community surrounding the SAM will be presented.
The principle of superposition is offered as a basis for visualization of the process involved
in creating a total exposure. A linear system will exist for respirable-PM PEM exposures if the
expected PEM response to a source emitting 2 mg/min of PM is exactly twice the PEM response
to that identical source emitting 1 mg/min of identical PM. If superposition applies, then we can
construct the total exposure by adding all the increments of exposures from the various source
classes and activities that a subject performs on a given day.
7-114
-------
Let the SAM be representative of the macroscale ambient PM concentration in the
community as shown on Figure 7-30a. This is the exposure that would be measured for a person
if they spent 24-h per day outdoors near the SAM site. Neglecting local microscale variation
(e.g. backyard barbecue or leaf burning), while people are outdoors they are exposed to 100% of
the SAM value (Figure 7-30b). Assume that this exposure is also the baseline PM for a location
in traffic which occurs outdoors. The increment produced by the local traffic is considered later.
While people are indoors, they are exposed to a variable fraction of time-lagged SAM PM.
This constitutes an amount of (1) the fresh PM which depends on recent SAM and the air
exchange rate between indoors and outdoors, and the PM deposition sinks (filtration of
recirculated air, surfaces, etc.), and (2) PM from outdoor sources that had been deposited in the
past but is resuspended due to human activity and air currents. PTEAM (Ozkaynak et al.,
1996), as cited in Section 7.2, found that outdoor air was the major source of indoor particles,
accounting for 75% of the fine fraction (<2.5 //m AD) and 67% of the thoracic fraction (< 10 jim
AD) in indoor air. It is noted that these average fractions will be lower in communities with
lower average SAM values. Lewis (1991) reported an apportionment of indoor air PM in 10
homes within a wood burning community in Boise, ID. The results showed that 50% of the fine
PM was of outdoor origin (SAM), and in 9 of 10 homes, 90% of the sulfur was from outdoors
(one home had an anomalous sulfate injection from a humidifier using tap water). This is
consistent with indoor sources varying independently of the SAM in a stationary manner
(constant mean and variance), so that the relative contribution of indoor sources to indoor
exposures decreases as SAM increases. Figure 7-30c represents the increment to PEM from
outdoor sources of SAM while the
7-115
-------
SAM
6 12 18 24
Time - Hours
E
"3)
SAM
SAM
Outdoors
6 12 18 24
Time - Hours
SAM
/ SIM \
, Indoors N
/ ^
•' J*^_
6 12 18 24
Time - Hours
0
\
N
N. ^ s
r
t,™
SAM
/ ^
/ x
\
' \
' \
/ \
i
'
Traffic
•<— Increment —>
to SAM
V
X
N.
6 12 18
Time - Hours
24
Occupational
Exposure
Increment
to SIM
SAM
X I
12 18 24
Time - Hours
ETS
Exposure
SAM
6 12 18 24
Time - Hours
"E
O)
3.
SAM
y~\
SIM s
/ Indoors x
/ non-ETS s
\ ' non-SAM Nx
^-'' ^^^_j| ^--
^)
E
>
"E
3.
20 ~~*
Cigarettes
Smoked
?
SAM
'
-
'
- .
^
.
0 6 12 18 24 0
Time - Hours
Figure 7-30. Components of personal exposure.
6 12 18
Time - Hours
24
7-116
-------
subjects are indoors at home and at work. The SAM value is shown as the dotted line for
reference in this and all the following Figures 7-30c to 7-30h.
While people are indoors, at home, and at work, they also are exposed to PM emitted by
indoor sources - other than ETS from passive smoking and specific occupational sources. These
sources, such as cooking, lint from clothing and furnishings, mold, insects, etc., create PM that
agglomerates and deposits as visible dust that can be continuously resuspended, which
constitutes an additional PEM increment. Figure 7-30d shows the additive effect of this source.
In traffic, or near vehicles in a parking garage or parking lot, people are exposed to an increment
of PM over and above the SAM value for that location. Figure 7-30e shows the additive PM for
this setting that would be added to Figure 7-30b for the local vehicular emissions.
At work in a "dusty trade" (e.g., welder, mechanic, or miner) there is an increment of
exposure associated with these occupational activities that generate PM. Figure 7-3 Of represents
the additive PM for these activities which are assumed to take place "indoors".
In an indoor setting, in the presence of a smoker or the wake of a smoker, a PEM will
record an increment of ETS associated with the act of smoking. Figure 7-3 Og shows the added
PM increment for this source.
Last, but not least, is the physical act of smoking itself. As described previously, the main
stream smoke from a cigarette, cigar, or pipe is inhaled directly without being sampled by a
PEM. The mass of PM directly inhaled from smoking one-pack-per-day of cigarettes rated as
delivering "1 mg 'tar' per cigarette by FTC method" is 20 mg per day (Federal Trade
Commission, 1994). If this were distributed into a nominal 20 m3 of air inhaled per day, it
would be an additive increment on the order of 1 mg/m3 to a 24-h PEM reading. Tar emissions
as rated by the Federal Trade Commission (1994) range from <0.5 mg/cigarette to 27
mg/cigarette. Therefore one-pack-per-day smokers can have a PM exposure standard deviation
that is much larger than the mean exposure to PM of non-smokers, simply from choice of brand.
Figure 7-30h represents the impact of the act of smoking as creating exposures represented by
the vertical spikes with an integral area > 1 mg-day/m3 per day.
For all subjects, by the principle of superposition, the sum of the areas shown in
Figures 7-30b and 7-30c represents the exposure of an individual to the PM constituents that are
characterized by a SAM PM concentration. The additional exposure categories that are
independent of the SAM concentration (Figures 7-30d through 7-30g) and are appropriate for
7-117
-------
that subject would represent the portion of 24-h PEM PM that is not associated with SAM.
Variance of SAM should explain much of the variance in the SAM related PEM fraction as
defined by Figures 7-30b and 7-30c. The summation over a full day for all categories 7-30b to
7-30g would be the PEM for any subject, such as is shown in Figure 7-2 (Repace and Lowery,
1980).
Although there are no data for PEM PM exposures of individuals living in homes without
any indoor sources of PM, there are data for PEM sulfate as discussed previously in Section
7.4.3. Given that there are negligible sources of sulfur (S) that originate in the home (matches,
low-grade kerosene, humidifiers using tap water), the high correlation of PEM sulfate and SAM
sulfate (R2 = 0.92) of Figure 7-27 reported by Suh et al. (1993), where no appreciable sources of
S were present, is an indication that the same relationship should hold for all SAM PM of that
size range. The data of Anuszewski et al. (1992) show that light scattering particles measured by
nephelometry had a very high correlation between indoor and outdoor concentrations (R2 > 0.9)
for one home, but were lower for others. Lewis (1991) and Cupitt et al. (1994) report that PM10
appears to penetrate with an average factor of 0.5 in Boise homes without woodburning. The
factor goes up to 0.7 with woodburning, and the authors assume that the factor would go up to
0.9 in the summer when homes are less tightly sealed. However, the authors did not consider the
deposition rate k. This is in contrast to the data of Thatcher and Layton (1995), who measured k
and found penetration factors of 1.0 for all PM sizes < 10 //m.
If the variance of the PEM PM portion which is uncorrelated to SAM (Figure 7-3 Od to 7-
30g) is very large, the percentage of the variance of the PEM PM that can be explained by the
variance of SAM PM will be very small. It may be possible that the different populations
sampled, cited in the studies of Table 7-26, have widely different home characteristics,
occupations, mode of commuting, and smoking exposures that contribute to the different PEM
vs SAM relationships. In some of the cleaner communities (such as Watertown, MA; Topeka,
KS; Waterbury, VT; and Kingston and Harriman, TN) SAM averaged less than 20 |ig/m3. The
non-SAM increments to PEM exposure in these locales were greater than the SAM and may
have been so variable between people (eg. ETS and non-ETS exposures pooled together) that the
PEM PM became insignificantly correlated with the SAM PM data. The exception is Houston,
TX, with a SAM = 16 |ig/m3 and a significant R2= 0.34 (0.005 < p < 0.05). However, Morandi
et al. (1988) note that deletion of two outlier observations would reduce R2 and make it
7-118
-------
nonsignificantly different from 0 (p > 0.2). This is in contrast to the three studies in
communities with high SAM levels (Tamura et al., 1996; Clayton et al., 1993; Lioy et al., 1990),
where the relations between PEM and SAM were significant.
All discussions above relate to nonsmokers. As for the smoker, the exposure from Figure
7-30h would outweigh the sum of all the other exposures, 7-30b through 7-30g. This smoking
increment may have an important implication for interpretation of epidemiology studies that
relate ambient PM, as a surrogate of exposure, to mortality or morbidity.
Because the daily amount of individual smoking and other exposures from indoor sources
(cooking, ETS, resuspension of settled dust by walking into carpeted rooms, hobbies) is
independent of the daily SAM value, the variance of the PM SAM value is a surrogate for the
variance component of total personal exposures to PM associated with PM SAM. For
nonsmokers ambient PM reflects about 50 to 70% of their PM10 exposure that by definition does
not contain directly inhaled smoke exposure (Tamura et al., 1996; Ozkaynak et al., 1996). This
relationship would also hold for the total PM exposure of smokers minus the effective increment
they receive from their direct smoking which is independent of PM SAM. Therefore, a
relationship between ambient PM (SAM) and human exposure to PM (PEM) that makes sense, is
that the SAM value is a surrogate for personal exposure to PM (PEM) from PM originating in
the ambient air. This relationship would apply to everyone, smokers and nonsmokers alike.
However, treating SAM as a surrogate for total personal exposure to PM from all sources,
including those major sources of PM that vary independently of SAM (active smoking and
occupational exposures), would be wrong.
7.6.2 Relation of Community Participate Matter Exposure to Ambient
Particulate Matter Concentration
For the morbidity/mortality studies described in Chapter 12 that use SAM as the
independent variable, that SAM can be interpreted to stand as a surrogate for the average
community exposure to PM from sources that influence the SAM data. These sources of
ambient PM do not include indoor sources such as the "personal cloud" of skin flakes and lint,
ETS, cooking fumes, and resuspended PM from walking on a dirty carpet. Thus, if we could
subtract off from each PEM measurement the contribution to the total exposure from the indoor
sources, such as smoking, cooking, carpets, and personal clouds, the residual PM from ambient
7-119
-------
sources would probably improve the correlation with SAM, as described by the data of Tamura
et al. (1996) for nonsmoking-noncarpeted homes occupied by elderly people. Mage and
Buckley (1995) tested the relationship of the mean PEM to SAM as a means to minimize the
affect of variations of these indoor sources of PM on the relation of PEM to SAM, and their
results, with modifications, are presented in the following section.
There are several different models for these analyses and although most describe the same
linear relationship, the models differ greatly in their assumptions about the error terms. The
discussion of the various models is followed by U.S. EPA reanalyses of five different
PEM-SAM data sets described previously in Section 7.4.
7.6.2.1 Methodology
Methods for Missing Data
One common difficulty in the use of aerometric data is the presence of missing data
elements. For example, consider the following PEM data from the study of Tamura et al.
(1996). The authors measured the 48-h personal exposure to PM10 for seven individuals living
near a main road for 11 periods in four seasons distributed over a complete year. This example
has a great deal of missing data, and for purposes of computation, the data were split into a
group living close to the road (persons A, B, C, and D), and a group living farther from the road
(persons E, F, and G). Their indoor and outdoor data were shown previously on Figure 7-24.
The PEM data for the first group are shown in Table 7-27.
Unless pairwise correlations are computed, the standard solution to the problem is to delete
all observations for which any of the variables are missing. This approach, known as a
complete-case analysis, is standard in the majority of the statistical packages. For this example,
we would be left with only 5 of the original 11 periods of observation. This section will
describe a model which will allow for the inclusion of all available data.
The reason for the missingness of the data is extremely important because it determines our
ability to obtain maximum likelihood estimates (MLE). The following definitions are
paraphrased from Little and Rubin (1987): If the probability of being missing is independent
7-120
-------
TABLE 7-27. 48-HOUR PERSONAL EXPOSURE TO PM10
(Data Taken by Subjects Living Along a Main Road in Tokyo)
Period
1
2
O
4
5
6
7
8
9
10
11
Person A
43.7
27.4
30.2
22.4
57.4
M
M
24.6
31.0
22.9
68.7
Person B
40.4
31.5
39.2
29.2
43.2
26.1
37.9
M
34.5
M
51.8
Person C
37.5
29.8
32.7
25.9
43.3
27.9
35.8
41.4
36.0
24.3
52.6
Person D
52.3
26.0
M
38.2
M
39.9
34.6
39.8
45.6
30.6
68.1
M = Missing observation.
Source: Tamuraet al. (1996).
of both the variables missing and the variables present, then the data are said to be missing
completely at random (MCAR). If the probability of being missing depends on the variables
present, but not on the variables missing, then the data are said to be missing at random (MAR).
If neither situation holds, then there are no general solutions to the problem. This would happen
if the value of the missing variable (which is not known to us) is directly related to its
probability of being missing. Laird (1988) discusses models used for maximum likelihood
estimation with missing data, as well as a detailed discussion of the non-response mechanism.
One solution is to assume that the measurements are distributed as a multivariate normal
distribution (or to assume that some transformation of the data give a multivariate normal
distribution). The estimation of the parameters of a multivariate normal model with missing
data is a problem which has been discussed for many years (see Afifi and Elashoff, 1966). The
first general solution to the problem of estimating a mean vector and covariance matrix from a
multivariate normal distribution with data missing at random was given by Woodbury and
Hasselblad (1970). The solution, referred to as the "Missing Information Principle", was
generalized to other missing data problems by Orchard and Woodbury (1972). Proof that the
7-121
-------
method always improved the likelihood was given by Dempster et al. (1977), and the
generalized solution method was named the E-M algorithm.
To describe the problem, the following notation will be used. Let x = x1,x2,...,xk be a
k-dimensional random vector from a multivariate normal distribution
f(x|M,s = (2n)
(7-10)
where £ is a symmetric positive definite matrix and ji is a vector. The mean of the vector x is ji
and its covariance is £. Assume that we have n observations from this distribution, X1,X2,...,Xn.
The E-M algorithm can be used to estimate the parameters of a multivariate normal
distribution. The method starts with any reasonable first estimate of the parameters. Assume
that we have initial estimates of the parameters ji and £, which can be obtained by filling in the
missing data with the column means and then estimating the parameters in the usual manner.
The E step consists of estimating the sufficient statistics. For this model, the sufficient statistics
are the sums and sums of squares of cross products.
Assume that at one particular point, X;, some of the observations are missing and some of
the observations are present. Without loss of generality, we will drop the subscript, i, and
rearrange the subscripts so that the vector X is [Xl3 X2] where all of the observations, Xl5 are
missing and all the observations X2 are present. Partition the mean vector ji and the covariance
matrix S in a similar fashion
M =
and E =
En E12
E21 E22
(7-11)
Compute the regression of the missing observations on the observations present
R - y T ~*
p - Zj12 Zj22 .
(7-12)
Estimate the missing values, Xl3 by their expected values
E(X^ = P! + p(/2 - |J2 ) .
(7-13)
7-122
-------
Compute the correction to the expected sums of squares
~ ^12 ^22 ^21 • ('"•'• v
Now add the vector X to the sums and XX' to the sums of squares and cross products using their
expected values for the missing values; remember to add Sn 2 to the cross products
corresponding to Xr
The M step consists of recomputing the estimates of ji and £ from the completed sums and
sums of squares and cross products. This procedure will converge, typically taking five to 20
iterations for a moderately sized problem. Using the methods just described, the estimates of
both the missing values and the parameters for the data of Tamura et al. (1996), based on
U.S. EPA reanalyses, are shown in Table 7-28.
This method was also used to fill in the missing values for persons E, F, and G (shown in
Table 7-29). Once the missing data were estimated, the average across all seven persons was
computed and compared with the ambient measurement monitor as shown in Table 7-30. These
data will be used as examples for the next section.
Linear Regression Models
The various linear regression models are illustrated next using the average personal
exposure values from the Tamura et al. (1996) data set which were described in the previous
section. For these examples, the average personal exposure will be considered the dependent
variable and the ambient concentration at the Itabashi site will be the independent variable.
The first model is often referred to as the fixed independent variable model (see Dunn and
Clark, 1974, p. 225). The model assumes that the dependent variable is a linear function of the
independent variable with random error which is normally distributed (this is a bad assumption
but this is the most commonly used model). This can be written as
7-123
-------
TABLE 7-28. PARAMETER ESTIMATES FOR 48-HOUR PM10 PERSONAL
EXPOSURE MONITOR DATA TAKEN BY SUBJECTS LIVING
NEAR A MAIN ROAD IN TOKYO Cug/m3)
(Estimated Missing Values Shown in Parentheses)
Day
1
2
O
4
5
6
7
8
9
10
11
Means
Person A
43.7
27.4
30.2
22.4
57.4
(29.3)
(28.9)
24.6
31.0
22.9
68.7
35.1
Person B
40.4
31.5
39.2
29.2
43.2
26.1
37.9
(43.3)
34.5
(26.7)
51.8
36.7
Covariance/Corr elation Matrix
Person A
Person B
Person C
Person D
215.8
0.745
0.819
0.888
83.9
58.9
0.949
0.731
Person C
37.5
29.8
32.7
25.9
43.3
27.9
35.8
41.4
36.0
24.3
52.6
35.2
(Correlation below
96.4
58.4
64.3
0.816
Person D
52.3
26.0
(37.4)
38.2
(58.4)
39.9
34.6
39.8
45.6
30.6
68.1
42.8
diagonal)
157.4
67.6
79.0
145.6
Source: Parameter estimates, including the calculation of estimated missing values, and covariance/correlation
matrix results from reanalyses by U.S. EPA of data from Tamura et al. (1996).
Y! = Po + Mi + e1f where (7-15)
i = l,2,...,n, n is the number of observations, and e; is normal with mean 0 and variance a2. No
assumption is made about the distribution of the independent variable since it is considered to be
fixed.
Using the previous example, the estimated coefficients are given in Table 7-31, and the
results are shown graphically in Figure 7-31.
The second model is often referred to as the bivariate normal model (see Dunn and Clark,
1974, p. 239). This model assumes that the dependent variable and the independent variable are
both normally distributed. Actually, the assumption is stronger—it assumes that
7-124
-------
TABLE 7-29. PARAMETER ESTIMATES FOR 48-H PM10 PERSONAL EXPOSURE
MONITOR DATA TAKEN BY SUBJECTS LIVING FARTHER FROM THE SAME
TOKYO MAIN ROAD DESCRIBED IN TABLE 7-28 (in
(Estimated Missing Values Shown in Parentheses)
Period
1
2
3
4
5
6
7
8
9
10
11
Person E
57.1
(30.9)
26.8
32.9
68.6
31.2
26.5
35.8
40.7
29.8
62.5
Person F
62.2
26.5
23.1
(30.6)
(69.2)
26.6
24.0
(28.7)
(36.9)
27.5
51.2
Person G
(37.1)
(29.0)
25.3
27.2
48.0
24.4
29.7
37.7
35.4
22.4
61.0
Source: Parameter estimates, including the calculation of estimated missing values, based on reanalyses by
U.S. EPA of data from Tamura et al. (1996).
the joint distribution of the two variables is bivariate normal. The bivariate normal distribution
is a special case of the multivariate normal distribution described earlier. The intercept, P0, and
regression coefficient, Pl3 are estimated by the same formulas as were used in the first model
even though the assumption is not the same. The R-squared term is also the same, but the
ANOVA Table no longer makes any sense.
The third linear model is the same as the first except that a lognormal error term is used.
This kind of model requires the use of a general linear model fitting routine. The model gives
less weight to large deviations about the predicted line where the predicted values are already
large. The model still assumes that the independent variable is fixed and measured without
error. The fit to the previous example is shown in Table 7-32. There is no measure comparable
to R2, but the log-likelihoods can be compared directly. Note that
7-125
-------
TABLE 7-30. AVERAGE PERSONAL EXPOSURE DATA COMPARED WITH
ITABASHI SITE MONITOR (PM,n;
Period
1
2
O
4
5
6
7
8
9
10
11
Itabashi Site
66.5
30.1
37.9
50.3
90.5
40.7
40.5
55.1
70.6
31.9
99.5
Average Personal
47.2
28.7
30.7
29.5
55.4
29.3
31.1
35.9
37.2
26.3
59.4
Source: Data from Tamura et al. (1996).
TABLE 7-31. RESULTS OF LINEAR REGRESSION ANALYSIS, ASSUMING A
NORMAL ERROR USING THE EXPOSURE DATA FROM JAPAN
Linear regression
Y = intercept + slope X
Variable Beta Std. Err. Beta
Intercept 11.32 3.025
Slope 0.466 0.050
ANOVA Table
Source Sum of Squares Mean Square Error D.F. F-value
Regression 1194.3 597.2 2 42.9
Error 125.3 13.9 9
TOTAL 1319.6 120.0 11
R-squared = 0.905
Log-likelihood = -28.99
Source: U.S. EPA reanalyses of data from Tamura et al. (1996).
7-126
-------
80
1 60
40
o>
Q.
0)
O)
flj
0)
< 20
30 60
Ambient
90
120
Figure 7-31. Plot of 48-h average personal PM10 exposure and ambient PM10 data from
Japan—linear regression.
Source: U.S. EPA reanalyses of data from Tamura et al. (1996).
TABLE 7-32. RESULTS OF LINEAR REGRESSION ANALYSIS, ASSUMING A
LOGNORMAL ERROR USING THE EXPOSURE DATA FROM JAPAN
Multiple log-linear regression analysis
Variable Mean
Ambient 55.78
Mean 1
Sum of squares for error = 0.089
Mean square error = 0.010, d.f. = 9
Log-likelihood = -28. 50
Beta Std.Err.Beta
0.43 0.06
13.07 3.26
Source: U.S. EPA reanalyses of data from Tamura et al. (1996).
7-127
-------
the linear model with a lognormal error fits slightly better than the normal error model, although
the difference of 0.49 in the log-likelihood is not statistically significant.
Orthogonal Regression Models
Orthogonal regression is also known as principle components regression. There is no real
assumption about the model. The purpose of the analysis is to pass a line through the data such
that as much of the variation is explained as possible. Variation is measured as the squared
distance from the points to the fitted line. Because no distributional assumptions are made, no
confidence limits can be placed on the estimated line. The measure of the total variation is
Total variation = ana22 - o2l2 . (7-16)
The fraction of the variation explained is derived from the eigenvalues of the covariance matrix,
and the regression line corresponds to the first eigenvector. That is, the eigenvalues are the
solution of
\ / X
11 n i
= 0. (7-17)
°22
1 0
0 1
The values of A which satisfy equation (7-17) are
°22 y°ii- °22)2 + 4oi2
The slope of the line corresponding to the largest eigenvalue, Al3 is
(7-19)
The intercept, P0, is easily calculated because the line must pass through the mean of the data.
7-128
-------
The measure, percent of variation explained, is a generalization of the multiple R2 measure
from a single dependent variable, but its behavior is somewhat different. For a two variable
problem it can be calculated as X1/(X1 + A2). In general, for correlations near 1, it will be about
twice as good (.975 to .98 instead of .95), but for correlations near 0, the behavior is not as
simple. As a result, it can only be used to compare one orthogonal regression with another.
Because the standard correlation coefficient is a non-parametric measure of association, it can be
used for orthogonal regression as well. The results of fitting by U.S. EPA of an orthogonal
regression model to the previous example are in Table 7-33. The slope and intercept are almost
identical to the normal error model values shown in Table 7-31.
TABLE 7-33. RESULTS OF AN ORTHOGONAL REGRESSION
ANALYSIS OF THE EXPOSURE DATA FROM JAPAN
Y = intercept + slope X
Variable Beta
Intercept 10.83
Slope 0.475
Total variation 5686.9
Percent explained 98.5
Source of data: U.S. EPA reanalyses of data from Tamura et al. (1996).
Measurement Error Models
In general, most linear regression analyses assume the independent variable has no
measurement error. When this error exists and no correction is made for it, the estimated
regression coefficients tend to be biased towards zero. Because we often have multiple monitors
we can often attempt to estimate these components of variation, and therefore correct our
estimated regression coefficients. The solution usually requires some additional assumptions—
in particular the assumption of multivariate normality is necessary for most of the solutions.
Additionally, some information must be available about the error variance. Either the error
variance of the independent variable or the dependent variable, or the ratio of the error variance
to the variance of the dependent variable must be known exactly. In some cases, these values are
7-129
-------
known with sufficient accuracy from other experiments so that the values can be treated as
known.
Much of the material on measurement error in continuous variables comes from the work
of Kendall and Stuart (1961) and Fuller (1987). Both authors make the same distinction that
was made in the earlier section regarding the fixed or random nature of the independent variable.
We will consider the more interesting case of measurement error in an independent random
variable.
This subsection assumes a model with a continuous dependent variable and a continuous
independent variable whose values are considered to be random and measured with error. For
example, Hasabelnaby et al. (1989) described an analysis of pulmonary function data using
measurements of NO2 exposure as a covariate. The true NO2 exposure was assumed to be a
random variable which was estimated by sampling NO2 levels in the home for two weeks out of
the year. The other terms in the model were height and gender of the individual, and these were
measured with little or no error.
The single random independent variable model assumes a single independent variable
whose values, X;, are random values. The model is
(7-20)
and we wish to estimate P0 and Pr Assume that the expected value of x is j^, the expected value
of y is jjy, and that the variance of x is oxx. We do not observe y; and x;, but rather Y; and X;,
where
Y1 = y1 + Y/ and (7-21)
Xi = X7. + 57. , (7-22)
and where Y; is normal with mean 0 and variance o^ and 5; is normal with mean 0 and variance
oxx. The covariances between x;, 5;, and Y; are assumed to be zero. This assumption implies that
the vector (Y,X) is distributed as a bivariate normal vector with mean
7-130
-------
E(Y,X) =
. M,
(7-23)
and covariance
OYY OXY
°XY axx
Plaxx+a
xx
>laxx
XX
(7-24)
Let (3j be the standard regression estimate based on the observed data,
V1 n
V V ~\ f V V ~\
Ai A )( ' 1 ' > •
7=1
(7-25)
7=1
The expected value of (^ is
-1
yy
'XX
(7-26)
Thus, for the bivariate normal model, the least squares regression coefficient is biased towards
zero. The ratio, oxx laxx i§ known by several names including the attenuation, the reliability
ratio, and in genetics as the heritability (Fuller, 1987).
Maximum likelihood equations can be set up for the bivariate normal model with
measurement error. The first and second moments, which are sufficient to determine the
distribution, will give five equations in the six unknown parameters, j^, oxx, oxx, o^, P0, and Pr
Clearly, some additional information is needed to make the problem identifiable. The three
possibilities for additional information are oxx, o^, or the ratio o^o^, which lead to three
different solutions. Two of these solutions are discussed in the following subsections.
If the measurement error in X, oxx, is known, then the solution is straightforward. For
example, assume we know the variation between the ambient monitors because we have multiple
monitors. Let Sxx be the maximum likelihood estimate of oxx, SYY be the maximum likelihood
estimate of OYY, and SXY be the maximum likelihood estimate of OXY. The maximum likelihood
estimate of Pj becomes
7-131
-------
! = SXY/(SXX - oxx). (7-27)
Note that this estimator reduces to equation (7-25) when the measurement error in x, oxx, is 0.
If the measurement error in Y, o^, is known, then there is a comparable solution. Let Sxx,
SYY, and SXY be defined as before. The maximum likelihood estimate of Px becomes
(7-28)
All of this was based on the assumption that there was a true relationship between x andy
that had no error. If, in fact, there was some error so that
// = Po + Pi*/ + ?/> (7-20)
where e; is normal with mean 0 and variance oee2, then the estimate of Px would still come from
equation (7-25), but the correlation would be estimated as
°xy
In order to estimate oxx and o^, we can use an analysis as described in the following section.
This correlation represents the upper bound to the observed correlation. That is, it is the
correlation of the personal and ambient monitors if we had an infinite number of both. Under
the assumption of equation (7-20), the value of this correlation is 1.
Components of Variance Models
7-132
-------
If we have measurements from several individuals over time or several ambient monitors
over time, then these measurements can be used in an analysis of variance (ANOVA) model.
The purpose of the model is to estimate the variation between individuals and/or the variation
between monitors. This information can then be used to adjust our slope estimates as described
earlier, as well as letting us estimate the correlation between ambient and personal monitors
assuming we had an infinite sample of both.
The logical analysis for this kind of data is a repeated measures design (see Winer, 1962,
pp. 105-124). For most examples, the necessary components can be obtained from the results of
a standard two-way ANOVA table. For example, consider the data of Tamura et al. (1996) after
the missing values have been estimated (Tables 7-28, 7-29). There are 7 individuals measured
over 11 48-h periods, resulting in the following ANOVA Table 7-34.
TABLE 7-34. RESULTS OF AN ANOVA ANALYSIS OF THE EXPOSURE DATA
FROM JAPAN
Source of Variation
date
person
date x person
Total
D.F.
10
6
60
76
S.S.
9235.41
634.53
2248.66
12118.60
M.S.
923.54
105.76
37.48
Source of data: U.S. EPA reanalyses of data from Tamura et al. (1996).
These results indicate that the mean square error for person is 105.76. This represents an
estimate of 7 o^ + oee (mean squared error). The value, 37.48, represents an estimate of oee, so
that Oyy can be estimated by (105.76 - 37.48) 11 = 9.75. Because we will actually use the mean
of 7 persons to estimate the average, the variance component we need for equation (7-28) is
estimated by 9.75/7 = 1.39.
For example, consider the data of Tamura et al. (1996). From the above analysis, we have
an estimate of the person variation, o^, of 1.39 (for the mean of 7 individuals). Thus using
equation (7-28), we can estimate Px as (119.97 - 1.39) / 232.83 = 0.509.
7-133
-------
7.6.3 U.S. EPA Analysis of Data Sets
7.6.3.1 Tokyo, Japan Data Set
The data set of Tamura and Ando (1994) and Tamura et al. (1996) presents an interesting
problem. Shown in Table 7-35 is the correlation matrix for average personal exposure with the
two nearby ambient sites as well as their average. The Yamato site is located near a highway
intersection 0.7 km from the central Itabashi site.
TABLE 7-35. COVARIANCE AND CORRELATION MATRIX FOR AVERAGE
PERSONAL EXPOSURE AND AMBIENT EXPOSURES FROM JAPAN
Covariance/Corr elation Matrix (Correlation
Average person
Itabashi site
Yamato site
Average site
Average Personal
119.97
(0.951)
(0.736)
(0.840)
below diagonal)
Itabashi Site
232.83
499.30
(0.874)
(0.949)
Yamato Site
308.81
748.50
1467.62
(0.983)
Average Site
270.82
623.90
1108.06
865.98
Source of data: U.S. EPA reanalyses of data from Tamura et al. (1996).
Note that the correlation of the average personal exposure is much higher with the Itabashi
site than with the Yamato Site or the Average of the two sites. The estimated components of
variance can give strange results when there are only two sites and one is much more highly
correlated. For this reason, only the Itabashi site is used in the following analyses. If there had
been additional sites it would have been possible to make all of the analyses in Table 7-36, but
only those single site analyses are included at this time.
7.6.3.2 Phillipsburg, New Jersey Data Set
The personal exposure data (Lioy et al., 1990) contained some missing values and three
outlier values, and they all were estimated as described earlier. The results of U.S. EPA
reanalyses are shown in Table 7-37. In order to estimate the error variances, these data were
used in an analysis of variance as described earlier. The results are shown in Table 7-38.
7-134
-------
TABLE 7-36. SUMMARY OF RESULTS OF THE
ANALYSIS OF THE EXPOSURE DATA FROM JAPAN
Regression Model
Linear, normal error
Linear, lognormal error
Orthogonal
Linear adjusted for person error
Linear adjusted for ambient error
Measures of Association
Correlation of personal averages with Itabashi site
Correlation adjusted for measurement error
Average correlation of ambient with mean person
Average correlation of person with mean ambient
Fraction of variation explained by orthogonal regression
P, Pn
0.466 11.3
0.431 13.1
0.475 10.8
0.509 8.9
(Not available)
Value
0.951
(Not available)
(Not available)
0.872
0.985
Source: U.S. EPA reanalyses of data from Tamura et al. (1996).
The site monitoring data contained some missing values, and they were estimated by U.S.
EPA as described in Section 7.6.2.1. The means, covariances and correlations were also
estimated. The results are in Table 7-39. In order to estimate the error variances, the same data
were used in an analysis of variance as described earlier. The results of the EPA analyses are
shown in Table 7-40. The individual exposure values were averaged as well as the site exposure
values. These means are shown in Table 7-41.
The same regression analyses described earlier were performed by U.S. EPA. A plot of
the linear regression is shown in Figure 7-32. The orthogonal regression gives virtually an
identical plot and is not shown. The results of the analyses are in Table 7-42.
Note that all estimated regression equations are quite similar. The interesting value is the
correlation adjusted for measurement error. This represents an estimate of the correlation
between the mean of an infinite number of personal samplers and the mean of an infinite number
of fixed site samplers. This value is relatively close to one, but we do not have good estimates
of its variance to tell if the value is really different from one.
7-135
-------
TABLE 7-37. PERSONAL EXPOSURE SUSPENDED PARTICULATE MATTER DATA FROM
PHILLIPSBURG, NEW JERSEY. MISSING VALUES ESTIMATED ( ); OUTLIER VALUES RECOMPUTED [ ].
Person Identifier (//g/m3)
Day
1
2
O
4
5
6
7
8
9
10
11
12
13
14
01
59
52
74
115
65
45
75
104
84
55
10
39
26
45
02
85
58
69
88
37
16
77
81
29
29
60
59
44
44
11
54
85
94
136
139
56
65
79
48
70
65
80
65
89
31
39
17
56
104
38
22
35
67
56
35
25
23
35
17
41
(53.2)
(76.7)
86
65
77
34
36
83
85
59
36
127
31
105
42
36
45
77
116
64
27
80
32
122
81
[48.1
57
47
117
51
41
50
90
112
56
28
27
69
30
25
49.4]
32
114
(24.8)
52
28
53
93
120
52
21
34
61
36
39
43
35
67
24
61
123
104
200
125
184
60
92
112
57
199
93
121
47
117
62
67
56
134
272
190
58
(110.2)
91
96
77
84
95
95
63
81
96
50
166
193
79
57
124
144
156
63
99
31
71
44
82
79
49
81
98
49
12
77
69
123
41
32
45
18
14
91
50
66
77
164
(95.7)
54
107
96
91
66
78
63
31
57
92
32
63
187
172
89
99
184
198
[100.6]
135
122
72
109
108
Source: Data from Lioy et al. (1990). Missing values estimates and recomputed outlier values calculated by U.S. EPA.
-------
TABLE 7-38. RESULTS OF AN ANOVA ANALYSIS OF THE PERSONAL
EXPOSURE DATA OF PHILLIPSBURG, NEW JERSEY
Source of Variation
Date
Person
Date x Person
Total
d.f.
13
13
169
195
s.s.
119,600
103,300
149,900
372,800
m.s.
9202
7942
887
Source: U.S. EPA reanalyses of data from Lioy et al. (1990).
TABLE 7-39. SAM SITE CONCENTRATIONS, PM10 DATA
FROM PHILLIPSBURG, NEW JERSEY
[Missing Values Estimated ()].
Day Site 101 Site 102 Site 103
01
02
03
04
05
06
07
08
09
10
11
12
13
14
Means
C ovari ance/C orrel ati on
Site 101
Site 102
Site 103
Site 020
26
51
94
148
76
15
44
101
59
46
37
28
27
21
55
41
(55.6)
(101.8)
155
81
17
47
105
67
52
36
33
27
23
.2 60.1
28
55
112
165
76
13
49
119
68
50
35
28
27
19
60.3
Site 020
24
46
98
209
85
50
51
99
66
57
34
28
25
38
65.0
Matrix (Correlation below diagonal)
1313.
0.
0.
0.
.9 1346.5
995 1393.8
996 0.994
943 0.935
1538.9
1581.4
1816.2
0.929
1596.6
1630.9
1850.1
2183.4
Source: U.S. EPA reanalyses of data from Lioy et al. (1990).
7-137
-------
TABLE 7-40. RESULTS OF AN ANOVA ANALYSIS OF THE SITE EXPOSURE
DATA OF PHILLIPSBURG, NEW JERSEY
Source of Variation
Site
Day
Site x Day
Total
d.f.
3
13
39
55
s.s.
671
90286
3615
94572
m.s.
223.6
6945.1
92.7
Source: U.S. EPA reanalyses of data from Lioy et al. (1990).
TABLE 7-41. AVERAGE PERSONAL PM10 EXPOSURE DATA COMPARED WITH
THE SITE EXPOSURE DATA FOR PHILLIPSBURG, NEW JERSEY
Day
1
2
o
J
4
5
6
7
8
9
10
11
12
13
14
Ambient Average (//g/m3)
29.75
51.55
101.45
169.25
79.5
23.75
47.75
106
65
51.25
35.5
29.25
26.5
25.25
Average Personal (//g/m3)
60.15
58.91
106
134.29
86.76
42.07
80.23
91.86
79.19
69.57
60.74
62.79
57.14
62.04
Source: U.S. EPA reanalyses of data from Lioy et al. (1990).
7.6.3.3 Beijing, China Data Set
The Beijing, China data set reported by the World Health Organization (1985) is listed in
Table 7-43. From these data, daily mean values of the ambient and personal exposure values
were computed. An U.S. EPA reanalysis of these data is shown in Table 7-44 and in Figure 7-
33. The results of the analysis indicate that there is not a significant linear relationship between
the personal and ambient monitoring data. For this reason, it does not
7-138
-------
200
150
100
a
a.
a
O)
2
0
< 50
50 100
Average Site PIVJo
150
200
Figure 7-32. Plot of relationship between average personal PM10 exposure versus ambient
PM10 monitoring data from Phillipsburg, NJ and regression line calculated
by U.S. EPA.
Source: Lioy et al. (1990).
make any sense to adjust the coefficient for measurement error. The subjects all worked at the
same institute so their daytime personal exposures may not have been independent of each other.
7.6.3.4 Riverside, California Data Set
Both the personal exposure and the monitoring data used in analyses by Clayton et al.
(1993) contained some missing values, and they were estimated by U.S. EPA as described
earlier. The estimated correlation/covariance matrix results of U.S. EPA reanalyses of these data
are shown in Table 7-45.
Because the individual monitors were placed on different individuals each period, we can't
really estimate the variation between individuals. Based on previous analyses, we know that
most of the residual is variation between individuals, and so we will use this as a
7-139
-------
TABLE 7-42. RESULTS OF THE ANALYSIS OF THE
EXPOSURE DATA FROM PHILLIPSBURG, NEW JERSEY
Regression Model
Linear, normal error
Linear, lognormal error
Orthogonal
Linear adjusted for person error
Linear adjusted for ambient error
Measures of Association
Correlation of averages
Correlation adjusted for measurement error
Average correlation of ambient with mean person
Average correlation of person with mean site
Fraction of variation explained by orthogonal regr.
P,
0.546
0.560
0.556
0.556
0.587
Pn
42.3
41.4
41.9
41.9
40.1
Value
0.955
0.974
0.944
0.633
0.984
Source: U.S. EPA reanalyses of data from Lioy et al. (1990).
surrogate. On average there were 3.5 persons per period and this number of individuals was
used in the analysis of variance shown in Table 7-46. The dichot monitoring data contained
little missing data, and so it was analyzed against the personal monitoring data for those days
with data. The results of the linear regression are in Table 7-47 and are shown graphically in
Figure 7-34. The individual exposure values were averaged so that they could be compared with
the site exposure values. These means are shown in Table 7-48. Note that the orthogonal
regression slope is larger than either of the linear regression slopes. Note also that the linear
regression slope adjusted for measurement error is larger than any of the other slopes.
7.6.3.5 Azusa, CA Data Set
The Azusa, CA data set for PM10 reported on by Wiener et al. (1990) was described earlier
in Section 7.4.1.1.1 and presented in Table 7-2la. The same regression analyses described
earlier in this section were performed on the 24-h cross-sectional data and the results are shown
in Table 7-49. A plot of the linear regression analysis, resulting in a
7-140
-------
TABLE 7-43. PERSONAL AND AMBIENT EXPOSURE
DATA FOR BEIJING, CHINA (mg/m3)
Day
1
2
2
2
2
2
3
3
3
4
4
4
4
4
4
5
5
5
5
5
5
5
Personal
0.13
0.15
0.10
0.12
0.23
0.14
0.11
0.09
0.09
0.31
0.12
0.13
0.35*
0.12
0.25
0.10
0.22
0.32
0.12
0.08
0.13
0.07
*The only personal value higher than
Source: World
Health Organization
Ambient
0.19
0.25
0.25
0.25
0.25
0.25
0.31
0.31
0.31
0.33
0.33
0.33
0.33
0.33
0.33
0.36
0.36
0.36
0.36
0.36
0.36
0.36
the ambient value.
(1985).
TABLE 7-44. RESULTS OF LINEAR
FOR THE BEIJING, CHINA
Day Personal Ambient
6
6
6
6
6
6
7
7
7
8
9
9
9
9
10
11
11
11
11
11
11
11
0.15
0.17
0.13
0.16
0.21
0.08
0.35
0.24
0.20
0.15
0.23
0.18
0.10
0.38
0.11
0.23
0.32
0.11
0.21
0.11
0.20
0.29
0.42
0.42
0.42
0.42
0.42
0.42
0.44
0.44
0.44
0.53
0.55
0.55
0.55
0.55
0.59
0.69
0.69
0.69
0.69
0.69
0.69
0.69
REGRESSION ANALYSIS
EXPOSURE
DATA
Linear regression analysis of average personal exposure versus ambient exposure
Y = intercept -+
Variable
Intercept
Slope
- slope X
Beta
0.116
0.142
Std. Error Beta
0.040
0.088
ANOVA Table
Source
Regression
Error
TOTAL
R-squared = 0.
Log-likelihood
Sum of Squares Mean Square Error
05925, r = 0.2434
= -46.95
0.0179
0.2835
0.3014
0.00893
0.00692
0.00701
D.F.
2
41
43
F-Value
1.2911
Source: U.S. EPA reanalyses of data from World Health Organization (1985).
7-141
-------
IB
C
O
a
a.
400-
300
200
100
• •
200 400
Ambient PIV|0
600
800
Figure 7-33. Plot of means of personal exposures and ambient PM10 from Beijing, China
and regression line calculated by U.S. EPA.
Source: U.S. EPA reanalyses of data from World Health Organization (1985).
TABLE 7-45. ESTIMATED MEAN VECTOR, COVARIANCE MATRIX,
AND CORRELATION MATRIX OF PERSONAL EXPOSURE
PM,n DATA FROM RIVERSIDE, CALIFORNIA (24-h,
Monitor
Means
Personal
109.9
Covariance/Correlation Matrix
Personal
Indoor
Backyard
Dichot
Wedding
PEM-SAM
1055.0
(0.849)
(0.725)
(0.707)
(0.721)
(0.736)
Indoor
79.9
Backyard
91.7
Dichot
71.2
Wedding
68.4
PEM-SAM
80.4
(Correlation below diagonal)
917.4
1107.6
(0.703)
(0.767)
(0.753)
(0.776)
1024.7
1017.9
1893.2
(0.821)
(0.832)
(0.858)
749.0
832.7
1165.6
1063.4
(0.956)
(0.989)
838.9
897.0
1296.9
1116.6
1282.8
(0.976)
913.7
987.4
1427.4
1232.9
1337.1
1462.3
Source: U.S. EPA reanalyses of data reported on by Pellizzari et al. (1992).
7-142
-------
TABLE 7-46. RESULTS OF AN ANOVA ANALYSIS OF
THE PERSONAL EXPOSURE DATA OF RIVERSIDE, CALIFORNIA
Source of Variation
period
residual
Total
D.F.
46
114
160
S.S.
167,400
275,000
442,400
M.S.
3640
2412
Source: U.S. EPA reanalyses of data reported on by Pellizzari et al. (1992).
TABLE 7-47. RESULTS OF THE ANALYSIS OF THE
EXPOSURE DATA FROM RIVERSIDE, CALIFORNIA
Regression Model
Linear, normal error
Linear, lognormal error
Orthogonal
Linear adjusted for person error
Linear adjusted for ambient error
Measures of Association
P,
0.6174
0.6185
0.8071
0.9675
(Not applicable)
Po
59.7
57.1
44.2
31.0
Value
Correlation of averages
Correlation adjusted for measurement error
Fraction of variation explained by orthogonal regr.
0.721
(Not applicable)
0.864
Source: U.S. EPA reanalyses of data reported on by Pellizzari et al. (1992).
negative slope, is shown in Figure 7-35. There clearly is no relationship between the pooled
PEM and SAM variables for this data set. The statistical explanation for the negative correlation
and slope (PEM decreases with increasing SAM) is that one of the observations (PEM = 273
|ig/m3, SAM = 48 |ig/m3, for House 9, Day 10, person 1, as shown in Table 7-2la) is an outlier
(273 |ig/m3 > mean + 3*SD). Removal of this single datum point changes both the correlation
and the slope to slightly positive values of similar magnitude. Because of the insignificance of
the slope and correlation, further adjustments for measurement error do not make sense.
7-143
-------
200
150
100
50
50 100
Ambient SAM
150
200
Figure 7-34. PTEAM mean 24-h PM10 data compared for personal PEM and SAM.
Source: U.S. EPA reanalyses of data reported on by Pellizzari et al. (1992).
7.6.4 Discussion of Statistical Analyses: Mean PEM Versus Mean SAM
The Beijing study had an insignificant positive slope and the Azusa study gave an
estimated slope less than zero that becomes insignificant positive with the removal of one
outlier. Possible explanations for the low slope of the Beijing study may be related to the
unusually low ratio of PEM to SAM of order 0.4. Either the SAM PM3 5 monitor that was used
may have been influenced by a local PM source, and thereby was not representative of the
Beijing locality where the subjects worked and lived, or the air exchange between indoors and
outdoors during the winter period was greatly minimized for personal comfort.
In the Beijing dataset of 44 pairs of simultaneous SIM and SAM (Table 7-43) only one
PM3 5 PEM value was greater than SAM, as opposed to Azusa where in the 50 pairs of
simultaneous SIM and SAM (Table 7-2Ib) only six PM2 5 PEM values were less than SAM. On
a day where SAM PM3 5 reached 690 |ig/m3 in Beijing, seven simultaneous PEM values all
ranged between 110 |ig/m3 and 320 |ig/m3. In relation to Figure 7-16, these PEM/SAM ratios
between 0.16 and 0.45 correspond to low air exchange rates of order 0.1 to 0.3 air changes per
hour. In the tightly-sealed poorly-heated building where all the subjects worked
7-144
-------
TABLE 7-48. AVERAGE 24-HOUR PM10 PERSONAL EXPOSURE DATA
COMPARED WITH THE PEM-SAM SITE EXPOSURE DATA
FOR RIVERSIDE, CALIFORNIA
Period
1
3
5
7
9
11
15
17
19
21
23
25
27
29
31
37
39
41
43
47
49
51
53
57
59
61
63
65
67
69
71
73
75
77
79
81
83
85
87
89
91
93
95
Average Personal
48.3
83.6
108.6
88.3
68.3
121.0
68.2
95.8
102.5
116.8
160.5
97.7
72.2
107.6
103.0
165.3
144.4
135.6
168.2
173.8
144.9
65.0
76.7
110.9
78.4
136.1
103.1
142.4
163.6
153.7
144.2
150.6
125.4
112.1
63.7
67.5
102.2
92.0
100.0
88.9
113.0
82.4
97.3
PEM-SAM Site
35.1
41.7
56.9
64.1
51.7
55.8
56.0
69.1
92.0
108.2
126.4
79.4
60.7
52.9
87.4
66.8
106.2
138.5
107.5
175.9
112.9
77.9
42.8
17.6
46.7
61.1
78.4
77.9
127.6
150.4
147.4
166.4
139.6
59.2
42.7
61.4
75.8
35.7
65.3
75.3
122.7
48.8
57.1
Source: U.S. EPA-calculated 24-h averages, based on 12-h data reported on by Pellizzari et al. (1992).
7-145
-------
TABLE 7-49. RESULTS OF THE LINEAR REGRESSION
ANALYSIS OF THE EXPOSURE DATA FROM AZUSA, CALIFORNIA
Variable
Beta
Std. Error Beta
Intercept
Slope
119.1
-0.054
13.77
0.201
Covariance Matrix of Parameter Estimates
Intercept
Slope
Log-likelihood = -263.4
Intercept
189.7
-2.543
ANOVA Table
Source: U.S. EPA reanalyses of data reported on by Wiener et al. (1990).
400
UJ
Q.
Q.
«
c
o
2
300
200
100
:r
Slope
-2.543
0.040
Source
Regression
Error
TOTAL
R-squared = 0.0015
Sum of Squares
111.2
76590
76700
Mean Square Error
55.6
1531.8
1475.1
D.F.
2
50
52
F-Value
0.0363
50 100 150
PM10 Ambient SAM |jg/nri
200
Figure 7-35. Plot of ambient and personal monitoring PM10 data from Azusa, CA and
calculated (slightly negative slope) regression line, which becomes positive if
single outlier value (*•) is deleted.
Source: U.S. EPA reanalyses of data reported on by Wiener et al. (1990).
7-146
-------
during the Beijing winter, a small variation in air exchange could result in a relatively large
difference in the indoor PM, which would result in PEM that appears to be uncorrelated with
SAM. If a contribution of PM generated by personal activity and ETS is subtracted from the
PEM values then the estimated air exchange rates would be even lower. The remaining
discussion will be based on the other three studies, realizing that the discussion is not supported
by these two studies.
The major conclusions which can be reached from the remaining three studies are as
follows.
(1) The average of several ambient monitors correlates better with mean personal
exposure than does an individual site (as would be predicted by the Central Limit
Theorem).
(2) The average of several personal monitors correlates better with mean ambient
exposure than does the ensemble of individual monitors.
(3) There is no evidence of the existence of a maximum (ceiling) correlation between
personal and ambient measurements. The only study with fixed multiple (n > 2)
ambient SAM locations and multiple personal monitors is the Phillipsburg, NJ, study.
The estimated correlation adjusted for measurement error was 0.97. The true
(unknown) correlation between an infinite average of personal monitors with an
infinite average of fixed site monitors may be different (smaller) in other locations,
but we do not have the data to evaluate that.
(4) The correlation coefficient is probably the best measure of association between
personal and ambient measurements. It can be used independent of the regression
technique or model and does not assume a distributional form. The "percent of
variation explained" as derived from orthogonal regression is not comparable to any
measure used for other models.
(5) The choice of a model (linear, linear with lognormal error, orthogonal) makes less
difference than the adjustment for measurement error.
(6) Based on the results of the Phillipsburg, NJ, analysis, one or more fixed site monitors
can do an excellent job of predicting the average of all personal exposures (if they
could be measured) even though the prediction for most individual exposures is quite
poor. This is also supported by the Tokyo, Japan, data set (Tamura et al., 1996). The
other data sets did not provide adequate information to either confirm or deny this
conclusion.
The value of the improvement of the mean PEM relationship to SAM is that it provides a
better visualization that helps in understanding how mean PEM varies with SAM. It thus
provides a measure of the validity of the use of a daily PM SAM as a surrogate for the mean PM
7-147
-------
PEM in the community for nonsmokers. It is clear that the uncertainty in predicting mean
personal exposure PM is much smaller than the uncertainty in predicting the personal exposure
PM for a nonsmoking individual when we note that the means have a much smaller variability
about the line as shown in Figures 7-31, 7-32, and 7-34.
There appears to be two distinct categories of cross-sectional exposure studies that were
examined: In the first type of study, such as Lioy et al. (1990), Clayton et al. (1993), and
Tamura et al. (1996), there is a significant R2 between individual PM PEM and PM SAM. In
this category, there is an appreciable improvement in correlation between the mean PEM and
SAM. It has been suggested that these cases with higher correlation of PEM PM with SAM PM
may arise where the fine portion of the ambient PM (PM2 5) is highly variable from day-to-day,
and the ambient coarse fraction is relatively constant (Wilson and Suh, 1995). In an urban area,
the fine particle composition and the fine particle concentration are often highly correlated from
site-to-site on any given day. This is due, in part, to the gas phase reactions of SOX and NOX,
associated with regional sources, to produce sulfates and nitrates in the submicron range.
Because of the long residence times of these species due to their negligible deposition velocities,
they are well mixed throughout the air mass (Suh et al., 1995; Burton et al., 1996).
On the other hand, ambient coarse particles are generated locally, and they have higher
deposition velocities than the fine particles. Their impact may then be limited by fallout to a
locality downwind of their emission point, as they are not readily transported across an urban
area. Therefore, during an air pollution episode, people living in an urban area may be exposed
to fine PM of similar chemical composition and concentrations, whereas they will be exposed to
coarse PM of ambient origin with a chemical composition that can depend on the location of the
exposure. Because ambient PM penetrates readily into a nonambient setting, the correlation
between the mean PM2 5 PEM and PM2 5 SAM would be high because all the people would have
similar exposure to the ambient fine PM - plus exposure to indoor generated PM2 5 which may
have less fluctuation in the absence of smoking.
In the second type of study, such as Sexton et al. (1984), Spengler et al. (1985), and
Wiener et al. (1990), there is negligible correlation between individual PEM PM and SAM PM,
and consequently there will be little correlation between their mean PEM and the SAM. In these
cases, if the fine fraction is not an appreciable portion of the ambient PM, or there are significant
7-148
-------
indoor sources, then the correlations between mean PM PEM and PM SAM will be lower and
possibly not significantly different from zero.
7.7 IMPLICATIONS FOR PARTICULATE MATTER AND MORTALITY
MODELING
PM related mortality may be specific to the most highly susceptible portion of the
population. Such a cohort may be the elderly people with the most serious chronic obstructive
pulmonary disease (COPD) and cardiac insufficiency. Smithard (1954) relates the findings of
Dr. Arthur Davies (Lewisham coroner) who autopsied 44 people who died suddenly during the
1952 London Fog:
"The great majority of deaths occurred in people who had pre-existing heart and lung
trouble, that is to say they were chronic bronchitic and emphysematous people with
consequent commencing myocardial damage. The suddenness of the deaths, Dr. Davies
thought, was due to a combination of anoxia and myocardial degeneration resulting in
acute right ventricular dilatation."
Mage and Buckley (1995) hypothesized that these people with compromised cardio-
pulmonary systems may be relatively inactive, while selecting to live in homes or institutional
settings without sources of indoor pollution. When their time is spent in clean settings (e.g.
where smoking is prohibited), they would have little exposure to PM other than from the
ambient pollution that intrudes into their living quarters (Sheldon et al., 1988a,b). The exposure
to PM of this cohort, would be highly correlated with PM SAM, and so would be their mortality,
if this ambient PM was reactive in their pulmonary tracts as described by West (1982).
However, there have been no results reported of an exposure study done on people with COPD
who correspond to the Lewisham mortality cohort. The cohort of five elderly housewives and
two male retirees in Tokyo (Tamura et al., 1996) may come close to this susceptible cohort.
Individual PM PEM of people outside these cohorts, who could be relatively insensitive to
ambient PM, might not be significantly correlated with PM SAM, as reported in most of the
other studies of nonsmokers cited in Table 7-26. This suggests a model to relate PM and
mortality as follows. Let any person (j) on a given day have a probability of mortality, p(m) = kj
Xj, where kj is the unit probability of mortality per |ig/m3 of PM per day, Xj is the daily average
exposure to PM, |ig/m3, independent of kj. Let us assume that each individual (j) has their own
7-149
-------
personal value of kj that can vary from day-to-day with changes in their respiratory health, such
as a transient pulmonary infection (West, 1982).
The expectation of total mortality (M) in a community of size N can be shown to be the
summation of k X over all individuals (j = 1 to N) as follows:
M = SkjXj (7-31)
If kj is independent of Xj, then we can define K as (1/N) H kj, and the mean community exposure
xas (1/N) H Xi, and it follows
M = NKX (7-32)
This implies that, given a linear relationship of mortality with PM PEM exposure (X) as
assumed in most studies discussed in Chapter 12, the expected mortality is proportional to the
mean community personal exposure to PM. The individual in the community, on any given day,
with the highest probability of dying from a PM exposure related condition is that individual
with the highest product kj Xj, not necessarily the highest exposed individual with the maximum
value of Xj (West, 1982).
The Phillipsburg, NJ, data set is a case in point. In this study, three subjects had
excessively high PEM PM. These values were caused by a hobby involving welding in a
detached garage (971 jig/m3), a home remodeling activity (809 |ig/m3) and usage of an unvented
kerosene heater (453 |ig/m3). Excessive PM generating activities are not expected of elderly
people who may have compromised pulmonary systems. In fact, the elderly and infirm husband
of the remodel er had a personal exposure of 45 |ig/m3 on the day of the remodeling activity. The
indoor monitors in the homes of the welder and remodel er only recorded 55 |ig/m3 and 19 |ig/m3,
respectively, during those events, indicating the specificity of the high exposure to only the
individual involved. These three outliers were removed from the analysis and were replaced by
the procedure for missing data of section 7.6.2.1, which estimates their exposures as if they had
not done those specific activities responsible for their noncharacteristic exposures (see Table 7-
37). This procedure is reasonable, since it is unlikely that these activities would be performed
by individuals with pulmonary conditions similar to those of the Lewisham mortality cohort
7-150
-------
(Smithard, 1954). As shown on Table 7-42 and Figure 7-32, the regression improves markedly
to a value of R2 = 0.914.
It is this relation of the average PM PEM exposure to PM SAM concentration, as shown in
Figure 7-32 that may be a better representation of the true situation underlying the PM vs
mortality relationships because of the "healthy worker" effect. Chronically ill people who are
sensitive to PM might change their behavior to minimize their exposure to irritants.
Consequently, healthy people with high PEM PM measures in occupations and indoor settings
can cause the regression R2 between PEM and SAM for nonsmokers to be low, but they may not
be the individuals at highest risk of the acute effects of PM exposure.
7.7.1 Relative Toxicity of Ambient Particulate Matter and Indoor Particulate
Matter
In the previous sections the SAM PM was evaluated as a predictor of PEM PM of
nonsmokers on the implied basis that the health effects of PM were only mass dependent, and
independent of chemical composition. It was shown in Table 7-26 that many early PM studies
of PEM had a low correlation between PEM and SAM on a cross-sectional basis that was often
not significantly different from zero. But, in the later studies (Tamura et al., 1996; Lioy et al.,
1990), a significant relationship was observed between PEM and SAM on an individual basis.
Further analysis showed that on a daily basis, SAM would appear to be a good predictor of mean
community exposure to ambient PM10 of nonsmoke exposed people from the results of the
Tokyo, Japan; Riverside, CA; and Phillipsburg, NJ; studies. However, there can be a large
difference in toxicity of PM per unit mass which is related to the chemical composition,
solubility and size of the particles. For example, mercury (Hg) and arsenic (As) have
significantly different toxicities in their inorganic and organic forms. Hexavalent chromium
(Cr) is more toxic than trivalent Cr. Anthropogenic PM, from combustion of fossil fuels, is
much more toxic than PM of natural origin (Beck and Brain, 1982; Mage et al., 1996). Fine
urban particulate matter generated by coal smoke during the 1952 London Fog at concentrations
of order 2,000 //g/m3 caused thousands to die (Holland et al., 1979; United Kingdom Ministry of
Health, 1954) but 2,000 Mg/m3 of soil dust from dust storms (Hansen et al., 1993) would not
have been as deadly.
7-151
-------
Soil constituents that are tracked-in to a home on shoes, and are subsequently resuspended,
contribute to the personal cloud (Roberts et al., 1990; Thatcher and Layton, 1995). "Even if this
crustal PM is relatively inert, its presence in the lung potentiates the toxicity of the
anthropogenic particles because it increases the residence time of the more toxic PM (WHO,
1995)" (Mage et al., 1996). This increase in soil constituents was also shown in the PTEAM
study (Ozkaynak et al., 1996) on Figure 7-22 "by observation that nearly all [soil] elements were
elevated in personal samples" but sulfur, which is in the ambient fine mode, was not a personal
cloud constituent. This is consistent with the observations of Wilmoth et al. (1991) that
"extremely small particles (below two micrometers) require local airflow (sampling) velocities
near 100 miles per hour [45 m/s] to overcome surface attraction forces and dislodge [them] for
sampling".
Figure 7-36 shows an example of resuspension of Pb in a Denver, CO, home
(Moschandreas et al., 1979). During the one-week sample, a wind shift brought a clean air mass
to below 0.01 //g/m3. In this time period, the average indoor Pb dropped from 0.085 to
0.048 //g/m3. The residual 0.048 //g/m3 represents the effect of resuspension by human activity.
When the wind shifted again, and ambient Pb rose to 0.360 //g/m3 the indoor Pb rose to
0.180 //g/m3. Note the peaks in the indoor concentration of Pb up to and above 0.10 //g/m3
during the clean air period which are indicative of variations in resuspension by human
activities.
There is also some indication in laboratory animal studies, using transpleural
catheterization and intratracheal instillation, that products of fossil fuel combustion are more
acutely toxic to animals than wood smoke and soil constituents (U.S. Environmental Protection
Agency, 1982, Table 12-6; Beck and Brain, 1982). Although these laboratory animal studies
may have no direct relation to toxicity in humans, they provide an indication of their relative
toxicity in animals when administered by those two routes.
In summary, there is evidence that not all PM constituents have the same toxicity per unit
mass. These differences are due to differences in aerodynamic diameter and chemical
composition. As shown on a Venn diagram (Figure 7-37, Mage [1985]), the focusing of the
description of a PM10 exposure increases the ability to estimate the potential toxicity of the
exposure. In the sequential description given below, the uncertainty in the toxicity of the
mixture is decreased as more information is provided.
7-152
-------
0.90
o
50.10
*J
C
0)
o
C
o
o
0.01
Pb Outdoors
V
0.90
O
20.10
o
o
40 80 120 160
Time, hours
0.01
Pb Indoors
0.90
0.10
40 80 120 160
Time, hours
0.01
Figure 7-36. Comparison of indoor and outdoor concentrations of lead in a home in
Denver, October 1976, for 1 week, starting at 1600 h. Mean values are given
by horizontal bars.
Source: Moschandreas et al. (1979).
1. 2 |ig/m3 of PM
10-
2. 2 |ig/m3 of PM10 in the size interval 2 to 2.5 jim.
3. 2 |ig/m3 of PM10 in the size interval 2 to 2.5 |im, 50% of automotive origin and 50% of
indoor source origin.
4. 2 |ig/m3 of PM10 in the size interval 2 to 2.5 |im, 50% of automotive origin and 50% of
indoor source origin, 0.5 |ig/m3 of Pb, 0.5 |ig/m3 of BaP and 1 |ig/m3 of unspecified
inorganic material.
As applied to human exposure to PM, this concept of differential toxicity suggests that data
collections might benefit by providing data that would allow the toxicity of a PM exposure to be
evaluated in terms of chemical information, in addition to the mass collected per unit volume.
7-153
-------
Figure 7-37. Venn diagram (Mage, 1985) showing focusing of information to more
completely specify toxicity of a given PM mixture: (1) universe of all possible
mixtures of PM with concentration of 2 /zg/m3; (2) subuniverse of all
combinations of PM with concentration of 2 Mg/m3 in size interval 2.0 to 2.5
jum; (3) subuniverse of all combinations of PM with concentration of 2 Mg/m3
in size interval 2.0 to 2.5 /j,m AD with 50% of automotive origin and 50%
from indoor sources; and (4) subuniverse of all combinations of PM with
concentration of 2 jUg/m3 in size interval 2.0 to 2.5 /j,m AD with 50% of
automotive origin and 50% from indoor sources; 25% Pb, 25% BaP and
50% unspecified inorganic materials.
7.7.2 Summary: Linkage of Ambient Concentrations of Particulate Matter
to Personal Exposures to Particulate Matter
As described by Wilson and Suh (1995), total exposure to ambient PM (Xae) of any given
size range is equal to the summation of exposures to ambient PM over both ambient (Xa) and
nonambient (X^) microenvironmental conditions. Total exposure to PM is equal to Xae plus
exposure to nonambient PM concentrations generated independently of personal activities
and nonambient PM concentrations generated dependency on personal activities (X^) which
7-154
-------
may correspond to smoking and the personal cloud effect. For a period (T) of constant ambient
PM a subject spends time Ta outdoors and time (T - TJ in n different nonambient
microenvironments. The total exposure to ambient PM can be expressed as:
Xae = [Ta ^ + (|"Ta) XnJ (7-33)
For a nonambient microenvironment, the equilibrium concentration of ambient particles in
it will be equal to
V D 3
A, r a
Xna = — (7-6)
(3 + fc)
where P = penetration fraction of PM in the ambient air entering the nonambient
microenvironment,
a = air exchange rate, h"1
k = deposition rate (a function of AD), h"1.
As discussed in section 7.2, the penetration factor/1 is virtually equal to 1 for all particles
less than 10 jim (Thatcher and Layton, 1995) and the fraction of Xna/Xa is as shown on Figure 7-
16. Combining equations 7-33 and 7-6, we obtain
x,. = . (7.34)
where T - Ta = S tj3 total time spent indoors,
j = 1 to n, index of indoor microenvironment visited.
Defining z as the overall ratio of exposure to ambient PM (X^) to the ambient
concentration (Xa), so that Xae = z Xa, letting _y = Ta/T, the fraction of time the subject is
outdoors, we obtain the average relation,
z =y + (l -y) (—^—] , (7-35)
( a + k
7-155
-------
where is a time weighted average .
\ a + k
As shown on Figure 7-38, on a daily basis, z can vary by an appreciable amount by
spending a fraction (y) of time outdoors. For^ = 1/3 (8 h), exposures to fine ambient PM25
increase by 100% for people living in homes with an air exchange rate a = 0.1 h"1.
The total exposure (X) can now be written as,
(7-36)
where S [(X^ + (X^] tj / T = P, the personal exposure increment produced by sources that do
not influence the ambient concentration as measured by a stationary ambient monitor (SAM).
Simplifying, we can rewrite Equation 7-36 as,
X = z Xa + p (7-37)
which gives a physical significance to the slope and intercepts of the regressions of PEM (X)
versus SAM (Xa) as discussed in Section 7.6.
The values of z, which depend on y, a, k and P can be determined from their independent
measurements described previously. P = 1 for all PM < 10 jim A.D. (Thatcher and Layton,
1995) andy = 0.074 [U.S. mean fraction of time spent outdoors per day; U.S. Environmental
Protection Agency (1989)]. From PTEAM (Wallace et al., 1993), a = 0.9 h"1 as a median value
for night and day. Ozkaynak et al. (1993a,b) have determined values for k as follows:
ForsulfateA:=0.16h"1
For PM25£= 0.39 h'1
ForPM10£= 1.01 h'1
7-156
-------
y = 1, fine and coarse
•o
0
in
o
Q.
X
0
(0
0)
n
Q.
i_
O
O
•o
+J
3
O
*-
O
o
m
Air exchange rate (air changes per hour)
Figure 7-38. Fraction of ambient PM to which people are exposed (z) as a function of
fraction of time outdoors (y) and air exchange rate for fine (PM2 5) and
coarse (PM10 - PM2 5) particles.
From the equation z = y + (1 -y) P a/(a + k)
for sulfate, z = 0.074 + 0.926 (0.9)7(0.9 + 0.16) = 0.859
for PM2 5 it is z = 0.074 + 0.926 (0.9)7(0.9 + 0.39) = 0.720
for PM10 it is z = 0.074 + 0.926 (0.9)7(0.9 + 1.01) = 0.512
These predicted values match closely to the reported values of z cited in this Chapter 7 as
follows:
Suh et al. (1993) report z = 0.87 ± 0.02 (r2 = 0.92) for SO4=
Tamura et al. (1996) [Table 7-32] report z = 0.466 (r2 = 0.905) for PM10,
Lioy et al.(1990) [Table 7-44] report z = 0.546 (r2 = 0.91) for PM10
It is not known what the average values of_y and a were for the State College, PA, and
Phillipsburg, NJ, cohorts of Suh et al. (1995) and Lioy et al. (1990), or the Tokyo, Japan, cohort
of Tamura et al. (1996). Therefore these results can only be considered as tentative at this time.
7-157
-------
The parameter B in Equation 7-37 represents the contribution to personal exposures (PEM)
from nonambient sources both independent of and dependent on personal activities. In general
the composition of the PM emitted by indoor sources (or resuspended by human activity) that
influence B will be different from the PM emitted into the ambient atmosphere from sources
controlled by State Implementation Plans (SIP)s. The nonambient |iE emissions are from the
activities of the subject (cooking, heating, smoking, resuspension of housedust, hobbies, etc.) or
independent activities of others in the same jiE that are independent of the ambient concentration
(Xa).
For the situation in Tokyo (Tamura and Ando, 1994; Tamura et al., 1996) the PM10 PEM
vs PM10 SAM correlation is good for all subjects individually, as well as their average PEM,
because the data were collected in a manner to minimize B. These data for the seven nonsmoke
exposed elderly subjects were culled to remove observations which were influenced by overt
particle generating activities such as visitors' smoking, burning of incense, and burning of
antimosquito coils. The custom of taking off shoes on entry into Japanese residences and use of
"tatami" mat flooring minimized resuspension of PM less than 10 |im AD, although indoor
activity did raise dust above 10 jam AD (Tamura et al., 1996).
For the U.S. cities of Phillipsburg, NJ, and Riverside, CA, with large numbers of
observations, the correlations of PEM vs SAM for PM10 were significantly positive but less than
for Tokyo, Japan, possibly due to the passive smoking and house dust generation in the
Riverside, CA, and Phillipsburg, NJ, studies. Even so, in Riverside, CA, ambient sources
provided about 67% of PM10 mass measured indoors (Ozkaynak et al., 1996). Finally, the
results of the studies in Beijing, China, and Azusa, CA, gave positive correlations of PEM and
SAM that were not significantly different from zero (If one outlier is included in the Azusa
analysis, the PEM vs SAM correlation is negative). These low correlations may be due to low
air exchange rates in Beijing during the winter as evidenced by the low PEM/SAM ratios, and
the presence of indoor sources in Azusa, as evidenced by the PEM almost double the SIM or
SAM. These latter studies are typical of the results in other U.S. cities such as Kingston and
Harriman, TN (Spengler et al., 1985), where ambient pollution is relatively low, so that the
personal cloud and indoor source effects predominate.
In summary, it appears that the first exposure conclusion of the previous PM criteria
document (U.S. Environmental Protection Agency, 1982), quoted in section 7.1.3, has been
7-158
-------
generally supported by recent studies. If the relation of equation 7-35 which appears to predict
the observed relations in several studies cited in this document is a reasonable model of the
personal exposure to ambient PM, then that conclusion can be adjusted more specifically as
follows:
1. Long-term personal exposures to fine PM sulfates of outdoor origin may be estimated
by approximately 85% of the sulfate in the fine fraction of ambient PM.
2. Long-term personal exposures to PM < 2.5 jim A.D. of outdoor origin may be
estimated by approximately 70% of the PM < 2.5 jim A.D. in the ambient PM.
3. Long-term personal exposures to PM < 10 jim A.D. of outdoor origin may be estimated
by approximately 50% of the PM < 10 jim A.D. in the ambient PM.
These relationships still need to be validated in populations other than those from which
they were derived. Variations will exist for cohorts with different fractions of time spent
outdoors (y) and air exchange rates (a) than the values chosen for representing the national
averages.
Ambient concentrations of PM10 measured at properly sited monitoring stations are highly
uniform in urban areas (Burton et al., 1996, Suh et al., 1995), have no losses in penetration into
jiEs (Thatcher and Layton, 1995), and may be highly correlated with personal exposures to PM10
(Tamura et al., 1996) where indoor sources of PM10 are minimal. Even where indoor sources of
PM10 exist, they tend to produce different chemical species than those found in the PM2 5
fraction, as shown by the sulfates which do not appear in the personal cloud (Ozkaynak et al.,
1996; Suhetal., 1993).
It is therefore concluded that the presence of variable indoor sources of PM10 tends to
lower the observed correlations between PEM PM10 (derived from both ambient and nonambient
sources) and SAM PM10 (derived only from ambient sources) and even achieve values
nonsignificantly different from zero. Consequently, the use of an ambient concentration
of PM25 or PM10 in relation to daily changes of mortality and morbidity may be a reasonable
surrogate for the average personal exposure of people in the community to the PM2 5 or PM10
generated by ambient sources. "The consistently higher R2 values observed in the longitudinal
regressions support the epidemiological findings more strongly than the poor correlations noted
in the standard cross-sectional regressions" (Wallace, 1996), as per the U.S. EPA reanalyses
shown in Tables 7-36 and 7-42.
7-159
-------
7.8 SUMMARY AND CONCLUSIONS
For PM, the total exposure of an individual consists of the summation of the individual's
exposure to PM in a variety of microenvironments. This typically includes exposures while (a)
outdoors and (b) indoors (at-home or in microenvironments such as shops and public buildings;
at-work in an office or factory; and in a vehicle). The principle of superposition is a useful
mechanism to visualize the summation process. A simplification of this summation process for
an arbitrary individual, described in detail by Figure 7-30, is illustrated in Figure 7-39. In each
sub-figure (a to d) of Figure 7-39, the shaded area represents PM exposure (in jig-h/m3) of
ambient origin appropriately indexed by a central (community) monitoring station. The clear
area represents that PM exposure (in |ig-h/m3) the individual is exposed to which is not
characterized by the PM measured at the central monitoring station.
Figure 7-39a shows that while outdoors, the subject can be exposed to (a) widely dispersed
ambient PM that is represented by the community monitoring station and, independently, also to
(b) proximal PM that does not markedly influence the monitoring station reading (from tobacco
smoking, standing over a grill at a backyard barbecue, "personal cloud", etc.). For example, in
the PTEAM Study, backyard concentrations of PM25 and PM10 had a correlation on the order of
0.9 with a central monitoring station. Also, in Tokyo (Figure 7-25), outdoor concentrations
immediate to the homes of subjects studied by Tamura et al. (1996) had a correlation of 0.9 with
the local ambient monitoring station.
Figure 7-39b shows that, while indoors (not at work), the subject can be exposed not only
to (a) ambient PM (represented by the monitoring station) that infiltrates indoors but also to (b)
PM of indoor origin that does not influence the ambient monitoring station reading (from
smoking, cooking, vacuuming, "personal cloud", etc.). Obviously, the proportion of exposure to
PM of ambient origin versus that of indoor origin can vary widely, depending on: outdoor
concentrations of the ambient PM; the air exchange rate of indoor spaces; the
7-160
-------
£
£
O>
0.
o
(0
o
Q.
X
UJ
14 hours
1 hour
8 hours
1 hour
(a) Outdoors
Figure 7-39.
(b) Indoors
At-home, etc.
(Non-work)
(c) Indoors
At-work
(d) In-traffic
(e) Total
Exposure
Conceptual representation of potential contributions of PM of ambient
origin and PM generated indoors to total human exposure of a
hypothetical individual. The total personal exposure (e) of that
individual will consist of the sum of exposures to widely dispersed PM of
ambient origin (shaded areas) characterized by measurements at a
centrally-located community monitoring site and all other exposures
(non-shaded areas) to proximally generated particles either outdoors or
indoors in situations designated for (a), (b), (c), and (d). Times of
exposure in the various situations reflect typical time-action patterns for
U.S. adults. Depicted exposures to PM of non-ambient origin may vary
greatly from those shown there for qualitative impression only,
depending on various factors described in the text.
presence or absence of indoor PM sources; and the removal efficiency of indoor sinks for
specific constituents of the respective PM of ambient or indoor origin. In the absence of major
indoor PM sources (e.g., smoking), the percentage of total exposure contributed by PM of
ambient origin can be substantial. For example, as shown in Table 7-2, between 60% and 80%
of indoor air PM was estimated by source apportionment methods to be of ambient origin in
non-smokers' homes in two U.S. cities (Steubenville, OH; Portage, WI) included in the Harvard
Six-City Study. Even in smokers' homes, it was estimated that 60% of the non-smoking related
PM was of ambient origin in the same two cities. The New York State ERDA Study (see page
7-161
-------
7-23) also showed that, in homes without combustion sources, 60% of the total indoor PM2 5 was
from outdoor sources. For homes with smokers in the same study, about 66% of the non-
tobacco smoke indoor particles were found to be of ambient origin. Similarly, based on the
Tamura et al. (1996) data shown in Figure 7-24, it can be estimated that as much as 80% of the
measured indoor PM10 in Japanese homes without combustion sources was of ambient origin.
Figure 7-39c shows that while indoors at work the subject can also be exposed to
(a) ambient PM (represented by the community monitoring station) which infiltrates indoors,
and (b) PM of indoor origin that does not influence the monitoring station reading (from
smoking, welding, machining, "personal cloud", etc.). It can be expected that, for office-type
work, similar relationships as described above for the other indoor conditions (e.g., smokers' or
non-smokers' homes) would apply. However, for work conditions involving particle generation
(e.g., wood working, welding, mining, etc.), the personal exposure of "dusty-trade" workers to
indoor-generated particles can be several orders of magnitude greater than their exposure to
indoor particles of ambient origin.
Figure 7-39d shows that while in traffic, the subject can be exposed to (a) ambient PM that
is represented by the monitoring station (via ambient air infiltration into the vehicle), and (b) PM
of on-board or proximal vehicle origin that does not directly influence the community
monitoring station reading (from smoking, exhaust penetration from nearby vehicles, etc.). For
example, in one study, Morandi et al. (1988) found that the average concentration of PM3 5 in
motor vehicles in traffic (55 //g/m3) was 60% higher than the average outdoor PM3 5 level (35
Mg/m3).
Figure 7-39e is a simple rearrangement of the shaded and non-shaded areas to show that an
individual's total daily exposure (|ig-h/m3) can be thought of as the sum of two quantities: (a)
exposure to PM characterized by the local community monitoring station, and (b) exposure to
PM of immediately proximal origin that varies independently of the PM measured at the
monitoring station. Conceptually, everyone in the community will be exposed to the mix of PM
represented by the shaded area that is characterized by the local monitoring station, due to their
time outdoors and the penetration of PM into indoor microenvironments and vehicles. However,
not everyone in the community will be exposed to the identical mix of PM represented by the
clear area, because this exposure and its chemical composition is idiosyncratically related to their
individual habits and practices (smoking, home cleanliness, hobbies, "personal cloud", etc.),
7-162
-------
their occupation (home maker, student, office worker, welder, miner, etc.) and their mode and
usage of transportation (car, bus, train, etc.).
Evaluation of information useful in determining relative contributions of ambient (outdoor)
and non-ambient (indoor) particles to total human exposures leads to the following key
conclusions:
(1) For PM, the ambient environment can be a major source of indoor pollution due to air
exchange and infiltration. Whether the ambient is the dominant source of indoor PM
depends on the relative magnitude of indoor sources of PM.
(2) For PM of size fractions that include coarse particles, some studies have identified
statistically significant relationships between personal exposures and ambient
concentrations, while other studies have not, probably due to overwhelming effects of
indoor sources, "personal clouds" and other individual activities.
(3) Cross-sectional regressions of personal exposure on outdoor PM25 and PM10
concentrations generally explain less than 25% of the variance (R2 < 0.25). However,
longitudinal regressions for each person in the study (in those cases where the person
was measured repeatedly) often show much better relationships between personal
exposure and outdoor air concentrations.
(4) Personal exposures to outdoor-generated PM of any size fraction < PM10 can be
estimated from the fraction of time spent indoors and an estimate of the air exchange
rate and deposition rate associated with that size fraction.
(5) The relationship between ambient concentration and personal exposure is better for
finer size fractions of ambient PM, than for coarser PM. Higher correlations between
ambient concentration and personal exposures have been found for fine PM
constituents (such as sulfates) without indoor sources.
(6) For a study population of nonsmokers in which there is a significant positive
correlation between personal exposures and ambient concentrations, the ambient
concentration can predict the mean personal exposure with much less uncertainty than
it can predict the personal exposure of any given individual.
(7) For Riverside, CA, where 25% of the nonsmoking population was estimated to have
personal exposures on the day they were monitored that exceeded the 24-h National
Ambient Air Quality Standard for PM10 of 150 //g/m3, approximately 50% of this
mass was found to be of ambient origin.
(8) The personal exposure to PM of smokers is dominated by the milligram quantities of
PM inhaled with each cigar, pipe, or cigarette smoked.
(9) For the U.S. studies, almost all personal exposures to PM are greater than the ambient
concentrations.
7-163
-------
(10) The penetration factor from outdoors to indoors for both PM2 5 and PM10 was found
to be unity in the PTEAM and Thatcher and Layton (1995) studies.
(11) Deposition rates in indoor microenvironments for PM10 and its fine and coarse
fractions were determined in the PTEAM Study. Similar deposition rates were found
by Thatcher and Layton (1995). Deposition reduces exposure to ambient PM; coarse
mode PM is removed more rapidly than PM2 5, which is removed more rapidly than
sulfate.
(12) Under equilibrium conditions, residential indoor concentrations of outdoor-generated
PM of any size fraction < PM10 can be estimated for any given air exchange rate, by
employing the deposition rate associated with that size fraction.
(13) For PM, studies have detected a "personal cloud" related to the activities of an
individual who may generate significant levels of airborne PM in his/her vicinity
which may not be picked up by an indoor PM monitor at a distance.
(14) There is some evidence that nonsmoke-exposed elderly people have lower residential
indoor PM concentrations than the simultaneous ambient PM concentrations, as
opposed to the general population who have indoor PM concentrations comparable to
or greater than ambient PM concentrations.
(15) Measured indoor air concentrations of PM25 and PM10 generally exceed outdoor air
concentrations (often by a factor of two) except in areas where outdoor
concentrations are high (e.g., Steubenville, OH and Riverside, CA).
(16) Indoor concentrations are higher during the day than at night.
(17) Correlations between indoor and outdoor particle mass concentrations were not
significant in two of the three major studies reviewed. In the third (PTEAM) study,
they ranged between 0.22 and 0.54.
(18) Regressions of indoor on outdoor PM2 5 and PM10 concentrations generally explain
less than half of the variance (R2 < 50%) if the regressions are carried out
simultaneously on all homes in the study. However, regressions for a single home (in
those cases where homes were measured repeatedly) often have much better indoor-
outdoor relationships (R2 up to 90%). Since most epidemiological studies deal with
repeated measurements overtime, "longitudinal" regressions by individual home may
be more relevant to these studies than "cross-sectional" regressions across all homes.
(19) The largest identified indoor source of particles in both homes and buildings is
cigarette smoking. Homes with smokers have an ETS-related PM2 5 concentration
increment ranging between 25 and 45 //g/m3.
(20) The second largest identified indoor source of particles is cooking. Homes with
cooking had increased levels of PM10 on the order of 10 to 20 //g/m3.
7-164
-------
(21) Unknown indoor sources accounted for a substantial fraction (25%) of indoor
concentrations of both PM2 5 and PM10 in the PTEAM Study. These sources appear to
be due to personal activities, including resuspension of house dust.
(22) Variations in personal exposure due to fluctuations produced by indoor sources of
PM are independent of the variations in personal exposure produced by ambient
sources.
7-165
-------
REFERENCES
Adams, W. C. (1993) Measurement of breathing rate and volume in routinely performed daily activities [final report].
Davis, CA: University of California, Physical Education Department, Human Performance Laboratory;
contract no. A033-205.
Afifi, A. A.; Elashoff, R. M. (1966) Missing observations in multivariate statistics: I. review of the literature. J. Am.
Stat. Assoc. 61: 595-604.
Alberts, W. M.; Brooks, S. M. (1992) Advances in occupational asthma. Clin. Chest Med. 13: 281-302.
American Academy of Allery & Immunology. (1994) Aeroallergen Monitoring Network 1994 pollen and spore report.
Milwaukee, WI: American Academy of Allery & Immunology.
Anuszewski, J.; Larson, T. V.; Koenig, J. Q. (1992) Simultaneous indoor and outdoor particle light-scattering
measurements at nine homes using a portable nephelometer. Presented at: 11th annual meeting of the American
Association for Aerosol Research; October; San Francisco, CA. Cincinnati, OH: American Association for
Aerosol Research; paper no. 3A.5.
Arlian, L. G. (1989) Biology and ecology of house dust mites, Dermatophagoides spp. and Euroglyphus spp. Immunol.
Allergy Clm. North Am. 9: 339-356.
Aso, N; Yoshizawa, S.; Tang, H. (1993) Behavior of particles generated from floor. In: Kalliokosi, P.; Jantunen, M.;
Seppanen, O., eds. Indoor air '93: proceedings of the 6th international conference on indoor air quality and
climate, v. 4, particles, microbes, radon; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 43-48.
Baxter, C. S.; Wey, H. E.; Burg, W. R. (1981) A prospective analysis of the potential risk associated with inhalation of
aflatoxin-contaminated grain dusts. Food Cosmet. Toxicol. 19: 765-769.
Beck, B. D.; Brain, J. D. (1982) Prediction of the pulmonary toxicity of respirable combustion products from residential
wood and coal stoves. In: Proceedings: residential wood & coal combustion specialty conference; March;
Louisville, KY. Pittsburgh, PA: Air Pollution Control Association; pp. 264-280.
Bennett, K. M. (1983) [Letter to H. H. Hovey, Jr., New York State Department of Environmental Conservation,
regarding the national ambient air quality standards]. Washington, DC: U.S. Environmental Protection Agency,
Assistant Administrator for Air, Noise, and Radiation; March 18.
Binder, R. E.; Mitchell, C. A.; Hosein, H. R.; Bouhuys, A. (1976) Importance of the indoor environment in air pollution
exposure. Arch. Environ. Health 31: 277-279.
Burge, H. A. (1995) Bioaerosols in the residential environment. In: Cox, C. S.; Wathes, C. M., eds. Bioaerosols
handbook. Boca Raton, FL: CRC Press, Inc.; pp. 579-597.
Burton, R. M.; Suh, H. H.; Koutrakis, P. (1996) Spatial variation in particulate concentrations within metropolitan
Philadelphia. Environ. Sci. Technol. 30: 400-407.
Chan, C.-C.; Huang, S.-K.; Chen, Y.-C.; Wang, J.-D. (1995) Personal, indoor, and outdoor concentrations of respirable
suspended particulates (RSP) and nicotine in seven smoker's families in Taiwan. Atmos. Environ.: submitted.
Clayton, C. A.; Perritt, R. L.; Pellizzari, E. D.; Thomas, K. W.; Whitmore, R. W.; Wallace, L. A.; Ozkaynak, H.;
Spengler, J. D. (1993) Particle total exposure assessment methodology (PTEAM) study: distributions of
aerosol and elemental concentrations in personal, indoor, and outdoor air samples in a southern California
community. J. Exposure Anal. Environ. Epidemiol. 3: 227-250.
7-166
-------
Code of Federal Regulations. (1994) National primary and secondary ambient air quality standards; definitions. C. F. R.
40: part 50, §50.1.
Cohen, B. S.; Chang, A. E.; Harley, N. H.; Lippmann, M. (1982) Exposure estimates from personal lapel monitors. Am.
Ind. Hyg. Assoc. J. 43: 239-243.
Cohen, B. S.; Harley, N. H.; Lippmann, M. (1984) Bias in air sampling techniques used to measure inhalation exposure.
Am. Ind. Hyg. Assoc. J. 45: 187-192.
Colome, S. D.; Kado, N. Y.; Jacques, P.; Kleinman, M. (1990) Indoor-outdoor relationships of particles less than 10
,um in aerodynamic diameter (PM10) in homes of asthmatics. In: Indoor air '90: precedings of the 5th
international conference on indoor air quality and climate, volume 2, characteristics of indoor air; July-August;
Toronto, ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air Quality and Climate, Inc.;
pp. 275-280.
Colome, S. D.; Kado, N. Y.; Jaques, P.; Kleinman, M. (1992) Indoor-outdoor air pollution relations: particulate matter
less than 10 //m in aerodynamic diameter (PM10) in homes of asthmatics. Atmos. Environ. Part A 26:
2173-2178.
Costle, D. M. (1980) [Letter to J. Randolph, U.S. Senate, regarding EPA's definition of "ambient"]. Washington, DC:
U.S. Environmental Protection Agency, Administrator; December 19.
Coultas, D. B.; Samet, J. M.; McCarthy, J. F.; Spengler, J. D. (1990) Variability of measures of exposure to
environmental tobacco smoke in the home. Am. Rev. Respir. Dis. 142: 602-606.
Cox, C. S. (1987) The aerobiological pathway of microorganisms. New York, NY: John Wiley & Sons.
Cox, C. S.; Wathes, C. M., eds. (1995) Bioaerosols handbook. Boca Raton, FL: CRC Press, Inc.
Croft, W. A.; Jarvis, B. B.; Yatawara, C. S. (1986) Airborne outbreak of trichothecene toxicosis. Atmos. Environ. 20:
549-552.
Crouse, W. E.; Ireland, M. S.; Johnson, J. M.; Striegel, R. M., Jr.; Wilhard, C. S.; DePmto, R. M.; Oldaker, G. B., Ill;
McBride, R. L. (1989) Results from a survey of environmental tobacco smoke (ETS) in restaurants. In: Harper,
J. P., ed. Combustion processes and the quality of the indoor environment: transactions of an international
specialty conference; September 1988; Niagara Falls, NY. Pittsburgh, PA: Air & Waste Management
Association; pp. 214-222. (A&WMA transactions series: TR-15).
Cupitt, L. T.; Glen, W. G.; Lewtas, J. (1994) Exposure and risk from ambient particle-bound pollution in an airshed
dominated by residential wood combustion and mobile sources. In: Symposium of risk assessment of urban
air: emissions, exposure, risk identification, and risk quantitation; May-June 1992; Stockholm, Sweden.
Environ. Health Perspect. 102(suppl. 4):75-84.
Daisey, J. M.; Spengler, J. D.; Kaarakka, P. (1987) A comparison of the organic chemical composition of indoor
aerosols during woodburning and non-woodburning periods. In: Seifert, B.; Esdorn, H.; Fischer, M.; Rilden,
H.; Wegner, J., eds. Indoor air '87: proceedings of the 4th international conference on indoor air quality and
climate: v. 1, volatile organic compounds, combustion gases, particles and fibres, microbiological agents;
August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of Germany: Institute for Water, Soil
and Air Hygiene; pp. 215-219.
7-167
-------
Dempster, A. P.; Laird, N. M.; Rubin, D. B. (1977) Maximum likelihood from incomplete data via the EM algorithm. J.
R. Stat. Soc. Ser. B 39: 1-38.
Diemel, J. A. L.; Brunekreef, B.; Boleij, J. S. M.; Biersteker, K.; Veenstra, S. J. (1981) The Amhem lead study: II.
indoor pollution, and indoor/outdoor relationships. Environ. Res. 25: 449-456.
Dockery, D. W.; Spengler, J. D. (1981a) Indoor-outdoor relationships of respirable sulfates and particles. Atmos.
Environ. 15: 335-343.
Dockery, D. W.; Spengler, J. D. (1981b) Personal exposure to respirable particulates and sulfates. J. Air Pollut. Control
Assoc. 31: 153-159.
Dockery, D. W.; Schwartz, J.; Spengler, J. D. (1992) Air pollution and daily mortality: associations with particulates
and acid aerosols. Environ. Res. 59: 362-373.
Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
Duan, N.; Mage, D. T. (1996) Combination of direct and indirect approaches for exposure assessment. J. Exposure
Anal. Environ. Epidemiol.: accepted.
Dunn, O. J.; Clark, V. A. (1974) Applied statistics: analysis of variance and regression. New York, NY: John Wiley &
Sons; pp. 221-251. (Wiley series in probability and mathematical statistics).
Edmonds, R. L., ed. (1979) Aerobiology: the ecological systems approach. Stroudsburg, PA: Dowden, Hutchinson &
Ross, Inc. (US/IBP synthesis series 10).
Federal Trade Commission. (1994) Tar, nicotine, and carbon monoxide of the smoke of 933 varieties of domestic
cigarettes. Washington, DC: Federal Trade Commission.
Fletcher, B.; Johnson, A. E. (1988) Comparison of personal and area concentration measurements and the use of a
manikin in sampling. In: Vincent, J. H., ed. Ventilation '88. New York, NY: Pergamon Press; pp. 161-165.
Fugas, M.; Wilder, B.; Paukovic, R.; Hrsak, J.; Steiner-Skreb, D. (1973) Concentration levels and particulate size
distribution of lead in the air of an urban and an industrial area as a basis for the calculation of population
exposure. In: Barth, D.; Berlin, A.; Engel, R.; Recht, P.; Smeets, J., eds. Environmental health aspects of lead:
proceedings [of an] international symposium; October 1972; Amsterdam, The Netherlands. Luxembourg:
Commission of the European Communities; pp. 961-968; report no. EUR 5004 d-e-f
Fuller, W. A. (1987) Measurement error models. New York, NY: John Wiley & Sons.
Gregory, P. H. (1973) The microbiology of the atmosphere. 2nd ed. New York, NY: John Wiley & Sons.
Groves, W. A.; Hahne, R. M. A.; Levine, S. P.; Schork, M. A. (1994) A field comparison of respirable dust samplers.
Am. Ind. Hyg. Assoc. J. 55: 748-755.
Guerin, M. R.; Jenkins, R. A.; Tomkins, B. A. (1992) The chemistry of environmental tobacco smoke: composition and
measurement. Chelsea, MI: Lewis Publishers.
Hansen, A. D. A.; Kapustin, V N; Kopeikin, V. M.; Gillette, D. A.; Bodhaine, B. A. (1993) Optical absorption by
aerosol black carbon and dust in a desert region of central Asia. Atmos. Environ. Part A 27: 2527-2531.
Hasabelnaby, N. A.; Ware, J. H.; Fuller, W. A. (1989) Indoor air pollution and pulmonary performance: investigating
errors in exposure assessment. Stat. Med. 8: 1109-1126.
7-168
-------
Heavner, D. L.; Morgan, W. T.; Ogden, M. W. (1995) Determination of volatile organic compounds and ETS
apportionment in 49 homes. Environ. Int. 21: 3-21.
Highsmith, V. R.; Hoffman, A. J.; Zweidmger, R. B.; Cupitt, L. T.; Walsh, D. (1991) The IACP: overview of the Boise,
Idaho, and the Roanoke, Virginia, field studies. Presented at: 84th annual meeting and exhibition of the Air &
Waste Management Association; June; Vancouver, BC, Canada. Pittsburgh, PA: Air & Waste Management
Association; paper no. 91-131.1.
Hodges, S. D.; Moore, P. G. (1977) Data uncertainties and least squares regression. Appl. Stat. 44: 185-195.
Holcomb, L. C. (1993) Indoor air quality and environmental tobacco smoke: concentration and exposure. Environ. Int.
19: 9-40.
Holland, W. W.; Bennett, A. E.; Cameron, I. R.; Florey, C. du V.; Leeder, S. R.; Schilling, R. S. F.; Swan, A. V;
Waller, R. E. (1979) Health effects of particulate pollution: reappraising the evidence. Am. J. Epidemiol. 110:
525-659.
Ingham, D. B.; Yan, B. (1994) The effect of a cylindrical backing body on the sampling efficiency of a cylindrical
sampler. J. Aerosol Sci. 25: 535-541.
Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1995) A relation between personal and ambient PM10.
Epidemiology 6(suppl): S45.
Jenkins, R. A.; Palausky, M. A.; Counts, R. W.; Guerm, M. R.; Dmdal, A. B.; Bayne, C. K. (1995a) Determination of
personal exposure of nonsmokers to environmental tobacco smoke in the United States. Lung Cancer:
submitted
Jenkins, R. A.; Palausky, M. A.; Guerin, M. R.; Dindal, A. B.; Counts, R. W.; Bayne, C. K. (1995b) Personal exposure
to environmental tobacco smoke in workplace and away-from work settings: a 16 city case study. In:
Measurement of toxic and related air pollutants, 1995 abstracts book of an EPA/A&WMA international
symposium; May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste Management Association; pp.
40-41.
Kamens, R.; Lee, C.-T.; Wiener, R.; Leith, D. (1991) A study to characterize indoor particles in three non-smoking
homes. Atmos. Environ. Part A 25: 929-948.
Kendall, M. G.; Stuart, A. (1961) The advanced theory of statistics, v. 2: inference and relationship. London, United
Kingdom: Charles Griffin & Co., Ltd.; pp. 378-415.
Kendrick, B. (1992) The fifth kingdom. 2nd ed. Newburyport, MA: Focus Information Group.
Kim, Y. S.; Stock, T. H. (1986) House-specific characterization of indoor and outdoor aerosols. Environ. Int. 12:
75-92.
Klepeis, N. E.; Ott, W. R.; Switzer, P. (1994) A Total Human Exposure Model (THEM) for Respirable Suspended
Particles (RSP). Presented at: 87th annual meeting & exhibition of the Air & Waste Management Association;
June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-WA75A.03.
Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics; contract
nos. 68-D9-0166 and 68-D3-0013.
Koutrakis, P.; Briggs, S. L. K.; Leaderer, B. P. (1992) Source apportionment of indoor aerosols in Suffolk and
Onondaga Counties, New York. Environ. Sci. Technol. 26: 521-527.
7-169
-------
Kulmala, M.; Saari, H.; Raunemaa, T.; Olin, M. (1987) Indoor air aerosols: outdoor air influence on indoor air. In:
Seifert, B.; Esdorn, H.; Fischer, M.; Rilden, H.; Wegner, J., eds. Indoor air '87: proceedings of the 4th
international conference on indoor air quality and climate, v. 1, volatile organic compounds, combustion gases,
particles and fibres, microbiological agents; August; Berlin, Federal Republic of Germany. Berlin, Federal
Republic of Germany: Institute for Water, Soil and Air Hygiene; pp. 564-568.
Laird, N. M. (1988) Missing data in longitudinal studies. Stat. Med. 7: 305-315.
Last, J. M., ed. (1988) A dictionary of epidemiology. 2nd ed. New York, NY: Oxford University Press; p. 56.
Leaderer, B. P.; Hammond, S. K. (1991) Evaluation of vapor-phase nicotine and respirable suspended particle mass as
markers for environmental tobacco smoke. Environ. Sci. Technol. 25: 770-777.
Leaderer, B.; Koutrakis, P.; Briggs, S.; Rizzuto, J. (1990) Impact of indoor sources on residential aerosol
concentrations. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and
climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
International Conference on Indoor Air Quality and Climate, Inc.; pp. 269-274.
Leaderer, B. P.; Koutrakis, P.; Briggs, S. L. K.; Rizzuto, J. (1994) The mass concentration and elemental composition
of indoor aerosols in Suffolk and Onondaga Counties, New York. Indoor Air 4: 23-34.
Lebret, E.; McCarthy, J.; Spengler, J.; Chang, B.-H. (1987) Elemental composition of indoor fine particles. In: Seifert,
B.; Esdorn, H.; Fischer, M.; Rilden, H.; Wegner, J., eds. Indoor air '87: proceedings of the 4th international
conference on indoor air quality and climate: v. 1, volatile organic compounds, combustion gases, particles and
fibres, microbiological agents; August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of
Germany: Institute for Water, Soil and Air Hygiene; pp. 569-574.
Lebret, E.; Boleij, J.; Brunekreef, B. (1990) Environmental tobacco smoke in Dutch homes. In: Indoor air '90:
precedings of the 5th international conference on indoor air quality and climate, volume 2, characteristics of
indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air
Quality and Climate, Inc.; pp. 263-268.
Lehmann, E.; Rentel, K.-H.; Allescher, W.; Hohmann, R. (1990) Messung von Dieselabgas am Arbeitsplatz
[Measurement of diesel exhaust at the workplace]. Zentralbl. Arbeitsmed. Arbeitsschutz Prophyl. Ergon. 40:
2-11.
Letz, R.; Ryan, P. B.; Spengler, J. D. (1984) Estimated distributions of personal exposure to respirable particles.
Environ. Monit. Assess. 4: 351-359.
Lewis, C. W. (1991) Sources of air pollutants indoors: VOC and fine particulate species. J. Exposure Anal. Environ.
Epidemiol. 1: 31-44.
Lewis, W. H.; Vinay, P.; Zenger, V. E. (1983) Airborne and allergenic pollen of North America. Baltimore, MD: The
Johns Hopkins University Press.
Lighthart, B.; Mohr, A. J. (1994) Atmospheric microbial aerosols: theory and applications. New York, NY: Chapman
& Hall.
Ligocki, M. P.; Liu, H. I. H.; Cass, G. R.; John, W. (1990) Measurements of particle deposition rates inside Southern
California museums. Aerosol Sci. Technol. 13: 85-101.
Lioy, P. J.; Waldman, J. M.; Buckley, T.; Butler, J.; Pietarinen, C. (1990) The personal, indoor and outdoor
concentrations of PM-10 measured in an industrial community during the winter. Atmos. Environ. Part B 24:
57-66.
7-170
-------
Little, R. J. A.; Rubin, D. B. (1987) Statistical analysis with missing data. New York NY: John Wiley & Sons.
Litzistorf, G.; Guillemin, M. P.; Buffat, P.; Iselin, F. (1985) Influence of human activity on the airborne fiber level in
paraoccupational environments. J. AirPollut. Control Assoc. 35: 836-837.
Lofroth, G.; Stensman, C.; Brandhorst-Satzkorn, M. (1991) Indoor sources of mutagenic aerosol particulate matter:
smoking, cooking and incense burning. Mutat. Res. 261: 21-28.
Mage, D. T. (1980) The statistical form for the ambient particulate standard annual arithmetic mean vs annual
geometric mean. J. AirPollut. Control Assoc. 30: 706-708.
Mage, D. T. (1983) Public health aspects of air quality surveillance. Public Health Rev. 11: 5-54.
Mage, D. T. (1985) Concepts of human exposure assessment for airborne particulate matter. Environ. Int. 11: 407-412.
Mage, D. T. (1991) A comparison of the direct and indirect methods of human exposure. In: Gledhill, B. L.; Mauro, F.,
eds. New horizons in biological dosimetry: proceedings of the international symposium on trends in biological
dosimetry; October 1990; Lerici, Italy. New York, NY: Wiley-Liss, Inc.; pp. 443-454. (Progress in clinical and
biological research: v. 372).
Mage, D. T.; Buckley, T. J. (1995) The relationship between personal exposures and ambient concentrations of
particulate matter. Presented at: 88th annual meeting of the Air & Waste Management Association; June; San
Antonio, TX. Pittsburgh, PA: Air & Waste Management Association; paper no. 95-MP 18.01.
Mage, D. T.; Ott, W. R. (1996) Accounting for nonuniform mixing and human exposure in indoor environments. In:
Tichenor, B. A., ed. Characterizing sources of indoor air pollution and related sink effects. Philadelphia, PA:
American Society for Testing and Materials; pp. 263-278.
Mage, D.; Ozolins, G.; Peterson, P.; Webster, A.; Orthofer, R.; Vandeweerd, V.; Gwynne, M. (1996) Urban air
pollution in megacities of the world. Atmos. Environ. 30: 681-686.
Mamane, Y. (1992) Characterization of PTEAM indoor aerosol samples (electron microscopy analysis). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment
Laboratory.
Marple, V. A.; Rubow, K. L.; Turner, W.; Spengler, J. D. (1987) Low flow rate sharp cut impactors for indoor air
sampling: design and calibration. JAPCA 37: 1303-1307.
Martinelli, C. A.; Harley, N. H.; Lippmann, M.; Cohen, B. S. (1983) Monitoring real-time aerosol distribution in the
breathing zone. Am. Ind. Hyg. Assoc. J. 44: 280-285.
Mathews, K. P. (1989) Inhalant insect-derived allergens. Immunol. Allergy Clin. North Am. 9: 321-338.
Matthews, T. G.; Thompson, C. V.; Wilson, D. L.; Hawthorne, A. R.; Mage, D. T. (1989) Air velocities inside domestic
environments: an important parameter in the study of indoor air quality and climate. Environ. Int. 15: 545-550.
Mayo, E. (1960) Hawthorne and the Western Electric Company. In: Merrill, H. F., ed. Classics in management. New
York, NY: American Management Association; pp. 417-436.
McKenzie, R. L.; Bright, D. S.; Fletcher, R. A.; Hodgeson, J. A. (1982) Development of a personal exposure monitor
for two sizes of inhalable particulates. Environ. Int. 8: 229-233.
Miesner, E. A.; Rudnick, S. N; Hu, F.-C.; Spengler, J. D.; Ozkaynak, H.; Preller, L.; Nelson, W. (1989) Particulate and
nicotine sampling in public facilities and offices. JAPCA 39: 1577-1582.
7-171
-------
Miller, A.; Thompson, J. M. (1970) Elements of meteorology. Columbus, OH: Charles E. Merrill Publishing Co.; p. 50.
Millette, J. R.; Hays, S. M. (1994) Resuspension of settled dust. In: Settled asbestos dust sampling and analysis. Boca
Raton, FL: CRC Press, Inc.; pp. 59-65.
Morandi, M. T.; Stock, T. H.; Contant, C. F. (1986) Characterization of indoor microenvironmental exposures to
respirable particulate matter. Presented at: 79th annual meeting of the Air Pollution Control Association; June;
Minneapolis, MN. Pittsburgh, PA: Air Pollution Control Association; paper no. 86-67.2.
Morandi, M. T.; Stock, T. H.; Contant, C. F. (1988) A comparative study of respirable particulate microenvironmental
concentrations and personal exposures. Environ. Monit. Assess. 10: 105-122.
Moschandreas, D. J.; Winchester, J. W.; Nelson, J. W.; Burton, R. M. (1979) Fine particle residential indoor air
pollution. Atmos. Environ. 13: 1413-1418.
Mumford, J. L.; Williams, R. W.; Walsh, D. B.; Burton, R. M.; Svendsgaard, D. J.; Chuang, J. C.; Houk, V. S.; Lewtas,
J. (1991) Indoor air pollutants from unvented kerosene heater emissions in mobile homes: studies on particles,
semivolatile organics, carbon monoxide, andmutagenicity. Environ. Sci. Technol. 25: 1732-1738.
Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
National Institute for Environmental Studies. (1994) Experimental studies on the health effects of air pollutants
containing diesel exhaust particles [FY 1988-1992 ]. Kokuritsu Kogai Kenkyusho Tokubetsu Kenkyu Hokoku
(SR-14-'94): 26-32.
National Research Council. (1986) Environmental tobacco smoke: measuring exposures and assessing health effects.
Washington, DC: National Academy Press.
Nazaroff, W. W.; Cass, G. R. (1989) Mathematical modeling of indoor aerosol dynamics. Environ. Sci. Technol. 23:
157-166.
Nazaroff, W. W.; Salmon, L. G.; Cass, G. R. (1990a) Concentration and fate of airborne particles in museums. Environ.
Sci. Technol. 24: 66-77.
Nazaroff, W. W.; Ligocki, M. P.; Ma, T.; Cass, G. R. (1990b) Particle deposition in museums: comparison of modeling
and measurement results. Aerosol Sci. Technol. 13: 332-348.
Nazaroff, W. W.; Gadgil, A. J.; Weschler, C. J. (1993) Critique of the use of deposition velocity in modeling indoor air
quality. In: Nagda, N. L., ed. Modeling of indoor air quality and exposure. Philadelphia, PA: American Society
for Testing and Materials; pp. 81-104. (ASTM special technical publication no. 1205).
Neas, L. M.; Dockery, D. W.; Ware, J. H.; Spengler, J. D.; Ferris, B. G., Jr.; Speizer, F. E. (1994) Concentration of
indoor particulate matter as a determinant of respiratory health in children. Am. J. Epidemiol. 139: 1088-1099.
Ogden, T. L.; Bartlett, I. W.; Pumell, C. J.; Wells, C. J.; Armitage, F.; Wolfson, H. (1993) Dust for cotton manufacture:
changing from static to personal sampling. Ann. Occup. Hyg. 37: 271-285.
Oldaker, G. B., Ill; Ogden, M. W.; Maiolo, K. C.; Conner, J. M.; Conrad, F. W., Jr.; DeLuca, P. O. (1990) Results from
surveys of environmental tobacco smoke in restaurants in Winston-Salem, North Carolina. In: Indoor air '90:
precedings of the 5th international conference on indoor air quality and climate, volume 2, characteristics of
indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada: International Conference on Indoor Air
Quality and Climate, Inc.; pp. 281-285.
7-172
-------
Oldaker, G. B., Ill; Taylor, W. D.; Parrish, K. B. (1993) Investigations of ventilation, smoking activity, and indoor air
quality at four large office buildings. In: Proceedings of IAQ '92. American Society of Heating, Refrigerating
and Air-Conditioning Engineers, Inc.; pp. 365-374.
Orchard, T.; Woodbury, M. A. (1972) A missing information principle: theory and applications. In: LeCam, L.;
Newman, J.; Scott, E., eds. Proceedings of the 6th Berkeley symposium on mathematical statistics and
probability, v. 1. Berkeley, CA: University of California Press; pp. 697-715.
Ott, W. R. (1982) Concepts of human exposure to air pollution. Environ. Int. 7: 179-196.
Owen, M. K.; Ensor, D. S.; Hovis, L. S.; Tucker, W. G.; Sparks, L. E. (1990) Particle size distributions for an office
aerosol. Aerosol Sci. Technol. 13: 486-492.
Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air. Atmos.
Environ. Part A 26: 2149-2162.
Ozkaynak, H.; Spengler, J. D.; Ludwig, J. F.; Butler, D. A.; Clayton, C. A.; Pellizzari, E.; Wiener, R. W. (1990)
Personal exposure to particulate matter: findings from the Particle Total Exposure Assessment Methodology
(PTEAM) prepilot study. In: Indoor air '90: precedings of the 5th international conference on indoor air quality
and climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
International Conference on Indoor Air Quality and Climate, Inc.; pp. 571-576.
Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Spengler, J. (1993a) The particle TEAM (PTEAM) study: analysis of the
data, volume III [draft final report]. Research Triangle Park, NC: U.S. Environmental Protection Agency; EPA
contract no. 68-02-4544.
Ozkaynak, H.; Spengler, J. D.; Xue, J.; Koutrakis, P.; Pellizzari, E. D.; Wallace, L. (1993b) Sources and factors
influencing personal and indoor exposures to particles, elements and nicotine: findings from the particle TEAM
pilot study. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds.
Indoor air '93: proceedings of the 6th international conference on indoor air quality and climate, v. 3,
combustion products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp.
457-462.
Ozkaynak, H.; Xue, J.; Spengler, J. D.; Wallace, L. A.; Pellizzari, E. D.; Jenkins, P. (1996) Personal exposures to
airborne particles and metals: results from the particle TEAM study in Riverside, California. In: Lippmann, M.,
ed. Papers from the ISEA-ISEE annual meeting; September 1994; Research Triangle Park, NC. J. Exp. Anal.
Environ. Epidemiol.: in press.
Pellizzari, E. D.; Thomas, K. W.; Clayton, C. A.; Whitmore, R. W.; Shores, R. C.; Zelon, H. S.; Pemtt, R. L. (1992)
Particle total exposure assessment methodology (PTEAM): Riverside, California pilot study, volume I [final
report]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and
Exposure Assessment Laboratory; EPA report no. EPA/600/R-93/050. Available from: NTIS, Springfield, VA;
PB93-166957/XAB.
Pellizzari, E. D.; Thomas, K. W.; Clayton, C. A.; Whitmore, R. C.; Shores, H.; Zelon, S.; Pentt, R. L. (1993) Particle
total exposure assessment methodology (PTEAM): Riverside, California pilot study - volume I [project
summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Atmospheric Research and
Exposure Assessment Laboratory; EPA report no. EPA/600/SR-93/050.
Pengelly, L. D.; Goldsmith, C. H.; Kerigan, A. T.; Furlong, W.; Toplack, S. (1987) The Hamilton study: estimating
exposure to ambient suspended particles. JAPCA 37: 1421-1428.
Pemtt, R. L.; Clayton, C. A.; Pellizzari, E. D.; Thomas, K. W.; Wallace, L. A.; Spengler, J. D.; Ozkaynak, H. (1991)
Particle Total Exposure Assessment Methodology (PTEAM) Pilot Study: personal, indoor, and outdoor
particulate concentration distributions for southern California fall 1990 - preliminary results. In: Measurement
7-173
-------
of toxic and related air pollutants: proceedings of the 1991 U.S. EPA/A&WMA international symposium, v. 2;
May; Durham, NC. Pittsburgh, PA: Air & Waste Management Association; pp. 665-671. (A&WMA
publication VIP-21).
Pickles, J. H. (1982) Air pollution estimation error and what it does to epidemiological analysis. Atmos. Environ. 16:
2241-2245.
Platts-Mills, T. A. E.; Chapman, M. D. (1987) Dust mites: immunology, allergic disease, and environmental control. J.
Allergy Clin. Immunol. 80: 755-775.
Pope, C. A., Ill; Thun, M. J.; Namboodm, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr. (1995)
Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J. Respir. Crit.
CareMed. 151:669-674.
Quackenboss, J. J.; Lebowitz, M. D.; Crutchfield, C. D. (1989) Indoor-outdoor relationships for particulate matter:
exposure classifications and health effects. Environ. Int. 15: 353-360.
Quackenboss, J. J.; Krzyzanowski, M.; Lebowitz, M. D. (1991) Exposure assessment approaches to evaluate
respiratory health effects of particulate matter and nitrogen dioxide. J. Exposure Anal. Environ. Epidemiol. 1:
83-107.
Raunemaa, T.; Kulmala, M.; Saari, H.; Olin, M.; Kulmala, M. H. (1989) Indoor air aerosol model: transport indoors and
deposition of fine and coarse particles. Aerosol Sci. Technol. 11: 11-25.
Repace, J. L. (1987) Indoor concentrations of environmental tobacco smoke: field surveys. In: O'Neill, I. K.;
Brunnemann, K. D.; Dodet, B.; Hoffmann, D., eds. Environmental carcinogens: methods of analysis and
exposure measurement, volume 9 - passive smoking. Lyon, France: World Health Organization; pp. 141-162.
(IARC scientific publications: no. 81).
Repace, J. L.; Lowrey, A. H. (1980) Indoor air pollution, tobacco smoke, and public health. Science (Washington, DC)
208: 464-472.
Repace, J. L.; Ott, W. R.; Wallace, L. A. (1980) Total human exposure to air pollution. Presented at: 73rd annual
meeting of the Air Pollution Control Association; June; Montreal, PQ, Canada. Pittsburgh, PA: Air Pollution
Control Association; paper no. 80-61.6.
Rippon, J. W. (1988) Medical mycology: the pathogenic fungi and the pathogenic actinomycetes. 3rd ed. Philadelphia,
PA: W. B. Saunders Company.
Roberts, J. W.; Camann, D. E.; Spittler, T. M. (1990) Monitoring and controlling lead in house dust in older homes. In:
Indoor air '90: precedings of the 5th international conference on indoor air quality and climate, volume 2,
characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada: International
Conference on Indoor Air Quality and Climate, Inc.; pp. 435-440.
Rodes, C. E.; Kamens, R. M.; Wiener, R. W. (1991) The significance and characteristics of the personal activity cloud
on exposure assessment measurements for indoor contaminants. Indoor Air 2: 123-145.
Santanam, S.; Spengler, J. D.; Ryan, P. B. (1990) Particulate matter exposures estimated from an indoor-outdoor source
apportionment study. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and
climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
International Conference on Indoor Air Quality and Climate, Inc.; pp. 583-588.
Sexton, K.; Spengler, J. D.; Treitman, R. D. (1984) Personal exposure to respirable particles: a case study in Waterbury,
Vermont. Atmos. Environ. 18: 1385-1398.
7-174
-------
Sheldon, L. S.; Handy, R. W.; Hartwell, T. D.; Whitmore, R. W.; Zelon, H. S.; Pellizzari, E. D. (1988a) Indoor air
quality in public buildings: v. I. Washington, DC: U.S. Environmental Protection Agency, Office of Acid
Deposition, Environmental Monitoring, and Quality Assurance; report no. EPA 600/6-88/009a. Available
from: NTIS, Springfield, VA; PB89-102503.
Sheldon, L.; Zelon, H.; Sickles, J.; Eaton, C.; Hartwell, T. (1988b) Indoor air quality in public buildings: volume II.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems
Laboratory; report no. EPA 600/6-88/009b. Available from: NTIS, Springfield, VA; PB89-102511.
Sheldon, L. S.; Hartwell, T. D.; Cox, B. G.; Sickles, J. E., II; Pellizzari, E. D.; Smith, M. L.; Pemtt, R. L.; Jones, S. M.
(1989) An investigation of infiltration and indoor air quality: final report. Albany, NY: New York State Energy
Research and Development Authority; New York State ERDA contract no. 736-CON-BCS-85.
Sheldon, L.; Clayton, A.; Keever, J. Pemtt, R.; Whitaker, D. (1992) PTEAM: monitoring of phthalates and PAHs in
indoor and outdoor air samples in Riverside, California, volume II [final report]. Sacramento, CA: California
Environmental Protection Agency, State Air Resources Board, Research Division; contract no. A933-144.
Available from: NTIS, Springfield, VA; PB93-205649/XAB.
Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J. (1988) Indoor/outdoor ratios and indoor surface accumulations of
ionic substances at Newark, New Jersey. Atmos. Environ. 22: 461-469.
Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J.; Shields, H. C. (1990) Measurement and modeling of airborne
concentrations and indoor surface accumulation rates of ionic substances at Neenah, Wisconsin. Atmos.
Environ. Part A 24: 627-638.
Sinclair, J. D.; Psota-Kelly, L. A.; Peins, G. A.; Ibidunni, A. O. (1992) Indoor/outdoor relationships of airborne ionic
substances: comparison of electronic equipment room and factory environments. Atmos. Environ. Part A 26:
871-882.
Smith, K. R.; Apte, M. G.; Ma, Y.; Wongsekiarttirat, W.; Kulkarni, A. (1994) Air pollution and the energy ladder in
Asian cities. In: Tyler, S.; Sathaye, J.; Goldman, N, eds. Energy use in Asian Cities. Energy (Oxford) 19:
587-600.
Smithard, E. H. R. (1954) The 1952 fog in a metropolitan borough. Mon. Bull. Minist. Health Public Health Lab. Serv.
(G. B.) 13:26-35.
Spengler, J. D.; Thurston, G. D. (1983) Mass and elemental composition of fine and coarse particles in six U.S. cities. J.
AirPollut. Control Assoc. 33: 1162-1171.
Spengler, J. D.; Dockery, D. W.; Reed, M. P.; Tosteson, T.; Quinlan, P. (1980) Personal exposures to respirable
particles. Presented at: 73rd annual meeting of the Air Pollution Control Association; June; Montreal, PQ,
Canada. Pittsburgh, PA: Air Pollution Control Association; paper no. 80-61.5b.
Spengler, J. D.; Dockery, D. W.; Turner, W. A.; Wolfson, J. M.; Ferris, B. G., Jr. (1981) Long-term measurements of
respirable sulfates and particles inside and outside homes. Atmos. Environ. 15: 23-30.
Spengler, J. D.; Treitman, R. D.; Tosteson, T. D.; Mage, D. T.; Soczek, M. L. (1985) Personal exposures to respirable
particulates and implications for air pollution epidemiology. Environ. Sci. Technol. 19: 700-707.
Spengler, J. D.; Ware, J.; Speizer, F.; Ferris, B.; Dockery, D.; Lebret, E.; Brunnekreef, B. (1987) Harvard's indoor air
quality respiratory health study. In: Seifert, B.; Esdom, H.; Fischer, M.; Rilden, H.; Wegner, J., eds. Indoor air
'87: proceedings of the 4th international conference on indoor air quality and climate: v. 2, environmental
tobacco smoke, multicomponent studies, radon, sick buildings, odours and irritants, hyperreactivities and
allergies; August; Berlin, Federal Republic of Germany. Berlin, Federal Republic of Germany: Institute for
Water, Soil and Air Hygiene; pp. 218-223.
7-175
-------
Spengler, J. D.; Ozkaynak, H.; Ludwig, J.; Allen, G.; Pellizzari, E. D.; Wiener, R. (1989) Personal exposures to
participate matter: instruments and methodologies P-TEAM. In: Measurement of toxic and related air
pollutants: proceedings of the 1989 U.S. EPA/A&WMA international symposium; May; Raleigh, NC.
Pittsburgh, PA: Air & Waste Management Association; pp. 449-463. (A&WMA publication VIP-13).
Sterling, T. D.; Dimich, H.; Kobayashi, D. (1982) Indoor byproduct levels of tobacco smoke: a critical review of the
literature. J. Air Pollut. Control Assoc. 32: 250-259.
Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993) Validation of personal exposure models for sulfate and aerosol strong
acidity. J. Air Waste Manage. Assoc. 43: 845-850.
Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
Tamura, K.; Ando, M. (1994) Suspended particulate matter in indoor and outdoor air along a main road and personal
exposure assessment. Presented at: Sixth conference of the International Society for Environmental
Epidemiology/fourth conference of the International Society for Exposure Analysis joint conference;
September; Research Triangle Park, NC. Chapel Hill, NC: University of North Carolina, School of Public
Health; abstract no. 116.
Tamura, K.; Ando, M.; Sagai, M.; Matsumoto, Y. (1996) Estimation of levels of personal exposure to suspended
particulate matter and nitrogen dioxide in Tokyo. Environ. Sci. (Tokyo) 6: accepted.
Thatcher, T. L.; Layton, D. W. (1995) Deposition, resuspension, and penetration of particles within a residence. Atmos.
Environ. 29: 1487-1497.
Thomas, K. W.; Pellizzari, E. D.; Clayton, C. A.; Whitaker, D. A.; Shores, R. C.; Spengler, J.; Ozkaynak, H.; Froehlich,
S. E.; Wallace, L. A. (1993) Particle total exposure assessment methodology (PTEAM) 1990 study: method
performance and data quality for personal, indoor, and outdoor monitoring. J. Exposure Anal. Environ.
Epidemiol. 3: 203-226.
Turk, B. H.; Brown, J. T.; Geisling-Sobotka, K.; Froehlich, D. A.; Grimsrud, D. T.; Harrison, J.; Koonce, J. F.; Prill, R.
J.; Revzan, K. L. (1987) Indoor air quality and ventilation measurements in 38 Pacific Northwest commercial
buildings: final report to the Bonneville Power Administration, volume I: measurement results and
interpretation [final report]. Berkeley, CA: U.S. Department of Energy, Lawrence Berkeley Laboratory, Indoor
Environment Program; report no. LBL-22315-1/2. Available from: DE88014064/XAB.
Turk, B. H.; Grimsrud, D. T.; Brown, J. T.; Geisling-Sobotka, K. L.; Harrison, J.; Prill, R. J. (1989) Commercial
building ventilation rates and particle concentrations. ASHRAE Trans. 95(part 1): 422-433.
U.S. Centers for Disease Control. (1988) Assessing exposures of health-care personnel to aerosols of
ribavirin—California. Morb. Mortal. Wkly. Rep. 37: 560-563.
U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
U.S. Environmental Protection Agency. (1986a) Second addendum to air quality criteria for particulate matter and
sulfur oxides (1982): assessment of newly available health effects information. Research Triangle Park, NC:
Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report
no. EPA-600/8-86-020F. Available from: NTIS, Springfield, VA; PB87-176574.
U.S. Environmental Protection Agency. (1986b) Air quality criteria for lead. Research Triangle Park, NC: Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
7-176
-------
U.S. Environmental Protection Agency. (1989) Exposure factors handbook. Washington, DC: Office of Health and
Environmental Assessment; EPA report no. EPA/600/8-89/043. Available from: NTIS, Springfield, VA;
PB90-106774.
U.S. Environmental Protection Agency. (1992) Respiratory health effects of passive smoking: lung cancer and other
disorders. Washington, DC: Office of Research and Development, Office of Health and Environmental
Assessment; EPA report no. EPA/600/6-90/006F. Available from: NTIS, Springfield, VA;
PB93-134419/XAB.
U.S. Environmental Protection Agency. (1995) PFT database for residential air flow measurements: database
documentation, final report. Washington, DC: Office of Toxic Substances; contract no. 68-D3-0013, task 3-18.
United Kingdom Ministry of Health. (1954) Mortality and morbidity during the London fog of December 1952.
London, United Kingdom: Her Majesty's Stationery Office. (Reports on public health and medical subjects no.
95).
Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
Wallace, L.; Clayton, A.; Whitmore, R.; Shores, R.; Thomas, K.; Whitaker, D.; Reading, P.; Pellizzari, E.; Spengler, I;
Ozkaynak, H.; Froehhch, S.; Jenkins, P.; Ota, L.; Westerdahl, D. (1991a) Initial results from the PTEAM study:
survey design, population response rates, monitor performance and quality control. In: Measurement of toxic
and related air pollutants: proceedings of the 1991 U.S. EPA/A&WMA international symposium, v. 2; May;
Durham, NC. Pittsburgh, PA: Air & Waste Management Association; pp. 659-664. (A&WMA publication
VIP-21).
Wallace, L. A.; Pelhzzan, E.; Sheldon, L.; Whitmore, R.; Zelon, H.; Clayton, A.; Shores, R.; Thomas, K.; Whitaker, D.;
Reading, P.; Spengler, J.; Ozkaynak, H.; Froehlich, S.; Jenkins, P.; Ota, L.; Westerdahl, D. (1991b) The TEAM
study of inhalable particles (PM10): study design, sampler performance, and preliminary results. Presented at:
84th annual meeting of the Air & Waste Management Association; June; Vancouver, British Columbia,
Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-171.3.
Wallace, L.; Nelson, W.; Ziegenfus, R.; Pelhzzan, E.; Michael, L.; Whitmore, R.; Zelon, H.; Hartwell, T.; Pemtt, R.;
Westerdahl, D. (1991c) The Los Angeles TEAM study: personal exposures, indoor-outdoor air concentrations,
and breath concentrations of 25 volatile organic compounds. J. Exposure Anal. Environ. Epidemiol. 1:
157-192.
Wallace, L. A.; Ozkaynak, H.; Spengler, J. D.; Pellizzari, E. D.; Jenkins, P. (1993) Indoor, outdoor, and personal air
exposures to particles, elements, and nicotine for 178 southern California residents. In: Jantunen, M.;
Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings of the
6th international conference on indoor air quality and climate, v. 3, combustion products, risk assessment,
policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 445-450.
Weschler, C. J.; Shields, H. C.; Kelty, S. P.; Psota-Kelly, L. A.; Sinclair, J. D. (1989) Comparison of effects of
ventilation, filtration and outdoor air at telephone office buildings. In: Nagda, N. L.; Harper, J. P., eds. Design
and protocol for monitoring indoor air quality. Philadelphia, PA: American Society for Testing and Materials;
pp. 9-34. (ASTM special technical publication 1002).
West, J. B. (1982) Pulmonary pathophysiology—the essentials. 2nd ed. Baltimore, MD: Williams & Wilkins; p. 168.
Wiener, R. W. (1988) Measurement and evaluation of personal exposure to aerosols. In: Measurement of toxic and
related air pollutants: proceedings of the 1988 EPA/APCA international symposium; May; Research Triangle
Park, NC. Pittsburgh, PA: Air Pollution Control Association; pp. 84-88. (APCA publication no. VIP-10).
7-177
-------
Wiener, R. W. (1989) Particle total exposure assessment methodology—an overivew of planning and
accomplishments. In: Measurement of toxic and related air pollutants: proceedings of the 1989 U.S.
EPA/A&WMA international symposium; May; Raleigh, NC. Pittsburgh, PA: Air & Waste Management
Association; pp. 442-448. (A&WMA publication VIP-13).
Wiener, R. W.; Wallace, L.; Pahl, D.; Pellizzan, E.; Whittaker, D.; Spengler, I; Ozkaynak, H. (1990) Review of the
particle TEAM 9 home field study. In: Measurement of toxic and related air pollutants: proceedings of the
1990 EPA/A&WMA international symposium, v. 2; May; Raleigh, NC. Pittsburgh, PA: Air & Waste
Management Association; pp. 452-460. (A&WMA publication VIP-17).
Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York, NY:
Van Nostrand Reinhold Publishers.
Wilmoth, R. C.; Powers, T. J.; Millette, J. R. (1991) Observations on studies useful to asbestos O&M activities.
Microscope 39: 299-312.
Wilson, W. E.; Suh, H. H. (1995) Differentiating fine and coarse particles: definitions and exposure relationships
relevant to epidemiological studies. In: Schmidt-Ott, A., ed. Trends in aerosol research IV: new approaches in
aerosol science and technology, proceedings of the seminar; January; Gerhard Mercator University, Duisburg,
Germany. Duisburg, Germany: Gerhard Mercator University of Duisburg; pp. 57-71.
Winer, B. J. (1962) Statistical principles in experimental design. New York, NY: McGraw-Hill Book Co., Inc.; pp.
105-124.
Woodbury, M.; Hasselblad, V. (1970) Maximum likelihood estimates of the variance covariance matrix from the
multivanate normal. In: Proceedings of SHARE XXXIV, v. 3; pp. 1550-1561.
World Health Organization. (1982a) Human exposure to carbon monoxide and suspended particulate matter in Zagreb,
Yugoslavia. Geneva, Switzerland: World Health Organization; GEMS (Global Environmental Monitoring
System) report no. EFP/82.33.
World Health Organization. (1982b) Human exposure to SO2, NO2 and suspended particulate matter in Toronto,
Canada. Geneva, Switzerland: World Health Organization; GEMS (Global Environmental Monitoring System)
report no. EFP/82.38.
World Health Organization. (1984) Human exposure to suspended particulate matter and sulfate in Bombay, India.
Geneva, Switzerland: World Health Organization; GEMS (Global Environment Monitoring System) report no.
EFP/84.86.
World Health Organization. (1985) Human exposure to carbon monoxide and suspended particulate matter in Beijing,
People's Republic of China. Geneva, Switzerland: World Health Organization; GEMS (Global Environmental
Monitoring System) report no. PEP/85.11.
World Health Organization. (1995) Interpretation of WHO air quality guidelines: report of an expert working group.
Geneva, Switzerland: World Health Organization; report no. WHO/EOS/95; in press.
Yoshimura, I. (1990) The effect of measurement error on the dose-response curve. Environ. Health Perspect.
87: 173-178.
7-178
-------