A Flood Inundation Modeling Approach for Urban and Rural Areas in Lake and Large-Scale River Basins

Round 1

Reviewer 1 Report

        In my point of view, the manuscript needs major revision before it can be accepted. I hope the authors are willing to modify their manuscript, also taking into account the comments provided here.

• Recently the state of the arts statistical machine/Deep learning techniques was applied for the flood modelling approach? What is the reason for choosing the HEC-RAS 2D model over other established ML/DL techniques? Also, is possible to compare the physical model vs the statistical model in your analysis? There are tons of ML-based forecasting models such as Quantile regression forest, SVM, boosted tree, decision tree, etc., which were successfully used to estimate meteorological variables such as soil moisture, temperature, precipitation, etc. Also, you need to provide more literature reviews in the introduction section associated with the research gap/limitation.

• As the model was applied for different climatic conditions, can you provide climatic information for the selected study areas? You can show Köppen–Geiger climatic zones on the map.

• You used multiple datasets with different sources for Hydraulic Modelling. How do you merge all those datasets into a common time series, please explain? Which interpolation method did you apply? It is necessary to report how the matching is carried out which will include information about the resolution(temporal/spatial), sources, data ranges, etc. Can you provide a table with source, resolution, and methods?
• Can you provide a high impactful schematic diagram to understand the proposed research framework where the big impact of the results can be presented?

• Can you explain geophysical and climatic variability in Greece?
• Why did you analyze your model only 3 design period (50, 100, 1000 years), why you did not perform it for 1-year, 2-year or other frequent events?
• You mentioned Block method (for T = 50 and 100), and Worst profile method( for T=1000 years) in your methodology section. Why did you choose these methods. Can you explain?
• How did you determine the time of concentration?
• In Urban area did you consider any underground drainage system? Or did you consider all surface flow?
• Did you validate your hydraulic result with active gage stations? Can you explain about this issue?

• Can you explain about the transferability of the model? How did you apply your hydraulic model on ungauged locations? What is the criteria about model set up?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Review of manuscript “A flood inundation and modelling approach for urban and rural areas in ungauges lake and large-scale river basins”, by G. Papaioannou et al., submitted to WATER

22 March 2021

The manuscript presents a flood hazard modelling exercise that must have required quite a lot of effort given the tough combination between wide areas and high resolution. These models are part of the work performed for implementing the European Directive on Floods at a national level (in this case in Greece).

The material fits the aims and scope of WATER and the manuscript is in general well written (even if, at some instances, the style is a bit more didactic than one would expect from a research paper). I provide a list of detailed comments below, and I do not expect that they will be much troubling for the authors. My main suggestions are to describe more clearly the modelling approach (I quite did not understand if the mentioned junctions separate different models or different parts of one implemented model) and the sensitivity analysis (it is, in the end, unclear if and how the uncertainty in rainfall was accounted for).

My recommendation is that the manuscript be acceptable after a minor revision. Congratulations to the authors for this nice piece of work and I hope that my comments will be useful.

Line-by-line comments

------------------------------

Title: What is a “flood inundation and modelling approach”? Likely, “and” is to be removed.

107-108: Statistics is not a physical process. In this sense, I do not get the three categories “statistical, hydrological and hydraulic”. Shouldn’t “hydrological and hydraulic” be better? To me the uncertainty in the parameters of the IDF curves, mentioned at the following lines, is hydrological and not statistical.

111-117: this period is too long and I suggest splitting it to make the reading smoother.

119: this line sounds as hindering the possibility to generalize the approach, that instead should not be site-specific if the value of the manuscript is to be claimed. I suggest to add some consideration on possible generalization of the method (that, before reading, sounds promising).

Fig 1(A): It is not clear why we need a junction J1 at an end of a reach. Please add a flow direction along the main stems.

178: it is unclear which this “broader region” is.

18-190: one does not get why an “inadequate natural discharge capacity” is considered a “minor cause of flooding”. Shouldn’t this be the core of the business?

205-212: these lines sound misplaced since they refer to both basins while they are still under the heading of sub-section 2.2.

215: based on the above, I am still uncertain if results on risk will be presented (estimating damages from depths, for example) or the study will be limited to hazard.

220: 1000 years is a huge return period. Out of curiosity: is it really considered in Greece? In Italy, from where I am, the maximum T considered is 500 while 1000 is used only for dams.

230-245: I suggest to rephrase this paragraph. Lines 231-232 suggest that only the uncertainty in initial soil moisture was considered. Lines 235-241 may be interpreted thinking that all the three sources were considered. Finally the reader gets that nine models were produced considering uncertainties (2) and (3). I think that the paragraph can be made clearer to avoid a reader’s misinterpretation.

261-267: uhm, here it seems that also an uncertainty in rainfall was considered. So, in the end, were the un models 27 rather than 9?

256: I guess that there is an underlying assumption that these 24 h are longer than any relevant time of concentration, isn’t it?

293-294: aren’t river paths and banks part of the geometry rather than of ICs and BCs?

335-337: Again out of curiosity since I have not been using it: is Hec-Ras unable to consider bridges if run a 2D mode?

339: “multistage impact of flooding” is obscure, please specify.

361-377: a simulation for a large area is thus performed splitting it into sub-areas and running models for the latter. It sounds like interface conditions need to be considered very carefully. Moreover, what about feedbacks?

378-394: similar comment for splitting the simulation in case of a large river. Furthermore, it looks like the breakpoints between successive simulations should be chosen at locations where the river is well contained by the cross section, to avoid complications in the treatment of upstream and downstream boundary conditions. Second, the simulation direction fixed (at line 381 it is mentioned that the computation starts with the upstream), which is reasonable considering that a flood hydrograph will be built from the upstream and reduced by upstream flooding, but complicates the business in the most likely case of subcritical flow; how are interface conditions checked?

A more general comment is the following. Since a river is split into several reaches, one may think that a 1D approach would be better chosen in the reaches without overflow to speed up the computation. In these reaches a 1D model would adequately represent the flood propagation. What is the authors’ opinion on this.

419: I suggest to reconsider the use of the term “analytical”, that is frequently used in contrast to numerical. As far as I understood, there is no analytical modelling in this work.

450-469: this paragraph is a bit confused and could be rephrased for clarity. Furthermore, are these junctions implemented as junctions in Hec? If this is the case, the model is not split. Or are these conceptual model junctions where the operator sets boundary conditions to transfer information from a previous model to a following one? That would be more consistent with the description in section 3.3.

472: same comment, are these conceptual junctions or are they implemented in the model?

507: how was the rainfall duration determined/chosen?

Figure 4: for T=1000 y, the outflow hydrograph is much lower than the inflow one. Does such a strong reduction happen due to flooding in the short distance (Figure 1) between these junctions?

Table 3: here we have again nine models, so the uncertainty in rainfall was not considered (or it seems so). Related to a prior comment (261-267).

649-650: unfortunately, no criteria “to select the optimum configuration” are provided. Can some discussion be added in this respect?

651: reaching the end of the main body of the manuscript I can answer the comment made at line 215. There is no conversion of hazard into risk in this paper, so a claim that risk modelling was accomplished here should be avoided. It is just for clarity and I do not intend to discount the value of the work, since I fully guess how painstaking this modelling must have been.

683: more than what else?

Arriving at the end of the manuscript, I would suggest to remove the word “ungauged” from the title. The issue of a very large basin is accounted for, while the issue of a lack of model validation (due to missing gauges) is not even mentioned in the manuscript and should be removed from the title.

A few grammatical suggestions

64: I would use “has been facing” instead of “faces”, sounds more correct.

73: Better to use “Consequently”?

77: I suggest to remove “are more general and they”.

82, 83: “can one”

135: “from 469 to 1808”

182-183: check sentence.

186: check and fix “are intensify”.

188: “with consequence” -> “thus inducing”.

196-200: check and fix “The area … is

211: “than” -> “from”

260: remove a “)”

437-438: was implemented

498-499: the sentence lacks a verb

500: associated

501: requiring

510: extremely

516: assumption

528: was performed

533: procedure

685: indicates

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors,

The review of this manuscript has left me with mixed feelings. On the one hand, the title of the manuscript seemed interesting enough, and it seemed to address or propose an interesting methodology for the analysis of flood hazard and risk. However, after reading and reviewing it, I have the feeling that the manuscript provides little that is really new.

The methodological section is excessively long and redundant, since with the extensive development of the same, the novel aspects are simply listed, without really entering into its description and analysis in depth. In addition, this section (section 3) presents a confusing and chaotic wording that does not allow for easy interpretation. This is undoubtedly helped by the fact that the authors jump randomly to comment on aspects related to the hydrological analysis, other times to the hydraulic-hydrodynamic analysis, and very rarely to the generation of the design hyetograms. It is evident that this section needs to be completely restructured and rewritten in order to reduce its length and increase the number of important and really novel explanations or methodological descriptions.

On the other hand, the results section is not very extensive, and describes very briefly the main results obtained, without going into detail in their description. In addition, I believe that a critical discussion section of the results is needed, which could also include aspects on the relationship of the results (including the uncertainty analysis carried out) with the flood risk (including aspects of the exposed elements and their vulnerability to floods, which have not been considered in this manuscript and therefore invalidate the use of the term "risk").

Without all this, the manuscript remains a simple flood hazard study, which does not really contribute anything new to this field of study.

For all these reasons, I consider that the authors should undertake a rewrite of the manuscript in which the novel aspects are detailed, not only from a theoretical point of view (enumeration of these novelties) but also from a practical point of view in which the advantages and improvements that the present study and the methodological framework that it advocates bring to the study of flood hazard and risk are really shown.

In addition, in the attached document the authors can find other comments related to the present manuscript.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors significantly improved the manuscript by addressing all the comments.