the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Correcting for filter-based aerosol light absorption biases at ARM’s SGP site using Photoacoustic data and Machine Learning
- 1Center for Aerosol Science and Engineering, Department of Energy, Environmental and Chemical Engineering, Washington 5 University in St. Louis, St. Louis, MO 63130, USA
- 2NOAA Global Monitoring Laboratory, Boulder, CO 80305, USA
- 3Earth and Environmental Sciences Division, Los Alamos National Laboratory, Los Alamos, NM 87545 USA
- 1Center for Aerosol Science and Engineering, Department of Energy, Environmental and Chemical Engineering, Washington 5 University in St. Louis, St. Louis, MO 63130, USA
- 2NOAA Global Monitoring Laboratory, Boulder, CO 80305, USA
- 3Earth and Environmental Sciences Division, Los Alamos National Laboratory, Los Alamos, NM 87545 USA
Abstract. Measurement of light absorption of solar radiation by aerosols is vital for assessing direct aerosol radiative forcing, which affects local and global climate. Low-cost and easy-to-operate filter-based instruments, such as the Particle Soot Absorption Photometer (PSAP) that collect aerosols on a filter and measure light attenuation through the filter are widely used to infer aerosol light absorption. However, filter-based absorption measurements are subject to artifacts which are difficult to quantify. These artifacts are associated with the presence of the filter medium and the complex interactions between the filter fibers and accumulated aerosols. Various correction algorithms have been introduced to correct for the filter-based absorption coefficient measurements toward predicting the particle-phase absorption coefficient (Babs). Since previously-developed correction algorithms have a fixed analytical form, fundamentally, they are unable to predict particle phase absorption coefficients with a high degree of accuracy universally: different corrections are required for rural and urban sites across the world. In this study, we analyzed three months of high-resolution ambient data collected using a 3-wavelength photoacoustic spectrometer (PASS) and PSAP on the same inlet; both instruments were operated at the Department of Energy’s Atmospheric Radiation Measurement (ARM) Southern Great Plains user facility in Oklahoma. We implemented the two most commonly used analytical correction algorithms, namely the Virkkula (2010) and the average of Virkkula (2010) and Ogren (2010)-Bond (1999), as well as a Random Forest Regression (RFR) machine learning algorithm to infer particle phase Babs values from PSAP data and estimate their accuracy in comparsion to the refernce Babs values measured synchronously by the PASS. The wavelength averaged Root Mean Square Error (RMSE) values of Babs predicted using the RFR algorithm are improved by an order of magnitude in comparison to those predicted by the Virkulla (2010) and average correction algorithms. A revised form of the Virkkula (2010) algorithm suitable for the SGP site has been proposed; however, its performance yields approximately two fold errors when compared to the RFR algorithm. To further improve the accuracy of our proposed RFR algorithm, we trained and tested the model on dataset of laboratory-generated combustion aerosols. The RFR model used as inputs the size distribution, uncorrected Tricolor Absorption Photometer (TAP)-measured Babs, and nephelometer-measured scattering coefficient Bscat and predicted particle-phase Babs values within 5% of the reference Babs measured by a PASS. Machine learning approaches offers a promising path to correct for biases in long term filter-based absorption datasets and accurately quantify their variability and trends needed for robust radiative forcing determination.
Joshin Kumar et al.
Status: closed
-
RC1: 'Comment on amt-2022-42', Anonymous Referee #1, 08 Apr 2022
The manuscript "Correcting for filter-based aerosol light absorption biases at ARM's SGP site using Photoacoustic data and Machine Learning" by J. Kumar et al. shows that a random forest tree machine learning algorithm can correct particle absorption coefficients measured with filter-based instruments. The analysis was performed for a specific measurement site, but the study itself can be used as a blueprint for testing ML algorithms for other stations with various aerosol types.
The manuscript addresses a relevant topic and falls within the scope of AMT. It is well written and the conclusions are sound. The reviewer recommends the manuscript for publication after considering the following minor comments.
Specific comments:
Page 1, line 18: The reviewer does not fully agree with the chosen explanation or wording why filter-based instruments have problems in predicting the particle absorption coefficient. The reviewer believes that the reason for the limitations is not that fixed analytical (*) forms were chosen, but rather that there are hidden influencing parameters. There is also no algorithm that takes into account all known influencing parameters, e.g. Nakayama et al. (2010) present a correction for particle penetration depth, the restricted two-stream method (Mueller et al., 2014) takes into account particle asymmetry but not particle size. It should not be concluded that solvers with fixed analytical functions are generally unable to predict particle absorption coefficients with high accuracy.
(*The reviewer means that iterative solvers for fixed parameterised functions are also included in the class of fixed analytic functions. )
Page 1, line 31: Does the RFR model use the particle size distribution as input? Cf. line 195, where it says that the total mass concentration is used as input. What does total mass concentration mean? Is the cumulative mass on a PSAP filter spot meant?
Line 59: Can PASS be considered a first principle method? A few lines later the authors describe the problem with liquid or multiphase particles, which is a fundamental problem of the method?
Line 98: Do the authors mean the absorption coefficients or the uncorrected absorption coefficients?
In Figures A3 and 1, the axis label and caption should indicate whether corrected or uncorrected absorption coefficients are shown.
General comments to chapter 2:
Because of the known artefact due to light scattering particles, it would be informative if the authors presented single scatter albedos.
Why was the Virkkula (2010) correction revised but not the Ogre (2010)-Bond (1999) correction?
Figure 5: It seems that there are fewer data points in Figure 5 than in Figures 3 and 4. Is the split of the data into training and test data sets the only reason?
Line 303: Does it have any influence that the laboratory dataset was taken with a TAP photometer and the data from the SPG site was taken with a PSAP photometer?
Appendix A4: check sentence: “Tune the parameters of the ML model if the performance to achieve desired level of accuracy.”
References:
Nakayama, T., et al. (2010). "Size-dependent correction factors for absorption measurements using filter-based photometers: PSAP and COSMOS." Journal of Aerosol Science 41(4): 333-343.
Mueller, T., et al. (2014). "Constrained two-stream algorithm for calculating aerosol light absorption coefficient from the Particle Soot Absorption Photometer." Atmos. Meas. Tech. 7: 4049-4070.
-
AC1: 'Reply on RC1', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Rajan Chakrabarty, 12 May 2022
-
RC2: 'Comment on amt-2022-42', Anonymous Referee #2, 09 Apr 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-RC2-supplement.pdf
-
AC2: 'Reply on RC2', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Rajan Chakrabarty, 12 May 2022
-
RC3: 'Comment on amt-2022-42', Anonymous Referee #3, 11 Apr 2022
General comments
Kumar et al. describe a very interesting comparison of “traditional” PSAP correction algorithms with a new machine learning algorithm. The work is important and can contribute significantly to the painful post-processing of the filter-photometer data in general. Filter photometers are used widely and in very different environments, so an algorithm that reduces the bias with no or very little assumptions is most welcome.
The reviewer will take the unusual action of sticking to general comments (under several titles) and have only two specific ones. The importance of the results is unfortunately influenced by the hastiness and shallowness of the writing. There is a definite lack of attention to detail. The paper should be heavily revised and reviewed afterwards. It definitely deserves publication – the improvement in the regressions between the PSAP and PASS is impressive.
Terminology and parameters
What is B_PSAP?
What are “uncorrected filter-based absorption raw signals”?
The statement (line 138) that “this overestimation…” – by filter pohotometers, “… is due to the enhancement of light absorption by the filter deposited aerosol due to scattering based artifacts” is misleading. The enhancement is due to scattering of light by the filter fibers. Part of the reduction of light intensity below the sample is due to scattering (away from the forward direction). This is correctly summarized in Eq. 1, but not here. The separation of these two effects is artificial (Mueller et al., 2010). This should be discussed, especially in light of the superiority of the RFR algorithm.
The authors should use notation and naming of the parameters consistent with Ogren et al (2017). I suspect that the “uncorrected filter-based absorption raw signals” are in fact the attenuation coefficient. If this is so, please use this parameter. The “raw signals” could be interpreted as raw intensity signals measured by the photodiode… Please be precise.
Also, the absorption coefficient is derived, not measured, by filter photometers. The authors correctly state this, but then relapse into claiming that the absorption coefficient is “measured” by PSAP.
What is he reason for including Eq 4 before Eq 5?
It is never referenced.
Measurements
Start by explicitly stating the period under investigation. This sets the stage. It seems there is 6.5 years of collocated PASS and PSAP data, yet the authors use only 3 months?!
Which filter was used in the PSAP?
Pall, Azumi, anything else… - see for example Ogren et al., 2017 for the difference in regression slope.
Was the inlet dried? Was there a cutoff? How was the flow split? Were the ACSM measurements performed in the same size fraction?
Please provide all relevant details!
The authors state that SSA is not available. This is probably not true, as the ARM web page:
https://www.arm.gov/capabilities/instruments?category=aerosol&type=armobs&site=sgp
states that there is a neph installed at SGP from 4 Oct 2010 onwards. Include the scattering data everywhere as it will improve the corrections, especially the Virkkula 2010 corrected algorithm.
SSA is fundamental in terms of the overestimation of derived absorption using filter-photometers (Weingartner et al., 2003; Virkkula et al., 2015; Yus-Díez et al., 2021). The analysis of the performance (see below on the comment which parameter to use) of the algorithms as a function of SSA should be investigated.
Similarly, ACSM measurements are supposed to be available from 18 Nov 2010 onwards. The selectio of only 3 months makes the huge OM event starting on 2015 07 07 very important when looking at the “average” picture – it heavily skews the data, if using averages, especially since the PASS and PSAP were not working consistently during this period. The ACSM measurements could be used to a higher degree in the interpretation of the results (see also below).
What is the relevance Fig A2, showing, among other things, negative OC?
The period is 2 years prior to the measurement campaign.
The authors use negative AAE derived from the blue/green wavelength pair for inter/extrapolation. Such values are highly unusual and require major attention. There must be an error somewhere, since other AAE values seem closer to what one would expect. The OM event probably has a huge influence on AAE. What happened, a large fire?
Data processing and Algorithms
How were the period with incorrect, suspect, and missing values identified? What were the criteria?
How was averaging performed? Was the Springston and Sedlacek (2007) algorithm used? It should be at least investigated, there is an interesting example by Backman et al. (2017) which could be followed.
The reason for using the average of Virkkula (2010) and Ogren (2010)-Bond (1999) is described only briefly. It would help to treat each separately and then show the average (which is used in processing of the data). One expects a comment also on why is this paper better than Arnott et al. (2005).
How is the training/testing split 70/30 determined?
More details need to be provided. Why is the learning period twice as long as the test one? What happens if the periods are extended (6.5 years of data!)? Is the 70/30 split pre-defined by a human or is this some sort of a Monte Carlo sampling?
The RFR method is empiricistic. It would be of great interest to check its performance in periods of different SSA… to see what are the real parameter of interest (see the back-scattering coefficient and SSA in Virkkula et al., 2015)?
The reviewer is not sure that RMSE is the correct parameter to estimate the performance of the correction algorithms. It assumes the error only in the PSAP measurements. While this is true for “bias” assuming that PASS is the “absolute truth” (see above), but it is not true for experimental noise.
What is the cause of RMSE wavelength variability? Noise? This is algorithm independent – green regressions are always best.
Results
Why is the number of points in the regression on Fig 5 (RFR) lower that for other algorithms?
Miscellaneous
I am curious: could you derive Virkkula parameters with the RFR algorithm?
The laboratory experiment (Section 3.5) is very different but interesting. There should be more experimental detail.
Specific comments
Please spell check the manuscript!
Please use the dates in global format (18 Nov 2010), so that our colleagues from outside the Americas will understand them without ambiguity.
References
Arnott, W. P., Hamasha, K., Moosmüller, H., Sheridan, P. J., and Ogren, J. A.: Towards aerosol light-absorption measurements with a 7-wavelength Aethalometer: Evaluation with a photoacoustic instrument and 3-wavelength nephelometer, Aerosol Sci. Technol., 39, 17-29, 10.1080/027868290901972, 2005.
Backman, J., Schmeisser, L., Virkkula, A., Ogren, J. A., Asmi, E., Starkweather, S., Sharma, S., Eleftheriadis, K., Uttal, T., Jefferson, A., Bergin, M., Makshtas, A., Tunved, P., and Fiebig, M.: On Aethalometer measurement uncertainties and an instrument correction factor for the Arctic, Atmos. Meas. Tech., 10, 5039–5062, https://doi.org/10.5194/amt-10-5039-2017, 2017.
Müller, T., Virkkula, A., and Ogren, J. A.: Constrained two-stream algorithm for calculating aerosol light absorption coefficient from the Particle Soot Absorption Photometer, Atmos. Meas. Tech., 7, 4049–4070, https://doi.org/10.5194/amt-7-4049-2014, 2014.
Ogren, J. A., Wendell, J., Andrews, E., and Sheridan, P. J.: Continuous light absorption photometer for long-term studies, Atmos. Meas. Tech., 10, 4805–4818, https://doi.org/10.5194/amt-10-4805-2017, 2017.
Springston, S. R. and Sedlacek, A. J.: Noise characteristics of an instrumental particle absorbance technique, Aerosol Sci. Tech., 41,1110–1116, https://doi.org/10.1080/02786820701777457, 2007.
Virkkula, A.: Correction of the calibration of the 3-wavelength Particle Soot Absorption Photometer (3λ PSAP), Aerosol Science and Technology, 44, 706-712, 2010.
Virkkula, A., Chi, X., Ding, A., Shen, Y., Nie, W., Qi, X., Zheng, L., Huang, X., Xie, Y., Wang, J., Petäjä, T., and Kulmala, M.: On the interpretation of the loading correction of the aethalometer, Atmos. Meas. Tech., 8, 4415–4427, https://doi.org/10.5194/amt-8-4415-2015, 2015.
Weingartner, E., Saathoff, H., Schnaiter, M., Streit, N., Bitnar, B., and Baltensperger, U.: Absorption of light by soot particles: determination of the absorption coefficient by means of Aethalometers, J. Aerosol Sci, 34, 1445-1463, 10.1016/S0021-8502(03)00359-8, 2003.
Yus-Díez, J., Bernardoni, V., MoÄnik, G., Alastuey, A., Ciniglia, D., IvanÄiÄ, M., Querol, X., Perez, N., Reche, C., Rigler, M., Vecchi, R., Valentini, S., and Pandolfi, M.: Determination of the multiple-scattering correction factor and its cross-sensitivity to scattering and wavelength dependence for different AE33 Aethalometer filter tapes: a multi-instrumental approach, Atmos. Meas. Tech., 14, 6335–6355, https://doi.org/10.5194/amt-14-6335-2021, 2021.
-
AC3: 'Reply on RC3', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Rajan Chakrabarty, 12 May 2022
Status: closed
-
RC1: 'Comment on amt-2022-42', Anonymous Referee #1, 08 Apr 2022
The manuscript "Correcting for filter-based aerosol light absorption biases at ARM's SGP site using Photoacoustic data and Machine Learning" by J. Kumar et al. shows that a random forest tree machine learning algorithm can correct particle absorption coefficients measured with filter-based instruments. The analysis was performed for a specific measurement site, but the study itself can be used as a blueprint for testing ML algorithms for other stations with various aerosol types.
The manuscript addresses a relevant topic and falls within the scope of AMT. It is well written and the conclusions are sound. The reviewer recommends the manuscript for publication after considering the following minor comments.
Specific comments:
Page 1, line 18: The reviewer does not fully agree with the chosen explanation or wording why filter-based instruments have problems in predicting the particle absorption coefficient. The reviewer believes that the reason for the limitations is not that fixed analytical (*) forms were chosen, but rather that there are hidden influencing parameters. There is also no algorithm that takes into account all known influencing parameters, e.g. Nakayama et al. (2010) present a correction for particle penetration depth, the restricted two-stream method (Mueller et al., 2014) takes into account particle asymmetry but not particle size. It should not be concluded that solvers with fixed analytical functions are generally unable to predict particle absorption coefficients with high accuracy.
(*The reviewer means that iterative solvers for fixed parameterised functions are also included in the class of fixed analytic functions. )
Page 1, line 31: Does the RFR model use the particle size distribution as input? Cf. line 195, where it says that the total mass concentration is used as input. What does total mass concentration mean? Is the cumulative mass on a PSAP filter spot meant?
Line 59: Can PASS be considered a first principle method? A few lines later the authors describe the problem with liquid or multiphase particles, which is a fundamental problem of the method?
Line 98: Do the authors mean the absorption coefficients or the uncorrected absorption coefficients?
In Figures A3 and 1, the axis label and caption should indicate whether corrected or uncorrected absorption coefficients are shown.
General comments to chapter 2:
Because of the known artefact due to light scattering particles, it would be informative if the authors presented single scatter albedos.
Why was the Virkkula (2010) correction revised but not the Ogre (2010)-Bond (1999) correction?
Figure 5: It seems that there are fewer data points in Figure 5 than in Figures 3 and 4. Is the split of the data into training and test data sets the only reason?
Line 303: Does it have any influence that the laboratory dataset was taken with a TAP photometer and the data from the SPG site was taken with a PSAP photometer?
Appendix A4: check sentence: “Tune the parameters of the ML model if the performance to achieve desired level of accuracy.”
References:
Nakayama, T., et al. (2010). "Size-dependent correction factors for absorption measurements using filter-based photometers: PSAP and COSMOS." Journal of Aerosol Science 41(4): 333-343.
Mueller, T., et al. (2014). "Constrained two-stream algorithm for calculating aerosol light absorption coefficient from the Particle Soot Absorption Photometer." Atmos. Meas. Tech. 7: 4049-4070.
-
AC1: 'Reply on RC1', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Rajan Chakrabarty, 12 May 2022
-
RC2: 'Comment on amt-2022-42', Anonymous Referee #2, 09 Apr 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-RC2-supplement.pdf
-
AC2: 'Reply on RC2', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Rajan Chakrabarty, 12 May 2022
-
RC3: 'Comment on amt-2022-42', Anonymous Referee #3, 11 Apr 2022
General comments
Kumar et al. describe a very interesting comparison of “traditional” PSAP correction algorithms with a new machine learning algorithm. The work is important and can contribute significantly to the painful post-processing of the filter-photometer data in general. Filter photometers are used widely and in very different environments, so an algorithm that reduces the bias with no or very little assumptions is most welcome.
The reviewer will take the unusual action of sticking to general comments (under several titles) and have only two specific ones. The importance of the results is unfortunately influenced by the hastiness and shallowness of the writing. There is a definite lack of attention to detail. The paper should be heavily revised and reviewed afterwards. It definitely deserves publication – the improvement in the regressions between the PSAP and PASS is impressive.
Terminology and parameters
What is B_PSAP?
What are “uncorrected filter-based absorption raw signals”?
The statement (line 138) that “this overestimation…” – by filter pohotometers, “… is due to the enhancement of light absorption by the filter deposited aerosol due to scattering based artifacts” is misleading. The enhancement is due to scattering of light by the filter fibers. Part of the reduction of light intensity below the sample is due to scattering (away from the forward direction). This is correctly summarized in Eq. 1, but not here. The separation of these two effects is artificial (Mueller et al., 2010). This should be discussed, especially in light of the superiority of the RFR algorithm.
The authors should use notation and naming of the parameters consistent with Ogren et al (2017). I suspect that the “uncorrected filter-based absorption raw signals” are in fact the attenuation coefficient. If this is so, please use this parameter. The “raw signals” could be interpreted as raw intensity signals measured by the photodiode… Please be precise.
Also, the absorption coefficient is derived, not measured, by filter photometers. The authors correctly state this, but then relapse into claiming that the absorption coefficient is “measured” by PSAP.
What is he reason for including Eq 4 before Eq 5?
It is never referenced.
Measurements
Start by explicitly stating the period under investigation. This sets the stage. It seems there is 6.5 years of collocated PASS and PSAP data, yet the authors use only 3 months?!
Which filter was used in the PSAP?
Pall, Azumi, anything else… - see for example Ogren et al., 2017 for the difference in regression slope.
Was the inlet dried? Was there a cutoff? How was the flow split? Were the ACSM measurements performed in the same size fraction?
Please provide all relevant details!
The authors state that SSA is not available. This is probably not true, as the ARM web page:
https://www.arm.gov/capabilities/instruments?category=aerosol&type=armobs&site=sgp
states that there is a neph installed at SGP from 4 Oct 2010 onwards. Include the scattering data everywhere as it will improve the corrections, especially the Virkkula 2010 corrected algorithm.
SSA is fundamental in terms of the overestimation of derived absorption using filter-photometers (Weingartner et al., 2003; Virkkula et al., 2015; Yus-Díez et al., 2021). The analysis of the performance (see below on the comment which parameter to use) of the algorithms as a function of SSA should be investigated.
Similarly, ACSM measurements are supposed to be available from 18 Nov 2010 onwards. The selectio of only 3 months makes the huge OM event starting on 2015 07 07 very important when looking at the “average” picture – it heavily skews the data, if using averages, especially since the PASS and PSAP were not working consistently during this period. The ACSM measurements could be used to a higher degree in the interpretation of the results (see also below).
What is the relevance Fig A2, showing, among other things, negative OC?
The period is 2 years prior to the measurement campaign.
The authors use negative AAE derived from the blue/green wavelength pair for inter/extrapolation. Such values are highly unusual and require major attention. There must be an error somewhere, since other AAE values seem closer to what one would expect. The OM event probably has a huge influence on AAE. What happened, a large fire?
Data processing and Algorithms
How were the period with incorrect, suspect, and missing values identified? What were the criteria?
How was averaging performed? Was the Springston and Sedlacek (2007) algorithm used? It should be at least investigated, there is an interesting example by Backman et al. (2017) which could be followed.
The reason for using the average of Virkkula (2010) and Ogren (2010)-Bond (1999) is described only briefly. It would help to treat each separately and then show the average (which is used in processing of the data). One expects a comment also on why is this paper better than Arnott et al. (2005).
How is the training/testing split 70/30 determined?
More details need to be provided. Why is the learning period twice as long as the test one? What happens if the periods are extended (6.5 years of data!)? Is the 70/30 split pre-defined by a human or is this some sort of a Monte Carlo sampling?
The RFR method is empiricistic. It would be of great interest to check its performance in periods of different SSA… to see what are the real parameter of interest (see the back-scattering coefficient and SSA in Virkkula et al., 2015)?
The reviewer is not sure that RMSE is the correct parameter to estimate the performance of the correction algorithms. It assumes the error only in the PSAP measurements. While this is true for “bias” assuming that PASS is the “absolute truth” (see above), but it is not true for experimental noise.
What is the cause of RMSE wavelength variability? Noise? This is algorithm independent – green regressions are always best.
Results
Why is the number of points in the regression on Fig 5 (RFR) lower that for other algorithms?
Miscellaneous
I am curious: could you derive Virkkula parameters with the RFR algorithm?
The laboratory experiment (Section 3.5) is very different but interesting. There should be more experimental detail.
Specific comments
Please spell check the manuscript!
Please use the dates in global format (18 Nov 2010), so that our colleagues from outside the Americas will understand them without ambiguity.
References
Arnott, W. P., Hamasha, K., Moosmüller, H., Sheridan, P. J., and Ogren, J. A.: Towards aerosol light-absorption measurements with a 7-wavelength Aethalometer: Evaluation with a photoacoustic instrument and 3-wavelength nephelometer, Aerosol Sci. Technol., 39, 17-29, 10.1080/027868290901972, 2005.
Backman, J., Schmeisser, L., Virkkula, A., Ogren, J. A., Asmi, E., Starkweather, S., Sharma, S., Eleftheriadis, K., Uttal, T., Jefferson, A., Bergin, M., Makshtas, A., Tunved, P., and Fiebig, M.: On Aethalometer measurement uncertainties and an instrument correction factor for the Arctic, Atmos. Meas. Tech., 10, 5039–5062, https://doi.org/10.5194/amt-10-5039-2017, 2017.
Müller, T., Virkkula, A., and Ogren, J. A.: Constrained two-stream algorithm for calculating aerosol light absorption coefficient from the Particle Soot Absorption Photometer, Atmos. Meas. Tech., 7, 4049–4070, https://doi.org/10.5194/amt-7-4049-2014, 2014.
Ogren, J. A., Wendell, J., Andrews, E., and Sheridan, P. J.: Continuous light absorption photometer for long-term studies, Atmos. Meas. Tech., 10, 4805–4818, https://doi.org/10.5194/amt-10-4805-2017, 2017.
Springston, S. R. and Sedlacek, A. J.: Noise characteristics of an instrumental particle absorbance technique, Aerosol Sci. Tech., 41,1110–1116, https://doi.org/10.1080/02786820701777457, 2007.
Virkkula, A.: Correction of the calibration of the 3-wavelength Particle Soot Absorption Photometer (3λ PSAP), Aerosol Science and Technology, 44, 706-712, 2010.
Virkkula, A., Chi, X., Ding, A., Shen, Y., Nie, W., Qi, X., Zheng, L., Huang, X., Xie, Y., Wang, J., Petäjä, T., and Kulmala, M.: On the interpretation of the loading correction of the aethalometer, Atmos. Meas. Tech., 8, 4415–4427, https://doi.org/10.5194/amt-8-4415-2015, 2015.
Weingartner, E., Saathoff, H., Schnaiter, M., Streit, N., Bitnar, B., and Baltensperger, U.: Absorption of light by soot particles: determination of the absorption coefficient by means of Aethalometers, J. Aerosol Sci, 34, 1445-1463, 10.1016/S0021-8502(03)00359-8, 2003.
Yus-Díez, J., Bernardoni, V., MoÄnik, G., Alastuey, A., Ciniglia, D., IvanÄiÄ, M., Querol, X., Perez, N., Reche, C., Rigler, M., Vecchi, R., Valentini, S., and Pandolfi, M.: Determination of the multiple-scattering correction factor and its cross-sensitivity to scattering and wavelength dependence for different AE33 Aethalometer filter tapes: a multi-instrumental approach, Atmos. Meas. Tech., 14, 6335–6355, https://doi.org/10.5194/amt-14-6335-2021, 2021.
-
AC3: 'Reply on RC3', Rajan Chakrabarty, 12 May 2022
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2022-42/amt-2022-42-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Rajan Chakrabarty, 12 May 2022
Joshin Kumar et al.
Joshin Kumar et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 413 | 123 | 22 | 558 | 7 | 7 |
- HTML: 413
- PDF: 123
- XML: 22
- Total: 558
- BibTeX: 7
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1