Reducing Flood Risk in Changing Environments: Optimal Location and Sizing of Stormwater Tanks Considering Climate Change

Round 1

Reviewer 1 Report

The paper presents the results of the project “Urban Drainage and Climate Change”, focusing on the analysis of an adaptation measure to climate change to be implemented within urban drainage systems, namely the use of stormwater tanks. The paper is divided into three main sections: the first presents the expected effect of climate change on the local rainfall regime in terms of updated Intensity-Duration-Frequency curves, that will be used as the input for flood modelling; the second sections presents two optimization strategies differing both for the objective functions and for the solution algorithm; the third section presents and discusses the results.

Generally speaking, I enjoyed reading the paper: it is clear, well-written, concise and scientifically sound. However, I have some concerns about how the paper is conceived. My main concern is that I cannot clearly understand the positioning of the paper in the research background. In other words, it is not clear what is the novelty introduced by the authors, maybe as part as the project (could the Authors give details about the project and its relevance?), and what is used as state-of-the-art methodology. This difficulty is also increased by the fact that there are not enough references, so it is not clear whether references are missing or they do not exist at all. I tend to the first option because this research field is prominent and the related literature is quite vast. In my opinion, this is a problem because the topic, the methodology and the results are interesting but they are not promoted enough, since their added value remains unnoticed.

Below you can find specific issues where I experienced this concern.

Climate Change

Nowadays, literature about climate change and its effects on the local extreme rainfall regime is very wide, but there is no reference at all in this section, except for an outdated IPCC Assessment Report (the fourth). Here, the Authors refer to RCP scenarios which are related to GHG concentrations, not emissions; this is a very delicate point because it entails a change of perspective from the source (emissions) to the destination of pollution (the atmosphere). The IPCC report defining RCP scenarios is the fifth Assessment Report, dated 2014. This, I feel, is the reference that should replace the outdated one. Secondly, there are many research papers investigating the problem of how to update IDF curves including the effect of climate change (see Padulano et al. 2019 for literature and an example) because it is not a straightforward operation with any consolidated methodology. The Authors propose a methodology referenced by a local Conference (by the way, is this reference produced within the project or is it previous research?): I am not saying it is not worthy, but since the reference is not accessible or widespread, at least the methodology should be positioned in related current literature. Does this method to integrate climate change into IDF curves build on a widespread approach? For example, estimating the effect of climate change on extreme daily rainfall and, successively, retrieving sub-daily duration events assuming their relationship with the daily duration (estimated with historic data) is still valid, is not widespread, but it is not uncommon (again, see Padulano et al. 2019). Moreover, the Authors do not perform any bias correction, but decide to rely on the models providing the most correct representation of reality. Again, this should be justified.

Optimization algorithms

Here the lack of references is very visible. For both the algorithms, it is not clear whether the optimization strategies (and related solutors) are conceived by the Authors for this research or they are inherited by some other research by some other authors. For example, OptiTank and OptiSu are barely mentioned, they are not given enough importance as project products. Are they available or commercialized? Are there other examples of integrating some optimizer in SWMM by means of the Toolkit? (See for example Del Giudice & Padulano 2016).

On the other hand, it is somehow hinted (page 9, line 330-332) that SWMM Toolkit is a product of your project, which is not in my understanding, as also shown by Figure 4. In turn, Figure 4 is not the “operating scheme of SWMM toolkit” because it does not explain how SWMM Toolkit works, but how it interfaces with your products (the optimizer), whereas this whole architecture is your product.

As concerns the optimization strategy, in my opinion the explanation of the equations is a little careless. For example, I can see the same name for the parameters in equation 12 and 13. Moreover, in this section the methodology and case study information are a little mixed. I suggest being theoretical here, and then explaining in the Case Study section which are the parameters that are already known and which will be calibrated. This also occurs in the Results section, where, for example, it is stated that an interview was performed with the local authority to understand where flood is expected to be critical and where it is particularly suitable to place stormwater tanks. Again, this is part of the Case study section.

Finally, the Authors propose three approaches (section 2.3.2). I see that the first is your research but what about the other two? They are not mentioned anywhere in the following sections. If this is part of previous or separate analysis, or if these results are not presented nor discussed here, I suggest removal.

Results

It would be interesting here to do a little bit more of a discussion. Did you expect those differences between the two algorithms? What are the technical consequences of relying on each of them?

 

References

Del Giudice, Giuseppe, and Roberta Padulano. Sensitivity analysis and calibration of a rainfall-runoff model with the combined use of EPA-SWMM and genetic algorithm. Acta Geophysica 64.5 (2016): 1755-1778.

Padulano, Roberta, Alfredo Reder, and Guido Rianna. An ensemble approach for the analysis of extreme rainfall under climate change in Naples (Italy). Hydrological Processes 33.14 (2019): 2020-2036.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors have prepared a paper that should be of interest to readers of Water.  However, this reviewer found the English difficult to read.  As this detracted from the clarity of the message, a review by a native English speaker would enhance the paper.

While the approach proposed in the paper is explained clearly, there are a number of implicit assumptions that the authors should discuss for clarity purposes.  These are:

1 - There appears to be an assumption that the return period of a flood level (or flow) is directly related to the causative rainfall.

2 - Furthermore, the addition of catchment storage does not change this relationship.

Both of these assumptions are dubious.  While the probability neutral assumption may be valid for the tail of the distribution (extreme events), it is unlikely to be true for events of the magnitude that the drainage system would be designed to manage - typically 1 in 5 to 20-year events.  Secondly, the addition of catchment storage changes the attenuation and lag of an event - increasing the lag and attenuation - thereby increasing the likelihood that an event with a different duration will now result in the critical level.  These changes need to be considered as the drainage system will be designed on a risk basis requiring the prediction of a level with a given probability, not the level arising from rainfall with a given probability.

The introduction of RTC in an urban drainage system warrants comment.  One issue precluding its use in the past has been the need for power during the event, and the high likelihood of power failure during the events where the RTC is required.  While the neglect of this aspect is reasonable, the lack of acknowledgment of this neglect needs rectification.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This manuscript evaluates the imapact of Climate Change on future precipitation and then use the resultant IDF curves to determine the optimal location and sizing of storage units through the application of heuristics such as Simulated Annealing and Pseudo-Genetic Algorithms. From an applied perspective, this work is beneficial for flood risk management as determining the location of stormwater tanks can provide additional information that whether the current facilities are sufficient and a better than to renovate all the pipes that compose the drainage system. However, from a scientific perspective, I did not find any clear novelty or additional message in this manuscript. the Optimal location and sizing of storage units in a drainage system has been widely discussed in the literature. In terms of methodology, the SWMM method used in this manuscript is the repetitive work of Maria C.Cunha, (2016) (https://www.sciencedirect.com/science/ article/abs/pii/S1364815216301517). The only two differences between this manuscript and the work by Cunha, (2016) is that here the authors add Annealing and Pseudo-Genetic Algorithms and use a different case study. Additionally, Stormwater Tanks is not a new topic in the literature. Considering the fact that this manuscript has give more details regarding climat change, it may be acceptable to ignore the lack of novelty. However, to compensate for this problem, the manuscript should be more attractive using a better language and more exotic illustrations. Overall, I think this manuscript is of interest to the Journal of water due to its applied contribution to flood inundation Management.

Comment #01: Why did RCP2.6 be not adopted? try to justify by this sentence : RCP2.6 was not considered in the current selection despite to be the least likely and the mitigation effort implied by this RCP, is unfeasible in the current circumstances, because it needs a sustained global CO2 mitigation rate of around 3% per year, not a likely prospect, at least in the near future.

Comment #02: There are furthersome major issues that need to be addressed before publication can be considered, is to use all 21 GCMs models projections in the analysis.

Comment #03: delete this sentence from the manuscript 'an initial fitness evaluation using the correlation coefficient in order to determine which of them represent better the historical series of precipitation in the study area'. For more information see the methodology used in this paper (https://link.springer.com/article/10.1007/s00382-018-4128-0)

Comment #04: The research question and the progress made in this research compared to previous research should be highlighted in the introduction, as well as throughout the document.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I think that the paper has significantly improved with respect to the initial version. I just feel that an extensive language review is still required, especially in the new paragraphs where there are a lot of language issues (missing verbs, repeated words, singular/plural and third person issues).

Author Response

Thanks a lot to our reviewer 1 for all his/her comments. We did an edition to our English Languaje and style.

Reviewer 3 Report

I think this manuscript is of interest to the Journal of water due to its applied contribution to flood inundation Management. Good job.

Author Response

Thanks a lot to our reviewer 3 for all his/her comments on our paper. We did a new edition of our English languaje and style.