An Urban Flash Flood Alert Tool for Megacities—Application for Manhattan, New York City, USA

Round 1

Reviewer 1 Report

The manuscript describes a new framework/workflow for rapid mapping of urban flooding. The topic is very important and timely so I approached this review with much interest. Unfortunately, the manuscript is poorly written both in terms of its overall English and, most importantly for the review, the description of the methodology. This means that I'm unable to robustly evaluate the core of this paper - the new methodology. Moreover, validation of the results is practically non-existent. There is a validation sub-section but its basically comparing the flood map to a random picture of a flooded street in the area (its location is not even reported). The authors acknowledge this limitation and leave to a future study. This adds to the sense that this manuscript is an attempt to published an initial (underbaked) part of a study. 

A more robust description of the methodology may overcome some of these issues as the focus of the paper can be steered toward that aspect of the study (maybe add sensitivity analyses of the workflow components).

Specific comments are included in the attached pdf file. Yellow highlight indicates sections which need to be revised, though major editing of English is needed throughout the manuscript.   

Comments for author File: Comments.pdf

Author Response

Rafea Al-Suhili

Cheila Cullen

Reza Khanbilvardi

June 7, 2019

Re: An Urban Flash Flood Alert Tool for Megacities – Application for Manhattan, New York City, U.S.A

Dear Reviewers,

We would like to thank you for your valuable suggestions, questions and comments. We have implemented many of your suggestions throughout the document. Please refer to the table below for detailed information.

Reviewer I

General

We agree with the reviewer about the importance and complexity of this topic. We would like to emphasize that at this point, we are presenting the first basic step of a new solution framework.

Regarding the validation process, we adjusted the verification section identifying the location of the flood area. At the moment, flood depth volumes are not available to compare spatial flood depth between prediction and reality. Nevertheless, as our objective for now is to only identify flooded areas, we leave this type of validation for future work. Nevertheless, as suggested by the reviewer, we have incorporated a section describing sensitivity analysis for the model.

Specific

Reviewer 2 Report

Overall recommendation

Major revision

General comments

This paper describes a model system for fast and simplified estimation of spatial extent of inundation in an urban area due to intensive rain. The model has two parts. One part is a water balance model which calculates the volume of water which is beyond the capacity of the normal pipe drainage system. The other part is a mapping device based on a DEM which uses the volumes from the water balance model as input and produces maps of inundated areas.

This is probably a useful approach for fast prognoses of inundation where time is of the essence and less emphasis is on accuracy. I am not totally convinced that the approach used is novel, but I have not found anything exactly similar.

The short computational times, and the ease of modelling comes at a prize. The main issues with the model system, as I see it, are the following:

- there is no dynamics in the modelled surface inundation; water is just allocated based on topography

- there is no feedback from the pipe system to the surface; once water goes down the drain it cannot reappear

- the hydraulic head driving the flow to the pipe system is just assumed; there is no feedback from the calculated inundation

It is not possible to deem the usefulness of the proposed model system as long as the inaccuracies due to the issues mentioned above (and other simplifications) have not been established based on comparison with observations.

You must elaborate on the issues listed above and/or explain why they are (or are not) important.

Specific comments

·       (line 114) The LIDAR DEM is presented as 1-ft. Assuming this is the horizontal resolution – what is the vertical resolution/ accuracy?

·       (line 115) The population given is based on residents. What is the daytime population? I would assume that for Manhattan the ration is quite large

·       (line 131) Please state very clearly how houses have been treated. Is the roof area totally excluded from the model?  

·       (line 139) Eq. 1. What is S ?

·       (line 159) The RR model used seems to be developed by the authors themselves. Although it is a simple model structure, it would be helpful if there were a reference to a detailed description of the model and some kind of verification

·       (line 195) Eq. 9. What is β ?

·       (line 206) Please check Fig. 3.  

·       (line 250) Table 2. Results are given with too many significant digits.

·       (line 260-263) Fig. 6,7. Please explain more in detail exactly where in the map the photo was taken.

·        (line 233, 245, 250) There are typing errors on these lines. Please check all text – there are more.

Author Response

Rafea Al-Suhili

Cheila Cullen

Reza Khanbilvardi

June 7, 2019

Re: An Urban Flash Flood Alert Tool for Megacities – Application for Manhattan, New York City, U.S.A

Dear Reviewers,

We would like to thank you for your valuable suggestions, questions and comments. We have implemented many of your suggestions throughout the document. Please refer to the table below for detailed information.

Reviewer II

General

We appreciate the reviewer’s valuable comments. We appreciate that she/he considered our work useful for fast prognoses of inundation where time is of the essence. This is exactly our research objective, to develop a model that quickly identifies flooded areas regardless of the depth. The model can be extended to find these depths by using the DEM model, but this is left as a future goal.

Specific

Round 2

Reviewer 1 Report

The revised manuscript is a major improvement over the initial submission. I think that the impact of the paper will increase considerably if it will offer a more explicit description of the methodology so that the readers could quite easily be able to reproduce (e.g. how exactly did you generate the flood maps) or even directly use (by sharing scripts) the framework. Now obviously the authors don't have to do it but that will greatly diminish the utility of this paper as the individual components of the methodology are not new or innovative by them selfs- the overall framework is. As it stands I think it will be hard for the readers to gain much from this paper.

Specific comments:

Line 300, 329-331, 370 - use metric units.

Flood map generation is still not clear to me. Figure 3 is not much help.

Line 437 - how does one estimate S?

Equation 2 and associated text - flood map generation method is still not clear enough. I suggest you used a "toy" example (let's say a 10x10 grid) to explain and visualize this process. How does one actually go about calculating these?

Line 561 -remove

Eq 7 - Dk should be defined here and not after eq 8. 

Line 580 - there is no equation 12.

Line 582 - how does the temporal variation in b calculate?

Line 583 - 'inlet flow head' is not mentioned or before. How is it calculated?

Line 584 - 'The actual and the simulated runoff depths are expected to have small differences, but they do not affect our objective of pre-identifying flood areas. The expected damages and consequences of severe floods, i.e. 40 cm or higher, will not be much affected by a few centimeters of flood depth.'  - maybe so but wouldn't it have a major impact on the calculated flow extent?

Line 628 and Table 1 caption - I don't think Table 1 shows the 'full' dataset - it is a subset of 5 basins (which is perfectly fine).

Figure 5 caption -  explain what is 'n'?

Figure 6 - can you specify the dL in this example? This is a good opportunity to clarify the methodology by demonstrating it with actual numbers. 

Line 948 - 'reasonable computational time' - provide specifics.

Table 3 is missing a caption.

Line 1017- 'Our approach presents reasonable flooding results for the sake of quick preparedness.' I don't disagree with this statement but it was not well-enough demonstrated in section 3.2. Add a 'concluding' comment in 3.2 or the discussion about the correspondence with the flooding reports.

Author Response

Rafea Al-Suhili

Cheila Cullen

Reza Khanbilvardi

June 18, 2019

Re: An Urban Flash Flood Alert Tool for Megacities – Application for Manhattan, New York City, U.S.A – Second review

Dear Reviewers,

We would like to thank you for all your valuable contributions in the review of this document. Please refer to the table below for detailed information.

Reviewer I

General

We have taken your suggestions on making the map methodology more clear. We have added examples in the description of the methodology, added numerical examples and change the flow chart describing the process. We hope that these steps make the whole process more clear.

Specific

Reviewer 2 Report

I am happy with all the corrections made to my original specific comments.

Regarding my general comments - denoted Point 1 - Point 4 by the authors:

The clarification/Changes made with regards to Point 3 and Point 4 are ok.

Regarding the Points 1 and 2, I do not find that the authors have answered properly or maybe they have misunderstood my Point. However, I feel confident that most readers will see the weak Points by themselves and then make their own judgment of the overall merit of the model presented.

Thus I feel that this paper is now ready for publication.

Author Response

Rafea Al-Suhili

Cheila Cullen

Reza Khanbilvardi

June 18, 2019

Re: An Urban Flash Flood Alert Tool for Megacities – Application for Manhattan, New York City, U.S.A – Second review

Dear Reviewers,

We would like to thank you for all your valuable contributions in the review of this document. Please refer to the table below for detailed information.

Reviewer II

General

We appreciate the reviewer’s valuable comments. We have made some changes to the paper that hopefully help clarify the reviewer’s concerns.

Specific