Dear Carmelo and co-authors,

thanks a lot for providing your online replies to the referee comments. I would like to invite you to resubmit a revised version of the manuscript for further consideration.

Both referees consider their comments to require major revisions. Even if the comments may not seem numerous, some are indeed substantial, and please consider them carefully in your revision. In particular provide relevant model characteristics pointed out (not down to the last detail but relevant concepts) and model credibility proof (at least a short account on validations and details on strengths weeknesses and quality in the relevant low flows) - so that it can be read without requiring to search out all previous publications. Also, please consider carefully the comment on statistical significance of presented trends and changes; I would agree that direction only is not such a strong proof of change; among other comments.

I am looking forward to seeing the revised version.

Best wishes
Kerstin

Dear Carmelo and co-authors,

the two reviewers have assessed the revisions you made and have come to different conclusions.

As you can see, you have not been able yet to convince Rev. 1 about the quality of the model performance regarding the specific type of drought characteristics changes that you are simulating in this study. I agree that the reviewer touches on a very important fundamental aspect in model assessment in general and that this is important. Please make an effort to provide some material to convince us (it can also be specific reference to an evaluation documented elsewhere but needs to relate to the very aspects of drought modelled here) and based on that discuss where/when uncertainties are higher/lower.

In addition, the reviewer had a number of minor comments which can help increase the impact of the paper. Please consider them carefully in this next revision round.

Thanks and best regards
Kerstin

Dear Carmelo, Gustavo and others,
thanks for revising your manuscript. Reviewer 1 still has a concern, that I also find very important. Model performance in the low flow/drought part of the discharge distribution needs to be made transparent. We all know that it is never as good as performance in other parts of the discharge range - so no worries.
But model performance in the target variables (here: "drought traits" (title!) needs to be fully transparent and well documented for readers, always.
If I understand correctly, the cited references are not sufficient, because they used a differently calibrated LISFLOOD than the version parametrized for this study here. If this is indeed the case, then there is no way around the need to show a validation of low flow range flows for THIS version here in addition and specifically (at least as a supplementary material).
I therefore ask you to please carefully address the reviewer's remaining concern to make the model validation for this study clear, accessible and specific, before I will then take the final decision.
Thanks and kind regards
Kerstin

Dear Carmello and co-authors,
thanks for compromising. The abilities and application potential of large-scale models are of course different than those of a calibrated basin scale model and I know the application is Macro-Governance rather than local management. But it is important to be transparent on what large scale models are able to simulate quantitatively also for readers and other modellers to know what is needed to advance such models. Hence I appreciate your providing flow deviations. From a water management perspective deviations up to an order of magnitude in a low flow situation are of course still too large to allow using the model for actual basin scale or even river reach decision making. Hopefully future work can reduce these errors.
Thanks again and best regards
Kerstin