I would like to warmly thank both Referees for their helpful reviews and for having already, through the second round of comments discussion, provided additional comments on how improving the paper, following the Authors’ first clarifications. I thank also the Authors for the exhaustive replies, where they carefully and appropriately answered to the most important points raised by the Referees concerning the scientific methods and the main objectives of the work.

I invite the Authors to submit the revised manuscript they are already working on, suggesting them to keep into account the Comments made by the Referees in their first and second comments and in particular I suggest them to revise the presentation structure, devising a way to better clarify the steps of the proposed procedure, so that the readers may fully appreciate your work and, if they want to, may be also able to implement it themselves.

Dear Authors,
I do thank you for the detailed reply letter and for the changes made in order to address the comments by the Referees.
I confess that, even if you clarified many points, I still find the presentation cumbersome and the procedure difficult to follow and not described very clearly, as demonstrated also by the last doubts raised by Ref#2 in his major comment (it is not clear the use of the indexes presented in section 3.2, where both RRMSE and heterogeneity are presented, even if it is said that only RRMSE is used for calibrating the neighbourhood size…). It is now probably difficult to re-structure it completely (and after all, none of the Referees asked you to do so in their last reviews), but you should try to clarify such doubts (if the Referee - who has carefully read the paper more than once - still find some points dubious, it will be certainly worse for a ‘regular’ reader…).
Therefore, I do hope you may find a way to further clarify the procedure; I would suggest adding a flow-diagram with all the steps and the variables and indexes used in the different phases.
In terms of minor comment I do agree with the Referee on the suggestion to modify the terminology on the distance (as already done in another point) and to try to improve the Quebec map, making its features more ‘recognisable’), whereas personally I prefer that the scatterplots are on separate panels, since I find them more readable.
I look forward to reading the last version of the paper,
Best regards,
Elena Toth
HESS Editor

Dear Authors,
I believe that the addition of Figure 1 will clarify the procedure and improve the readability and the readers’ positive reception of your work, and I do appreciate also the other clarifications you made in the paper as suggested by the last comments made by the Referee.
I believe the paper may now be accepted for publication, after you will have done a close check of possible language improvements in the text.
Best regards,
Elena Toth
HESS Editor