Abstract

Most oil characterisation procedures are time consuming, labour intensive and utilise only part of the acquired chemical information. Oil spill fingerprinting with multivariate data processing represents a fast and objective evaluation procedure, where the entire chromatographic profile is used. Methods for oil classification should be robust towards changes imposed on the spill fingerprint by short-term weathering, i.e. dissolution and evaporation processes in the hours following a spill. We propose a methodology for the classification of petroleum products. The method consists of: chemical analysis; data clean-up by baseline removal, retention time alignment and normalisation; recognition of oil type by classification followed by initial source characterisation. A classification model based on principal components and quadratic discrimination robust towards the effect of short-term weathering was established. The method was tested successfully on real spill and source samples.

Links

Publisher Full Text

Authors

Nielsen NJ, Ballabio D, Tomasi G, Todeschini R, Christensen JH

Institution

Department of Basic Sciences and Environment, Faculty of Life Sciences, University of Copenhagen, Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark. njn@life.ku.dk

Source

Journal of chromatography. A 1238: 2012 May 18 pg 121-7

MeSH

Algorithms
Bayes Theorem
Calibration
Discriminant Analysis
Flame Ionization
Least-Squares Analysis
Petroleum
Petroleum Pollution
Principal Component Analysis
Reproducibility of Results
Weather

Pub Type(s)

Journal Article

Language

eng

PubMed ID

22503620