Characteristics and Sources of Fine Particulate Matter in Urban Air

Particulate matter is an important component of the air pollution mixture that has been linked to various adverse health outcomes as well as having general environmental effects. It is not yet clear which causative agents and underlying mechanisms are responsible for these adverse health effects. However, it is likely that the source of the particles, a factor which influences their characteristics and composition, plays a role in determining the health effects of particles. The focus of this thesis is first on characterizing urban ambient particulate matter with the emphasis on resuspended soil and street dust, and second, on the identification and comparison of major sources of urban particulate matter in Europe. The data for this work has been collected during three epidemiological studies conducted during 1995-1999 in Finland (Kuopio 1995, Number of measurement days=36; Helsinki 1996-97, N=185 and 1998-99, N=182), The Netherlands (Amsterdam 1998-99, N=237) and Germany (Erfurt 1998-99, N=177). Mass and number concentrations of several size fractions of particulate matter were monitored on a daily basis mainly during winter and spring months. Concentrations of gaseous air pollutants, elemental composition of PM samples and absorption coefficients of PM filters, a marker for carbonaceous particles from combustion sources, were also measured. Average mass concentrations of particulate matter (PM2.5 and PM10) were notably lower in Helsinki compared to Amsterdam and Erfurt, but the difference was less clear for number concentrations of ultrafine particles (in this work, 10-100 nm) and absorption coefficient of PM2.5, both of which are markers mainly for traffic-related fine particulate matter at the measurement sites used in this study. In Helsinki, the absorption coefficients of PM2.5 and PM10 samples were rather similar and clearly higher than that of PM1 samples, indicating that the fraction of particles between 1-2.5 m in diameter contained substantial amounts of carbonaceous material. Absorption coefficients were better correlated with the number than with the mass of particles. Sources of PM2.5 were resolved from two measurement periods in Helsinki and one period in Amsterdam and Erfurt. Source categories with very similar elemental profiles were obtained for both periods in Helsinki and Amsterdam, with the exception that a component related to industrial activities was detected only in Amsterdam. In all cities, secondary and primary particles from long-range sources and local traffic were clearly the most important determinants of PM2.5 concentrations. In Helsinki, the relative contribution of long-range transported air pollution to average PM2.5 levels was clearly higher (50%) than in the two Central European cities (32-34%). On the other hand, traffic made more similar relative contributions to PM2.5 in each city (23-36%). Owing to the higher PM2.5 concentrations in Amsterdam and Erfurt, the PM2.5 mass contributions from long-range transport were equal and contributions from traffic notably larger in these two cities compared to Helsinki. The other major components of PM2.5 identified included soil, oil combustion particles, sea salt and particles from industrial sources. A comparison of a statistical multivariate method (principal component analysis with multiple linear regression) and a deterministic method (chemical mass closure), indicated that the agreement of the two source apportionment methods was good for sources whose chemical composition is stable and well defined, such as crustal dust and sea salt. Source contribution estimates for the more complex sources of particulate matter, such as local combustion and long-range transported air pollution, differed substantially between the two methods. Notable episodes of high resuspended dust concentrations took place at all monitoring sites. However, a seasonal influence of these episodes was seen only in Finland, shown by high concentrations of crustal dust during the spring months. These dust episodes affected the PM2.5/PM10 ratio more than the PM1/PM2.5 ratio. Also, the correlation of PM2.5 with PM10 during winter was clearly higher than during spring, while the correlation of PM2.5 with PM1 did not change from season to season. The stability of the PM1/PM2.5 ratio suggests that in terms of variation of fine particle mass, monitoring of PM1 did not significantly add to the information content already obtained from monitoring of PM2.5. The major cause for elevated concentrations of resuspended dust in urban environments seems to be the turbulence and tyre stress related to traffic. Results from Kuopio also showed that resuspended particulate matter could contain considerable amounts of trace elements from anthropogenic sources. In summary, the two most distinctive features in the Finnish cities (Helsinki and Kuopio) compared to Amsterdam and Erfurt were, in addition to the lower average mass concentrations of particles, the large relative contribution from secondary and long-range transported particulate matter and the seasonally dependent episodes of resuspended dust in urban environments.