lable at ScienceDirect

Atmospheric Environment 45 (2011) 650e658

Contents lists avai

Atmospheric Environment

journal homepage: www.elsevier .com/locate/atmosenv

Contrast in air pollution components between major streets and backgroundlocations: Particulate matter mass, black carbon, elemental composition, nitrogenoxide and ultrafine particle number

Hanna Boogaard a,*, Gerard P.A. Kos b, Ernie P. Weijers b, Nicole A.H. Janssen c, Paul H. Fischer c,Saskia C. van der Zee d, Jeroen J. de Hartog a, Gerard Hoek a

a Institute for Risk Assessment Sciences (IRAS), Division Environmental Epidemiology, Utrecht, The Netherlandsb Energy Research Centre of the Netherlands (ECN), Department of Air Quality, Petten, The Netherlandsc Institute for Public Health and the Environment (RIVM), Center for Environmental Health Research, Bilthoven, The NetherlandsdMunicipal Health Service Amsterdam, Department of Environmental Health, Amsterdam, The Netherlands

a r t i c l e i n f o

Article history:Received 6 July 2010Received in revised form15 October 2010Accepted 20 October 2010

Keywords:Air pollutionElemental compositionStreetsContrastPM10

PM2.5

NO2

NOx

Black carbonChromiumCopperIron

* Corresponding author. Hanna Boogaard, IRAS, Dimiology, Jenalaan 18d, 3584 CK Utrecht, The Netherlafax: þ31 30 253 94 99.

E-mail address: j.m.c.boogaard@uu.nl (H. Boogaar

1352-2310/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.atmosenv.2010.10.033

a b s t r a c t

Policies to reduce outdoor air pollution concentrations are often assessed on the basis of the regulatedpollutants. Whether these are the most appropriate components to assess the potential health benefits isquestionable, as other health-relevant pollutants may be more strongly related to traffic.

The aim of this study is to compare the contrast in concentration between major roads and (sub)urbanbackground for a large range of pollutants and to analyze the magnitude of the measured difference inthe street e background for major streets with different street configurations.

Measurements of PM10, PM2.5, particle number concentrations (PNC), black carbon (BC), elementalcomposition of PM10 and PM2.5 and NOx were conducted simultaneously in eight major streets and nine(sub)urban background locations in the Netherlands. Measurements were done six times for a weekduring a six month period in 2008.

High contrasts between busy streets and background locations in the same city were found for chro-mium, copper and iron (factor 2e3). These elements were especially present in the coarse fraction of PM.In addition, high contrasts were found for BC and NOx (factor 1.8), typically indicators of direct combustionemissions. The contrast for PNCwas similar to BC. NO2 contrast was lower (factor 1.5). The largest contrastwas found for two street canyons and two streets with buildings at one side of the street only.

The contrast between busy streets and urban background in NO2 was less than the contrast found forBC, PNC and elements indicative of non-exhaust emissions, adding evidence that NO2 is not representing(current) traffic well. The study supports a substantial role for non-exhaust emissions including brake-and tyre wear and road dust in addition to direct combustion emissions. Significant underestimation ofdisease burden may occur when relying too much on the regulated components.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Numerous studies have documented the effects of air pollutionon morbidity and mortality from respiratory and cardiovasculardiseases (Pope and Dockery, 2006; WHO, 2006). Efforts to reduceair pollution emissions from stationary and mobile sources aresubstantial in Europe (Harrison et al., 2008). The evaluation of theeffectiveness and health benefits of traffic policies is usually

vision Environmental Epide-nds. Tel.: þ31 30 253 95 87;

d).

All rights reserved.

modeled instead of measured, focusing on regulated components.It is well-documented that proximity to traffic is much betterreflected in contrasts in particle number concentrations (PNC) orblack carbon (BC) than in PM10 and PM2.5 levels (Fischer et al.,2000; Harrison et al., 2004b; Janssen et al., 1997; Smargiassiet al., 2005). Omission of these components in an evaluation ofa traffic policy measure may lead to an underestimation of thehealth impacts of traffic policies. Often the easy to measure andregulated pollutant nitrogen dioxide (NO2) is used to characterizetraffic-related air pollution e.g. in the evaluation of the Londoncongestion charge zone (Tonne et al., 2008).

The issue of concentration contrasts of various pollutants is alsoof importance in the interpretation of studies that have shown

Fig. 1. Geographical location of the measurement sites in the Netherlands.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658 651

increased cardio-respiratory morbidity and mortality related toliving near major roadways (HEI, 2010). It has been argued thatultrafine particles may play a role in this association, partly basedupon the large contrast in concentration (Hoek et al., 2010).However, the contrast of other components may be even higher.

With the reduction of tailpipe emissions, interest in non-tail-pipe emissions is increasing. Johansson et al. (2009) reported highelemental concentrations of Cu and Sb in a street canyon comparedto an urban background location. Another study showed that inparticular Fe, Cu, Ba and Sbwere elevated at a roadside compared tobackground locations (Gietl et al., 2010). Currently few exposurestudies have measured an extensive set of traffic-related airpollutants at various urban roads simultaneously. From thedifferent elements, in particular Cu, Fe, and Zn are of toxicologicalinterest since these transition metals may play a role in the pro-inflammatory effects of particulate matter. There is also consider-able evidence that ultrafine particles and BC are related to adversehealth effects (Donaldson et al., 2005).

We recently started a study to evaluate the air quality and healthbenefits of environmental zoning traffic policies directed at heavyduty vehicles in several Dutch cities. The study includes compre-hensive measurements of air quality and population health statusprior to the implementation of the policy and two years after theimplementation. Here we report on the 2008 baseline air qualitymeasurements of traffic-related air pollution at moderately-busystreets, urban and suburban backgrounds locations. The aims of thepaper are to compare the quantitative contrast in concentrationbetween the major roads and background for a large range ofpollutants: PM10, PM2.5, PNC, BC, NO2, NOx and elemental compo-sition of PM10 and PM2.5; to analyze themagnitude of the measureddifference in the street e background for major streets withdifferent street configurations.

2. Methods

2.1. Study design

The evaluation of the air quality and health benefits of envi-ronmental zoning traffic policies in the Netherlands includescomprehensive measurements of air quality in the second half of2008 and 2010. The major component of the policy was directed atforbidding old heavy duty vehicles (minimal EURO 2) to enter theenvironmental zone of the city. Inner-city streets were selected thatwere expected to be affected by the policy. Eight street locationswere selected in five Dutch cities: Amsterdamwith about 750 000inhabitants, The Hague with about 480 000, Utrecht 280 000inhabitants, Den Bosch 136 000 and Tilburg 220 000 inhabitants(Fig. 1). Most important local source of air pollution in those cities istraffic, since in none of the cities, large industrial sources werepresent. Shipping can be another local source, especially in thechannels of Amsterdam. The smallest distance from a sampling sitein Amsterdam was 30 m away from a minor canal.

For each busy street location, a city background and a suburbanlocation in the neighbourhood of the city were chosen to disen-tangle temporal trends related to the local policy and more genericemission trends or weather patterns. In total 17 monitoring siteswere selected: eight busy streets, five urban background sites (oneper city) and four suburban background sites. One suburbanlocation was used for two nearby cities (Den Bosch, Tilburg). Anadditional monitoring site was selected as a reference site wherecontinuous measurements were made. At each site, one-weekmeasurements of PM10, PM2.5, BC, NO2 and NOx have been con-ducted six times during June 2008eJanuary 2009. All PM10 andPM2.5 filters were analysed for elemental composition usingenergy dispersive X-ray fluorescence spectrometry. PNC and the

size distribution of particles were measured at two busy streetsand the two nearby suburban backgrounds because of the healthstudy involving the population in the street and in the suburbanarea.

2.2. Characteristics of the monitoring sites

The eight street locations had a traffic intensity varying betweenapproximately 10 000 and 19 000motorized vehicles passing every24 h (Table 1). The sampling site was located between 9 and 15 mfrom the road axis. Two street locations can be considered ascanyon type of streets. Equipment at the street locations was placedoutside a dwelling or at a balcony at the first floor. Inlets wereplaced 4e5 m above ground level.

The five city background locations were located in the citycentres but not in the direct vicinity of busy streets (>50 m).Distance to the street sites ranged between 300 m (Koningsweg)and 4 km (Haarlemmerweg). The four suburban locations were invillages (w30 000 inhabitants) in the neighbourhood of theselected cities (10e30 km) and are background locations as well(Fig. 1). Nearest major road at the suburban background locationwas 100 m. Equipment of all but two of the background locationswas placed in the backyard of a dwelling and inlets of 7 backgroundlocations were placed 3e5 m above the ground. At two backgroundlocations, inlets were placed 1 m above ground level. The referencebackground site was located at the Utrecht University Campus, inthe centre of the country. None of the locations was near industrialareas, small industries, gas stations or large parking areas.

Table 1Detailed characteristics of the different streets.

Cities Streets Traffic intensity Fractiona Speed (km h�1) Road type Distance to road axis

Per 24-h Middle Heavy Per 24-h In rush hours

Amsterdam Haarlemmerweg 15 253 0.03 0.02 41 38 Adjoining buildings on one side 11Amsterdam Hoofdweg 9774 0.01 0.06 41 41 Adjoining buildings on two sides 16The Hague Stille Veerkade 17 438 0.05 0.02 34 32 Canyon 11Den Bosch Brugstraat 17 896 0.05 0.05 32 28 Canyon 9Den Bosch Koningsweg 17 138 0.05 0.03 46 39 Adjoining buildings on two sides 14Tilburg HVB 18 812 0.03 0.07 51 49 Adjoining buildings on two sides 12Utrecht Vleutenseweg 13 553 0.06 0.05 39 36 Adjoining buildings on two sides 15Utrecht Weerdsingel Wz 14 831 0.06 0.03 35 31 Adjoining buildings on one side 9

a Fraction of total traffic intensity. Traffic is characterized as middle traffic when the distance between the wheel axes of a vehicle was between 3.5 and 7 m.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658652

2.3. Ambient concentration measurements

Six one-week average measurements were conducted on 17locations. Due to equipment constraints locations were measuredin two rounds, Amsterdam plus the Hague sites in one round andDen Bosch, Tilburg and Utrecht in another round. For a validcomparison of the spatial contrast between sites, an additionalreference site was operated during all twelve sampling weeks tolimit potential bias due to temporal variation. This samplingschedule with discontinuous measurements spread over a year hasbeen used before to assess spatial variation in outdoor air pollution,often in the framework of providing the monitoring bases for landuse regression studies (Hoek et al., 2008). Though efficient andvalid estimates of the spatial contrast are obtained, the absoluteconcentrations can deviate from the annual average. In additionPNC and the size distribution of particles were measured at onestreet and suburban background location simultaneously.

2.3.1. PM10, PM2.5 and BCPM10 measurements were conducted with MSP PM10 sampling

heads. PM2.5 was measured with PM2.5 GK2.05 cyclones (BGI Inc.,Waltham, MA). Flow-controlled battery operated pumps (BGI-400)were used at a flow rate of 4 l min�1 (PM10) or 3.5 l min�1 (PM2.5).The sampling flow was measured before and after each samplingperiod using calibrated rotameters. In order to prevent filter over-load, timers were used to sample 5 min during every 20 min. Thusin one week, effectively a 42-h representative sample was taken.Elapsed-time indicators were used to calculate average samplevolumes. For all measurements the same type of filters were used(Anderson Teflon 37 mm).

Procedures for weighing and reflectance measurements havebeen described before (Brunekreef et al., 2005; Hoek et al., 2002).Both before and after sampling, all filters were double weightedwith a Mettler MT5 micro-balance (Mettler-Toledo, Greifensee,Switzerland). Filters were stored at 4 �C and were weighted afterequilibrating for 24-h in a temperature (20e23 �C) and humidity(30e40%) controlled room. Reflectance of the PM10 filters wasmeasured at five standard spots and the average reflectance wascalculated. Next, the reflectance of the PM10 filters was transformedinto absorption coefficients (a) according to the following formula(ISO 9835, 1993):

a ¼ ðA=2VÞ*LnðR0=RsÞwhere a¼ absorption coefficients (10�5 m�1), A¼ loaded filter area(m2), V ¼ sampled volume (m3), R0 ¼ average reflectance of fieldblank filters, and Rs ¼ reflectance of the sampled filter.

Absorption coefficients have been shown to be highly correlatedwith measurements of elemental carbon (EC) (Cyrys et al., 2003),consistent with the observation of high correlation betweenAethaolometer BC measurements using light transmission and EC

(ten Brink et al., 2004). In this paper wewill further use the term BCfor the absorption coefficient measurements.

For quality assurance 12 field blanks and 19 duplicate sampleswere collected at multiple sites on a rotating basis. Mean fieldblank weight changes were subtracted from all sample weights.

On two of our sites there were also fixed site monitors of NOx(chemiluminescence monitors, EcoPhysics, Switzerland, type CLD700AL) and PM10 (gravimetrically, low volume samplers, Derenda,Germany) running operated by the Municipal Health ServiceAmsterdam (urban background site) and Haarlemmerweg (streetsite). Therefore a comparison was made to test accuracy.

2.3.2. NOx

Average one-week NO and NO2 concentrations were measuredwith OGAWA passive samplers (OGAWA & Company USA, Inc.) withprecoated NO2 and NO collectionpads. Theywere placed at the sameheightnear the PMequipment. Sampling andanalysiswasperformedaccording to procedures reported before (Van Roosbroeck et al.,2007). 21 duplicate samples were taken to test precision. From eachbatchof40samples, four labblankswere retained. Inaddition, in total20 field blanks were taken spread over the sampling period. Batchaverage blank levels were subtracted from all samples.

2.3.3. Elemental compositionAll filters were analysed for different elemental components

with energy dispersive X-ray fluorescence spectrometry (ED-XRF)at the Cooper Environmental Services, Portland, USA.

Quality assurance included testing of standardmaterial from theNational Institute of Standards and Technology (SRM 1228, SRM987) and with a Micromatter Multi-elemental Quality ControlStandard (Multi 13391). Measurements of certain elements oncertified contents of standardmaterial showed percentages of errorup to 7% except for Cu. Cu was underestimated with about 13%. Wehave not corrected for this since there is no indication that thisdiffers between locations. To test the precision of the ED-XRFanalysis 18 (9%) of the samples have been re-measured. In thispaper, 16 elements are reported which were consistently above thelimit of detection and had a good reproducibility. Mean field blankconcentrations were subtracted from all samples. Details of theelemental analyses have been reported earlier (EPA, 1999).

2.3.4. Size specific particle number concentrationsAdditional particle number measurements have been per-

formed at two street locations (Haarlemmerweg, Amsterdam andVleutenseweg, Utrecht) and the two matching suburban locationfor 4 and 3 weeks respectively in the period September2008eJanuary 2009. Particle size distributions were measuredusing a Scanning Mobility Particle Sizer (SMPS) comprising a TSIModel 3071 Electrostatic Classifier and Model 3022 CondensationParticle Counter. Procedures have been reported before (Mirme

Table 2Averages of different traffic-related air pollutants (mg m�3) at the various streets andmedian ratio street vs urban background location in the same city.

Streets PM10 PM2.5 BCa NO2 NOx

Haarlemmerweg 27.5 17.8 4.1 54.3 100.51.2 1.2 2.1 1.5 2.0

Hoofdweg 22.4 15.1 2.7 47.5 75.61.1 1.0 1.7 1.3 1.5

Stille Veerkade 32.2 19.4 4.4 54.1 109.91.3 1.3 2.1 2.0 2.8

Brugstraat 32.9 18.0 3.7 53.8 99.31.3 1.3 2.1 2.0 2.8

Koningsweg 30.2 17.4 2.8 44.2 78.31.1 1.2 1.9 1.6 2.1

HVB 29.4 17.2 2.4 36.6 60.61.0 1.1 1.5 1.3 1.4

Vleutenseweg 25.9 15.4 2.1 41.6 70.31.0 1.1 1.5 1.2 1.6

Weerdsingel Wz 29.1 16.6 3.4 52.4 87.81.2 1.1 2.2 1.5 2.0

Overall ratiob 1.2 1.2 1.9 1.5 1.8P value <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

First row is concentration. Second row is ratio with respective urban background.a Black carbon (BC) is actual absorbance (‘soot’) and is expressed in 10�5 m�1.b Calculated as the median ratio street vs urban background location for the

individual samples (N PM10 ratio ¼ 45, N PM2.5 ratio ¼ 40, N BC ratio ¼ 45, N NOratio ¼ 48). Overall ratio tested on significant with the use of the non-parametricsigned rank test.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658 653

et al., 2002). The SMPS distinguished particles with a diameter of16e470 nm in 48 predefined classes. From the derived size distri-butions, PNC were calculated.

2.4. Traffic counts and meteorological data

Automated traffic counts were done at the street locations forone week in November 2008 by making use of a standardMinuteman EVR (Counters & Accessories Ltd, England), which isa two-tube counter. Apart from total traffic counts, speed andclassification (light, middle and heavy) of the motorized vehicleshas been registered as well.

Data on hourly standard meteorological parameters wereobtained from the Royal Dutch Meteorological Institute, the Biltlocated in the middle of the Netherlands (<5 km from Utrecht), tocharacterize the weather conditions during the sampling weeks.

2.5. Data management and analysis

To limit potential bias in the comparison between locations dueto temporal variation, an adjustment was made using data from thecontinuous reference site (Hoek et al., 2002). Correlation of thestandard air pollutants was high between all sites, thus the criticalassumption that temporal patterns are correctly reflected at thereference site seems justified. This adjustment was not made forelemental concentrations, as correlation in time between sites wastypically quite low (<0.8) for most elements. Comparison of abso-lute elemental concentrations between streets can only be madefor sites in the same sample round.

For NO2 and NO all samples were available for analysis. Some PMmeasurements were lost for various technical reasons, leaving 209(92%) filters available for analysis. Most missing samples were atthe reference location (N ¼ 7). In the absence of available fixed sitemonitors from the National Air Quality Monitoring Network in theneighbourhood, we have estimated missing PM reference valueswith the use of the nearest background site. We have madea regression line per pollutant and calculated predicted concen-trations. Correlation was high between those sites (>0.96).

After checking the SMPS data, 18 and 16 paired measurementdays were available for analysis. Missing data was due to powersupply problems, site-related problems and initial instrumentalproblems (cracked electrical isolator in classifier of SMPS).

Descriptive statistics were calculated. Median ratios betweenindividual simultaneously measured samples were calculated tocompare streets to the matching urban and suburban sites. Themedian is less influenced by outlier values. The overall median ratiowas tested on significance by the use of the non-parametric signedrank test.

2.6. Data quality

2.6.1. Limits of detectionsLimit of detection (LOD), calculated as three times the SD of all

field blanks divided by the nominal sampling volume, was3.5 mg m�3 for PM10, 0.5 mg m�3 for PM2.5 concentrations and0.7 10�5 m�1 for BC. None of the filters were below LOD. The LODswere 3.6 and 10.1 mg m�3 for NO2 and NOx respectively, afterexclusion of one very high field blank (NO2: 14.4 mg m�3). All NO2

measurements were above the LOD. All but one NOxmeasurementswere above LOD, after exclusion of one high field blank. Supple-ment Table 1 reports the LODs for the various elements. Allreported PM10 elemental concentrations were above LOD apartfrom some samples of Al (85% > LOD), Ba (72% > LOD) and V(99% > LOD). For the PM2.5 samples percentages detected weresomewhat lower ranging from 67% for Cl to 100% for Fe, Pb and S.

Values below the LOD were retained rather than replaced by e.g.0.5*LOD.

2.6.2. Precision and comparison with fixed site monitorsPrecision expressed as coefficients of variability (CV) from

duplicate measurements was good: PM10 (CV ¼ 6%), PM2.5 and BC(CV both 7%), NO2 (4%) and NOx (10%). We found somewhat higherPM10 (mean difference: 14%) and NO2 (mean difference: 8%) urbanbackground concentrations thenmeasured by thefixed sitemonitorin Amsterdam, running on exactly the same location. Correlationswere high (>0.93). NO2 concentrations were somewhat lower (4%)on the street location, compared to another fixed site monitor inAmsterdam, measured at the same street but closer to the kerb.

Supplement Table 2 shows CVs between replicate measure-ments for different elements. For the laboratory replicates, CVswere under the 10% for Ca, Cu, Fe, S, Si, and Zn for both fractions. Pbconcentrations of the PM2.5 fraction are not reported due to thelarge CV. For the PM10 field replicate measurements, CVs werebelow the 12.5% for Al, Ba, Ca, Cu, Fe, Mn, S, Si, and Zn. Ni has oneunexplained outlier field replicate measurement. The CVs for PM2.5replicates were mostly somewhat larger.

3. Results and discussion

3.1. Concentration differences between locations

3.1.1. PM mass, BC and NOx

Averages of PM10, PM2.5, BC, NO2 and NOx concentrations for thedifferent street locations are given in Table 2. Substantial contrastwas found between the moderately-busy streets and the respectiveurban background locations, especially for BC (factor 1.9) and NOx(factor 1.8). A small contrast (factor 1.2) was found for PM10 andPM2.5, which is in line with previous studies (Fischer et al., 2000;Harrison et al., 2004b; Janssen et al., 1997: Lianou et al., 2007;Smargiassi et al., 2005). The contrast for NO2 was in between PMand BC (factor 1.5). Table 2 further illustrates that the ratio between

Table

3Ave

rage

PM10

elem

entalco

ncentrations(ngm

�3)at

theva

riou

sstreetsan

dmed

ianratiostreet

vsurban

back

grou

ndlocation

inthesamecity.

Streets

Rou

nd

Al

Ba

Br

Ca

Cl

Cr

Cu

FeMn

Ni

PbS

SiTi

VZn

Haa

rlem

merweg

A99

.217

.16.6

435.8

1017

.37.9

63.2

1316

.018

.42.9

11.0

1753

.344

8.8

21.8

3.4

52.3

1.3

0.8

1.1

1.3

1.3

3.5

5.1

3.4

1.7

1.2

0.6

1.1

1.2

1.1

0.7

1.4

Hoo

fdweg

A71

.38.9

7.3

365.0

1414

.64.7

28.4

664.9

11.2

2.8

10.6

1535

.436

9.0

21.6

3.6

43.1

1.3

0.5

1.1

1.2

1.1

2.0

2.4

1.8

1.2

1.2

0.8

1.0

1.3

1.1

0.9

1.3

StilleVee

rkad

eA

89.8

14.5

9.9

475.4

2148

.28.5

56.5

1332

.420

.35.0

13.5

1841

.853

3.7

27.0

6.1

1.7

2.6

1.3

1.1

1.5

1.5

2.9

3.6

2.8

2.5

1.2

1.4

1.2

1.7

1.7

1.0

1.7

Brugstraa

tB

52.3

14.8

6.8

276.3

1553

.58.7

47.6

1072

.415

.13.3

12.6

1153

.333

8.9

11.8

3.4

55.1

3.2

3.8

1.2

1.3

1.2

2.4

3.6

2.9

1.8

1.4

1.1

1.1

1.5

1.3

1.0

1.3

Kon

ingsweg

B47

.410

.26.0

587.3

1104

.35.1

35.6

908.0

15.1

2.4

11.3

1210

.942

3.6

15.1

3.7

67.1

1.7

2.1

1.1

1.0

1.0

1.5

2.1

2.0

1.4

1.2

1.0

1.1

1.8

1.4

0.9

1.3

HVB

B42

.19.1

7.3

354.2

1208

.95.3

33.1

734.0

12.6

3.3

16.4

1266

.831

9.4

14.4

4.9

50.2

1.1

1.2

1.1

1.2

1.1

2.2

2.6

2.6

1.5

1.1

1.3

1.1

1.3

1.5

1.1

1.3

Vleutenseweg

B23

.95.3

5.5

311.0

1081

.93.7

18.8

439.2

10.7

2.6

7.6

986.5

260.8

12.2

3.8

37.4

1.0

0.6

0.9

1.1

1.1

1.3

1.5

1.1

1.0

0.9

1.0

1.0

1.2

1.0

0.9

1.0

Wee

rdsinge

lWz

B33

.99.4

6.1

331.3

1432

.95.7

40.7

997.7

15.5

3.3

7.7

1058

.030

9.3

15.4

3.8

46.0

1.2

0.9

1.0

1.1

1.1

2.3

3.1

2.7

1.6

1.1

0.8

1.0

1.4

1.1

0.9

1.1

Ove

rallratioa

1.4

1.2

1.1

1.2

1.2

2.2

2.8

2.5

1.5

1.2

1.1

1.1

1.3

1.3

0.9

1.4

Pva

lue

0.01

0.45

0.04

0.00

020.00

8<0.00

01<0.00

01<0.00

01<0.00

010.07

0.14

0.02

0.00

20.04

0.07

<0.00

01

Firstrow

isco

ncentration.S

econ

drow

isratiowithresp

ective

urban

back

grou

nd.

aCalcu

latedas

themed

ianratiostreet

vsurban

back

grou

ndlocation

fortheindividual

samples(N

ratio¼

45).Ratio

tested

onsign

ificantwiththeuse

ofthenon

-param

etricsign

edranktest.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658654

busy street and respective urban background varies substantiallybetween streets, especially for BC. For BC ratios above 2.0 are foundfor the two canyon streets (Stille Veerkade and Brugstraat) and thetwo streets which are built on one side of the street only (Haar-lemmerweg, Weerdsingel Wz). Section 3.3 further discusses thesedifferences.

3.1.2. Elemental compositionSubstantial differences between streets were found in the

elemental composition of PM10 (Table 3). Especially, Cr, Cu and Fewere clearly elevated at all street locations (overall ratio with urbanbackground> 2.0). The highest ratioswere againobservedat the twocanyon and one-side built streets. Especially Cu differs substantiallyfrom a factor 1.5 at the Vleutenseweg to a factor 5.1 at the Haar-lemmerweg, both compared to background levels. The contrast ofthese three elements was much larger than found for Al, Si and Ti,frequently used indicators of soil dust, suggesting a traffic source.

In the PM2.5 fraction, only the elements Cu and Fe wereconsistently and highly elevated at all streets (Table 4). Cr and Bawere not detected sufficiently to be reported neither was Pb (highCV). Mn, Si and Ti are also somewhat higher at most street loca-tions. No consistent difference was found for the other elementalconcentrations. The increased concentration ratios for Cu and Fecould be explained as the tail of the coarse particle fraction sampledby the PM2.5 cyclone. Alternatively, it could reflect direct tailpipeemission of fine particles. Elemental concentrations in PM2.5samples were much smaller than in the PM10 samples, with theexception of Ni, S and V, elements typically associated with thePM2.5 fraction.

The contrasts in Fe and Cu were even more pronounced whencompared to the simultaneously measured suburban locations,especially in the PM10 fraction. Cu at the street locations rangedfrom a factor 2.1 to 6.5 compared to the matching suburban loca-tion (Supplement Tables 3e5).

3.1.3. Particle size distributionsParticle numbers and particle size distributions, measured in

two streets also differed substantially from the nearby suburbanlocation (Table 5, Fig. 2). PNC was more elevated at the one-sidedbuilt street location (Haarlemmerweg: factor 2.4), than at thewider, two sided built street location (Vleutenseweg: factor 1.3),consistent with the contrast for other components. At the Haar-lemmerweg, especially PNC in the 30e100 nm range were signifi-cantly elevated (factor 3.2), the typical size range for particlesfreshly emitted by motorized traffic. Particles at the Haarlemmer-weg were generally peaking at 40e50 nmwhereas at the suburbanlocation the peak was at a somewhat larger particle diameter(70e80 nm). At the Vleutenseweg and the suburban location, mostparticles were in the 60e70 nm range (Fig. 2).

3.2. Interpretation of contrasts in concentration of the variouspollutants

3.2.1. NO2 vs BC, PNC and elemental compositionThe contrast in NO2 between busy roads and background was

smaller than the contrast for BC, PNC and elements indicative ofnon-tailpipe emissions such as Cr, Cu and Fe. NO2 has often beenused as an indicator for traffic-related air pollution in epidemio-logical studies and traffic management evaluations (Tonne et al.,2008). NO2 is widely used because an air quality guideline existsand it is easy to measure with passive samplers. Some indicativeevidence that NO2 is not representing (current) traffic well comesfrom a few studies that documents stagnant NO2 concentrationsover time, despite considerable reductions in NOx concentrations(Carslaw, 2005; Keuken et al., 2009). We were not able to

Table 4Average PM2.5 elemental concentrations (ng m�3) at the various streets and median ratio street vs urban background location in the same city.

Streets Round Al Br Ca Cl Cu Fe Mn Ni S Si Ti V Zn

Haarlemmerweg A 29.3 4.2 40.4 159.2 9.1 200.1 4.0 1.8 1306.7 86.6 3.4 2.7 24.60.8 1.0 1.3 1.3 3.2 2.2 1.6 1.2 1.0 1.1 0.8 1.1 1.3

Hoofdweg A 23.8 4.3 36.3 194.7 4.5 116.8 2.4 1.3 1377.9 146.3 3.2 1.9 30.91.1 1.0 1.2 1.1 1.8 1.5 1.0 1.0 1.0 1.8 0.7 0.8 1.1

Stille Veerkade A 15.1 4.7 36.8 212.7 7.4 159.2 4.0 2.3 1426.6 60.4 2.6 3.7 24.00.9 0.9 1.6 1.3 1.8 1.6 1.3 0.9 1.0 1.2 1.7 1.0 1.2

Brugstraat B 17.6 2.8 33.4 88.7 5.2 140.4 3.3 1.3 831.1 105.0 1.1 1.7 28.21.3 0.6 1.2 1.0 2.4 1.9 2.0 0.6 1.0 1.4 0.6 0.6 1.1

Koningsweg B 16.7 2.9 53.4 62.0 6.9 153.0 3.0 1.1 872.8 116.8 1.6 2.4 35.51.4 0.8 1.5 1.0 2.6 1.7 2.1 0.8 1.0 1.5 0.9 1.1 1.3

HVB B 11.3 3.6 33.2 108.2 5.0 125.7 2.0 1.8 888.8 44.9 0.7 2.8 22.80.9 0.8 0.9 0.9 2.3 2.0 1.2 1.4 1.0 1.1 0.0 1.2 1.0

Vleutenseweg B 13.9 3.6 37.4 58.2 3.1 97.4 2.7 1.7 765.9 52.5 2.0 2.8 18.21.0 1.3 0.9 1.0 1.4 1.3 1.0 1.0 1.1 1.7 0.9 1.1 1.6

Weerdsingel Wz B 10.1 5.4 33.4 63.9 5.5 142.2 3.1 1.4 764.4 58.0 1.7 2.5 20.41.1 1.3 1.0 1.0 2.8 2.0 1.6 0.9 1.0 1.4 2.0 1.1 1.5

Overall ratioa 1.0 1.1 1.0 1.0 2.3 1.9 1.4 1.0 1.0 1.2 0.8 1.0 1.1P value 0.72 0.43 0.15 0.46 0.0007 <0.0001 0.0007 0.87 0.64 0.0022 0.08 0.87 0.0064

First row is concentration. Second row is ratio with respective urban background.a Calculated as the median ratio street vs urban background location for the individual samples (N ratio¼ 40). Ratio tested on significant with the use of the non-parametric

signed rank test.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658 655

distinguish between primary (directly emitted) and secondary NO2

(formed secondary by NO and O3). Emission of primary NO2 isrising due to an increase in diesel cars and their particulate filters inEuropean countries. Our data add to the evidence that also withNO2, traffic-related contrasts are underestimated. The reported NO2contrast may differ from country to country due to the difference incar park, e.g. the fraction of diesel vehicles emitting higher fractionsof primary NO2.

3.2.2. Elemental composition and non-tailpipe emissionsThe contrast in Cu, Fe and Cr was at least as large as that in PNC

and BC. Cu at traffic sites originated mostly from non-tailpipeemissions, in particular from brake wear, together with Sb and Ba(Gietl et al., 2010; Lough et al., 2005; Sternbeck et al., 2002). Thecontributionof non-tailpipe emissions inPMmasshas been recentlyhighlighted (Thorpe and Harrison, 2008). Because of substantialefforts to reduce exhaust emissions, non-exhaust emissions havebecomemore important. Non-exhaust particles typically arise fromabrasive sources which include brake wear, tyre wear, road surfaceabrasion and resuspension in thewakeof passing traffic. Fe has oftenbeen associated with brake wear and resuspension of road dust aswell (Thorpe and Harrison, 2008). In addition, Fe could originatefrom direct tailpipe emissions. The elevated concentrations of Cr atthe streets are not totally understood, although Cr may reflectresuspended road dust or brake lining as well (Rodríguez et al.,2004). In addition, Cr can originate from a variety of other

Table 5Average particle number concentrations (p cm�3) at two streets and median ratiostreet vs suburban background location.

Streets N 16e30 nm 30e100 nm 100e470 nm Total

Haarlemmerweg 18 1257 10 727 4207 16 1910.8 (0.48) 3.2 (<0.0001) 2.3 (<0.0001) 2.4 (<0.0001)

Vleutenseweg 16 1540 6365 2538 10 4431.6 (0.02) 1.4 (0.0005) 1.4 (0.08) 1.3 (0.0005)

First row is concentration. Second row is ratio with respective suburban back-ground.Ratios tested on significant with the use of the non-parametric signed rank test. Inbrackets the p value.

(industrial) sources. However, our results suggests a common trafficsource, as Cr was highly correlated (>0.8) with more traffic-relatedindicators like NOx, NO2, BC, and in Cu and Fe.Mn is also higher at allstreet locations, but to a somewhat lesser extent (factor 1.5). Wefound only very modestly increased Zn concentrations at the streetsites (factor 1.4). Zn may derive from tyre wear, but has multipleatmospheric sources (Thorpe and Harrison, 2008), which may havecontributed to the relatively small contrast in our study. In none ofthe cities, large industrial sources were present of the elementsreported here. Recent measurements in London also did not showa large contrast between a major road and urban background for Zn(Gietl et al., 2010).

Typically, most elements are somewhat lower as has been foundin other recent studies investigating busier traffic roads (Dongarràet al., 2007; Gietl et al., 2010; Johansson et al., 2009; Oliveira et al.,2010). Hence sampling was often much closer to the road as well.Comparison of absolute concentrations is limited given the largedifferences in sample location, sampling period and time, andmonitoring equipment.

Concentration contrasts were also found in the PM2.5 fraction, inline with previous studies. Götschi et al. (2005) found substantiallyelevated concentrations of Ca, Cu, Fe, Mn, Ti and BC in PM2.5 parti-cles at an Antwerp busy street location, compared to a backgroundsite (factor 1.4e1.9) (Götschi et al., 2005). Hueglin et al. (2005)reported large contrasts in Fe, Cu and Sb (factor 7e9) betweena busy street in Bern and a suburban site in Basel in the PM10fraction. In the PM2.5 fraction those contrasts lowered considerably(Hueglin et al., 2005). Johansson et al. (2009) reported much higherratios than in the present study. The largest differencewas found forCu and Sb in the PM10 fraction, with on average more than seventimes higher levels in a street canyon compared to an urban back-ground location on a rooftop in Stockholm. Cr was elevated withabout a factor 2e3, comparable with our study. Fe was not reportedin the study (Johansson et al., 2009). Harrison et al. (2004b)reported at roadside sites a strong elevation in EC, organiccompounds and Fe-rich dust in comparison to the simultaneouslycollected urban background samples. Another study showed that inparticular Fe, Cu, Ba and Sb were elevated at a London roadsidecompared to background locations (factor 5e6) (Gietl et al., 2010).

0

5000

10000

15000

20000

25000

30000

10

Diameter (nm)

3^

mc

/n

.D

go

ld

/N

d

Haarlemmerweg

Suburban

0

2000

4000

6000

8000

10000

12000

14000

16000

Diameter (nm)

3^

mc

/n

.D

go

ld

/N

d

Vleutenseweg

Suburban

1000100

10 1000100

Fig. 2. Mean particle size distribution measured by SMPS at two street locations andthe respective suburban background locations. Note that the scales are different.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658656

3.2.3. Particle number contrastsThe contrast in PNC at the two streets where PNCwas measured

was slightly smaller than observed for BC and concentrations of Cuand Fe (Supplement Tables 3e5). An additional study at the samestreet in Utrecht (Weerdsingel Wz) also showed slightly smallercontrasts for PNC than for BC (Boogaard et al., 2010). PNC and BCwere both more than three times higher at the street locationcompared to the suburban location. Differences between thesuburban and urban background location (both the same as inpresent study) were much smaller and statistically non-significant(Boogaard et al., 2010).

Several studies have shown that the spatial gradient nearmotorways is largest for the nucleation mode particles (<10 nm)(Buonanno et al., 2009). The relatively low contrast in the smallestparticle range in our study might be the result of coagulation of thesmallest particles with the larger chainlike particles. In addition,the instrument collected particles from 16 nm, hence the particlessmaller than 16 nm were not sampled. The larger agglomeratedparticles (with a diameter of some hundreds of nm) have a rela-tively large surface where nucleation aerosol can easily deposit onby diffusion. Because there is much more of that type of aerosol instreet locations, the deposition by diffusion might be enhanced atthese locations with high aerosol concentrations.

Apart from the PM composition, other PM characteristics likeparticle size and surface area may determine the health risks posedbyPMaswell. Ultrafineparticlesmight playan important role in thisassociation, both because of the large PNC contrast and a rapiddecline with increasing distance from the motorway (Zhu et al.,2002). However, most studies of PNC near major roads were

conducted near or close to motorways up to 300 000 motorizedvehicles per 24-h (Weijers et al., 2004; Westerdahl et al., 2005; Zhuet al., 2002). There is much less information about PNC concentra-tion contrasts for less busy urban roads (Aalto et al., 2005; Lianouet al., 2007) to which a significant amount of urban inhabitantsare exposed. The latter studies suggested a modest contrast in PNCat urban streets as well, comparable with BC or NOx contrasts.

3.3. Factors influencing street concentrations

The observation that contrasts with urban background differedsubstantially between the eight traffic sites is probably explained bythe poorer dilution in street canyons; the larger fraction of windfrom the road to the sampling location in streets with only one sideof the road built and a smaller distance to the road. Traffic intensitieswere quite similar for all streets apart from the Hoofdweg showinglesser traffic (Table 1). The average speed was also lowest at thecanyon streets and the Weerdsingel Wz. The average speed in rushhours dropped significantly, and hardly any difference was shownbetween average daytime and rush hour speed, consistent with theobserved continuous amount of congestion and stop-and-go trafficduring daytime. Stop-and-go traffic increases emissions; especially(hard) acceleration contributes to this. Acceleration reducestemporarily the efficiency of catalytic converters (Fruin et al., 2008).

In the sampling period, average temperature was about 8 �C(2e18 �C). Wind directionwas predominantly south west (5 times),and mean wind speed was 3.6 m s�1 (2.5e4.6 m s�1). Especially,wind direction plays an important role in the large variationbetween measurements. The hourly weather data from the nearbystation de Bilt could not explain the observed contrasts betweendifferent streets. Field observations suggested that especially withwinds parallel to the street limited contributions of traffic werefound. We do not see different patterns of behaviour betweendifferent street types. Especially canyon type of streets is known fortheir enclosed nature leading to a de-coupling of the dispersionwithin the street canyon from those above the building canopy thatdetermine the background concentrations (Harrison et al., 2004a).At the two streets with adjoining buildings at one side, the windwas never coming from that side. With that wind, contrasts areexpected to be even greater due to recirculation and entrapment inthe lee of the buildings (Harrison et al., 2004a). There was noinformation available about subtle weather conditions at thespecific locations.

3.4. Limitations

A limitation of our sampling schedule with discontinuousmeasurements spread over a half year is that the absoluteconcentrations can deviate from the annual average. However,efficient and valid estimates of the spatial contrast between loca-tions are obtained (Hoek et al., 2008).We further could not afford toperform particle size monitoring on more than two of the eightstreets. Sampling height was generally the same for all but twobackground locations. Exact sampling height was slightly differentbetween the streets in Utrecht (4e5 m) and the suburban back-ground location (1 m). This was also the case for the Stille Veerkadeand the matching urban background location. The reportedcontrast between those street and the background sites wasprobably not substantially affected by this difference in samplingheight, as in the well-mixed urban atmosphere, vertical gradientsare generally small for sampling sites away from local sources suchasmajor roads. In contrast, nearmajor roads large vertical gradientshave been identified (Väkevä et al., 1999).

The background locations were chosen away from busy streets(>50 m). Exact distance to nearest major roads at background

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658 657

locations differed. Distance to nearest major road was smallest atthe urban background sites for Amsterdam (50m,17 000motorizedvehicles per 24-h), Den Bosch and Utrecht (70 and 100 m, both16 000 motorized vehicles per 24-h). For the suburban backgroundlocations, distance to nearest major road was somewhat furtheraway (between 100 and 300 m). Difference in distance to nearestmajor road has unlikely had more than a minor influence on thereported contrasts, as all the locations were in residential areassurrounded by other buildings.

We measured only the ambient concentrations. Since peoplespend a large percentage of their time indoors, infiltration intoindoor air is important as well for actual exposure (Monn, 2001).Various studies have investigated indoor/outdoor relationships andidentified important determinants such as air exchange rate,building characteristics, filtration and particle size distribution andcomposition (Monn, 2001; Brunekreef et al., 2005; Sarnat et al.,2006). Infiltration rates of the different components differ. Forexample components associated with PM2.5 penetrate well intoindoor environments with typical percentages of 70e80%, althoughhighly depending on ventilation status. The infiltration of ultrafineparticles is associated with higher agglomeration, with moreagglomeration reported for the nucleation mode particlescompared to particles in the larger end of the ultrafine particlerange (Zhu et al., 2005).

4. Conclusion

Simultaneousmeasurements of an extensive set of air pollutantsat eight moderately-busy urban streets and urban and suburbanbackground locations, showed that the contrast between busystreets and urban background in NO2 was substantially less thanthe contrast found for BC, PNC and elements indicative of non-exhaust emissions, adding to indicative evidence that NO2 is notrepresenting (current) traffic well. The quantitative contrast for theelements Cr, Cu and Fe (factor 2.2 to 2.8) was even larger than thecontrast found for BC and NOx (factor 1.8), typically indicators ofdirect combustion emissions. This supports a substantial role fornon-exhaust emissions, with brake wear, tyre wear and road dustas the potentially important components. The contrast in PNCmeasured in two streets was similar to BC. The variability incontrast between streets was as large as that between streets andbackground locations, with the largest contrast found for two streetcanyons and two streets with buildings at one side of the streetonly. Several of the components showing a large contrast havedocumented health effects, such as BC, PNC and transition metalslike Cu and Fe, and may be associated ultimately with variousadverse health effects e in addition to PM10 effects. Significantunderestimation of disease burden may occur when relying toomuch on the regulated components.

Appendix. Supplementary material

Supplementary material related to this article can be foundonline at doi:10.1016/j.atmosenv.2010.10.033.

References

Aalto, P., Hameri, K., Paatero, P., Kulmala, M., Bellander, T., Berglind, N., et al., 2005.Aerosol particle number concentration measurements in five European citiesusing TSI-3022 condensation particle counter over a three-year period duringhealth effects of air pollution on susceptible subpopulations. J. Air WasteManag. Assoc. 55, 1064e1076.

Boogaard, H., Montagne, D., Brandenburger, A., Meliefste, K., Hoek, G., 2010.Comparison of short-term exposure to particle number, PM10 and sootconcentrations on three (sub) urban locations. Sci. Total Environ. 408,4403e4411.

Brunekreef, B., Janssen, N.A., de Hartog, J.J., Oldenwening, M., Meliefste, K., Hoek, G.,et al., 2005. Personal, indoor, and outdoor exposures to PM2.5 and its compo-nents for groups of cardiovascular patients in Amsterdam and Helsinki. Res.Rep. Health Eff. Inst., 1e70.

Buonanno, G., Lall, A.A., Stabile, L., 2009. Temporal size distribution and concen-tration of particles near a major highway. Atmos. Environ. 43, 1100e1105.

Carslaw, D.C., 2005. Evidence of an increasing NO2/NOX emissions ratio from roadtraffic emissions. Atmos. Environ. 39, 4793e4802.

Cyrys, J., Heinrich, J., Hoek, G., Meliefste, K, Lewne, M., Gehring, U., et al., 2003.Comparison between different traffic-related particle indicators: elementalcarbon (EC), PM2.5 mass, and absorbance. J. Expo. Anal Environ. Epidemiol 13,134e143.

Donaldson, K., Tran, L., Jimenez, L.A., Duffin, R., Newby, D.E., Mills, N., et al., 2005.Combustion-derived nanoparticles: a review of their toxicology followinginhalation exposure. Part Fibre Toxicol. 2 (10).

Dongarrà, G., Manno, E., Varrica, D., Vultaggio, M., 2007. Mass levels, crustalcomponent and trace elements in PM10 in Palermo, Italy. Atmos. Environ. 41,7977e7986.

EPA. Determination of metals in ambient particulate matter using X-RAY Fluores-cence (XRF) spectroscopy. Method IO-3.3. 1999.

Fischer, P.H., Hoek, G., van Reeuwijk, H., Briggs, D.J., Lebret, E., van Wijnen, J.H., et al.,2000. Traffic-related differences in outdoor and indoor concentrations ofparticles and volatile organic compounds in Amsterdam. Atmos. Environ. 34,3713e3722.

Fruin, S., Westerdahl, D., Sax, T., Sioutas, C., Fine, P.M., 2008. Measurements andpredictors of on-road ultrafine particle concentrations and associated pollut-ants in Los Angeles. Atmos. Environ. 42, 207e219.

Gietl, J.K., Lawrence, R., Thorpe, A.J., Harrison, R.M., 2010. Identification of brakewear particles and derivation of a quantitative tracer for brake dust at a majorroad. Atmos. Environ. 44, 141e146.

Götschi, T., Hazenkamp-von Arx, M.E., Heinrich, J., Bono, R., Burney, R., Forsberg, B.,et al., 2005. Elemental composition and reflectance of ambient fine particles at21 European locations. Atmos. Environ. 39, 5947e5958.

Harrison, R.M., Jones, A.M., Barrowcliffe, R., 2004a. Field study of the influence ofmeteorological factors and traffic volumes upon suspended particle mass aturban roadside sites of different geometries. Atmos. Environ. 38, 6361e6369.

Harrison, R.M., Jones, A.M., Lawrence, R.G., 2004b. Major component composition ofPM10 and PM2.5 from roadside and urban background sites. Atmos. Environ. 38,4531e4538.

Harrison, R.M., Stedman, J., Derwent, D., 2008. New directions: why are PM10concentrations in Europe not falling? Atmos. Environ 42, 603e606.

HEI, 2010. Traffic-related Air Pollution: a Critical Review of the Literature onEmissions, Exposure, and Health Effects. HEI Special Report 17. Health effectsInstitute, Boston, MA.

Hoek, G., Beelen, R., de Hoogh, K., Vienneau, D., Gulliver, J., Fischer, P., et al., 2008. Areview of land-use regression models to assess spatial variation of outdoor airpollution. Atmos. Environ. 42, 7561e7578.

Hoek, G., Boogaard, H., Knol, A., de Hartog, J.J., Slottje, P., Ayres, J.G., et al., 2010.Concentration response functions for ultrafine particles and all-cause mortalityand hospital admissions: results of a European expert panel elicitation. Environ.Sci. Technol. 44, 476e482.

Hoek, G., Meliefste, K., Cyrys, J., Lewne, M., Bellander, T., Brauer, M., et al., 2002.Spatial variability of fine particle concentrations in three European areas.Atmos. Environ. 36, 4077e4088.

Hueglin, C., Gehrig, R., Baltensperger, U., Gysel, M., Monn, C., Vonmont, H., 2005.Chemical characterisation of PM2.5, PM10 and coarse particles at urban, near-city and rural sites in Switzerland. Atmos. Environ. 39, 637e651.

ISO, 1993. Ambient air-determination of a black smoke Index (ISO 9835). Interna-tional Organization for Standardization.

Janssen, N.A.H., Van Mansom, D.F.M., Van Der Jagt, K., Harssema, H., Hoek, G., 1997.Mass concentration and elemental composition of airborne particulate matterat street and background locations. Atmos. Environ. 31, 1185e1193.

Johansson, C., Norman, M., Burman, L., 2009. Road traffic emission factors for heavymetals. Atmos. Environ. 43, 4681e4688.

Keuken, M., Roemer, M., van den Elshout, S., 2009. Trend analysis of urban NO2concentrations and the importance of direct NO2 emissions versus ozone/NOxequilibrium. Atmos. Environ. 43, 4780e4783.

Lianou, M., Chalbot, M.C., Kotronarou, A., Kavouras, I.G., Karakatsani, A.,Katsouyanni, K., et al., 2007. Dependence of home outdoor particulate mass andnumber concentrations on residential and traffic features in urban areas. J. AirWaste Manag. Assoc. 57, 1507e1517.

Lough, G.C., Schauer, J.J., Park, J.S., Shafer, M.M., Deminter, J.T., Weinstein, J.P., 2005.Emissions of metals associated with motor vehicle roadways. Environ. Sci.Technol 39, 826e836.

Mirme, A., Kreyling, W., Khlystov, A., ten Brink, H.M., Ruuskanen, J., Tuch, T., et al.,2002. Intercomparison of aerosol spectrometers for ambient air monitoring.Aerosol Sci. Technol. 36, 866e876.

Monn, C., 2001. Exposure assessment of air pollutants: a review on spatialheterogeneity and indoor/outdoor/personal exposure to suspended particulatematter, nitrogen dioxide and ozone. Atmos. Environ. 35, 1e32.

Oliveira, C., Pio, C., Caseiro, A., Santos, P., Nunes, T., Mao, H., et al., 2010. Road trafficimpact on urban atmospheric aerosol loading at Oporto, Portugal. Atmos.Environ. 44, 3147e3158.

Pope, C.A., Dockery, D.W., 2006. Health effects of fine particulate air pollution: linesthat connect. J. Air Waste Manag. Assoc. 56, 709e742.

H. Boogaard et al. / Atmospheric Environment 45 (2011) 650e658658

Rodríguez, S., Querol, X., Alastuey, A., Viana, M.M., Alarcón, M., Mantilla, E., et al.,2004. Comparative PM10ePM2.5 source contribution study at rural, urban andindustrial sites during PM episodes in Eastern Spain. Sci. Total Environ. 328,95e113.

Sarnat, S.E., Coull, B.A., Schwartz, J., Gold, D.R., Suh, H.H., 2006. Factors affecting theassociation between ambient concentrations and personal exposures to parti-cles and gases. Environ. Health Perspect. 114, 649e654.

Smargiassi, A., Baldwin, M., Pilger, C., Dugandzic, R., Brauer, M., 2005. Small-scalespatial variability of particle concentrations and traffic levels in Montreal:a pilot study. Sci. Total Environ. 338, 243e251.

Sternbeck, J., Sjödin, A., Andréasson, K., 2002. Metal emissions from road traffic andthe influence of resuspension: results from two tunnel studies. Atmos. Environ.36, 4735e4744.

ten Brink, H., Maenhaut, W., Hitzenberger, R., Gnauk, T., Spindler, G., Even, A., et al.,2004. INTERCOMP2000: the comparability of methods in use in Europe formeasuring the carbon content of aerosol. Atmos. Environ. 38, 6507e6519.

Thorpe, A., Harrison, R.M., 2008. Sources and properties of non-exhaust particulatematter from road traffic: a review. Sci. Total Environ. 400, 270e282.

Tonne, C., Beevers, S., Armstrong, B., Kelly, F., Wilkinson, P., 2008. Air pollution andmortality benefits of the London Congestion Charge: spatial and socioeconomicinequalities. Occup. Environ. Med. 65, 620e627.

Väkevä, M., Hämeri, M., Kulmala, M., Ruuskanen, J., Laitinen, T., 1999. Street levelversus rooftop concentrations of submicron aerosol particles and gaseouspollutants in an urban street canyon. Atmos. Environ. 33, 1385e1397.

Van Roosbroeck, S., Jacobs, J., Janssen, N.A.H., Oldenwening, M., Hoek, G.,Brunekreef, B., 2007. Long-term personal exposure to PM2.5, soot and NOx inchildren attending schools located near busy roads, a validation study. Atmos.Environ. 41, 3381e3394.

Weijers, E.P., Khlystov, A., Kos, G.P.A., Erisman, J.W., 2004. Variability of particulatematter concentrations along roads and motorways determined by a movingmeasurement unit. Atmos. Environ. 38, 2993e3002.

Westerdahl, D., Fruin, S., Sax, T., Fine, P.M., Sioutas, C., 2005. Mobile platformmeasurements of ultrafine particles and associated pollutant concentrations onfreeways and residential streets in Los Angeles. Atmos. Environ. 39, 3597e3610.

WHO, 2006. Air Quality Guidelines. Global Update 2005. WHO Regional Office forEurope, Copenhagen.

Zhu, Y.F., Hinds, W.C., Kim, S., Shen, S., Sioutas, C., 2002. Study of ultrafine particlesnear a major highway with heavy-duty diesel traffic. Atmos. Environ. 36,4323e4335.

Zhu, Y.F., Hinds, W.C., Krudysz, M., Kuhn, T., Froines, J., Sioutas, C., 2005. Penetrationof freeway ultrafine particles into indoor environments. J. Aerosol Sci. 36,303e322.