An introduction to receptor modeling

Abstract

A major problem facing air quality management personnel is the identification of sources of airborne particles and the quantitative apportionment of the aerosol mass to those sources. The ability to collect particle samples and analyze these samples for a suite of elements by such techniques as neutron activation analysis or X-ray fluorescence provides the data for the problem of resolving a series of complex mixtures into its components based on the profiles of the elements emitted by the various sources in the airshed. If all of the sources and their composition profiles are known, then the mass balance model becomes a multiple regression problem. If a series of samples have been analyzed without substantial information being available on the sources, factor analysis methods can be employed. In both situations, the analysis is complicated by higher levels of measurement error in these analyses than in typical spectrochemical problems. In addition, the source profiles can vary as the composition of input materials for the emission sources change in time. Thus, there are limitations to the ability of statistical methods to resolve sources in real world problems. The physical and statistical basis of these methods and their application to representative problems will be reviewed.