Thank you for your important revisions. The main three issues raised by both reviewers concerned the lack of strong hypothesis to introduce the paper, the lack of evidence that dust deposition did occur and important issues regarding the statistical analyses. In addition, reviewers also made many minor comments. Following the very detailed report of both referees, the authors have carefully answered point by point to major and minor issues raised. Consequently, authors have substantially rewritten some sections, new sections have been added, new figures and tables proposed and finally the references have been updated. The paper reads very well now also because a native English speaker proofread the manuscript.

I therefore consider that your paper can be now published after some minor revisions.
I made many edits and comments directly on your text, in addition, I recommend that you answer the following points:

Abstract:
- I wonder if updates could be done considering that several important points and conclusions are now supported using different statistical tools: I find it a pity that the efforts made on the statistical aspects to consolidate the interpretations of the results are not at all reflected in the abstract and the conclusion. I strongly encourage the authors to do so.
- I found the section starting L18 quite confusing as it somehow contradicts the observations analyzed as a 'virtual time series' under the influence of dust deposition with different time lag. I suggest removing that sentence and to focus on a clear statement about the 'virtual time series concept' results that I think are very interesting.

Results and Table 1.
I would be more precise concerning the sediment traps data and I made suggestions directly in the text. Please report in the text the corrections that I suggest for the lithogenic cumulated fluxes. The point here is that (1) at TYR the sediment trap sampling likely missed a part of the export as it started 6 days after the event occurred and (2) at FAST, the last trap represented 24 hours collection between June 5 and 6, so the collection likely missed the main lithogenic export that likely occurred after. Lithogenic fluxes at both sites are thus likely well below the actual export following the events.
There is one remark from Referee#2 concerning the sampling 0-300m: I agree with you answer but I think your final statement “Also note that the observed impact on zooplankton is more significant because it integrates the whole water column” should be somewhere in the text, conclusion maybe (need to rephrase the sentence).

There are a number of additional figures and table that are presented in response to reviewers comments. I would definitely recommend adding this pertinent material to the Supp. Info.

Dear Authors,
I am happy to accept publication of your manuscript pending the revisions suggested by the associate editor Cécile Guieu.
Sincerely,
Christine Klaas