Elemental Composition of Airborne Particulates And Source Identification: Data Analysis Techniques

Abstract

Approaches are discussed and illustrated for the analysis of the very large trace element data set obtained by instrumental neutron activation and emission spectroscopic analysis of over 700, 24 hour ambient air particulate samples collected at 16 sites in Cleveland, OH from August 1971 to October 1972. Examples are presented of the use of pairwise correlation statistics, cluster analysis algorithms, enrichment factors, and pollutant concentration roses to identify sources. It is demonstrated that some elemental constituents of the Cleveland suspended particulate can be related to specific source types: earth’s crust (Al, Si, Sc, Ti, V, La, Sn, Eu, Oy, Th), automotive exhaust (Pb, Br), metallurgy (Fe, Cr, Co) and specific industries (Sb, Hg). Even with 24 hr samples, some sources can be geographically located by the directional characteristics of the element.