Regression analysis of aerosol optical properties with long-term MODIS data using forward selection method

In this work, a new filtering method of forward selection (FS), has been employed for linear and multiple regression analysis of aerosol optical properties with meteorological parameters using long-term moderate resolution imaging spectroradiometer (MODIS) data for New Delhi area. Long-term observation (15 years, March 2000–December 2014) of aerosol optical depth (AOD) at 550 nm, fine mode fraction particles (FMF) and Angstrom exponent (AE) are analyzed to study their variability on annual and seasonal basis. The analysis reveals relatively high mean AOD (0.71 ± 0.25), associated with moderate AE (0.69 ± 0.13) for the overall period. Considerable higher values of AOD (0.89 ± 0.33) were reported in monsoon and AE (0.87 ± 0.10) during winter season. Likewise, seasonal fraction for AOD was also found to be higher (42%) during the monsoon season and for AE (31%) and for FMF (55%) during winter season. The peak value of AOD was during June–July and the lowest in the transitional months of February and September, while AE was high in January–February and low in May–June suggesting significant urban and biomass burning contribution. In general, the analysis shows a rather well-mixed type of aerosols present in the urban environment, which affect the regional air quality as well as climate associated with long-range transport of pollutants through the westerly winds from the Thar Desert and biomass burning in the western parts of India. The presence of AOD trend is evidence of air quality deterioration particularly in highly populated areas. The aerosol classification with relationship between AOD and alpha shows that urban/biomass burning aerosols are dominant in Delhi during winter and pre-monsoon. It is clearly seen that irrespective of constant emissions occurring each month, the estimated pollution is much higher in winter months and lower in summer months. It is concluded that fine particles exhibit much higher variations than urban aerosols annually and seasonally.