This manuscript reports and discusses the source apportionment of organic and inorganic aerosols in Fairbanks, Alaska during winter, as measured by HR-ToF-AMS and CHARON-PTR-ToF-MS. Fairbanks is of particular interest as during winter, it frequently has poor air quality due to enhanced biomass burning emissions (residential heating) combined with poor air dispersion caused by temperature inversions. The focus of the paper is on comparison of the source apportionment results from the AMS and CHARON-PTR data, emphasizing their complementary nature.  While AMS data have been  extensively used for source apportionment, that is not the case for CHARON-PTR or even PTR-MS. and To my knowledge, there is only one (very recent) source apportionment study that uses both CHARON-PTR and AMS data. The paper by Ijaz et al provides a very detailed source apportionment study and it will serve as a useful resource for future source apportionment studies using both CHARON-PTR and AMS data. It also demonstrates the ability of CHARON-PTR, due to its molecular level characterisation of OA, to distinguish additional residential heating sources. This is something that would not be possible to identify based only on the AMS data. Based on the arguments outlined above I recommend this paper for publication in the ACP.

Below are my specific comments (most of which I would describe as minor).

Line 120: I suggest “molecular level information” instead of “good molecular resolution”

Line 134 Please use sulfate instead of sulfur.

Line 174: what was m/z range?

Line 188: no need for “PeTeR version 6.0…” as it has already been mentioned

Line 195: As AMS’ come with either capture or standard vaporiser, it is worth mentioning which one you had.

Line 199: ..nitrate, ammonium, sulfate, chloride.

Line 208-9: nitrate, sulfate, chloride (please use singular for these ions).

Line 2016: why max m/z was only 120; why were PAHs excluded?

Line 220: not needed to mention that empty rows and columns were removed (same for line 188).

Line 271: have you tried to look at the correlation between aerosol loadings and WS for the periods with strong temperature inversion? This would be a nice visual confirmation of your statement.

Line 272: low instead of slow

Line276: daily average instead of campaign-averaged?

Line 284: sulfate instead of sulfur. I am assuming the authors chose to use sulfur as species other than SO₄^2-  can result in fragments that are used to calculate the overall SO₄^2- mass loadings (e.g. organosulfates). The choice of naming should be either explained in the Exp section, or use sulfate and discuss the contribution of organosulfates to SO₄^-2 later in the text (page 15). Note that Fig 1 has SO4, not S.

Line 318: fragmentation increases dramatically when?

Line 359: Considering that ResH1_4  “closely co-varied in time”, how strongly do they correlate with each other? Could they all, or some of them be a part of the same factor (i.e. be a result of factor splitting)?

Line 369: what is in the same order? I assume peak intensity?

Line 385: atmospheric processing. Could it be that ResH1 is related more to low temperature combustion?

Line 483: CHARON can see species that can evaporate at 150oC (I am assuming this is the thermodenuder temp.; please mention operating TD temperature in the Experimental section); there are also species that will not be efficiently ionised by proton transfer. This should also be mentioned as potential reasons for the observed discrepancy.

how do overall mass loadings of AMS Org and Charon Org compare?  Are there any other studies reporting concurrent measurements by CHARON and AMS? How did their mass loadings compare?  And doing source apportionment? If yes, you did their factor loadings compare?

Line 513: How well do OOA and BBOA correlate with each other?

Line 524: I think it should be the other way around 0.4 and 0.2%

Line 591: remove “of”

Line630: If the goal of this discussion and Fig 8 + S18-21 was to see which factor “responds” the most to ADEC advisories, then looking at relative contributions of ResH factors could be misleading as a decrease in absolute concentrations of only one factor will impact rel. contributions of all factors.

Figure 1 B) Personally I find it hard to make distinction between WS and WD colours on the graph. How was J(NO2) calculated? I do not think that is mentioned anywhere. E) I do not think Fig1E is discussed anywhere in the text. How is PM2.5 number concentration measured?

Figure 2: if data points are sized, then size legend should be included

All mass spectra figures: considering CHARON resolution of around 5000, not more than 2 decimal places should be given for m/z.

Figure 4 and 5: I suggest to combine these two figures to look the same as Fig 3 and put the last column in figure 5 in the supplement.

Figure 7A: on y-axis next to HR-ToF AMS put the ion in question.

Table 1: I suggest moving it in the supplement

Supplementary:

S3: I am assuming the sentence “ This introduces a sust the campaign, and so must the EF” is there by mistake

S5 and S6: Can you please briefly justify the number of bootstrap replicates for both PTR(CHARON) and AMS factorisation?

Figure S3: Why is not starting from 15nm?

Figure S5: why should scaled residuals ideally be +/- 2?

General comment: the font in most of the figures (axis numbers, labels,legend) is too small. Please consider increasing the font.

The manuscript titled "Complementary aerosol mass spectrometry elucidates sources of wintertime sub-micron particle pollution in Fairbanks, Alaska, during ALPACA 2022 " uses aerosol mass spectrometry techniques, including CHARON PTR-ToF MS and HR-AMS, to analyze non-refractory particulate matter and its sources. Positive Matrix Factorization was applied to identify and estimate the contributions of various pollution sources, with a particular focus on residential heating emissions, on-road traffic, and secondary aerosol formation. The findings highlight the significant role of residential heating, alongside secondary organic aerosol processes, notably organosulphur, on air quality.

These observations offer valuable insights into this subarctic region, which is prone to significant particulate pollution events during winter. However, I have several concerns regarding the manuscript’s structure, novelty, and methodology, which I outline in the general comments. A list of specific comments follows thereafter.

Structure and presentation:

1. The text is at times lacking conciseness and its structure could be optimized. In some instances, a potential issue is identified, but its possible causes are only discussed several paragraphs later, making the argumentation harder to follow. For example, the relatively low CHARON_traffic mass mentioned in L.486 is its potential causes debated only a few paragraphs later in the text, on the following page.

2. The inclusion of three PMF analyses for two instruments is not easy to follow, particularly due to the way the acronyms are introduced. The results section lacks clarity, as multiple PMF analyses are presented in a fragmented manner. Moreover, the added value of applying PMF to both AMS OA and OA+Inorganics is unclear. Upon reading, one has the feeling that PMF AMS organics could be removed from the manuscript without any meaningful loss while improving readability.

3. I don’t find the figures particularly suitable for publication in their current form. Their layout, font sizes and colour choices should be refined to meet article-level standards.

Novelty

4. The authors repeatedly justify in the abstract and introduction that “significant uncertainties exist about aerosol sources, formation, and chemical processes during cold winter” in Fairbanks, Alaska (e.g. L.31). However, an extensive body of literature spanning over a decade has already attributed residential heating as a major driver of air quality degradation in that region, as well studies focusing on sulphur-containing particles. The manuscript should more precisely articulate the specific knowledge gaps that remain unresolved. Conversely, care must be taken not to extrapolate its findings to other arctic regions given the localized representativeness of their findings (e.g. L.138).

5. There appears to be little “complementarity” between CHARON and HR-AMS analyses in providing new insights into wintertime PM levels. CHARON independently proposes a novel separation of residential heating sources (discussed further below), while HR-AMS produces results that are largely consistent with prior literature (traffic, residential heating, and OS as key winter contributors). Some of these findings have already been reported in studies from related or the same field campaigns (e.g., Campbell et al., 2022; Robinson et al., 2024; Yang et al., 2024). Given the study’s title, I expected some type of combined analysis of CHARON and HR-AMS to yield results beyond the simple sum of their individual analysis, but this expectation is not met.

Methodology:

6. The PMF on CHARON data is difficult to verify, notably with this complex 8-factor solution. As far as I can gather, I am not completely convinced that the four residential heating factors are well-separated. This is often an issue for co-varying factors, which is the case here where Pearson is between 0.6 and 0.75. Furthermore, there seems to be some issues among the four residential heating factors. For example, when normalized to their OA contributions, ResH2 (attributed here to hardwood combustion) and ResH3 (attributed to heating oil combustion) have the same relative fraction of levoglucosan—a compound that is a unique tracer of cellulose pyrolysis. Furthermore, all four factors exhibit a fairly high and comparable correlation with SO₂, while ResH1 “mixed wood burning” even shows a stronger correlation with "sulphate" than heating oil combustion, discussed here and previous references to be the main source of sulphur dioxide. I’ll refrain from commenting on the separation between softwood vs. hardwood emissions, i.e. factors ResH2 and ResH4, as I am not an expert on the topic, but it also seems to be at least somewhat debatable. I believe this needs to be strengthened. As I discussed above, the separation of residential heating sources is a key novelty aspect of this work.

7. L.216-219: It’s unclear why mz>120 was used for OA calculation but removed from PMF analysis, notably when co-emission of PAH is expected with factors such as BBOA and HOA. Their use as “external tracers” raises some questions about the robustness of the analysis. Please provide arguments to justify this (as well as re-phrase sentence for clarity).

Specific comments 

1. Please rework the abstract providing a more quantitative view of the results presented here, highlighting their novelty. 
2. L. 35: “Which” instead of “that”.
3. L101: remove “in detail”
4. L111: Do you mean “more selective”? Or just selective, since it depends on proton affinity.
5. L.120: “Information on NR-PM1 OA” or something along those lines.
6. L.127: “Not well understood”.
7. L134: I’d advise maintaining sulphate, or rewrite “chloride-, nitrogen- and sulphur-containing species”. Albeit well known by the community (and underlying current understanding of sulphate at Fairbanks), sulphur is the element, not the species or the aerosol type.
8. L.134: “What does “good mass resolution” mean? ToF-ACSM would be enough? V-mode? W-mode? I suggest being more accurate here, notably on the role of mass resolution in the findings of the manuscript.
9. L.150: “recorded”
10. L.182: What is the validity of Leglise fragmentation correction into Fairbanks OA?
11. L.199: RIE for sulphate was 1.93. Is that consistent with previous characterizations of the instrument? Has estimated SO4 from HR-AMS been compared with other observations?
12. L.203: Confirm that the CDCE algorithm on PIKA calculated a correction factor down to 0.35.
13. L.237: please develop “to understand the data” for clarity.
14. L.239-242: Re-phrase for improved readability.
15. L.272: the correct SI notation for the second is “s” and not “sec”.
16. L.274: Indicate what the variability range stands for. 
17. L.276: remove “campaign”.
18. L.276-L.278: Methodology section, on ancillary observations?
19. L.291-298: I find this paragraph somewhat unclear, albeit highly relevant. Is 9% of the CHARON mass attributed to heteroatomic ions, being roughly 7% oxygenated ions (CHO) and the rest ON and OS? Are all OS and ON removed from the source apportionment analysis? Also, it’s unclear what the relevancy of those “prominent peaks” is, they don’t seem to be identifiable in Figure S8.
20. L.278-281: eBC has already been estimated at 15% of PM1, how can PM2.5 be 99% of NR-PM1? There is a quantification issue with one (or both) methods, besides other refractory PM1 species that have not been considered, and the PM1-PM2.5 fraction. The analysis seems to completely miss those basic considerations. 
21. L.304: avoid “could unequivocally be identified”
22. Figure 2: To improve readability I suggest not to colour code against PMF factors that have not yet been presented in the manuscript but some standard tracers (whether levoglucosan from CHARON or more “usual” tracers from AMS like f60 and f55, for example). Otherwise, panes D and E can go into supplementary, where results from different sections can be combined without impacting the flow of the manuscript. 
23. L.308-312: Shouldn’t the size-dependent EF correct for that effect?
24. L.316-318: Has total OA been observed to decrease during laboratory experiments in the lab or just the concentration of those particular species? Fragmentation does not necessarily induce OA mass loss, it will of course depend on the PA of the fragments.
25. L.335: Clarify PMF of OA and OA+inorganics. Given that OA+inorganics analysis has seen added one factor for NO3 and one for SO4, it’s not unexpected that the main source of OA has remained unchanged.  
26. L.346: It’s curious about the low correlation with BC for such a dominating OA source, has it been corroborated also on previous studies for this site? Is there an explanation for the complete lack of correlation with CO?
27. L.374: Rephrase the analysis (also L.369), where 14% (ResH2) of LEV is considered robust and 9% (ResH3) minor.
28. L.412: present the relative contribution in the form of an equation.
29. L.460: The journal and DOI of this reference are lacking, I could not find it.
30. L.466: Please include already here possible explanations for discrepancies on mass loadings between HOA from AMS and traffic factor from CHARON from page 14. 
31. L.475-477: Rephrase “reliable tracer for it is yet to be identified” as a direct sentence.
32. S3: correct “sust”
33. Fig. S3: Most of the EF calculated is above 7, which if I read correctly implies a volume-weighted distribution generally >200nm for submicrometric aerosols. Is that correct? 

References

Campbell, J. R., Battaglia, M. J., Dingilian, K., Cesler-Maloney, M., St Clair, J. M., Hanisco, T. F., Robinson, E., DeCarlo, P., Simpson, W., Nenes, A., Weber, R. J., and Mao, J.: Source and Chemistry of Hydroxymethanesulfonate (HMS) in Fairbanks, Alaska, Environ. Sci. Technol., 56, 7657–7667, https://doi.org/10.1021/acs.est.2c00410, 2022.

Robinson, E. S., Battaglia, Jr, M., Campbell, J. R., Cesler-Maloney, M., Simpson, W., Mao, J., Weber, R. J., and DeCarlo, P. F.: Multi-year{,} high-time resolution aerosol chemical composition and mass measurements from Fairbanks{,} Alaska, Environ. Sci. Atmos., 4, 685–698, https://doi.org/10.1039/D4EA00008K, 2024.

Yang, Y., Battaglia, M. A., Mohan, M. K., Robinson, E. S., DeCarlo, P. F., Edwards, K. C., Fang, T., Kapur, S., Shiraiwa, M., Cesler-Maloney, M., Simpson, W. R., Campbell, J. R., Nenes, A., Mao, J., and Weber, R. J.: Assessing the Oxidative Potential of Outdoor PM2.5 in Wintertime Fairbanks, Alaska, ACS ES\&T Air, 1, 175–187, https://doi.org/10.1021/acsestair.3c00066, 2024.

The authors describe analysis of a mass spectrometry dataset collected in Fairbanks, Alaska, aimed at investigating sources of ambient PM₁ that contribute to local poor air quality events resulting in non-attainment of health-based air quality standards. Local meteorology, in particular temperature inversions, lead to accumulation of emissions from local activities including home heating, traffic and cooking. The novelty of this work lies in the additional insights obtained from leveraging the PTR-CHARON molecular compositional data. Most real-time organic aerosol apportionment studies rely on analysis of heavily fragmented EI mass spectra from AMS instruments. While these are useful for mass closure, and for informing air quality policy (eg traffic vs woodburning mass contributions) relatively little information on organic aerosol composition is obtained. In this work, separate, thorough analyses of the AMS organic, AMS organic+inorganic and the PTR-CHARON datasets through PMF are used to maximize the value of each technique. There are also synergistic aspects. For example, the CHARON dataset reveals multiple distinct local residential combustion source profiles featuring marker ions that can be associated with hardwood, softwood, resin and even oil combustion, whereas only a single residential combustion factor (BBOA) is extracted from the AMS data. The identification of specific marker ions (eg furfural, guiacol, eugenol, vanillin, coniferaldehyde) will undoubtedly be useful for CHARON users going forward. On the other hand, mass closure appears to be an issue for the CHARON dataset, which is limited by poor particle transmission at lower sizes and significant fragmentation of alkanes despite the soft ionization used. In the case of this work I don’t believe the mass concentration underestimation of the CHARON for the various factors is too concerning because the AMS functions as a quantitative instrument for closing PM₁ OA. However, I think the conclusions would benefit from a broader discussion on where the authors feel that CHARON systems fit in to future source apportionment efforts that aim to instruct policy. Should an AMS system always be co-deployed for example, to ensure accurate OA mass loading measurements, with the emphasis on the CHARON data to speciate the AMS “factors” in more detail? Is there a potential for the CHARON data to target and reasonably accurately quantify specific air toxics that represent health risks, for example PAHs or quinones? Or does the size-dependent transmission efficiency introduce too much uncertainty in the absence of co-located SMPS measurements? This would be useful for the community. Overall, I find the manuscript to be well written and the analysis is rigorous, and I have only minor comments below.      

Specific comments:

How often was zeroing performed for the CHARON mass spectral signals? Any humidity dependence of instrumental response?

For the PTR, for the species in the gas calibration mix, how well do the reported mixing ratios based on transmission efficiency and dipole moment/polarizability-derived k values agree with the expected mixing ratios?

Are the signals for BTEX observed in the constrained CHARON traffic factor from the condensed fraction of those species in OA or breakthrough of gases from the denuder?

If the residential heating factors co-vary how were they resolved by PMF? Do they have slightly different temporal dependencies? Presumably home heating emissions from all fuel types peak at the same times of day. Or is there a slightly different wind dependence that results in changes in the relative contributions to OA from these different fuel types because of different uses in different parts of the community.

Figure 4: What is the SM factor? Small carboxylic acids? Is the acetone from denuder breakthrough? Does the lack of a diurnal trend mean this is background OA?

Was there an external validation of the AMS total mass concentrations from a gravimetric/nephelometer particle mass monitor?

How is the fragmentation correction performed for the CHARON data? Are smaller fragment ions quantified or is there a fraction from the MH+ signal that is assumed to be “missing”?

The description of the varying enrichment factors is hard to follow. How is it dependent on concentration? Isn’t the EF selected for each hour based on the mass-size mode of the SMPS? Although there seems to be bimodal distributions for some of the residential factors in the SI. Are the EF values relatively low compared to other studies?

Figure 8 and the corresponding figures in the SI are very hard to follow. I think to show how the relative mass contributions of the residential factors changes before, during and after an advisory period, the format in Figure 1D might be useful. If it’s noisy the temporal resolution could be reduced. It’s hard to track all of the box plots and in the SI versions it’s hard to track the temporal order with the curved arrows.

The mass size mode information in Figure S12 is great. Confirms that there are distinct residential combustion sources.  

The externally derived size distributions for the different factors suggest that COA and the traffic factor have similar size distributions. Does that mean that the large difference in the slopes benchmarking against the AMS in Figure 5 (0.016 vs 0.13) is driven almost entirely by composition?