the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Jul 2026
Method-Dependent Variability in Hygroscopicity Parameter (κ) of Particles from Biomass-Burning Across Fuel Types and Burn Phases
Abstract. The hygroscopicity parameter κ is important for assessing biomass-burning aerosol impacts on visibility, direct radiative forcing, and cloud-related indirect effects. Discrepancies among κ values derived from different methods have been reported, but often separately across fuel types, burn phases, and experimental systems, limiting assessment of their consistency and controlling factors. Here, we compare κ derived from a cloud condensation nuclei counter (CCNC), a hygroscopicity tandem differential mobility analyzer (HTDMA), and AMS–SP2 composition-based predictions within a single laboratory framework. Fresh and aged biomass-burning particles were produced from four representative fuels: hardwood, softwood, peat, and leaves, burned under smouldering or flaming conditions. CCNC-derived κ systematically exceeded HTDMA-derived κ, with discrepancy magnitudes strongly dependent on fuel type and burn phase. Composition-based κ generally fell between CCNC and HTDMA values. Coupled effects of κ size dependence and size representation, together with externally mixed non-hygroscopic black carbon, affected discrepancies among the three approaches but only partly explained them. Using O:C-informed κorg values instead of a constant κorg = 0.1 consistently improved AMS–SP2 prediction accuracy relative to HTDMA values in peat experiments only, by 16–59 %, while substantial residual discrepancies remained. These results suggest the presence of additional controlling factors not tested here, motivating measurements of surface-active species, particle phase state, and co-condensation of soluble gases to potentially explain the remaining discrepancies in future work. They further highlight the need for regime-aware κ parameterizations in atmospheric and climate models, alongside careful treatment of size representation, black-carbon inclusion, and κorg assumptions in composition-based hygroscopicity predictions.
Competing interests: The authors declare that they have no conflict of interest.
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
(1129 KB) - Metadata XML
-
Supplement
(3407 KB) - BibTeX
- EndNote
Status: open (until 24 Aug 2026)
-
RC1: 'Comment on egusphere-2026-2723', Anonymous Referee #1, 21 Jul 2026
reply
-
AC2: 'Reply on RC1', Sara Aisyah Syafira, 15 Aug 2026
reply
1. The sentence, “Further discussion regarding the more detailed patterns of the HTDMA-derived κ values, along with investigation of some observed outliers and large propagated error bars, is in the Supplementary Information, Notes S4 and Fig. S5,” has now been replaced by the following text in the manuscript. The first paragraph of Note S4 was moved to this section, with some wording edits to improve readability:
“Notably, the smouldering peat experiment exhibited κ values close to those of flaming hardwood, with both representing the lowest κ values. In contrast, flaming softwood showed the highest κ values. The κ values for the other three experiment types fell between these three experiment types. Among the wood fuels, softwood exhibited higher κ values than hardwood under both burn phases. Further investigation of the observed outliers and large propagated error bars is provided in the Supplementary Information. ”2. We noticed that the words " pre-aging stage" have just started to appear in the CCNC discussion, while they should already appear in the HTDMA discussion too as part of the description about the used datasets. Once checked, it was written as “collected from all repetitions and all hourly fresh-aging stage period intervals” (line 329 and 331).
To make it clearer, we changed that sentence to “collected from all repetitions and all aging stage period intervals, including the dark aging experiments and the pre-aging period of the photoaging experiments. Pre-aging stage refers to the time period since the start of the experiment at which the chamber cooling was first turned ON, i.e., directly after both the smoke injection and chamber air filling finished, until the time point at which the lights were turned on.” This clarification is intended to make the distinction between the pre-aging, dark-aging, and photoaging periods clearer to the reader.3. Thank you for the question. Before answering it, we’re going to re-clarify some major information provided in Table 3 first, as below:
Size-dependent κ variability in the HTDMA measurements exhibits a larger magnitude than the supersaturation-dependent κ variability in CCNC measurements across almost all experiment types, except smouldering softwood, based on the range of the absolute values.When comparing the different fuel types and burn phases, the values were normalized to a 0–100 range to compare the relative magnitude of the factor-dependent κ effect (size dependence for HTDMA and supersaturation dependence for CCNC) across the different experiment types. The individual normalized values should not be interpreted as measures of fuel-type variability. Rather, they reflect the relative magnitude of the factor-dependent κ effect (size dependence for HTDMA and supersaturation dependence for CCNC) across different fuel type/burn-phase conditions. The narrower range of these normalized values for CCNC (83) than for HTDMA (100) indicates a smaller variation in the magnitude of the supersaturation-dependent effect across six experiment types for CCNC than in the magnitude of the size-dependent effect across six experiment types in HTDMA measurements.
When comparing the mean κ values obtained from CCNC and HTDMA, we find that the absolute mean κ values span a narrower range across the HTDMA measurements (0.103) than across the CCNC measurements (0.133). However, because the mean CCNC κ values are consistently higher than the mean HTDMA κ values, the absolute ranges are not directly comparable as a measure of relative fuel/burn-phase-dependent variability. By normalising the range of κ values by the mean κ across the six experiment types for each measurement method, values of 172% for HTDMA and 75% for CCNC were obtained.
Regarding the direct answer to the question, HTDMA and CCNC measurements have different conditions and mechanisms, so that particles from the same experiment type, i.e., same burn phase and fuel type, may act differently and thus also have different relative κ value positions to the other experiment type under measurements of the two instruments. An obvious example in the current study is peat particles, which had the lowest HTDMA κ but the highest CCNC κ. Therefore, the reliability of the comparison across different fuel types or different burn phases (different experiment types) depends on the use/application of results. HTDMA measurements should be more reliable for use in a subsaturated conditioned environment regarding the hygroscopic growth factor, e.g., for ambient air visibility observation in the low-level atmosphere. Meanwhile, CCNC measurements should be more accurate to be used in a supersaturated environment observing cloud droplet activation, relevant to aerosol–cloud interaction studies and cloud activation modelling.
4. Thank you very much. The original sentences:
“The suspected cause of these both reversed and higher values of the BC-excluded predicted κ of flaming experiments was the higher inorganic: organic ratio of these experiments compared to the smouldering experiments. Across the flaming experiments, the prediction accuracy errors changed from underpredicting the CCNC κ as much as 23.17% and 4.91% to overpredicting the CCNC κ as much as 164.93% and 233.41% after the BC exclusion, for flaming hardwood and flaming softwood experiments, respectively.”
It is now changed to:“Specifically, the predicted κ changes from 23.17% and 4.91% lower than the CCNC κ to 164.93% and 233.41% higher than the CCNC κ after the BC exclusion, for flaming hardwood and flaming softwood experiments, respectively. The suspected cause of this flipped pattern of the flaming versus smouldering was the higher inorganic: organic ratio of the flaming experiments compared to the smouldering experiments.“
5. The mass-weighted HTDMA κ values were lower in most experiments, but this was not the original reason for adopting number-weighted κ in Fig. 3. The original reason was that CCNc also measures κ using the number size distribution instead of the particle mass size distribution, specifically in determining the D50 values. The exploration using mass-weighted mean HTDMA κ is just for comparison with the AMS-SP2 mass-based predicted κ. By comparing the first right picture in Figure S7 (Leaves) and Figure S8.b, the mean size in number-weighted mean HTDMA κ has values closer to the SMPS number mean size than the mean size in mass-weighted mean HTDMA κ has to the SMPS mass mean size. It may partly explain the majority of the worse predictions in measured HTDMA κ comparisons. Additionally, the mass size distribution was obtained from the assumption of constant and uniform density across all particles, that maybe inappropriate in certain experiment types. (Notes: there was a mistyping in the Figure Label of Fig. S8, which should be SMPS mass mean size instead of number mean size).
6. Based on Table S12, comparing the measured and predicted κ gaps/differences at the HTDMA sizes of 110 nm and 75 nm, no consistent preference is revealed across all individual experiments. However, the majority is that the κ values at 75 nm were closer to the mean predicted κ than those at 110 nm, except in one flaming-softwood experiment and in one non-photo-ageing flaming-softwood experiment.
The SMPS particle mass distribution mean size is generally higher than the SMPS particle number distribution mean size. For example, in the leaves experiment, the number size distribution mean sizes are approximately less than 350 nm, but the mass size distribution mean sizes are around 400-500 nm. However, the predicted mean AMS-SP2 mass-weighted mean κ is closer to the smaller (75 nm) size than to the larger (110 nm) size in most experiments. This may indicate that the direction of the κ–size dependence is not consistent across the particle size distribution within the same experiment type.
7. Thank you for the suggestion. We will try to conduct the ion-pairing scheme, thus species-specific inorganic κ assignment and volume-fraction-based κ calculation following the method in Hu et al (2021) https://doi.org/10.1039/d0fd00077a, for some (if not all) experiments, and then compare with the corresponding mass-fraction-based κ
Citation: https://doi.org/10.5194/egusphere-2026-2723-AC2
-
AC2: 'Reply on RC1', Sara Aisyah Syafira, 15 Aug 2026
reply
-
RC2: 'Comment on egusphere-2026-2723', Anonymous Referee #2, 22 Jul 2026
reply
This manuscript investigates the method-dependent variability of the aerosol hygroscopicity parameter (κ) for biomass-burning aerosols by comparing three commonly used approaches: HTDMA measurements, CCNC-derived κ values, and composition-based predictions using AMS-SP2 data. The authors systematically evaluate the effects of fuel types and combustion phases on κ variability within a consistent experimental framework.
The study addresses an important issue in aerosol hygroscopicity research, namely the limited comparability of κ values derived from different methods. The results demonstrate that κ is not solely controlled by bulk chemical composition but is influenced by multiple factors, including measurement conditions, particle size, and mixing state. The finding that CCNC-derived κ values are generally higher than HTDMA-derived values highlights the differences between sub-saturated water uptake and super-saturated cloud activation processes. The combination of HTDMA, CCNC, AMS, and SP2 measurements provides a valuable dataset for understanding aerosol–water interactions.
However, several issues should be addressed to further improve the clarity and scientific significance of this manuscript.
Major comments：
1. Manuscript clarity and organization
The manuscript contains valuable information; however, the current version is sometimes difficult to follow due to lengthy descriptions and repeated explanations. The authors should further streamline the Introduction, Methods, and Results and Discussion sections, clarify the main scientific questions, and emphasize the key findings. Some detailed descriptions could be moved to the SI to improve readability.2. Figure quality and consistency
The quality and presentation of figures require further improvement. The formatting, font size, and resolution should be standardized throughout the manuscript. Some figures are currently difficult to interpret due to insufficient clarity, which may not fully meet the current presentation standards of ACP. The authors should revise all figures accordingly.3. Limitation of using O/C ratio to predict hygroscopicity
The manuscript uses AMS-derived organic composition and O/C ratio as important indicators for predicting aerosol hygroscopicity. However, this approach may oversimplify the chemical control of water uptake. Different oxygen-containing functional groups have distinct interactions with water: hydroxyl and carboxyl groups generally enhance hydrophilicity, while other oxygen-containing groups may have substantially different contributions despite increasing the O/C ratio.I encourage the authors to further investigate specific oxygen-containing functional groups using FTIR, NMR, or high-resolution mass spectrometry approaches and establish quantitative relationships between functional groups and hygroscopicity. Such analysis would provide a more mechanistic explanation for the differences in aerosol hygroscopicity among different fuel types and improve the current prediction framework.
4. Need for direct evidence of size-dependent effects
The authors suggest that particle-size-related factors may contribute to the discrepancy between HTDMA measurements and composition-based predictions. However, the current discussion remains largely speculative because the were not sufficiently characterized in a size-resolved manner. Size-resolved chemical measurements would provide stronger evidence to determine whether the observed size dependence of κ is associated with variations in chemical composition, oxidation state, or mixing state among different particle size ranges.5. I appreciate the authors’ efforts in investigating a broad range of fuel types, which provides a valuable dataset for understanding biomass-burning aerosol hygroscopicity. However, related observations have been reported in previous studies. (ACS EST Air 2026, 3, 3, 697–709; Journal of Geophysical Research: Atmospheres, 131, e2025JD044564.) Therefore, if the authors are unable to provide additional direct evidence to validate the proposed mechanisms further, the manuscript should include a more detailed comparison with previous work, particularly regarding differences in experimental design, scientific insights, and novel contributions. Such discussion would be important for clearly establishing the originality and significance of this study.
6. Fuel moisture content can affect combustion efficiency, flame temperature, aerosol formation pathways, and particle properties. The authors should clarify how fuel moisture content was controlled before combustion.
Citation: https://doi.org/10.5194/egusphere-2026-2723-RC2 -
AC1: 'Reply on RC2', Sara Aisyah Syafira, 15 Aug 2026
reply
- Thank you for the suggestion. Following this, we are going to do further proofreading; then streamline the “Introduction, Methods, and Results and Discussion" sections, clarifying the main scientific questions, emphasizing the key findings, and moving some detailed descriptions to the SI. Detailed changes will be described later along with the submission of the revised version.
- The MATLAB figure files are available through the Zenodo link we state in the manuscript. We will try to find a better way to export or reproduce the figures, resulting in higher resolution and better clarity of the figures in the manuscript.
- We agree that the use of the O/C ratio, especially across experiments involving different fuel types and burn phases, can be an oversimplification. We therefore still prefer to use the widely used mean organic κ value of 0.1, with an extension of this value for a plausibility assessment in our investigation of the inter-method κ discrepancies. Some published studies have shown good linear correlations (R²) between O/C and hygroscopicity, whereas others have not. This indicates that O/C is not the only factor governing organic aerosol hygroscopicity and that the detailed molecular properties of the organic fraction may also play an important role. Quantitative characterization of oxygen-containing functional groups could provide an additional and potentially more mechanistic approach to the O/C ratio for estimating organic κ. However, the suggested measurements for that factor were not conducted during the experimental campaign of the current study. Offline chemical composition analysis using UHPLC–ESI high-resolution Orbitrap mass spectrometry was conducted on the filtered particles of wood-type fuels during the experiments in the current study (Evans et al., 2025). The results showed the presence of organonitrogen species in the OA in small fractions, while approximately 90% of the OA consisted of oxygenated compounds (CHO), with more than 50% of the CHO mass in primary organic aerosol (POA) being aromatic, largely in the form of functionalized monoaromatic species, with the aromatic contribution subsequently decreasing upon ageing (Evans et al., 2025). However, no dedicated investigation was done to detect and quantify the oxygen-containing functional groups. We will add the recommended organic κ investigation using oxygen-containing functional groups in the proposed future set of experiments described in the Supplementary Information, Note S8.
- Thank you for the comments. We will try to extract/obtain size-resolved chemical composition data from some experiments as examples and include them either in the main text of the manuscript or in the supplementary information. Currently, we have an AMS size-resolved chemical composition graph from one experiment, as well as some SP2 black carbon (BC) size-distribution graphs, available that could be included as representative examples.
- Thank you for the reference article. As described in the introduction, our study aims not only to inventory the κ values across different fuel types and burning phases, but also to systematically compare the κ values from three different methods, investigating how they differ and how the behaviour of the differences varies across fuel types and burning phases. All the experiments were conducted within the same experimental framework, procedures, and facilities, minimizing inter-study variability caused by different facility design-dependent gas-to-particle partitioning, ageing protocols, etc. We will re-check and try to make it clearer in the introduction while also adding Peltokorpi, S., et al., 2026 (and probably other additional relevant references)
- Thank you for the suggestion. We did not intentionally dry the fuels before combustion. We just put the fuels directly after collection in an indoor laboratory room with constant air-conditioning temperature. We measured the fuel moisture through coarse measurements using a pinless, non-destructive moisture meter, mainly just as a check, especially across repetitions of the same fuel type. However, the measurements were not used to subsequently control or prescribe a specific fuel moisture content before combustion.
Citation: https://doi.org/10.5194/egusphere-2026-2723-AC1
-
AC1: 'Reply on RC2', Sara Aisyah Syafira, 15 Aug 2026
reply
-
RC3: 'Comment on egusphere-2026-2723', Anonymous Referee #3, 03 Aug 2026
reply
This study presents a comprehensive experimental intercomparison of three methods for determining the hygroscopicity parameter κ for biomass burning aerosols. The experimental design covers multiple fuel types and combustion phases within a single chamber framework, and the dataset is substantial. Nevertheless, the current version of the manuscript mainly focuses on the discrepancies in κ values derived from different measurement methods and techniques, as well as the underlying causes of such differences. While this aspect is undoubtedly important, to strengthen the novelty of this study, the work could be further focused on the differences in hygroscopic κ values obtained from various biomass fuels. Comprehensive comparisons should be conducted against κ results reported in previous studies for fossil fuel combustion emissions, ambient atmospheric aerosols, and other organic aerosol systems to elaborate the unique characteristics of biomass-burning aerosol hygroscopicity.
The total duration of each experiment is not specified in the manuscript. It remains unclear whether hygroscopicity measurements were only conducted during the final 2–3 hours of each experiment. If so, the observed discrepancies in aerosol hygroscopicity may be largely attributed to atmospheric aging effects. A 2–3 hour aging period is sufficient to alter the chemical composition and physicochemical properties of freshly emitted biomass-burning particles. Therefore, analyzing the temporal evolution of particle hygroscopicity and other key properties throughout the aging process would provide valuable scientific insights and greatly enrich the current analysis.
Due to the lack of density data, the manuscript employs mass fractions rather than volume fractions in the chemical composition-based prediction using the ZSR mixing rule. The authors acknowledge this limitation. However, the subsequent interpretation of the discrepancies between predicted and measured κ values as reflecting real physicochemical differences is only valid if the error introduced by this mass-fraction approximation is negligible. For combustion experiments with high black carbon content, the substantial density difference between black carbon and organic matter implies that using mass fractions systematically underestimates the volume fraction of black carbon and overestimates that of organic material. The magnitude of this bias could be considerable. The authors do not quantify this uncertainty. A sensitivity analysis using reasonable density ranges for each component would help determine whether the observed discrepancies persist after correcting for this approximation.
The statistical treatment is inappropriate, with autocorrelated hourly data from the same experiment treated as independent replicates, which likely inflates the significance of the reported differences. The formatting and figure quality are also below the standard expected for a professional journal. For these reasons, the manuscript cannot be recommended for publication in its current form.
In Figure 4, many κ values derived from HTDMA measurements are below zero, which requires explicit explanation. In addition, the authors state that the fuel-type dependence of the size effect is markedly stronger than that of the supersaturation effect. However, this trend cannot be clearly observed from Figure 4. To better illustrate the size-dependent effect, we suggest moving the relevant figures from the supplementary information to the main text, which would substantially support the interpretation and discussion of Figure 4.
Many studies have compared the three methods for determining hygroscopicity. The finding that CCNC-derived κ exceeds HTDMA-derived κ for organic-dominated particles has been reported previously. The systematic documentation across multiple fuel types is a useful addition, but the central claim that κ is regime-dependent is not sufficiently novel. A more significant contribution would be a quantitative mechanistic explanation for the observed regime dependence.
The manuscript devotes a large portion of the text to the section “Investigation of the sources of discrepancy in κ”, in which the inconsistencies and uncertainties in κ values derived from different techniques and methods are discussed in detail from various perspectives. We recommend that the authors summarize these uncertainty sources and incorporate a schematic figure in this section to intuitively illustrate the origins and mechanisms underlying the discrepancies in κ determinations.
The formatting and figure quality require improvement. For example, there are unexplained colour highlights at several locations (e.g., lines 334). In Figures 2, 3, and 4, the individual panels are not properly aligned, and font sizes are inconsistent across panels of the same type. The authors are advised to thoroughly revise the formatting and improve the quality of the figures to enhance the overall presentation of the manuscript.
The authors may move the Table 4. To SI.
Citation: https://doi.org/10.5194/egusphere-2026-2723-RC3
Data sets
Dataset for "Method-Dependent Variability in Hygroscopicity Parameter (κ) of Particles from Biomass-Burning Across Fuel Types and Burn Phases" Sara Aisyah Syafira https://doi.org/10.5281/zenodo.20695234
Model code and software
Dataset for "Method-Dependent Variability in Hygroscopicity Parameter (κ) of Particles from Biomass-Burning Across Fuel Types and Burn Phases" Sara Aisyah Syafira https://doi.org/10.5281/zenodo.20695234
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 190 | 92 | 28 | 310 | 46 | 12 | 21 |
- HTML: 190
- PDF: 92
- XML: 28
- Total: 310
- Supplement: 46
- BibTeX: 12
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study aims to address why do there exist significant differences in hygroscopicity parameters (κ) measured using different methods when evaluating the hygroscopicity of biomass combustion aerosols. The authors conducted a systematic study on fresh and aged aerosols generated during both smoldering and flaming combustion stages of four representative fuels (hardwood, softwood, peat, and leaves) with a chamber.
It is a fascinating and meaningful research topic, and the author has elaborated on the experimental process in great detail. However, the full text is overly lengthy. If possible, please consider condensing it.
I have provided a few suggestions for improvement.
Line 346-348 I believe that further discussion of the detailed patterns of HTDMA-derived κ values is very important for analyzing and understanding the effects of different fuel types and burning types on κ values. This discussion can certainly be included in the manuscript.
Line 381-382 Could you elaborate in detail on what the dark aging and pre-aging stages?
Line 500-506 The size dependence of HTDMA exerts a greater influence on their measured κ values than does the SS dependence of CCNC. When it comes to assessing κ values across different fuel types, does CCNC offer greater reliability?
Line 598-601 This sentence does not help explain the increase in the BC-excluded predicted κ of flaming experiments; rather, it confuses the reader as to whether the values are under- or overestimated.
Line 633-642 Does the HTDMAmass-weighted mean size lead to lower κ values, and is this the reason why the number-weighted were adopted in Fig. 3?
Line 694-706 Based on the results of this study, for the discrete size, which HTDMA κ size should be used as the basis for comparison with the bulk predicted κ?
Section 3.5.8 The authors employed an approximation based on the ZSR mixing rule with mass-fraction weighting to predict κ. While this approach is justifiable, the impact of this choice on the extent of prediction deviation should be discussed more explicitly. For example, consider the difference between using a single κ value for all inorganic species and assigning species-specific κ values.