Review of the manuscript entitled «Estimates of mass absorption cross sections of black carbon for filterbased absorption photometers in the Arctic» submitted to Atmos. Meas. Techn. by Ohata et al.

This manuscript addresses observations of black carbon aerosol in the Arctic. Knowing spatio-temporal variation of black carbon (BC) concentration and resulting light absorption is of general atmospheric and climatic interest as motivated by the authors. They present multi-annual and multi-site data sets of parallel measurements with two or more methods. This is a great effort and of large value as starting point for many studies. The methods and data treatment are sound and the manuscript is well written and it provides ample and important information towards consistent interpretation of observations made with different instrument types commonly deployed to quantify black carbon mass concentrations, and also provides approximate values for the conversion factor between light absorption coefficient and black carbon mass concentration, i.e. the MAC value, as far as achievable with filter-based methods. Therefore matching the scope of this journal.

This manuscript was initially submitted to the journal “Atmospheric Chemistry and Physics” (https://acp.copernicus.org/preprints/acp-2020-1190/) where it underwent a first round of reviews, which resulted in transfer to the current journal (Atmos. Meas. Techn.). I already provided a review at that stage (https://acp.copernicus.org/preprints/acp-2020-1190/acp-2020-1190-RC2.pdf). The authors already implemented these comments in an appropriate manner. Therefore, I only have few additional minor and technical comments listed below. Other than that, I can only stress that this valuable manuscript warrants publication.

Minor and technical comments:

1. Equation 1: The value(s) applied for f_{fil} should be reported. And/or the ratio “f_{fil}/MAC(COSMOS,SP2)”, which is the final and only factor applied to measured b_{0} for inferring M_{BC}.
2. Caption of Figure 2: I suggest to add something along the line: “D_{m} is the mass equivalent diameter of bare BC or FeOx”.
3. Section 2.2.3.3: It seems that actual wavelength of the MAAP instrument at Fukue was measured and found to be 639 nm, which is a very slight difference from 637 nm reported in the literature. It might be worthwhile to mention that 639 nm is an actual measured value.
4. Figure 3c: choosing logarithmic instead of linear axis scaling might provide better visualization of performance in the lower concentration range, where the majority of data points appear. In the current variant it is not perfectly clear whether the value of the regression slope, which is driven by the high concentration data points due to fixing the axis intercept at the origin, only matches data well at high concentrations or at low concentrations too.
5. Choosing logarithmic instead of linear axis scaling might be preferable for several figures for the reasons laid down in the previous graph.
6. Section 3.1.1: please put emphasis on the fact that quoted diameter values are BC mass equivalent diameter.
7. Line 358: “These results show that on average, the agreement between MBC (COSMOS) and MBC (SP2) at Alert was within 10 %.” – What exactly means “on average within 10%”?
8. Figure 4: It is very good that the instrument comparison is presented in different ways. Furthermore, distinction of performance in different concentration ranges as done in panels 4e and 4f is valuable. Having said so, the threshold is chosen at a very low level (2 ng/m3), which separates the data set somewhere in the single digit percentile range. I’d rather suggest to split somewhere between 15th and 25th percentile. In any case, the histogram of the low concentration data only should also be added (besides all data and high concentration data, which are already shown). This comment also applies to several other figures).
9. Line 393: “The babs, and therefore the MAC, for the PSAP and the two CLAP instruments (CLAP1, CLAP2) agree to within 8%”. – Based on the values reported in Table 2, the difference appears to be somewhat larger?
10. Caption of Table 3: please explain the variable “V” in the caption (i.e. inter-quartile range in relative terms). This comment may apply to other table captions too.
11. Line 462: check consistency of quoted V-value with Table 3.
12. Line 475: PSAP derived absorption coefficients are around a factor of 1.5 larger than aethalometer derived absorption coefficients at Ny Alesund. Zanatta et al. (2016) reported a systematic difference the commercial and ITM variants of the PSAP, which is approximately in this range (Sect. 2.3.2 and Table 3 in their manuscript). Which variant of the PSAP is operated at the Ny Alesund site? 
13. Line 586: repetition of “at Alert”.
14. The authors discussion that BC mass from MAAP default instrument output systematically differs from their direct measurements due to a difference in factory default and actual MAC value. A brief discussion for BC mass from aethalometer default instrument output compared to their direct measurements should also be included, e.g. in Sect. 4.

Review of the manuscript "Estimates of mass absorption cross sections of black carbon for filterbased absorption photometers in the Arctic" submitted to Atmos. Meas. Techn. by Ohata et al.

The presented article is an important contribution to the understanding of the relationship between black carbon and light absorption in Arctic regions. The methods used are appropriate for the available data and the results are well supported with figures and tables. In general, the manuscript is well written.

However, the reviewer disagrees with the authors in one particular point in the interpretation of the results. The reviewer recommends the manuscript for publication after considering the comments.

General comment:

The authors use a filter-based absorption photometer to determine the mass concentration of black carbon. In line 136, the authors write: "We critically re-examine the concepts underpinning the use of filter-based instruments to estimate M_BC." This approach should have been discussed more critically.

The COSMOS is basically a filter-based absorption photometer that evaporates the volatile components by heating, so that only the absorption of a soot core is measured. This device was "calibrated" (compare lines 641 to 644) using an SP2 with one type of aerosol.  This is critical because two different properties of the soot were compared. However, the determination of a value for MAC(COSMOS) for this aerosol type is correct. In this study, an application to arctic aerosols was tested.  Although comparisons of M_BC with COSMOS and SP2 for Alert and Fukue Island stations agreed within the uncertainties, the conclusions should not be that the COSMOS using a universal MAC value is a "secondary reference device" for M_BC.  As with any filter-based absorption photometer, the result should be labelled as "equivalent mass concentration" (see Petzold et al. 2013) using a specific MAC.  

The comparison of two filter-based absorption photometers, where the volatile shell is removed in one device, thus shows the light absorption enhancement factor, which is an important component of the MAC. Furthermore, like all other filter-based absorption photometers, the COSMOS is subject to certain errors such as uncertainty due to particle penetration depth (see Nakayama et al. 2010).  

In the view of the reviewer, in this study only the comparisons of a light absorption photometer with an SP2 strictly fulfil the conditions to derive MAC values. The other comparisons of COSMOS with other filter-based absorption photometers are rather studies on the enhancement factor.

The reviewer does not want to criticise the quality and overall results of this study, but to point out that soot in particular is a sensitive issue due to different metrics and the common nomenclature should therefore be followed very strictly. The reviewer proposes to refer to the facts described above and to denote the BC mass concentrations derived from  COSMOS as "equivalent" mass concentrations.

Further comments:

Line 177: better use attenuation coefficient instead of extinction coefficient

Line 188: give values for f_fil

Lines 191 – 207:  It is common for filter based instruments to report equivalent black carbon assuming an MAC. But in this case it is unfortunate to have to different MAC values when changing the filter type. The MAC is a property of the particle and not of the instrument. The reviewer suggests to attribute the changes of the sensitivity to the value of f_fil instead of MAC.

Line 271: Magnetite is not a good proxy for iron oxides to the reviewer's knowledge. The imaginary part of magnetite seems to be very high. This value cannot explain the reddish colour of many minerals. Alternatively, the refractive indices of hematite could be considered (e.g. Sokolik et al. 1999). 

Lines 315–320: How large are the uncertainties of the correction value C0?

Line 335: The MAAP was originally calibrated for the wavelength 670 nm. Was the correction factor of 1.05 taken into account to adjust the wavelength (see Müller et al. 2011)?

Line 877: Line break between Petzold (2004) und Petzold (2005)

References:

Petzold, A., Ogren, J. A., Fiebig, M., Laj, P., Li, S.-M., Baltensperger, U., Holzer-Popp, T., Kinne, S., Pappalardo, G., Sugimoto, N., Wehrli, C., Wiedensohler, A., and Zhang, X.-Y.: Recommendations for reporting "black carbon" measurements, Atmos. Chem. Phys., 13, 8365–8379, https://doi.org/10.5194/acp-13-8365-2013, 2013.

Nakayama, T.,Kondo,Y.,Moteki,N.,Sahu,L.K.,Kinase,T.,Kita,K.,&Matsumi,Y.(2010).Particle size-dependent correction factors for filter-based aerosol absorption photometers: PSAP and COSMOS. Journal of Aerosol Science, doi:10.1016/j.jaerosci.2010.01.004.

Sokolik, I.N. and Toon, O.B. (1999) Incorporation of mineralogical composition into models of the radiative properties of mineral aerosol from UV to IR wavelengths, J. Geophys. Res., Vol.104, 9423-9444