Relationship of extinction coefficient, air pollution, and meteorological parameters in an urban area during 2007 to 2009

Light extinction, which is the extent of attenuation of light signal for every distance traveled by light in the absence of special weather conditions (e.g., fog and rain), can be expressed as the sum of scattering and absorption effects of aerosols. In this paper, diurnal and seasonal variations of the extinction coefficient are investigated for the urban areas of Tehran from 2007 to 2009. Cases of visibility impairment that were concurrent with reports of fog, mist, precipitation, or relative humidity above 90 % are filtered. The mean value and standard deviation of daily extinction are 0.49 and 0.39 km^?1, respectively. The average is much higher than that in many other large cities in the world, indicating the rather poor air quality over Tehran. The extinction coefficient shows obvious diurnal variations in each season, with a peak in the morning that is more pronounced in the wintertime. Also, there is a very slight increasing trend in the annual variations of atmospheric extinction coefficient, which suggests that air quality has regressed since 2007. The horizontal extinction coefficient decreased from January to July in each year and then increased between July and December, with the maximum value in the winter. Diurnal variation of extinction is often associated with small values for low relative humidity (RH), but increases significantly at higher RH. Annual correlation analysis shows that there is a positive correlation between the extinction coefficient and RH, CO, PM₁₀, SO₂, and NO₂ concentration, while negative correlation exists between the extinction and T, WS, and O₃, implying their unfavorable impact on extinction variation. The extinction budget was derived from multiple regression equations using the regression coefficients. On average, 44 % of the extinction is from suspended particles, 3 % is from air molecules, about 5 % is from NO₂ absorption, 0.35 % is from RH, and approximately 48 % is unaccounted for, which may represent errors in the data as well as contribution of other atmospheric constituents omitted from the analysis. Stronger regression equation is achieved in the summer, meaning that the extinction is more predictable in this season using pollutant concentrations.