A Holistic approach Combining Factor Analysis, Positive Matrix Factorization and UNMIX Applied to Receptor Modeling

Abstract

The contribution of various sources over the air pollution enables to establish effective control policies. The present work focuses on the holistic approach of combining factor analysis (FA), positive matrix factorization (PMF) and UNMIX in order to identify the sources and their contributions through analysis of ions and elements in particulate matter (PM₁₀), organic carbon (OC) and elemental carbon (EC). Insight from the emission inventory, various models has been used to remove subjectivity in source identification. In the first method the proposed approach verified using a synthetic data and then the method has been again repeated with field (R.K Nagar, Chennai, India) study data. Factor analysis identified four factors with 99.7% of the variation in the measured data. PMF and UNMIX identified marine aerosol, fuel oil combustion, coal combustion and soil dust based on source profile and contribution results of synthetic dataset. The R.K Nagar data provided four sources such as marine aerosol, soil dust, coal combustion and secondary aerosols through the PMF and UNMIX. The obtained source contributions of these sources from PMF and UNMIX are 30.13% and 56.44%, 9.43% and 3.83%, 39.99% and 19.08%, 20.45% and 20.65% respectively.