Determining the influence of different atmospheric circulation patterns on PM10 chemical composition in a source apportionment study

This study combines a set of chemometric analyses with a source apportionment model for discriminating the weather conditions, local processes and remote contributions having an impact on particulate matter levels and chemical composition. The proposed approach was tested on PM10 data collected in a semi-rural coastal site near Venice (Italy). The PM10 mass, elemental composition and the water soluble inorganic ions were quantified and seven sources were identified and apportioned using the positive matrix factorization: sea spray, aged sea salt, mineral dust, mixed combustions, road traffic, secondary sulfate and secondary nitrate. The influence of weather conditions on PM10 composition and its sources was investigated and the importance of air temperature and relative humidity on secondary components was evaluated. Samples collected in days with similar atmospheric circulation patterns were clustered on the basis of wind speed and direction. Significant differences in PM10 levels and chemical composition pointed out that the production of sea salt is strongly depending on the intensity of local winds. Differently, typical primary pollutants (i.e. from combustion and road traffic) increased during slow wind regimes. External contributions were also investigated by clustering the backward trajectories of air masses. The increase of combustion and traffic-related pollutants was observed when air masses originated from Central and Northwestern Europe and secondary sulfate was observed to rise when air masses had passed over the Po Valley. Conversely, anthropogenic contributions dropped when the origin was in the Mediterranean area and Northern Europe. The chemometric approach adopted can discriminate the role local and external sources play in determining the level and composition of airborne particulate matter and points out the weather circumstances favoring the worst pollution conditions. It may be of significant help in designing local and national air pollution control strategies.

► Elements and ions in 193 PM10 samples were studied by Positive Matrix Factorization.
► Results were geochemically interpreted in relation to weather data.
► Days with similar circulation patterns were clustered on the basis of wind data.
► Samples with similar air mass histories were grouped using back-trajectories.
► The major components of PM10 chemistry and dynamics in NE Italy were identified.