APPLICATION OF STATISTICAL ANALYSIS IN THE SELECTION OF DIAGNOSTIC RATIOS FOR FORENSIC IDENTIFICATION OF AN OIL SPILL SOURCE

Chun Yang,Zhendi Wang,Bruce Hollebone,Carl E. Brown,Mike Landriault
DOI: https://doi.org/10.7901/2169-3358-2008-1-297
2008-01-01
Abstract:ABSTRACT In this work, 14 fresh crude oils of different types and origins were analyzed by gas chromatography with mass-selective detection, and over 80 potentially diagnostic ratios were calculated based on the quantitation of isoprenoids, polycyclic aromatic hydrocarbons (PAHs), biomarkers, diamondoids, bicyclic sesquiterpanes and aromatic steranes, etc. Diagnostic power (DP) was calculated for the selection of the candidate source-sensitive diagnostic ratios and used to determine which ratios were most diagnostic among the crude oils studied. In order to investigate the effect of evaporative and biodegradative weathering on diagnostic ratios and thereby to differentiate weathering-resistant ratios from weathering-sensitive ratios, triplicate analyses were performed for two suites of reference oils, laboratory-evaporated Prudhoe Bay crude oils and laboratory-biodegraded Alberta Sweet Mixed Blend (ASMB) crude oils, respectively. Student'S t-test was used to statistically evaluate whether diagnostic ratios were significantly affected by weathering and to ensure that the observed change is not due to analytical variance. It was found that, diagnostic ratios generally remained consistent for oils with slight to medium evaporative weathering, only the ratios of those compounds with lower boiling points such as adamantanes changed greatly. For biodegraded oils, most of diagnostic ratios remained constant for lightly to moderately biodegraded oils; while most of diagnostic ratios with exception of certain triaromatic steranes and high-molecular-weight terpane and sterane biomarkers demonstrated significant changes for heavily biodegraded oils.
What problem does this paper attempt to address?