Comparison of PM₁₀ Sources Profiles at 15 French Sites Using a Harmonized Constrained Positive Matrix Factorization Approach

Round 1

Reviewer 1 Report

The work is interesting, scientifically sounding, and well presented. A number of advanced statistical techniques are also employed to enable robustness of the achieved results, which were fully addressed and compared with similar studies in the literature. The study fits the scopes of the Journal and could be accepted after the following amendments have been implemented.

Major Remarks

In my opinion, paper's conclusions should be further elaborated in order to better highlight main findings achieved in the framework of the SOURCES project. For example, in addition to highlight main PM10 source apportionment from an overall perspective, this could be highlighted by seasonality, type of site, and regional scale (as addressed within the Results and Discussions section). Furthermore, future research directions and possible work's perspectives could be included at the end of the Conclusions.

Minor Remarks / Typos

The Abstract needs to be shortened as it is too long.

Table 1. “0°35'17.01" O” should likely be corrected to “0°35'17.01" W”.

In Table 1, it could be more helpful to include elevation above mean sea level for each site.

P5.L92. “S/N‘0.2” should likely be corrected to “S/N<0.2”.

P5.L100. “were applied” should be corrected to “was applied”.

P6.L117. “report” should likely be corrected to “ratio”.

P7.L153. “sulfhate” should be corrected to “sulfate”.

Fig. 2. In the caption, remove the first “(” before “in μgm-3”.

P15.L352. The comma after “despite” should be removed.

 Suppl. Mat. Fig. SI2. In the caption, the font should be changed from bold to normal text.

Author Response

We greatly thank you for your quick reading and very valuable comments. As you suggested, the conclusion has been modified to better highlight the important results of the study. We also give explicitly a possible future work as an opening. In this respect, the conclusion now reads as follows:

This paper presents results obtained from PMF analyses conducted in a harmonized way and taking advantage of datasets corresponding to more than 2200 samples collected at 15 different sampling sites in France. To the best of our knowledge, this is the first study using such a large database over several years of monitoring. Thirty-seven input variables have been selected to cover a wide range of major expected PM₁₀factors and to optimize total mass reconstruction at all sites. The operator-related subjectivity is reduced by using homogeneous statistical and geochemical criteria for the source discrimination and identification.

Overall, the main PM₁₀factors are found to be primary traffic (15 ± 7 % of the PM₁₀mass on a yearly average) and biomass burning (17 ± 9 %) as well as two secondary aerosol fractions dominated by ammonium nitrate (17 ± 8 %) and ammonium sulfate (15 ± 4 %), respectively. These findings are in good agreement with what has been commonly observed by many previous studies over Europe and illustrates –if still needed– the major impact of anthropogenic emissions on urban air quality. Nevertheless, substantial contributions of mineral dusts (13 ± 4 %), fresh (6 ± 2 %) and/or aged sea-salts (8 ± 4 %), as well as of primary biogenic aerosols (7 ± 3 %) are also observed at almost all sites, reflecting that natural emissions should not be neglected as other important contributors to PM₁₀levels on a yearly basis.

During wintertime, high contribution of the biomass burning is observed, together with the nitrate-rich factor in early spring (March). The sea-salt also presents higher contribution during this period of the year, potentially linked to stronger westerlies winds. In summer, all sites present important contributions from biogenic sources: the primary biogenic and marine SOA factors. Conversely, the primary traffic, dust, aged sea-salt and sulfate rich factors do not present clear seasonality pattern.

Uncertainties of each PMF results were investigated thanks to the bootstrap and displacement methods, both for the base and the constrained runs. The use of a set of constraints –taken as limited as possible– allows to decrease these uncertainties and to strengthen the output statistical representativity of factors commonly observed all over the French continental territory. The internal variability (i.e. for a given site) assessment shows high confidence for the concentration of the main proxies of sources (i.e. specific tracers) and for the PM₁₀mass apportionment, but larger uncertainties are found for concentrations of non-tracer species and caution should be taken concerning the contributions of such chemical species to the different factors.

Finally, the external variability (i.e. site between sites) was investigated thanks to both the SID and PD metrics. In particular, the discrepancy between the chemical profiles of sources could be quantified thanks to both the SID and PD metrics, notably showing that fresh sea salt, nitrate-rich, sulfate-rich and biomass burning factor profiles can be considered as chemically similar at all the sites studied. Conversely, some significant variabilities are observed for other factor chemical profiles, such as the ones related to traffic and industrial emissions as well as mineral dusts and marine SOA. The latter finding is calling for the use of additional variables, such as organic molecular markers, within studies aiming at scrutinizing specific local sources at individual sites and/or for a better knowledge on the origins of SOA fractions.

Overall, results obtained in this comprehensive study may help to increase the completeness of the SPECIEUROPE database, and can be used to assess the contribtutions of the different sources to the oxidative potential of PM following methodology proposed by Weber et al. [49].

Conversely, and as suggested, the abstract has been shortened, deleting both following sentences:

Furthermore, all the outcomes of the project, including time- series and source profiles along with their corresponding uncertainties, have been made available through a website interface, proposed as supporting material for this article, at https://pm-sources.fr. This methodology offers a unique opportunity to compare main PM factor’s contributions and chemical footprints at a national scale.

We also corrected each typo following your suggestions (please see revised version). Thank you very much again for your careful review.

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript studied the PM₁₀ source profile across the nation and the information is valuable for air quality control and health studies as well. Overall, the manuscript is well written with detailed information on data and methodology and thorough analysis of the results. I have one comment: For the Introduction section, the authors may provide a summary of previous work that studied PM₁₀ profiles using PMF and how this work fits into the big picture, such as how this work is different from or improves on previous work. The SOURCES project is mentioned in Section 2 of Material and methods, one suggestion is to mention the SOURCES project in the introduction along with the objective of this project. My point is to better show the significance of this work. 

Author Response

We thank you very much for your reading and comment, that has been addressed in the revised manuscript. In this respect, the introduction has been reworded to emphasize previous studies as well as the SOURCES project in this section.

In particular, the following paragraph has been added, when presenting the SPECIEUROPE database:

This database could be implemented thanks to various recent PMF studies, such as those described in Karagulian et al. [1] Belis et al. [2], Viana et al. [3], Mooibroek et al. [4], Perrone et al. [5] and reference therein. However, the robustness of these source profiles needs to be assessed more deeply and outputs from additional source apportionment studies are still needed to increase the European coverage and to document various aerosol fractions and/or geographical areas.

And the SOURCES project has been introduced as follows:

In particular, a very limited number of comprehensive PM_2.5 and/or PM₁₀ source apportionment studies were conducted in France before 2012 [6–9]. To fill this gap, the research community along with regional monitoring networks put common efforts into extensive filter samplings, off-line chemical analyses and statistical data treatments for numerous French locations [10]. Moreover, the SOURCES project has been set-up to delivering a comprehensive overview of these filter-based source apportionment studies achieved in France for the 2012-2016 period [11].

Finally, the very first part of the “Material and methods” section has been changed accordingly (please see revised manuscript), and the following references have been added:

Karagulian, F.; Belis, C.A.; Dora, C.F.C.; Prüss-Ustün, A.M.; Bonjour, S.; Adair-Rohani, H.; Amann, M. Contributions to cities’ ambient particulate matter (PM): A systematic review of local source contributions at global level. Atmos. Environ. 2015, 120, 475–483.

Belis, C.A.; Karagulian, F.; Larsen, B.R.; Hopke, P.K. Critical review and meta-analysis of ambient particulate matter source apportionment using receptor models in Europe. Atmos. Environ. 2013, 69, 94–108.

Viana, M.; Kuhlbusch, T.A.J.; Querol, X.; Alastuey, A.; Harrison, R.M.; Hopke, P.K.; Winiwarter, W.; Vallius, M.; Szidat, S.; Prévôt, A.S.H.; et al. Source apportionment of particulate matter in Europe: A review of methods and results. J. Aerosol Sci. 2008, 39, 827–849.

Mooibroek, D.; Staelens, J.; Cordell, R.; Panteliadis, P.; Delaunay, T.; Weijers, E.; Vercauteren, J.; Hoogerbrugge, R.; Dijkema, M.; Monks, P.S.; et al. PM₁₀ Source Apportionment in Five North Western European Cities—Outcome of the Joaquin Project. In Airborne Particulate Matter; 2016; pp. 264–292.

Perrone, M.G.; Zhou, J.; Malandrino, M.; Sangiorgi, G.; Rizzi, C.; Ferrero, L.; Dommen, J.; Bolzacchini, E. PM chemical composition and oxidative potential of the soluble fraction of particles at two sites in the urban area of Milan, Northern Italy. Atmos. Environ. 2016, 128, 104–113.

El Haddad, I.; Marchand, N.; Wortham, H.; Piot, C.; Besombes, J.-L.; Cozic, J.; Chauvel, C.; Armengaud, A.; Robin, D.; Jaffrezo, J.-L. Primary sources of PM_2.5 organic aerosol in an industrial Mediterranean city, Marseille. Atmospheric Chem. Phys. 2011, 11, 2039–2058.

El Haddad, I.; Marchand, N.; Temime-Roussel, B.; Wortham, H.; Piot, C.; Besombes, J.-L.; Baduel, C.; Voisin, D.; Armengaud, A.; Jaffrezo, J.-L. Insights into the secondary fraction of the organic aerosol in a Mediterranean urban area: Marseille. Atmospheric Chem. Phys. 2011, 11, 2059–2079.

Bressi, M.; Sciare, J.; Ghersi, V.; Mihalopoulos, N.; Petit, J.-E.; Nicolas, J.B.; Moukhtar, S.; Rosso, A.; Féron, A.; Bonnaire, N.; et al. Sources and geographical origins of fine aerosols in Paris (France). Atmospheric Chem. Phys. 2014, 14, 8813–8839.

Waked, A.; Favez, O.; Alleman, L.Y.; Piot, C.; Petit, J.-E.; Delaunay, T.; Verlinden, E.; Golly, B.; Besombes, J.-L.; Jaffrezo, J.-L.; et al. Source apportionment of PM₁₀ in a north-western Europe regional urban background site (Lens, France) using positive matrix factorization and including primary biogenic emissions. Atmospheric Chem. Phys. 2014, 14, 3325–3346.

Amodeo, T.; Favez, O.; Jaffrezo, J.-L. Programmes de recherche expérimentaux pour l’étude des sources de PM en air ambiant; INERIS, 2017 ; 76 pages.

Favez, O.; Jaffrezo, J.L.; Salameh, D.; Amodeo, T. Etat des lieux sur les connaissances apportées  par les études expérimentales des sources de particules fines  en France - Projet SOURCES.; ADEME, 2017; 135 pages.

Author Response File: Author Response.pdf