the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Jul 2026
Determination of Aerosol Particle Organic Carbon (OC) by Aerosol Mass Spectrometry (AMS) and Thermo-Optical Transmission (TOT) Analysis: Spatial and Seasonal Differences
Abstract. Reliable quantification of organic carbon (OC) is essential for the comparability of long-term aerosol measurements across monitoring networks. A multi-site intercomparison of OC measurements was conducted at three Central European background stations, Melpitz (MEL), Frýdlant (FRY), and Košetice (NAOK), using aerosol mass spectrometry (AMS/ACSM) and thermal-optical transmission (TOT) methods (offline and semi-continuous). Semi-continuous and offline TOT measurements showed high consistency across sites and seasons, suggesting negligible influence of particle size differences on OC comparability. A pronounced seasonal contrast was observed between AMS derived OC (OCAMS) and TOT based OC (OCTOT). During summer, good agreement was found with OCTOT:OCAMS slopes of 1.27 (MEL), 0.86 (FRY), and 0.83 (NAOK). In winter, AMS systematically underestimated OC, with slopes decreasing to 0.34 (MEL), 0.24 (FRY), and 0.63 (NAOK). The discrepancy increased under combustion dominated conditions and was associated with low organic aerosol to equivalent black carbon (OA:eBC) ratios (<≈5), identifying a regime where OCAMS and OCTOT diverge substantially. Source dependent effects were evident, with the largest deviations observed at the coal influenced site (FRY) and during long-range transport events, while improved agreement coincided with higher OA:eBC ratios and enhanced secondary organic aerosol formation in summer. Evaluation of potential uncertainties (including pyrolytic carbon formation in TOT and AMS related limitations such as refractory organics and ionization efficiency) showed that no single factor explains the winter discrepancy, suggesting combined matrix and source effects. The results show that OA:eBC can serve as a practical indicator for OCAMS-OCTOT agreement and provide recommendations to improve OC quantification under winter combustion conditions.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Measurement Techniques.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(1629 KB) - Metadata XML
-
Supplement
(1905 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- CC1: 'Summer/Winter vs Clusters, wildfire influence, raw data', J. C. Corbin, 20 Jul 2026
-
RC1: 'Comment on egusphere-2026-2028', Anonymous Referee #2, 31 Aug 2026
The manuscript examines the seasonal difference between AMS-derived organic carbon (OCAMS) and thermal-optical OC (OCTOT) measurements from three Central European sites using AMS / ACSM and thermal-optical transmission methods. These comparisons are relevant to long-term aerosol monitoring, and the study has the potential to provide useful insights into differences between these two commonly used OC measurement approaches However, there are several important issues that should be addressed before the manuscript is suitable for publication.
Major comments:
The observational finding of substantial wintertime differences between OCAMS and OCTOT appears reasonably robust, particularly because the agreement between online and offline TOT measurements suggests that the large winter discrepancies are unlikely to arise solely from filter-analysis analytical artifacts. However, the manuscript proceeds from this observational result to relatively specific mechanistic interpretations involving coal combustion, refractory organics, dust transport, BrC, and ultimately an OA threshold for identifying conditions under which OCAMS maybe biased low. These interpretations are not yet sufficiently supported because the analyses do not adequately distinguish among instrumental effects, circular reasoning arising from shared variables, sampling differences, and genuine aerosol-composition effects. In several cases, the analyses are partly circular or do not directly test the proposed mechanisms. The manuscript should be revised by first rigorously establishing the robustness and causes of the inter-method differences and then providing independent evidence for the proposed explanations. Otherwise, the corresponding interpretations and conclusions should be removed or substantially toned down.
A key example is the proposal of OA:eBC as a diagnostic threshold for the discrepancy between OCAMS and OCTOT. OCAMS is calculated directly from AMS OA as OCAMS = AMS OA/(OM:OC). Thus, the two quantities are intrinsically coupled through the shared AMS OA measurement. If the AMS underestimates OA for any reason, both OCAMS and OA will decrease simultaneously. Consequently, the observation that low OA is associated with low OCAMS/OCTOT cannot, by itself, demonstrate a source- or matrix-dependent effect.
This concern is further strengthened by Figure 8, which pools observations from three sites and two distinct seasons. FRY winter appears to occupy much of the low-OA:eBC/large-discrepancy region, whereas summer observations largely occupy the high-OA:eBC/small-discrepancy region. The apparent relationship may therefore reflect differences among site-season groups rather than a continuous relationship within a given aerosol regime. In addition, the proposed OA:eBC threshold of approximately 5 appears to be visually selected rather than statistically established.
Given these concerns, it is important that the authors redo this analysis using independent source indicators. For example, examine against eBC, , independently derived source tracers, PMF factors, PAHs, potassium, levoglucosan, elemental tracers, or other available source information. At minimum, the relationship should be demonstrated separately within each site and season and tested using a multivariable or mixed-effects model. Any threshold should be determined statistically and validated rather than inferred visually from the pooled dataset.
Instrument-related uncertainties also require more careful evaluation before the OC discrepancies are attributed primarily to aerosol composition or source characteristics. A major source of AMS measurement uncertainty is its collection efficiency for particles. While the manuscript discusses RIEs, high-molecular-weight compounds, ion transmission, organonitrates, and the Pieber effect, it does not quantitatively assess uncertainties associated with CE. This is particularly important because three different instruments were used—an HR-ToF-AMS at MEL, a C-ToF-AMS at FRY, and a ToF-ACSM at NAOK—and the largest OCAMS–OCTOT discrepancy occurs at FRY. The manuscript should clearly state how CE was determined or assumed for each instrument, site, and season; whether composition-dependent CE was applied; what calibrations were performed; and whether independent measurements (such as SMPS and particle mass concentration measurements) are available to inform about possible differences in inlet transmission, vaporizer performance, and overall mass quantification among the instruments.
The MEL winter volume closure shown in Figure S13 is useful, but it does not establish quantitative closure for the other instruments or seasons. In particular, no equivalent mass or volume closure appears to be presented for FRY, where the largest discrepancy is observed. Providing AMS/ACSM mass or volume closure for each site and season, especially FRY winter, would substantially strengthen the assessment of whether instrument-specific measurement biases contribute to the observed differences.
The OM:OC calculation and HR-versus-UMR data treatment need to be carefully explained as well. here appear to be internal inconsistencies between the equations provided in the manuscript and several values reported in Table S1. This directly affects the quantity being evaluated, so it is not a minor presentation issue. For example, the manuscript states: and . However, several numbers in Table S1 do not reproduce from these equations. For example, as printed in the Supplement:
MEL winter: f_44=0.16 -> OM:OC ≈ 2.16, not 2.64.
MEL summer: f_44=0.14 -> OM:OC ≈ 2.05, not 2.57.
NAOK summer: f_44=0.18 -> OM:OC ≈ 2.27, not 2.81.
Table S1 nevertheless reports 2.64, 2.57, and 2.81, respectively. These differences are large enough to substantially alter OCAMS.
A related inconsistency occurs in Section 3.2, which is presented as an evaluation of the influence of mass spectral resolution on OCAMS, but Figure S3 actually plots HR OA against UMR OA, not derived OC. The figure caption says “organic carbon,” whereas the axes say OA. Agreement in OA does not establish agreement in OC because the HR and UMR OM:OC estimates can differ. It is important that the authors explicitly document the exact time-resolved equation used to calculate OCAMS for every site, identify which MEL calculation was used in Figures 5–8, reproduce Table S1, and then redo the HR-versus-UMR comparison using derived OC rather than OA.
More generally, important information needed to evaluate the comparability of the datasets is scattered throughout the manuscript and Supplement. A detailed summary table should therefore be provided that compiles the relevant information for all datasets analyzed in this study. At a minimum, the table should include the measurement sites, sampling period and season, instrument types and model, measured quantity, sampling size cut (PM1 vs. PM2.5), number of valid offline samples or paired observations used in the comparisons, and the native and/or averaged time resolution of the online measurements. For the AMS/ACSM measurements, it would also be useful to indicate the instrument configuration and any important differences among sites that may affect comparability.
Finally, the conclusions and recommendations should be more closely aligned with what the present dataset can establish. The recommendations that TOT should serve as the primary reference for winter OC trends, that AMS-derived OC should under certain conditions be regarded only as a lower bound, and that OA:eBC < 5 identifies problematic measurement conditions are broad conclusions based on three Central European sites, two relatively short seasonal campaigns, three different AMS/ACSM instruments, and a mechanism that remains unresolved. A more defensible and still important conclusion is that, under the winter combustion-influenced conditions encountered in this study, OCAMS was substantially lower than EUSAAR-2 OCTOT, whereas agreement was considerably better in summer. The causes of this seasonal discrepancy remain uncertain and likely involve multiple instrumental, sampling, and aerosol-composition factors. The broader recommendations should therefore be moderated accordingly unless they can be supported by additional independent analyses.
Specific comments:
- Abstract, lines 15–16: remove “negligible influence of particle size differences.” FRY demonstrates that size-cut mismatch cannot explain the large winter discrepancy, but it does not show that size effects are negligible everywhere.
- Methods, Sect. 2.4: provide more information about AMS/ACSM calibration, CE, RIE-OA, fragmentation-table treatment, and instrument-specific QA/QC. Referring readers primarily to another publication is insufficient for a measurement-comparison paper.
- Provide sample sizes (data point numbers) in all the scatter comparison figures.
- Table S1: verify all OM:OC, O:C, and f44 values against Eqs. 1–2 and specify whether the table contains means of time-resolved ratios or ratios calculated from mean values. Several entries appear inconsistent with the stated equations.
- 3.1: slopes of 0.79–0.88 should not simply be described as “close to unity” without discussing the systematic 10–20% difference and its relevance to subsequent comparisons.
- Lines 187–207: the statement that the lower MEL R2 “could be attributed to enhanced SOA production” needs evidence. Enhanced SOA does not inherently reduce agreement between two OC methods.
- 3.2/Fig. S3: clarify OA versus OC. The current figure evaluates OA, despite the section and caption referring to OC.
- Lines 289–291: saying the MEL comparison demonstrates “reliable and reproducible OC estimation across instruments and years” is too strong. Similar slopes in different campaigns do not establish accuracy.
- Line 300: “size effects were excluded” should be changed to something like “nominal inlet size-cut differences cannot explain the FRY discrepancy.” AMS aerodynamic-lens transmission and particle properties still matter.
- 3.4.1: explain whether C1W/C2W grouping was defined independently before examining the OC discrepancy. Otherwise there is a risk of post hoc classification.
- Lines 367–374: direct evidence is needed for the claim that transported mineral dust carried enough refractory organic material to influence the OC comparison.
- 3.5.1: the section title repeats the title of Sect. 3.5.
- Lines 469–476/Fig. S4: remove the “PC subtraction” as a correction or justify it rigorously from thermal-optical theory.
- Lines 527–545: the RIE sensitivity analysis is performed mainly at MEL, while FRY has the most severe discrepancy. Its generalization to FRY and NAOK therefore needs qualification.
- Lines 539–541: an OSc range from –1 to +1 should not simply be described as “fresh, less-oxidized POA.” OSc approaching +1 is highly oxidized. Interestingly, the Supplement states that most values were –1 to 0; the main text and SI should be made consistent.
- 3.5.2: the Pieber-effect discussion is overly qualitative. This can potentially be evaluated looking of periods of The argument that three instruments are unlikely to show the same artifact does not exclude a systematic seasonal dependence associated with nitrate-rich winter aerosol.
- 3.6: distinguish clearly among BrC absorption, inferred BrC mass, MACOA, and MACBrC; explain uncertainty and independence of these quantities.
- Conclusion, line 575: “(range of slopes and R2)” is an unfinished placeholder and needs to be replaced.
- Conclusion: replace statements implying that OCTOT is the absolute truth with language such as “AMS-derived OC was systematically lower than EUSAAR-2 TOT OC.” Both are operationally defined measurements with different uncertainties.
- The Supplement title differs from the manuscript title, and Figure S3 says “organic carbon” while plotting OA. These should be harmonized.
Citation: https://doi.org/10.5194/egusphere-2026-2028-RC1 -
RC2: 'Comment on egusphere-2026-2028', Anonymous Referee #3, 26 Sep 2026
The manuscript titled "Determination of Aerosol Particle Organic Carbon (OC) by Aerosol Mass Spectrometry (AMS) and Thermo-Optical Transmission (TOT) Analysis: Spatial and Seasonal Differences" fundamentally focuses on measurement-comparability challenges across instruments and addresses this by integrating observations from three Central European sites over two seasons. I think comparing offline and semi-continuous thermal-optical measurements is useful for the broader community, and the authors made a comprehensive effort to examine possible sources of disagreement for OC between AMS/ACSM and thermo-optical technique.
Still, there are avenues where further improvements are necessary. For example, the conclusion that winter combustion conditions and more particularly, the low OA:eBC ratios, determine when AMS-based OC diverges from TOT OC is somehow undersupported with the current analysis. This, I think is my primary comment. The measurements have different particle-size cuts and time resolutions whose treatment should be more clearly described. The OA:eBC ratio is not independently validated. The paper would benefit from a stronger consideration of these limitations and more careful interpretations. Following are my complete comments:
-
pp 4, lines 85–94; pp 5, lines 112–145; pp 10, lines 245–250: The compared measurements do not consistently represent the same aerosol size fraction. AMS/ACSM measures PM₁. Offline TOT uses PM₂.₅ filters. The online TOT cut is PM₂.₅ at MEL and NAOK but PM₁ at FRY; meanwhile, the AMS/ACSM cuts also vary among sites. The statement that size effects are excluded at FRY does not resolve the size-cut mismatch at the other comparisons. It would be good to constrain the size-cut contributions, clearly note which comparisons are size matched, and qualify conclusions that rely on PM₁–PM₂.₅ comparisons.
-
pp 4, lines 97–110; pp 5, lines 112–120; pp 10–11, lines 245–275: The time matching and regression methods need to be described in enough detail to reproduce the analysis. Please specify how AMS/ACSM data were averaged to match the 12-hour offline filters and the 2-hour or 4-hour online TOT measurements; how incomplete or missing intervals were handled; and how many paired observations enter each fit.
-
pp 14, lines 339–355; pp 15, lines 370–377; pp 22, lines 589–593: The proposed OA:eBC threshold is not yet established as a practical indicator. The conclusion recommends caution when winter OA:eBC is below approximately 5, but the evidence appears to be a visual or site-season association across a limited number of campaigns. OA:eBC also contains AMS OA—the quantity used to calculate OCAMS—so its relationship with OCAMS–TOT differences may partly suggest mathematical coupling. If possible, the relationship should be quantitatively tested using paired observations, accounting for uncertainty and sensitivity to site and season, and assess whether it remains after considering the measurement-size mismatch. At the very least, the OA:eBC < 5 should be presented as a preliminary observation specific to these campaigns rather than a general threshold or recommendation.
-
pp 12, lines 293–315; pp 13–15, lines 321–377; pp 20–21, lines 535–565: Source attribution and proposed mechanisms should be framed as hypotheses. The authors attribute the largest discrepancy in measurements to coal influence at FRY, transported dust, refractory organics, and combustion matrix effects, but the evidence presented is largely indirect. The merged trajectory clusters include dust episodes but it is acknowledged that mean trajectories do not identify the dust source. OA:eBC is not a unique source tracer and brown carbon absorption does not directly establish the mass of refractory OC missed by AMS. I suggest to meticulously distinguish measured results from interpretations.
-
pp 5, lines 124–145; pp 9, lines 227–243; pp 17–20, lines 432–508: I think the AMS/ACSM OC estimation discussion should be further expanded. The authors apply the f₄₄-based equations to UMR data at FRY and NAOK, while MEL uses HR data and the I-A method. The UMR–HR comparison at MEL is used to suggest similar behavior elsewhere, but that somehow does not validate the UMR conversion at other sites or for their aerosol compositions. Please state precisely how OC was calculated at each site, identify the applicable OM:OC method and corrections, and discuss the limitations of extrapolating the MEL HR–UMR result to FRY and NAOK. The reported uncertainty should include the uncertainty in the f₄₄-to-O:C and OM:OC parameterizations.
-
pp 16–17, lines 385–430; pp 22, lines 598–602: The interpretation of pyrolytic carbon (PC) subtraction and TOT as a reference needs clarification. Subtracting PC from winter OC changes the slopes substantially, including a NAOK slope above unity, but the manuscript does not make clear how PC is defined, measured, and assigned to OC, or how the uncertainty of that subtraction is handled. It may be useful to show that calculation (SI-relevant) and clarify whether the reported PC quantity is a measured fraction or an operationally defined fraction of the thermogram. A standardized protocol is not necessarily an unbiased absolute reference as the authors also discuss possible charring and split-point uncertainty. Please qualify the conclusion that TOT provides “a more stable reference for absolute OC levels” and justify what “stable” means in this context.
-
pp 20–21, lines 525–565: The brown carbon analysis is not sufficiently connected to the OC comparison. Brown carbon is used to help explain AMS–TOT disagreement, but the section mainly interprets seasonal and site-level absorption differences. The reported brown carbon absorption standard deviations are large relative to some means, particularly in summer, and MAC values also vary strongly. The SI shows brown carbon absorption and MAC calculations but further expansion is needed to describe uncertainty and data filtering, and test whether either quantity covaries with the OC residuals. If no such relationship is tested, present this section as contextual optical characterization rather than evidence explaining the method discrepancy. Some source-specific interpretations (for example, “coal-derived” or “tar-rich”) should be softened unless supported by independent source or molecular measurements.
Minor comments
-
pp 6, lines 167–170; pp 7, lines 185–193; pp 8, lines 207–225: The subsection numbering in Section 3 repeats “3.1.1” for TC, OC, and EC. Please correct the numbering.
-
pp 15, lines 349 and 362–365: Please resolve the winter NAOK OA:eBC inconsistency: section 3.4.2 gives 10.21 in one place and 10.75 later (lines 349 and 362–365). Ensure the abstract, text, tables, and figures use the same value.
-
pp 11, lines 260–275: The online OC comparison at MEL is described as showing OCOnline lower than OCAMS (section 3.3), whereas the offline comparison has a slope of 1.27 for OCAMS relative to OCOffline. Please make the regression orientation explicit and explain the difference between the two comparisons.
-
pp 7, lines 185–190: The statement that the lower R² at MEL in the online/offline OC comparison “could be attributed to enhanced SOA production” (section 3.1) is speculative. Please either support this with analysis or remove/qualify it.
-
pp 1–2, lines 12–27 and 51–73; pp 10–11, lines 245–275: Please check terminology and notation throughout: use OCAMS/OCACSM consistently, define “OCTOT” when it first appears outside the abstract, and standardize “semi-continuous,” “online,” and “offline.”
Citation: https://doi.org/10.5194/egusphere-2026-2028-RC2 -
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 261 | 117 | 39 | 417 | 71 | 38 | 33 |
- HTML: 261
- PDF: 117
- XML: 39
- Total: 417
- Supplement: 71
- BibTeX: 38
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dear Authors,
I looked very briefly at your interesting article, and have some brief comments. I apologize for not having enough time to expand these comments into a full review.