Modified Polytopic Vector Analysis To Identify and Quantify a Dioxin Dechlorination Signature in Sediments. 1. Theory
2004-03-15T00:00:00Z (GMT) by
Risk-based sediment management decisions require the characterization of contamination sources and fate processes in the field. Polytopic vector analysis (PVA) is a multivariate technique based on a linear mixing model, used to resolve chemical fingerprints and suited for forensic investigations of environmental contamination. The traditional algorithm is constrained to positive fingerprint (end-member) components and cannot resolve fingerprints with both positive and negative values required for a reactive end-member. We developed a modified algorithm (M-PVA) to resolve a dioxin dechlorination fingerprint, indicative of biotic/abiotic transformations in field samples of sediments. The new procedure isolates from the dioxin pattern net compositional changes due to dechlorination in a separate end-member. Using two artificial data sets for which the composition and sample contribution of all end-members are known, the dechlorination fingerprint was reproduced with a root mean square error of 28−41%. The dechlorination end-member contribution to the total variability (set at 4.0 and 10.0%, respectively) was overestimated 1−5-fold. The ability of M-PVA to reproduce the dechlorination pattern and its variability contribution depends on the actual contribution of dechlorination to variability. At an actual contribution of 4.0%, the model outcome deviates more strongly from the original than is the case for a contribution of 10.0%. As such, application of M-PVA to environmental data should include an uncertainty analysis to distinguish variability due to dechlorination from variability due to error. The development of the modified PVA procedure is an important step toward the field characterization of fate processes in dioxin-impacted sediments.