Peer-Review Record
Predicting Outflow Hydrographs of Potential Dike Breaches in a Bifurcating River System Using NARX Neural Networks
Hydrology 2021, 8(2), 87; https://doi.org/10.3390/hydrology8020087
Reviewer 1:
Reviewer 2:
Anonymous
Hydrology 2021, 8(2), 87; https://doi.org/10.3390/hydrology8020087
Submission received: 30 April 2021
/
Revised: 25 May 2021
/
Accepted: 28 May 2021
/
Published: 3 June 2021
(This article belongs to the Special Issue Advances and Perspectives in Flood Risk Modeling, Assessment and Communication)
Round 1
Reviewer 1 Report
Dear authors, I find this is an excellent piece of work. Nevertheless, I made a number of comments you may want to consider to improve the paper: see the attached file.
In addition, please consider the following:
- when the proposed approach is presented, the discourse could be clearer and simpler by clarifying the purpose (real time forecasting), the simplifying hypothesis (known potential breach locations, so the models do not predict where, but just which; deterministic breach approach based on overtopping criterion only, independent of type of structure and conditions; ...) and the procedure (calibrating a battery of parallel NARX for the set of potential breaches or, more precisely, for the set of potential breaches that the hydraulic models showed as failing indeed during at least one of the multiple events adopted in training)
- the concept exposed quite marginally at lines 196, 197 on the implicit consideration of the sequential effects of any breach downstream by the battery of NARX thanks to the fact that the physical model does it too should be given the appropriate emphasis and space because, in my view, it addresses a quite delicate point
- it is important to stress the limitations of the approach: for instance, in reality, levees may fail because of other mechanisms and a story of events may have debilitated certain locations more than others. Moreover, potential breach locations -exactly because they are identified- are the object of reinforcement which make them no longer belong to the potential sites set...calling into play new...unknown sites
- discussion at Par 6.1: it would be very interesting to show how the system actually behaved in the historical past: did it experience multiple failures even for events that, under the hypothesis of this study, would not?
kind regards
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 2 Report
Please see the attached file for comments.
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
