• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Aerosol Particle Number Concentration Measurements in Five European Cities Using TSI-3022 Condensation Particle Counter Over a Three- Year Period During Health Effects of Air Pollution on Susceptible Subpopulations

Posted on: Tuesday, 16 August 2005, 03:00 CDT

ABSTRACT

In this study, long-term aerosol particle total number concentration measurements in five metropolitan areas across Europe are presented. The measurements have been carried out in Augsburg, Barcelona, Helsinki, Rome, and Stockholm using the same instrument, a condensation particle counter (TSI model 3022). The results show that in all of the studied cities, the winter concentrations are higher than the summer concentrations. In Helsinki and in Stockholm, winter concentrations are higher by a factor of two and in Augsburg almost by a factor of three compared with summer months. The winter maximum of the monthly average concentrations in these cities is between 10,000 cm^sup -3^ and 20,000 cm^sup -3^, whereas the summer min is ~5000-6000 cm^sup -3^. In Rome and in Barcelona, the winters are more polluted compared with summers by as much as a factor of 4- 10. The winter maximum in both Rome and Barcelona is close to 100,000 cm^sup -3^, whereas the summer minimum is >10,000 cm^sup - 3^ During the weekdays the maximum of the hourly average concentrations in all of the cities is detected during the morning hours between 7 and 10 a.m. The evening maxima were present in Barcelona, Rome, and Augsburg, but these were not as pronounced as the morning ones. The daily maxima in Helsinki and Stockholm are close or even lower than the daily minima in the more polluted cities. The concentrations between these two groups of cities are different with a factor of about five during the whole day. The study pointed out the influence of the selection of the measurement site and the configuration of the sampling line on the observed concentrations.

INTRODUCTION

In recent years a number of studies have focused on the relationship between ambient particulate matter and adverse health effects. Elevated concentrations of aerosol particles have been associated with increases in all-cause mortality, mortality for respiratory and cardiovascular diseases, hospital admissions, and exacerbation of respiratory symptoms in chronically ill patients.1- 3 An extensive summary of >100 studies can be found in an article by Pope and Dockery.4 These studies can be subdivided in two main categories: acute and chronic exposure studies, focusing on either the short term variation or the long term changes in particulate air pollution levels.

Standards to limit the concentrations of particle mass with particle aerodynamic diameter <10 m are widely established in many countries. In addition, progress toward criteria for regulating the fine fraction of particulate matter (aerodynamic diameter <2.5 m) has been made both in the United States and Europe. However, both of these standards are based on particle mass and neglect to explicitly address ultrafine particles (<100 nm in aerodynamic diameter) because of the small contribution of such particles on total aerosol particle mass. Ultrafine particles can easily be inhaled and deposited in the deeper regions of the respiratory tract and can, therefore, pose a health risk. In fact, both toxicological and epidemiological studies have indicated that a high number concentration of ultrafine particles may cause serious health effects.5-8

Recent literature presents several studies focusing on the properties of ultrafine particles in urban areas. The findings indicate that the behavior of the particulate mass and of the total number concentration, which is dominated by the ultrafine fraction of aerosol particles, have to be investigated separately. Ruuskanen et al.9 found that ultrafine particles and the fine fraction of particulate matter differed both in sources and temporal variability in urban air. They presented a principal component analysis showing that particulate pollution can be divided in two categories, dominated either by aerosol particle mass or total number.

Several long-term datasets on aerosol particle number concentrations and size distributions have been published recently, reflecting the paucity of data on long-term particle number concentrations and difficulty in establishing meaningful standards for the concentration of total particle number. Long-term measurements of aerosol particle number size distributions over 4 years in Leipzig, Germany, have been presented by the Institute for Tropospheric Research group.10 A similar data set over 6 years in Helsinki, Finland, is presented by Hussein et al.11 and for a roadside site in London over >3years by Charron and Harrison.12 These studies have been performed with stationary monitoring equipment in a fixed site. Bukowiecki et al.13 had a different approach and published seasonal data measured using a mobile laboratory in the Zurich area. Whereas more information on long- term ultrafine aerosol properties, thus, has accumulated recently, all of the studies can be viewed as individual case studies in which the details of the instrumentation, site characteristics, and study periods are all different. There is clearly a need for systematic comparison between different cities in different climatic conditions.

In addition to temporal variability on different time scales, information on the spatial variation of particle number concentration is also essential in studies of human exposure of pollutants. Few studies have looked into this topic in detail. Buzorius et al.14 investigated the total aerosol particle number concentration at four different places in Helsinki. They found that during working days, when traffic intensity is relatively high, correlation between different concentration time series was >0.7. This indicates that one can use a central monitoring site to predict the temporal behavior at another site in the same city, but the total number concentration may be significantly different. Besides correlation, this study showed rather similar values in three measurement sites in town area, located within 3 km from each other. The concentration differed by a factor of two between the two extreme sites, whereas the other two had similar average concentrations. In addition, comparison measurements performed at the same site, but 60-m apart showed similar values within <10% difference.

In previous field studies, several common features can be identified. First of all, traffic is found to be the most important source for ultrafine particles in urban areas. Traffic produces primary particles with the highest numbers at ~70-100 nm in aerodynamic diameter. In addition, traffic is responsible for producing a high number of particles close to 10 nm in diameter. These particles are formed in the atmosphere close to the tail pipe of the vehicle, and their rate of formation has been shown to depend on both the type of the vehicle and the meteorological conditions.10,12

This study is part of a European Union-funded project named Health Effects of Air Pollution on Susceptible Subpopulations. The aims of the project were as follows: (1) to quantify the risk of hospitalization and of death because of air pollution, in particular airborne ultrafine particles, in individuals with coronary heart disease, and (2) to quantify the attributable risk of environmental exposures among a sensitive subgroup to facilitate appropriate public health strategies for the prevention of air pollution related health effects. For the part of the study that is presented in this article, the task was specifically to provide a database of cross- European data on ultrafine particles for the study areas in the Health Effects of Air Pollution on Susceptible Subpopulations study.

In this article, we present long-term aerosol particle total number concentration measurements in five metropolitan areas across the Europe. The measurements have been carried out using identical, well-characterized, and calibrated instrumentation. In addition, the measurement sites have been selected to facilitate intercomparison between the cities. Also, the influence of the specific monitoring environment and the sampling inlet construction were studied using a reference site and reference instrumentation during part of the study. The data can be useful in estimating the exposure of large urban populations to fine and ultrafine aerosol particles in different climatic conditions.

INSTRUMENTATION

In this study particle number concentration was measured with a TSI model 3022 condensation particle counter (CPC).15 The schematics of the instrument is presented in Figure 1. In this instrument aerosol is drawn into the instrument with a built-in pump. The user may select an inlet flow of either 0.3 L/min or 1.5 L/min. In our experiments we used the higher flow rate to achieve lower sampling losses. The flow rate is controlled with a pressure gauge connected to an orifice flow element. Aerosol is at first drawn into a saturator. The saturator walls are kept saturated with butanol with the aid of an automatically filled pool and a fibrous felt. A resistance heater keeps the saturator temperature at 35 C. When the aerosol exits from the saturator, it is saturated with butanol vapor. On exiting the saturator, aerosol \is drawn into the condenser where the wall temperature is kept at 10 C with a thermoelectric device. When aerosol starts to cool, butanol begins to condense on the nearest surface available, including aerosol particle surfaces. Depending on the particle concentration, particles can grow up to a diameter of 10 m before they exit the condenser and are optically detected. The condenser exit nozzle accelerates the droplets through a laser beam. Part of the light scattered by the droplet is detected with a photodetector.

Figure 1. Schematic illustration of the condensation particle counter TSI 3022.

The CPC 3022 has three modes of operation. The real-time single- particle counting mode counts the electrical pulses generated by photodetector individually over some time period and calculates the concentration using the given aerosol flow rate. It is used for particle concentrations <1000 cm^sup -3^. In single-particle live- time counting mode the counter clock is advanced just when the electronics is ready to detect a particle. The electronics is stopped for a few microseconds each time a particle is detected. In other words, only the time when the instrument is ready to detect a new particle is taken into account. This so called live time is then used to calculate the final concentration. The live time counting mode is used for particle concentration between 1000 and 10,000 cm^sup -3^. After 10,000 cm^sup -3^ photometric calibration is used to calculate the concentration. In this mode the intensity of light scattered from multiple particles in the scattering volume is detected, and the concentration is calculated using a calibration table saved to the instrument. The CPC is not calibrated for concentrations >10^sup 7^ cm^sup -3^. The calibration is saved to the CPC's in steps of 0.1 V around 10,000 cm^sup -3^ corresponding to a concentration step of 5,000 cm^sup -3^. About 15 calibration points are used over the concentration span typical for city areas. Fluctuations of 0.005 V are common in the photodetector voltage, which corresponds to concentration fluctuations of ~250 cm^sup -3^. It is quite characteristic of the instrumentation that when counting mode shifts from live time to photometric calibration, mode a concentration step either down or up is observed. The reasons for this step are mainly the errors in concentrations from the live time counting mode as the concentration gets higher and also the changes of the calibration in the photometric mode with time. The magnitude of this step is usually rather small, <10%, and is within the overall accuracy of the instrument.

According to the standard operation procedure of the Health Effects of Air Pollution on Susceptible Subpopulations study, the instrument used should be either a new one or factory calibrated before the campaign. During the 3-year campaign, each instrument should be compared against a "standard" instrument at least once. In practice each instrument was either changed to a fresh one, factory calibrated, or calibrated by the University of Helsinki more than once during the campaign. Each instrument was compared against a reference at least once during the campaign. Each week the instrument was checked. Data was copied from the logging computer, and the computer clock was set to the correct reference time (local wintertime), status lights were checked, and butanol was drained and filled. A butanol bottle was kept connected to the instrument all of the time. The pump was switched off, and photodetector voltage and aerosol particle concentration were checked when the flow was off. Flow rates were measured with a low-pressure drop flow calibrator. If the flow was off by >10%, flow was adjusted. The CPC was kept in a "high flow mode," 1.5 L/min. The data was transferred to a common database in Helsinki and saved as 1-min averages.

Instrument calibrations were made by the University of Helsinki for some instruments. Charged monodisperse silver aerosol particles were used for the calibration. Concentration calibration was made against an aerosol electrometer and a TSI model 3025 instrument. The model 3025 was used also as a standard when model 3022 cut-off diameters were determined. The typical 50% cut-off diameter for model 3022 was around 7 nm.

According to the standard operating procedure, sampling lines should be made of stainless steel tube with an i.d. >4 mm. The length of the sampling line should be <5 meters, and it should be as short as possible. Later it was noticed that, especially in Rome and Barcelona, adding a drier to the sampling line was essential. Water started to accumulate inside the instrument despite the scheduled butanol change. At first Permapure driers were tested, but silicagel driers were found to be more reliable, although more laborious. The change of the sampling lines in the middle of the campaign caused some artifacts to the data because of the larger diffusion losses after adding the drier. The losses in the Permapure drier are similar to that of 2 m of regular tubing, and losses in the silicagel drier correspond to 5 meters of regular tubing. The losses were studied by operating another CPC parallel to the one with the dryer. The use of dryers leads to 10% losses of 15-nm particles for the Permapure drier and 20% losses for the silicagel drier.

The field comparisons against standard instrument recently calibrated by the University of Helsinki were performed at least once during the campaign in each city. The comparisons were made with ambient air by running the instruments parallel with as similar sampling lines as possible. In Helsinki and Barcelona the comparison appeared to be fairly satisfactory. Hourly average concentrations differed in most cases by <10%, which is the accuracy stated by manufacturer. At the other sites the comparisons were not so successful. In Augsburg the difference was sometimes >60%, in Rome, >50%, and in Stockholm, 30%. During the second comparison period in Stockholm during autumn 2002 the difference was <10%. In Rome a more detailed comparison study was made with three CPC's during the spring 2003. The discrepancies were found to be ~20% with identical sampling lines. This discrepancy might partly be explained by the fact that two of the instruments were calibrated in Helsinki and one by the manufacturer. Pearson correlation coefficients were, however, >0.98. When instruments were compared with different sampling lines and 7 m apart from each other, the discrepancy increased. The hourly correlations decreased to 0.75 and concentration differences between the instrument rose to 60%. The length of the sampling line has an effect on the total concentration because in polluted environments the mean particle size can be quite small, close to 10 nm in diameter. The small particles are strongly affected by diffusion and are deposited on the walls of the sampling lines. The instruments were compared again some months later, and the results were very similar. Instrument calibration was not changed during this time.

Particle number concentration measurements with CPC's have many sources of possible errors. The instrument response is linearly proportional to the sample flow rate. The flow rate might change or become unstable when the flow-measuring element becomes contaminated. In most cases the reason for unstable flow is the penetration of butanol or water into the sampling lines. It is essential to keep the water out of the instrument. In some centers the filling system of the butanol failed a few times during the measurement period causing the instrument flooding. The photometric calibration of the instrument can also change in time. This might happen because of contamination of the optics, laser power drop, or saturator contamination. Optics contamination and laser power drop directly affect the instrument response. Optics can get contaminated because of high large particle concentrations. This happened once during the period because of asphalt works close to the measurement site in one center (Stockholm). Water or butanol leakage to the optics might also contaminate it. Laser power will decrease over time, and the instrument should be serviced and the laser current checked by the manufacturer every third year. Saturator contamination affects the maximum supersaturation inside the saturator, which leads to increase in cut-off diameter of the device and smaller final droplet sizes. If the droplet size after the condenser falls below 0.5 m, droplets are not detected with the optical counter. Typical contaminants are, again, water and large dust particles. Data acquisition problems were also fairly common. Computers can fail, and power breaks may cause trouble. These problems were observed in all of the cities, but they did not produce major problems in the dataset.

Table 1. Summary of the site descriptions.

SITE DESCRIPTIONS

The study period was from May 2001 to December 2003. In some centers, the experiments were started already earlier, and all of the centers managed to start measurements by May 2001. It was decided before the campaign that the measurement sites should be urban background sites. This means locations at some distance from direct sources and broadly representative of citywide background concentrations. The sites could be situated in elevated locations, parks, or urban residential areas. The sites are summarized in Table 1. Three of the primary sites, in Helsinki, Stockholm, and Barcelona, are elevated sites close to the city center. The Augsburg site is situated inside a garden and the Rome site by a moderately busy street close to the downtown area. The secondary site in Rome is situated inside a park. In Stockholm, the secondary site is located in a street canyon.

Helsinki

Helsinki has a population of ~0.5 million people. It is bounded in the south by the Gulf of Finland. The measurement site is located ~3.5 km north of the city center, 10\1 m west of one of the main roads carrying traffic into and out from the city. One of the city harbors and several power plants are located 2 km south of the site. The instrument is located on the fourth floor level ~13 m above the nearest small street and ~100 m from the closest major street. The sample was taken through a wall, and the inlet extends 1 m outside of the wall. The sample line is a 2.5-m long 4-mm i.d. stainless steel tube. Sampling started May 1, 2001, with a plain stainless steel line, but on August 8, 2001, the Permapure drier (MD-070-24 sec-4) was added to the line because of water accumulation inside the CPC. Later, on April 22, 2003, instrument maintenance was taken over by the RUPIOH project (relationship between ultrafine and fine particulate matter in indoor and outdoor air and respiratory health, European Union-funded research project), and the drier was changed to a 30-cm-long silicagel drier, and the tube length was increased to 4 m.

Stockholm

Stockholm has a population of ~1 million people. The city is limited by the Baltic Sea in the east. The instrument is located on the roof of the Stockholm Air Quality and Noise Analysis measurement site on Rosenlundsgatan a few kilometers south from the city center. This area constitutes the southern edge of the Stockholm downtown area. Main roads surround the site a few hundred meters away. The inlet is 20 m above the street and 50 m from the closest bigger road. The inlet line is a 2-m long stainless steel tube with an i.d. of 4 mm. On July 4, 2002, the Permapure drier was added to the line, and it was removed on November 26, 2002. No major problems with sampling occurred. Only one month of data was lost during the summer 2003 because of instrument problems. In Stockholm particles are measured also at a secondary street canyon site on Hornsgatan. The inlet is 4 m from the building and 3 m from the middle of the nearest lane. The height above ground level is 3.5 m. It extends 1.5 m outside the sidewalk over the vehicle lanes. The size of the inlet is 4 mm.

Augsburg

Augsburg is a city in Germany with a population of ~0.45 million people, located some 60 km west of Munich. The Augsburg measuring site is in the orchard of the monastery St. Stephan. This orchard is enclosed by a wall that is ~3 m high. The measurement device is in the orchard, at a distance of 15 m from the nearest wall. The monastery is located <1 km north of downtown Augsburg, on the high banks of the river Lech. The nearest street is at a distance of 50 m, being a minor street with low traffic intensity, and the nearest major street is -150 m away. The nearest stationary particle source is a thermal power station (powered by gas) at a distance of 300 m. The sample inlet is a 120-cm long SS tubing 3 mm in diameter located 2 m above the ground. No dryer was used.

Measurements in Augsburg suffered major problems because of instrument problems between July 10, 2003, and December 1, 2003. Data from another nearby measurement site was used to fill the gap in primary site data between September 8, 2003, and December 1, 2003.

Barcelona

The city of Barcelona has a population of ~1.5 million people. It is located on the north coast of the Mediterranean. The instrument is located in downtown Barcelona by the sea, a few hundred meters west of the Port Olimpic de Barcelona. The city center is north of the site. The main road (B10) is located just 30-40 m north of the site. The city harbor is located some kilometers west of the site, and the main power plant about the same distance to the east. The site is influenced by sea breeze. The instrument location is very similar to that of the Helsinki site. It is in the fourth floor of the Institut Municipal d'Investigaci Mdica. Initially the sampling line was 2.1 m long and 4 mm in diameter. A similar Permapure drier to that in Helsinki was installed to the sampling line on October 24, 2001, but it did not function as expected. Because we were not able to find out the reason, a new 47-cm long Topas silcagel drier replaced the Permapure drier on March 21, 2003.

Measurements in Barcelona suffered from CPC problems and computer problems several times. Between July 4 and September 29, 2001, the instrument was being serviced because of water accumulation inside the instrument. This caused some breaks in the data. Between May 6 and July 22, 2002, the CPC experienced some fluctuations in the flow rate, which caused a 2-week break in July. The instrument was changed, but problems with the instrument continued until September 7, 2002, when the serviced instrument was reinstalled. The next breakdown occurred on December 24, 2002, because of flow rate problems. The CPC was replaced January on 15, 2003, and subsequently the new instrument behaved fairly well, indicating that the drier worked well and prevented water entering the instrument. However, the whole November 2003 was lost because of instrument problems.

A second instrument was installed ~500 m from Institut Municipal d'Investigaci Mdica, on the fourth floor of a building of the Universitat Pompeu Fabra, and was run between February 28, 2002, and May 8, 2002. The distance from the nearest busy street is 300 m, and the sampling height is 20 m. The length of the sampling line was - 1 m, and there was no dryer in the sampling line.

Rome

Rome has a population of ~2.6 million people. It is the capital of Italy located by the river Tevere (Tiber) ~20 km from the Mediterranean Sea. It covers a densely populated area ~20 km in diameter surrounded by ringway E80. The primary site instrument is located on the front yard of the Italian National Institute of Health, inside an instrument container. The inlet is located few meters from the street Viale Regina Elena at a height of 3 m from the ground. A traffic intensity of 25,000 cars/days was estimated, constant throughout the year except for August. The city center is located few kilometers west of the site but still within the city area. Traffic pollution was presumed to be roughly intermediate relative to the city, considering the traffic intensity and location of the site between the center and suburban areas. The area is not subject to industrial emission. This site could be considered as a traffic-influenced site.

The secondary site is located in a protective container inside a botanical park located in the area surrounding the city center. To the east of the sampling site there is a little hill (Gianicolo Hill) characterized by many different plants and trees. To the south there is the famous district of Trastevere (~400 m from the sampling point), which is a limited traffic area (only for residents). To the west and north there is the river of Rome (Tevere) and ~2 km to the north, the San Peter Square. The nearest busy traffic street is located at -400 m from the sampling inlet. We can assume the site as an urban background site.

Both sites have identical stainless steel inlet tubes, with 4-mm Ld., and the total length of the tube (from the inlet to the instrument) is 2.25 m. The tube extrudes from the cover of the containers for 1.2 m. The distance between the inlet and the ground is 2.5 m at the primary site, and 3 m at the secondary site. The Permapure dryers (model MD-070-24 sec-4) that were later installed did not substantially modify measurements from these instruments, because it was inserted after cutting off an appropriate portion of the original tube.

The driers were installed to primary site June 6, 2002, and to secondary site at the end of March 2002. In spite of the drier, water continued to cause some trouble during the summer of 2002 at both sites and during the summer 2003 at the primary site. The primary site CPC had to be serviced during the summer 2003 and the secondary site CPC during the summer 2002.

RESULTS

Annual Variation

The sites can be divided to two subgroups. Helsinki, Stockholm, and Augsburg belong to a group that could be characterized as relatively clean urban areas. In all of these three cities, yearly average concentrations are -10,000 cm^sup -3^. The second group consists of the cities of Rome and Barcelona and may be regarded as polluted urban areas. Yearly averages in these two cities are about five times higher than in the "clean" cities. Typical values are ~50,000 cm^sup -3^ and monthly means reaching values are as high as 100,000 cm'3 (in Barcelona). The summary of the concentration statistics at all of the cities is given at Table 2. The variation of the aerosol particle number concentration in both clean and polluted cities (Helsinki and Barcelona) are shown in Figures 2 and 3. In both cities, the concentration varies both annually and diurnally. Qualitatively, the cities look relatively similar. However, the absolute values are markedly different. In addition, a more detailed look at the temporal variation also shows clear differences. In Helsinki, the high concentrations appear during winter months (October to March) and between daytime (6:00 a.m. to 6:00 p.m.). In Barcelona, the annual behavior shows similarly higher concentrations during winter months. However, the diurnal pattern seems to be weaker with high concentrations also during the nighttime. The temporal variation is discussed in more detail below.

Table 2. Summary of the daily mean concentration statistics at the five sites.

In all five of the cities, the winter concentrations are higher than the summer concentrations. The annual variations of the concentration in all cities are plotted in Figures 4-8. The monthly mean values, as well as 25 and 75 percentiles and minimum and maximum values of the daily average concentrations are shown. In Helsinki and Stockholm, winter concentrations are higher by a factor of two and in Augsburg almost by a factor of three compared with summer months. The cleanest month in Stockholm and Helsinki is typically July and in August in Augsburg. The dirtiest month is typically \January, but there is some variation between years, and high concentrations are observed typically from November to March. The winter maximum (monthly mean) in these clean cities is between 10,000 cm^sup -3^ and 20,000 cm^sup -3^, whereas the summer min is ~5000-6000 cm^sup -3^. The effect of the addition of a drier into the sampling line in Helsinki can be seen as decreasing concentration during spring 2003. The decrease is notable, and the issue of dryer will be discussed in more detail later in the article.

In Rome and Barcelona, the winters are more polluted compared with summers by as much as a factor of 4-10. The cleanest month in these cities is usually August. The dirtiest month is also typically January with high concentrations during all the months between November and March. The winter maximum in both Rome and Barcelona is close to 100,000 cm^sup -3^, whereas the summer minimum is >10,000 cm^sup -3^. The summer values in Rome and Barcelona are close to the winter values in clean cities.

Figure 2. Dependence of the particle concentration on month and hour in Helsinki.

Diurnal Behavior

The diurnal behavior was investigated by calculating the concentration pattern for weekdays as hourly averages. In all cities the five weekdays are similarly dirty. This can be seen in Figure 9, which presents the average concentrations during 8:00 a.m. each day of the week. Sunday is the cleanest day. In Helsinki, Sunday concentrations are 60% of the weekday concentrations. In Augsburg, Stockholm, and Rome they are ~75% and in Barcelona, 65% of the weekday concentrations. The Saturday average concentrations lie between the weekday and Sunday concentrations. Summary of the average concentrations during different days of the week in different cities is presented in Table 3.

Figure 3. Dependence of the particle concentration on month and hour in Barcelona.

Figure 4. Monthly statistics of daily means in Helsinki.

Figure 6. Monthly statistics of daily means in Augsburg.

During the weekdays the maximum concentration in all of the cities is detected during the morning hours between 7 a.m. and 10 a.m. in the wintertime. The average diurnal behavior is shown in Figure 10. The shift to daylight savings time can be clearly detected in March and October (data not shown). In Helsinki the concentration reaches a maximum between 7:00 and 8:00 a.m., and then gradually decreases until 2:00 a.m. the following night. No afternoon maximum is detected. In Barcelona the concentration maximum is also detected between 7:00 and 8:00 a.m., and a clear evening maximum is detected around 9:00 p.m. The late maximum could be explainable by the decrease of the sea breeze. In Rome the maximum concentration is detected between 8:00 and 9:00 a.m. and the evening maximum around 7:00 or 8:00 p.m. In Augsburg the morning maximum is detected between 8:00 and 9:00 a.m. and the evening maximum between 9:00 and 10:00 p.m. In Stockholm the morning maximum is achieved quite late, around 9:00 a.m., and similarly to Helsinki, no clear evening maximum is detected. Instead, a weak afternoon rush hour maximum can be seen. It can be seen in Figure 10 that the daily maxima in the clean cities are close or even lower than the daily minima in the more polluted cities. The concentrations between these two groups of cities are different with a factor of about five during the whole day.

Figure 5. Monthly statistics of daily means in Stockholm.

Figure 7. Monthly statistics of daily means in Barcelona.

The diurnal pattern of the particle number concentration is caused mainly by the variation in traffic during the day and by the strength of air mixing. The morning concentration maximum is caused by increased particle emissions from the traffic. It is noted that the mixing of the boundary layer is also increased during the early morning hours because of increasing sun radiation. Later in the day the mixing height of the boundary layer increases, and the evening peak is not as strong as the morning maximum. Also, the relative amounts of petrol and diesel vehicles may have diurnal patter, which may result to the absence of the afternoon peak.16,17 In coastal sites of Barcelona, Helsinki, and Stockholm, sea breeze also causes concentration decrease during low wind speed days.

Figure 8. Monthly statistics of daily means in Rome.

The average diurnal behavior during Saturdays (Figure 11) and Sundays (Figure 12) is very different to that of weekdays, indicating the role of traffic as the strongest source influencing the aerosol particle number concentration. However, the differences between the cities are at the similar level as during the weekdays.

Spatial Variation of the Total Concentration

To study the influence of the actual location of the measurement site, parallel measurements were conducted in three of the cities: Barcelona (60 days), Rome (293 days), and Stockholm (534 days). It is shown earlier by Buzorius et al.14 that correlation between sites may be expected. However, the absolute values of the sites are probably different because of different distances from local emissions (streets) and micrometeorological parameters. The correlations between the simultaneous measurements in two sites are presented in Figures 13-15. The average concentrations at the secondary sites were in Barcelona, 52,840 cm^sup -3^, in Rome, 49,967 cm^sup -3^, and in Stockholm, 11,136 cm"3. The concentration ratios between the parallel sites were rather similar in Barcelona and Rome (2.21 and 2.53, correspondingly), whereas in Stockholm the ratio was much larger-0.16. Note that in Stockholm the secondary site was closer to the traffic emissions compared with the primary site, whereas in Barcelona and in Rome the situation was the opposite.

Figure 9. Average concentrations at 8 a.m. each day of the week for all studied cities.

Figure 10. Hourly dependence of particle concentration during weekdays.

Table 3. Weekday dependence of particle concentration (cm^sup - 3^; daily average).

DISCUSSION

This study shows that clear annual and diurnal behavior of the number concentration of aerosol particles is found in the studied European cities. It is probable that the same is valid for other cities, which have similar emission patterns and climatic conditions. The max values are always found during the wintertime, whereas min is in the middle of the summer. The diurnal concentration pattern is also similar in all of the studied cities indicating the major role of the traffic as the source of fine and ultrafine particles. Whereas similarities between the cities are obvious, there are also differences that are significant and may have several explanations.

Figure 11. Hourly dependence of particle concentration Saturdays.

Figure 13. Pearson correlation coefficient between the Barcelona sites.

One of the main differences between the five cities is the absolute values of the concentrations. The two southern cities had clearly higher concentrations during the whole study period. The reasons for that might include meteorological differences, as well as differences in the actual emissions. The latter include factors like size and population of the city, traffic intensity, the relative amounts of different types of vehicles, and so forth. All of these have probably significant differences but are not studied within this study. However, more investigations are needed before target values for aerosol particle number concentrations can be meaningfully set.

Figure 12. Hourly dependence of particle concentration during Sundays.

Figure 14. Pearson correlation coefficient between the Rome sites.

The measurements were carried out using similar instrumentation and at sites that were designed to be comparable. The main idea was to obtain data that is comparable between the sites. However, several major issues that need to be considered were found during the measurements. First of all, the actual location of the site appears to be important. Within this study, three of the cities had additional measurements for some time at a nearby site. The comparison between the two sites of the same city indicate that the mean concentrations can be several times higher at one of the two sites. This finding is based on only three measurements and, therefore, the magnitude of this problem needs more attention. However, for example, the classification of Stockholm as a clean city would be different if the other site with higher concentration was selected as the study site. Similarly, Rome and Barcelona would appear much cleaner if the secondary sites were selected for the study.

Figure 15. Pearson correlation coefficient between the Stockholm sites.

An important factor to be considered is also the details of the inlet. In principle, the inlet can be specified in detail, and all the measurements could be done using the same setup. However, in practice the characteristics of the site (e.g., room size, orientation, and the place for the instruments in the room) determine the min length of the sampling line and also the orientation and need for bends, and so forth. If identical lines are preferred, the most complicated needs to be used. This will cause additional problems in those sites, where much simpler setup is possible. The sampling line details affect the losses in aerosol sampling (e.g., the diffusional losses in laminar flow depend directly on the length of the tube). Therefore, it needs consideration whether one makes shortest possible lines to minimize the losses or longer lines to have comparable losses with other measurements. We note also that whereas diffusion losses can be taken into account in size distribution measurements, the losses cannot be corrected for total concentration only. Finally, electrically conductive material needs to be used always when sampling aerosol particles, and stainless steel or copper are typically preferred.

The dryer connected to the sampling line needs additional c\onsideration. The need for a dryer is obvious for climatic conditions where water condensation is possible inside the CPC. Mainly this is the case in warm and humid weather conditions, because the counter is placed typically in an air-conditioned room. In colder climates (in this study, Helsinki and Stockholm), the dryer is needed only for a short time period during middle of the summer. The dryers that were used in this study were seen to have significant losses for the total particle number. This is clearly seen in the Helsinki data during spring 2003 when the dryer was installed. The losses inside the dryer cannot be corrected for the total concentration.

CONCLUSIONS

We have presented long-term aerosol particle total number concentration measurements in five metropolitan areas across the Europe. The measurements have been carried out using a CPC (TSI model 3022) that is able to count the total aerosol particle number concentration between ~7 nm and a few micrometers.

The results show that in all of the studied cities the winter concentrations are higher than the summer concentrations. In Helsinki and in Stockholm winter concentrations are higher by a factor of two and in Augsburg almost by a factor of three compared with summer months. In Rome and in Barcelona the winters are more polluted compared with summers by as much as a factor of four or five.

During the weekdays the maximum concentration in all of the cities is detected during the morning hours between 7:00 and 10:00 a.m. The evening maxima were present in Barcelona, Rome, and Augsburg, but these were not as pronounced as the morning ones. The daily maxima in Helsinki and Stockholm are close or even lower than the daily minima in the more polluted cities. The concentrations between these two groups of cities are different with a factor of about five during the whole day.

This study showed that the details of the sampling line (length, geometry, etc.) and the use of a dryer in the sampling line influence significantly to the observed concentration. In addition, the location of the sampling site within the city was shown to be significant in determining the measured concentration level. These issues need consideration when planning similar experiments and also future investigations to obtain comparable data from different urban areas.

ACKNOWLEDGMENTS

This work is supported by European Union contract no. QLRT-2000- 00708. The financial support from "Institute de Salud Carlos III" Red de Centras RCESP, C03/09, Red RESPIRA, C03/011 and Red de Grupos INMA G03/176 is acknowledged.

IMPLICATIONS

This manuscript gives an overview of aerosol particle number concentrations in five European cities over about a two-year period. These data are valuable for: (1) providing basic information about fine and ultrafine particulate pollution concentrations in European cities; (2) showing the diurnal and annual variation of the aerosol concentrations; (3) providing data for future legislation concerning the particulate pollution; and (4) giving a database that can be used in epidemiological studies related to aerosol particles and health effects.

REFERENCES

1. Air Quality Criteria for Paniculate Matter; EPA/600//P0 -95/ 001cF; United States Environmental Protection Agency Office of Research and Development: Washington, DC, 1996.

2. Dockery, D.W.; Pope, C.A. Acute Respiratory Effects of Paniculate Air Pollution; Ann. Rev. Public Health 1994, J5, 107- 132.

3. Pope C.A., 111; Burnett, R.T.; Thun, MJ.; Galle, E.E.; Krewski, D.; ItO, K.; Thurston, G.D. Lung Cancer, Cardiopulmonary Mortality, and LongTerm Exposure to Fine Paniculate Air Pollution; JAMA 2002, 287, 1132-1141.

4. Pope C.A., Ill; Dockery D.W. Epidemiology of Particle Effects. In Hoigate, S.T., Koren, H., Maynard, R., Samet, J., eds. Air Pollution and Health. Academic Press: London, 1999, 673-705.

5. Oberdrster, G.; Gelein, R.; Ferin, J.; Weiss, B. Association of Paniculate Air Pollution and Acute Mortality: Involvement of Ultrafine Particles? Inhalation Toxicol. 1995, 7, 111-124.

6. Oberdrster, G. Toxicology of Ultrafme Particles: In Vivo Studies; Philosophical Transactions of the Royal Society of London A 2000, 358, 2719-2740.

7. Pekkanen, J.; Timonen, K.L.; Ruuskanen, J.; Reponen, A.; Mirme, A. Effects of Ultra-Fine and Fine Particles in Urban Air on Peak Expiratory Flow Among Children With Asthmatic Symptoms; Environ. Res. 1997, 74, 24-33.

8. Wichmann, H.E.; Peters, A. Epidemiological Evidence of the Effects of Ultrafine Particle Exposure; Philosophical Transactions of the Royal Society of London A 2000, 358, 2751-2769.

9. Ruuskanen, J.; Tuch, Th.; Ten Brink, H.; Peters, A.; Khlystov, A.; Mirme, A.; Kos, G.P.A., Brunekreef, B.; Wichmann, H.E.; Buzorius, G.; Vallius, M.; Kreyling, W.G.; Pekkanen, J. Concentrations of Ultrafine, Fine and PM25 Particles in Three European Cities; Atmos. Environ. 2001, 35, 3729-3738.

10. Wehner, B.; Wiedensohler, A. Long Term Measurements of Submicrometer Urban Aerosols: Statistical Analysis for Correlations with Meteorological Conditions and Trace Gases; Atmos. Chem. Phys. 2003, 3, 867-879.

11. Hussein, T.; Puustinen, A.; Aalto, P.P.; Mkel, J.M.; Hmeri K. and Kulmala, M. Urban Aerosol Number Size Distributions; Atmos. Chem. Phys. Discuss 2003, 3, 5139-5184.

12. Charron, A.; Harrison, R.M. Primary Particle Formation from vehicle Emissions during Exhaust Dilution in the Roadside Atmosphere; Atmos. Environ. 2003, 37, 4109-4119.

13. Bukowiecki, N.; Dommen, J.; Prevot, A.S.H., Weingartner E.; BaItensperger, U. Fine and Ultrafine Particles in the Zurich (Switzerland) Area Measured With a Mobile Laboratory: an Assessment of the Seasonal and Regional Variation Throughout a Year, Atmos; Chem. Phys. 2003, 3, 1477-1494.

14. Buzorius, G.; Hmeri, K.; Pekkanen, J.; Kulmala, M. Spatial Variation of Aerosol Number Concentration in Helsinki City; Atmos. Environ. 1999, 33, 553-565.

15. Agarwald J.K.; Sem, GJ. Continuous Flow, Single-Particle- Counting Condensation Nucleus Counter; J. Aerosol Sd. 1980, 11, 343- 357.

16. Wehner, B.; Birmili, W.; Gnauk T.; Wiedensohler, A. Particle Number Size Distributions in a Street Canyon and Their Transformation into the Urban-Air Background: Measurements and a Simple Model Study; Atmos. Environ. 2002, 36, 2215-2223.

17. Whlin, P.; Palmgren F.; Van Dingenen, R. Experimental Studies of Ultrafine Particles in Streets and Their Relationship to Traffic; Atmos. Environ. 35 Supplement No. 2001, 1, S63-S69.

Pasi Aalto, Kaarle Hmeri, Pentti Paatero, and Markku Kulmala

Department of Physical Sciences, University of Helsinki, Helsinki, Finland

Tom Bellander and Niklas Berglind

Department of Occupational and Environmental Health, Stockholm, Sweden

Laura Bouso, Gemma Castano-Vinyals, and Jordi Sunyer

Institut Municipal d'lnvestigaci Mdica, Barcelona, Spain

Giorgio Cattani and Achille Marconi

lnstituto Superiore di Sanit, Rome, Italy

Josef Cyrys, Stephanie von Klot, Annette Peters, and Katrin Zetzsche

Forschungszentrum Institut f. Epidemiologie, Neuherberg, Germany

Timo Lanki and Juha Pekkanen

National Public Health Institute, Department of Environmental Health, Unit of Environmental Epidemiology, Kuopio, Finland

Fredrik Nyberg

Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden

Billy Sjvall

Stockholm Air Quality and Noise Analysis, Stockholm, Sweden

Francesco Forastiere

Department of Epidemiology, Local Health Authority Rome E, Rome, Italy

About the Authors

Pasi Aalto, laboratory manager, Kaarle Harneri, professor, Pentti Paatero, laboratory manager, and Markku Kulmala, professor, are from the Department of Physical Sciences, University of Helsinki, Helsinki, Finland. Tom Bellander, associate professor, and Niklas Berglind, statistician, are from the Department of Occupational and Environmental Health, Stockholm, Sweden. Laura Bouso, research assistant, Gemma Castano-Vinyals, PhD student, and Jordi Sunyer, senior researcher and professor, are from the Institut Municipal d'lnvestigaci Mdica, Barcelona, Spain. Giorgio Cattani, research assistant, and Achille Marconi, research director, are from the Institute Superiore di Sanit, Rome, Italy. Josef Cyrys, research associate, Stephanie von Klot, research associate, Annette Peters, senior researcher, and Katrin Zetzsche, data manager, are from the Forschungszentrum Institut f. Epidemiologie, Neuherberg, Germany. Timo Lanki, researcher, and Juha Pekkanen, professor, are from the National Public Health Institute, Department of Environmental Health, Unit of Environmental Epidemiology, Kuopio, Finland. Fredrik Nyberg, lecturer, is from the Institute of Environmental Medicine, Karolinska Institute, Stockholm, Sweden. Billy Sjvall, technician, is from the Stockholm Air Quality and Noise Analysis, Stockholm, Sweden. Francesco Forastiere, head of Environmental Epidemiology Unit, is from the Department of Epidemiology, Local Health Authority Rome E, Rome, Italy. Address correspondence to: Kaarle Hmeri, Division of Atmospheric Sciences, Department of Physical Sciences, P.O. Box 64 (Gustaf Hllstrmin Katu 2), FIN-00014, University of Helsinki, Finland; phone: +358-40-5684487; fax: +358-9-19150860; e- mail: kaarle.hameri@helsinki.fi.

Copyright Air and Waste Management Association Aug 2005

Source: Journal of the Air & Waste Management Association

More News in this Category