Comparison of lead isotopes with source apportionment models, including SOM, for air particulates

We have measured high precision lead isotopes in PM2.5 particulates from a highly-trafficked site (Mascot) and rural site (Richmond) in the Sydney Basin, New SouthWales, Australia to compare with isotopic data from total suspended particulates (TSP) from other sites in the Sydney Basin and evaluate relationships with source fingerprints obtained from multi-element PM2.5 data. The isotopic data for the period 1998 to 2004 show seasonal peaks and troughs that are more pronounced in the rural site for the PM2.5. samples but are consistent with the TSP. The Self Organising Map (SOM) method has been applied to the multi-element PM2.5 data to evaluate its use in obtaining fingerprints for comparison with standard statistical procedures (ANSTO model). As seasonal effects are also significant for the multi-element data, the SOM modelling is reported as site and season dependent. At the Mascot site, the ANSTO model exhibits decreasing 206Pb/204Pb ratios with increasing contributions of fingerprints for “secondary smoke” (industry), “soil”, “smoke” and “seaspray”. Similar patterns were shown bySOMwinter fingerprints for both sites. At the rural site, there are large isotopic variations but for the majority of samples these are not associated with increased contributions from the main sources with the ANSTO model. For two winter sampling times, there are increased contributions from “secondary industry”, “smoke”, “soil” and seaspray with one time having a source or sources of Pb similar to that of Mascot. The only positive relationship between increasing 206Pb/204Pb ratio and source contributions is found at the rural site using the SOM summer fingerprints, both of which show a significant contribution from sulphur. Several of the fingerprints using either model have significant contributions from black carbon (BC) and/or sulphur (S) that probably derive from diesel fuels and industrial sources. Increased contributions from sources with the SOM summer fingerprints could explain the summer-time peaks in isotopic ratio at both sites and, at the rural site, be associated with meteorological influences. Nevertheless, the SOM results indicate that there are multiple overlapping ‘weak’ sources.