Characterizing aerosol sources based on aerosol optical properties and dispersion modelling in a Scandinavian Coastal Area (Aarhus, Denmark)

Teng, Zihui; Skønager, Jane Tygesen; Massling, Andreas; Skov, Henrik; Evangeliou, Nikolaos; Eckhardt, Sabine; Bilde, Merete; Rosati, Bernadette

doi:10.5194/ar-2025-34

Preprints

https://doi.org/10.5194/ar-2025-34

Preprints

07 Nov 2025

| 07 Nov 2025

Status: this preprint is currently under review for the journal AR.

Characterizing aerosol sources based on aerosol optical properties and dispersion modelling in a Scandinavian Coastal Area (Aarhus, Denmark)

Zihui Teng, Jane Tygesen Skønager, Andreas Massling, Henrik Skov, Nikolaos Evangeliou, Sabine Eckhardt, Merete Bilde, and Bernadette Rosati

Abstract. Coastal aerosols are formed through the complex mixing between marine air masses and continental emissions, which originate from both natural and anthropogenic sources. The properties of coastal aerosols are decisive for their interaction with sunlight and influence on clouds as well as the potential health implications for the population in these areas. In this study, the aerosol properties and sources at Aarhus Bay, Denmark, were investigated by combining in-situ aerosol light scattering and absorption with size distribution measurements and footprint analysis by FLEXPART. Our analysis demonstrates a considerable contribution of anthropogenic aerosols from both fossil fuel combustion and biomass burning as well as periods with highly scattering aerosols. Furthermore, good agreement was found between in-situ and modelled black carbon data. Combining in-situ measurements and FLEXPART analysis further evidenced a major impact of local emissions as well as a few long-range transport intrusions.

Received: 28 Oct 2025 – Discussion started: 07 Nov 2025

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.

Download & links

Preprint (PDF, 10425 KB)

Supplement (2497 KB)

Download & links

Preprint (10425 KB)
Metadata XML
Supplement (2497 KB)
BibTeX
EndNote

Zihui Teng, Jane Tygesen Skønager, Andreas Massling, Henrik Skov, Nikolaos Evangeliou, Sabine Eckhardt, Merete Bilde, and Bernadette Rosati

Status: open (until 19 Dec 2025)

Post a comment Subscribe to comment alert

RC1:
'Referee’s comments on the manuscript “Characterizing aerosol sources based on aerosol optical properties and dispersion modelling in a Scandinavian Coastal Area (Aarhus,Denmark)” by the authors Zihui Teng, Jane Tygesen Skonager, et al.', Anonymous Referee #1, 25 Nov 2025 reply
The present study characterises aerosol sources through dispersion modelling on optical and size distribution data measured near the coast of Denmark which overlooks Aarhus Bay.
The study includes literature comparisons of the 1-month-worth of measured data and a focus on three different showcase scenarios. This leads to a complicated and very interesting picture of the different aerosol sources at play in the Aarhus Bay area and with the FLEXPART model authors also investigate their origin.
Gaining better insights into the complex system of atmospheric aerosol is fundamental. This study focuses on coastal regions where the interaction between marine and continental air masses is still largely understudied. I recommend this article be published after the authors address the following list of comments and issues.
General comments
Environmental data is often distributed following a log-normal distribution. The heavy-tailed nature of log-normal distributions causes relatively large standard deviations compared to the distribution’s mean value. This is especially true when considering extensive parameters like σ_sca, σ_abs and eBC. To make the comparisons of section 3.1 and of the three chosen cases more statistically relevant, the authors should consider including other statistical parameters other than mean±standard deviation in the main text and not just in Table 1, e.g. medians and other percentiles.

The authors should double check all the reported wavelengths in the paper. As an example, the correct triplet of nephelometer (Aurora 3000, Ecotech) wavelengths is 450nm, 525nm and 635nm and not 470nm, 525nm and 630nm as the authors stated in page 4, line 75, and neither is it 440nm, 525nm and 635nm as stated in page 7, line 185-186. Still, Figure 2(a) and other figures and tables in the paper list the correct triplet of nephelometer wavelengths.

Specific issues and comments
Page 3, line 73: Figure S1 shows no drying of the sampling line. However, the authors should state it explicitly in the main text.

Page 3, line 76: did the authors apply the no cut correction or the sub-μm correction?

Page 4, line 80: the authors should include what autocalibration setting was used for the nephelometer measurements and how they treated nephelometer zero checks.

Page 4, line 81: the aethalometer measures the attenuation of light shone through a filter tape which is progressively loaded with atmospheric aerosol. I suggest the authors make this explicit by, e.g., writing “filter light attenuation signal”. Furthermore, how to pass from the filter attenuation signal to the absorption coefficient should be clarified.

Page 5, line 102-103: what percentage of data was eliminated due to RH above 40%? Furthermore, the authors should include a comment as to why this data filter was applied as it is the only data filter which refers to environmental values and not inherently bad data due to tube disconnections and instrument self-calibrations.

Page 5, line 107: the authors should include what hypotheses were used for the particle loss calculator (inertial impaction, gravitation deposition, diffusion, …). Moreover, maximum inlet efficiencies are usually found for particles with diameters of approximately 100-200nm and aerosol particles of interest for this paper (such as BC, BrC and secondary aerosols) can be much smaller than 1μm. I suggest the authors consider switching to a logarithmic scale on the x-axis of Figure S2 as this can help the reader better view the inlet efficiency plot at all diameters of interest.

Section 2.4: I suggest writing a more general definition of AAE, SAE and ΔSAE (as was done for SSA, b and g in equations (2), (6) and (7)) and anticipate those definitions to the beginning of the section, e.g. in line 121. Indeed, the authors throughout the whole paper use and show results found in the literature for AAE and SAE values calculated at different pairs of wavelengths to the ones of equations (3) and (4).

Page 6, line 138: isn’t the Ångström matrix obtained by subdividing the AAE-SAE biplot? I suggest the authors explain what the Ångström matrix is more thoroughly and add a reference herein to Figure S17 to better help the reader.

Page 6, line 155: while it is true that g=0 occurs for aerosol particles in the Rayleigh-scattering regime where scattering is symmetrical for the front and back hemisphere, there are cases in optics where the forward flux of cosine-weighted scattered radiation is equal to the backward one but with no symmetry between the forward and backward scattered light. As such, g=0 is more an indicator of balanced scattering between the forward and backward directions rather than symmetrical scattering.

Page 11, line 204-208: qualitative terms like “approximately” and “on average” should be replaced by more accurate scientific terminology, are the reported values means, modes or medians?

Page 12, line 254-256: Figure 4 shows median daily trends where traffic peaks are evident for weekdays and weekends only in the absorption signal. I suggest the authors refrain from using the words “similar trends” as weekday scattering is completely different to any other trend and weekend scattering peaks are not found at the same time as weekend absorption peaks. The difference between the absorption and scattering trends, both for weekdays and the weekend, suggests that non-absorbing and/or weakly-absorbing aerosols play a significant role in measured optical data.

Figure 3 shows a shorter measurement period compared to Figure 2. Furthermore, Figure S3 shows daily patterns of 7 different weeks but only 5 full weeks are shown in Figure 2 and Figure 3. Could the authors comment on this? I suggest including the whole period of seven weeks for all figures. If data are not available for the whole seven weeks, FigureS17 should either show the trends of weeks where there is complete data coverage, or the caption should explicitly state which weeks have complete data coverage.

Page 11, line 216: the authors state that the aethalometer model was used to separate the fossil fuel component from the biomass burning component at the beginning of the text. Why do the authors mention compatibility of measured AAE values to fossil fuel AAE values if they expect to find a biomass burning component and intend to determine its contributing percentage it with the aethalometer model at a later stage?

Page 14, line 283-288: the authors should include a statistical parameter which quantifies the “very good agreement” between modelled and measured BC concentrations. To give more weight and importance to the FLEXPART analysis, the authors should also consider including comments on: (1) why the modelled results of Figure 5 show a smoothed-out trend compared to measured values; (2) why sharp BC peaks are never correctly modelled; (3) why the model underestimates BC concentrations and if this underestimation is found in any other literature.

Pages 14-15, lines 290-304: could the authors also include the same statistical parameters of the previous point for the three showcase scenarios and comment on how the modelled results of the selected cases compare to the modelled results of the whole period? Is there a better or a worse agreement for Case1, Case2 and Case3 compared to the whole period?

Figure S8 and Figure S10 show missing NOx data for a considerable portion of the Case 1 scenario (which is the local pollution case) and NOx is a tracer of local emissions as stated by the authors in page 15, line 310. Maybe this is a stupid question, but can the lack of NOx data influence modelled results substantially? If this is the case, please add a comment in the paper.

Page 14, line 287: I would consider a percentage of 3% for the industrial emissions as being a minor contribution too.

Page 15, line 308-309: I would revise the phrase as follows “NOx and CO are often used as tracers of pollution as they are mainly emitted in fossil fuel combustion and biomass burning”. Any type of open-air combustion occurring at a high temperature can break the bonds of N₂ and O₂ (which are the main atmospheric gases) thus forming NO which subsequently can oxidise to NO₂. Moreover, nitrogen is essential for plant growth and is found abundantly in any plant, for these reasons NOx can be emitted in wood-burning processes too and should also be included as possible contributing sources on page 16, line 316. One should also note that BC and not only BrC is emitted in wood burning processes too.

Page 17, line 344: averages are a more general definition than mean. I suggest that authors use more accurate statistical terminology.

Page 19, line 392-394: This is the first occurrence where the authors discuss possible limitations of the aethalometer model. Indeed, the aethalometer model assumes that the absorption signal is due to only two absorbing aerosol ensembles, one with a high wavelength dependence and the other with a low wavelength dependence (AAE=2 and AAE=1 was the pair of Ångström exponents employed by the authors respectively for biomass burning and fossil-fuel combustion). Model limitations list the need to find site-specific AAE values to fit the data correctly, that particles with low AAE values can be emitted in wood burning processes too and that there are other absorbing aerosols other than BB and FF aerosol in the atmosphere such as mineral dust. For the above reasons, the aethalometer model offers only a coarse view of what is truly found in the atmosphere and its results should be taken with a pinch of salt. I suggest the authors explain the aethalometer model more thoroughly and discuss some of its limitations in the main text (e.g. when it is introduced in section 2.2 in page 4, line 88-89) as the aethalometer model is extensively used in the study.

Typographical and grammatical issues
The following is a list of revised phrases where typos and grammatical errors were revised.
Page 2, line 33: please revise the last phrase. An option could be the following “have contributed to decades-worth of in-situ aerosol optical …”.

Page 3, line 54: please revise “carries out” to “carried out”

Page 3, line 57: please revise to “habitats which include wildlife such as herds of deer and…” possibly other wildlife.

Page 4, line 86: “deployed” should be changed to “employed”.

Page 6, line 154: please eliminate "the" and revise to “angular distribution of scattered radiation”.

Page 7, line 178: please eliminate "the" and revise to “of fire emissions” and “of BB dispersion”.

Page 7, line 178: “crucial for an accurate simulation of …” or “accurate simulations of …” or “crucial for the accurate simulation of …”.

Page 7 line, 187-188: “Many sharp peaks were only visible in σ_abs and were absent in σ_sca, indicating the presence of strongly absorbing particles for short periods of time”.

Page 14, line 273: “during weekends”.

Page 14, line 277: “a lower PBL hinders dispersion thus favouring higher concentrations”.

Page 18, line 358: I would revise to “thus supporting the hypothesis of marine-particles intrusion”.

Page 18, line 362: the wording “this is also clearly visible” suggests that a different analysis was carried out and confirmed the FLEXPART results. I suggest the authors revise the phrase to “The FLEXPART footprint analysis for Case 3 is shown in Figure S7” as the authors are still talking about FLEXPART model results.

Page 19, line 385-386: “reconstructed by the model” or “extracted from the model”.

Page 20, line 408: “with the FLEXPART model was carried out”.

Reply
Citation: https://doi.org/10.5194/ar-2025-34-RC1
RC2: 'Comment on ar-2025-34', Anonymous Referee #2, 25 Nov 2025 reply

General Comments
The manuscript presents an ambitious and multifaceted characterization of aerosol optical and microphysical properties in a Scandinavian coastal environment, supported by dispersion modelling using FLEXPART. The combination of in-situ observations with source–receptor modelling is valuable, and the multi-instrument dataset collected over several weeks offers substantial scientific potential.
However, some methodological aspects require clarification—particularly the inlet configuration, humidity handling, nephelometer corrections, aethalometer assumptions, and instrument calibration procedures. Since these factors directly affect σ_sca, σ_abs and eBC, their explicit and consistent reporting in the main text is essential.
In addition, several optical parameters and time series are summarized using mean ± standard deviation, but the underlying distributions are likely skewed. Including median and percentile metrics directly in the main text would improve the scientific robustness of comparisons among the three case studies.
Finally, the selection criteria for the defined periods, the interpretation of optical exponents, and the model–measurement comparisons would benefit from clearer justification and more transparent presentation. With these refinements, the study would provide a more solid contribution to coastal aerosol research.

Specific Comments
Page 2, line 58–63: the inlet description does not clearly state whether the sampling line included drying or conditioning. Since RH strongly affects σ_sca, this should be explicitly clarified in the main text.
Page 3, line 65: the Introduction refers to a field campaign lasting 6.5 weeks. However, the Methods section specifies measurement dates from March 3rd to April 11th, 2023, which corresponds to approximately 5.7 weeks (40 days). The authors should revise the reported study duration in the Introduction to ensure consistency with the dates provided in the Methods section.
Page 3, line 72–74: please specify whether the nephelometer data were processed using the “no-cut” correction or the “sub-micron” correction. This choice significantly affects scattering coefficients and Ångström exponents.
Page 4, line 78–82: the autocalibration configuration for the Aurora 3000 should be described, along with the treatment of zero checks (frequency, duration, and how zero offsets were applied).
Page 4, line 85–88: the wavelengths reported for the nephelometer appear inconsistent across the manuscript. Please verify and harmonize all wavelength triplets for scattering and absorption instruments.
Page 4, line 95–97: the WELAS Optical Particle Spectrometer (OPS) measures particle optical diameters in the range of approximately 0.2 µm to 10 µm. This lower limit 0.2 µm excludes the nucleation mode and a substantial fraction of the Aitken mode (ultrafine particles). Given the study's focus on combustion-related sources (traffic, domestic heating), which are major contributors to the ultrafine fraction, the authors must explicitly discuss the potential impact of this 0.2 µm cut-off on the derived total particle number concentrations and the source apportionment results, acknowledging the likely underestimation of ultrafine particle counts.
Page 5, line 115–120: clarify whether the sampling system and tubing efficiency calculations were used to correct measured concentrations, or only to estimate potential biases. Currently, the text does not explicitly state how inlet losses were applied.
Page 6, line 150–160: a more detailed explanation of the criteria used to define the three case periods would improve reproducibility. Thresholds for SSA, SAE, AAE or FLEXPART footprint characteristics should be reported.
Page 7, line 185–195: optical parameters such as AAE and SAE are interpreted as indicators of dominant particle size and source. Consider including a short discussion of the sensitivity of these quantities to mixing state and chemical composition.
Page 8, line 210–220: the interpretation of high scattering periods should consider the role of sea-salt coarse particles. Because the inlet figures show size-dependent losses above ~6–8 μm, these limitations should be acknowledged in the discussion.
Page 9, line 240–255: the comparison between measured eBC and FLEXPART BC would benefit from including additional statistical indicators (median difference, percentiles), since eBC distributions are typically log-normal.
Page 10, line 265–275: if FLEXPART under- or overestimates BC for certain case studies, potential reasons (emission inventory uncertainties, domestic heating patterns, atmospheric mixing) should be explored more explicitly.
Technical Corrections
Page 1, line 20–22: minor grammar refinement recommended to improve flow of introductory paragraph.
Page 2, line 55: replace “was carries out” with was carried out.
Page 3, line 90–92: ensure consistency in unit formatting (e.g., Mm⁻¹, μg m⁻³, # cm⁻³).
Page 4, line 75: correct wavelength notation for nephelometer channels; ensure format is uniform throughout the text.
Page 5, line 130–132: several sentences are overly long; consider splitting for readability.
Page 6, line 155: abbreviations for FLEXPART sectors (DOM, TRA, SHP, etc.) should be introduced at first occurrence in the main text, not only in figure captions.
Page 7, line 182–185: ensure spacing between numbers and units is consistent (e.g., “450 nm” instead of “450nm”).
Page 8, line 210: check figure cross-references; some appear out of order (e.g., S15–S18 referenced before S12–S14).
Page 9, line 235–240: equation formatting should be standardized to match journal style (subscripts, λ-notation, exponents).
Page 10, line 280–285: typographical inconsistencies in references to case-study periods; ensure identical naming across text, figures and captions.

Reply

Citation: https://doi.org/10.5194/ar-2025-34-RC2

Zihui Teng, Jane Tygesen Skønager, Andreas Massling, Henrik Skov, Nikolaos Evangeliou, Sabine Eckhardt, Merete Bilde, and Bernadette Rosati

Supplement

https://doi.org/10.5194/ar-2025-34-supplement

Zihui Teng, Jane Tygesen Skønager, Andreas Massling, Henrik Skov, Nikolaos Evangeliou, Sabine Eckhardt, Merete Bilde, and Bernadette Rosati

Viewed

Total article views: 238 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
181	42	15	238	14	12	16

HTML: 181
PDF: 42
XML: 15
Total: 238
Supplement: 14
BibTeX: 12
EndNote: 16

Views and downloads (calculated since 07 Nov 2025)

Month	HTML	PDF	XML	Total
Nov 2025	181	42	15	238

Cumulative views and downloads (calculated since 07 Nov 2025)

Month	HTML	PDF	XML	Total
Nov 2025	181	42	15	238

Viewed (geographical distribution)

Total article views: 241 (including HTML, PDF, and XML) Thereof 241 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 27 Nov 2025

Download

Preprint (10425 KB)
Metadata XML

Short summary

Coastal aerosols in Scandinavian urban areas remain understudied. We examined aerosol optical properties and size distributions along Denmark's coastline. Combining in-situ data with dispersion modelling, we identified two main aerosol types: carbonaceous aerosols from fossil fuel and biomass burning and large, highly-scattering aerosols (potentially sea salt). Black carbon from in-situ data and dispersion modelling correlated well while FLEXPART underestimated the concentration.


Total:	0
HTML:	0
PDF:	0
XML:	0