Integrated Methodology for Forensic Oil Spill Identification JAN H. CHRISTENSEN,* ,†,‡ ASGER B. HANSEN, † GIORGIO TOMASI, § JOHN MORTENSEN, ‡ AND OLE ANDERSEN ‡ Department of Environmental Chemistry and Microbiology, National Environmental Research Institute, Frederiksborgvej 399, P.O. Box 358, 4000 Roskilde, Denm ark, Departm ent of Life Sciences and Chem istry, Roskilde University, Universitetsvej 1, P.O. Box 260, 4000 Roskilde, Denm ark, and Department of Food Science, The Royal Veterinary & Agricultural University, Frederiksberg, Denm ark A new integrated methodology for forensic oil spill identification is presented. It consists of GC-MS analysis, chromatographic data processing, variable-outlier detection, multivariate data analysis, estimation of uncertainties, and statistical evaluation. The methodology was tested on four groups of diagnostic ratios composed of petroleum biomarkers and ratios within homologous PAH categories. Principal component analysis (PCA) was employed and enabled the simultaneous analysis of many diagnostic ratios. Weathering was taken into account by considering the sampling uncertainties estimated from replicate spill samples. Statistical evaluation ensured an objective matching of oil spill samples with suspected source oils as well as classification into positive match, probable match, and nonmatch. The data analysis is further refined if two or more source oils are classified as probable match by using weighted least squares fitting of the principal components, local PCA models, and additional information relevant to the spill case. The methodology correctly identified the source of two spill samples (i.e.,crude oils from Oseberg East and Oseberg Field Centre) and distinguished them from closely related source oils. Introduction Waterborne oilspills ofunknown origin often occur in rivers, in open waters, and in coastal waterways. These spills range from the continuous leakage from land sources and illegal tank washings at sea to larger spill accidents. The liability associated with oil released into the environment warrants a comprehensive chemical characterization to defensibly determine the oil source(s), distinguish spilled oil from background hydrocarbons, and assess the impact on the ecosystem. Consequently, defensible differentiation and characterization of spilled oil is a critical part of many oil spill assessments. Chemical fingerprinting is an important selection of techniques for solving such liability issues. Gas chromatog- raphy-flame ionization detection (GC-FID)is typicallyused for screening followed by a more comprehensive chemical analysis by gas chromatography-mass spectrometry (GC- MS) (1, 2). GC-MS is the standard analytical technique for oil chemical fingerprinting, because it can resolve a broad range of petroleum biomarkers and polycyclic aromatic hydrocarbons (PAHs) including NSO compounds and be- cause of the low cost of quadrupole instruments. Other instrumental techniques include metastable ion reaction monitoring GC-MS-MS (3), two-dimensional gas chroma- tography (GC-GC) (4), gas chromatography-isotope ratio massspectrometry(GC-IRMS) (5),and excitation -emission fluorescence spectroscopycombined with multivariate data analysis (6). Data analysis is a vital part of chemical fingerprinting, and a wide range of biomarkers are used in this respect. Particular attention has been given to tri-pentacyclic trit- erpanes and steranes, because theyare indicators ofsource, maturation,and in-reservoirweatheringand biodegradation (7), are found in high concentrations in most oils (7),and are recalcitrant when released into the environment (2). Petro- genic PAHs are dominated almost exclusively by the C1-C4 alkylated homologuesofthe parent compounds(in particular, naphthalene, phenanthrene, dibenzothiophene, fluorene, and chrysene) (8). The effect of short-term weathering processes (mainly evaporation and water washing) can be considerable on the distribution of C0-C4 homologues but is limited on the distribution ofPAH isomers within a homo- logue category (2, 9). Consequently, it can be kept to a mini- mum byfocusingthe data analysison ratiosbetween concen- trations, or peak areas, of biomarkers and of PAH isomers within a homologue category(diagnostic ratios).Conversely, the distribution of biomarkers and of PAHs within a homo- logue categorycan be altered bybiodegradation because the individual isomers have varying susceptibility to microbial degradation (10). Developmentsin oilfingerprintingproceduresduringthe 1990s have concentrated on analytical techniques and on the application of new diagnostic ratios. Less attention has been paid to data analysis, which is often limited to visual comparison of ion chromatograms, bar charts of alkanes and PAH concentrations,and double plotsofdiagnosticratios (8, 11).Multivariate methodshave been used for data analysis in organic geochemistry since the 1980s (12, 13) and have only recently been adopted for oil spill identification (2, 14, 15). The advantages of such methods as compared to uni- variate ones are manifold;in particular,theyallow for simul- taneous analysis of a vast number of correlated variables by proper handling of redundant information. In this paper, we present an integrated methodology for forensic oil spill identification based on GC-MS fingerprint- ing, multivariate data analysis, and statistical comparisons of the chemical composition of oil spill samples with that of source oils. The paper is organized in two sections: first, the methodology is outlined, focusing on multivariate data analysis and statistical evaluation; then it is tested on spill samples and source oils from a Nordtest Round-Robin oil spill exercise (16). Methodology Forensic oil spill identification deals with the process of defensibly linking spilled oil with its generic source. True identity in this context would imply all data to be identical, but as mineral oils consist of thousands of individual compoundsand theircomposition are affected byweathering processes, this is an impossible task. Hence, the rationale of forensic oil spill identification is that identity prevails if no significant chemical differences can be demonstrated by *Corresponding author telephone: +45-46301200; fax: +45- 46301114; e-mail: jch@dmu.dk. † National Environmental Research Institute. ‡ Roskilde University. § The Royal Veterinary & Agricultural University. Environ. Sci. Technol. 2004, 38, 2912-2918 2912 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 10, 2004 10.1021/es035261y CCC: $27.50  2004 American Chemical Society Published on Web 04/17/2004