A new radiosonde system for measuring the vertical variability of aerosol single scattering properties

Broda, Michalina; Zawadzka-Mańko, Olga; Chiliński, Michał T.; Nurowska, Katarzyna; Kłapiński, Szymon; Makuch, Przemysław; Markowicz, Krzysztof M.

doi:10.5194/egusphere-2025-6467

Preprints

https://doi.org/10.5194/egusphere-2025-6467

Preprints

05 Jan 2026

| 05 Jan 2026

A new radiosonde system for measuring the vertical variability of aerosol single scattering properties

Michalina Broda, Olga Zawadzka-Mańko, Michał T. Chiliński, Katarzyna Nurowska, Szymon Kłapiński, Przemysław Makuch, and Krzysztof M. Markowicz

Abstract. This study presents a newly developed balloon-borne AeroSonde system equipped with calibrated low-cost optical particulate matter sensors (SPS30) for vertically resolved aerosol optical measurements. Calibration of the sensors against reference instruments (Aurora 4000 nephelometer and LAS 3340) enabled the estimation of scattering coefficients (ASC), scattering Ångström exponent (SAE), and effective particle radius (R_eff), with Pearson correlation coefficients exceeding 0.88. The best agreement was obtained for ASC, while SAE and R_eff showed substantially larger uncertainties, limiting their use to qualitative discrimination between fine- and coarse-mode particles.

The AeroSondes were deployed for the first time during measurement campaigns in June and August 2025 over southeastern Poland. Soundings with vertical resolution below 10 m reached altitudes of 10–27 km a.g.l. during episodes of long-range transport, including North American biomass-burning plumes and Saharan dust intrusions. The vertical distribution of aerosol layers observed by the AeroSonde is consistent with active remote sensing measurements, showing good agreement in both the optical layering structure and the absolute values of extensive aerosol properties, with correlation coefficients ranging from 0.74 to 0.81. The extended vertical sounding range allows the estimation of column-integrated quantities, such as aerosol optical depth, as well as their contributions from individual layers, for example, within the planetary boundary layer.

These results demonstrate that low-cost, balloon-borne AeroSonde sensors can provide reliable vertical profiles of aerosol properties, supporting enhanced spatio-temporal coverage and improved evaluation of atmospheric aerosol models.

Received: 24 Dec 2025 – Discussion started: 05 Jan 2026

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

Michalina Broda, Olga Zawadzka-Mańko, Michał T. Chiliński, Katarzyna Nurowska, Szymon Kłapiński, Przemysław Makuch, and Krzysztof M. Markowicz

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-6467', Anonymous Referee #1, 09 Mar 2026
The manuscript presents a system for aerosol vertical profiling, but the current form suffers from some structural, methodological, and interpretative weaknesses that prevent the scientific results from emerging clearly. A recurring issue throughout the text is the imbalance between extensive descriptive passages and the lack of corresponding scientific analysis. The new system was tested with only five vertical profiles. More comparisons with lidar would be expected to assess optical properties, and comparisons with aerosounding data would also be useful to support the presented results. Many sections, particularly Section 4.1, read more like a campaign or field‑log report than a scientific study. Long situational descriptions are repeated multiple times, while essential methodological information (e.g., uncertainties, calibration procedures, justification of regression shape, model references, data quality criteria) is missing or insufficiently explained.
The manuscript is also very difficult to navigate. Figures and tables are placed far from the referencing text, many figures duplicate information already shown elsewhere, and several figures or panels are not referenced. The appendices and supplemental materials are similarly hard to follow, with unclear placement and redundant content.
Overall, the manuscript requires major restructuring, removal of redundant descriptive content, clearer methodological explanations, improved figure organization, and stronger scientific interpretation.
Specific comments
L33: Is there any link or reference available for the cited measurements?
L47–49 could be moved to the Methods section.
Section 2.2.1: The description would be clearer if the sensor were described first, followed by Vaisala, then data acquisition, then the GPS, and finally the detachment system. The current order is misleading. L75: The power consumption is given only for the GPS sensor; why is the power consumption of the other system components not provided?
L85: Can the height of the lower and middle atmosphere be estimated in kilometers? Can the distance from the launching site also be given in kilometers? What happens to the sample if it cannot be retrieved? Are there any height restrictions or approvals required to operate the new system?
Section 2.2.2: An introductory explanation of the use of the presented systems could be added. Citations to studies that used the lidar and sunphotometer could be included. Bernardoni et al. (2021) is not an original citation describing the working principle. L129 What mass absorption coefficient was used for the equivalent black carbon calculations?
Section 2.2.3: An introduction explaining the purpose of the presented models could be provided. L159: Are there any recent works on the LAGRANTO model that could be cited?
L172: Why is the correction not introduced first, followed by the description of how the corrected data are used? Is there a citation for the noise detection algorithm? Why was the flagging threshold set at 45% and not, for example, 50%? How many data points were removed due to flagging?
What is meant by the “transported layer”? L191 How is the transported layer defined within the atmosphere?
L227: What is the origin of the horizontal wind speed data?
L235: What is the correlation at the original time resolution of the aerosol data? The correlation at the original time resolution would be more relevant for high-frequency measurements than hourly averages.
L238: Can the term “very high concentration” be quantified?
L243: Could the 5-second sensor response time be a limiting factor for atmospheric measurements?
Why are all regressions non-linear? What is the physical reason for the chosen regression shapes? Can the regression choices be supported by references? It is unclear where Equation 5 comes from.
L270: The SPS30 does not appear to be very useful for typical particle sizes below 0.5 µm. Therefore, the SPS30 may not be very useful for biomass burning aerosol?
Why are the data averaged across all instruments in Figure 3 instead of being shown individually? Each sounding uses only one unit.
June 10: The near-surface inversion is not acknowledged in the text (L318).
Why are cloud presence and cloud descriptions so emphasized (e.g., L346–349), if they do not play a role in the results?
A large part of the text is devoted to describing trajectories, while the key information is simply whether they originated in Canada or not. Therefore, most of this text could be shortened.
Why were the vertical profiles from the system not compared with WMO station data?
L399: It is not possible to claim that the stratosphere was unaffected because no measurements were taken there.
It is not clear why two different Klett approaches were used. One would expect a methodological suggestion at the end of the MS?
It is unclear how the measurement uncertainty was calculated in Figure 10.
Was the temporary reduction in Reff artificial due to abrupt atmospheric changes and the low spatial response of SPS30, or do the authors expect it to represent real behavior?
L500: The upper limit of the aerosol layer appears to be clearly connected to the existence of the inversion layer, as AOD decreases from 20 to 3 Mm⁻¹. Values above the inversion would likely not be connected to dust.
L626: If the new system used Raman lidar, what additional information would it provide beyond using the lidar alone?
Figures:
Figure 1 could be combined with Figure 2 if rotated by 90 degrees.

Figure 3C is not referenced in the text.

In Figure 3, the Pearson correlation, RMSE, and equation are not necessary because they are already given in the text.

Figures 5 and 6 could be combined into a single figure for better comparison.

Figure B3: Why is the “anthropogenic” category written in capital letters?

Figure 5B is not referenced in the text.

The information in Figure 4 is completely duplicated in Figures B4, C4, C3, and D3. Figure 4 could therefore be removed.

Figure 7C presents the same information already included in the C3 plot and therefore duplicates the information. This duplication also applies to other figures with trajectories.

Figures 10 and 11 could be combined into a single figure. Part of the Figure 10 label would be more appropriate in the Methods section rather than in the figure label.

Figure 13: Why is the agreement of SAE and RR with the reference instruments and ground measurements so poor?

Figure 15 could be moved to the Supplementary Information because the information it contains is already described in the text.

Figure A1b: The resolution of the y-axis should be improved because the values are currently impossible to read.

Figure B1 could be combined with Figure B2. Figure B2: The starting height of the model trajectory could be highlighted. The same highlighting could be applied to other similar figures.

Figures D2 and C2 atc. appear more relevant to flight planning than to the results and may not be necessary.

Tables:
Table 1 could be moved to the Supplementary Information. In the main text, the information from Table 1 could be summarized in a sentence.
Table 3: Does it mean that there is a maximum three-hour time difference? This is unclear from the description.
Minor comments
The manuscript does not describe single scattering properties but rather bulk scattering properties of the aerosol; therefore, the term “single scattering” should be removed from the entire manuscript, including the title.

Abbreviations are not defined at their first appearance (L3, L9, 29, 48, 58, 61, 88, 154, 244, etc.). If an abbreviation is not used later in the text, it is not necessary to define it.

L39: It is unclear what is meant by “high rise installations.”

Note 1 is not necessary and could be removed.

L191: The description of mean and standard deviation could be written as mean ± standard deviation instead of a full sentence. The same issue applies to L193, L194, L206, L207, etc.

L218: The abbreviation AEC stands for aerosol extinction coefficient.

L238 and L309: Why is the mean difference given in micrometers rather than nanometers?

L398: The inversion does not increase with height; it disappears.

Lines 433–438 and 486–492 repeat similar information and could be omitted.

Lines 438–443 repeat the layer positions already described in Section 4.1 without a clear purpose.

Sentences starting at L445 and L548 convey the same information; one of them could be removed.

RMSE has a unit, and this unit should be included in Figure 15.

L576: The uncertainties mentioned in the text have units, but the units are missing.

L581: The PBL itself does not contribute to AOD, but the aerosol trapped within the PBL does.

In the appendices, both text and supplementary figures are included in the references, which makes navigation very difficult.

Appendix D is not necessary because the situation description is already fully given in the text.
Citation: https://doi.org/10.5194/egusphere-2025-6467-RC1
RC2:
'Comment on egusphere-2025-6467', Anonymous Referee #2, 12 Mar 2026
General assessment
This manuscript presents an observational analysis of aerosol and dust layers using the novel balloon-borne AeroSonde system and multiple ground-based remote-sensing instruments, trajectory analysis, and supporting observations. The study aims to demonstrate that low-cost balloon-borne AeroSonde sensors can provide reliable vertical profiles of aerosol properties and improve spatio-temporal observational coverage for aerosol model evaluation.
The topic is relevant for the atmospheric science community, particularly for studies of long-range aerosol transport, and vertical aerosol structure. The use of multiple datasets and observational approaches has the potential to provide useful insights into aerosol layering and transport pathways.
However, in its current form the manuscript is overly lengthy and contains a large amount of descriptive material and repetitive figures, which makes it difficult to clearly identify the main scientific contributions and new findings. Several sections require clarification of methodology, improved structure, and stronger interpretation of the results. In addition, instrument uncertainties and data interpretation aspects are either missing or insufficiently discussed.
For these reasons, I believe the manuscript requires major revision before it can be considered for publication. Below I provide detailed comments intended to help the authors improve the clarity, scientific robustness, and readability of the work.
General comments
The manuscript is very dense and contains a large amount of descriptive information, some of which does not directly contribute to the main scientific findings. The authors are encouraged to:
Shorten descriptive sections where possible.

Focus on the key scientific objectives and new findings.

Reduce repetition between sections.

Move supporting or secondary information to supplementary material if appropriate.

The manuscript should include a clearer discussion of sensor calibration, measurement uncertainties, instrument intercomparability and possible biases in aerosol optical and particle measurements. This is particularly important when combining multiple datasets.
Since the Sensirion SPS30 is primarily designed as a particulate matter (PM) sensor, it would be important for the manuscript to present the PM measurements obtained. Currently, the discussion focuses mainly on derived optical quantities (e.g., ASC, SAE), while the primary measurements reported by the sensor are not shown or analyzed in detail. Including the PM time series and/or vertical profiles would help readers better understand the original sensor response, its variability with altitude, and its relation to the retrieved optical properties. Furthermore, presenting the PM data would allow a clearer assessment of coarse-mode particle contributions and potential sensor limitations under dust-dominated conditions, which are relevant for several of the case studies discussed in the manuscript. I recommend that the authors include representative PM profiles or distributions in the results section and discuss how they relate to the derived aerosol optical parameters.
Figures 5–9 appear repetitive and may not all be necessary in the main manuscript. The authors may consider presenting one representative figure in the main text to illustrate the methodology and key findings, while moving the remaining figures to the Appendix or Supplementary Material. This would help improve the readability and flow of the manuscript while still making the full set of results available to interested readers.
Several parts (particularly in the results) read more like descriptive reporting of meteorological conditions rather than scientific analysis. The authors should improve the discussion by interpreting the observations more clearly, linking results to the study objectives, explaining the physical mechanisms behind the observations, highlighting the new insights provided by the study. For instance, Sections 4.1 and 4.2 require improvement. Currently it reads more like a general meteorological description rather than a scientific analysis.
Specific comments
Line 35: also balloon-borne observations, recommended papers to cite:
Kalnajs, L. E., & Deshler, T. (2022). A new instrument for balloon-borne in situ aerosol size distribution measurements, the continuation of a 50 year record of stratospheric aerosols measurements. Journal of Geophysical Research: Atmospheres, 127, e2022JD037485. https://doi.org/10.1029/2022JD037485
Kezoudi, M., Tesche, M., Smith, H., Tsekeri, A., Baars, H., Dollner, M., Estellés, V., Bühl, J., Weinzierl, B., Ulanowski, Z., Müller, D., and Amiridis, V.: Measurement report: Balloon-borne in situ profiling of Saharan dust over Cyprus with the UCASS optical particle counter, Atmos. Chem. Phys., 21, 6781–6797, https://doi.org/10.5194/acp-21-6781-2021, 2021.
Line 50: “in the absence of reference measurements”, here authors mean reference in-situ?
Sections 4.1 and 4.2 require improvement. It would be beneficial to shorten and merge the sections in order to create a more coherent and clear narrative that connects meteorological conditions, aerosol observations and transport pathways.
Line 323: “revealed aerosol layers from ground up to 6.5 km”. Please clarify: Which instrument detected these layers? How were aerosol layers distinguished from clouds?
Line 327: Transport from Canada is mentioned. Please clarify: At which altitude this transport occurred. Whether the authors checked for wildfire activity in Canada on that day. If so, please provide supporting references or datasets.
Line 329: AERONET observations are mentioned. Please specify: The exact site of the AERONET station, distance from the measurement location, whether the observations were collocated in time with the event
Lines 330–333: When discussing cloud layers, it would be useful to include: Cloud base and top heights, cloud thickness, cloud phase (if known). This would provide a clearer picture of the atmospheric structure during the event.
Line 351: “All LAGRANTO trajectories…” Please clarify whether the word “all” refers to: all altitudes, all trajectory starting points, all time steps. The trajectory configuration using LAGRANTO trajectory model should also be described more clearly.
Line 352: “the trajectories were relatively straight and consistent”. This statement is vague. Please clarify in terms of what metric, for example: directional variability, wind shear, transport pathway stability
Line 360: “What does this mean for the Poland region?” The interpretation of the trajectory results over Poland needs clarification. It is not clear what implication the authors intend to draw for this region.
Line 369: The term “dust cloud” is used. Please clarify whether this refers to: a dust layer, dust mixed with cloud droplets or a dust plume. The terminology should be more precise.
Line 384: Strong winds are mentioned. Please discuss whether instrument performance may have been affected by strong wind conditions, including sampling efficiency, flow stability, measurement bias.
Line 392: “The reported ASC value…”. Please clarify: From which instrument this ASC value is derived. How it was calculated
Line 409: states that Reff increases with altitude. Please explain: Why is this behavior observed? Up to which altitude does this trend persist? Whether this could be related to size-dependent transport or another reason?
Lines 535–539: The explanation of discrepancies needs to be expanded. Reason 1 mentions that mineral dust can substantially increase AEC due to absorption. Please quantify the associated uncertainty and provide references. Reasons 2 and 3: Why was information from the ceilometer depolarization ratio not used to correct or constrain the observations from the SPS30 sensor? This could potentially improve the interpretation of the particle measurements.
Paragraph starting line 542: states that large particles were present in the layer, but no direct PM measurements are shown. How was this conclusion derived? Whether PM size distribution data are available? If so, these measurements should be included in the manuscript.
Lines 608–614: The conclusions provide a useful summary of the study; however, several statements appear somewhat stronger than what is supported by the uncertainty analysis presented earlier in the manuscript. Given the relatively large uncertainties reported for some derived parameters (particularly SAE and Reff), the authors should ensure that the conclusions remain fully consistent with the quantitative limitations of the measurements and retrieval methods.
Lines 615–617: mentions that the vertical profiles of ASC measured by the AeroSonde are consistent with simultaneous lidar observations. While the comparison shown earlier in the manuscript is promising, this statement would benefit from explicit quantitative metrics, such as correlation coefficients, mean bias or normalized bias, RMSE values. Including such metrics in the conclusions would provide stronger support for the claimed agreement.
Lines 618–620: The reported AOD bias range of 0.01–0.12 relative to AERONET observations would benefit from additional context. For example, discussing whether this level of agreement falls within the expected uncertainty range of measurements.
Lines 621–623: The statement that the AeroSonde system is suitable for quantitative estimation of SAE and Reff should be reconsidered or slightly moderated. Earlier sections of the manuscript report substantial uncertainties, particularly for SAE (up to ±1.0), which may limit its applicability for robust aerosol-type classification. The authors may therefore wish to clarify that these parameters are primarily useful for qualitative interpretation or indicative characterization, rather than precise quantitative retrieval.
Lines 624–628: The future work section appropriately highlights the need for improved calibration strategies. Given the importance of coarse-mode aerosols (e.g., mineral dust) in several of the presented case studies, it would be useful to explicitly mention improved calibration for coarse particles, further validation against reference-grade instruments, and expanded deployments under different aerosol regimes.
Technical comment: The ordering of references in the text appears inconsistent. In some parts of the manuscript, references appear to be sorted by relevance, while in other cases they appear alphabetical or chronological. The authors should choose one consistent citation style and apply it.
Citation: https://doi.org/10.5194/egusphere-2025-6467-RC2

Michalina Broda, Olga Zawadzka-Mańko, Michał T. Chiliński, Katarzyna Nurowska, Szymon Kłapiński, Przemysław Makuch, and Krzysztof M. Markowicz

Viewed

Total article views: 489 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
312	155	22	489	21	18

HTML: 312
PDF: 155
XML: 22
Total: 489
BibTeX: 21
EndNote: 18

Views and downloads (calculated since 05 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	180	75	12	267
Feb 2026	43	33	3	79
Mar 2026	89	47	7	143

Cumulative views and downloads (calculated since 05 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	180	75	12	267
Feb 2026	43	33	3	79
Mar 2026	89	47	7	143

Viewed (geographical distribution)

Total article views: 459 (including HTML, PDF, and XML) Thereof 459 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 25 Mar 2026

Short summary

This study shows that balloon-borne instruments with low-cost particle sensors can reliably measure how airborne particles are distributed with height in the atmosphere. The sensors were tested during field measurements in southeastern Poland which identified particle layers transported over long distances including wildfire smoke and desert dust. The results indicate that this approach can improve aerosol observations and support better representation of aerosol processes in atmospheric models.


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