Understanding the complexities of PM10Source apportionment using Positive Matrix Factorization at three kerb-sites and control site of Mumbai, India
Indrani Gupta, Abhaysinh Salunkhe, Rakesh Kumar
NEERI, Mumbai Zonal Centre
Abstract Number: 134
Working Group: Urban Aerosols
Last modified: April 1, 2011
Particulate Matter (PM$_(10)) has been the main air pollutant exceeding the ambient standards in most of the major cities in India. Source apportionment has been considered a key process to determine the steps for development of action plan in these cities. The methodologies used for source apportionment pose some complexities with regard to understanding of local sources and its impact on the outcome. During last few years, receptor models which include Chemical mass balance (CMB), positive matrix factorization (PMF), PCA–APCS and Unmix have been accepted for developing effective and efficient air quality management plans. Rizzo and Scheff, 2007 compared the magnitude of source contributions resolved by MB and PMF models and examined correlations between PMF and CMB resolved contributions. They found the major factors correlated well and were similar in magnitude; additionally, PMF resolved source profiles were generally similar to measured source profiles. Callen et al. 2009, carried out source apportionment of PM10 by three receptor models: PCA–APCS, Unmix and PMF. They concluded that greater requirements of measure of uncertainty in PMF permitted to obtain better results than with the other two models. In the present study, samples of PM10 were collected from three kerbsites of Mumbai and a background site during April 2007 to March 2008. Samples were analyzed for elemental concentrations by ICP-AES; EC/OC using DRI’s Thermal/Reflectance Optical Carbon Analyzer and ions using Ion Chromatograph. The data sets were subjected to PMF process to identify the possible sources of atmospheric aerosols in these areas. The best solutions were found to range between seven, to nine factors. The identified factors effectively predicted the measured PM10 concentrations (R2 varying between 0.7 to 0.8). The percentage and composition of the sources varied widely among the sites depending on the land use. Despite assuming that three kerb sites will provide a similar results being dominated by vehicles, however. it was seen that all three sites provided different source profiles which were based on variation in landuse as also effect of some distant sources. The complexities of sources indicated that for development of action plan sources other than local sources also will need better attention. Some of the major sources at kerb sites are nitrate, motor vehicles, re-suspended road dust, oil fired power plant, vegetative burning etc. whereas at background site the major sources are natural soil, marine aerosol, wood burning and secondary ammonium sulphate. The paper presents the source apportionment (SA) profiles of widely differing sources at kerb-sites. SA is more important in Mumbai and other cities in India, as large population lives close to road side leading to enormous population exposure.