Particle number size distribution in the eastern Mediterranean: Formation and growth rates of ultrafine airborne atmospheric particles

Atmospheric Environment 77 (2013) 790e802

Contents lists available at SciVerse ScienceDirect

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

Particle number size distribution in the eastern Mediterranean: Formation and growth rates of ultrafine airborne atmospheric particles I. Kopanakis a, S.E. Chatoutsidou a, K. Torseth b, T. Glytsos a, *, M. Lazaridis a, * a b

Department of Environmental Engineering, Technical University of Crete, Polytechneioupolis, Chania 73100, Greece Norwegian Institute for Air Research, Instittutveien 18, N-2027, Norway

h i g h l i g h t s
Particle number concentration was measured at Akrotiri research station (Greece).
The data were analysed according to the origin of the air masses reaching the site.
The characteristics of size distributions were analysed.
New particle formation events were investigated.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 August 2012 Received in revised form 23 May 2013 Accepted 27 May 2013

Particle number concentration was measured between June 2009 and June 2010 at Akrotiri research station in a rural/suburban region of western Crete (Greece). Overall, the available data covered 157 days during the aforementioned period of measurements. The objectives were to study the number size distribution characteristics of ambient aerosols and furthermore to identify new particle formation events and to evaluate particle formation rates and growth rates of the newborn particles. Aerosol particles with mobility diameters between 10 and 1100 nm were measured using a Scanning Mobility Particle Sizer (SMPS) system. Measurements were performed at ambient relative humidities. The median total particle number concentration was 525 #/cm3 whereas the number concentration ranged between 130 #/cm3 and 9597 #/cm3. The average percentage of particles with diameters between 10 nm and 100 nm (N10e100) to total particles was 53% during summer and spring, but reached 80% during winter. Maximum average contribution of nano-particles (10 nm < Dp < 50 nm) to total particles was recorded also in winter and was attributed partly to the effect of local heating. Furthermore, back trajectories (HYSPLIT model) showed that different air mass origins are linked to different levels of particle number concentrations, with higher values associated with air masses passing from polluted areas before reaching the Akrotiri station. Modal analysis of the measured size distribution data revealed a strong nucleation mode during winter (15e25 nm), which can be correlated with emissions from local sources (domestic heating). The nucleation mode was observed also during the spring campaigns and was partly linked to new particle formation events. On the contrary, an accumulation mode (80e120 nm) prevailed in the measurements during summer campaigns, when the station area was influenced by polluted air masses arriving mainly from Eastern Europe. In total, 13 new particle formation events were recorded during the 157 days of measurements. Nucleation events were associated with low values of N100 particle number concentration and reduced coagulation sinks. Mean growth and formation rates were calculated and showed values equal to 6 nm hr1 and 13 cm3 s1, respectively. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Particle number concentration Size distribution Nucleation events Eastern Mediterranean

1. Introduction

* Corresponding authors. E-mail address: [email protected] (T. Glytsos). 1352-2310/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atmosenv.2013.05.066

Atmospheric aerosols influence substantially life on Earth. Climate change, human health, visibility and air quality are affected by atmospheric particles. More precisely, particles with aerodynamic

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

diameter below 1 mm are affecting the radiative budget in the troposphere (Seinfeld and Pandis, 2006) and penetrate deep into the human respiratory tract and affect pulmonary health (Housiadas and Lazaridis, 2010; Lazaridis et al., 2001). Moreover, particles with diameter lower than 20 nm (nucleation mode) can be generated in the atmosphere due to nucleation processes. New particle formation can be an important source of cloud condensation nuclei (CCN) and thus influence climate and water cycling (Seinfeld and Pandis, 2006; Sun and Ariya, 2006). The formation of new particles is followed by condensational growth up to the 50e200 nm size range and occurs in almost any part of the troposphere (Holmes, 2007; Kulmala et al., 2007b; Kulmala and Kerminen, 2008). Recent studies showed that nucleation events can occur during the whole year at the Mediterranean area. Such events were recorded both at urban and remote  sites (Dall’Osto et al., 2012; Pikridas et al., 2012; Zdímal et al., 2010; Reche et al., 2011). The atmospheric aerosol distribution depends on the type of a region (urban, rural continental, remote continental, maritime, polar, desert) and on processes occurring at the area and result to the emission, removal or size change of local aerosol particles. Each mode has homogeneous origin and derives from a specific process (e.g. nucleation, condensation, emission). Typically, the aerosol size spectra is characterized by two modes, but unimodal size distributions are also common. Three or four modes in a distribution are rare in the ambient atmosphere (Seinfeld and Pandis, 2006). New particle formation bursts can occur on a large scale. Several mechanisms have been suggested to explain nucleation in the atmosphere (Jung et al., 2006). These include binary homogeneous nucleation of sulphuric acid e water (Vehkamäki et al., 2002), ternary homogeneous nucleation of sulphuric acid e water e ammonia (Kulmala et al., 2000; Lazaridis, 2001), ion-induced nucleation (Laakso et al., 2002; Lee et al., 2003; Winkler et al., 2008), nucleation of organic vapours (Zhang et al., 2004) and heterogeneous nucleation (nucleation on pre-existing particles) (Lazaridis and Kulmala,1992; Lazaridis et al., 1991; Winkler et al., 2008). Furthermore, different particle formation mechanisms might be prevailing in different environmental conditions (Dal Maso et al., 2005; Lehtipalo et al., 2009; Rimnacova et al., 2011; Yoon et al., 2006; Vuollekoski et al., 2010). The last 20 years researchers have managed to identify and reliably measure particles with electrical mobility diameters below 20 nm. Recently, new condensation particle counters (CPCs) have been developed, so clusters with diameter below 3 nm can be measured (Kulmala et al., 2007a; Sipilä et al., 2008), even though the cluster concentration does not seem to be the main critical factor determining whether a nucleation burst occurs (Lehtipalo et al., 2009). On the other hand, the concentrations of sulphuric acid, ammonia, water and organic compounds, precursors to particle nucleation and growth, are believed to be important factors (Boy et al., 2008; Jeong et al., 2010; Kulmala et al., 2005; Paasonen et al., 2010). The current paper presents an analysis of measured particle number size distributions during several measurement campaigns conducted between June 2009 and June 2010, at Akrotiri station, Crete, Greece. The main objectives of this work were to identify the levels of particle number concentration in the region and correlate them to the possible particle sources, to study the average atmospheric particle size distributions for Nucleation, Aitken and Accumulation modes, to investigate the occurrence of nucleation events (NEs) and to evaluate particle formation and growth rates.

791

of east Mediterranean. Detailed description of the Akrotiri research station can be found in Lazaridis et al. (2008) and Kopanakis et al. (2012). The measurement site is approximately 5 Km from the city of Chania (108,000 inhabitants) and 140 Km from the city of Heraklion (173,000 inhabitants). A weather station is located at the measurement site and meteorological data for temperature, humidity, wind speed and wind direction, solar radiation and rainfall are recorded at a 10 min time interval. There are no major industrial facilities in the area, except of a power plant placed within 10 Km from the station. The weather in the area is characterized by hot dry summers and mild winters with low precipitation (401e662 mm yearly average values for the last decade). Clear sky conditions prevail most of the days while mean temperatures for winter and summer are 12  3.1  C and 26  3.59  C respectively. The yearly average relative humidity is 65  12% with minimum average values during summer (56  14%) and maximum average values during winter (71  11%). Particle number size distributions for aerosols with mobility diameters between 10 and 1100 nm (scanned range) have been measured during several campaigns between June 2009 and June 2010. A Scanning Mobility Particle Sizer (SMPS) system, consisting of a Condensation Particle Counter (CPC) and a Differential Mobility Analyzer (DMA) (SMPS, Model 5.401, Grimm Inc., Germany), was used to measure the particle size distributions. Particles sizes were scanned every 406 s into 43 consecutive size bins. The entire system is automated and is equipped with customized software, provided by Grimm, which was used for the inversion of the raw size distributions and for exporting the data to other applications. The inlet of the system was placed 2.2 m above the ground and a cover was used to protect it from the rain. The SMPS system was located inside a laboratory container, which was air conditioned. The indoor temperature was 22  C during all measurements campaigns. Sheath air was circulated in a closed loop through a drying system, designed to keep relative humidity below 20%. Aerosol measurements were conducted at ambient relative humidities. All the collected data were examined for flawed measurements due to instrumentation malfunctions, which were removed from the data set. Diffusional losses were calculated according to Hinds (Hinds, 1999) and the penetration efficiency was 96% for particles with diameters below 30 nm, 97% for particles with diameters between 30 nm and 50 nm and more than 97% for bigger particles. On the whole, the collected data covered a period of 54 summer days (38 days between 23 June and 18 August, 2009 and 16 between 2 June and 26 June, 2010), 41 autumn days (between 9 October and 30 November, 2009), 21 winter days (between 12 January and 12 February 2010) and 51 spring days (between 3 April and 31 May, 2010). 2.2. Back trajectories Back trajectories were studied to reveal the origin of air mass reaching the study area. The trajectories were obtained using the HYSPLIT4 Model (HYbrid Single-Particle Lagrangian Integrated Trajectory), developed by the Air Resources Laboratory of the National Oceanic and Atmospheric Administration (NOAA) (Draxler and Rolph, 2003). The 3-dimensional trajectories were computed for the Akrotiri station (coordinates 35.53 N, 24.06 E). The trajectories were 120 h (5 days) long. 2.3. Data processing e the AMANpsd algorithm

2. Materials and methods 2.1. Measurement site e campaigns Measurements were conducted at Akrotiri station, a coastal site in the northwest part of Crete, Greece. Crete is located in the middle

An aerosol algorithm (AMANpsd algorithm) was used to evaluate the modal structure of the measured particle number size distribution data (Ondracek et al., 2009). The algorithm can be used 1) for merging data derived from instruments measuring particle number concentrations in different size ranges or operating under

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

2 3 2  n logDp  logDpg;i X dN Ni 5 pffiffiffiffiffiffiffi ¼ exp4    dlogDp 2log2 sg;i i ¼ 1 2plogsg;i

percentile 25%

min

median

max

percentile 75%

100000

Particle number (#/cm3)

different physical principles (SMPS and Aerodynamic Particle Sizer (APS)) and 2) for calculating and evaluating the variables that describe the multilognormal size distributions of the experimental data sets. In this work the measurements were performed using only a SMPS and therefore the algorithm was used just for data parameterization and more specifically for calculating the geometric mean diameter (GMD), the geometric standard deviation (GSD) and the number concentration of each of the modes comprising the aerosol size distribution. The multilognormal model is widely used for describing the aerosol size distribution (Hussein et al., 2004; Seinfeld and Pandis, 2006) and is also used in the AMANpsd algorithm. The multilognormal distribution is mathematically expressed as:

10000

1000

100

10

Summer 2009 Autumn 2009 Winter 2010

Spring 2010 Summer 2010

Fig. 1. Median particle number concentration, standard deviation and descriptive statistics for campaigns during each month of the sampling period.

(1)

where n is the number of modes, Ni the number concentration in each mode, Dp is the aerosol diameter, Dpg,i is the geometric mean diameter in each mode, and sg,i is the geometric standard deviation. The AMANpsd algorithm reconstructs the experimental data set by identifying the different modes of each size distribution using a constrained minimization method (Neadler-Mead simplex method). The method is applied to minimize the difference between the multilognormal model and the experimental data. Boundary conditions for each parameter computed by the algorithm are set, and these conditions are used as constrains in the minimization procedure. Finally the resulting parameters are evaluated by the algorithm and if these parameters do not fulfil certain physical criteria the procedure is repeated and new parameters are computed. 3. Data analysis and parameterization results 3.1. Particle number concentrations The median value of the total particle number concentration in the diameter range 10e1100 nm was 525 #/cm3. The number concentration ranged between 130 #/cm3 and 9597 #/cm3 and the maximum concentration was recorded on 15/4/2010. Particle concentration measurements above 1000 #cm3 accounted for 20% of total measurements. The median value was at least 20 times lower compared to the concentrations measured in urban stations in the Mediterranean region, which is expected since these stations are strongly affected by local traffic emissions and by emissions from the nearby industrial areas (Costabile et al., 2010; Reche et al., 2011). On the other hand, the measured values in Akrotiri station are closer to the concentrations in the remote Finokalia station (northern eastern Crete, Kalivitis et al., 2008). Particle number concentration in Finokalia ranged between 900 #/cm3 and 2500 #/ cm3 during two month period in 2005. This is also in agreement with the work of Putaud et al. (2010), presenting particle measurements from different sites in Europe. The comparison of the data sets in the study by Putaud et al. (2010) depicted a clear decrease in particle concentration from urban to rural sites. In Fig. 1, a statistical analysis of seasonal data during the sampling period is presented. The fluctuations in particle concentrations were broader in the winter period. High variations also occurred in the spring. On the contrary, during the summer period, the particle number concentrations presented lower variability, probably due to the presence of stagnant atmospheric conditions. Moreover, the analysis of the data showed that at Akrotiri station the particle number concentrations during weekends was equal to the corresponding values of workdays. The median particle

number concentrations for the whole measurement period were 537 #/cm3 [160 #/cm3 e 6100 #/cm3] for weekends and 512 #/cm3 [130 #/cm3 e 9600 #/cm3] for workdays. Similar conclusions can be deduced from the diurnal cycles of particle number size distributions measured during weekends and weekdays, which are presented in Fig. 2. The same trend of the particles concentration appeared when examining the data sets for the measurements conducted during each season of the year separately, with the exception of October and November 2009, when the average particle number concentration for weekends was higher by 10% in comparison to the workdays. Nevertheless, the similarities in particle number concentrations for weekdays and weekends along with the low particle concentration values indicate that emissions from the vehicle fleet did not result in changes to the particle concentration in the study area, since traffic load was reduced during weekends. Following previous works (Gómez-Moreno et al., 2011; Hussein et al., 2004) the particle number distributions were divided in two classes: N10e100 (10 nm < Dp < 100 nm) and N100 (100 nm < Dp < 1000 nm) particles. The two classes were selected according to the differences in the chemical and physical properties of the particles and to the diverse emission paths that lead them to the atmosphere. Special attention was given to particles with diameters between 10 nm and 50 nm (nano-particles, NP), since the presence of these particles is related to adverse health effects (Franck et al., 2011; Wichmann et al., 2000). The particle number concentrations showed seasonal variability, with maximum values observed during spring campaigns for total and N100 particles and during winter campaigns for NP and N10e100 particles. The average particle number concentrations for all the campaigns, according to the season that these campaigns were performed, along with the corresponding meteorological conditions are summarized in Table 1.

Weekdays Particle number concentration (#/cm3)

792

Weekends

800 Total

600 400

N10-100

N100

200 0 05:00

08:00

11:00

14:00

17:00 20:00 Time of day

23:00

02:00

05:00

Fig. 2. Diurnal cycles (median values of particle number concentration) for total particles and for N10e100 and N100 particles for weekdays and weekends.

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

793

Table 1 a) Median particle number concentration and ranges, b) average temperature average relative humidity and average wind speed for each sampling period. Measurements were conducted at ambient relative humidities. (a) Sampling campaigns

Total number concentration (#/cm3)

N10e100 number concentration (#/cm3)

N100 number concentration (#/cm3)

NP number concentration (#/cm3)

Summer 2009 Autumn 2009 Winter 2010 Spring 2010 Summer 2010

515 476 391 747 443

274 303 310 402 260

250 [10e790] 174 [10e834] 74 [10e1325] 213 [20e1375] 174 [56e674]

68 [12e2399] 125 [11e2067] 184 [20e2589] 186 [15e3416] 76 [17e932]

[227e2933] [130e2872] [120e6425] [120e9596] [180e2702]

(b) Average temperature ( C) 25 18 13 21 24

    

[35e2586] [20e2317] [30e5100] [55e5697] [81e2080] Average RH (%)

3.4 3.2 3.7 3.4 3.8

58 73 70 60 60

    

Results for the summer period of 2010 incorporate measurements only during June 2010. Maximum values of particles number concentration were not recorded during days with nucleation events. Elevated particles concentration values, in remote or suburban stations, can be caused from the transport of polluted atmospheric masses and also from new particle formation events (Lazaridis et al., 2005). N10e100 particles number concentration was higher, during all the measurements periods, than that of the N100 particles. Their difference fluctuated from only 8.5% during summer period campaigns to 165% during autumn period campaigns. Representative data (hourly average values) for April and May 2010 are presented in Fig. 3. The contribution of N10e100 particles to the total particle concentration was season dependent. More precisely, in spring and summer sub micrometer particles consisted of 53% by N10e100 particles, while the percentage increased to 80% in winter. A corresponding increase was observed also for NP and their percentage in respect to total particles increased from 22% in summer to 52% in winter. The increase in the percentage of NP cannot be fully explained by local pollution sources but it can partly attributed to emissions from domestic heating during winter. These emissions, in conjunction with the lower boundary layer during winter, can contribute to the increase in the N10e100 particles. Nevertheless, when considering all the available data, it can be concluded that the station area is mostly affected by polluted air masses from Athens and continental Europe (Lazaridis et al., 2008). To investigate the effect of local meteorological conditions to the particle number concentration, correlations between all available concentration values versus wind speed and temperature were performed. No significant correlations were observed between number concentration and temperature both for N10e100 and N100 particles. Therefore, it can be assumed that temperature variations do not affect significantly the particle number concentration at the

Particle Number (#/cm3)

3000

2000

1000

0 2/4

12/4

22/4

2/5

12/5

22/5

1/6

11/6

Date Fig. 3. N10e100 (black line) and N100 (grey line) particle number concentrations for the period AprileMay 2010.

13 11 13 13 14

Average wind speed (km h1) 8.8 7 9.8 9.4 9.4

    

4 3.8 4.1 5.5 7

Akrotiri station. Fairly good negative correlations were obtained between number concentration and wind speed with r values ranging for total particles from 0.55 for summer and spring to 0.71 for winter (P < 0.05). High wind speed favours the atmospheric dispersion and lead to lower particle concentrations in the ambient atmosphere. Similar results were obtained by Galindo et al. (2011) who studied PM mass concentration in the city of Elche (Spain). Elche has very similar characteristics to Chania regarding both climatic conditions and population density. On the other hand, the effect of the wind to the particle number concentration seems to be stronger in larger and most polluted cities (Mejia et al., 2008; Hussein et al., 2006; Wu et al., 2008). 3.2. Analysis of concentration variations according to the origin of air masses The collected wind data at the station in conjunction with the analysis of the back trajectories, provided by the HYSPLIT model, showed that air masses were reaching the station mainly from West, North and Northwest in a total percentage of 78% for the above three directions. Northeast directions followed with a share of 15%. This is in agreement with previous studies (Guerzoni et al., 1990; Lazaridis et al., 2008). Further attempts to correlate the results of the HYSPLIT model with the presence of particulate matter showed that different paths of the arriving air mass can be associated to different levels of particle number concentrations. Fig. 4 presents the daily median number concentrations versus the origin of air masses Median daily concentrations over 1000 #/cm3 were recorded only for air masses coming from North and Northwest, after travelling over the mainland of central and Western Europe. Elevated concentrations were observed also for air masses reaching the station from Northeast, crossing land areas of the Eastern Europe. In these cases, daily median particle number concentrations varied between 300 and 850 #/cm3. The highest median value of the daily median concentrations [653 #/cm3] corresponded to air masses arriving at the station from Northeast, as it can be seen in Fig. 5, where the 25% and 75% percentiles and the maximum and minimum values are also presented. Air masses arriving from the West after travelling mainly over sea area contributed less to the particle load in the station area. Air masses passing over the North African coasts (South, Southeast and Southwest directions) are shown together, since they represent less than 7% of the total observed cases. These air masses led to particle concentrations with median values almost equal to those of the air masses coming from polluted areas of the central and Western Europe. Such air masses are characterized with increased coarse particle loading during Saharan dust events.

794

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

station for a four months period and reported higher concentration values (1800e2500 #/cm3) for air masses arriving at Finokalia from Turkey and Black sea in comparison to particles concentrations influenced by marine air masses (900e1200 #/cm3). 3.3. New particle formation events

Fig. 4. Daily median values of particle number concentrations versus the origin of the air masses reaching Akrotiri research station.

A characteristic case showing the effect of the particle load of the transported air masses to the study area is discussed below. Back trajectories plots referring to the summer period are presented in Fig. 6 along with particle number concentration versus time (Fig. 7). Fig. 6a depicts the plots for the air mass trajectories during the periods 23e26/6/2009 and 1e6/7/2009 (“Period I”), whereas Fig. 6b shows air mass trajectories during the period 14e 22/7/2009 (“Period II”). The particle number concentrations during the examined time periods are showed in Fig. 7. In “Period I”, air masses were transported on rather low altitudes and arrived at the Akrotiri station from Northwest directions, through mainly sea areas. These air masses presented particle number concentrations in the range 230e1400 #/cm3, with average value 400 #/cm3. On the other hand, “Period II” consists of air masses originating from North and Northeast directions with number concentrations ranging between 190e2935 #/cm3 and higher average value (699 #/cm3) compared to period I. Air masses coming from Northeast directions were influenced by pollution sources located in Eastern Europe and were carrying gaseous pollutants that affected the particle load at the measurement’s site. These results are in agreement with the results of Kalivitis et al. (2008), who analyzed the correlation between the number concentration and the origin of air masses in the Finokalia

percentile25%

min

median

max

percentile75%

During the sampling period, several NEs took place, resulting in the formation of new particles. Days with NEs were identified using the ultrafine particles number concentration and the results of the particles size distributions analysis performed with the AMANpsd algorithm. In order to specify a time period as NE three criteria should be met: a) the ultrafine particles number concentration should exceed twice the monthly average value b) a mode with GMDs below 50 nm should be present for at least 1 h c) the GMDs in this mode should present an increasing trend. The end of the NEs was set at the time when the nano-particles number concentration returned to the values before the beginning of the nucleation burst. The NEs were classified as short lived (SL) if their duration was less than 6 h and long lived (LL) if their duration exceeded 6 h Table 2 presents the NEs which took place during the sampling period (June 2009eJune 2010). The duration, atmospheric temperature and relative humidity for every nucleation event are also presented. In total, 13 new particle formation events were recorded during the sampling periods. Nucleation events during night were also observed. In this study, nocturnal nucleation particle formation event is defined as the burst of new particles occurring at nighttime, after sunset and in the absence of photochemical activity. These events were registered throughout all the seasons (2 in summer, 3 in spring and 1 in winter). Three of the night time events were SL. Nocturnal NEs are associated with the presence of oxidations products of monoterpens (Ortega et al., 2012). Days with NEs represented less than 10% of the total measurement days. Kalivitis et al. (2008) reported only 3 NEs during 4 months measurements in Finokalia station and Pikridas et al. (2012) reported 58 NEs during a one year measurement campaign at the same site. Place et al. (2010) observed 13 NEs during 3 months campaign in a rural area. Also, Gómez-Moreno et al. (2011) studied 127 NEs during 2 years of measurements in the urban station of CIEMAT (Madrid, Spain). On the other hand, in remote areas the percentage of days with the appearance of NEs can reach the values of 45% (Dal Maso et al., 2005) or even 90% (Hallar et al., 2010). 4. Discussion 4.1. Particle size distributions

1400

Particle number (#/cm3)

1200 1000 800 600 400 200 0 N

NW

W

NE

S/SW/SE

Wind Direction Fig. 5. Median particle number concentration, standard deviation and descriptive statistics of median daily number concentrations versus the origin of air masses reaching Akrotiri station.

Particle size distributions were analysed in order to identify possible particle modes and determine their characteristics. Modal analysis of particle size distributions can be a powerful tool for specifying the origin and possible sources of particles found in the atmosphere (Costabile et al., 2009; Mejia et al., 2008). The analysis of the collected data with the AMANpsd algorithm revealed that particle size distributions were mainly bimodal with the exception of the time period referring to the measurements of the summer of 2009 where unimodal size distributions prevailed. The percentage of the bimodal size distributions ranged between 42% and 57%. The number of the unimodal size distributions was higher during spring and summer compared to the autumn and winter. Periods with nucleation events were treated separately and the apportionment of the size distributions was the same to the non nucleation periods (the percentage of the bimodal size distributions during nucleation events was 57%).

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

795

Fig. 6. Typical back trajectories calculated by HYSPLIT4 Model for two different summer periods (a) Period I: 23/6e26/6 2009 and 1/7e6/7 2009 (b) Period II: 14/7e22/7 2009).

The characteristics of the aerosol size distributions were examined separately for weekdays and weekends. The median GMDs, number concentrations in each mode and the corresponding standard deviations of each mode, for each season and for periods with NEs are presented in Table 3a,b. Twelve of the thirteen NEs occurred on weekdays and therefore the statistics for days with NEs were incorporated in the weekdays (Table 3a). No significant differences in GMDs values were observed between weekdays and weekends indicating that variations in emissions from the vehicle fleet cannot explain changes in the size distributions characteristics. Minimum GMD values were computed for the measurements during the winter of 2009. The GMD in the first mode, for all bimodal distributions, presented the same median value with the exception of the autumn period. On the other hand, the median values for the unimodal distributions and for the second modes of bimodal distributions were lower in the winter period. The GMD values for unimodal and bimodal distributions for nucleation events were closer to the corresponding winter median GMDs. Considering the fact that 11 of the 13 nucleation events were reported during spring and summer, the presence of smaller particles during nucleation can be verified from the size distribution

Fig. 7. Particle number size concentration for the summer period 23/6/2009e14/7/ 2009.

analysis. The median particle number concentration in unimodal distributions and in the first mode of bimodal distributions for NEs was at least 1.5 higher than the corresponding concentrations for periods without NEs due to the presence of very small particles. Aerosol size distributions were season dependant, as it can be seen in Fig. 8, where the median size distributions for each season campaigns are depicted. The summer and spring median distributions presented peaks in higher diameters, suggesting that accumulation mode particles prevailed in the station area during these time periods. Furthermore, frequency plots were made for the GMDs of all size distributions (including nucleation events) according to the season of the year that the measurements were conducted (Fig. 9). It can be clearly seen that different particles sizes dominated during each season measurement campaigns. The frequency plots were combined with back trajectories data and several particles modes were identified. During winter, a strong nucleation mode was present and it was centered at diameters in the range of 15e25 nm. Only two nucleation events were recorded during the winter campaign and the observed nucleation mode can be considered to be a result of emissions from local heating. An Aitken mode was identified with the corresponding GMDs ranging between 40 and 60 nm. Particles in this mode can be correlated to local emissions but can also be products of coagulation and condensation of nucleation particles. On the contrary, an accumulation mode prevailed during campaigns conducted in summer. GMDs ranged between 80 and 110 nm and the presence of the mode can be attributed to the effect of polluted air masses transported from central and Eastern Europe. A nucleation mode was also present, that in most of the cases was comprised from all distributions with GMDs to 50 nm, and included all the fresh and aged particles generated during nucleation events. The nucleation mode, identified in spring campaigns, was observed during nucleation processes (6 NEs occurred in spring), but was also present in days without nucleation. Possible sources of particles in this mode could be local emissions, although the use of domestic heating is significantly reduced in the city of Chania in spring compared to winter. The accumulation mode in spring

796

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

Table 2 Characteristics of nucleation events at Akrotiri station, for the whole sampling period (June 2009eJune 2010). LL refers to NEs with duration higher than 6 h, while SS to NEs with duration less than 6 h. Event (#)

Start

Finish

Duration (hr)

Temperature ( C)

Rel. hum. (%)

Type

1 2 3 4 5 6 7 8 9 10 11 12 13

07/07/09 (10:00) 09/07/09 (09:00) 03/08/09 (20:50) 13/08/09 (06:30) 17/08/09 (21:30) 03/02/10 (13:30) 09/02/10 (20:10) 06/04/10 (21:45) 07/04/10 (13:49) 09/04/10(16:45) 10/04/10(15:20) 19/04/10 (21:15) 26/05/10 (22:00)

07/07/09 (17:30) 09/07/09 (23:05) 04/08/09 (03:00) 13/08/09 (11:00) 18/08/09 (01:30) 04/02/10 (01:30) 10/02/10 (04:00) 07/04/10 (11:12) 08/04/10 (10:36) 10/04/10(00:20) 11/04/10 (03:58) 20/04/10 (02:13) 27/05/10 (05:36)

7.5 13 6 4.5 4 12 8 13.5 21 7.5 12 5 7

31 28 24 24 23 6 16 14 13 14 13 17 16

53 57 74 62 66 60 86 73 70 65 66 83 83

LL LL SL SL SL LL LL LL LL LL LL SL LL

(GMDs between 80 and 120 nm) was the result of the influence of polluted air masses to the station area and of the ageing of nucleation and Aitken particles. No clear conclusions can be drawn from the analysis of size distributions in autumn. GMDs were almost equally shared to a nucleation mode, an Aitken mode and an accumulation mode. No nucleation events were recorded during the autumn campaigns. The back trajectories analysis showed that in 50% of the measurement’s days, the air masses arriving at the station were either travelling above sea areas or remained in the Aegean Sea for more than 3 days. Therefore, autumn measurements

could be regarded as background to the station, since particle load was not significantly influenced by local sources or from heavily polluted air masses reaching the Akrotiri station. 4.2. Nucleation events 4.2.1. Methodology Growth rate (GR) and formation rate (JD) were calculated for each NE. The Growth Rate (GR) refers to the increase of particle size with time. In general, it can be calculated from the following formula:

Table 3 a, b: Median values of GMDs, number concentration and standard deviation sorted according to the season period for weekdays (a) and weekends (b). Corresponding median values during nucleation events are presented separately in the weekdays table, since 12 out of 13 nucleation events occurred on weekdays. The numbers 1, 2, 3 refer to the corresponding mode (1 for 1st mode, etc.). Number concentration [#/cm3]

GMD [nm] Campaigns

1

2

(a) Median values for weekdays for the whole data set Unimodal Summer 2009 106 Autumn 2009 75 Winter 2010 59 Spring 2010 99 Summer 2010 96 Nucleation 77 Bimodal Summer 2009 25 111 Autumn 2009 37 132 Winter 2010 25 79 Spring 2010 29 109 Summer 2010 26 107 Nucleation 25 98 Trimodal Summer 2009 9 42 Autumn 2009 16 52 Winter 2010 15 46 Spring 2010 14 46 Summer 2010 10 42 Nucleation 16 50 (b) Median values for weekends for the whole data set Unimodal Summer 2009 113 Autumn 2009 78 Winter 2010 67 Spring 2010 94 Summer 2010 90 Bimodal Summer 2009 24 119 Autumn 2009 26 99 Winter 2010 22 82 Spring 2010 26 105 Summer 2010 26 92 Trimodal Summer 2009 8 44 Autumn 2009 14 54 Winter 2010 15 49 Spring 2010 12 48 Summer 2010 7 35

3

1

120 146 135 126 123 130

475 450 496 504 587 843 46 133 158 128 58 239 23 53 74 60 30 158

122 144 144 124 110

478 418 586 530 565 29 93 155 103 61 17 55 69 49 38

2

434 99 223 415 530 464 60 130 152 143 105 229

440 243 364 435 472 40 132 161 135 51

Standard deviation 3

1

2

3

365 94 67 280 411 158

1.74 1.85 1.93 1.82 1.69 1.73 1.55 1.50 1.52 1.50 1.50 1.41 1.60 1.54 1.49 1.56 1.60 1.45

1.66 1.50 1.65 1.64 1.60 1.67 1.53 1.48 1.54 1.47 1.49 1.49

1.60 1.49 1.50 1.59 1.56 1.61

454 103 89 292 442

1.76 1.87 1.92 1.84 1.67 1.57 1.49 1.54 1.52 1.44 1.59 1.62 1.59 1.57 1.61

1.67 1.67 1.66 1.67 1.64 1.49 1.56 1.54 1.51 1.51

1.62 1.50 1.51 1.63 1.62

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

dN/dlog(Dp) (#/cm3)

Summer

Autumn

Winter

Spring

JD ¼

GMD: 102 nm GSD: 1.79

1000 GMD: 70 nm GSD: 2

750

GMD: 87 nm GSD: 1.78

DN Dt

797

(4)

(3)

where N is the number of particles at a specific time period and t is the time period. Since there were available experimental data for particles bigger than 10 nm, the particle formation rate corresponds to particles resulted from nucleation and not to the actual particles born during the nucleation bursts. JD was calculated for each time interval of the nucleation and the total formation rate was obtained from the sum of the individual formation rates. The instrument used in the measurement campaigns can only detect particles with aerodynamic diameters bigger than 10 nm. This limitation created a challenge in separating the burst of new particles from the particle growth. Although nucleation and growth are considered to be decoupled (Kulmala et al., 2004), the lack of data regarding a) the concentration of particles with aerodynamic diameters less than 10 nm and b) the concentration of sulphuric acid and ammonia which play an important role to the nucleation process, made it difficult to accurately calculate the duration of the nucleation bursts. In this study, in order to separate the burst of new particles from the particle’s growth, the measurements of the first six size bins, that include the particles with aerodynamic diameters less than 20 nm, were used to identify the newly formed particles. The formation rate (JD) of these particles was used to evaluate the duration of the nucleation burst and separate it from the particle growth time. More precisely, when the JD of the particles with aerodynamic diameters less than 20 nm was reaching values close to zero, the nucleation burst was assumed to be completed.

where k is the number of size bins in the distribution, Nk is the number of particles which belong to diameter Dk and Nt is the sum of particles which belong to the modes of the diameter, as discussed previously. While Dp was calculated for every time interval of the nucleation duration, for the calculation of GR, two values are needed, the first and the last one. The new particle formation rate refers to the increase in the number of particles during the nucleation burst and it is given from the equation:

4.2.2. Analysis of results The duration of a nucleation phenomenon at Akrotiri station, varied between 4 and 21 h, with an average value of 9.3 h. Furthermore, growth and formation rates during the NEs were computed from the original sampling data (Table 4). It was found that the average GR and JD were 6  4 nm h1 and 13  10 # cm3 sec1, respectively. Computed GR and JD values for different environments are presented in Table 5. Growth rates are depended on the concentrations of precursor gasses, such as sulphuric acid and

500

GMD: 55 nm GSD: 2.1

250 0 10

100

1000

Diameter (nm) Fig. 8. Median size distributions for each season campaigns.

GR ¼

DDp Dt

(2)

where DDp is the size range of the particles from D to Dmax and t is the time required for the particles to reach Dmax. According to Qian et al. (2007), in order to calculate GR, every size distribution of each time interval of the collected data has to be examined manually. Most of the size distributions had a minimum in a mode less than 100 nm. It was considered that these particles participated in the nucleation burst. For those distributions that had no minimum it was assumed that particles, right after the maximum of the curve contributed to the phenomenon. The mean particle diameter Dp, was obtained from the equation:

Pk Dp ¼

Nk Dk Nt

i¼1

Fig. 9. Frequency plots for the geometric mean diameters of the measured size distributions. The values in x-axis refer to the lower limit of each size bin.

798

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

Table 4 Computed Growth rates (GR) and Formation Rates (JD) of the NEs reported in Akrotiri station. LL refers to NEs with duration higher than 6 h, while SS to NEs with duration less than 6 h. Event (#)

GR (nm h1)

JD (cm3 s1)

Type

1 2 3 4 5 6 7 8 9 10 11 12 13

6.3 4.7 8.4 12.4 14.1 1.4 3.1 1.6 1.5 6.6 3.1 5.9 6.8

37.5 19.9 1.4 9.2 6.5 18.1 10.8 13.2 22.0 14.3 5.4 0.3 12.0

LL LL SL SL SL LL LL LL LL LL LL SL LL

ammonia (Holmes, 2007; Pikridas et al., 2012). Further measurements must be performed in order to fully identify the conditions that can lead to nucleation bursts in the study area. Based on the calculated formation rates, the “burst time” corresponded to 20e80% of the total time of the nucleation phenomenon, with median length 53% (or in absolute terms 5 of 9 h). The main characteristic during a NE is the increase in the aerosol number concentration at the ambient atmosphere. Fig. 10a, c describe a NE that took place on July 7, 2009 while Fig. 10b depicts a nocturnal NE on May 26, 2010. The formation of new particles resulted in an increase in the number concentration of particles smaller than 50 nm due to condensation of semivolatile species on new particles (Fig. 10a). In many cases, newly fresh particles can quickly shrink due to evaporation and eventually disappear (Harrison et al., 2012). Preexisting larger particles at Akrotiri station did not act as condensation sinks, since the median concentrations during all NEs ranged between 200 #/cm3 and 400 #/cm3. Low number concentration values are very important for the occurrence of NEs since the surface of these particles is a major factor for the disappearance of newly formed particles. Larger particles can also decrease the number of fresh particles due to coagulation on their surface (Kulmala et al., 2001). Coagulation sinks for particles with mobility diameters 1 nm, 3 nm and 5 nm were computed during all the reported NEs. The coagulation coefficients were computed by the algorithm described by Seinfeld and Pandis using the Fuchs correction coefficient (Seinfeld and Pandis, 2006). The ranges of the computed sinks at Akrotiri station were 3$103 e 1.4$103 s1, 4.7$104 e 9$104 s1 and 2$104 e 3.8$104 s1for 1 nm, 3 nm and 5 nm particles respectively. The above values are similar to the values reported by Kulmala et al. (2001), who studied NEs at Hyytiala (Finland). The combination of low coagulation sinks and low N100 particles number concentrations can lead to nucleation bursts at the study area.

Table 5 Recent observations on new particle formation events characteristics. Reference

Region

GR (nm h1)

JD (cm3 s1)

RH (%)

Current study Jeong et al., 2010 Salma et al., 2010 Lee et al., 2008 Kalivitis et al., 2008 Hamed et al., 2007 Wu et al., 2007 Qian et al., 2007 Dal Maso et al., 2005 Kulmala et al., 2004

Rural/suburban Urban e Rural Urban Coastal Remote coastal Continental Boreal forest Urban Continental Review

5.8  4.0 3.0 7.7  2.4 5.9

13.1  9.9 11e17 4.2  2.5

69  10 61

7.0 0.1e11.2 5.9 3.0 1e20

<67 1.1e1.7 6.0 3.3e81.4 17 0.8 0.01e10

The presence of a mode with GMDs below 50 nm, with an increasing trend, is clearly depicted in Fig. 9b during the first 2 h of the phenomenon. The GMDs in this mode were shifted towards bigger diameters and this increase can be fitted by a first order polynomial (r2 ¼ 0.9). The size distribution was bimodal during the first 2 h, with the first mode (nucleation mode) describing the newly formed particles, while the second mode was referring to the preexisting particles in the atmosphere. When the GMD in the nucleation mode reached the value of 46 nm, the size distribution turned to unimodal with GMDs between 70 and 105 nm. Moreover, during the event, the average number concentration was 793  281 # cm3, whereas for over 5 h, the concentration was higher than the monthly average value of 600 # cm3, with the maximum value to exceed 1200 # cm3. The ultrafine particles number concentration presented a maximum at 246 # cm3, which was 2.7 times higher than the corresponding average concentration calculated for July 2009. The percentage of ultrafine particles to total particles was ranging between 50% and 80% while the GMDs values in the nucleation mode were increasing. In addition, in all the nucleation burst cases examined, the above percentage ratio of ultrafine to total particles was higher than the background level (ranging by 20%e56%). Furthermore, during the nucleation outbreaks, the percentage of particles in diameters 13e35 nm was higher by 20% in relation to its season average value, which is in agreement with other results in eastern Mediterranean (Kalivitis et al., 2008). Increased number concentrations, during nucleation events, for particles with diameters below 30 nm in comparison to their average seasonal number concentrations were also reported in other studies (Stanier et al., 2004; Wu et al., 2007). The temporal increase of particles concentration during NEs is depicted in Fig. 11, where a LL nucleation event is described. The nucleation phenomenon was divided in four time periods according to the GMDs evolution and the median concentration (dN/d log Dp) is plotted for every time period. Fig. 11 refers to a nucleation burst on 9th of April, 2010. Its duration was 7.5 h. At first (Phase 1, duration 108 min), the median particle size distribution showed a bimodal distribution with two modes with mean diameters located at 18 nm and 94 nm. During this phase, the GMD in the nucleation mode was rapidly increased from 14 nm to 20 nm. During second and third phase, the particle number distribution still had a bimodal structure with two modes, at 21 nm and 90 nm (Phase 2) and at the 25 nm and 90 nm (Phase 3). Throughout these two phases, the rate of increase in the GMD of the first mode was lower and at the same time a raise in the particle concentration of the second mode was observed due to the coagulation of new particles and condensation. At the last phase, the median size distribution was unimodal (the mean value of the mode was at 62 nm and it is presented at the last plot of Fig. 11). In addition, an attempt was made to correlate nucleation events with atmospheric conditions (temperature; relative humidity; wind direction; solar radiation) and the presence of air pollutants (PM10 mass concentration e FH 62 I-R Beta gauge, ESM Andersen Instruments GmbH, Erlangen, Germany; O3 e APOA 360, Horiba, Japan) which are continuously recorded in the Akrotiri Station. Although, there was not significant correlation between meteorological parameters and the new particle formation events recorded  at Akrotiri station, as shown also in other studies (Zdímal et al., 2010), it can be noted that from the observed nucleation episodes, only two (both during summer period) occurred at mean relative humidity values lower than 60% (see Table 2). This is in agreement with the study by Lee et al. (2008), where the observed NEs occurred for relative humidities above 67%. Mean wind speed was below 9 km h1 in all nucleation cases with the exception of the number 8 NE (see Table 2), when the corresponding value was

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

799

Fig. 10. Description of the NEs on a) 7/7/2009 and on b) 26/5/2010 (nocturnal NPF). c) GMDs of the first and second mode versus time for the NE on 7/8/2009. The points on 10a,b refer to the mid-diameters of each mode for every distribution. The nucleation mode on 10c refers to the first mode of the size distribution during the initial 2 h of the phenomenon.

15 km h1. However, wind speeds higher than 20 km h1 are frequently observed in the Akrotiri Station. No NEs were detected on cloudy days. The solar radiation values during the NEs were always lower than the maximum daily values during the same

month that the NE was observed. Furthermore, 5 NEs started at night time, as it was mentioned before. This is in contrast with other studies where an increase in solar radiation is observed during nucleation bursts (Kroll and Seinfeld, 2008; Lee et al., 2008;

800

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802 800

600

t1= 108 min DP= 18 nm

-3

dN/dlogDp (#/cm )

500

700

= 1.29

t1= 108 min

= 1.23

t2 = 216 min

DP= 21 nm

t2 = 108 min

600

400

500

300

400

DP= 94 nm

200

300

DP= 90 nm

= 1.80 200

100

= 1.72

100

0

0

10

100

1000

10

100

1000

1200

DP= 25 nm

= 2.83

t3 = 115 min

800

t4 = 440 min t4 = 109 min

DP= 62 nm

t3 = 331 min

= 1.36

-3

dN/dlogDp (#/cm )

1000

800 600

DP= 90 nm

400

400

= 1.65

200

0

0

10

100

Dp (nm)

1000

10

100

1000

Dp (nm)

Fig. 11. Dynamic evolution of a nucleation event at 09/04/2010 (number of particles (# cm3) vs. mobility diameter (nm)).

Qian et al., 2007). On the other hand, solar radiation values are generally high in the area, due to the geographical location of the station. It is known, that ozone, participates actively along with other species such as hydroxyl and nitrate radicals, in photochemical atmospheric reactions during the primary oxidation process by which organic aerosols produced from gas phase organic clusters (Kroll and Seinfeld, 2008). In our study, mean ozone concentrations, during NEs, ranged between 22 and 53 ppb and they were below the 75th percentile of the corresponding monthly ozone concentration. Furthermore, no NEs were recorded during time periods with Sahara dust events, which are common in the station area. Enhanced presence of large particles in the atmosphere results in huge condensation sinks for semivolatile species and prevents the ignition of nucleation bursts. The analysis of the back trajectories retrieved with the HYSPLIT4 model showed that the air masses that had reached the study area before the nucleation burst have passed through polluted European regions. More precisely, northeastern and northwestern winds were prevailing during NEs and the origin of the air masses was traced in Russia and Ukraine (7 events), Germany (1 event) and Italy/France (5 events). No southern winds were recorded during NEs, indicating that the necessary precursor chemical compounds were transported to the study area before the beginning of the phenomenon. 5. Conclusions Intensive field campaigns were carried out at Akrotiri station in a rural/suburban region of western Crete (Greece) between June 2009 and June 2010. Number concentrations and size distributions

of particles in the size range of 10e1100 nm were measured and 13 nucleation events were recorded. During all measurement periods, total particle number concentrations were lower than the corresponding number concentrations in urban Mediterranean sites. Particle number concentration presented similar values on weekends and weekdays which indicate that emissions from the vehicle fleet did not significantly influence the particle load in the study area. The analysis of the collected data with the AMANpsd algorithm showed that size distributions were mostly bimodal with GMDs presenting lower values during the winter period. Moreover, back trajectories showed that different air mass origins were associated with different levels of particle number concentrations. The effect on the particle load in the area was higher for air masses arriving from North and Northwest, through polluted areas of western and central Europe. The modal analysis of the measured size distributions revealed the presence of several particle modes. A strong nucleation mode was observed during winter campaigns. Particles in this mode were derived from emissions from local heating or from nucleation bursts. This nucleation mode was present also in spring, when 6 NEs were recorded. On the other hand, the accumulation mode during summer campaigns can be correlated to the influence of polluted air masses arriving from central and Eastern Europe at the station area. Formation of new particles was observed, leading to an increase of aerosol number concentration. The duration of nucleation events varied from 4 to 21 h with a mean value at 9.3 h. Moreover, mean growth and formation rates during the nucleation events were computed to be 6  4.00 nm h1 and 13  10 cm3 sec1, respectively. The 9 out of the 13 nucleation events took place during night time or occurred both during day and night. Air masses were

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

arriving at the station from northeastern and northwestern directions during NEs, carrying gaseous pollutants that favoured the initiation of nucleation. Furthermore, N100 median particle number concentrations ranged between 200 # cm3 and 400 # cm3 during nucleation bursts which resulted to low order coagulation sinks during the nucleation process. References Boy, M., Karl, T., Turnipseed, A., Mauldin, R.L., Kosciuch, E., Greenberg, J.P., Rathbone, J., Smith, J., Held, A., Barsanti, K., Wehner, B., Bauer, S., Wiedensohler, A., Bonn, B., Kulmala, M., Guenther, A., 2008. New particle formation in the front range of the Colorado Rocky Mountains. Atmospheric Chemistry and Physics 8, 1577e1590. Costabile, F., Birmili, W., Klose, S., Tuch, T., Wehner, B., Wiedensohler, A., Franck, U., König, K., Sonntag, A., 2009. Spatio-temporal variability and principal components of the particle number size distribution in an urban atmosphere. Atmospheric Chemistry and Physics 9, 3163e3195. Costabile, F., Amoroso, A., Wang, F., 2010. Sub-mm particle size distributions in a suburban Mediterranean area. Aerosol populations and their possible relationship with HONO mixing ratios. Atmospheric Environment 44, 5258e5268. Dal Maso, M., Kulmala, M., Riipinen, I., Wagner, R., Hussein, T., Aalto, P.P., Lehtinen, K.E.J., 2005. Formation and growth of fresh atmospheric aerosols: eight years of aerosol size distribution data from SMEAR II, Hyytiälä, Finland. Boreal Environmental Research 10, 323e336. Dall’Osto, M., Beddows, D.C.S., Pey, J., Rodriguez, S., Alastue, A., Harrison, R.M., Querol, X., 2012. Urban aerosol size distributions over the Mediterranean city of Barcelona, NE Spain. Atmospheric Chemistry and Physics 12, 10693e10707. Draxler, R.R., Rolph, G.D., 2003. HYSPLIT (HYbrid Single-particle Lagrangian Integrated Trajectory). Model access via NOAA ARL READY Website. NOAA Air Resources Laboratory, Silver Spring, MD. http://www.arl.noaa.gov/ready/hysplit4. html. Franck, U., Odeh, S., Wiedensohler, A., Wehner, B., Herbarth, O., 2011. The effect of particle size on cardiovascular disordersethe smaller the worse. Science of Total Environment 409, 4217e4221. Galindo, N., Varea, M., Gil-Moltó, J., Yubero, E., Nicolás, J., 2011. The influence of meteorology on particulate matter concentrations at an urban Mediterranean location. Water Air & Soil Pollution 215, 365e372. Gómez-Moreno, F.J., Pujadas, M., Plaza, J., Rodríguez-Maroto, J.J., MartínezLozano, P., Artíñano, B., 2011. Influence of seasonal factors on the atmospheric particle number concentration and size distribution in Madrid. Atmospheric Environment 45, 3169e3180. Guerzoni, S., Correggiani, I., Miserocchi, S., 1990. Windblown particles from ships and land-based stations in the Mediterranean Sea: a review of trace metal sources. Water Pollution Research Reports 20, 377e386. Hallar, G.A., Lowenthal, D.H., Chirokova, G., Borys, R.D., Wiedinmyer, C., 2010. Persistent daily new particle formation at a mountain-top location. Atmospheric Environment 45, 4111e4115. Hamed, A., Joutsensaari, J., Mikkonen, S., Sogacheva, l., Dal Maso, M., Kulmala, M., Cavalli, F., Fuzzi, S., Facchni, M.C., Decesari, S., Mircea, M., Lehtinen, K.E.J., Laaksonen, A., 2007. Nucleation and growth of new particles in Po Valley, Italy. Atmospheric Chemistry and Physics 7, 355e376. Harrison, R.M., Dall’Osto, M., Beddows, D.C.S., Thorpe, A.J., Bloss, W.J., Allan, J.D., Coe, H., Dorsey, J.R., Gallagher, M., Martin, C., Whitehead, J., Williams, P.I., Jones, R.L., Langridge, J.M., Benton, A.K., Ball, S.M., Langford, B., Hewitt, C.N., Davison, B., Martin, D., Petersson, K.F., Henshaw, S.J., White, I.R., Shallcross, D.E., Barlow, J.F., Dunbar, T., Davies, F., Nemitz, E., Phillips, G.J., Helfter, C., Di Marco, C.F., Smith, S., 2012. Atmospheric chemistry and physics in the atmosphere of a developed megacity (London): an overview of the REPARTEE experiment and its conclusions. Atmospheric Chemistry and Physics 12, 3065e3114. Hinds, W.C., 1999. Aerosol Technology. John Wiley & Sons, New York. Holmes, N.S., 2007. A review of particle formation events and growth in the atmosphere in the various environments and discussion of mechanistic implications. Atmospheric Environment 41, 2183e2201. Housiadas, Lazaridis, 2010. Inhalation dosimetry modelling. In: Lazaridis, M., Colbeck, I. (Eds.), Human Exposure to Pollutants via Dermal Absorption and Inhalation. Springer, Dordrecht, pp. 185e237. Hussein, T., Puustinen, A., Aalto, P.P., Mäkelä, J.M., Hämeri, K., Kulmala, M., 2004. Urban aerosol number size distributions. Atmospheric Chemistry and Physics 4, 391e411. Hussein, T., Karppinen, A., Kukkonen, A., Harkonen, J., Aalto, P.P., Hameri, K., Kerminen, V.-M., Kulmala, M., 2006. Meteorological dependence os sizefractionated number concentration of urban aerosol particle. Atmospheric Environment 40, 1427e1440. Jeong, C.H., Evans, G.J., McGuire, M.L., Chang, R.Y.W., Abbatt, J.P.D., Zeromskiene, K., Mozurkewich, M., Li, S.M., Leaitch, W.R., 2010. Particle formation and growth at five rural and urban sites. Atmospheric Chemistry and Physics 10, 7979e7995. Jung, J., Adams, P.J., Pandis, S.N., 2006. Simulating the size distribution and chemical composition of ultrafine particles during nucleation events. Atmospheric Environment 40, 2248e2259. Kalivitis, N., Birmili, W., Stock, M., Wehner, B., Massling, A., Wiedensohler, A., Gerasopoulos, E., Mihalopoulos, N., 2008. Particle size distributions in the

801

eastern Mediterranean troposphere. Atmospheric Chemistry and Physics 8, 6729e6738. Kopanakis, I., Eleftheriadis, K., Mihalopoulos, N., Lydakis-Simantiris, N., Katsivela, E., Pentari, D., Zarmpas, P., Lazaridis, M., 2012. Physico-chemical characteristics of particulate matter in the eastern Mediterranean. Atmospheric Research 106, 93e107. Kroll, J.H., Seinfeld, J.H., 2008. Review: chemistry of secondary organic aerosol: formation and evolution of low-volatility organics in the atmosphere. Atmospheric Environment 42, 3593e3624. Kulmala, M., Kerminen, V.-M., 2008. On the formation and growth of atmospheric nanoparticles. Atmospheric Research 90, 132e150. Kulmala, M., Pirjola, L., Mäkelä, J.M., 2000. Stable sulphate clusters as a source of new atmospheric particles. Nature 404, 66e69. Kulmala, M., Dal Maso, M., Mäkelä, J.M., Pirjola, L., Väkevä, M., Aalto, P., Miikkulainen, P., Hämeri, K., O’Dowd, C.D., 2001. On the formation, growth and composition of nucleation mode particles. Tellus B 53, 479e490. Kulmala, M., Vehkamaki, H., Petajä, T., Dal Maso, M., Lauri, A., Kerminen, V.M., Birmili, W., McMurry, P.H., 2004. Formation and growth rates of ultrafine atmospheric particles: a review of observations. Journal of Aerosol Science 35, 143e176. Kulmala, M., Petäjä, T., Mönkkönen, P., Koponen, I.K., Dal Maso, M., Aalto, P.P., Lehtinen, K.E.J., Kerminen, V.M., 2005. On the growth of nucleation mode particles: source rates of condensable vapor in polluted and clean environments. Atmospheric Chemistry and Physics 5, 409e416. Kulmala, M., Riipinen, I., Sipilä, M., Manninen, H., Petäjä, T., Junninen, H., Dal Maso, M., Mordas, G., Mirme, A., Vana, M., Hirsikko, A., Laakso, L., Harrison, R.M., Hanson, I., Leung, C., Lehtinen, K.E.J., Kerminen, V.M., 2007a. Towards direct measurement of atmospheric nucleation. Science 318, 89e92. Kulmala, M., Mordas, G., Petäjä, T., Gronholm, T., Aalto, P.P., Vehkamaki, H., Hienola, A.I., Herrmann, E., Sipilä, M., Riipinen, I., Manninen, H.E., Hameri, K., Stratmann, F., Bildec, M., Winklerd, P.M., Wagnerd, P.E., 2007b. The condensation particle counter battery (CPCB): a new tool to investigate the activation properties of nanoparticles. Journal of Aerosol Science 38, 289e304. Laakso, L., Mäkelä, J.M., Pirjola, L., Kulmala, M., 2002. Model studies on ion-induced nucleation in the atmosphere. Journal of Geophysical Research 107, 4427. http:// dx.doi.org/10.1029/2002JD002140. Lazaridis, M., 2001. New particle formation of ternary droplets in the atmosphere e a steady state nucleation kinetics approach. Atmospheric Environment 35, 599e607. Lazaridis, M., Kulmala, M., 1992. A theoretical description of the heterogeneous nucleation in binary systems. Journal of Aerosol Science 23, S137eS140. Lazaridis, M., Kulmala, M., Laaksonen, A., 1991. Binary heterogeneous nucleation of a water-sulphuric acid system: the effect of hydrate interaction. Journal of Aerosol Science 22, 823e830. Lazaridis, M., Broday, D.M., Hov, Ø., Georgopoulos, P.G., 2001. Integrated exposure and dose modeling and analysis system. 3. Deposition of inhaled particles in the human respiratory tract. Environmental Science Technology 35, 3727e3734. Lazaridis, M., Spyridaki, A., Solberg, S., Smolík, J., Zdímal, V., Eleftheriadis, K., Aleksanropoulou, V., Hov, O., Georgopoulos, P.G., 2005. Mesoscale modeling of combined aerosol and photo-oxidant processes in the eastern Mediterranean. Atmospheric Chemistry and Physics 5, 927e940. Lazaridis, M., Dzumbova, L., Kopanakis, I., Ondracek, J., Glytsos, T., Aleksandropoulou, V., Voulgarakis, A., Katsivela, E., Mihalopoulos, N., Eleftheriadis, K., 2008. PM10 and PM2.5 levels in the eastern Mediterranean (Akrotiri research station, Crete, Greece). Water Air & Soil Pollution 189, 85e101. Lee, S.H., Reeves, J.M., Wilson, J.C., Hunton, D.E., Viggiano, A.A., Miller, T.M., Ballenthin, J.O., Lait, L.R., 2003. Particle formation by ion nucleation in the upper troposphere and lower stratosphere. Science 301, 1886e1889. Lee, Y.G., Lee, H.W., Kim, M.S., Choi, Ch.Y., Kim, J., 2008. Characteristics of particle formation events in the coastal region of Korea in 2005. Atmospheric Environment 42, 3729e3739. Lehtipalo, K., Kulmala, M., Sipilä, M., Petäjä, T., Vana, X., Ceburnis, V., Dupuy, V., O’Dowd, C., 2009. Nanoparticles in boreal forest and coastal environment: a comparison of observations and implications of the nucleation mechanism. Atmospheric Chemistry and Physics 10, 7009e7016. Mejia, J.F., Morawska, L., Mengersen, K., 2008. Spatial variation in particle number size distributions in a large metropolitan area. Atmospheric Chemistry and Physics 8, 1127e1138.  Ondracek, J., Zdímal, V., Smolik, J., Lazaridis, M., 2009. A merging algorithm for aerosol size distribution from multiple instruments. Water Air & Soil Pollution 199, 219e233. Ortega, I.K., Suni, T., Boy, M., Grönholm, T., Manninen, H.E., Nieminen, T., Ehn, M., Junninen, H., Hakola, H., Hellen, H., Valmari, T., Arvela, H., Zegelin, S., Hughes, D., Kitchen, M., Cleugh, H., Worsnop, D.R., Kulmala, M., Kerminen, V.M., 2012. New insights into nocturnal nucleation. Atmospheric Chemistry and Physics 12, 4297e4312. Paasonen, P., Nieminen, T., Asmi, E., Manninen, H.E., Petaja, T., Plass-Dulmer, C., Flentje, H., Birmili, W., Wiedensohler, A., Horrak, U., Metzger, A., Hamed, A., Laaksonen, A., Facchini, M.C., Kerminen, V.M., Kulmala, M., 2010. On the roles of sulphuric acid and low-volatility organic vapours in the initial steps of atmospheric new particle Formation. Atmospheric Chemistry and Physics 10, 11223e 11242. Pikridas, M., Riipinen, I., Hildebrandt, L., Kostenidou, E., Manninen, H., Mihalopoulos, N., Kalivitis, N., Burkhart, J.F., Stohl, A., Kulmala, M., Pandis, S.N., 2012. New particle

802

I. Kopanakis et al. / Atmospheric Environment 77 (2013) 790e802

formation at a remote site in the eastern Mediterranean. Journal of Goephysical Research 117, D12205. http://dx.doi.org/10.1029/2012JD017570. Place Jr., P.F., Ziemba, L.D., Griffin, R.J., 2010. Observations of nucleation-mode particle events and size distributions at a rural New England site. Atmospheric Environment 44, 88e94. 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., Huglin, C., Jones, A.M., Kasper-Giebl, A., Kiss, G., Kousa, A., Kuhlbusch, T.A.J., Loschau, 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 e 3: physical and chemical characteristics of particulate matter from 60 rural, urban, and kerbside sites across Europe. Atmospheric Environment 44, 1308e1320. Qian, Shi, Sakurai, H., McMurry, P.H., 2007. Characteristics of regional nucleation events in urban East St. Louis. Atmospheric Environment 41, 4119e 4127. Reche, C., Querol, X., Alastuey, A., Viana, M., Pey, J., Moreno, T., Rodriguez, S., Gonzalez, Y., Fernandez-Camacho, R., Sanchez de la Campa, A.M., de la Rosa, J., Dall’Osto, M., Prevot, A.S.H., Hueglin, C., Harrison, R.M., Quincey, P., 2011. New considerations for PM, black carbon and particle number concentration for air quality monitoring across different European cities. Atmospheric Chemistry and Physics 11, 6207e6227. Rimnacova, D., Zdimal, V., Schwarz, J., Smolik, J., Rimnac, M., 2011. Atmospheric aerosols in suburb of Prague: the dynamics of particle size distributions. Atmospheric Research 101, 539e552. Salma, I., Borsós, T., Weidinger, T., Aalto, P., Hussein, T., Dal Maso, M., Kulmala, M., 2010. Production, growth and properties of ultrafine atmospheric aerosol particles in an urban environment. Atmospheric Chemistry and Physics 11, 1339e1353. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics: from Air Pollution to Climate Change. John Wiley, New York. Sipilä, M., Lehtipalo, K., Kulmala, M., Petäjä, T., Junninen, H., Aalto, P.P., Manninen, H.E., Kyrö, E.-M., Asmi, E., Riipinen, I., Curtius, J., Kurten, A., Borrmann, S., O’Dowd, C.D., 2008. Applicability of condensation particle counters to measure atmospheric clusters. Atmospheric Chemistry and Physics 8, 4049e4060.

Stanier, C.O., Khlystov, A.Y., Pandis, S.N., 2004. Ambient aerosol size distributions and number concentrations measured during the Pittsburgh air quality study (PAQS). Atmospheric Environment 38, 3275e3284. Sun, J., Ariya, P.A., 2006. Atmospheric organic and bio-aerosols as cloud condensation nuclei (CCN): a review. Atmospheric Environment 40, 795e820. Vehkamäki, H., Kulmala, M., Napari, I., Lehtinen, K.E.J., Timmreck, C., Noppel, M., Laaksonen, A., 2002. An improved parameterization for sulfuric acid e water nucleation rates for tropospheric and stratospheric conditions. Journal of Geophysical Research 107, 4622e4632. Vuollekoski, H., Nieminen, T., Paasonen, P., Sihto, S.L., Boy, M., Manninen, H., Lehtinen, K.E.J., Kerminen, V.M., Kulmala, M., 2010. Atmospheric nucleation and initial steps of particle growth: numerical comparison of different theories and hypotheses. Atmospheric Research 98, 229e236. Wichmann, H.E., Spix, C., Tuch, T., Wölke, G., Peters, A., Heinrich, J., Kreyling, W.G., Heyder, J., 2000. Daily mortality and fine and ultrafine particles in Erfurt, Germany part I: role of particle number and particle mass. Research Report Health Effects Institute 98, 5e86. Winkler, P.M., Steiner, G., Vrtala, A., Vehkamäki, H., Noppel, M., Lehtinen, K.E.J., Reischl, G.P., Wagner, P.E., Kulmala, M., 2008. Heterogeneous nucleation experiments bridging scale from molecular ion clusters to nanoparticles. Science 319, 1374e1377. Wu, Z., Hu, M., Liu, S., Wehner, B., Bauer, S., Massling, A., Wiedensohler, A., Petaja, T., Dal Maso, M., Kulmala, M., 2007. New particle formation in Beijing: statistical analysis of a 1-year data set. Journal of Geophysical Research 112, D09209. Wu, Z., Hu, M., Lin, P., Liu, Sh., Wehner, B., Wiedensohler, A., 2008. Particle number size distribution in the urban atmosphere of Beijing, China. Atmospheric Environment 42, 7967e7980. Yoon, Y.J., O’Dowd, C.D., Jennings, S.G., Lee, S.H., 2006. Statistical characteristics and predictability of particle formation events at mace head. Journal of Geophysical Research 111, D13204. http://dx.doi.org/10.1029/2005JD006284. Zhang, R.I., Suh, I., Zhao, J., Zhang, D., Fortner, E.C., Tie, X., Molina, L.T., Molina, M.J., 2004. Atmospheric new particle formation enhanced by organic acids. Science 304, 1487e1490.  Zdímal, V., Smolík, J., Eleftheriadis, K., Wagner, Z., Housiadas, C., Mihalopoulos, N., Mikuska, P., Ve cera, Z., Kopanakis, I., Lazaridis, M., 2010. Dynamics of atmospheric aerosol number size distributions in the eastern Mediterranean during the “SUB-AERO” Project. Water Air & Soil Pollution 214, 133e146.