the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Nov 2025
Quantitative uncertainty and post-processing for micro-aethalometers measuring black carbon
Abstract. Aethalometers measure black carbon mass concentrations by monitoring light attenuation through a particle filter as it becomes laden with aerosols. As the uncertainties in the resulting measurements are not easily quantified via a bottom-up traceable approach, there is a need for inter-device comparisons to provide operationally defined uncertainties. The present work compared five micro-aethalometers to known mass concentrations of laboratory-generated soot, formed using an inverted ethylene flame and a Centrifugal Particle Mass Analyzer-Electrometer Reference Mass Standard (CERMS). Uncertainties were found to scale with mass concentration, with multiplicative errors between devices of approximately 10 % in the best case of long sampling times and/or high mass concentrations. A quantitative expression is provided for the uncertainty in the aethalometer measurements as a function of mass concentration, sampling interval, and flow rate. An open-source algorithm is also provided for the unsupervised reanalysis of aethalometer or other filter photometer data over varying periods to reach a specified target uncertainty.
- Preprint
(4131 KB) - Metadata XML
-
Supplement
(242 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4209', Anonymous Referee #1, 01 Dec 2025
- AC3: 'Reply on RC1', Timothy Sipkens, 13 Feb 2026
-
RC2: 'Comment on egusphere-2025-4209', Anonymous Referee #2, 02 Dec 2025
This study explores the uncertainties involved in measuring equivalent black carbon mass concentration using micro-aethalometers. Under controlled laboratory conditions, five micro-aethalometers (MA200/MA300) were evaluated for measuring black carbon aerosols with known mass concentrations, generated using a reference mass standard coupling MISG and CPMA. This study successfully isolates 3 types of uncertainties and develops a comprehensive uncertainty model that considers factors such as mass concentration, sampling intervals, and flow rates. Additionally, this study presents an open-source algorithm designed to optimize sampling strategies based on targeted uncertainty levels. The experiments were carefully designed and the manuscript is well written. The following issues need to be addressed.
Major Comments.
- The title should be revised, as this study is limited to laboratory conditions and does not consider ambient-related factors such as temperature, relative humidity, and low EC/TC ratio samples. An alternative title can be “Laboratory-based uncertainty quantification and post-processing strategies for micro-aethalometers measuring black carbon”. The following statement should be included in the conclusions. “It should be noted that the uncertainty model derived in this work is based on laboratory-generated, non-volatile soot under stable conditions. Its applicability to ambient measurements, where factors such as fluctuating temperature, relative humidity, and the presence of externally mixed non-BC aerosols or coatings on BC particles (leading to low EC/TC ratios) can significantly influence the measurement, requires further validation.”
- The main finding of this study is the identification of three types of uncertainties: multiplicative inter-device bias, random noise due to dual-spot correction, and additive systematic bias. This finding must be explicitly stated in the abstract.
- At the end of Section 1, the authors state that “Since our aerosol model represents a simple source of eBC, with negligible content of non-absorbing PM and stable gas-phase composition, our results provide a lower limit on between-instrument reproducibility.” Compared to mini-CAST, MISG tends to generate samples with a high EC fraction (>90%). Did the authors verify the actual level of EC fraction in this study? If the EC fraction reported in this paper is indeed greater than 90%, then the results apply only to source emission scenarios. For ambient atmospheric aerosols, where the EC fraction is typically less than 10%, could the authors provide insights into how the EC/TC ratio might affect the uncertainty of micro-aethalometers when applied to ambient aerosols? Although this is not the primary focus of the current study, understanding this relationship would be valuable.
- The statement in Section 2 that “The correction factor is taken as that provided by the instrument firmware and applied during post-processing” is incorrect. The k value is derived by solving the system of equations from Eq. 1, which was obtained under different flow rates at the two sampling spots. For each data point at time t, a corresponding k value can be calculated. The k value is not preset in the instrument's firmware.
- The authors mentioned, “We tested five microAeths (two MA200, three MA300, AethLabs, USA) that were previously deployed at ambient monitoring sites across Canada.” Did the authors calibrate the flows of the MA200 and MA350 before the experiments? The accuracy of the flow readings can significantly affect the inter-device bias.
- Did the authors check the firmware version of the five micro-aethalometers? The reviewer's experience with the MA series indicates that the firmware version can influence the inter-device bias.
Citation: https://doi.org/10.5194/egusphere-2025-4209-RC2 - AC1: 'Reply on RC2', Timothy Sipkens, 13 Feb 2026
-
RC3: 'Comment on egusphere-2025-4209', Anonymous Referee #3, 02 Dec 2025
Sipkens et al. conducted an inter-device comparison using five micro-aethalometers and identified two fundamentally different types of uncertainties. In addition, the study provides an open-source algorithm to mitigate these uncertainties. The manuscript is logical and well written. However, the following issues should be clarified:
Major comments:
- The accuracy of filter photometers is influenced by particle size, while this study uses a fixed particle mobility diameter of 235 nm. How would using other particle diameters affect the reported uncertainties?
- Lines 285–289: Please explain what causes the difference in inter-device variability between the micro-aethalometers (MA200, MA300, and MA350) and the AE33.
Minor suggestion:
- Suggest combining Section 2 (Experimental setup) and Section 3.2 (Model fitting and uncertainty quantification) into a single “Materials and methods” section.
Citation: https://doi.org/10.5194/egusphere-2025-4209-RC3 -
AC2: 'Reply on RC3', Timothy Sipkens, 13 Feb 2026
We thank the referee for the comments. Please find our item-by-item responses below.
- The accuracy of filter photometers is influenced by particle size, while this study uses a fixed particle mobility diameter of 235 nm. How would using other particle diameters affect the reported uncertainties?
It is well established that the device will have a different response to particles of different sizes, as well as different compositions, etc. It is noted, however, that this is not a noise or inter-device reproducibility issue. It will impact the slope of the line in the parity plots, but this should be accounted for using calibration factors that are designed to accommodate particle size and composition effects. Quantifying the inter-device variability requires that these contributions be minimized, motivating the current experimental design.
This is now explicitly acknowledged following the list of observations from the parity / error plot (Sec. 3.1):
“As noted in the introduction, this model represents a minimum uncertainty that does not take into account various causes of artifacts that can occur in aethalometer measurements (such as particle size, composition, and rapid changes in gas composition). Such additional artifacts will add further uncertainty.”
We have also added a note in the conclusions:
“It should be noted that these expressions only capture repeatability and inter-device variability. They do not intend to account for other systematic artifacts, including cross-sensitivity to scattering, humidity, and temperature. These effects are often location-specific and should be addressed by using appropriate calibration factors. Rather, these expressions give the minimum uncertainty that can be achieved after the aethalometers have been calibrated for the properties of the specific particles being measured. We also do not assert that this model will account for all uncertainties in the measurements, as, while systematic artifacts in the data can largely be removed by way of calibration (similar to calibration to remove inter-device variability), the calibration factors used to perform this correction will themselves have uncertainties (whether due to physical fluctuations during the measurements or incomplete knowledge of the artifacts) that must be considered alongside the uncertainties here.”
- Lines 285–289: Please explain what causes the difference in inter-device variability between the micro-aethalometers (MA200, MA300, and MA350) and the AE33.
The present analysis does not explicitly provide information to answer this question, both in terms of the underlying cause (with only hints, due to the nature of the noise, i.e., multiplicative) and its extendibility to other classes or aethalometers. One would reasonably expect that instrument design, including the size of the optics and detectors, would influence the variability between the devices. However, we cannot say for certain and do not want to add conjecture to the manuscript. We have added the following note to the manuscript:
“More investigation would be required to understand the source of the inter-device variability and verify the form of the model for other aethalometer classes.”
Without knowing the source of the inter-device variability, it would also be hard to extend the model directly to AE33. While we suspect that a very similar model would apply, this would have to be explicitly verified.
Minor suggestion:
- Suggest combining Section 2 (Experimental setup) and Section 3.2 (Model fitting and uncertainty quantification) into a single “Materials and methods” section.
We would argue that Sec. 3.1 belongs where it does, as the subsequent 3.2 requires some understanding of the model structure. However, we can see the argument for moving 3.2 to a “Materials and methods” section, so we opted to move Sec. 3.1 and 3.2 to a new “Materials and methods” section (Sec. 2).
Citation: https://doi.org/10.5194/egusphere-2025-4209-AC2
- AC1: 'Reply on RC2', Timothy Sipkens, 13 Feb 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 291 | 128 | 37 | 456 | 38 | 22 | 17 |
- HTML: 291
- PDF: 128
- XML: 37
- Total: 456
- Supplement: 38
- BibTeX: 22
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Egusphere-2025-4209 - Review
The article presents new statistical methods for determination of uncertainties connected to inter-instrument differences and measurement noise and applies them to micro-aethalometers MA300 and MA200. The experiment was performed by comparing tested instruments with Centrifugal Particle Mass Analyzer-Electrometer Reference Mass Standard (CERMS). The article is well written with appropriate introduction and methodology and description of the results. There are two issues which need to be addressed.
Filter photometers suffer from particle size dependent response, which makes it important to use realistic particle size for testing. The soot particles selected by CERMS with a mobility diameter of 235 nm are larger compared to the diesel engine and wood stove emissions (Laborde et al., 2012) and are more comparable to wildfire emissions. Measurement of larger particles is expected to result in higher “quantum” noise (from discrete arrival of particles on the filter), as already observed by the authors. Particle size is expected to influence the instrument response to BC mass.
The second issue is the influence of the filter loading effect. Figure 2 show jumps in instrument response during tape advance which indicate the presence of filter loading effect. Same can be concluded by analyzing relative instrument response (eBC/ref_mass) as a function of attenuation using data in the Supplement (please see the graph in the attached pdf). Authors should discuss the implications of filter loading effect, for example the increased variation of the instrument response.
Line by line comments
Page 1, Line 17: “A quantitative expression is provided for the uncertainty in the
aethalometer measurements as a function of mass concentration, sampling interval, and flow rate.”
It should be noted that these uncertainties present one part of the uncertainties connected to the filter photometers. Uncertainties connected to filter loading effect, particles size, cross-sensitivity to scattering and mass absorption cross-section can be higher than inter-instrument variability and noise.
Page 2., Line 42:
There have been more measurements of the influence of particle size on instrument response. It seems to be a general feature of filter photometers: it was observed for AE33, CLAP and MAAP (Ramshoo et al, 2022; Drinovec et al., 2022; Yus-Dies et al., 2025).
Ramshoo et al, 2022; https://doi.org/10.5194/amt-15-6965-2022-supplement
Drinovec et al., 2022; https://doi.org/10.5194/amt-15-6965-2022
Yus-Diez et al., 2025; https://doi.org/10.5194/amt-18-3073-2025
Page 4, Line 114: “size distribution had a geometric mean mobility diameter of 235 nm”
The selected particle size depends strongly on charge distribution from the unipolar charger. How stable was the charge distribution during the experiment?
Page 7. Fig 3.
In Figure 3. eBC measured by aethalometers is compared to particle mass concentration of CERMS. What is the contribution of organics to the sample mass? What would be the effect of thermodenuder or catalytic stripper?
Page 7. Line 158. “with structured artifacts as a function of attenuation appearing in the measurements when not correcting the data (cf. Figure 2b)”
This is in contrast with the caption of Figure 2 which indicates that the data is corrected for the filter loading effect (except for data where ATN<3). Please clarify which data is corrected for the filter loading effect.
Page 9, Line 202: “While not shown in Figure 3, it is noted that the attenuation coefficient has minimal effect above an attenuation of 3,”.
Data on Figure 3 and in the supplement suggest that the filter loading effect is still present.
Page 18, Line 376: “While the dual spot correction algorithm was found to be effective in correcting biases in the measurements”
Please see the comments on filter loading effect above.
Page 19, Line 394. “Partially processed data is included alongside the manuscript.”
Please provide both raw and processed data.