Extension of self-modeling curve resolution to mixtures of more than three components: Part 3. Atmospheric aerosol data simulation studies

In the previous two papers [R.C. Henry, B.M. Kim, Extension of self-modeling curve resolution to mixtures of more than three components: Part 1. Finding the basic feasible region, Chemom. Intell. Lab. Syst. 8 (1990) 205–216; B.M. Kim, R.C. Henry, Extension of self-modeling curve resolution to mixtures of more than three components: Part 2. Finding the complete solution, Chemom. Intell. Lab. Syst. 49 (1999) 67–77.], a method was described to extend the self-modeling curve resolution (SMCR) technique to mixtures of more than three components, one set of data was created without errors to examine the performance of the model when no random measurement errors exist. In this paper, simulation studies were conducted to examine the performance of the model and to determine the effects of random measurement errors, both in source compositions and in ambient concentrations, on the estimated source compositions. The bias and errors in the estimated source compositions were also determined. Five sources (roadway, marine, secondary, crustal, and residual oil) were assumed in the simulation. It was shown that the model estimates source compositions with acceptable error and bias. The maximum percentage uncertainties in the estimated compositions of the roadway and secondary sources were less than 10% for most of the major species, except elemental carbon in the roadway and organic carbon (OC) in the secondary sources, which are 20%. The maximum uncertainties in the estimated marine compositions were about 30% for major species.