The Science of the Total Environment 257 Ž2000.

199᎐211

Transport of traffic-related aerosols in urban areas

U
´ a, , E. Rokitaa,b, W. Maenhautc
A. Wrobel
a
Institute of Physics, Jagiellonian Uni¨ ersity, Reymonta 4, 30-059 Krakow,
´ Poland
b
Regional Laboratory of Physicochemical Analyses and Structure Research, Jagiellonian Uni¨ ersity, Ingardena 3,
´ Poland
31-060 Krakow,
c
Instituut ¨ oor Nucleaire Wetenschappen, Uni¨ ersiteit Gent, Proeftuinstraat 86, B-9000 Gent, Belgium

Received 8 September 1999; accepted 9 May 2000

Abstract

This study was undertaken to assess the influence of traffic on particulate air pollution in an urban area, and to
´ a city
characterise the short-range transport of the aerosols generated by traffic. The study was conducted in Krakow,
located in southern Poland with a population of approximately 800 000. Aerosol samples were collected using
automatic sampling equipment at five sites located at different distances from the main road in Krakow,´ ranging from
5 to 1500 m. The sampling set-up allowed standardisation of the results due to continuous determination of the
meteorological parameters Žtemperature, atmospheric pressure, wind speed and direction, rainfall and humidity..
Aerosol particles were separated according to aerodynamic diameter into two size fractions: ) 1.9 ␮m Žcoarse
fraction.; and 1.9᎐72 ␮m Žfine fraction.. The concentrations of 27 elements were measured in both size fractions ŽNa,
Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr, Zr, Ba, Pb.. The multielement
analyses were performed by Particle-Induced X-ray Emission ŽPIXE. spectrometry. Traffic contribution to particu-
late air pollution was determined on the basis of 13 elements which were present above the detection limit in all
samples ŽMg, Al, Si, P, S, K, Ca, Ti, Mn, Fe, Cu, Zn, Pb.. It was found that the traffic contribution in the coarse size
fraction was approximately 80% up to 150 m from the road; it dropped abruptly by a factor of 2 over a distance of
150᎐200 m and declined further to 20% at 1500 m from the road. Traffic contribution for the fine particle
concentrations of individual elements was 50᎐70% in the close vicinity of the road Ž5 m.; then there was a decrease,
followed by an increase at a greater distance from the road. Possible explanations for this behaviour of the fine
particles are given. 䊚 2000 Elsevier Science B.V. All rights reserved.

Keywords: Traffic-derived particulate air pollution; Elemental composition of aerosols; Short-range aerosol transport; Motorway
emission; Vehicle

U
Corresponding author. Tel.: q48-12-633-93-76, ext. 5697; fax: q48-12-633-70-86.
´ ..
E-mail address: ufawrobe@cyf-kr.edu.pl ŽA. Wrobel

0048-9697r00r$ - see front matter 䊚 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 8 - 9 6 9 7 Ž 0 0 . 0 0 5 1 9 - 2
200 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
1. Introduction
Particulate air pollution can be generated by
natural and anthropogenic sources. Anthro-
pogenic sources are classified as stationary and
mobile ones. In recent years, a general downward
trend in aerosol emission from stationary sources
has been observed in most of the developed coun-
tries. It has been estimated in highly industri-
alised countries that the traffic-related aerosol air
pollution constitutes more than 50% of the total
particulate air pollution ŽEnvironmental Protec-
tion Year-book, 1996; Harrison et al., 1996.. Since
the reduction of heavy industry output the intro-
duction of new technologies is now common prac-
tice in Poland. It may be expected that the traffic
component will also dominate in the near future.
An increase in the traffic component is particu-
larly compelling in urban areas.
Air pollution is a cause of serious concern for
the public, especially inhabitants of heavily popu-
lated regions ŽWilliams and McCrae, 1995.. These
authors indicate that hazards associated with ve-
hicle-derived air pollution are considered to be
comparably as important as unemployment, law
and order, and rising prices. A similar trend is
also observed in Poland, manifested mainly as
opposition of citizens against planned highway
construction in the vicinity of residential areas or
new transit roads in the cities.
A number of articles and reports have been
published regarding the contribution of traffic to
total air pollution ŽKuhler et al., 1994; Harrison
and Jones 1995; Harrison et al., 1996; Janssen et
al., 1997; Hitchins et al., 2000.. There are some
literature data on behaviour of emissions gener-
ated by traffic at increasing distances from the
road which include principally the studies of
gaseous emissions ŽKuhler et al., 1994. and mass
concentrations ŽHitchins et al., 2000. while the
present study attempted to evaluate traffic impact
on the basis of elemental concentrations.
The studies were conducted in Krakow, ´ a city
located in southern Poland with a population of
approximately 800 000. The city lies in a valley
and its centre is highly built up, so disadvanta-
geous phenomena connected with air pollution
are especially significant.
The main goal of this research was to de-
termine the contribution of traffic to the particu-
late air pollution and to characterise transport of
aerosols in urban area in close vicinity of a road.
For this purpose, elemental concentrations were
measured in aerosol samples collected at several
distances from a road with high traffic intensity.
The work was designed as a pilot study constitut-
ing a basis for future, more specific investigations.
It is obvious that the more elements andror
species are determined in aerosols, the more ef-
ficient and unequivocal the characterisation of
the particulate pollution sources is. Therefore, we
used the multielement technique Particle-In-
duced X-Ray Emission ŽPIXE. spectrometry to
measure the elements.
2. Aerosol collection
At each sampling site, aerosol samples were
collected separately during the day and night.
‘Day’ and ‘night’ sampling hours were defined
taking into consideration traffic intensities. Dif-
ferences in ‘day’ and ‘night’ elemental concentra-
tions were used as a measure of traffic impact on
particulate air pollution. The sampling time was
24 h, so each individual sample was collected for
2 consecutive days or nights. A total of 30 coarse
and 30 fine aerosol samples were collected using
a stacked filter unit wNorwegian Institute for Air
Research ŽNILU. type ‘open face’ filter holderx
with two 47-mm diameter Nuclepore polycarbon-
ate filters Ža coarse Apiezon-coated filter with a
pore size of 8 ␮m and a fine filter of 0.4-␮m pore
size.. At the flow rate used Ž2.3 m3rh., the coarse
filter has a cut-off point Ž d50 . of 1.9 ␮m equiva-
lent aerodynamic diameter ŽEAD. ŽHeidam, 1981;
Lodge et al., 1981.. Thus, the coarse filter col-
lected the particles larger than 1.9 ␮m, whereas
the fine filter collected the particles smaller than
1.9 ␮m.
The stacked filter unit constituted a part of the
sampling set-up, consisting also of a vacuum pump
and a flow meter. The sampling set-up was con-
nected to a weather station allowing the record-
ing of the following parameters: wind direction
and velocity, temperature and humidity of air,
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
rainfall, and barometric pressure. The station was
controlled by a microprocessor SAB80C537 ŽSie-
mens.. The control unit allowed recording of me-
teorological data and air flows, and enabled sam-
pling according to a pre-established schedule; for
example, variability ranges of meteorological
parameters could be defined. To make measure-
ments more flexible, a portable sampling set-up
was used. A detailed description of the sampling
set-up is given elsewhere ŽWrobel ´ and Rokita,
1998..
Previous experiments revealed a strong depen-
dence of the measured elemental concentrations
on selected meteorological parameters, namely
wind speed and direction ŽHarrison and Williams,
´ et
1982; Sharan et al., 1995; Sharan, 1996; Wrobel
al., 1999. as well as rainfall ŽOrlic et al., 1999;
´ et al., 1999.. The other measured meteo-
Wrobel
rological parameters Žtemperature, barometric
pressure, and humidity. had a smaller influence
on the results in late spring and summer, when
the tests were conducted. Based on the results of
the test experiments, the following ranges of sam-
pling parameters were established having the
sampler activated in the current study: tempera-
ture 9᎐22⬚C; barometric pressure 999᎐1006 hPa;
absolute humidity 7᎐9 grm3; and wind direction
SW to NW, wind speed 0᎐3 mrs. Furthermore,
´
our previous studies ŽWrobel et al., 1999. have
shown that rainfall could decrease the atmo-
spheric concentrations of the elements by as much
as 50% in the coarse size fraction, the collections
were made only during dry weather.
It is obvious that sequential sampling at dif-
ferent sites under exactly the same weather con-
ditions is not possible. Therefore, a simple stan-
dardisation was made with regard to wind speed
and direction, as described in detail elsewhere
´ et al., 1999.. Briefly, as the concentration
ŽWrobel
at a particular sampling site is inversely propor-
tional to the speed Ž ¨ . of the wind component
´
from the source to the site ŽJuda and Chrusciel,
1974; Sharan et al., 1995; Sharan, 1996., with the
source in our case being the road, all concentra-
tions were multiplied by ¨ r ² ¨ :, with ² ¨ : the
average wind speed for the whole sampling pe-
riod.
201
3. Area description
To study the short-range aerosol transport, the
main transit road in the city of Krakow ´ was
selected on the basis of traffic intensity measure-
ments ŽTurzanski,´ 1997.. This road is charac-
terised by a traffic density of 2500᎐3000 carsrh
during the rush hours. The traffic density remains
around this range from 07.00 to 19.00 h, and
decreases to 200᎐500 carsrh in the night. The
traffic densities on neighbouring streets during
the day are shown in Fig. 1; the night traffic on
these streets is - 50 carsrh. These traffic density
data determined the sampling schedule at each
site, i.e. ‘day’ samples were collected from 07.00
to 19.00 h and night samples from 19.00 h to
07.00 h.
The five selected sampling sites ŽFig. 1. were
located in an area of Krakow ´ where the transit
road is expected to be the main source of particu-
late air pollution. There were no other major
aerosol sources in their close vicinity. Obviously,
it would have been preferable to conduct the
collections at sites where the air flow from the
transit road is not hindered by any obstacles, but
it is impossible to find such places in the city
centre. Therefore, the sampling sites were located
along a street with traffic density of - 500 carsrh
during the day. The sampling sites were similar
with regard to the influence of other emission
sources and degree of sheltering, and they were
located at the following distances from the transit
road: 5, 150, 200, 270, and 1500 m.
4. Multielement analysis
To assess the contribution from traffic to the
particulate air pollution, the aerosol samples were
subjected to multielemental analysis by Particle-
Induced X-Ray Emission ŽPIXE. spectrometry.
The PIXE analyses were undertaken at the Uni-
versity of Gent and at the Institute of Nuclear
´ Details of the PIXE set-ups
Physics in Krakow.
and the procedures used in Gent and Krakow ´ are
given elsewhere ŽMaenhaut et al., 1981, 1987;
´ et al., 1996..
Wrobel
The PIXE technique was selected because it
202 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
´ with location of the main transit road ŽMickiewicza Str.. and of the five sampling sites. Different line
Fig. 1. Map of part of Krakow
patterns are used to indicate different daytime traffic densities.
enables analysis of samples with very low mass Žof
approx. 0.1 mg.. This low mass requirement al-
lowed us to reduce the size of the sampling set-up
and to make it portable. Furthermore, the filter
samples Žor sections thereof. are directly bom-
barded as such in PIXE, and neither dissolution
nor another chemical sample preparation proce-
dure is required. Concentrations of up to 27 ele-
ments Ži.e. Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V,
Cr, Mn, Fe, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb,
Sr, Zr, Ba, Pb. could be determined with a 15-min
PIXE bombardment, with detection limits down
to 0.1 ngrm3. For the coarse size fraction, each of
the 27 elements was above the detection limit in
at least one sample, but for the fine fraction, four
elements Ži.e. Ge, Sr, Zr, Ba. were always below
the detection limit.
Examples of PIXE spectra for a coarse filter
sample and a blank filter are shown in Fig. 2. In
the blank filter, the following elements were de-
tected: Cr, Mn, Fe, Ni, Cu, and Br. There were
only trace amounts of Fe, Mn, and Cu in the
blanks. Moreover, these elements were present at
much higher levels in the samples than in the
blanks Žfor Fe over 1000 times higher, and for Mn
and Cu approx. 100 times higher., so that the
blank correction was small. For Cr, Ni, and Br,
however, the levels in the blank filters approxi-
mated those in the samples wCr: ; 4 ngrm3
Žfine. ᎐40 ngrm3 Žcoarse.; Ni: ; 0.1 ngrm3
Žfine. ᎐; 1 ngrm3 Žcoarse.; Br: 4 ngrm3 Žfine. ᎐
; 8 ngrm3 Žcoarse.x. Therefore, the atmospheric
concentration data for these elements were not
retained for further examination.
5. Results and discussion
In the present study, evaluation of traffic-
derived air pollution was based on elemental con-
centrations determined by Proton Induced X-ray
Emission ŽPIXE.. This method has been widely
used in aerosol studies all over the world, e.g. in
the United Kingdom ŽHarrison and Jones, 1995.,
Italy ŽPrati et al., 1999., Chile ŽArtaxo et al.,
1999., USA ŽCahill, 1990., India ŽBandhu et al.,
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel 203

Fig. 2. PIXE spectra for a coarse filter sample and a blank filter ŽApiezon-coated Nuclepore polycarbonate filter, 8 ␮m pore size..
X-Ray intensity is plotted against X-ray energy ŽE, keV.. NrQ is the number of counts per channel per unit charge. The inserted
table gives the concentrations of various elements Žngrm3 ., determined on the basis of this spectrum.

2000., Brasilia ŽGerab et al., 1998., Atlantic Ocean
ŽMaenhaut et al., 1981., Singapore ŽOrlic et al.,
1999.. It is characterised by high sensitivity if
small samples are to be analysed.
In this work, elemental concentrations were
determined in two aerosol fractions. The coarse
fraction, equivalent aerodynamic diameter ŽEAD.,
resulting from flow rate ranged from 1.9 to 72
␮m, it contained PM10 particles and black smoke.
Such a definition of the coarse fraction was ex-
pected to reveal distance-related differences in
traffic-derived particulate air pollution. The fine
fraction ŽEAD- 1.9 ␮m. consisted of the inhal-
able particulate fraction, the most hazardous for
human health ŽHarrison et al., 1996..
To present elemental concentrations for the
aerosol fractions, the elements were divided into
three groups: major elements, heavy metals, and
minor elements. Most of the elements included in
the group of major elements ŽNa, Mg, Al, Si, P, S,
Cl, K, Ca, Fe. are present in high concentrations
in crustal rock and soil; the group of heavy metals
contains metals with atomic number ) 25 which
are frequently included in environmental moni-
toring programs ŽMn, Cu, Zn, Pb., while the
remaining elements were classified as minor ele-
ments ŽTi, V, Ga, Ge, As, Se, Rb, Sr, Zr, Ba..
Table 1 presents the elemental concentrations
Žbefore applying the normalisation for wind speed
described in Section 2. in the samples collected at
5 m from the road. In the coarse size fraction, two
elements dominated, i.e. Si and Ca, whose maxi-
mum concentrations exceeded 1 ␮grm3. Concen-
trations of Na, Mg, Al, S, Cl, K and Fe were in
the range of several hundred ngrm3, while P, Ti,
Ba and heavy metals concentrations ranged from
10 to 100 ngrm3. As to the minor elements,
concentrations of coarse V, Sr and Zr fluctuated
from 1 to 10 ngrm3, while those of Ga, Ge, Se,
and As were approximately 0.1 ngrm3. Elemental
concentrations in the fine size fraction were con-
siderably different from those in the coarse frac-
204 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel

Table 1
Average day and night elemental concentrations ŽAvg., in ngrm3 , and associated standard deviations ŽS.D.., based on up to four
´ and respective values measured in Arnhem ŽJanssen et al., 1997. and London
samples, at 5 m from the transit road in Krakow
ŽHarrison et al., 1996.

Element Fine size fraction Coarse size fraction

´
Krakow Arnhem ´
Krakow London
Range annual
Day Night Day Night
Avg
Avg S.D. Avg S.D. Avg S.D. Avg S.D.

Na ᎐ ᎐ 72.6 41.4 25᎐420 172 95 143 24

Mg 16.7 a ᎐ 16.4 10.4 107 63 67.9 4.0
Al 13.9 6.7 13.5 1.1 355 247 185 40 456 b
Si 46.5 27.0 89.8 51 9᎐164 1614 1176 817 168
P 7.38 a ᎐ 7.14 2.41 42.7 9.8 37.6 24.6
S 406 202 771 521 297᎐4321 293 67 277 116
Cl 6.52 1.19 10.7 6.4 172 107 145 32
K 40.5 22.2 93.6 23.1 46᎐471 206 122 143 43
Ca 32.0 19.0 32.6 5.6 1149 653 671 127
Fe 39.2 19.4 56.0 29.2 120᎐367 749 443 459 77 913
Mn 1.34 0.39 2.20 0.97 7᎐40 15.7 9.3 9.99 1.12 15
Cu 1.77 0.75 2.24 0.78 6᎐61 16.9 8.6 8.27 1.14 24
Zn 11.0 2.7 32.8 16.9 18᎐174 49.5 36.2 44.1 21.0 65
Pb 16.3 7.5 28.8 11.0 29.6 16.2 20.4 9.4 99
Ti 1.38 0.21 1.17 0.31 43.9 27.8 21.3 3.6
V ᎐ ᎐ ᎐ ᎐ 2.19 a ᎐ 0.67 0.07 15
Ga ᎐ ᎐ 0.34 0.22 0.87 0.38 0.41 0.05
Ge ᎐ ᎐ ᎐ ᎐ 0.58 0.23 0.35a ᎐
As ᎐ ᎐ ᎐ ᎐ ᎐ ᎐ 0.86 a ᎐ 2.65 b
Se ᎐ ᎐ 0.36 0.21 0.77 a ᎐ 0.33a ᎐ 2.0 b
Rb 0.55a ᎐ ᎐ ᎐ ᎐ ᎐ ᎐ ᎐
Sr ᎐ ᎐ ᎐ ᎐ 3.45 2.69 2.14 0.32
Zr ᎐ ᎐ ᎐ ᎐ 5.89 3.46 1.83a ᎐
Ba ᎐ ᎐ ᎐ ᎐ 11.9 7.0 5.86 1.93
a
Detected in one sample.
b
According to AEA-Technology Trace Element Survey.

tion. Sulfur was the dominant element in the fine
fraction, with concentrations exceeding 500
ngrm3. Concentrations of all other major ele-
ments Žwith the exception of Cl and P. as well as
these of two heavy metals ŽZn and Pb. fell in the
range 10᎐100 ngrm3, while fine Cl, P, Mn, Cu,
and Ti were present at 1᎐10 ngrm3. The concen-
trations of fine Ga and Se were less than 1
ngrm3. The average concentrations, after normal-
isation for wind speed, for the three groups of
elements in the day and night samples at two
distances from the transit road Ž5 and 1500 m. are
displayed in Fig. 3.
A number of studies have been published
concerning elemental concentrations in urban
aerosols Že.g. Thurston and Spengler, 1985; Har-
rison et al, 1996; Artaxo et al., 1999; Orlic et al.,
1999; Prati et al., 1999; Bandhu et al., 2000..
However, direct comparison with the literature
data is hindered by the fact that particulate ele-
mental compositions depend strongly on particle
size under analysis. Therefore, the data obtained
in the present study were compared with those
which analysed similar particulate fractions col-
lected in European urban areas. Elemental levels
in the fine fraction were compared with their
respective values measured in the samples ŽPM2.5 .
collected at Arnhem 0.5 m from a street Ž15 000
carsr24 h. ŽJanssen et al., 1997.. Coarse fraction
data were related to the results obtained in Cen-
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
Fig. 3. Average elemental concentrations Žngrm3 ., normalised to wind speed, in the coarse and fine aerosol size fractions in the
samples collected at 5 and 1500 m from the transit artery.
tral London since in the latter study the samples
were not size fractionated ŽHarrison and Jones,
1995.. The perusal of Table 1 indicates that gen-
erally the data are comparable. The largest dif-
ferences were noted in the lead levels in the
coarse fraction which was three times higher in
London than in Krakow. ´ Literature data indicate
that aerosol lead concentrations show wide dif-
ferences in different cities, e.g. in ngrm3: in the
coarse fraction, 0.53 in Birkenes, Norway, 19.81 in
Lund, Sweden ŽCornille et al., 1987., 610 in Hus-
ton, USA, 1225 in Chandigarh, India ŽBandhu,
2000., and in the fine fraction 3.3 in Birkenes
ŽCornille et al., 1987., 31 in Singapore ŽOrlic et
al., 1999., ; 100 in Santiago, Chile ŽArtaxo et al.,
1999., which may result from different traffic den-
sities, population, topography, etc., in various
cities. It should be taken into account that the
compared data were obtained using different ana-
205
lytical methods, and the results from London are
for annual averages.
The main objectives of this study were to evalu-
ate the contribution from traffic to the particulate
air pollution and to describe the short-range
aerosol transport. For this work we only retained
the elements Ži.e. Mg, Al, Si, P, S, Cl, K, Ca, Ti,
Mn, Fe, Cu, Zn, Pb. which could be determined
Žin the coarse size fraction. at all sampling sites
during the day and night in at least three repli-
cates, and we used the concentrations that were
normalised for wind speed and direction. Normal-
isation for wind direction was applied earlier by
Harrison and Williams Ž1982. with respect to
long-range transport of cadmium, lead and zinc.
Hitchins et al. Ž2000. demonstrated strong effect
of wind speed and direction on short-range
transport of traffic-related aerosols ŽPM2.5 or less..
Therefore, in our work we used normalisation for
206 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel

Fig. 4. Traffic contribution to the particulate air pollution at different distances from the transit artery for the coarse ŽA. and fine
ŽB. size fractions. Difference between day w cŽ d .x and night w cŽ n.x elemental concentrations normalised to the concentration at 5 m
from the road w cŽ dy 5 m.x wi.e. p s 100 = w cŽ d . y cŽ n.xrcŽ d᎐5 m.x plotted against distance from the transit road ŽD..

both these parameters to be able to compare
results of measurements conducted at different
locations and times.
For coarse Cl, the variation with distance from
the road differed substantially from the changes
of all other coarse fraction elements, which sug-
gests that coarse Cl originated from sources not
connected with traffic. Therefore, the further
analysis of the traffic impact is based on 13 ele-
ments. For several elements concentrations were
below detection limits in most of the samples ŽNa,
V, Ga, Ge, As, Se, Rb, Sr, Zr, and Ba in the
coarse fraction, and the same elements plus Mg,
P, and Cl in the fine fraction.. To be able to
measure these elements consistently, a longer
sampling time would be necessary. However, a
longer sampling time involves the risk of larger
changes in weather conditions during the collec-
tion. So, a compromising approach was adopted
whereby the sampling time was restricted to have
fairly constant meteorological conditions during
it.
With the exception of Cl, the daytime concen-
trations of all elements in the coarse size fraction
were approximately two to five times higher at the
site at 5 m from the transit road than at 1500 m
from this road ŽFig. 3.. In the fine size fraction,
concentrations of most elements decreased by a
factor of approximately 2 over the distance from 5
to 1500 m; the concentration of fine sulfur de-
creased by a factor of 3.
To describe the short-range transport of the
traffic-related aerosol component, the differences
between the daytime and night-time elemental
concentrations were adopted as a measure of the
traffic contribution ŽFig. 4.. This assumption is
justified by the fact that the traffic density was
6᎐10 times higher during the day than in the
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
night. From Fig. 3 it appeared that the elemental
concentrations were substantially higher during
the day than in the night for both size fractions.
The day᎐night difference is particularly pro-
nounced for the coarse size fraction.
As the absolute concentrations of the different
elements differ by up to two orders of magnitude,
a normalisation was performed in order to facili-
tate a comparison of the results for the different
elements. This was accomplished by setting the
daytime concentration at 5 m from the transit
road at 100% for each of the elements. The data
in Fig. 4 then present the normalised distance-
dependent change in the traffic contribution for
the various elements.
As seen from Fig. 4a, the contribution from
traffic at 5 m from the transit road ranges from
89% for coarse Al to 71% for coarse Pb. The
traffic contribution to coarse particle air pollution
for the various elements Žwith the exception of
Cl. drops to 25% ŽCu. ᎐55% ŽCa. at 270 m from
the road, and it declines further at 1500 m from
the pollution source to values from 6% ŽP. to
38% ŽMg.. It is noteworthy that an abrupt drop in
the coarse fraction elemental concentrations Žap-
prox. a factor of 2. is observed over a distance of
between 150 and 200 m from the transit artery.
The difference between the patterns of distance-
dependent change in the various elements, may in
part be due to differences in the aerodynamic
diameter Žand thus dry deposition velocity. of the
particles with which the elements are associated.
Assessing the traffic contribution for the fine
size fraction is much more complex than for the
coarse fraction because of the longer atmospheric
lifetime of the fine particles. From Fig. 4b it can
be seen that the approximate contributions from
traffic at 5 m from the road are in the range
53᎐68% for most elements Ž39% for Zn.. Fur-
thermore, although the contributions from traffic
at any particular site differ markedly for the
different elements, the patterns of change with
distance from the road are rather similar: a drop
at 150 m, an increase at 200᎐270 m, and a slow
decrease over the distance to 1500 m from the
pollution source.
From a comparison of Fig. 4a,b it is evident
that the traffic-related percentages Žparameter p .
207
of the different elements at any particular site are
much closer to each other for the coarse size
fraction than for the fine fraction, except for the
site at 150 m from the main road. When averag-
ing the p-values of the various elements at each
of the sites for the coarse size fraction ŽFig. 4a.,
the highest standard deviation was obtained at
150 m from the road Ž31%. and the lowest at 5 m
Ž5.0%.. This suggests that the various elements
are either associated with roughly the same aero-
sol particles or otherwise with particle popula-
tions of rather similar size distribution and source.
The aerodynamic diameter determines the parti-
cle settling velocity; for example, for particles
with aerodynamic diameters of 1 and 100 ␮m the
settling velocities in still air Žat 0⬚C and 1 atm.
are 10 mmrs and 25 cmrs, respectively.
The average p-values for the coarse fraction
elements can be considered as the traffic compo-
nent of the coarse particle air pollution. These
average p-values and the variability ranges in p
are shown in Fig. 5. This figure suggests that the
transport of the traffic-related coarse particles
has two distinct components: a short-distance
component, with a rapid decline in concentration
up to 150 m, and a long-distance component with
a much smaller decrease beyond 150 m.
It is necessary to emphasise that the combus-
tion of gasoline and diesel fuel in motor vehicle
engines gives rise to particles with aerodynamic
diameters ranging from less than 0.1 to 2 ␮m
ŽLodge et al., 1981; Harrison et al., 1996.. Conse-
quently, the traffic-related aerosol should espe-
cially be seen in the fine size fraction. The
traffic-related coarse fraction aerosol consists
predominantly of resuspended road dust. The road
dust is composed of particles of natural, indus-
trial, andror municipal origin, and it is repeatedly
introduced into the atmosphere by moving cars.
The elemental composition of the coarse fraction
aerosol appeared to be rather similar to that of
soils and dust emitted by power plants and steel
works in the Krakow ´ area ŽManecki et al., 1981;
´
Turzanski, 1996.. Thus, the traffic resuspends
coarse particulate air pollutants from a variety of
sources, but this effect remains limited to a rela-
tively small area.
The transport of the fine aerosol is quite dif-
208 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel

Fig. 5. Traffic component of the coarse particulate air pollution. Average p-value Žsee Fig. 4 and text for explanation of parameter
p . and associated standard deviation plotted against distance from the transit road. Boundaries of the grey area indicate the
variability range in p Žminimum and maximum values..

ferent from that of the coarse aerosol. The low
settling velocity of the fine particles and, conse-
quently, their longer atmospheric lifetime have
the effect that the fine aerosol is a mixture of
particles from a large area and from various
sources. Furthermore, fine particles are more sub-
ject to turbulence and atmospheric mixing
processes than coarse particles are, which makes
the atmospheric transport of the former more
complicated. Also, fine particles are constantly
formed inside the atmosphere from various
gaseous species by gas-to-particle conversion
Žnucleation .. The emissions from motor vehicle
engines permanently supplement the atmosphere
with primary fine particles and with various aero-
sol precursor gases. The introduction of the parti-
cles and gases into the atmosphere is subject to
local turbulence Ždepending on the speed of ejec-
tion of combustion gas and local temperature
gradients., which causes rapid rise and dispersion
of the fine particles, so that they do not settle in
the close vicinity of the pollution source. This
phenomenon may help to explain why the highest
concentrations of fine fraction elements were not
observed close to the road Ž5᎐150 m. but further
away Ž200᎐270 m, see Fig. 4b..
Aerosol studies widely utilise models which are
aimed to determine the contribution of individual
emission sources to total air pollution Žreceptor
modelling. or to establish patterns of the propa-
gation of particulate air pollution on the basis of
basic physical parameters and laws of hydrody-
namics Žanalytical models..
Receptor modelling consists basically in the
determination of elemental composition of partic-
ulate fractions and identification of particular pol-
lution sources by factor analysis ŽKeiding et al.,
1986; Dillon and Goldstein, 1984.. However, to be
able to apply these statistical techniques, it is
necessary to perform many sample collections. It
is assumed that such analysis is valid if n q 50
samples were collected, where n is a number of
the determined elements ŽThurston and Spengler,
1985.. Therefore, if sampling time is 24 h and
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
n s 10, then sampling time at one sampling site
should be 120 days. Application of receptor mod-
elling was not necessary in the present work since
its main goal was the measurement of traffic
impact at the defined site in the city. Receptor
models have been used to identify sources of air
pollution in different cities, e.g. in the USA in
Los Angeles and New York ŽHenry and Hidy,
1979., Boston ŽHopke et al., 1976; Thurston and
Spengler, 1985., Salt Lake City, St. Louis ŽHenry
and Hidy, 1982., in the UK in Belfast, Birming-
ham, Bristol, Edinburgh, Leeds, Liverpool, Lon-
don, Newcastle ŽHarrison et al., 1996.. The stud-
ies in the UK report the percentage contribution
of each identified source to total concentration of
each component of total suspended particles and
traffic-related fraction of PM10. Although for
obvious reasons caution should be exercised in
comparing the data obtained in the present work
and the data resulting from multivariate analysis
in the above-mentioned study, it is interesting to
note that the contribution of lead Žone of the
markers of vehicular source. was estimated in the
latter at 60% while in the present work Žthough
obtained by different method. it was 70%, 1500 m
from the road.
The studies concerning analytical modelling of
the transport of particulate air pollutants utilise a
number of models describing behaviour of aero-
sols in air. Literature contains reports on ad-
vanced models, e.g. of Davidson and Slawson
Ž1982. that calculated effective source flux
parameters for use in the analytical plume rise
model, which describes transport from point
source. Spatial and temporal distribution of vehic-
ular pollutants from line and area sources was
modelled by Hao et al. Ž2000. in the urban area in
Beijing, in dependence on traffic intensity. This
study was based on the investigations of Sharan
ŽSharan et al., 1995; Sharan, 1996. who developed
the model for the dispersion of air pollutants in
low wind conditions in relation to distance, which
could be described by Gaussian-shaped curve.
Moreover, Huang et al. Ž2000. constructed a two-
dimensional air quality numerical model in an
urban street canyon which showed that the con-
centration distribution of pollutants emitted from
209
the street depends on inflow wind and street
canyon geometry. The model was tested using the
data from the experiments carried out in Japan.
The above-mentioned models can be applied in
future for explaining the behaviour of both aero-
sol fractions observed in the present study.
Besides model development, traffic-related
aerosol transport was studied in terms of horizon-
tal distribution of elemental and particulate con-
centrations. The study of Janssen et al. Ž1997.
demonstrated 1.3- and 2.6-fold decreases in PM2.5
and black smoke, respectively, and a two- to sev-
enfold drop in Si, Fe, Cu concentrations, at a
background site in compression with a location
near the road. They also showed significant posi-
tive correlation between elemental and particu-
late concentrations for most elements. The study
of Hitchins et al. Ž2000. investigated the changes
in the concentration of particles with the fol-
lowing EAD: 0.005᎐0.9 ␮m, 0.5᎐20 ␮m and 2.5
␮m ŽPM2.5 . over a distance 15᎐375 m from a
major road in Australia, dependent on wind di-
rection. The conclusion of the authors is: ‘ex-
posure to submicrometre particles is significantly
increased within the distance from 15 to 150 m
from a major road.’ The results for PM2.5 and
larger particles were similar. Therefore, the find-
ings of the aforementioned authors agree with
the results of the present work regarding the
coarse fraction. Differences in the behaviour of
the fine fraction between these two studies can be
explained by the various topography of the study
area and seasons of sample collection.
6. Conclusions
The main findings and conclusions of the pre-
sent study are:
䢇 The traffic component of coarse particle air
pollution was estimated at approximately 80%
up to 150 m from the road, ; 40% at 200᎐270
m, and ; 20% at 1500 m.
䢇 The region up to 150 m from the road is at the
highest risk of contamination with coarse par-
ticles from resuspended road dust; further-
more, the coarse particle air pollution from
210 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
the traffic is mostly confined to an area within
- 300 m from the road. Therefore, it is possi-
ble to reduce the pollution from traffic-related
coarse aerosol particles by changing traffic
management and by the construction of ap-
propriate shielding.
䢇 Transport of fine aerosol particles is much
more complicated due to their longer atmo-
spheric lifetime. The contribution from traffic
to the fine particle concentrations of individ-
ual elements was 53᎐68% in the close vicinity
of the road. However, the traffic-related fine
aerosol pollution affects the entire city. Re-
duction of this type of pollution is only possi-
ble by reducing the automotive emissions.
䢇 The approaches used in this work can be
applied in studies on other sources of coarse
particles. They result in making such studies
simple, flexible, and effective.
References
Artaxo P, Oyola P, Martinez R. Aerosol composition and
source apportionment in Santiago de Chile. Nucl Instr
Methods B 1999;150:409᎐416.
Bandhu HK, Puri S, Garg ML et al. Elemental composition
and sources of air pollution in the city of Chandigarh,
India, using EDXRF and PIXE techniques. Nucl Instr
Methods B 2000;160:126᎐138.
Cahill TA. Analysis of air pollutants by PIXE: the second
decade. Nucl Instr Methods B 1990;49:345᎐350.
Cornille P, Maenhaut W, Pacyna JM. PIXE analysis of size-
fractionated aerosol samples collected at Birkenes, Norway,
during spring 1987. Nucl Instr Methods B 1990;49:376᎐382.
Davidson GA, Slawson PR. Effective source flux parameters
for use in analytical plume rise models. Atmos Environ
1982;16:223᎐227.
Dillon WR, Goldstein M. Multivariate analysis ᎏ methods
and applications. New York: John Wiley & Sons, 1984:23.
˛d
´
Environmental Protection Year-book in Polish. Głowny Urza
Statystyczny, Warszawa, 1996:169.
Gerab F, Artaxo P, Gillett R, Ayers G. PIXE, PIGE and ion
chromatography of aerosol particles from northeast Ama-
zon Basin. Nucl Instr Methods B 1998;136r138:955᎐960.
Hao J, He D, Wu Y, Fu L, He K. A study of the emission and
concentration distribution of vehicular pollutants in the
urban area of Beijing. Atmos Environ 2000;34:453᎐465.
Harrison RM, Williams CR. Airbone cadmium, lead and zinc
at rural and urban sites in north-west England. Atmos
Environ 1982;16:2669᎐2681.
Harrison RM, Jones M. The chemical composition of airbone
particles in the UK atmosphere. Sci Total Environ
1995;168:195᎐214.
Harrison RM, Brimblecombe P, Derwent RG et al. Airborne
particulate matter in the United Kingdom. Third report of
the quality of urban air review group. Department of the
Environment. London, 1996;47:38᎐130.
Heidam NZ. Review: aerosol filtration by sequential filtration
by nuclepore filters. Atmos Environ 1981;15:891᎐904.
Henry RC, Hidy GM. Multivariate analysis of particulate
sulphate and other air quality variables by principal compo-
nents-I. Annual data from Los Angeles and New York.
Atmos Environ 1979;13:1581᎐1596.
Henry RC, Hidy GM. Multivariate analysis of particulate
sulphate and other air quality variables by principal compo-
nents-II. Salt Lake City, Utah and St. Louis Missouri.
Atmos Environ 1982;16:929᎐943.
Hitchins J, Morawska L, Wolff R, Gilbert D. Concentrations
of submicrometre particles from vehicle emissions near a
major road. Atmos Environ 2000;34:51᎐59.
Hopke PK, Gladney ES, Gordon GE, Zoller WH, Jones AG.
The use of multivariate analysis to identify sources of
selected elements in the Boston urban aerosol. Atmos
Environ 1976;10:1015᎐1025.
Huang H, Akutsu Y, Arai M, Tamura M. A two-dimensional
air quality model in an urban street canyon: evaluation and
sensitivity analysis. Atmos Environ 2000;34:689᎐698.
Janssen NAH, Van Mansom DFM, Van Der Jagt K, Harssema
H, Hoek G. Mass concentration and elemental composition
of airborne particulate matter at street and background
locations. Atmos Environ 1997;31:1185᎐1193.
´
Juda J, Chrusciel S. Protection of atmospheric air in Polish.
WNT, Warszawa, 1974:85.
Lodge JP, Waggoner AP, Klodt DT, Crain CN. Non-health
effects of air-borne particulate matter. Atmos Environ
1981;7:431᎐482.
Keiding K, Palmgren Jensen F, Heidam NZ. Absolute mod-
elling of urban aerosol elemental composition by factor
analysis. Analitica Chimica Acta 1986;181:79᎐85.
¨
Kuhler M, Kraft J, Bess H, Heeren U, Schurmann D. Compar-
ison between measured and calculated concentrations of
nitrogen oxides and ozone in the vicinity of a motorway. Sci
Total Environ 1994;146r147:387᎐394.
Maenhaut W, Selen A, Van Espen P, Van Grieken R, Winch-
ester JW. PIXE analysis of aerosol samples collected over
the Atlantic Ocean from a sailboat. Nucl Instr Methods
1981;181:399᎐405.
Maenhaut W, Raemdonck H, Andreae MO. PIXE analysis of
marine aerosol samples: accuracy and artifacts. Nucl Instr
Methods B 1987;22:248᎐253.
Manecki A, Kłapyta Z, Scheibal-Chwastek M, Skowronski ´ A,
Tarkowski J, Tokarz M. Influence of industrial air pollu-
tants on geochemical changes in Nipołomicki Forest Žin
Polish.. Pol Acad Sci, Prace Mineralogiczne 1981;71:18᎐25.
Orlic I, Wen S, Ng TH, Tang SM. Two years of aerosol
pollution monitoring in Singapore: a review. Nucl Instr
Methods B 1999;150:457᎐464.
Prati P, Putti P, Zucchiatti A, Lucarelli F, Mand PA. Study of
particulate emissions near a steel plant in Genova by
 ´ et al. r The Science of the Total En¨ ironment 257 (2000) 199᎐211
A. Wrobel
continuous sampling and PIXE hourly analysis. Nucl Instr
Methods B 1999;150:428᎐432.
Sharan M, Yadav A, Singh MP. Comparison of the sigma
schemes for estimation of air pollutant dispersion in low
winds. Atmos Environ 1995;29:2051᎐2059.
Sharan M. A mathematical model for the dispersion of air
pollutants in low wind conditions. Atmos Environ 1996;
30:1209᎐1220.
Thurston GD, Spengler JD. A quantitative assessment of
source contribution to inhalable particulate matter pollu-
tion in metropolitan Boston. Atmos Environ 1985;19:9᎐25.
´
Turzanski P. Soil pollution with heavy metals in Krakow ´
province in Polish. Biblioteka Monitoringu ´ Srodowiska,
1996:45.
211
´ P. Problems of traffic-related environmental pollu-
Turzanski
tion in Polish. Biblioteka Monitoringu ´Srodowiska, 1997:
223.
Williams ID, McCrae IS. Road traffic nuisance in residential
and commercial areas. Sci Total Environ 1995;169:75᎐82.
´ A, Rokita E, Kwiatek WM, Dutkiewicz E. The new
Wrobel
´
PIXE set-up at the Institute of Nuclear Physics in Krakow.
Nucl Instr Methods B 1996;109r110:109᎐112.
´
Wrobel A, Rokita E. Traffic component of urban aerosols
determined by PIXE method. Nucl Instr Meth B
1998;136r138:990᎐993.
´ A, Rokita E, Maenhaut W. The influence of meteoro-
Wrobel
logical parameters on short range transport of aerosols.
Nucl Instr Methods B 1999;150:403᎐408.