Organic aerosol source apportionment by using rolling positive matrix factorization: Application to a Mediterranean coastal city

Abstract

We investigated the contributions and the evolution of organic aerosol (OA) sources at the Marseille-Longchamp supersite (MRS-LCP, France) based on Time-of-flight Aerosol Chemical Speciation Monitor (ToF-ACSM) measurements of non-refractory PM 1 over a fourteen-month period (1 February – 3 April 2018). The OA source apportionment was performed by positive matrix factorization (PMF) using the novel “rolling window” approach implemented in the Source Finder Professional (SoFi Pro). Here, PMF is performed over a 14-days window moving over the entire OA dataset, in order to account for the temporal variability of the source profiles. Six factors were resolved, including hydrocarbon-like organic aerosol (HOA) which is related to traffic exhausts, cooking-like organic aerosol (COA), biomass burning aerosol (BBOA), less oxidized organic aerosol (LOOA), more oxidized organic aerosol (MOOA) and a new defined source related to the mix between shipping and industrial plumes (Sh-IndOA). While HOA and COA consistently contribute to the total OA with on average 11.2% (ranging between 9.2 and 12.1% over the seasons) and 11.5% (11–12.1%), respectively, BBOA (11.7% on average) shows a larger seasonal variability with 18% in winter and no contribution in summer. BBOA profiles during winter were attributed to fresh biomass burning emissions from domestic heating, and more oxygenated profiles were assigned to regional land and agricultural waste burning for spring and early autumn. Sh-IndOA fraction is estimated to 4.5% (3.7–6.1%) and contributes to the total OA mass concentrations to a minor extent. The secondary organic aerosol (SOA) fraction includes both LOOA with 21.5% (18.8–27.2%) and MOOA with 39.6% (36.8–42.6%). Based on the f44/f43 analysis these sources appeared to be more linked to biogenic influences in summer, whereas the concentrations were associated with oxidized anthropogenic sources (biomass burning and road traffic) for the rest of the year. The investigation of MOOA geographical origins suggests some influence of air masses transported from the Rhône Valley and the west basin of the Mediterranean Sea. • OA source apportionment over 14 months at an urban background site. • Positive matrix factorization using the novel rolling window approach. • A new defined source of OA related to shipping and industrial plumes. • Highest contribution of SOA during every season.