Disparities in particulate matter (PM ₁₀ ) origins and oxidative potential at a city scale (Grenoble, France) – Part 1: Source apportionment at three neighbouring sites

Abstract. A fine-scale source apportionment of PM10 was conducted in three different urban sites (background, hyper-center, and peri-urban) within 15 km of the city in Grenoble, France using Positive Matrix Factorization (PMF 5.0) on measured chemical species from collected filters (24 h) from February 2017 to March 2018. To improve the PMF solution, several new organic tracers (3-MBTCA, pinic acid, phthalic acid, MSA, and cellulose) were additionally used in order to identify sources that are commonly unresolved by classic PMF methodologies. An 11-factor solution was obtained in all sites, including commonly identified sources from primary traffic (13 %), nitrate-rich (17 %), sulfate-rich (17 %), industrial (1 %), biomass burning (22 %), aged sea salt (4 %), sea/road salt (3 %), and mineral dust (7 %), and the newly found sources from primary biogenic (4 %), secondary biogenic oxidation (10 %), and MSA-rich (3 %). Generally, the chemical species exhibiting similar temporal trends and strong correlations showed uniformly distributed emission sources in the Grenoble basin. The improved PMF model was able to obtain and differentiate chemical profiles of specific sources even at high proximity of receptor locations, confirming its applicability in a fine-scale resolution. In order to test the similarities between the PMF-resolved sources, the Pearson distance and standardized identity distance (PD-SID) of the factors in each site were compared. The PD-SID metric determined whether a given source is homogeneous (i.e., with similar chemical profiles) or heterogeneous over the three sites, thereby allowing better discrimination of localized characteristics of specific sources. Overall, the addition of the new tracers allowed the identification of substantial sources (especially in the SOA fraction) that would not have been identified or possibly mixed with other factors, resulting in an enhanced resolution and sound source profile of urban air quality at a city scale.