Multivariate statistical methods applied to interpretation of saturated biomarkers (Velebit oil field, SE Pannonian Basin, Serbia)

Twenty-five crude oils originating from the Velebit oil field (SE Pannonian Basin), the most important oil field in Serbia, were investigated. Saturated biomarkers (n-alkanes, isoprenoids, steranes and triterpanes) were analyzed by gas chromatography-mass spectrometry (GC-MS). Based on the distribution and abundance of these compounds, a large number of source and maturation parameters were calculated, particularly those most often used in correlation studies of oils. The examined samples were classified according to their origin and level of thermal maturity using factor, cluster and discriminant analyses. According to the source and maturation parameters, combined factor and cluster analyses using the Ward method enabled the categorization of the investigated oils into three groups. The cluster Ward analysis was shown to be of greater susceptibility and reliability. However, in addition to the two aforementioned methods, K-Means cluster analysis and discriminant analysis were shown to be necessary for a more precise and detailed categorization in the case of a large number of samples in one group. Consequently, it was concluded that factor and cluster K-Means and Ward analyses can generally be used for the interpretation of saturated biomarkers in correlation studies of oils, but the observed results have to be checked, i.e., confirmed by discriminant analysis