Optimization of Artificial Neural Network (ANN) for Maximum Flood Inundation Forecasts
Round 1
Reviewer 1 Report
Water-1301065 review:
Title: Optimization of Artificial Neural Network (ANN) for Maximum Flood Inundation Forecasts
Summary:
The authors presented a framework of optimizing Artificial Neural Network (ANN) by training the system based on synthetic flood events; extending the training datasets with the use of hybrid sets, and selecting the best trained function based on possible functions.
Reviewer comments:
I would like to preface my review by stating that my expertise is not of ANN, but more on urban flood modeling. I would like to prefer the Literature review and Methods be reviewed by another reviewer who is more familiar with the procedure, and hopefully get the suggestions from them. However, I can provide some additional suggestions with regards to the overall construction of the research paper and the general requirements for publication.
Major comments:
- Abstract
Provide a brief description of the methodology used
- Introduction
Provide short description of data-driven models and highlight the use of ANN for flood forecasting
- Methods
Show the map of the domain and the gauge stations.
- Discussion
Why are Figure A1 and A2 put in Appendix and not in Results?
- References
Follow MDPI format for citations: https://www.mdpi.com/authors/references.
Ex. Journal articles
Author 1; Author 2; Author 3; etc. Title of the article. Journal Abbreviation Year, Volume, Firstpage–Lastpage, doi:prefix/suffix.
Minor comments:
The whole paper needs to be proof read to fix minor formatting errors (Capitalization, inconsistent punctuation marks, referencing format).
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1 – The introduction section is a little bit lacking in proving the most recent studies in this area. The followings are good examples that can enrich the introduction section:
Tamiru, H., & Dinka, M. O. (2021). Application of ANN and HEC-RAS model for flood inundation mapping in lower Baro Akobo River Basin, Ethiopia. Journal of Hydrology: Regional Studies, 36, 100855.
Gavahi, K., Abbaszadeh, P., & Moradkhani, H. (2021). DeepYield: A Combined Convolutional Neural Network with Long Short-Term Memory for Crop Yield Forecasting. Expert Systems with Applications, 115511.
Zhou, Y., Wu, W., Nathan, R., & Wang, Q. J. (2021). A rapid flood inundation modelling framework using deep learning with spatial reduction and reconstruction. Environmental Modelling & Software, 105112.
2 – The authors have chosen the size of 50 for their grid size. What is the effect of a higher or lower grid size? Why not choosing 25?
3 – “After that, the appropriate number of neurons and hidden layers are 107 determined by varying the number of hidden layers from 2 to 12, and the neurons number 108 in each layer from 10 to 60” what is the rational behind choosing these numbers. I suggest that authors elaborate more on the reasons for choosing these values.
4 – How did the author selected the nearby river for each pixel? What distance is considered as near?
5 – I suggest that authors consider ConvLSTM for their future works since the method is most suitable for flood inundation mapping applications.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors have fully addressed my comments and I suggest the manuscript publication in Water.
