Gas chromatography-pattern recognition techniques in pollution monitoring

Gas chromatography and pattern recognition methods (GC-PR) constitute a powerful tool for investigating complex environmental problems, e.g., source identification of pollutants. Multivariate methods are necessary to realistically analyze large chromatographic data sets and to seek meaningful relationships between chemical constitution and environmental variables. Recently, our laboratory has investigated the potential of GC-PR as a method for typing jet fuels so a spill sample could be related to its source. Fuels that had undergone weathering in a subsurface environment were correctly identified as to type using disjoint principal component models developed from gas chromatograms of unweathered jet fuels. The constituents of synthetic binary fuel mixtures were also correctly identified by fuel type using the principal component models.