Analysis of aerosol particles and coarse particulate matter concentrations in Chillán, Chile, 2001-2003.

Daily particle samples were collected in Chillán, Chile, at six urban locations from September 1, 2001, through September 30, 2003. Aerosol samples were collected using monitors equipped with a Sierra Andersen 246-b cyclone inlet on Teflon filters. Average concentrations of coarse particulate matter (PM10) for the 2001-2003 period ranged from 43.4 microg/m3 to 81.8 microg/m3 across the six sites. Annual PM10 concentration levels exceeded the European Union concentration limits. Mean PM10 levels during the cold season (April through September) were more than twice as high as those observed in the warm season (October through March). Average contributions to PM10 from organic matter, soil dust, nitrate (NO3-), elemental carbon, ammonium (NH4+), and sulfate (SO4(2-)) were 31%, 27%, 11%, 8%, 7%, and 5%, respectively. The chemical analyses indicated that carbonaceous substances were the most abundant components of PM10 in cold months, whereas crustal material was the most abundant component of PM10 during warm months. Higher concentration levels were observed in the downtown area suggesting a clear anthropogenic origin, whereas in the rural sites the source was mainly natural, such as resuspended soil dust associated with traffic on unpaved roads and agricultural activities.

Source Apportionment of Soil Samples by the Combination of Two Neural Networks Based on Computer-Controlled Scanning Electron Microscopy.

The apportionment of ambient aerosol mass to different sources of airborne soil is a difficult problem because of the similarity of the chemical composition of crustal sources. However, additional information can be obtained using individual particle analysis. A novel approach based on the combination of two neural networks, the adaptive resonance theory-based neural network (ART-2a) and the back-propagation (BP) neural network with electron microscopy data, has been developed to apportion the mass contributions of the crustal sources to ambient particle samples. The crustal source samples were analyzed using computer-controlled scanning electron microscopy (CCSEM). CCSEM provides elemental compositions and size parameters for individual particles as well as estimates of the shape and density from which the volume and mass of each particle can be estimated. The ART-2a neural network was first used to partition particles into homogeneous classes based on the elemental composition data. After the different particle type classes were produced by ART-2a, their mass fractions were calculated. In this way, the source profiles for the crustal dust sources can be obtained in terms of the mass fractions for different particle types. Then the BP neural network was applied to build the model between the mass fractions of different particle types and the mass contributions. Using the three physical source samples prepared for this study, artificial ambient samples were generated by randomly mixing particles from the three source samples. These samples were then used to examine the proposed method. Satisfactory predictions for the mass contributions of the three sources to the ambient samples have been obtained, indicating the proposed method is a promising tool for the source apportionment of chemically similar soil samples.