The applicability and challenges of black carbon sensors in monitoring networks

Elomaa, J. Tapio; Luoma, Krista; Harni, Sami D.; Virkkula, Aki; Timonen, Hilkka; Petäjä, Tuukka

doi:https://doi.org/10.5194/ar-3-293-2025

Articles | Volume 3, issue 1

https://doi.org/10.5194/ar-3-293-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/ar-3-293-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 3, issue 1

Research article

|

03 Jun 2025

Research article |

| 03 Jun 2025

The applicability and challenges of black carbon sensors in monitoring networks

J. Tapio Elomaa, Krista Luoma, Sami D. Harni, Aki Virkkula, Hilkka Timonen, and Tuukka Petäjä

Download

Final revised paper (published on 03 Jun 2025)
Preprint (discussion started on 15 Apr 2024)

Interactive discussion

Status: closed

RC1:
'Comment on ar-2024-12', Anonymous Referee #1, 24 May 2024

The manuscript presented a study of the spatial resolution of ambient black carbon collected in parallel with different types of black carbon sensors. This study is relevant and useful for atmospheric researcher. The study is very informative and well-structured. I, therefore, congratulate the authors for their efforts.
That is why I will suggest publication with minor corrections, namely:
Careful checking of the English language and the text as a whole is necessary.
There are also a few misstatements which I would advise correcting.
The author should change the title. It does not only examine the spatial variability of ambient black carbon.
In introduction: Please correct PM2.5 to PM_2.5.
Page 11, Paragraph 210: A sharp ATN change |ATN| >30 was manually identified. What could be the reason for this sharp change? Please explain it.
Page 11, 3.1 section: The authors used different flow rates for the same type of sensors. This is sometime hard to follow in the article. Please create a table summarizing this. How much during each campaign, etc.
Page 12, Paragraph 255: The authors wrote that the 1st intercomparison has on average lower concentrations compared to the 2nd intercomparison. Is it not because of the different meterological condition? Please explain it.
Page 13, Figure 2: Please check the labels (date) at the xaxis. Please correct it.
Page22, Paragraph 400: The authors wrote the following: With the MA-series sensors (MA200, MA350) the change of the temperature and RH caused clearly erroneous data as seen in Fig. 12
However, we cannot see the results of MA200 sensor in the Figure 12. Please include its results in the figure.
The conclusion contains some statements that need to be clarified. For istance: what is DST? please explain it.

Citation: https://doi.org/10.5194/ar-2024-12-RC1
- AC1: 'Reply on RC1', Tapio Elomaa, 30 May 2024
  
  We thank th Referee #1 for the valuable comments on the manuscript. Please find the point-to-point answers attached.
  
  Citation: https://doi.org/10.5194/ar-2024-12-AC1
RC2:
'Comment on ar-2024-12', Anonymous Referee #2, 28 Jun 2024

The manuscript presented a comparative study on understanding the performance of portable and low-cost black carbon monitors available in the scientific community. This study is relevant and useful as it provides information on BC monitoring devices that can be used for large-scale spatial and temporal monitoring across regions. Results from this manuscript can enrich the scientific community and regulatory bodies. Hence, I recommend this manuscript be published with some major corrections.
General suggestions:
1. A language check is necessary for the entire manuscript. Some segments of the text do not follow common practices of reporting data or manuscript writing, such as consistently reporting units of mass concentration or time.
2. Section titles require renaming (based on the updated context), particularly in Section 2 (Methods) and Section 3 (Results).
3. I highly recommend adding more text discussing the results and making section 3 from "Results" to "Results and Discussions," which is a typical nature of manuscript framing in similar types of scientific journals.
4. The title of the manuscript can be updated as the manuscript content does not presently match with the title.
5. Use PM_2.5 everywhere
6. Line 71: Why is MAAP chosen as a reference device to compare with Aethalometers, which work slightly differently? AE33 could be a great reference device in this work. If the authors have AE33's data, adding a comparison might help understand the unit-specific offsets in measuring BC. If they don't, I highly recommend providing sufficient justification (here in Line 71) or in the methods section why MAAP is used or better suitable.
7. Line 85: Some theoretical mistakes were identified. Please check the literature and correct it. Aethelometers measure light intensity and calculate light attenuation (ATN). From ATN measurements, ATN coefficients (b_ATN) are derived, followed by the absorption coefficient (b_abs). Drinovec 2015 explained this well.
8. Line 105: There are inconsistencies around the assumptions made. If any scattering correction is made, why are they assumed to be unity? More explanations are required. Also refer to Line 152.
9. Line 118: Leakage factor changes over time. Please refer to Drinovec 2015. This assumption might change the corrections, and so might the final corrected BC concentration. Please reconsider adapting to such changes.
10. It is not clear if the MA series device's inbuilt Dual-spot corrections were used for comparison or not. Typically, MA devices have their own correction mechanism, which is not the same as Drinovec 2015. Please check and confirm with a table of corrected data or add it to Table 2.
11. Some comments on figures have been mentioned in the attached file, which mainly focuses on the visibility of the graphics and texts.
12. Section 3.1.2: A separate segment on the sensor calibration in methods sections is recommended. Also, mention how the calibrations are assessed (metrics used, such as slope, MAE). Reporting the calibration results are helpful for future studies. While presenting the data in "Result" section, please compare it with previous literature. A table might be helpful for reporting metrics from uncalibrated data, calibrated data, and literature data - with the type of calibration procedure adopted. All these elements will expand this section, which I believe is going to improve the quality of the manuscript.
13. Generally, the results and discussion section is missing references from previous literature. Please compare the reported BC levels with different devices under this work and the levels reported in previous studies in similar regions or of similar spatial characteristics.
14. Section 3.2.4: Change the section title. I recommend removing this section and adding a detailed discussion explaining the variability in previous sections. If the authors want to keep this section, will this require some detailed analysis explaining how spatial variability is captured by different devices? If there are any differences in performances? Also, some discussion was required about how spatial variability can be captured by these devices. If the true spatial variability is higher/lower than the inter-device variability studied. Finally, some recommendations /comments would be helpful for the community, such as which device performed best in what context.
15. After restructuring the result and discussion section - please update the conclusion accordingly. Avoid including conclusions that have not been discussed well in previous sections (For example, Line 436).
Some minor comments can be found in the attached file.

Citation: https://doi.org/10.5194/ar-2024-12-RC2
- AC2:
  'Reply on RC2', Tapio Elomaa, 17 Sep 2024
  We thank the Referee #2 for the valuable comments on the manuscript. Please find the point-to-point answers to the comments below.
  A language check is necessary for the entire manuscript. Some segments of the text do not follow common practices of reporting data or manuscript writing, such as consistently reporting units of mass concentration or time.
  
  The manuscript has been checked in order to catch any deviancies from the guidelines.
  Section titles require renaming (based on the updated context), particularly in Section 2 (Methods) and Section 3 (Results).
  
  Following section titles have been updated. Section 2.2 “Dualspot correction algorithms”. 2.4 “Deployment area”. 3.1 “Intercomparison periods”. 3.1.1 “Applicability of the dualspot corrections”. 3.2.2 “Weekly features in BC concentration”
  I highly recommend adding more text discussing the results and making section 3 from "Results" to "Results and Discussions," which is a typical nature of manuscript framing in similar types of scientific journals.
  
  The title has been adjusted to the suggestion.
  The title of the manuscript can be updated as the manuscript content does not presently match with the title.
  
  The title has been changed to:
  The applicability and challenges of black carbon sensors in dense monitoring networks.
  Use PM_2.5everywhere
  
  Corrected.
  Line 71: Why is MAAP chosen as a reference device to compare with Aethalometers, which work slightly differently? AE33 could be a great reference device in this work. If the authors have AE33's data, adding a comparison might help understand the unit-specific offsets in measuring BC. If they don't, I highly recommend providing sufficient justification (here in Line 71) or in the methods section why MAAP is used or better suitable.
  
  Unfortunately, no AE33 data was available during this campaign. MAAP was utilized as the reference instrument as it was the highest-grade instrument available during this period.
  Line 85: Some theoretical mistakes were identified. Please check the literature and correct it. Aethelometers measure light intensity and calculate light attenuation (ATN). From ATN measurements, ATN coefficients (b_ATN) are derived, followed by the absorption coefficient (b_abs). Drinovec 2015 explained this well.
  
  Error fixed. “The measured variable by the instrument is the attenuation coefficient b_atn(λ) [m^-1] calculated from the measured attenuation and the operational parameters of the instrument as described in Eq. 2.”
  Line 105: There are inconsistencies around the assumptions made. If any scattering correction is made, why are they assumed to be unity? More explanations are required. Also refer to Line 152.
  
  Scattering correction is the 𝑠(𝜆)𝜎_sp(𝜆) part of the general correction scheme (Eq. 3). This refers to the scattering from the aerosols and is ignored in this study which was meant by saying “assumed unity”. The multiple scattering correction (C_ref) refers to the enhanced attenuation due to the filter fibers. For this the manufacturer recommendations are used. This statement is rephrased to:
  “The aerosol scattering correction requires measurement of the scattering coefficient, which in many cases is not possible due to the lack of instrumentation. Due to this the aerosol scattering correction is often voided as in this study”
  Text has been adjusted that in reference to these to corrections, aerosol scattering correction is used for the 𝑠(𝜆)𝜎_sp(𝜆) part and multiple scattering correction for the C_ref.
  Line 118: Leakage factor changes over time. Please refer to Drinovec 2015. This assumption might change the corrections, and so might the final corrected BC concentration. Please reconsider adapting to such changes.
  
  The line referring to the leakage factor has been removed from the text and equations. This was an error in the manuscript. The leakage factor was tested during data processing of the manuscript but later removed.
  It is not clear if the MA series device's inbuilt Dual-spot corrections were used for comparison or not. Typically, MA devices have their own correction mechanism, which is not the same as Drinovec 2015. Please check and confirm with a table of corrected data or add it to Table 2.
  
  The MA series inbuilt corrections are used. Meaning the dualspot corrected data that the instrument gives is used as is.
  Confusion to this is related that the authors were not aware of the exact correction algorithm used by the MA-series instruments. It was assumed that Dualspot refers to the correction outlined in Drinovec et al. (2015). This is now considered and refered in the modified version of the manuscript.
  Some comments on figures have been mentioned in the attached file, which mainly focuses on the visibility of the graphics and texts.
  
  These comments have been taken into account and the plots updated.
  Section 3.1.2: A separate segment on the sensor calibration in methods sections is recommended. Also, mention how the calibrations are assessed (metrics used, such as slope, MAE). Reporting the calibration results are helpful for future studies. While presenting the data in "Result" section, please compare it with previous literature. A table might be helpful for reporting metrics from uncalibrated data, calibrated data, and literature data - with the type of calibration procedure adopted. All these elements will expand this section, which I believe is going to improve the quality of the manuscript.
  
  Calibration methodology is added to section 2.5. The results of the calibration are showed in section 3.1.2
  Generally, the results and discussion section is missing references from previous literature. Please compare the reported BC levels with different devices under this work and the levels reported in previous studies in similar regions or of similar spatial characteristics.
  
  Some references to previous studies in Helsinki have been added to section 3.2.1.
  Section 3.2.4: Change the section title. I recommend removing this section and adding a detailed discussion explaining the variability in previous sections. If the authors want to keep this section, will this require some detailed analysis explaining how spatial variability is captured by different devices? If there are any differences in performances? Also, some discussion was required about how spatial variability can be captured by these devices. If the true spatial variability is higher/lower than the inter-device variability studied. Finally, some recommendations /comments would be helpful for the community, such as which device performed best in what context.
  
  This section is removed and merged with section 3.2.1.
  After restructuring the result and discussion section - please update the conclusion accordingly. Avoid including conclusions that have not been discussed well in previous sections (For example, Line 436).
  
  The conclusions have been updated to correspond to the results and discussion provided in the manuscript.
  
  The comments provided in the attached file have also been answered with point-to-point answers and can be found in the attached file.
  
  Citation: https://doi.org/10.5194/ar-2024-12-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Tapio Elomaa on behalf of the Authors (23 Sep 2024) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (27 Sep 2024) by Naďa Zíková

RR by Anonymous Referee #3 (30 Sep 2024)

RR by Anonymous Referee #4 (29 Nov 2024)

Suggestions for revision or reasons for rejection

The comparison of the performance of low-cost black carbon (BC) sensors, the correction algorithms and their impact on the measurements, as well as the sensitivity of these sensors to changes in relative humidity and temperature presented in this manuscript is valuable to the scientific community. Thus I find this article suitable for publication, once the remaining points that need further attention are addressed.

• Although the language in the manuscript has improved, there are still instances where the writing lacks clarity. I suggest another round of language revision. For example, sentences like “Unfortunately, the Observair sensors are not being produced as of the end of 2023…”, “The assumptions are that with 880 nm light source the absorption” or “However, still in the field study, several issues were observed...” need to be revised for better readability and precision.
• In Figure 2, the dates on the x-axis have not been corrected as suggested by Referee #1. There is still a shift in the dates between late May and early June. I recommend using consistent labeling, as with the other dates (every 2 days). Additionally, there is a language error: "Timeseries" should be corrected to "Time series."
• In Figures 12 and 13, there are errors in their descriptions. The units should be placed outside the parentheses, or the word "in" should be included within the parentheses.
• The subsections on methodology and results were not fully renamed to reflect the context, as suggested by both referees. Below are some examples for improved subsection titles:
◦ 2.1. Measuring principle to obtain BC mass concentration with low-cost sensors
◦ 2.3. Deployment of small BC sensors at the Kumpula Campus
◦ 2.4. Description of the sampling site
◦ 3.1. Inter comparison of BC sensors
◦ 3.1.2. Adjusting differences between sensors for comparison
◦ 3.2. Temporal and spatial variability during deployment
◦ 3.2.1. BC levels during days of the week
◦ 3.2.2. Diurnal variation in BC concentration
◦ 3.2.3. Artifacts caused by sensor overheating
• The authors did not provide the strong justification in line 71 as suggested by referee #2 on the use of the MAAP as the reference instrument.
• While presenting the calibrations results (F and orthogonal fit), a comparison with previous literature, as suggested for referee #2 is missing. The authors report the results of the orthogonal fit (slope, correlation coefficients and intercept) in Table 4 but they do not include metrics from literature data with the type of calibration procedure adopted.
• The added paragraph discussing the results with literature on temporal and spatial variability has enhanced the revised version of the article, as suggested by the referee #2. However, the discussion could be further enriched by comparing these results in greater detail with findings from other studies conducted in urban areas across Europe using the same sensors.
• The discussion on spatial variability has improved; however, there are still questions raised by referee #2 that have not been yet fully addressed, which would be of interest to the reader. For instance: "How is spatial variability captured by different devices?" or "Are there differences in performance, and how can spatial variability be effectively captured by these devices?" Additionally, some recommendations or comments are missing, such as identifying which devices performed best in specific contexts.
• The authors have revised the conclusions in the manuscript; however, they need to update them by incorporating a few lines addressing the points raised here (previously asked by the referees). This would enhance the manuscript by providing valuable information and recommendations for the community. For example, identifying which device performed best for monitoring the spatial and temporal variability of BC mass concentrations under the specific ambient conditions of the site (e.g., high relative humidity) would be particularly useful. Additionally, positioning the sampling site as having high, medium, or low BC levels in comparison to other urban sites using small BC sensors would further enrich the conclusions. This revision would avoid discussing aspects not explored in the manuscript, as noted by Referee #2, such as BC source apportionment, which refers to determining BC sources through spectral dependency or positive matrix factorization in the scientific community.

Referee Report: PDF

Hide

ED: Publish subject to minor revisions (review by editor) (29 Nov 2024) by Naďa Zíková

AR by Tapio Elomaa on behalf of the Authors (16 Apr 2025) Author's response Author's tracked changes Manuscript

ED: Publish as is (28 Apr 2025) by Naďa Zíková

AR by Tapio Elomaa on behalf of the Authors (06 May 2025)

Short summary

Black carbon (BC) is a pollutant from combustion that affects the climate and is harmful to health. We tested four different small BC sensors with a reference in Helsinki. The sensors compared well with the reference. As a sensor network they were able to capture differences in BC. Changes in temperature (T) and relative humidity (RH) caused error in the measurements. To reduce the effects of T and RH on BC sensors, more robust boxes should be developed, or corrections should be applied.