Qualitative and Quantitative Evaluation of Biodegradation Data Using Multivariate Modeling

Bent Skaare Pedersen¹, Egil Nodland¹, Tanja Barth¹, Heinz Wilkes², Andrea Vieth², and Hege Ommedal^3
1 University of Bergen, Bergen, Norway
² GeoForschungsZentrum Potsdam, D-14473 Potsdam, Germany
³ CIPR - Centre for Integrated Petroleum Research, Bergen, Norway

Biodegradation of crude oil has been shown to be a complicated process, with parallel or overlapping progressive changes of the hydrocarbon composition. Studies of such alterations, especially using chromatographic methods, result in a very large amount of data. By using multivariate tools, the extraction of systematic information from such large data sets can be made easier and more quantitative.

The sequential and quantitative changes in oil composition caused by anaerobic biodegradation in a controlled laboratory study has been mapped and subjected to multivariate analysis using the program package Sirius™. A Principal Component (PCA) model was first established, systematically identifying the differences between oils altered by biodegradation and oils which were not subject to the same microbial activity. The distinct differences in the composition of these two groups were also shown using SIMCA (Soft Independent Modeling of Class Analogies) modeling.

PLS (Partial Least Squares) modeling was then used to define a quantitative axis of biodegradation, and showing the relation between the individual components on this axis. The usefulness of this model relative to field data will be evaluated for incorporation into field studies and biodegradation rate models.

AAPG Search and Discovery Article #90039©2005 AAPG Calgary, Alberta, June 16-19, 2005