System of Air quality and weather Forecasting And Research (SAFAR-Mumbai) network hourly observations of air pollutants spanning half a decade are subjected to a Generalized Additive Model (GAM) to investigate the non-linear relationship of PM2.5 to other precursor gases and meteorological parameters. The Effective Degrees of Freedom (EDF), a statistic of GAM quantifying the non-linear relationship, reveals that PM2.5 is highly non-linearly dependent on NH3, SO2, and CO as inferred from the response curves exhibiting at least one inflexion point and threshold responses. PM2.5 is linearly related to NO2 (EDF = 1), while it is non-linearly related to other gases and meteorological parameters (EDF > 2). The frequency distribution of PM2.5 shows a sharp peak at 25 μg m−3 in monsoon and broad peaks in pre-monsoon (peak at 50 μg m−3), post-monsoon (peak at 60 μg m−3) and winter (peak at 100 μg m−3) indicating relatively large variability in winter and post-monsoon. Chemical analysis of PM2.5 shows a high contribution of four major ions in the order of SO42− > Cl− > Na+ > NH4+, mainly associated with anthropogenic and marine sources. The average sulphate oxidation ratio (SOR), nitrogen oxidation ratio (NOR), and neutralization ratio (NR) are up to 0.3, 0.04, and 0.27, respectively, indicating that SO2 is contributing majorly to secondary aerosol production. Monthly concertation of PM2.5, Ozone (O3), CO, NO2, SO2, and NH3 shows U-shaped variability with maximum in winter and minimum in monsoon.
Read full abstract