the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Mar 2025
Unraveling Arctic submicron organic aerosol sources: a year-long study by H-NMR and AMS in Ny-Ålesund, Svalbard
Abstract. Understanding the chemical composition and sources of organic aerosol (OA) in the Arctic is critical given its importance for particle climate-relevant properties. This study presents a year-long analysis (May 2019–June 2020) of PM1 filter samples collected in Ny-Ålesund, Svalbard. A multi-instrumental approach is employed to characterize the comprehensive chemical composition of PM1, with a specific focus on its water-soluble organic fraction (WSOA) depicted combining proton nuclear magnetic resonance spectroscopy (H-NMR) and high-resolution time-of-flight aerosol mass spectrometry (HR-TOF-AMS), which provide complementary insights on nature and structure of the organic aerosol classes characterizing the bulk OA mixture. Positive Matrix Factorization (PMF) source apportionment identifies consistent OA sources from the H-NMR and AMS datasets, showing a pronounced seasonality in OA contributions. Winter-spring aerosol is dominated by long-range transport of Eurasian anthropogenic pollution (up to 70 %), while summer is characterized by biogenic aerosols from marine sources (up to 44 %), including sulfur compounds, amines, and fatty acids. Occasional summertime high OA loadings are associated with wildfire aerosols enriched in levoglucosan and humic-like substances (HULIS; averagely 27–28 %). Eventually, about 28–40 % of the OA is attributed to an unresolved mixture of extremely oxidized compounds of difficult specific source-attribution. This integrated approach provides valuable insights into the seasonal dynamics of OA sources in the Arctic.
- Preprint
(1703 KB) - Metadata XML
-
Supplement
(2939 KB) - BibTeX
- EndNote
Status: open (until 28 Apr 2025)
-
RC1: 'Comment on egusphere-2025-760', Anonymous Referee #1, 29 Mar 2025
reply
Paglione et al. present a detailed analysis of PM1 filter samples collected over the course of a year at Ny-Ålesund, Svalbard. Through offline analysis via NMR, AMS, and other techniques, the authors characterize the water soluble fraction of PM1. The AMS and NMR datasets independently yielded similar PMF factors related to marine OA (further resolved as POA and SOA for the NMR data), wildfire OA, Arctic haze OA, and general atmospheric background OA. Factor identities were supported by correlations with tracers, as well as source analysis by backward airmass trajectories. Identified factors align with current understanding of Arctic aerosol sources composition and sources, but with added chemical information and long-term measurements to further assess seasonality. This study by Paglione et al. makes a meaningful contribution to the field and understanding of Arctic aerosol composition. I recommend this manuscript for publication following revisions in response to my following comments.
General Comments:
The use of acronyms should be cleaned up throughout the manuscript. For example, WSOC is defined in the introduction (line 93), but why reiterate the meaning of WSOC in section 2.2 (line 129) rather than 2.1 (line 114)? Similarly, WSOA is indirectly defined in the abstract as “…PM1, with a specific focus on its water-soluble organic fraction (WSOA)…” (line 24), later indirectly again as “…organic aerosol (OA) by analyzing its water-soluble fraction (WSOA)…”, and finally explicitly “…water-soluble organic aerosol (WSOA)” (line 318). I recommend being explicit with acronyms to avoid potential confusion. Long-range transport is defined twice back-to-back on lines 45 and 56. BSOA is defined on line 54, “biogenic secondary organic aerosol” is redefined as just “SOA” on line 74, then “biogenic secondary organic aerosol” is used with no acronym on line 78. This list is non-exhaustive. For clarity, be explicit and consistent with acronym definitions and use.
Specific Comments:
Line 24 – for clarity, I recommend defining WSOA with “water soluble organic aerosol” fully written out, rather than relying on the reader to interpret “…PM1, with a specific focus on its water-soluble organic fraction…”
Line 29 – Can you clarify what the percentages refer to? I assume PM1 contributions by mass?
Line 38 – I suggest rephrasing “…given the fast rate of temperature growth in this area…” to specify that the faster temperature increase is relative to the rest of the globe.
Lines 47-48 – While long-range transport of anthropogenic pollution may be declining, it is also important to acknowledge the increasing local pollution from resource extraction and shipping (e.g., Peters et al., 2011; Pizzolato et al., 2016).
Line 64-65 – “Therefore, current efforts have so far been unable to provide an understanding of the sources and formation pathways of the pan-Arctic OAs in different seasons.” This is a strong statement, somewhat misleading, and contradicted in the following paragraphs. More long-term measurements are certainly necessary, but this statement implies there is no knowledge on Arctic OA sources/formation/seasonality. This statement should be rephrased.
Line 93 – Would WSOA fit better than WSOC since you get non-carbon species in the AMS and IC?
Section 2 - You provide LODs for the EC/OC method (line 124), but what about LODs for your other methods? Did you perform any replicate measurements (e.g., did you run multiple aliquots of the extracts through the IC for an average and standard deviation?)
Line 104 – Does the sampler have a model (Echo?)?
Line 105 – Please provide more details on washing and baking the filters.
Line 108 – What was the range of collection times?
Line 108 – Were the field blanks evenly spaced throughout the campaign? Could you mark them in Fig. S1?
Line 111-113 – Can you clarify the filter portions? I could interpret this as “half of each filter was used in Bologna, the other half at Zurich” or “half of each filter was used in Bologna, while the other half was cut and a part of that other half was sent to Zurich.” I assume the former interpretation is the intention.
Line 151-159 – The sentence describing the identified functional groups is difficult to read. This may be easier to understand as a table. If possible, the addition of approximate chemical shift ranges would be useful too since NMR is less common for atmospheric measurements.
Line 163 – I suggest adding a reference to Fig. S2a for “on average 30%.”
Section 2.4 – Which meteorology data did you use? Did you use isobaric or isentropic trajectories?
Section 2.4 – The authors should acknowledge the general uncertainties of backward airmass trajectories in the Arctic due to a lack of meteorological measurements to constrain the model (e.g., Harris et al., 2005; Kahl, 1993).
Section 2.5 – Perhaps I missed it in the SI, but what fraction of the AMS and NMR signals is accounted for using PMF? In other words, how much of the measured signal on each instrument is not included in any of the factors?
Line 261 – Should the binning of the spectra also contribute to the error matrix? Averaging over the 0.02 ppm bins should include some manner of standard deviation?
Section 3.1 – How did you define your seasons? For example, was winter based on polar night? Was spring based on polar sunrise followed by snowmelt in mid May?
Line 324 – Can you provide standard deviations for O:C and OM:OC?
Line 358 – Figure S14 shows correlations between timeseries of molecular tracers with the NMR factors. A similar figure for the AMS factors would be useful (e.g., IC MSA for the marine biogenic OA factor).
Line 384 – Regarding “… less methylenic long chains and a higher degree of functionalization…”, are you referring to the lack of signal around 3.5 ppm? If not, do you have an explanation for that lack of signal in the marine POA factor compared to the other studies? It may help to also include references to chemical shift ranges (here and elsewhere) for the reader.
Line 390 – It would be worth adding a statement that the AMS factor better correlated with the sum of the NMR factors than with either NMR factor individually (per Table S1).
Line 409 – The comparison of factor F1b and ground types (Fig S15c) should be discussed in the main text. Currently, you discuss the role of sea ice and open ocean broadly, but your trajectory analysis provides further support for the factor identity.
Line 438 – I struggle to see a correlation between the NMR Arctic haze factor and vanillic acid / levoglucosan. To me, it appears as a comparison of noise during a time with low signal. Please provide further discussion to clarify.
Line 448 – The reference to Fig. S14 seems out of place since you discuss the AMS factor while Fig. S14 shows the NMR factors.
Line 463 – The brief explanation of the multilinear regression is hard to follow. Assuming I understand correctly, I suggest rephrasing in a manner similar to “Total OC mass was fit using a linear combination of the WSOC PMF factors (using AMS and NMR factors independently). The multiplicative coefficients are considered to be recovery coefficients (RC), which are inversely related to solubility.”
Line 465 – Following from the previous comment, the reader would benefit from a simple explanation on how to interpret the fitting coefficients. For example: “higher coefficients mean the corresponding factor was less water soluble, and is associated with a higher fraction of insoluble OC.” Again, assuming my understanding is correct.
Line 503 – Table 1 should be mentioned earlier and discussed in the previous paragraphs. When discussing each factor, mention the average mass contributions during relevant seasons (e.g., line 487 marine OA in summer (11% and 14% in AMS and NMR, respectively) and spring (9% and 10%)).
Fig. 4 – Are the NMR F1a and F1b factor timeseries stacked on top of one another? If so, it would be helpful to note that in the caption.
Table 1 – I suggest reformatting for clarity. At first, I thought this table was trying to show the variability of these factors within a particular season. I suggest adding two columns per factor, one for AMS and one for NMR to more clearly show that this table is meant to compare factors between the two instruments. With that said, does this table not simply repeat information in Fig. 6? To help this table add information, you should also include the standard deviations (i.e., the variability within each season and annually).
SI Fig. S2 – What kind of regression is being used (least-squares, orthogonal distance, etc.)? Is it weighted by measurement uncertainties? Similar for regressions in Fig. S3.
SI Fig. S4 – Include a legend.
Fig. S5a – For clarity, start the y axis at 0.5.
SI Lines 70 – For clarity, provide the downweighting factor.
SI Fig S14 – The plots comparing NMR factor 4 with various tracers all show a reduction in those tracers around August and September. Factor 4 doesn’t show a similar reduction. Please include discussion on (i) why these tracers might be lower during this time frame, and (ii) your thoughts on why factor 4 isn’t also lower during that time frame.
SI Fig. S16 – Could you add some annotations for key peaks, similar to Fig. S13?
SI Fig. S16 – There appears to be signal ~6.5-8.5 ppm in the “standards” and EUCAARI factor analysis. Is there a proposed reason why this signal is absent in this study’s wildfire factor?
SI Line 250 – The factors’ contributions to the total OA “…varied by less than 30 %...” (SI line 251). These uncertainties should be acknowledged in the main text and used in the discussion of comparing the AMS and NMR factor contributions (e.g., line 505). The presented factor (Fig. 4) does not seem to be the average presented in the SI (Fig. S19). How does the presented factor compare to the bootstrapping average?
SI Line 251 – Should the reference to factors’ averages and standard deviations be Fig. S19, not S17?
SI Line 301 – The claim that using ugC and umolH yielded similar results needs further discussion. Be quantitative in how they are similar. Perhaps normalized ratios of the fitting parameters? Or comparisons of the reconstructed total OC?
Table S4 – I’m not sure where “beta” is used? Is it meant to refer to a value in equation S4?
Technical Comments
Line 49-50 – Two instances of “increased.”
Line 63 – EC has not yet been defined.
Lines 71 and 74 – I think references to Moschos et al. (2022b) should be just (2022) since there are no references to other studies by Moschos et al. that I can see.
Line 112 – TOC should be defined.
Line 213 – Define OM.
Line 307 – GVB is not defined.
Line 355 – “…(for details see Supplementary Section S2, Figure S6-S18).” For clarity, I recommend moving this portion of the statement up to line 353 (maybe insert alongside “4 for AMS, 5 for NMR”). Section S2 does not discuss the agreement between the techniques, and instead is a more general discussion of the PMF analysis.
Line 442 – Ny-Ålesund is missing a hyphen.
Line 453 – VOC is not defined.
Line 504 – “With” respect to?
Line 516 – “…summertime OA resulted the less oxidized…” Should this say something like “summertime OA was less oxidized”?
Fig. 2a, 3a – The -3 in the y axis label should be superscript.
Figs. 2, 4, 5 - Ny-Ålesund is missing the hyphen and accent on the Å in the captions.
SI Fig. S2 – Panels b and c use AMS HROrg while the caption uses AMS WSOM.
SI Fig. S7 – caption says 5 factors while the figure shows 4.
References
Harris, J. M., Draxler, R. R., and Oltmans, S. J.: Trajectory model sensitivity to differences in input data and vertical transport method, Journal of Geophysical Research: Atmospheres, 110, https://doi.org/10.1029/2004JD005750, 2005.
Kahl, J. D.: A cautionary note on the use of air trajectories in interpreting atmospheric chemistry measurements, Atmospheric Environment. Part A. General Topics, 27, 3037–3038, https://doi.org/10.1016/0960-1686(93)90336-W, 1993.
Peters, G. P., Nilssen, T. B., Lindholt, L., Eide, M. S., Glomsrød, S., Eide, L. I., and Fuglestvedt, J. S.: Future emissions from shipping and petroleum activities in the Arctic, Atmospheric Chemistry and Physics, 11, 5305–5320, https://doi.org/10.5194/acp-11-5305-2011, 2011.
Pizzolato, L., Howell, S. E. L., Dawson, J., Laliberté, F., and Copland, L.: The influence of declining sea ice on shipping activity in the Canadian Arctic, Geophysical Research Letters, 43, 12,146-12,154, https://doi.org/10.1002/2016GL071489, 2016.
Citation: https://doi.org/10.5194/egusphere-2025-760-RC1 -
RC2: 'Comment on egusphere-2025-760', Anonymous Referee #2, 14 Apr 2025
reply
General comments
This paper presents a one-year data set of PM1 samples obtained in Ny-Ålesund with the bulk chemical analytical results of H-NMR spectroscopy and off-line HR-ToF-AMS. The authors show that the analytical results of the two measurements are consistent in terms of source contributions to OAs. Their data suggested that the observed OAs in winter and spring are dominated by long-range transport of anthropogenic pollution in Eurasia, while the aerosol in summer is characterized by biogenic aerosols from marine sources. Overall, the paper provides new insights into our understanding of the seasonality in the source contributions to OAs in the Arctic region and confirms some of the findings already reported in previous studies. While the data presented are valuable and interesting, there are some issues that need to be clarified before its publication in ACP.
Specific comments
(1) From the text, the contributions of terrestrial BSOA traced by oxidation products of terpenes are not clear. The authors attributed these BSOAs to emissions from wildfires but this is not always the case. Indeed, Moschos et al (2022) reported that significant or non-negligible amounts of BSOA (not necessarily related to biomass burning) from forests for the observed OA in pan-Arctic in summer. I think that the authors should add more discussions on this point (e.g., if the author’s result is different from Moschos et al., why?).
(2) P.7, L.215: WSOM = WSOC×(OM:OC)_AMS
I understand that the advantage of the use of OM:OC at the time of each sampling is to be able to expect more realistic abundance of WSOM rather than by use of a constant value of the factor. Meanwhile, the composition of OM minus OC can also include water-insoluble compound mass. How can this use of OM:OC ratio be verified (or is there any evidence) to represent water-soluble mass? The authors should add some more description including uncertainty in this calculation.
(3) P.10, L.295: “The major chemical mass …, followed by..”
I think that this statement may cause misunderstandings of readers: some may think that sulfate was the most abundant and the second most is seasalt, followed by OM. However, there seems to be no statistical difference in the fraction among sulfate, seasalt, and OM. As the authors described in the conclusion section, these three components had similar contributions to the PM1 mass. Please modify the sentence.
(4) P.11, after L. 329: Regarding the statement staring with “MSA and …,” which figure is referred to? Maybe Fig. 4? Please clarify it.
(5) Figure 5: This figure is very hard to see. For example, the color code is not clear what it represents. The authors explain it in the caption with quantitative information, but they should show the color code in the figure panel in addition to describing it in the caption. Moreover, geographical lines (map) in the figure are not clear at all.
(6) Figure 6 and P.16: As the author described, the scaled contributions of the NMR and AMS factors to total OC showed generally good agreement. However, the relative contributions of background OA, Arctic haze OA, and Aged wildfires OA between NMR and AMS particularly in spring are significantly different. The authors should add more discussion on the possible reason for this difference.
Citation: https://doi.org/10.5194/egusphere-2025-760-RC2
Data sets
Main datasets of "Unraveling Arctic organic aerosol sources: a year-long study by H-NMR and AMS in Ny-Ålesund, Svalbard" Marco Paglione https://doi.org/10.71761/0e110925-1f3d-4013-b048-e5a47ca3be6f
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 186 | 51 | 10 | 247 | 29 | 7 | 7 |
- HTML: 186
- PDF: 51
- XML: 10
- Total: 247
- Supplement: 29
- BibTeX: 7
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 68 | 29 |
| Italy | 2 | 27 | 11 |
| China | 3 | 22 | 9 |
| Germany | 4 | 17 | 7 |
| Switzerland | 5 | 11 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 68