Improving receptor models for ambient air particulate matter source apportionment.

This paper discussed solutions for two of the major technical issues of source apportionment of ambient air particulate matter:the impact of secondary organic carbon(SOC) in the accuracy of CMB modelling;the impact of collinearity problem due to the extent of incompatibility between sources and receptor.For the first issue,the CMB model was modified by deducting SOC from the receptor.For the secondary issue,a PCA/MLR-CMB model was developed.The PCA/MLR-CMB model started with a PCA/MLR appointment,and thus,significantly improved the compatibility between the sources and receptor by reducing of the impact of unknown sources in receptor.