Dear authors,

I want to start by apologizing for the time it took me for the decision, this was not an easy one to take. I had to carefully read the manuscript a couple of times, and also carefully weigh the reviewers comments, since although I like the general idea of the paper a lot, I found that there are some serious weaknesses with the methodology. Unfortunately, I cannot accept the paper for publication in its present form. Although I do appreciate your effort in addressing the reviewers comments, especially the most critical one, I was not able to be convinced that the underlying assumptions are totally sound, and that makes the comparison with AERONET weak. I will try to explain here why, based primarily on the comment on refractive indices of the reviewer of the revised version.

AERONET assumes internally mixed aerosols, with the exact same composition for both the fine and the coarse mode. "Always!", as the reviewer states, regardless of what happens in the real atmosphere. This means that there is always some coarse mode absorption, which might not be negligible, and the absorption from any constituent (BC and dust) are always in volume mixing with the scattering aerosols. By assuming different components, the manuscript assumes that when the "better" parametrization with externally mixed aerosols deviates from the "not so good" AERONET retrieval with internally mixed aerosols, indicates the presence of BrC. The big problem here is that the internally mixed aerosols do not behave the same way with the assumed externally mixed aerosols when BC/BrC explicit optical properties exist, neither when a core-shell assumption is made, when it comes to optics calculations. For this, the explicit modeling of absorption presented in the manuscript is inconsistent with the AERONET assumptions, so one can't directly compare the two, not even qualitatively. This is especially the case for AAE, which is very sensitive to the coarse mode volume fraction, even when fine mode dominates, which is not considered in the manuscript.

On a positive side, I believe the work presented is really worth pursuing a bit further. Brown carbon emerged to be an important aerosol component, and our understanding of it is really poor, if not worse. I would very much like to see a paper that you work with someone from AERONET, use your assumptions and repeat the AERONET retrievals, so create your own product, and then be able to say quantitatively something about BrC. Repeating the AERONET retrievals by paying attention to internal/external mixing and fine/coarse mode absorption (by using different refractive indices or different size assumptions) will allow you to compare apples with apples and make a study that I believe will be very highly cited in the future. I would strongly recommend you to try to address this weakness and resubmit the manuscript for publication.

Best regards,
Kostas Tsigaridis

Dear authors,

Following your appeal to my decision, I will proceed with assigning a couple more referees to the paper. I will ask them to try to submit their review quickly, to avoid any further delays. I will let you know as soon as I have the reviews back on whether further revisions are required from your side.

Best regards,
Kostas Tsigaridis

Dear authors,

I did request one additional review, from a new reviewer who is a well known expert on brown carbon, asking to explicitly comment on both the paper and the past reviews. As you can see from their report, the paper has strong weaknesses, and the new reviewer agrees with the past critical reviews. The interesting point the reviewer makes though is that "the data to do this accurately doesn’t exist at this point, making the efficacy of the increasing number of papers utilizing AAE approaches, like this one, questionable". This means that based on their opinion there is no way to improve the results of the manuscript, due to limitations of the currently available data used as input.

I regret to say that I intend to retain my original decision of rejection.

Best regards,
Kostas Tsigaridis

Dear Authors,

As agreed in response to your appeal, I will guide the review of the preceding editorial decision and of the revised manuscript. In order to be able of consulting reviewers through the online editorial support system, I am choosing the option "reconsider after major revisions", regardless of the extent of revisions actually performed in this step.

With best regards,
Uli Pöschl

Dear Colleagues,

After carefully considering the arguments raised in the course of the peer review process as well as the adjustments and clarifications implemented in the latest version of the manuscript, I am ready to accept the manuscript for publication in ACP subject to additional adjustments as detailed below.

Please distill the most important messages/statements from the text you added at the end of the manuscript (copied below) into a few sentences to be added at the end of the abstract:

"However, these results are subject to uncertainties associated both with the methodology and with the dataset to which it is applied. The core issue for all methods that use the AAE to estimate the absorption from BrC is the uncertainty associated with the absorption of BC. Our method improves upon previous studies using this approach by using the information from the wavelength dependent measurements themselves and by allowing for an atmospherically-relevant range of BC properties, rather than fixing these at a single assumed value. Additional constraints on BC optical properties and mixing state would help further improve the method.

Given the large uncertainties associated with AERONET retrievals of AAOD, the most challenging aspect of our analysis is that an accurate globally distributed multiple-wavelengths aerosol absorption measurement data set is unavailable at present. Thus, while our study provides qualitative global constraints, and insight into BrC aging processes from the aethalometer observations, achieving a better understanding of the properties, evolution, and impacts of global BrC will rely on the future deployment of accurate multiple-wavelength absorption measurements to which AAE-methods, such as the approach developed here, can be applied."

The entire text appears rather long for being added to the abstract, but on the other hand, there is no hard limit on the length of manuscript abstracts in ACP. Thus, please transfer most if not all of the information from this text into the abstract, especially a clear statement about the limited reliability of the presented results and how more reliable results may be obtained in future studies building on the presented method.

Moreover, please note that in the current PDF version of your manuscript most figures are of low quality (low resolution, and sometimes low contrast/large overlap of labels, lines and symbols). Please improve the figure quality upon revision and make sure that high quality is obtained and maintained in the course of proofreading and final publications.

Many thanks and best regards,
Uli Pöschl