Chemical Composition of Aerosol Size Fractions at a Coastal Site in Southwestern Italy: Seasonal Variability and Transport Influence

ABSTRACT

The present study focuses on the elemental characterization of fine and coarse particles collected at a coastal site of southwestern Italy, in a suburban area of the Calabria region. A chemical tracer analysis was carried out to identify the major emission sources influencing on the atmospheric aerosol levels. Size-resolved particulate samples were collected during three 2-week seasonal sampling campaigns: autumn (19 October to 2 November 2003), winter (19 January to 2 February 2004) and spring (26 April to 10 May 2004). Ambient concentrations of selected elements (Fe, Mn, Mg, Ca, V, Cu, Cr, Ni, Zn, Pb, and Cd) associated to fine and coarse size fractions were determined using atomic absorption spectrometry (AAS). The enrichment factor method was applied, suggesting a prevailing anthropogenic component for all the detected elements, with Fe, Mg, Mn, and Ca as exceptions. Trajectory sector analysis was used in order to discriminate the influence of different air mass origins and paths. Long-range transport from both the continental Europe and the Saharan region proved to be the main influencing factors. African dust outbreaks, whose occurrence frequency was greater during the autumn and spring seasonal monitoring periods, gave rise to a total of eight exceedances of the European Commission (EC) PM₁₀ daily limit value as well as an increase in values of the crustal-derived elements (Fe, Mg, and Ca). Long-range transport from the heavily industrialized area of Central/Eastern Europe contributed to the high levels of Zn, Cd, and Pb that were recorded during the winter sampling campaign. Seasonal trend and comparison with measurements previously performed across the Mediterranean basin were also presented and discussed.

IMPLICATIONS

This paper, which investigates both natural and anthropogenic aerosol sources, contributes to the characterization of fine and coarse particles in a suburban area of the southern Italian region of Calabria whose air quality assessment is still being defined. The monitoring site involved in this study, located at the interface between continental Europe and the Mediterranean basin, offers an important platform to study the influence of seasonal meteorology and predominant circulation patterns on the aerosol concentration levels and their chemical composition.

INTRODUCTION

Particulate matter (PM) plays a key role in creating health hazards,1–3 influencing air quality and climate change.4–8 The detailed knowledge of the aerosol physicochemical properties is becoming increasingly important because size and composition characterization of atmospheric particles is relevant to atmospheric process modeling9 and for environmental control purposes.10 Due to its wide implications, PM constitutes a growing challenge even though it is a complex subject of study. The understanding of its atmospheric evolution is indeed made difficult by the wide number of emission sources and by the numerous physical and chemical processes that are correlated to local, mesoscale, and synoptic meteorological conditions. In recent years, a growing number of studies detailing coarse and fine fractions of PM have been produced, although there is still a lack of comparative data among sampling sites. In particular, observation gaps are related to aerosol chemical composition and size distribution,11 with particular regard for the southern Italian regions whose elaboration of effective Regional Air Quality Management Plans (AQMPs) becomes thus more difficult. AQMPs represent a fundamental phase of the entire air quality process that has been assigned to the 20 Italian administrative regions, where measures should be defined and adopted in order to meet established air quality targets.12 The new Directive 2008/50/EC,13 in assessing compliance with air quality limit values, gives the possibility of skipping exceedances due to natural sources with regard to the amount of their contribution. In addition, as already observed in Escudero et al.,14 long-range transport of PM from the heavily industrialized area of Central/Eastern Europe can further contribute to PM₁₀ exceedances. Over the last decade, the estimate of natural dust and continental Europe's contribution to PM₁₀ concentrations has become a key issue in air quality assessment and policy formulation.15 To that purpose, PM₁₀ and PM_2.5 daily average ponderal values, alone, are not sufficient to give significant information on source apportionment analysis.16 More information can be obtained by studying the correlation between the coarse and fine PM modes and their temporal evolution. Chemical composition also represents a key tool for understanding the origin of particles, anthropogenic and/or natural, and it reveals the atmospheric processes that are involved.17–21 The aim of this study is to investigate the relationships between coarse and fine mass concentrations, and to identify the trace elements contained therein during three 2-week seasonal field campaigns at a suburban site on the southern Tyrrhenian coast of Italy. The site is located at the boundary between continental Europe and the Mediterranean basin, where, due to the variety of the regions surrounding the basin itself, different classes of particles can be found.22,23 In this context, we identified the relevant PM episodes and processes that gave rise to daily PM levels. In order to investigate the sources affecting the aerosol concentration and composition, we compared these results with both backward trajectories and satellite derived products.14,16 Specifically, we used trajectory sector analysis to study the influence of long-range transport from both industrial areas of Central/Eastern Europe and the arid areas of the Sahara.

MATERIALS AND METHODOLOGIES

Monitoring Site

The sampling site (39°25′N; 16°44′E) is located in a small village on the coast of Tyrrhenian Sea, about 35 km from Cosenza, the main urban site in the southern Italian province of Calabria (Figure 1). According to Decision 2001/752/EC and the “Criteria for EUROAIRNET” document,24 this air quality monitoring site has been classified as suburban. The surrounding area is characterized by several activities, including open construction areas, structural steel manufacture, and the production of plastics. Prevailing land uses are agriculture and woodland. There are few commercial activities, and the number of residential units is not significant. A railway crosses this zone; the highway (SS 18) is about 300 m away from the sampling site, with an appreciable traffic flow (about 4000 vehicals per day)25 and a moderate commercial traffic consisting mostly of diesel-powered vehicles (trucks).

Figure 1. Map of the sampling area showing the monitoring site location.

Sampling and Analytical Procedures

Fine and coarse measurements were performed during three 15-day sampling periods spread over three seasons: autumn (19 October to 2 November 2003), winter (19 January to 2 February 2004) and spring (26 April to 10 May 2004). In these samples, meteorological parameters (i.e., temperature, relative humidity, and wind velocity and direction) were recorded at regular intervals. Fine and coarse size fractions were collected by a manual Andersen dichotomous sampler (model 241) on 37-mm Teflon filters over a 24-hr sampling period at flow rates of 1.67 L∙min⁻ 1 and 15 L∙min⁻ 1, respectively, for a total operational flow rate of 16.7 L∙min⁻ 1. Prior to sampling, filters were conditioned for 48 hr at 25°C and 50% relative humidity. Filters were preweighed using a 0.1-μg-sensitivity microbalance. Internal calibration of the balance was performed at the beginning of each weighing session, following standard procedures.26 Samples and blank filters were kept in polyethylene bags before and after sampling until analysis. After sampling, fine and coarse filters were digested with a mixture of 5 mL HNO₃ 65% and 2 mL H₂O₂ 5%, following the hot digestion method (200 Series EPA Procedures, Metals Analysis). The resulting solutions were brought to 25 mL volume with Milli-Q water. Digested samples were analyzed using a GBC 932 Plus model atomic absorption spectrometer (AAS). Fe, Mn, Mg, Ca, V, Cu, Cr, Ni, Zn, Pb, and Cd concentrations were measured in all samples. Mg, Ca, and Zn concentrations were analyzed using flame atomic absorption spectroscopy (FAAS), whereas Fe, Mn, V, Cu, Cr, Ni, Pb, and Cd concentrations were determined by graphite furnace atomic absorption spectroscopy (GFAAS) by coupling the GBC 932 Plus AAS to a GBC GF 3000 electrothermal atomization system. A parallel analysis of blanks was also measured under similar experimental conditions to quantify possible contamination. We computed the limit of detection (LOD) and the limit of quantification (LOQ) following the International Union of Pure and Applied Chemistry (IUPAC) protocol.27

Data Analysis and Treatment

The recorded data set consists of fine and coarse daily mass concentrations (μg·m⁻ 3), daily atmospheric concentrations of Fe, Mn, Mg, Ca, V, Cu, Cr, Ni, Zn, Pb, and Cd (ng·m⁻ 3), and daily meteorological parameters (temperature [T; °C], relative humidity [RH; %], precipitation [mm], daily mean wind vector, component wind speeds [WS; m∙s⁻ 1], and wind direction [ϕ; deg]). For each measured variable, explorative statistical parameters were performed and then compared with previous studies, both with respect to sampling location and results. To get an indication of the relative contribution of crustal and noncrustal sources, the enrichment factor method28–31 was applied. The enrichment factors (EFs) were calculated for each detected element, considering Fe as the reference element and using the crustal composition given by Greenwood and Earnshaw.32 In addition, attempts were made to verify long-range transport contributions from different regions to the aerosol burden with qualitative considerations based on trajectory sector analysis.14,16 To this end, 3-day backward air trajectories at different altitudes (500, 1000, and 1500 m above sea level [a.s.l.]) were calculated for each sampling date with our monitoring site as the starting point. This backward-trajectory analysis was carried out using the Hybrid Single-Particle Lagrangian Integrated Trajectory Model,33 available at the National Oceanic and Atmospheric Administration (NOAA) Air Resources Laboratory (ARL).34 Trajectories ending at 500 m were indicative of circulation in the lowest troposphere within the boundary layer. Trajectories ending at 1500 and 1000 m were used to identify aerosols originating from North Africa and other long-range air mass transport events. This assessment of long-range air mass was also supported by the aerosol maps created by the Navy Aerosol Analysis and Prediction System (NAAPS) and provided by the Naval Research Laboratory (NRL) in Monterey (CA, USA).35 In particular, the sulfate and dust concentration maps were used to confirm episodes of transboundary air pollution transport from continental European and North Africa, respectively.14,16 Our interpretation of the temporal evolution of fine and coarse PM fractions was complemented with the inspection of the National Centers for Environmental Prediction (NCEP) based 700 mbar level geopotential height (m) (http://www.cdc.noaa.gov/). All the NCEP maps include the area 10°–60°N and 30°W–40°E.

RESULTS AND DISCUSSION

Explorative Statistical Analysis

Fine and Coarse Mass Concentrations

Daily values of fine and coarse concentrations are reported in Figure 2. PM coarse mass concentrations, which significantly increased during the spring sampling campaign, showed a larger variability compared to that observed in the fine PM levels. We observed some exeedances of the daily European Commission (EC) PM₁₀ limit value (50 μg·m⁻ 3)13: three during the autumn campaign, and five during the spring campaign. A peak value of 114.4 μg·m⁻ 3 was recorded on 7 May 2004. Descriptive statistics for fine and coarse PM size fractions (as well for meteorological parameters) are summarized in Table 1. The lowest values of both fine and coarse mean levels were recorded in the winter sampling period, whereas the highest were in autumn, for fine particulate concentrations, and in spring for coarse particulate concentrations. The PM seasonal trend that we observed was different from that generally observed in Northern Italy.36–38 This is due to a greater occurrence of natural aerosol sources in southern Italy during the spring-summer seasons.39,40 As reported in recent studies,41–44 the highest PM₁₀ values recorded over the Mediterranean area occur during the warmer seasons, and are generally correlated with anticyclonic conditions45–50 that favor the re-buildup of particles, particularly the advection of Saharan dust.51,52 In the winter season, the typical meteorological conditions in the central Mediterranean basin are influenced by the presence of cyclones, the remixing of air masses and the reinforcement of winds.53 These conditions give rise to a heightened aerosol dispersion and lower PM₁₀ values. The synoptic chart of NCEP-based 700 mbar geopotential heights (m) for our sampling periods showed consistent results with the prevailing pressure systems described above, leading to a decrease in both fine and coarse PM levels during the winter field campaign with respect to those recorded during the spring and autumn sampling periods. Also, the local weather conditions were equally important in affecting aerosol content at our sampling site. According to a previous study, the intensity of mountain valley and sea breeze circulations in Calabria are comparable to each other, and as such the observed local mesoscale flow is driven by both of these factors.54 A significant negative correlation with wind speed, either as average or maximum value, was found giving prominence to the efficiency of the atmospheric horizontal mixing as a dilution mechanism. Average daily precipitation was higher in autumn and spring (maximum values = 22.0 and 9.9 mm·day⁻ 1, respectively) than in winter (maximum value = 1.0 mm·day⁻ 1). This meteorological occurrence might influence the coarse/fine distribution. Indeed, an increase in precipitation might yield an increase of the washout ratio for the coarse fraction and therefore a decrease of its ambient concentration.55 Our analysis showed that the lowest average coarse contribution (∼68%) to PM₁₀ levels, was recorded during the first sampling campaign (autumn 2003) when the precipitation events were more intense than those observed during the second (winter 2004) and third (spring 2004). Other meteorological parameters did not show any relevant relationship to atmospheric particulate levels. A comparison with data in existing literature that refers to the same sampling period56 showed that fine particle mass concentrations observed in this study are in agreement with those recorded at different rural Mediterranean sites during each season. This was in contrast to the coarse PM levels, which were higher during the autumn and spring periods because of greater soil erosion in our study area.

Figure 2. Fine and coarse mass concentrations during the three 15-day sampling periods.

Table 1. Statistical summary for both fine and coarse mass concentrations and meteorological parameters

	Autumn				Winter				Spring
	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.
Fine (μg m^−3)	10.9	5.4	2.8	22.4	6.3	3.9	0.9	17.3	7.5	3.7	2.8	15.3
Coarse (μg m^−3)	23.6	9.8	9.7	44.8	17.4	7.8	8.0	29.8	34.7	31.9	3.3	99.8
Temperature (°C)	19.6	1.9	16.6	23.1	8.7	3.6	2.9	14.8	16.5	2.0	13.9	21.0
Wind speed (m s^−1)	4.6	3.0	1.8	13.1	5.6	3.5	2.1	11.4	4.1	2.1	2.6	9.9
Relative humidity (%)	74.5	5.6	65.3	83.3	64.4	15.1	40.1	91.4	70.2	9.2	52.9	83.2
Precipitation (mm)	3.5	6.0	0.0	22.0	0.2	0.3	0.0	1.0	4.1	2.1	2.6	9.9

Elemental Concentration

A statistical summary concerning both transition and heavy metals (Fe, Mn, Mg, Ca, V, Cu, Cr, Ni, Zn, Pb, and Cd ) associated with both fine and coarse PM fractions is reported in Table 2. The absolute concentration of various elements was highly differentiated, ranging from few ng·m⁻ 3 up to some μg·m⁻ 3 as a result of the combined effects of climatic factors and source strength. The elements showed different distributions between the fine and coarse PM fractions. Regarding mass closure, elemental concentrations in coarse fraction amounted to 9.1–13.2%, whereas in fine fraction they had a percentage of 18.4%, 20.8%, and 53.5% in the autumn, winter, and spring sampling campaigns, respectively. Enrichment factor (EF) average values were calculated for each element associated with both fine and coarse PM fractions and the related results are reported in Table 3. Mn, Mg, and Ca were characterized by low EF, confirming their crustal origin. A strong anthropogenic component was found for elements such as Cd, Zn, and Pb, whose high EFs (see Table 3) were comparable to those already reported in Ragosta et al.57 and in Heimbürger et al.58 As Table 3 shows, the EF average values for Ni, Cr, Cu, and V indicate they were enriched, but to a lesser extent. Overall, crustal elements such as Fe, Mg, and Ca revealed seasonal patterns, with the highest concentrations in spring and the lowest in winter. In particular, Ca levels were exceptionally high compared to those observed at other locations in southern Italy.40 This was due to erosion from cultivated agricultural land and from the sandy soil surrounding our sampling site. Conversely, as already observed in both the eastern and northwestern Mediterranean basin,58,59 anthropogenic trace metals did not show a clear seasonal pattern. Specifically, as Table 2 summarizes, elements with a prevalent anthropogenic origin (Ni, Pb, Cd, V, and Zn) were abundant in the coarse fraction during the winter period. Regarding fine fraction, Ni levels were highest in autumn, whereas Cd, V, and Zn were highest in springtime. Even though Mn had low EFs, it reached its maximum values during the winter period, thus behaving differently than other crustal-derived elements in both fine and coarse fractions. The V, Cu, Fe, and Mn results were comparable with those obtained in other European and Italian background/suburban sites.56,57,60,61 In addition, the range of Cd, Cr, Ni, Pb, and Zn levels were in agreement with other urban and/or urban roadside sites.56,57,62 These findings may be explained by specific activities in this particular area, such as structural steel manufacturing (Ni and Cr) and by the presence of a local road (SS 18), with exhaust emissions from both gasoline and diesel fuelled vehicles (mainly characterized by Pb, Zn, and Cd).63 Ongoing building activities and degradation of building elements can be considered as additional sources leading to a further increases in Cd, Cr, Ni, Pb, and Zn levels with respect to rural and suburban background values. The contribution of long-range transport of air pollutants should be also taken into account. A trajectory sector data analysis will take this into account in the next section.

Table 2. Statistical summary for fine and coarse elemental concentrations

	Autumn				Winter				Spring
Fine (ng m^−3)	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.
Fe	257.0	140.3	72.5	668.2	26.7	24.6	1.4	86.5	220.8	294.4	34.0	1171.5
Mn	4.2	2.1	1.2	7.8	4.9	3.3	0.3	10.6	3.2	4.4	0.0	18.5
Mg	110.2	67.9	8.3	230.2	134.2	78.8	44.8	311.1	186.9	163.9	24.9	535.6
Ca	1029.6	896.0	186.0	3285.7	607.6	637.0	62.7	2577.9	2617.6	1072.4	1138.8	5385.7
V	4.8	3.1	0.9	9.5	4.1	3.7	0.0	11.4	13.9	8.6	0.8	39.3
Cu	6.2	1.8	3.0	9.3	2.5	1.6	0.3	6.5	6.3	3.3	3.1	17.0
Cr	35.9	11.6	13.9	55.7	5.5	3.4	0.6	10.7	11.6	7.0	3.0	28.4
Ni	30.4	10.1	15.9	51.0	18.5	8.5	8.0	36.1	27.8	21.6	4.9	97.7
Zn	191.8	191.5	16.6	769.6	188.1	135.9	51.4	537.2	341.5	157.0	117.5	621.5
Pb	44.9	14.3	24.5	67.2	42.4	27.3	4.9	96.5	40.1	13.4	14.8	65.6
Cd	2.8	1.0	1.2	4.2	2.3	1.7	0.5	6.6	3.2	0.9	1.6	5.2
Coarse (ng m^−3)	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.	Mean	SD	Min.	Max.
Fe	250.8	104.1	84.0	473.7	138.3	72.8	10.2	247.4	177.9	107.2	71.3	472.4
Mn	3.8	1.6	1.1	6.7	31.7	15.4	4.3	56.6	2.5	1.9	0.0	7.5
Mg	516.1	193.9	241.6	871.0	441.3	258.9	147.3	1022.6	821.5	821.3	82.3	2652.4
Ca	929.3	279.1	565.6	1532.0	827.4	594.6	195.3	2613.8	1176.9	711.5	339.9	2310.6
V	3.6	2.8	0.7	9.1	9.0	4.7	0.7	16.0	3.4	2.2	0.0	9.2
Cu	5.1	1.8	2.3	9.1	3.9	1.9	0.4	7.9	5.7	1.4	3.3	8.6
Cr	19.7	9.5	3.3	35.9	7.5	4.6	0.8	14.4	8.9	5.3	1.1	20.5
Ni	22.5	12.0	2.6	54.6	24.3	9.6	12.6	44.1	14.9	11.2	0.3	42.9
Zn	114.8	126.6	36.0	551.3	240.7	138.3	42.7	567.4	104.4	86.2	21.4	286.0
Pb	45.5	21.6	5.1	81.0	54.7	24.6	21.1	103.6	35.6	19.8	0.0	72.8
Cd	2.6	1.2	0.9	4.5	4.8	1.9	2.0	8.6	2.6	0.9	1.5	5.3

Table 3. Enrichment factor average values for both fine and coarse PM fractions

	Autumn		Winter		Spring
	Fine	Coarse	Fine	Coarse	Fine	Coarse
Fe	1.0	1.0	1.0	1.0	1.0	1.0
Mn	1.4	1.1	1.5	9.6	1.1	0.7
Mg	1.1	5.8	1.5	4.9	2.2	9.5
Ca	7.0	5.5	3.6	4.8	19.3	7.1
V	6.9	3.2	6.6	17.6	30.4	1.9
Cu	25.1	17.2	6.0	12.3	72.3	24.8
Cr	106.8	51.4	15.3	18.8	34.3	23.3
Ni	95.5	58.7	56.5	69.5	95.0	41.9
Zn	906.4	474.3	876.7	1007.3	1631.4	428.2
Pb	1170.2	1056.2	1104.5	1294.5	1082.9	851.6
Cd	6141.7	5089.7	5038.2	9523.9	6378.0	4397.1

Trajectory Sector Data Analysis

To discriminate amongst the influence of different aerosol sources, a trajectory sector data analysis was performed on the prevailing air mass origin. Three-day backward trajectories were calculated for each sampling date, with our monitoring site as the starting point. Four broad geographical sectors, displayed in Figure 3, were defined in relation to different aerosol sources. These sectors match previous studies performed over the Mediterranean basin.22 The identified sectors are

Figure 3. Sectors for the origin of air masses: (1) West Mediterranean (WMED); (2) North Africa (NAF); (3) Central–Eastern Europe (CEU); and (4) Atlantic advection (NATL).

1.	WMED—Western MEDiterranean basin: air masses originating over West Mediterranean.
2.	NAF—North AFrica: air masses that were loaded with Saharan dust over North Africa prior to arriving at the sampling site.
3.	CEU—Continental EUrope: air masses crossing the European highly industrialized and densely populated area (Central–Eastern Europe).
4.	NATL—North ATLantic: air masses having a 3-day origin over the North Atlantic at altitudes above 3000 m a.s.l. that descended lower before reaching the measurement site.

The number of samples assigned to sectors 1–4 are reported in Figure 4, whereas Table 4 reports the seasonal occurrence frequency of the identified sectors. As Table 4 shows, the influence of the Saharan dust, with 46.7% and 40.0% of the total occurrence frequency, was quite higher in the autumn and spring seasons, respectively. The highest Continental Europe (CEU) occurrence, with a value of 66.7%, was recorded in the winter campaign during which the Saharan outbreak influence was otherwise less frequent (6.7%). Both WMED and NATL occurrences, with the same values ranging from 13.3% and 26.7%, were quite low in all seasonal sampling campaigns. Mean levels of chemical elements related to the total PM₁₀ mass fraction for each considered sector were also calculated and compared with the trajectory sector profiles reported in previous studies.64,65 In order to identify and separate the anthropogenic contribution originating from Central and Eastern Europe (CEU influence) from that coming from North Africa (NAF influence), we computed the mean percentage contributions of chemical elements for each identified sector. In this way, the larger contribution by NAF and CEU sectors for all the detected trace elements has been highlighted (see Table 5). An in-depth investigation of these two influences is discussed in the following sections.

Figure 4. Assignment of each sampling day to the main identified origin sectors: WMED, Mediterranean advection; NAF, African dust outbreaks; CEU, European air masses transport; NATL, Atlantic advection.

Table 4. Occurrence frequencies of the main sectors where air masses originated prior to ending at our sampling site

	Autumn	Winter	Spring
WMED	26.7%	0.0%	20.0%
NAF	46.7%	6.7%	40.0%
CEU	13.3%	66.7%	26.7%
NATL	13.3%	26.7%	13.3%

Table 5. Mean percentage contributions of chemical elements related to the total PM₁₀ mass fraction for each considered sector

	WMED	NATL	CEU	NAF
PM10	19.0%	21.2%	15.8%	43.9%
Ni	26.2%	21.8%	23.3%	28.2%
Pb	27.3%	23.9%	29.5%	19.3%
Cr	21.8%	21.3%	27.3%	29.6%
Cd	22.0%	25.0%	27.6%	25.5%
Fe	27.6%	15.5%	12.3%	44.4%
Mn	11.5%	26.3%	35.7%	26.8%
V	25.1%	21.2%	21.2%	31.9%
Cu	26.2%	22.1%	20.6%	31.0%
Zn	24.3%	16.0%	33.7%	26.4%
Ca	30.4%	13.6%	19.2%	37.3%
Mg	17.6%	17.7%	16.9%	48.2%

Saharan Dust Outbreaks

Based on trajectory sector analysis, the presence and traceability of Saharan dust contribution to the atmospheric aerosol burden in southern Italy has been searched for. As reported in Figure 4, NAF episodes were recorded during autumn, for periods of 4 (from 19 to 22 October 2003) and 3 (from 31 October to 2 November 2003) days, during winter on the 19 January (2004) and in spring for periods of 6 consecutive days (from 2 to 7 May 2004). Typically crust-derived elements such as Fe, Mg, and Ca were considered as tracers of Saharan dust and their temporal series reported in Figure 5a. PM₁₀ levels are also overlapped in the same figure, highlighting exceedances over the EC PM₁₀ daily limit value.13 Even if Saharan episodes generally occur in spring and summer,53 the appearance of significant episodes at other times of the year is not exceptional (they are already observed across the Mediterranean basin).66 Therefore, it is not surprising to observe peaks of crustal trace elements not only in spring but also in autumn. All the PM₁₀ exceedances recorded at our sampling site appeared in conjunction with the above identified NAF episodes. These findings have been corroborated by both 3-day HYSPLIT backward trajectories and NAAPS-based dust concentration maps, as reported in Figure 5b and c for two selected days (22 October 2003 and 5 May 2004). Furthermore, as Figure 5a shows, throughout the identified NAF episodes, both the PM₁₀ mass concentrations and the typical crustal element (Fe, Mg, and Ca) levels increased, and were exceptionally high from the 4 to 7 May. In this last period, according to the pressure scenario described in Meloni et al.,53 a well-defined low over Great Britain appeared concurrently with a high pressure above the Sahara (as is typical in the springtime), causing air masses heavy with dust. In addition to the general pattern common to all crustal elements described above (Fe, Mg, and Ca), intermediate trace metals such as V, Cu, Cr, and Ni had a greater contribution during the identified NAF episode (see Table 5). On the other hand, the anthropogenic component of Ni and V concentrations in marine environments is known to be substantial. In particular, oil combustion from shipping emissions is responsible for spreading significant amounts of Ni and V.67–69 As observed over the Tyrrhenian Sea,70 Saharan material also contains significant amounts of Ni and V. The relatively low enrichment factor for Ni and V obtained in this study might thus suggest that a combination of Saharan crustal inputs and anthropogenic emissions is the prevailing source of Ni and V for the Tyrrhenian coast of Calabria. In addition, it is known that Libya is the major emitter of Cu in the Mediterranean basin71 and Cu emissions might thus be associated to southerly crustal inputs, such as Saharan dust events. Nevertheless, Saharan inputs are also likely to bring relatively high amounts of anthropogenic material to the northwestern Mediterranean basin.72 This may partly explain how, in this data set, relatively high concentrations of trace metals with a recognized anthropogenic character (such as V, Cu, Cr, and Ni) were sometimes associated with Saharan dust events.

Figure 5. Saharan dust outbreaks occurred over southern Italy. (a) Crustal elements (Fe, Ca, and Mg) and PM₁₀ time series. As an example, NAAPS-based surface dust concentration maps and 3-days HYSPLIT backward trajectories are also reported for (b) 22 October 2003 and (c) 5 May 2004.

Anthropogenic European Events

European episodes that occurred over the Calabria region during well-defined periods are reported in Figure 4. Figure 6a displays the days with an identified CEU influence (grey bars) and the time series for elements whose enrichment factors were the highest, such as Zn, Pb, and Cd. High episodic peaks (for Zn, Cd, and Pb) in the time series diagram may be ascribed to long-range transport processes. As Table 5 shows, the CEU influence, with values of 29.3% for Pb, 27.6% for Cd, and 33.7% for Zn, was the highest within the other identified sectors. This conclusion is also supported by the NAAPS-based sulfate maps. Backward air-mass trajectories were computed for all the sampling dates with episodic peaks (highlighted by the grey bars in Figure 6a) and these suggested that air quality during those days was mainly influenced by the air masses arriving at the sampling site through sector 3 (covering the northeast direction). For an example of CEU influence, Figure 6b shows both the NAAPS-based sulfate map and the backward trajectory relating to 25 January 2004. With regard to seasonality, the European events were more frequent in winter than in autumn, and no events occurred in spring (see Table 4). This result is consistent with other studies that showed the influence of European pollutants in Spain.73,74 As reported in the literature, an anthropogenic event, generally, involves a larger increase in PM fine fraction.14 At our sampling site, during the identified European events, we observed that PM coarse fraction levels, in relation to fine fraction, ranged from 51.0% (26 January) to 96.4% (28 January). This apparent contradiction, as is observed in other coastal European sites,14 may be explained by considering the location of the sampling site close to the Tyrrhenian sea. Winds associated with the European events may indeed increase sea-surface movement, and consequently increase the production of sea spray aerosols, which are generally in coarse size. These winds may also affect the local soil suspension process, which is an additional source of course particles.1

Figure 6. Anthropogenic European events occurred over southern Italy. (a) Anthropogenic elements (Zn, Pb, and Cd) time series, with grey bars highlighting days with CEU influence. As an example, NAAPS-based surface sulfate concentration maps and 3-days HYSPIT backward trajectories are also reported for 25 January 2004 (b).

CONCLUSION AND RESEARCH NEEDS

Based on both the enrichment factor and trajectory sector analysis, three groups of elements within our data set were associated with different aerosol sources: crustal-derived elements (Fe, Ca, and Mg), trace metals of anthropogenic origin (Pb, Cd, and Zn) and a third intermediate group of trace metals that presented both anthropogenic and natural/crustal influences (Ni, Cu, Mn, V, and Cr). High EFs observed in this study for anthropogenic trace metals stressed the importance of anthropogenic sources on Pb, Cd, and Zn concentrations in the western Mediterranean atmospheric aerosols, consistent with previous studies. For Ni, Ci, Mn, V, and Cr, we found that although these trace metals are classified as anthropogenic, they have also a natural component because they are present in trace amounts of lithogenic material. In terms of total PM₁₀ size fraction, Saharan dust outbreaks appeared to be the main factor in exceeding the EC PM₁₀ daily target. During the whole sampling period, eight exceedances over the NAF occurrence were recorded. Regarding seasonal trends, the highest concentrations of anthropogenic trace metals occurred mostly in autumn/winter, whereas crustal-derived elements were more concentrated in the springtime. A particularly intense transport event from North Africa occurred in May, yielding high concentrations (about 100 μg·m⁻ 3) of PM₁₀. Long-range transport events from continental Europe were the other main contributors to the high Zn, Cd, and Pb levels recorded during the winter sampling campaign. The findings presented here highlight the need of more measurements in order to gain a more comprehensive evaluation of the trace metal cycling in the Mediterranean basin. Results presented in this work should be considered as an exploratory study that suggests the need for additional work to be performed. In particular, an in-depth chemical analysis including both inorganic and organic compounds of aerosol, coupled with long-term measurements, is necessary in order to better identify the presence of sources affecting PM levels at this sampling site. The new coastal site of the National Research Council of Italy (CNR) within the European Monitoring and Evaluation Programme (EMEP) (more information can be found at http://www.iia.cnr.it/rende/), located at about 30 km south of the sampling site used in this work, will allow for additional measurements on key species, and it will operate on a yearly basis. Aerosol speciation, specifically, will include tracers such as EC/OC, Si/Al, and Na/Cl for a better identification of biomass burning, crustal material, and sea-spray source, respectively. Yet, to the best of our knowledge, this paper presents the first results on airborne particulate matter characterisation over a coastal area in southwestern Italy, highlighting the deficiency of point measurements in the region of Calabria whose air quality assessment is still being defined.75 Therefore, this paper, along with other results currently being prepared for publication, offers an important contribution.

ACKNOWLEDGMENTS

The authors acknowledge the NOAA Air Resources Laboratory (ARL) for providing of the HYSPLIT transport and dispersion model, and of the READY website (http://www.arl.noaa.gov/ready.html), whose results are used in this publication. The NCEP-based images are provided by the NOAA CIRES Climate Diagnostics Center, Boulder, Colorado, USA, from their Web site at http://www.cdc.noaa.gov/. The authors would like to express their gratitude also to the NASA/Goddard Space Flight Center, NOAA Air Resources Laboratory (ARL) and to the Naval Research Laboratory for the provision of the TOMS-AI maps and the NAAPS aerosol maps, respectively.