Stefano Rimoldi, Marcus Acton-Bond, Vera Bernardoni, Laura Cadeo, G. Valli, Cristina Colombi, Rosario Cosenza, Manousos-Ioannis Manousakas, Benjamin Chazeau, R. Vecchi
In this work, aerosol absorption and scattering properties were investigated to gather additional information on aerosol types and sources. During a one-year campaign in Milan (Italy), a multi-wavelength aethalometer and a nephelometer were operated in parallel. Scattering coefficients are in general scarcely available at urban sites and this was the very first dataset for Milan and, more generally, for an urban site in the Po valley which is a well-known pollution hot-spot in Europe. Time patterns of the absorption and scattering coefficients throughout the year were investigated and Absorption/Scattering Ångström Exponents (AAE and SAE, respectively) were also computed to relate the optical properties to fossil vs. wood fuel combustion emission sources and to small vs. large particle size. The average daily trends for the absorption coefficients reflected the behaviour of the sources, with peaks in the morning due to traffic emissions observed in all seasons, and high values during the night as a consequence of lower mixing heights and, in winter, of domestic heating emissions, too. To gain knowledge about aerosol types and origin an innovative cluster analysis approach using AAE and SAE led to the identification of different aerosol types; this is an easy-to-implement approach which could be applied in monitoring networks for near-real time aerosol types identification. Moreover, episodes affected by specific aerosol types (i.e., fire event, desert dust air mass transport…) were singled out applying a more refined approach combining five intensive optical parameters. • Scattering and absorption coefficients measured in a European pollution hot-spot. • A novel approach using cluster analysis on AAE and SAE provided aerosol typing. • Fossil fuel, wood burning, and mineral dust aerosol were clearly identified. • A more refined approach previously implemented allowed short episode detection.