Target-Transformation Factor Analysis of Airborne Particulate Samples Selected by Wind-Trajectory Analysis

Abstract

In a companion paper, Rheingrover and Gordon showed that samples significantly influenced by single point sources could be identified by examining concentration and meteorological values in the data base for the Regional Air Pollution Study of St. Louis, MO. They examined the relationships between measured elemental concentrations by simple linear regression analysis to deduce compositions of particles emitted by the identified sources. To obtain more complete source profiles, the same sets of sample data were subjected to target-transformation factor analysis. Results of target-transformation factor analysis are compared with those of other approaches. As expected, target-transformation factor analysis extracts concentrations for more elements than do linear regressions of the same data sets. Also, the application of target-transformation factor analysis to groups of samples previously selected by wind-trajectory analysis yields more satisfactory results than the application of target-transformation factor analysis to samples taken under unspecified wind conditions.