Apportionment of black and brown carbon spectral absorption sources in the urban environment of Athens, Greece, during winter.

Abstract This study examines the spectral properties and source characteristics of absorbing aerosols (BC: Black Carbon; BrC: Brown Carbon, based on aethalometer measurements) in the urban background of Athens during December 2016–February 2017. Using common assumptions regarding the spectral dependence of absorption due to BC (AAEBC = 1) and biomass burning (AAEbb = 2), and calculating an optimal AAEff value for the dataset (1.18), the total spectral absorption was decomposed into five components, corresponding to absorption of BC and BrC from fossil-fuel (ff) combustion and biomass burning (bb), and to secondary BrC estimated using the BC-tracer minimum R-squared (MRS) method. Substantial differences in the contribution of various components to the total absorption were found between day and night, due to differences in emissions and meteorological dynamics, while BrC and biomass burning aerosols presented higher contributions at shorter wavelengths. At 370 nm, the absorption due to BCff contributed 36.3% on average, exhibiting a higher fraction (58.1%) during daytime, while the mean BCbb absorption was estimated at 18.4%. The mean absorption contributions due to BrCff, BrCbb and BrCsec were 6.7%, 32.3% and 4.9%, respectively. The AbsBCff,370 component maximized during the morning traffic hours and was strongly correlated with NOx (R2 = 0.76) and CO (R2 = 0.77), while a similar behavior was seen for the AbsBrCff,370 component. AbsBCbb and AbsBrCbb levels escalated during nighttime and were highly associated with nss-K+ and with the organic aerosol (OA) components related to fresh and fast-oxidized biomass burning (BBOA and SV-OOA) as obtained from ACSM measurements. Multiple linear regression was used to attribute BrC absorption to five OA components and to determine their absorption contributions and efficiencies, revealing maximum contributions of BBOA (33%) and SV-OOA (21%). Sensitivity analysis was performed in view of the methodological uncertainties and supported the reliability of the results, which can have important implications for radiative transfer models.