This manuscript is focused on the source apportionment of the oxidative potential (OP) of atmospheric PM₁₀ from a long-term large-scale sampling campaign in France. More than 1700 samples from 14 sites and 15 years sampling period were analyzed for OP in two endpoints (OP^AA and OP^DTT) and various chemical species. The authors implemented a positive matrix factorization (PMF) coupled with multiple linear regression (MLR) protocol for identifying the sources from individual sites contributed to measured OP. The importance of different sources was discussed based on both intrinsic OP (OPm) and their contributions to the total OPv. The authors also attempted to explain the seasonality of certain sources in the context of OP. Overall, the manuscript presented very interesting and significant results of PM₁₀ OP in West Europe area, and strongly supported the health effect study and legislation of air quality control in France. However, it lacks a discussion on OP’s spatial distribution and the linkages of chemical composition to OP. Following are my specific comments.

Specific comments:

1. Generally, fine particles (PM₅) have been linked to the adverse effect caused to human respiratory and cardiovascular system, since these particles have highest penetration efficiency in lower respiratory tract. However, the PM samples collected in this manuscript have involved larger sized particles which have shown lower oxidative potential (Hu et al., 2008;Ntziachristos et al., 2007) and somewhat lesser relevance to human health. Thus, a justification should be provided for using PM₁₀ in this health-related study, i.e. source apportionment of PM OP.
2. Lines 85 – 89: Please provide more details on the geographical information of all sites, e.g. distance to local highways, industries and/or coast, wind direction and relative location in the valley/peak of Alpine area etc. This can help to directly identify some potential sources.
3. Line 119: Why 1,4-naphthoquinone was chosen as the positive control for AA? Does it have a consistent and high AA activity?
4. The apportionment of SOA factors in this paper is ambiguous. The authors did not explain the contributions of OC in three highlighted SOA-related factors, i.e. nitrate-rich SIA, sulfate-rich SIA and MSA-rich. We suggest a better explanation should be provided on their formation and the differences of their contributions to OP and mass of PM₅.
5. The subsections in Section 3.3 are confusing. I suggest combining Sections 3.3.2 – 3.3.5 to a single subsection “Intrinsic OP of main PMF sources”, and create a new subsection for Sections 3.3.6 – 3.3.12 (like Section 3.4), “Profiles of OPm sources”.
6. Lines 367 – 370: The authors tend to tone down the intrinsic OP of MSA-rich factor. In fact, I found very high OP^DTTm for this factor at GRE-cb, NGT and STG-cle with reasonable CV (<0.6), while OP^AAm at these sites were near 0. I would suggest the authors to further explore the redox-active components and explore the reasons for different activities between two endpoints for this factor at these sites.
7. Lines 391 – 395: Please specify the factors with low variability of OPm and explain why the other factors show large variability.
8. Lines 410 – 415: the authors should highlight the high contribution of sulfate-rich SIA factor to OP^DTT In comparison to its marginal contribution to OP^AAv, it is also important to find out the ROS-active components in this factor for the difference between two OP endpoints.
9. Lines 421 – 425: The explanations on the mean and median values of contribution of sources are unclear. The authors should explain the meaning of these two values and justify why using them under different circumstances.
10. Lines 473 – 476: I suggest the authors further investigate the sensitive chemicals (like transition metals) to both OP endpoints at all the sites with the industrial factor, in order to explain the huge difference of OPm found among all six sites.
11. Line 507: A considerable contribution from sulfate-rich factor has been noted in Figure 4 – 6. Hence, the author made a wrong statement here saying both SIA sources contribute barely to OP which should be corrected.
12. Line 512: The paper showed many sources with different spatial variability in intrinsic OP, which is not discussed. Therefore, caution should be exercised when making statements like “relatively stable intrinsic OP at a large geographical scale”. The authors should discuss it in two types of sources – sources with low variability and stable intrinsic OP (e.g. road traffic), and sources with high variability and varied intrinsic OP (e.g. biomass burning).

Technical corrections:

1. Line 14: to prioritized – insert “be” between these two words.
2. Line 76: “multilinear” should be changed to “multiple linear” or “multivariate linear”.
3. Line 115: the abbreviation in the parenthesis should be “RTLF”.
4. Line 174, 237, 257: “specie” should be changed to “species”.
5. Line 266, 459: The word “inversion” in the title is confusing. I suggest replacing “OP’s inversion” with “intrinsic OP”.
6. Line 431: I suggest moving “in Figure 5” to the end of the sentence or include it in parenthesis.
7. Legend of Figures 5 and 6: the error bars __ the 95% confidence level – insert “represent” at the underline.
8. Legend of Figure A2: “Searman” should be changed to “Spearman”.

References

Hu, S., Polidori, A., Arhami, M., Shafer, M., Schauer, J., Cho, A., and Sioutas, C.: Redox activity and chemical speciation of size fractioned PM in the communities of the Los Angeles-Long Beach harbor, Atmospheric Chemistry and Physics, 8, 6439-6451, 10.5194/acp-8-6439-2008, 2008.

Ntziachristos, L., Froines, J. R., Cho, A. K., and Sioutas, C.: Relationship between redox activity and chemical speciation of size-fractionated particulate matter, Particle and fibre toxicology, 4, 5, 2007.

In this manuscript, the authors presented source apportionment results of OP-DTT and OP-AA measured on PM₁₀ filter samples collected at 14 different locations in France using PMF and multiple linear regression models. The authors mainly focus on discussing the intrinsic OP, the variability of different sources, and the daily mean and median contribution of sources to OP. The limitation of the study is well discussed. This study has a unique dataset. However, in my opinion, the authors did not fully utilize their dataset. One of the uniqueness of this study is that the dataset spans 15 years and covers a wide range of environments. Seasonal variations of OP-DTT and OP-AA are discussed but the authors may consider looking into other aspects that are more interesting, for example, the spatial homogeneity or the historical changes of OP. How do the OP sources and source contributions change over the 15-year period? Are there differences in the historical trends of OP-DTT and OP-AA and how does it compare to PM mass? What chemical components are the most important drivers to the changes of OP and PM mass? It is well known that biomass burning, traffic emissions, secondary processing are important OP sources. Presenting something other than sources would enhance the scientific significance of this manuscript. Below are my comments:

Major comments:

1. Please provide detailed protocols for both the DTT and AA assays. Multiple versions of DTT assays are currently used in the community. It is important to show which DTT protocol was used in this work. Perhaps the most important is the initial concentration of DTT as studies have shown that the initial DTT concentrations can have a large impact on the DTT consumption rates e.g. (Lin and Yu 2019).
2. Line 110, particles removed from the filters were added to 96-well plates for DTT analyses, and the authors claimed that this included “soluble and insoluble” particles. What is the extraction efficiency? Could there be particles that are not extractable and attached to the filters. Other studies that measured total DTT run the extract along with the filter in DTT solutions. E.g. (Gao, Fang et al. 2017). A note should be added to emphasize the differences in the protocol and state that the DTT activities may not be “total”.
3. In Figure 3, OP-AA from road traffic, biomass burning, dust and OP-DTT from biomass burning and dust sources are bi-model distributions. Why? It would be interesting to look into details in chemical components to figure out the observed distributions.
4. Section 3.3.6-3.3.11, I appreciate the authors’ efforts in discussing the variability of the intrinsic OP. However, it is not clear to me what do variabilities of different OP sources really bring about. It seems more interesting to compare the intrinsic OP or the contribution of different sources to OP with those from other studies or those from other regions of the world. These subsections lack in-depth discussion on each source. For example, for road traffic, transition metals (non-exhaust) and quinones from PAHs or soot (exhaust) can contribute to OP-DTT and OP-AA. How does your source profile from road traffic differ from others? What are the linkages of traffic-related chemical components to measured OP? What new insights does this work bring? One interesting questions is whether you can differentiate the contribution of exhaust and non-exhaust emissions to OP.
5. It would be useful to present how are metals (especially Fe and Cu) apportioned into each factor. Atmospheric metals can be found in biomass burning, road traffic, dust, or sulfate particles by acid processing. Discussions on how metals are distributed in these factors may help to interpret the contribution of different sources to OP.
6. It is not clear that what are the new findings in this manuscript compared to previous work from the same group, e.g. (Weber, Uzu et al. 2018).

Other comments:

1. Line 257, “organic specie” should be “organic species”
2. Line 258, define “HULIS”
3. Use the same font type throughout, for example, line 464, numbers seem to be a different font than other contents

References:

Gao, D., T. Fang, V. Verma, L. Zeng and R. Weber (2017). "A method for measuring total aerosol oxidative potential (OP) with the dithiothreitol (DTT) assay and comparisons between an urban and roadside site of water-soluble and total OP." Atmos. Meas. Tech. Discuss. 2017: 1-25.

Lin, M. and J. Z. Yu (2019). "Dithiothreitol (DTT) concentration effect and its implications on the applicability of DTT assay to evaluate the oxidative potential of atmospheric aerosol samples." Environmental Pollution 251: 938-944.

Weber, S., G. Uzu, A. Calas, F. Chevrier, J. L. Besombes, A. Charron, D. Salameh, I. Ježek, G. Močnik and J. L. Jaffrezo (2018). "An apportionment method for the oxidative potential of atmospheric particulate matter sources: application to a one-year study in Chamonix, France." Atmos. Chem. Phys. 18(13): 9617-9629.