Comprehensive analysis of PM10 in Belgrade urban area on the basis of long-term measurements.

In this study, we investigated the impact of potential emission sources and transport pathways on annual and seasonal PM10 loadings in an urban area of Belgrade (Serbia). The analyzed dataset comprised PM10 mass concentrations for the period 2003-2015, as well as their chemical composition (organic/elemental carbon, benzo[a]pyrene, As, Cd, Cr, Mn, Ni, Pb, Cl(-), Na(+), Mg(2+), Ca(2+), K(+), NO3 (-), SO4 (2-), and NH4 (+)), meteorological parameters, and concentrations of inorganic gaseous pollutants and soot for the period 2011-2015. The combination of different methods, such as source apportionment (Unmix), ensemble learning method (random forest), and multifractal and inverse multifractal analysis, was utilized in order to obtain a detailed description of the PM10 origin and spatio-temporal distribution and to determine their relationship with other pollutants and meteorological parameters. The contribution of long-range and regional transport was estimated by means of trajectory sector analysis, whereas the hybrid receptor models were applied to identify potential areas of concern.