Definition and characterization of petroleum compositional families in Williston Basin, North America using principal component analysis

Petroleum hydrocarbons in the gasoline range (GRH) and saturate (SFH- >210 degreesC boiling point) fractions carry information that is often obscured by compositional diversity and multiple processes working simultaneously. Multivariate statistical methods can enhance the analysis and interpretation of compositional data from these fractions, especially in conjunction with independent geological information. In the present study, Principal Component Analysis (PCA) was applied to the GRH and SFH data for 171 oil samples from the Williston Basin. These oils were previously classified using polycyclic terpane and sterane biomarker traits. Our results indicate that only Family A oils can be uniquely classified using PCA. Families B, C and D oils show GRH and SFH characteristic compositions consistent with biomarker-defined families, but these characteristics are insufficient for independent classification. However, the PCA analyses of the GRH and SFH compositional traits proves to be a useful technique ill recognizing the effect of mixing of oils derived from different sources. (C) 2004 Elsevier Ltd. All rights reserved.