Scholarly article on topic 'Concentrations and emission factors for PM2.5 and PM10 from road traffic in Sweden'

Concentrations and emission factors for PM2.5 and PM10 from road traffic in Sweden Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Atmospheric Environment
Keywords
{"PM10 " / "PM2.5 " / "Air dispersion" / Gothenburg / Umeå / "Traffic-related air pollution"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Martin Ferm, Karin Sjöberg

Abstract PM10 concentrations exceed the guidelines in some Swedish cities and the limit values will likely be further reduced in the future. In order to gain more knowledge of emission factors for road traffic and concentrations of PM10 and PM2.5, existing monitoring stations in two cities, Gothenburg and Umeå, with international E-road thoroughfares, were complemented with some PM2.5 measurements. Emission factors for PM10 and PM2.5 were estimated using NOX as a tracer. Monitoring data from kerbside and urban background sites in Gothenburg during 2006–2010 and in Umeå during 2006–2012 were used. NOX emissions were estimated from the traffic flow and emission factors calculated from the HBEFA3.1 model. PM2.5 constitutes the finer part of PM10. Emissions of the coarser part of PM10 (PM10–PM2.5) are suppressed when roads are wet and show a maximum during spring when the roads dry up and studded tyres are still used. Less than 1% of the road wear caused by studded tyres give rise to airborne PM2.5–10 particles. The NOX emission factors decrease with time in the used model, due to the renewal of the vehicle fleet. However, the NOX concentrations resulting from the roads show no clear trend. The air dispersion is an important factor controlling the PM concentration near the road. The dispersion has a minimum in winter and during midnight. The average street level concentrations of PM10 and PM2.5 in Gothenburg were 21 ± 20 and 8 ± 6 μg m−3 respectively, which is 36% and 22% higher than the urban background concentrations. Despite the four times lower traffic flow in Umeå compared to Gothenburg, the average particle concentrations were very similar; 21 ± 31 and 7 ± 5 μg m−3 for PM10 and PM2.5 respectively. These concentrations were, however, 108% and 55% higher than the urban background concentrations in Umeå. The emission factors for PM10 decreased with time, and the average factor was 0.06 g km−1 vehichle−1. The emission factors for PM2.5 are very uncertain due to the small increments in PM2.5 concentration at the thoroughfares, and were on average 0.02 g km−1 vehichle−1.

Academic research paper on topic "Concentrations and emission factors for PM2.5 and PM10 from road traffic in Sweden"

ELSEVIER

Contents lists available at ScienceDirect

Atmospheric Environment

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

Concentrations and emission factors for PM2.5 and PM10 from road traffic in Sweden

Martin Ferm*, Karin Sjöberg

IVL Swedish Environmental Research Institute Ltd, P.O. Box 53021, SE-400 14 Gothenburg, Sweden

HIGHLIGHTS

• Less than 1% of the road wear caused by studded tyres produce airborne PM10 particles.

• The PM10 emission factor has a maximum in spring when studded tyres are still used.

• NOX concentrations do not decrease despite renewal of the vehicle fleet.

ARTICLE INFO

ABSTRACT

Article history:

Received 13 January 2015

Received in revised form

12 August 2015

Accepted 13 August 2015

Available online 15 August 2015

Keywords:

Air dispersion

Gothenburg

Traffic-related air pollution

PM10 concentrations exceed the guidelines in some Swedish cities and the limit values will likely be further reduced in the future. In order to gain more knowledge of emission factors for road traffic and concentrations of PM10 and PM2 5, existing monitoring stations in two cities, Gothenburg and Umea, with international E-road thoroughfares, were complemented with some PM25 measurements. Emission factors for PM10 and PM2.5 were estimated using NOX as a tracer. Monitoring data from kerbside and urban background sites in Gothenburg during 2006—2010 and in Umea during 2006—2012 were used. NOX emissions were estimated from the traffic flow and emission factors calculated from the HBEFA3.1 model. PM2 5 constitutes the finer part of PM10. Emissions of the coarser part of PM10 (PM10—PM2 5) are suppressed when roads are wet and show a maximum during spring when the roads dry up and studded tyres are still used. Less than 1% of the road wear caused by studded tyres give rise to airborne PM2.5—10 particles. The NOX emission factors decrease with time in the used model, due to the renewal of the vehicle fleet. However, the NOX concentrations resulting from the roads show no clear trend. The air dispersion is an important factor controlling the PM concentration near the road. The dispersion has a minimum in winter and during midnight. The average street level concentrations of PM10 and PM2.5 in Gothenburg were 21 ± 20 and 8 ± 6 mg m-3 respectively, which is 36% and 22% higher than the urban background concentrations. Despite the four times lower traffic flow in Umea compared to Gothenburg, the average particle concentrations were very similar; 21 ± 31 and 7 ± 5 mg m-3 for PM10 and PM2 5 respectively. These concentrations were, however, 108% and 55% higher than the urban background concentrations in Umea. The emission factors for PM10 decreased with time, and the average factor was 0.06 g km-1 vehichle-1. The emission factors for PM2 5 are very uncertain due to the small increments in PM2 5 concentration at the thoroughfares, and were on average 0.02 g km-1 vehichle-1.

© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Airborne particles of the size fractions PM25 and PM10 are associated with health problems, such as cardiovascular disease and chronic obstructive pulmonary disease (Caiazzo et al., 2013;

* Corresponding author. E-mail address: Martin.Ferm@ivl.se (M. Ferm).

Brunekreef and Forsberg, 2005). However, health risks related to wear particles have not yet received much attention (Denier van der Gon et al., 2013). The main source of PM25_10 and NOX in Swedish cities is the local urban traffic (Sjödin et al., 2010). Longrange transport of particles belonging to the accumulation mode is, however, the main source of urban PM2 5 in the southern parts of Sweden (Furusjö et al., 2007). A high background concentration of PM2 5 is not only an issue when measures to reduce the particle concentration are to be taken, but also when PM2 5 emissions from

http://dx.doi.org/10.1016/j.atmosenv.2015.08.037

1352-2310/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

the road are quantified.

Vehicles mainly emit PM25 from the exhaust pipe while road abrasion mainly gives rise to larger wear particles (PM2.5-10). Particulate matter from road traffic also comes from other sources such as tyre wear, brake-wear and vehicle-induced resuspension of road dust (Lenschow et al., 2001; Abu-Allaban et al., 2003a; WHO, 2013). There is also a wind induced resuspension of particulate matter (Harrison et al., 2001). The term resuspension is used both for freshly formed particles from abrasion and older deposited road dust brought into the air.

Several approaches to measure emissions of wear particles have been used previously. Kupiainen et al. (2005) and Gustafsson et al. (2009) used a circular road simulator to measure emission factors. Cowherd and Englehart (1984) used a mass balance technique. They measured the particle flux on each side of the road by multiplying the vertical particle concentration profile with the wind speed. Kristensson et al. (2004) used a similar approach in a road tunnel.

The most common way to measure particle emission factors is by estimating the dispersion of the air by using a trace gas. A tracer can either be added (Claiborn et al., 1995; Kantamaneni et al., 1996) or a gas emitted from the vehicles, such as CO2, can be used as a tracer (Hallquist et al., 2013). The most commonly used tracer is NOX, which is emitted by the traffic (Nickel et al., 2013; Wang et al., 2010; Jones and Harrison (2006); Gehrig et al., 2004; Thorpe et al., 2007).

2. Experimental

Particle concentrations were measured with TEOM (Tapered Element Oscillating Microbalance) instruments, model 1400 AB. Different sampling heads were used for PM10 and PM2.5 according to the US Federal Register (1987). The results were corrected for evaporation, using an algorithm from the Swedish national reference laboratory: 1.19*measured mass concentration +1.15 mg m-3. The same algorithm was used in Sweden when PM10 concentrations were compared with the guidelines. The weight loss of collected particles due to the heating of the micro balance may vary with time. Negative hourly concentrations down to three times the standard deviation of the noise (-3 mg m-3) were therefore kept in the database.

Ketzel et al., 2007, argue that no correction for evaporation should be made when differences between concentrations near streets and urban background are calculated, as the weight losses are similar at the two sites. The particles emitted from the road, however, are not as volatile.

NOX was measured using chemiluminescent instruments with Mo-converters. In Umea, Monitor labs 9841 was used and at the urban background site in Gothenburg, an Ecophysics 700AL was used. At the E4 in Gothenburg, NOX was measured using a DOAS (Differential Optical Absorption Spectroscopy) instrument (OPSIS AR500) with the beam parallel to the road, starting from the point of particle measurements. The DOAS instrument was calibrated and serviced by the manufacturer. Hourly averages of all measured concentrations were registered and used in this study.

Two cities with thoroughfares belonging to the international E-road network were selected for this study. The E6 runs through Gothenburg in the south-western part of Sweden and the E4 runs through Umea situated in the north-eastern part of Sweden (Fig. 1). NOX, PM2.5 and PM10 concentrations were measured on an hourly basis at the two thoroughfares and the corresponding urban background stations.

Monitoring in Gothenburg was performed at the western roadside ofthe E6 (57°42'4'3.3"N, 11°59' 40.7"E), approximately 4 m from the kerbside, see Fig. 2. The E6 has three lanes in each direction and a total average traffic flow of 3800 vehicles per hour. The urban background site is situated on top of a roof 1.8 km north

Fig. 1. Map over Sweden showing the two cities and their closest regional measurement sites.

west of the measurement site at the E6, where also one of the two weather stations that are used in the study is situated. Precipitation quantities were monitored using a tipping bucket (Vaisala RG13H) and wind speed was monitored with an ultrasonic wind sensor (Gill instruments). Monitoring in Umea was performed at the eastern kerbside of the E4 in central Umea (63°49'43.9"N, 20°15'30.7"E), see Fig. 3. This thoroughfare has two lanes in each direction with a total average traffic flow of 900 vehicles per hour. The background site and the weather station are situated on top of a roof 400 m southeast of the street site and wind speed and direction were measured using an ultrasonic wind sensor (WindSonic M). The TEOM and chemiluminescent instruments were maintained by the retailer (Oleico). The traffic flow of different vehicle categories was measured by inductive-loop sensors embedded in the pavement. While this is not the most accurate technique, it was chosen due to the fact that it requires very little maintenance. Hourly averages of the speed and traffic flow were measured for each vehicle category in each lane in the two cities. The measurements started in 2006 in both cities and ended in 2010 in Gothenburg and in 2012 in Umea.

Measurements of PM2.5 and PM10 were made at two regional background sites (Rao near Gothenburg and Bredkalen which is the closest regional station to Umea with PM measurements). Regional background concentrations of NOX were not measured, but NO2 which is the major NOX species in regional background air in Sweden were measured at the two regional background sites, see Fig. .

2.1. Calculation of emission factors for PMX

When a tracer gas is used for calculation of an emission factor, it is assumed that the emitted particles and the trace gas are transported in exactly the same way to the measurement point at the kerbside.

Hourly emission factors in grams per km per vehicle (independent of the mix of vehicle categories) for PM2 5 and PM10 (EFPMx) can be calculated from hourly mean values of the emission factor for NOX (EFNOx) and the incremental NOX and PM concentrations at the roadside in comparison with the urban background concentrations (DPMX and DNOX), see Eq. (1). In the current study, the emission factor for NOX was calculated on an hourly basis from

Fig. 2. Map showing the two monitoring sites in Gothenburg and an additional site with meteorological measurements.

emission factors for the different vehicle categories (EFNOx,C) that were counted in each city, see Eq. (2). In the equation, FC is the traffic flow for category C in number of vehicles per hour in all lanes and both directions. Average emission factors for particulate matter for longer periods than 1 h can be calculated in several ways. The most common way is to use Eq. (3), in which an arithmetic mean is calculated from the same time interval that the data are stored.

'PMx = EFNOx

DPMX DNOX

NOx,C' FC

NOx = -

£Fc C

N xx eF

The average emission factor for the particles can for instance be used to estimate the incremental PMX concentration (DPMX) near the road. By combining Eqs. (1) and (2), Eq. (4) is obtained.

DPMx = EFpmx'E FcV EF F

q EFNOx,C' FC

When Eq. (3) is used to calculate the average emission factor, EFPMx, used in Eq. (4), hours with high emission factors will affect the average EFPMx to a large extent, causing Eq. (4) to overestimate the DPMX concentration for longer periods. By using longer average concentrations than in Eq. (3), a better agreement between estimated and observed DPMX is obtained. This calculation of EFPMx is shown in Eq. (5).

EF EF DPMx

EF PMx = EtNOx-==

A detailed description of how the emissions of NOX were calculated is given in the Supporting information. The average emission factors and how they vary from year to year depending on renewal of the vehicle fleet are given in Table 1.

Eq. (5) was therefore used in the current study to calculate the average emission factor for PMX that is illustrated in Figs. 6-11. Note that Eq. (5) is equal to Eq. (1) except for the averaging time. In

Fig. 3. Map showing the two monitoring sites in Umea.

Eq. (3), the averaging time was here 1 h whereas in Eq (5), the time can be chosen freely. For example, weekdays, a specific hour during all days for several years or a certain month during several years may be chosen as the averaging time. As the averaging period must be the same for all parameters in Eq. (5), only hours with complete data on all three parameters and positive values on the incremental concentration at the street level were used. On a monthly basis, the average emission factor calculated from Eq. (5) was of the order of 25—75% lower than that calculated from Eq. (3). Longer time periods than one month did not affect the relative difference between average emission factors for PMX calculated from Eq. (5) and Eq. (3) considerably. Eq. (5) resulted in more consistent averages compared to Eq. (3), when different time periods such as weekdays or hour of the day were compared.

2.2. Air dispersion

The total NOX emissions from all vehicle categories and lanes of the road (X] EFNOxC-FC) divided by the incremental NOX concentration thaC arises at the roadside (DNOX) can be expressed as a

Table 1

Emission factors used for NOX in Umea (upper part) and Gothenburg (lower part) expressed as g NO2 km-1 vehicle-1. The factor for trucks was calculated assuming 25% HGV and 75% LCV. Stop-and-go factors have been used when the average velocity was below 40 km h-1. The average emission factors have been weighted with the traffic flow in each category.

Category 2006 2007 2008 2009 2010 2011 2012

Passenger car 0.33 0.30 0.28 0.26 0.26 0.23 0.25

Stop-and-go 0.66 0.61 0.60 0.58 0.57 0.52 0.57

LCV 0.86 0.83 0.78 0.74 0.72 0.69 0.67

Stop-and-go 1.21 1.16 1.08 1.04 1.01 0.97 0.95

HGV 9.82 9.33 8.49 7.80 7.46 7.07 6.59

Stop-and-go 19.9 19.6 19.1 18.8 18.8 18.8 18.7

Weighted average 0.59 0.56 0.53 0.48 0.47 0.48 0.48

Passenger car 0.29 0.26 0.24 0.23 0.22

Stop-and-go 0.66 0.61 0.60 0.58 0.57

Lorry 2.5 2.4 2.2 2.0 1.9

Stop-and-go 5.9 5.8 5.6 5.5 5.4

Bus 6.3 6.0 5.7 5.3 4.7

Stop-and-go 18.3 17.7 17.3 16.5 15.3

Weighted average 0.53 0.53 0.51 0.45 0.45

Gothenburg

■ Street

□ City

□ Region

Fig. 4. Annual average contribution from the street, the city and the region to the concentrations of PM25 and PM10 in Gothenburg and Umeâ (2009-2010).

160 140 120 m 100

60 40 20 0

Gothenburg

NO2 NOx

NO2 NOx

Fig. 5. Annual average contribution from the street, the city and the region to the concentrations of NO2 and NOX (calculated as NO2) in Gothenburg and Umeâ (2009-2010). From the region, only NO2 concentrations are plotted.

Fig. 7. Monthly average emission factors for particles and air temperature in Umea (2006-2012).

■ Street Jl ¡> 0.10

□ City M 0.08

□ Region 0

O 0.06

J -L 0.04

0.02 0.00

PM10, dry PM10, wet PM2.5, dry PM2.5, wet

Fig. 8. Average seasonal variation of the emission factors for PM10 and PM2 5 for wet and dry road respectively (see the text) in Gothenburg.

-EFPM10 -EFPM2.S » air temp.

—i-1-1-1-1-1-1-1-1-1-1—

J FMAMJ JASOND

Fig. 6. Monthly average emission factors for particles and air temperature in Gothenburg (2006-2010).

dispersion factor (m2 s-1). The incremental particle concentration at the roadside equals the average particle emission factor multiplied by the traffic flow (= the particle emissions from the road) divided by the air dispersion factor (Eq. (4)). The dispersion factor is specific for the distance between the road and the measurement point. The measurement points have not been changed in any of the two cities during the entire measurement period.

^ 0.15

H 0.05

PM10, dry PM10, wet PM2.5, dry PM2.5, wet

-1-1-1-1-1-1-1-1-1-1-1-

J FMAMJ JASOND

Fig. 9. Average seasonal variation of the emission factors for PM10 and PM2 5 for wet and dry road respectively (see the text) in Umea.

3. Results and discussion

3.1. Traffic contribution

In Sweden, there are guidelines for the air concentrations of PM10, PM25, as well as for tail pipe emissions. The annual average PM10 concentration should not exceed 40 mg m-3 and the PM2.5 concentration limit is 25 mg m-3. In addition, the 90th percentile of

Time of day

Fig. 10. Average emission factors for PMi0 during the spring maximum (c.f. Fig. 6) and during the rest of the year in Gothenburg 2006—2010. The percentage of heavy vehicles (right axis) and EFNOx/3.5 and EFPM2.5 (left axes) are calculated as annual averages.

Time of day

Fig. 11. Average emission factors for PM10 during the spring maximum (c.f. Fig. 7) and during the rest of the year in Umea 2006—2012. The percentage of heavy vehicles (right axis) and EFNOx/3 and EFPM2.5 (left axes) are calculated as annual averages.

the diurnal average PM10 concentration should not exceed 50 mg m-3. There is no regulation for NOX concentrations in cities, however NO2 concentrations are regulated. The opposite is valid for vehicle emissions, for which no legislation for NO2 exists, however NOX emissions are regulated.

Fig. 4 illustrates the particle concentration caused by the thoroughfares, other sources in the city (urban background) and more distant sources (regional background) during two years for which data for all components are available simultaneously. As can be seen from the figure, the regional background concentrations are more important in Gothenburg than in Umea. In addition, the background concentrations are more important for PM2.5 than for PM10 in both cities, due to the fact that this particle fraction can be transported over longer distances than PM10. This implies that it is more difficult to quantify the local contribution of PM2.5 than of PM10 from the thoroughfares. The PM2.5 concentration is on average only 3.7 mg m-3 higher than the urban background in Gothenburg, seen over all investigated years, and 4.2 mg m-3 in Umea. These values are approximately three times the noise level for the TEOM instruments, which makes the results more uncertain.

Fig. 5 illustrates the concentrations of NO2 and NOX caused by the thoroughfares, other sources in the city and more distant sources during the same 2 years as for the data in Fig. 4. As can be seen from the figure, the regional background is insignificant in

both cities and is only 0.9 mg m-3 in Umea. The contribution from the street to the total concentration is higher for NOX than for particulate matter. Putaud et al. (2010) has published PM2 5 and PM10 concentrations in different parts of Europe. The urban background concentrations observed in the current study are in the lower range for North-western Europe according to those results.

3.2. Emission factors for particulate matter

3.2.1. Seasonal variations

There is no seasonal variation in the traffic flow. The emission factor for PM10 peaks in spring and has also a smaller maximum in autumn in Gothenburg (Fig. 6) as well as in Umea (Fig. 7). The peaks in spring and autumn are repeated every year in the two cities. The time and magnitude of the spring maximum vary annually, especially in Gothenburg. Studded tyres are only allowed from October to April in Sweden and there is a small peak in EFPM10 in both cities when the season for studded tyres begins. After this peak, the emission factor for PM10 decreases and is low in December and January before it starts to increase again. Roads are often wet during winter and it takes longer for them to dry up. Road salt that is used for de-icing purposes during winter also keeps the roads damp. An additional explanation to the very high peak in EFPM10 in spring could be a high frequency of dry roads and an accumulation of dust on the roads (Omstedt et al., 2005). Gustafsson et al. (2014) measured the amount of dust on roads in Stockholm and found a maximum in early spring.

Resuspension of road dust is a very efficient process (Gehrig et al., 2010) and the available dust is removed rapidly from the road (Nicholson and Branson, 1990). Another reason for the monthly PM10 emission peaks could be the use of road salt during the period when studded tyres are allowed. If there is road salt left on the street when it dries up, it can be emitted in large quantities as airborne particles (Ferm et al., 2006). NaCl is often used for de-icing purposes in Umea, while several different hygroscopic salts are occasionally applied for dust control in Gothenburg. Sampling for subsequent chemical analysis of PM10 has not been carried out at these sites. In Bavaria, PM10 exceeds the limit value in winter mainly due to the use of road salt (Schlachta et al., 2013). The use of traction sand can also cause formation of PM10 particles (Johansson et al., 2007; Kupianen et al., 2005; Kuhns et al., 2003). Studded tyres are forbidden in Germany and Croatia, and Ketzel et al. (2007) have compared the APMi0/ANOX ratios from streets in Copenhagen, Berlin, Leipzig, Halle, Stockholm, Helsinki and Klagenfurt. There are pronounced peaks in spring from the streets in Stockholm and Helsinki where studded tyres and salt are used, however not in Copenhagen, Berlin and Leipzig where salt but not studded tyres are used. Toth et al. (2011) have published PMi0 concentrations from a residential area in Zagreb, which did not show any spring or autumn peaks. Hussein et al., 2008 measured PM10 concentrations behind tyres using a mobile measurement system. They found 2.0—6.4 times higher concentrations behind studded tyres compared to friction tyres. Lowest concentrations were found behind summer tyres. It therefore seems likely that the use of studded tyres, traction sand and/or road salt are possible reasons for the peaks observed in the current study.

The emission factors for PM2.5 are not very accurate due to the small incremental concentration between street level and urban background. The emission factors for PM2 5 are mainly presented to distinguish particles coming from the exhaust pipes and brakes from that part of PM10 coming from the road. From Figs. 6 and 7 it is obvious that there is no peak in the emission factor for PM2.5 in the spring, which indicates that mainly particles larger than 2.5 mm are emitted from road wear. Gustafsson et al. (2008) investigated emissions using a road simulator and found that studded tyres emit

tens of times more particles than friction tyres. Contrary to our findings, they also found that there were significant emissions of particles with an aerodynamic diameter smaller than 2.5 mm.

The surface moisture was not monitored in the current study. However, the precipitation quantity was measured on an hourly basis in Gothenburg, and in Umea it was modelled using a statistical interpolation program (Hoaggmark et al., 2000) with Umea airport as the closest observation point. When there was precipitation the hour before the observation was carried out, the road was considered to be wet. When there was neither precipitation the hour before the observation nor during the observation, the road was considered to be dry. The results are shown in Figs. 8 and 9. The emission factor for PM10 is considerably lower for wet roads than for dry roads. For PM2.5, there is very little difference between wet and dry roads, indicating that most of the PM2.5 particles originate from the exhaust pipes. The difference between PM10 and PM2.5 is the coarse fraction of PM10 that comes from the road. The emissions of PM10 and PM2.5 are therefore plotted in the same scale in Figs. 6-11 to illustrate the mechanically formed fraction.

3.2.2. Variations during the week

During weekdays, the traffic flow is similar, however there is less traffic on Saturdays and Sundays. The daily average emission factors for both PM2.5 and PM10 during the weekends are, however, very similar to the factors for weekdays.

3.2.3. Diurnal variations

At night, the traffic flow is very low in both cities (midnight to 5 a.m.). It has a morning maximum around 7-10 a.m. and a global maximum around 4-5 p.m. In addition, heavy vehicles have a long flow maximum between 8 a.m. and 4 p.m. The heavy vehicle fraction, and thus the average NOX emission factor during free flow, has a maximum between midnight and 4 a.m. in Gothenburg and between 3 and 6 a.m. in Umea. The diurnal variation of EFPM10 in spring and during the rest of the year is shown in Figs. 10 and 11 respectively. The emissions are given in g km-1 vehicle-1 (independent of the vehicle category). Since there is a morning peak in the fraction of heavy vehicles (dotted lines), EFNOx in Eq. (2) will also peak (dashed bold lines). When APM10/ANOX is constant (heavier vehicles emit more PM10 per vehicle in a similar manner as they emit more NOX), EFPMx (see Eq. (5)) also shows a morning peak. To see if higher EFPM10 for HGV is due to tail pipe emissions or road emissions, EFPM2.5 was added to figs (the dashed fine lines in the bottom of the figs). The figs show that heavy goods vehicles emit more fine particles from the tail pipe and more coarse particles from the road compared to lighter vehicles.

The particle emissions have in some studies been estimated for LCV and HGV separately. For example, Gehrig et al. (2010) used e.g. a load simulator to distinguish between emission factors for LCV and HGV.

In Gothenburg, more PM10 is emitted in the afternoon in spring than during the rest of the year (Fig. 10). Possible reasons for this could be dryer roads due to higher temperature or higher air dispersion. In Umea, the elevated emission factor for PM10 during spring shows only a smaller afternoon peak compared to all the other months (Fig. 11). In both cities, EFNOx and DNOX show similar values in spring as during the rest of the year, while DPM10 concentrations are elevated during spring.

3.3. Temporal variation of air dispersion

Air dispersion is the relationship between emissions and concentration, and as it is the concentration of particles that must not exceed the guidelines, it is important that particles are not emitted

■H 9

C O) 6

Fig. 12. Seasonal variation of the air dispersion factor.

when the air dispersion is low.

The guidelines specify the total concentration (contribution from the road + contribution from the city + regional background). The contribution from the city is reduced by taking measures against emissions from all roads and other sources in the city. The dispersion factor is largely dependent on the meteorology (Rost et al., 2009) and will thus influence the contribution from the city in a similar manner as the road in question. The regional background is more difficult to influence and is significant in the southern parts of Sweden.

Fig. 12 illustrates the seasonal variation as monthly means of the dispersion factor. The factor is more than twice as high in the summer compared to the winter. This implies that a certain particle emission level will give rise to more than twice the concentration contribution in winter compared to summer in both cities. The dispersion factor in the spring, when the emission factor for PM10 has maximum, is close to the annual average dispersion factor.

The diurnal variation of traffic flow and air dispersion affects the diurnal variation of the air concentration. If the air dispersion were mainly caused by the meteorology, the diurnal dispersion factor would be a continuous function. The calculated average air dispersion factors for each hour of the day in Gothenburg and Umea are shown in Fig. 13. The morning peak of the air dispersion coincides with the morning rush hours. During weekends there is no morning rush hour and thus no morning peak of the air dispersion.

A combination of high traffic flow and low air dispersion causes episodes of very high DNOX concentrations near the roads. At the E6 in Gothenburg, such episodes are most common in December

14 12 10 8 6

0.5 ^ <

0 2 4 6 8 10 12 14 16 18 20 22 Time of day

Fig. 13. Average diurnal air dispersion in Gothenburg and Umea during the entire measurement periods.

Table 2

Average emission factors for PM10 and PM25 in g km-1 vehichle-1 in Gothenburg. The averages are calculated using Eq. (5) and the uncertainties as explained above.

EFPM10 Umeä EFPM2.5 Umeä EFPM10 Gothenburg EFPM2.5 Gothenburg

2006 0.088 ± - 0.012 0.027 ± 0.011 0.068 t 0.010 0.016 ± 0.009

2007 0.082 ± 0.009 0.023 ± 0.008 0.063 0.009 0.028 ± 0.008

2008 0.082 ± 0.008 0.023 ± 0.008 0.068 0.011 0.032 ± 0.010

2009 0.046 ± 0.006 0.013 ± 0.005 0.067 0.011 0.024 ± 0.010

2010 0.051 ± 0.006 0.014± 0.006 0.036 0.007 0.010± 0.006

2011 0.047 ± 0.007 0.020 ± 0.007

2012 0.049 ± 0.009 0.036 ± 0.008

average 0.064 ± 0.008 0.022 ± 0.008 0.060 0.009 0.021 ± 0.009

and between 7 a.m. and 9 a.m., when the DNOX can be as high as 1000 mg m-3 as an hourly mean. Such episodes also occur at the E4 in Umea with even higher concentrations of DNOX. They are most common during the same morning hours as in Gothenburg as well as between 4 p.m. and 6 p.m., and are most frequent in October to February.

3.4. Annual emission factors for PM25 and PM10

The annual emission factors for PM25 and PM10 in Gothenburg and Umeä, calculated using Eq. (5), are given in Table 2. As earlier mentioned the emission factor for PM2 5 is very uncertain due to the small incremental concentration. The accuracy for mass measurement is ±0.75% for the TEOM instruments and the noise level is 1 mg m~3. The measurement uncertainty was here estimated from the square root of the squared measurement uncertainty for PM2.5 at the street level plus the squared measurement uncertainty at the urban background level. This uncertainty is also shown in Table 2 for EFPM25 as well as for EFPM10.

For comparison, the same emission factors calculated with Eq. (3) are given in the Supporting information. The inter-annual variations are much smaller when Eq. (5) is used compared to when Eq. (3) is used, with one exception (EFPM10 in Umeä).

From Table 2 it can be seen that the EFPM10 decreases after 2008 in Umeä and after 2009 in Gothenburg. There is a small decrease in APM10 in Umeä and almost no change in ANOX between the two periods. The main reason for the decreasing EFPM10 in Umeä is the decrease in calculated EFNOx. In Gothenburg the main reason is the decrease in APM10 during 2010. There is no change in ANOX between the two periods and a smaller decrease in EFNOx in Gothenburg than in Umeä.

Ketzel et al. (2007) has compared different emission factors for PM2 5 and PM10, which are higher than those presented here. Jones and Harrison (2006) measured emission factors in a similar way as in this study but with the urban background station situated at street level on a different road instead of at roof level. They also used multiple linear regressions to differentiate the emission factors for LCV and HGV. Our results are between estimates for LCV and HGV. Bukowiecki et al. (2010) also differentiated between LCV and HGV and our average emission factors for PM10 are again within the range of their figures. Gehrig et al. (2004) measured the emission factor for PM10 at six streets in Zürich. Our results are similar to the street with the lowest emission factor in that study. Thorpe et al. (2007) measured the emission factors for PM10 and PM2 5-10 during four years in London using two urban background sites at street level. They found very similar emissions for PM2.5-10 as our results, but higher for the PM2.5 fraction. Abu-Allaban et al. (2003b) and Gertler et al. (2006) have published emission factor ranges that are higher than the averages in this study. Many of the published emission factors represent, however, shorter periods than one year. The standard deviation of the emission factor increases with decreasing measurement time, being small for annual

averages while for hourly averages it can be substantial.

The road wear is mainly caused by cars with studded tyres. The pavement rutting has been measured with a non-contact laser profilometer that was placed on fixed devices cast in the roadway, and a prediction model of wear caused by studded tyres has been developed (Jacobson and Wagberg, 2007). The total road wear for the E6 in Gothenburg can be estimated to 3.3 g/km per car with studded tyres and 2.2 g/km per car for the E4 in Umea (Jacobson, pers. comm.). The fraction of cars with studded tyres has been determined for parked cars in Gothenburg and Umea and was 71% in 2010 in Gothenburg and 97% in Umea in 2011. Studded tyres are not used on heavy vehicles. The road wear decreased during the '90 s when older steel studded tyres were replaced with tyres equipped with lightweight studs. The road wear increased later when the roads wearing courses were replaced with less wearing resistant aggregates to avoid polishing and thus reduced friction (Jacobson, 2007).

To estimate the fraction of the road abrasion caused by studded tyres that will give rise to PM10 particles, EFPM10-EFPM2.5 was calculated for the periods November—April and May—October (see Figs. 8—9). The difference between the two periods was 0.05 g km-1 vehichle-1 in Gothenburg and 0.02 g km-1 vehichle-1 in Umea. These emission factors for EFPM2.5—10 are only 1.5% and 0.8% of the total road wear caused by studded tyres in Gothenburg and Umea respectively. These percentages may also include emission of road salt and reemission of road dust and abrasion caused by e.g. sand on the roads.

4. Conclusions

Particles emitted from road wear are of a larger aerodynamic diameter than PM2.5 particles in the current study. The emission factor for PM10 has two annual maxima; a small one in the autumn when the use of studded tyres begins and a large maximum in spring prior to the period for studded tyres ends. In the period in between these maxima, the roads are often wet, which suppresses the emission of large particles to the air. Overall, less than 1% of the road wear caused by studded tyres are emitted to the air as PM2 5—10 particles. Furthermore, a certain emission level leads to higher concentrations in winter than in summer as well as higher concentrations at night than in daytime, due to a lower vertical mixing of the air when the lapse rate is smaller due to the lower irradiance.

Acknowledgements

We would like to thank the personnel at the Environmental Administration in Gothenburg and Department of Environmental Health, Umea municipality for operating all the urban measuring stations. The Swedish Environmental Protection Agency has funded the background stations. This project has been funded by the Swedish Transport Administration. We would also like to thank the anonymous reviewers for valuable comments.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2015.08.037.

References

Abu-Allaban, M., Gillies, J.A., Gertler, A.W., Clayton, R., Proffitt, D., 2003a. Tailpipe, resuspended road dust, and brake-wear emission factors from on-road vehicles. Atmos. Environ. 37, 5283-5293.

Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003b. Application of a multi-lag regression approach to determine on-road PM10 and PM2.5 emission rates. Atmos. Environ. 37, 5157-5164.

Brunekreef, B., Forsberg, B., 2005. Epidemiological evidence of effects of coarse airborne particles on health. Eur. Respir. J. 26, 309-318.

Bukowiecki, N., Lienemann, P., Hill, M., Furger, M., Richard, A., Amato, F., Prévôt, A.S.H., Baltensperger, U., Buchmann, B., Gehrig, R., 2010. PM10 emission factors for non-exhaust particles generated by road traffic in an urban street canyon and along a freeway in Switzerland. Atmos. Environ. 44, 2330-2340.

Caiazzo, F., Ashok, A., Waitz, I.A., Yim, S.H.L., Barrett, S.R.H., 2013. Air pollution and early deaths in the United States. Part I: quantifying the impact of major sectors in 2005. Atmos. Environ. 79, 198-208.

Claiborn, C., Mitra, A., Adams, G., Bamesberger, L., Allwine, G., Kantamaneni, R., Lamb, B., Westberg, H., 1995. Evaluation of PM10 emission rates from paved and unpaved roads using tracer techniques. Atmos. Environ. 29,1075-1089.

Cowherd Jr., C., Englehart, P.J., 1984. Paved Road Particulate Emissions. EPA-600/7-84-077. U.S. Environmental Protection Agency, Washington, District of Columbia.

Denier van der Gon, H.A.C., Gerlofs-Nijland, M.E., Gehrig, R., Gustafsson, M., Janssen, N., Harrison, R.M., Hulskotte, J., Johansson, C., Jozwicka, M., Keuken, M., Krijgsheld, K., Ntziachristos, L., Riediker, M., Cassee, F.R., 2013. The policy relevance of wear emissions from road transport, now and in the Future—An international workshop report and consensus statement. J. Air Waste Manag. Assoc. 63,136-149.

Ferm, M., Watt, J., O'Hanlon, S., De Santis, F., Varotsos, C., 2006. Deposition measurement of particulate matter in connection with corrosion studies. Anal. Bioanal. Chem. 384, 1320-1330.

Furusjo, E., Sternbeck, J., Anna Palm Cousins, A., 2007. PM10 source characterization at urban and highway roadside locations. Sci. Total Environ. 387, 206-219.

Gehrig, R., Hill, M., Buchmann, B., 2004. Separate determination of PM10 emission factors of road traffic for tailpipe emissions and emissions from abrasion and resuspension processes. Int. J. Environ. Pollut. 22, 312-325.

Gehrig, R., Zeyer, K., Bukowiecki, N., Lienemann, P., Poulikakos, L.D., Furger, M., Buchmann, B., 2010. Mobile load Simulators—A tool to distinguish between the emissions due to abrasion and resuspension of PM10 from road surfaces. Atmos. Environ. 44, 4937-4943.

Gertler, A., Kuhns, H., Abu-Allaban, M., Damm, C., Gillies, J., Etyemezian, V., Clayton, R., Proffitt, D., 2006. A case study of the impact of winter road sand/salt and street sweeping on road dust re-entrainment. Atmos. Environ. 40, 5976-5985.

Gustafsson, M., Blomqvist, G., Gudmundsson, A., Dahl, A., Swietlicki, E., Boghard, M., Lindbom, J., Ljungman, A., 2008. Properties and toxicological effects of particles from the interaction between tyres, road pavement and winter traction material. Sci. Total Environ. 393, 226-240.

Gustafsson, M., Blomqvist, G., Gudmundsson, A., Dahl, A., Jonsson, P., Swietlicki, E., 2009. Factors influencing PM10 emissions from road pavement wear. Atmos. Environ. 43, 4699-4702.

Gustafsson, M., Blomqvist, G., Janhall, S., Johansson, C., Norman, M., 2014. Operational Measures against PM10 Pollution in Stockholm - Evaluation of Winter Season 2012-2013. VTI report 802 (in Swedish).

Hallquist, À.M., Jerksjo, M., Fallgren, H., Westerlund, J., Sjodin À., 2013. Particle and gaseous emissions from individual diesel and cNg buses. Atmos. Chem. Phys. 13, 5337-5350.

Harrison, R.M., Yin, J., Mark, D., Stedman, J., Appleby, R.S., Booker, J., Moorcroft, S., 2001. Studies of the coarse particle (2.5-10 mm) component in UK urban atmospheres. Atmos. Environ. 35, 3667-3679.

Hussein, T., Johansson, C., Karlsson, H., Hansson, H.-C., 2008. Factors affecting non-tailpipe aerosol particle emissions from paved roads: on-road measurements in Stockholm, Sweden. Atmos. Environ. 42, 688-702.

Haggmark, L., Ivarsson, K.I., Gollvik, S., Olofsson, P.O., 2000. Mesan, an operational mesoscale analysis system. Tellus A 52, 1 -20.

Jacobson, T., 2007. Investigation of Pavement Wear of Studded Tyres in Stockholm during the Winter 2006/2007. VTI report 23-2007 (In Swedish).

Jacobson, T., Wägberg, L.-G., 2007. Developing and Upgrading of a Prediction Model of Wear Caused by Studded Tyres and an Overview of the Knowledge of the Factors Influencing the Wear. VTI report 7—2007 (In Swedish).

Johansson, C., Norman, M., Gidhagen, L., 2007. Spatial and temporal variations of PM10 and particle number concentrations in urban air. Environ. Monit. Assess. 127, 477—487.

Jones, A.M., Harrison, R.M., 2006. Estimation of the emission factors of particle number and mass fractions from traffic at a site where mean vehicle speeds vary over short distances. Atmos. Environ. 40, 7125—7137.

Kantamaneni, R., Adams, G., Bamesberger, L., Allwine, E., Westberg, H., Lamb, B., Claiborn, C., 1996. The measurement of roadway PM10 emission rates using atmospheric tracer ratio technique. Atmos. Environ. 30, 4209—4223.

Ketzel, M., Omstedt, G., Johansson, C., Düring, I., Pohjola, M., Oettl, D., Gidhagen, L., Wählin, P., Lohmeyer, A., Haakana, M., Berkowicz, R., 2007. Estimation and validation of PM25/PM10 exhaust and non-exhaust emission factors for practical street pollution modelling. Atmos. Environ. 41, 9370—9385.

Kristensson, A., Johansson, C., Westerholm, R., Swietlicki, E., Gidhagen, L., Ulla Wideqvist, U., Vesely, V., 2004. Real-world traffic emission factors of gases and particles measured in a road tunnel in Stockholm, Sweden. Atmos. Environ. 38, 657—673.

Kuhns, H., Etyemezian, V., Green, M., Hendrickson, K., McGown, M., Barton, K., Pitchford, M., 2003. Vehicle-based road dust emission measurements Part II: effect of precipitation, wintertime road sanding, and street sweepers on inferred PM10 emission potentials from paved and unpaved roads. Atmos. Environ. 37, 4573—4582.

Kupiainen, K., Tervahattu, H., Raisanen, M., Makela, T., Aurela, M., Hillamo, R., 2005. Size and composition of airborne particles from pavement wear, tires, and traction sanding. Environ. Sci. Technol. 39, 699—706.

Lenschow, P., Abraham, H.J., Kutzner, K., Lutz, M., Preu, J.D., Reichenbacher, W., 2001. Some ideas about the sources of PM10. Atmos. Environ. 35, S23—S33.

Nicholson, K.W., Branson, J.R., 1990. Factors affecting resuspension by road traffic. Sci. Total Environ. 93, 349—358.

Nickel, C., Kaminski, H., Bellack, B., Quass, U., John, A., Klemm, O., Kuhlbusch, T.A.J., 2013. Size resolved particle number emission factors of motorway traffic differentiated between heavy and light duty vehicles. Aerosol Air Qual. Res. 13, 450—461.

Omstedt, G., Bringfelt, B., Johansson, C., 2005. A model for vehicle-induced non-tailpipe emissions of particles along Swedish roads. Atmos. Environ. 39, 6088—609 .

Putaud, J.-P., van Dingenen, R., Alastuey, A., Bauer, H., Birmili, W., Cyrys, J., Flentje, H., Fuzzi, S., Gehrig, R., Hansson, H.C., Harrison, R.M., Herrmann, H., Hitzenberger, R., Hüglin, C., Jones, A.M., Kasper-Giebl, A., Kiss, G., Kousa, A., Kuhlbusch, T.A.J., Löschau, G., Maenhaut, W., Molnar, A., Moreno, T., Pekkanen, J., Perrino, C., Pitz, M., Puxbaum, H., Querol, X., Rodriguez, S., Salma, I., Schwarz, J., Smolik, J., Schneider, J., Spindler, G., ten Brink, H., Tursic, J., Viana, M., Wiedensohler, A., Raes, F., 2010. A European aerosol phenomenology - 3: physical and chemical characteristics of particulate matter from 60 rural, urban, and kerbside sites across Europe. Atmos. Environ. 44,1308—1320.

Rost, J., Holst, T., Saöhn, E., Klingner, M., Anke, K., Ahrens, D., Mayer, H., 2009. Variability of PM10 concentrations dependent on meteorological conditions. Int. J. Environ. Pollut. 36, 3—18.

Schlachta, R., Diemer, J., Schmid, M., Ott, H., 2013. Exceedances of the PM10 limit value due to the application of road salt on roads (in German). Immissionsschutz 3, 96—102.

Sjödin, Ä., Ferm, M., Björk, A., Rahmberg, M., Johansson, C., Gudmundsson, A., Swietlicki, E., Gustafsson, M., Blomquist, G., 2010. Wear Particles from Road Traffic. Paper presented at the Transport Research Arena (TRA) 7—10 June 2010 Brussels, Belgium.

Thorpe, A.J., Harrison, R.M., Boulter, P.G., McCrae, I.S., 2007. Estimation of particle resuspension source strength on a major London Road. Atmos. Environ. 41, 8007—8020.

Toth, I., Hercog, P., Peternel, R., Vojnikovic, B., 2011. Seasonal, daily and intradiurnal variation of PM10, NO2, NO and O3 in residential part of Zagreb. Croat. Coll. Antropol. 35 (2), 25—30.

U.S. Federal Register, 1987. Ambient Monitoring Reference and Equivalent Methods. United States Environmental Protection Agency. Federal Register 40 CFR Part 53, 1 July 1987.

Wang, F., Ketzel, M., Ellermann, T., Wählin, P., Jensen, S.S., Fang, D., Massling, A., 2010. Particle number, particle mass and NOX emission factors at a highway and an urban street in Copenhagen. Atmos. Chem. Phys. 10, 2745—2764.

WHO Technical report, 2013. Review of Evidence on Health Aspects of Air Pollution. REVIHAAP Project. www.euro.who.int/__data/assets/pdf_file/0004/193108/ REVlHAAP-Final-technical-report-final-version.pdf.