Application of SAFER model to the Los Angeles PM10 data

The chemical mass balance (CMB) model has been used widely in the PM10 source apportionment study and the PM10 state implementation plan (SIP) development. In its modeling application, the CMB model requires source composition profiles, which can either be measured by source testing or estimated by a multivariate receptor model such as source apportionment by factors with explicit restrictions (SAFER) from the ambient data only. The SAFER model is a multivariate receptor model utilizing a series of linear programming methods to estimate source compositions with explicit physical constraints. The SAFER model was applied to ambient PM10 data collected in 1986 in the South Coast Air Basin. Source compositions of selected major source categories were estimated. The SAFER model-estimated source profiles were compared with the measured source profiles by checking some important elemental ratios. Estimated source compositions were consistent with the measured profiles. Then the SAFER model-estimated source profiles were used for CMB analysis to estimate source contributions from each source category. Contributions from the roadway source range from 20 to 34 mu g m(-3), from the secondary source from 17.75 to 31.40 mu g m(-3), from the marine source from undetectable to 2.50 mu g m(-3), and from the crustal source from 4.06 to 8.13 mu g m(-3). Organic carbon seems to be mainly (81%) contributed by the primary roadway source, and sulfates and nitrates are mainly from the secondary source, although 32% of the sulfate is from primary sources such as roadway, crustal, and marine sources. Published by Elsevier Science Ltd.