抄録
Cases of dioxin pollution where the dioxin levels are over the environmental quality standard have been found in many places in Japan. For the establishment of their remediation plans, polluters should be identified and the source apportionments among polluters should to be carried out. In this paper, the authors applied Positive Matrix Factorization (PMF), one of new receptor models, to an actual river sediment pollution case and discussed the obtained results.
First, the authors mixed four known source profiles of dioxins by Monte Carlo simulation and produced pseudo environmental pollution data. Then PMF was applied to the dataset to see if the original source profiles could be extracted and also to see what kind of data transformation was appropriate to compare the source profiles extracted by PMF with the original source profiles. Then, PMF was applied to the actual dioxin pollution case in Furuayase River, Saitama Prefecture (Dataset: 40 congeners X200 samples). The results showed that 5 factors were contributing to the pollution. Those were identified as incineration, PCB formulation, chloronitrophen (CNP) and pentachlorophenol (PCP) derived dioxins in addition to one unknown source that had lower-chlorinated dibenzofurans in high concentration and dioxin-like PCBs in lower concentration as its characteristic congeners. The ratio of contributions from PCB formulation and from unknown source, however, were uniform for the highly contaminated area and varied for lower contaminated area, indicating that these two sources might have come from a same source in highly contaminated area.
The current study showed that PMF could be successfully applied to a river sediment dioxin pollution case, however, careful interpretation of PMF results was necessary.
