dear Authors

after the last review and after checking your responses to the referee's comments, we believe that you have addressed almost all comments and critical issues in the manuscript.
Before going ahead with publication, though, I am left with a possibly important, yet secondary, doubt.
From your manuscript, as well as from the comments and responses to referee #1, at Page 8 Line 21, I assume that you confirm that the limit to consider a SU as unstable is >0.15% since this is the average landslide density in your area.
As you may be aware, however, landslides are not a nicely autocorrelated property of the landscape and they are usually a pulse-type process (that is, 1/0 binary process of presence or absence). You are not obviously the only one using the landslide density as a measure of susceptibility in space. I am not saying the approach isn't correct in theory.
I am only suggesting that you should probably look at smaller neighbourhoods of any single point, or SU and, possibly, adopt a more flexible approach (such as e.g. a moving window estimation of the average landslide density in a much smaller area than the whole study area) to define the threshold for instability. That is, to shift from a constant to a spatially variable threshold. Otherwise, my impression is that you may get highly overestimated hazard levels while having very nice AUCs and validation data for true positives.

This is the only thing that I find it remains to be cleared out in the manuscript. But it is not an irrelevant one. For this reason, I ask you a last effort in trying to modify this part and in comparing new results with the one you obtained.

After the new response the manuscript will only be reviewed by myself, without requiring any new review round.

Best regards
FC

dear Authors

thank you for the last modifications to your manuscript and for waiting our processing time.
I consider all your responses and modifications adequate, even though I cannot agree with some of your statements on the choice of a fixed landslide density figure. But this is just my personal view and carries no weight in this case. Something that we could discuss afterwards on a personal basis maybe, since I consider the topic very relevant in general.

So, I apologise for the delay but I am now glad to communicate that the paper is accepted for publication in the present form and that the manuscript will be taken in charge by the processing unit.

Thank you again for submitting to NHESS,
my best wishes
FC