Abstract
The multivariate receptor models Positive Matrix Factorization (PMF) was used along with Chemical Mass Balance (CMB) model to evaluate source apportionment of dioxins in air samples collected in Fukuoka prefecture. A total of 210 air samples were obtained from 1999 to 2004 and were analyzed for 116 isomers including coplanar PCBs. The PMF model predicted that the major sources of dioxins in the air samples were waste incineration, PCBs, pesticide pentachlorophenol (PCP) and chloronitrophen (CNP) . The estimated mean source contribution rate of dioxins by the PMF model were as follows: incineration 21%, PCB 35%, PCP 27% and CNP 16%, respectively.
The source profiles calculated from the PMF model were compared with measured source data. Source profiles of incineration, PCB and CNP resembled each other, but PCP was slightly different.
A total of 9 cases including both PMF modeling and measured sources underwent evaluation by the CMB method to determine source contribution for individual air samples. As a result, five cases obtained good estimation. But taking some conditions into consideration, it was thought that the case of three mesaured source data and one PMF profile data (that of PCB) produced relatively better results. According to these source data, source contribution rate of dioxins in air were as follows: incineration 40%, PCB 44.1%, PCP 2.1% and CNP 7.2%, respectively.
