dear authors

thank you for the submission of your intesting manuscript on "A systematic exploration of satellite radar coherence methods for rapid landslide detection" to our journal.

As you know, two reviewers have provided detailed reports, which you have replied in detail in the discussion phase. Both comments were very positive and one of them suggested some minor revisions. The second one only suggested minor comments and corrections. Therefore, I would now like to invite you to submit a revised version of your manuscipt, incorporating all the said modifications.
Would you please also provide an ‘author’s reply’ to the reviewers. Please can you also include a track changes document between the old manuscript and the new one (you can include this as part of your ‘author’s reply’).

I look forward to seeing the next version of your manuscript which I will not send out for further review, but rather, will make the decision myself, assuming no major items come up in the revised manuscript for which I need outside reviewers to aid me in my decision.
 
Regards
FC

dear authors

after seen your replies to the comments of the two reviewers, with particular reference to Reviewer #2, I am pleased to inform you that the revised version of your manuscript NHESS-2020-168 can now be accepted for publication in the journal.

As a side note, I have much appreciated your approach on using SAR data for this kind of mapping and change detection and I think that the method could be potentially powerful in helping post-disaster assessment and residual risk estimation with special reference to rainfall-induced landslides, as well, due to the capability of SAR to work under bad weather conditions and dense cloud cover.

As a minor issue, that I consider a technical thing, I have seen your reply to reviewer#2 comment on the usage of ROC curve for imbalanced data. Although your reply can be considered sufficiently detailed, I would like you to suggest the usage of ROC convex hull approaches in identifying possible differences in ROC curves due to a different imbalance rate, as suggested e.g. by Chawla et al. (2002) in their SMOTE method. I do not deem necessary a resampling of your dataset for rebalancing but a look at this approach may render your accuracy assessment more robust, if applicable.

This is only a suggestion. In case you want to follow it, please make the modifications and upload a new manuscript as soon as possible, along with a reply to my comment. Otherwise, just upload the present version in replying to the "Publish subject to technical corrections" message.
After that, the paper will directly go to the production office for publication.

Thank you again for submitting to NHESS,
best
Filippo