Distribution patterns and characterization of outdoor fine and coarse particles

Simultaneous measurements of fine (PM2.5) and coarse (PM10) particles have been performed during three sampling period: summertime, early winter before the heating period, and during the winter heating. The sampling was done at an urban site, Harbin City, China, to apportion pollution sources and to investigate the impacts of weather parameters and emission sources on distribution patterns of PM2.5 and PM10 and their chemical compositions. Two chemometric techniques, hierarchical clustering analysis (HCA) and discriminant analysis (DA), and a receptor model (PMF5) were applied. The average mass concentrations of fine and coarse particles during the whole study period (i.e. annual averages), were 82.4 μg/m3 and 120.1 μg/m3, with relative humidity and wind speed being the main factors influencing distribution patterns of PM2.5 and PM10, while sum of major ions (NO3−, NH4+, SO42−), and organic carbon (OC) were the greater contributor to total mass concentrations. According to DA, the distribution patterns of PM2.5 and PM10 during the winter heating period were distinct but overlapping in summer and early winter before heating, implying that, PM2.5 and PM10 were originated from different sources during the winter heating period but from similar sources during the other two sampling periods. This claim was further supported by PMF results. Canonical discriminant function coefficients of DA have shown species such as OC, EC, SO42−, NO3−, Cl−, Ti, Sr, Ca2+, Ni, and Ba to be good predictors/discriminants responsible for the differences between the three sampling periods. HCA visualized the interrelations among the source markers, while the PMF5 modeling confirmed HCA findings. •PM2.5 and PM10 showed similar distribution pattern but different sources in winter.•In cold zones, biomass burning, coal use along with traffic could have health impacts.•Hierarchical clustering analysis (HCA) visualized the interrelations among tracers.•HCA agreed with results of Positive Matrix Factorization (PMF).•Linear discriminant analysis, is suggested to be a useful tool for aerosols' studies.