Air pollution is an important public health problem, as environmental epidemiology studies have consistently shown. Airborne particulate matter (PM) is currently one of the principal air pollutant health concerns in urban areas. Several epidemiological studies have found consistent associations between exposure to ambient particulate matter and adverse health effects, including increased mortality, hospitalization for respiratory or cardiovascular disease, and respiratory symptoms and decreased lung function1–3. Metal exposure, related to industrial emissions or traffic, is associated with heart rate variability4, increase in hospital emergency visits for respiratory conditions5 and increased risk of lung cancer after long term exposure6.

Particulate matter is emitted by a wide variety of natural and anthropogenic sources which influence its physical properties (e.g. size, specific surface, density or number-density) and chemical composition7. The relative proportions of the different components and the impact on human health are highly variable according to the type of sampling station8,9. For example, air in industrial and metropolitan areas is more contaminated with heavy metals than air in rural areas10,11. According to Nerriere et al.12, metal traces vary in relation to particle size (PM2.5 or PM10). Elements emitted through combustion processes (Ni, V, Pb) are present in higher proportions in PM2.5 than in PM1013. The main source of particle emissions in urban environments is usually traffic, although in highly industrial areas, traffic pollution may be secondary, depending on the composition of the particles emitted by the industrial activity14–16. Seasonal variability can also be observed in PM metal composition due to fluctuations in emission source17.

Particulate matter with aerodynamic diameter of less than 2.5mm (PM2.5) has been proven to be more harmful than PM10, given the greater ease with which it may penetrate the alveoli18,19. Time-series epidemiological studies have highlighted that exposure to PM2.5 leads to an increase in morbidity and mortality3. As a consequence, health protection legislation has begun to include smaller particles, an example of this being European Directive 2008/50/EC, which sets a limit of 25mg/m3 for PM2.5 over the course of one year, with target compliance by 2015.

The objectives of this study were: 1) to describe seasonal variation in PM2.5 levels in the area of study; 2) to characterize the trace elements composition (Cd, As, Mn, Cu, Cr, Ni, Fe, Pb, Hg and Zn; 3) to analyze the relationship between PM2.5 and PM10; and 4) to compare the trace elements in the PM2.5 and PM10 fractions.

Table 2 shows the mean PM2.5 and trace element concentrations recorded in each monitoring site over the two years of the study. The highest PM2.5 concentrations were recorded in the towns in the Urola Medio Valley during 2006. The limit value set by the European Community (25μg/m3) was exceeded only by Azkoitia in 2006. In general, a fall in PM2.5 concentrations was observed between 2006 and 2007, which was significant in the towns of Azkoitia (28.9μg/m3 in 2006 vs. 12.2μg/m3 in 2007, a 58% decrease), Azpeitia (23.4μg/m3 in 2006 vs. 13.6μg/m3 in 2007, a 42% decrease) and Legazpi (17μg/m3 in 2006 vs. 12.5μg/m3 in 2007, a 26% decrease). None of the trace elements exceeded the guideline value. However, the mean Mn concentrations found in Urola Medio, in 2006, and in the Oria Valley, in 2007, are comparable to those found in industrial and not urban environments.

On comparing the mean levels of trace elements in 2006 and 2007 for each valley, a significant decrease was observed (p<0.05) in the concentrations of all trace elements in the Urola Medio Valley and a significant increase (p<0.05) in Mn, Pb, As and Fe concentrations in the Oria Valley. In Urola Alto Valley, Ni was the only metal to display a significant decrease (p<0.05) between 2006 and 2007, in both towns.

Table 3 shows the seasonal variation per year in the main pollutant levels for all towns. Significant differences (p<0.05) were observed between the seasonal means for all pollutants; the highest concentrations in 2006 were registered in autumn except for PM2.5 with the highest levels in winter and in 2007, the highest mean values were registered in autumn and winter, and the lowest in summer in both years. During the winter months, winds of SSW and ENE prevailed in Beasain-Network, and WSW/SW/SSW and NE/ENE in Azpeitia and during the summer months, ENE in Beasain and NE/ENE in Azpeitia.

In general, all the air pollutants gathered in the same valley showed high correlations. Levels at different sites from in Urola Medio (Azkoitia and Azpeitia) showed good correlation levels for all trace elements (0.7<r<0.8) except Fe (r≈0.5). In Oria Valley, a high degree of correlation was observed between Beasain and Olaberria sites (correlations of 0.8–0.9 for all metals and PM2.5) and between Beasain and Ataun (r values between 0.6 and 0.7 for all trace elements, except Ni and Cd, r=0.5). A more moderate correlation was noted between Olaberria and Ataun (r values between 0.5 and 0.6, except Ni and Cd, r=0.4). The level of association between the towns in the Urola Alto Valley (Zumarraga and Legazpi) was low (0.2<r<0.5).

The percentage content of each trace element in PM10 and PM2.5 was calculated from 2 high-volume samplers placed in Urola Medio Valley. Significant differences (p<0.05) were found in the percentages of Pb, Fe, Ni and Cr between the particle sizes. The percentage composition of Pb is greater in PM2.5 than in PM10, whilst the percentage of Fe, Cr and Ni is greater in PM10. Mn and Cu, however, were distributed equally between both fractions.

Pearson’s linear correlation between the Network’s PM10 (continuous measurement) series and the PM2.5 study series for Beasain was r=0.86, with an adjusted R2 value of 0.71 (Table 4). In Azpeitia the linear correlation between the Network’s two continuous samplers (PM10 and PM2.5) was 0.91; and between the Network’s PM10 and the study’s PM2.5, 0.87. The adjusted R2 values were 0.85 and 0.78 respectively. The correlation between the Network’s PM2.5 and the study’s PM2.5 was 0.98, with an adjusted R2 value of 0.94.

In general, the concentrations of PM2.5 observed in the study area (12.2–28.9μg/m3) were lower than those set by the European Community; the limit of 25μg/m3 was exceeded only in Urola Alto Valley, in 2006. PM2.5 values were lower in 2007 than in 2006 in all locations, which could be attributed to a range of factors such as, for example, more favourable weather conditions (a higher percentage of rainy days). With regard to trace element content associated to PM2.5 there is not any regulation. The only existing guideline values for trace elements are associated to PM10 and not to PM2.5. Therefore this study was based only on descriptive analysis of trace elements.

In this case, similar behaviour was observed among towns located in the same valley. The notable drop in PM2.5 and associated trace elements observed in the Urola Medio Valley, for example, could be due to the installation of a particle capture device in January 2007 in the main steel plant in Azkoitia and to the termination of the plant’s activity in July 2007. This could also be the reason for the different seasonal variations observed for all pollutants between two years. While the distance between this plant and the sampling site in either town is considerable (1.5km to Azkoitia and 4.5km to the Azpeitia sampler), the prevailing SW/NE (nocturnal drainage) winds transport the pollutants throughout the valley, evidence that can be found in the high correlation between the two sites (Pearson r values between 0.7 and 0.8). Fig. 1 shows Azpeitia-Network PM2.5 time-series of the whole period study.

Figure 1.

Azpetia’s Network PM2.5 time serie (01/01/2006–31/12/2007).

(0.19MB).

The trace element levels in Beasain, Olaberria and Ataun (Oria Valley) also show similar behaviour. In this case an increase in Fe, Pb, As and Mn concentrations was observed in the three sites between 2006 and 2007. The wind direction data explains the high level of association between Olaberria and Beasain. In the case of Ataun, the mean values for trace elements found in the air (which were greater than those that have been reported for European cities) can only be explained by the fact that the town is located in the same valley.

The results for Zumarraga and Legazpi (Urola Alto Valley), were not as might have been expected. Legazpi, one of the towns with the highest levels of iron and steel foundries, displayed values lower than all other towns in the study area, both for PM2.5 and for the majority of the associated trace elements. The level of correlation between the trace elements recorded in Zumarraga and Legazpi was low (r<0.5), although the distance between these two samplers was the same as between those in Azkoitia and Azpeitia: 4.5km. This suggests, once again, that distance is not the only factor to be taken into account when evaluating air quality for a particular area. In June 2008 the Basque Government Air Quality Network placed a new equipment in Zumarraga. Wind roses have been obtained for the same months but in this case, for 2008 and 2009. The prevailing wind in winter is predominantly from the West, which is opposite to Legazpi's sampling site, and in the summer months, from the North and West. This information suggests that the air masses tend to move down the valley, towards the town of Zumarraga (especially the night drainage winds, which blow W and WSW), thus impeding the upward flow of air.

The main limitation of this study lies in the incomplete series of trace elements for all municipalities because only three high-volume samplers were available. For this reason the samplers were rotated monthly among the three valleys in order to gather information about temporal variability.

In general, a high degree of seasonal variability was observed. This could be attributed to the role played by the meteorological variables in the three valleys, as well as changes in industrial activity throughout the year, how was happened in Azkoitia. According to Querol et al, 200816, in autumn and winter, atmospheric conditions favour the accumulation of pollutants in the valleys of the Basque Country, in a phenomenon known as air mass stagnation.

As commented before, since there is no existing regulation, few investigations have analyzed the concentrations of trace elements in the PM2.5fraction13,21,22,23,24 (Table 5). These authors describe concentrations of PM2.5 and/or various associated trace elements sampled in a range of medium to large European cities (160,000 to 4 million inhabitants): Athens, Rome, Amsterdam, Helsinki, Prague, Erfurt, Basel and Frankfurt. If one compares the range of mean PM2.5 values for the two years for the towns in Gipuzkoa (12.2–28.9μg/m3), with the results of the above-mentioned investigations (10.5–29μg/m3, except for Athens: 37.3mg/m3), the results are similar. The main source of contamination could be different, however, this being traffic in the European cities22 and industry in the present study.

The mean concentrations of trace elements in the PM2.5 fraction in the study area greatly exceeded those present in European cities: Pb (34.2–154.5ng/m3 vs. 2.3–11.2ng/m3), Cr (4.1–19ng/m3 vs. 0.35–1.65ng/m3), Cu (8.2–45.5ng/m3 vs. 1.6–2.5ng/m3), Mn (18.9–89.8 60ng/m3 vs. 2.5–6.4ng/m3) and Fe (210–720ng/m3 vs. 60–259ng/m3), Ni (2.3–7.1ng/m3 vs. 0.8–4.6ng/m3) and Cd (0.3–0.8ng/m3 vs. 0.1–0.4ng/m3). In contrast, the results for trace elements associated to PM10 obtained in another Basque town, Llodio, with similar demographical, environmental and industrial characteristics to the towns in our study area, display similar levels16, except for Ni, which is higher.

The percentage composition of each of the trace elements in the PM2.5 and PM10 fractions is similar to that found by other authors13. The percentages of Fe, Cr and Ni were higher in PM10, whilst the percentage of Pb was greater in the PM2.5 fraction.

The high degrees of correlation obtained between the PM2.5 values from the high-volume samplers installed specifically for this study and the Network’s PM10 and/or PM2.5 values, indicate that the data gathered by both methods–gravimetric analysis and β-radiation attenuation method–are comparable in our area of investigation.

In general, the levels of particles found in the study area–a group of small towns with high levels of industrial activity–are similar to or higher than those of European cities while levels of analyzed trace elements are higher than those registered at other non-industrialized European cities. This has implications for the evaluation of health risks in environments that might be considered, in principle, as less polluted.