the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Integrating social, economic, and environmental risk into flood management of aging dam infrastructure by combining cost-benefit and multi-criteria decision analyses
Abstract. Management planning for aging dam infrastructure is typically conducted through the lens of a traditional cost-benefit analysis, in which flood characteristics are related to implementation costs while lacking endogenous consideration of environmental risks (i.e., pollutant dispersion, habitat disruption) and social impacts (i.e., vulnerability, community buy-in, hazard resiliency). To address this gap, we integrate cost-benefit ratios into a spatial multi-criterion decision analysis framework that amalgamates a suite of social and environmental criteria with stakeholder-defined weights and inundation outputs from standard flood control modelling. We use this framework to assess the costs and trade-offs for eight (8) alternative mitigation strategies associated with the Addicks and Barker Reservoir System in Houston, Texas, USA under extreme rainfall conditions. This case study illustrates how the total effectiveness of flood management scenarios may shift when flood modelling outputs are combined with spatially distributed environmental and social risks. We merge quantitative and qualitative data for high-risk decision-making, thereby fostering stakeholder collaboration amongst conflicting goals.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(2979 KB)
-
Supplement
(712 KB)
-
This preprint has been withdrawn.
- Preprint
(2979 KB) - Metadata XML
-
Supplement
(712 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
RC1: 'Comment on nhess-2021-144: an interesting approach', Mariana de Brito, 24 Jun 2021
General comments
In this interesting manuscript, the authors compare different flood risk management options by considering both MCDA and CBA. The topic is interesting and meaningful. Furthermore, the methodology applied is robust and innovative, and the final outputs of good quality. The graphs and Figures produced are in general easy to follow and summarize well the outcomes. However, I have some major concerns regarding the text readability. The entire paper, especially the methods section, could be reduced by 20-25% without losing relevant information. Another criticism is regarding the discussion section, which should be improved to add the limitations of the study. Hence, I recommend moderate revisions.
Specific comments:
- A main problem is that the reasoning for model assumptions is not clearly stated (e.g. for the weighting of the criteria, selection of criteria, as well as the points listed in lines 187 to 216). It is not clear how many stakeholders participated, for instance. For more specific comments see below.
- The methodology section reads too long and it is a mix of literature review and methods. It should be shortened. It is also not linear and difficult to follow. The authors first describe the weighting procedure and only then detail the criteria used. After that, the weighting procedure is explained in detail. I understand that section 3.2 now tries to add an overview, but it is confusing because many details are not explained. Hence, I would suggest to follow a linear description and incorporate the lines 230 to 255 in the other subsections. This way you will avoid repetition.
- Similarly, the results section is wordy. The last paragraphs should be included in the discussion. Furthermore, limitations and future work section should be added. Bellow you can find some suggestions on this (e.g. lack of sensitivity and uncertaintly analyses).
- Line 40: I would say that MCDM is also a tool for traditional flood management. What is innovative of your research is combining both. I would reformulate this paragraph, stating the advantages/disadvantages of each approach, and how their integrated use can provide better answers for an adequate flood risk management. Here you also need to show previous literature that has also followed a similar approach. Try to find flood-related articles that combine both approaches. If they do not exist, you can also list this as an innovation from your paper. These could be relevant articles, but you have to check if they fit the scope as they are not for FRM. https://www.sciencedirect.com/science/article/pii/S2352146515002197 https://www.sciencedirect.com/science/article/pii/S221204161630420X https://link.springer.com/article/10.1007/s40070-019-00098-1
- Line 43: “considered secondary in management frameworks” I disagree that this information is considered secondary. There are hundreds of flood vulnerability studies that show otherwise.
- Line 85-86: This should be in the methods section.
- Line 94-95: This is an important gap you are helping to fill. This should be mentioned in the introduction section
- Table 1: It is not clear how you extrapolated the cost estimate. I suggest adding a new column to the table where you can summarize the impacts you describe in section 2.2.2.
- Line 163-166: this is also a gap. I suggest mentioning it in the introduction.
- Line 195: how did you arrive at this number of 10.000 houses? Please detail it more. The same goes for all other quantitative assumptions
- Line 233: How exactly did you determine the relevant criteria? Was there a systematic procedure in the literature review you conducted? This is a gap that should be listed in the discussion section as different scientists would choose different criteria leading to completely different outcomes. Also, what is this “local knowledge”? Did you consult experts in the field? Or it was based on the author’s opinion. This should be clarified.
- Line 239: You need to explain how these weights were defined. How many stakeholders were involved? How were they selected? Where do they work and what is their expertise? If they were based only on the opinion of the authors, this should be stated. Furthermore, this should be added as a limitation in the discussion section.
- Line 256: What do you mean by “exploratory geospatial review”?
- Line 260: By consolidated, do you mean you aggregated several criteria into one? If yes, which and how? You should be clearer on the method used to combine these criteria.
- Line 270: Doesn’t the SoVI includes already population density? Wouldn’t there be then a redundancy? Ideally you should conduct a PCA or other data reduction techniques. See this article, it may be helpful for the discussion section: https://nhess.copernicus.org/articles/21/1513/2021/
- Line 272: It was assumed based on what? On the information provided by Klotzbach et al?
- Line 355: The validation against the stream gauge heights is not mentioned in the methods section. Also, why have you conducted validation for some alternatives and not for some? The validation procedure should be described in the methods.
Technical corrections
- Line 9: As a non-native speaker, I had to google what “community buy-in” means. It may be my ignorance, but perhaps you could just frame it as “community acceptance and support” or something similar? Still regarding to this, I do not understand why buy-in and resilience are social impacts. For me they are actually the opposite. I would keep "vulnerability" and use other examples here.
- Line 12: remove the (8)
- Line 84: Remove the word “qualitative”
- Line 154: please write “third reservoir (A2, Table 1), so the readers can understand that this is one of the 8 alternatives.
- Table 2: you should add the spatial resolution of these data.
- Line 255: What do you understand by “comprehensive risk dataset” and “ancillary risk datasets”? The difference between the two should be introduced.
- Line 290: I am not sure, but perhaps you can make a table with this information? Right now the text is too dense and difficult to have an overview of the many assumptions.
- Figure 2: the figures have a very low resolution. For the final version please use a pdf or similar graphs.
- Line 300: the information regarding the weighting should come before.
- Line 301: remove “general”
- Line 301: How exactly were these “discussions”. How many stakeholders? How did you achieved consensus between these stakeholders? Was one weighting derived for each participant and then you made an average? The procedure should be clarified.
- Line 305-307: If I understood correctly, you have not done this. Hence, it should be removed from the methods section. I would add this to the discussion, saying what future research could do/limitations in your study.
- Line 307-308: This should be in the discussion section.
- Line 310-312: this is literature review, not methods… I would remove all together.
- Figure 3: the color of the high risk easement should be changed as it is now the same as the color used for the study area border.
- Line 339: the normalization is mentioned 2 times in this paragraph.
- Line 354: Please provide this information in a table format. This way it is easier for the readers to compare the different alternatives.
- Line 354 to 364: The text reads too long and should be cut.
- Table 3: Please add to the legend of the figure what Ci, Ai, CBi, etc mean. It is easier for the readers not to need to search back in the text.
- Figure 5: I like the figure as it summarizes the outcomes and is easy to understand. However, I do not understand why some alternatives are in orange and some in blue. Please add this information to the legend.
- Figure 6: very important figure, but difficult to read because is twisted. Please use portrait orientation. Also, add the legend to the y axis. What do high and low z scores represent? Low z scores represent low social risk, for instance?
- Line 469-486: This is discussion, not results.
- Line 506: Please add a section called conclusion and add the text from here there.
Citation: https://doi.org/10.5194/nhess-2021-144-RC1 - AC1: 'Reply on RC1', Hanadi Rifai, 03 Jul 2021
-
RC2: 'Comment on nhess-2021-144', Anonymous Referee #2, 06 Aug 2021
The manuscript presents a MCA analysis of different flood mitigation options for a complex basin in Texas, USA. Topic is interesting and fits the scope of the journal. The idea itself is (obviously) not new in the literature, nor relevant methodological issues are here developed, but the case study is interesting and potentially deserves for publication, once a series of points are modified and/or clarified.
MAJOR POINTS
Equation (1) – Composite Risk.
Factors R (Environmental, Social) are weighted averages of evaluation scores. Concept is clear, but we have no information on how such scores are given, we only know about their general meaning (sources of contaminants, soil erodibility, medical facilities …). Lines 256-261 provide a long list of items to be considered in the risk evaluation. However, it is not clear which of them has been really considered in the environmental and social criteria, and how.
In different words: what do factors “ej” exactly represent? Are they binary quantities (e.g.: presence of a source of contaminant in a cell), extensive quantities (e.g.: length of inundated road in a cell), intensive quantities (percentage of flood insurances among residents)? How the scoring 0-100 is attributed to each factor for each cell? Do cells have a uniform extension?
In particular “Stream samples were obtained from field campaigns following Hurricane Harvey, which were used in this study to validate the areas of environmental burdens associated with contamination in local waterways.” I have not understood how such data were used to define values for the environmental factors in each cell.
Finally, I strongly suggest providing a short description of the SoVI (Social Vulnerability Index) and variables involved in the index.
Equation (2) – Impact Functions.
1) what are the “zonal statistics for the composite risk and the modeled inundation area of each alternative”?
2) what are the “zones”? the inundated areas for each scenario? I understand that “ai” are the corresponding inundated surfaces, is this correct? Or do they also comprehend the areas impacted by the “ancillary risk” (see below)?
3) what is the summation index in equation (2)?
At the end of the story, I cannot understand IF. If “Rbar” is an average (I guess, spatial average) over the zone and “a” is the area of the zone, than Rbar is constant over the zone and IF=Rbar+ancillary risk. But this has no sense, therefore I conclude that I was not able to understand equation (2).
Section 3.2.2 - Ancillary risk
I see conceptual inconsistencies here. Soil use / buyouts due to mitigation measures are not risks, are deterministic impacts; they have 100% probability along the lifetime of the mitigation measure, and they last for all such time. On the other hand, the damage from a flood scenario has a probability of exceedance less than 100% in the given period. What is here called “ancillary risk” is not a risk at all, and should not simply be added to the flood risk by assigning certain values of R-scoring for the areas impacted by the measures, as here done.
On the other hand, the extra-flooding expected along the Cypress Creek is an additional/ancillary risk, and it is correct to handle it as such. Why is such area simply accounted by a RS=100 scoring? It should be added to the flooded area, and evaluated with respect to CB and IFs.
Section 3.2.3 – Weight determination
This is THE key point of any MCA, making the difference between an exercise and a relevant field case. I cannot really understand how the authors determined the weights. I read about “discussions with Houston-area flood risk stakeholders, including governmental entities, interest groups, and specialized consulting firms”; the description continues with principles derived from the literature (lines 302-307); authors conclude by saying that “As participatory modelling is inherently qualitative, individual criterion weights will differ according to local conditions and stakeholder goals”. All this is true, but what did they really do? From lines 308-312 I may understand that weights in table 2 are somehow just a reasonable proposal, not yet evaluated by stakeholders?
It is also very important to clarify that weights are not general, but linked to the definition (and consequent variation ranges) for the indicators to be weighted. All this information set should be part of the discussion (with stakeholders) devoted to fixing the weights.
Again, weights are the key point. Selection criteria should be discussed. If such criteria are not robust, an extensive sensitivity analysis should be provided in order to give real value to the MCA.
Section 3.3 – CBA
Benefits are evaluated in terms of a fixed damage/hectare, without considering any specification for the soil use / exposed elements (residential, agriculture, industrial, …). Please, add some consideration about the accuracy and robustness of the used fixed value (=0.478 M$/he).
Section 3.4 – Integrated CBA + MCA
“Since the unique indicators contained different units of measurement ($/hectares, 0-100 risk) we used z-score normalization to transform the values to equivalent scales”: this is not fully true. CB is non-dimensional (ranging 0-1 if benefits exceed costs, but later I understand expressed as 0-100); R and IF are also non-dimensional (ranging 0-100). Why was then the z-normalization used? This point should be clarified.
Moreover, the choice of the uniform weighting of the three indicators (CB, IFE, IFS) is an important part of the weight determination: what is the rationale under such choice? was it discussed with stakeholders? Was any sensitivity analysis performed?
I have a wider doubt here, about the consistency of T, presumably as a consequence of not having understood how factors IFs (or Rs) are formed. Let us take two mitigation alternatives (A1, A2), with different costs C, different benefits B, different impacts IF which, for simplicity, we here limit to the number of people affected by the flood (we call it N). Let us imagine that C1>C2 but also B1>B2 (A1 reduces the flooded area more than A2) so that CB1=CB2. Thus, the CB component of T will be equal for the two cases, thus suggesting that T is an indicator of efficiency (intensiveness) rather than of efficacy (extensiveness). Let us imagine that N2>N1 (as A1 reduces the flooded are more than A2): what would happen to IF1 and IF2? Is their contribution to T consistent with an indicator of efficiency? What for all the other components of IFE and IFS?
MINOR POINTS
Line 25: “Hurricane Harvey”: add date of the event.
Line 150: “Wealthy and middle-income populations face higher risks when located outside of federally-designated floodplains where flood insurance is voluntary”. I cannot understand the reason of this higher risk.
Analysed mitigation strategies (tab. 1) are those proposed in the USACE 2020 report. In particular, costs of the mitigation actions are derived from such reports. The reader would expect that also hydrological / hydraulic scenario are derived from the report but, apparently, it is not so, as modelling of such scenarios is discussed along the manuscript. Authors should clarify this point and discuss consistency of scenarios with associated costs. In particular, for scenario A7 I read (line 400) that the authors used a wider channel extension with respect to the USACE proposal: what about costs, were they modified accordingly?
Line 235: “To standardize the point and polyline feature classes into spatially varied datasets, the Euclidean Distance method was applied. Euclidean distances convert feature layers into gridded datasets by assigning a value to each cell that indicates the distance of that cell to the nearest criterion, thus standardizing space and creating hotspots in multi-criteria decision making”. Meaning and impact of this procedure is not clear to me.
Line 271: “The spatial risk associated with flood insurance was derived from national flood hazard zones and a repository of damaged structures in the community. It was assumed that residents within the FEMA 1% and 0.1% flood zones carried flood insurance, while 20% of all other residents had purchased voluntary insurance”. Please, provide comment about soundness of such assumption.
Table 3: I cannot reconstruct values for CB. Let us take alternative A3 for Addicks as an example. Reduction of flooded area is 466 he; when multiplied by 0.478 M$/he we obtain a damage reduction B = 223 M$; cost is here C = 5000 M$ (please, make indication of units coherent in tables 1 and 3) with a consequent CB = C/B = 2200% … not comparable with values in table 3 and with common sense. This clearly means that there is something I have not understood (I honestly tried) … please, clarify.
Figure 6: I cannot understand spider graphs (c) and (d); I expect the same values as in plots (a) and (b) to be represented, but there is no coherence.
SOME SUGGESTIONS ON THE ORGANIZATION OF THE MANUSCRIPT
Section 2.1: this relatively long paragraph appears as a continuation/specification of the literature review provided in the Introduction rather than a description in the methodology here used. Moreover, the scheme for “integrated flood management decision-making” in fig. 1 is here presented but not really used (or, perhaps, not clearly explained). Consider better focusing all information within lines 18-104 with respect to the specific aim of the paper (case study).
Along the manuscript discussions are alternated between the hydrological / hydraulic scenarios (sections 2.2.1, 3.1, 4.1) and the impact/damage/cost scenarios (remaining sections). I suggest to re-organize the material so that the sections for the two groups are presented all together, and the flux of information should become more coherent. This would also avoid some repetitions.
Lines 314-319: here a general discussion about principles of a CBA is provided. However, here a simplified version of benefits evaluation is used. The general discussion could be omitted or moved to some introductory section.
Citation: https://doi.org/10.5194/nhess-2021-144-RC2 - AC2: 'Reply to RC2', Hanadi Rifai, 28 Aug 2021
Interactive discussion
Status: closed
-
RC1: 'Comment on nhess-2021-144: an interesting approach', Mariana de Brito, 24 Jun 2021
General comments
In this interesting manuscript, the authors compare different flood risk management options by considering both MCDA and CBA. The topic is interesting and meaningful. Furthermore, the methodology applied is robust and innovative, and the final outputs of good quality. The graphs and Figures produced are in general easy to follow and summarize well the outcomes. However, I have some major concerns regarding the text readability. The entire paper, especially the methods section, could be reduced by 20-25% without losing relevant information. Another criticism is regarding the discussion section, which should be improved to add the limitations of the study. Hence, I recommend moderate revisions.
Specific comments:
- A main problem is that the reasoning for model assumptions is not clearly stated (e.g. for the weighting of the criteria, selection of criteria, as well as the points listed in lines 187 to 216). It is not clear how many stakeholders participated, for instance. For more specific comments see below.
- The methodology section reads too long and it is a mix of literature review and methods. It should be shortened. It is also not linear and difficult to follow. The authors first describe the weighting procedure and only then detail the criteria used. After that, the weighting procedure is explained in detail. I understand that section 3.2 now tries to add an overview, but it is confusing because many details are not explained. Hence, I would suggest to follow a linear description and incorporate the lines 230 to 255 in the other subsections. This way you will avoid repetition.
- Similarly, the results section is wordy. The last paragraphs should be included in the discussion. Furthermore, limitations and future work section should be added. Bellow you can find some suggestions on this (e.g. lack of sensitivity and uncertaintly analyses).
- Line 40: I would say that MCDM is also a tool for traditional flood management. What is innovative of your research is combining both. I would reformulate this paragraph, stating the advantages/disadvantages of each approach, and how their integrated use can provide better answers for an adequate flood risk management. Here you also need to show previous literature that has also followed a similar approach. Try to find flood-related articles that combine both approaches. If they do not exist, you can also list this as an innovation from your paper. These could be relevant articles, but you have to check if they fit the scope as they are not for FRM. https://www.sciencedirect.com/science/article/pii/S2352146515002197 https://www.sciencedirect.com/science/article/pii/S221204161630420X https://link.springer.com/article/10.1007/s40070-019-00098-1
- Line 43: “considered secondary in management frameworks” I disagree that this information is considered secondary. There are hundreds of flood vulnerability studies that show otherwise.
- Line 85-86: This should be in the methods section.
- Line 94-95: This is an important gap you are helping to fill. This should be mentioned in the introduction section
- Table 1: It is not clear how you extrapolated the cost estimate. I suggest adding a new column to the table where you can summarize the impacts you describe in section 2.2.2.
- Line 163-166: this is also a gap. I suggest mentioning it in the introduction.
- Line 195: how did you arrive at this number of 10.000 houses? Please detail it more. The same goes for all other quantitative assumptions
- Line 233: How exactly did you determine the relevant criteria? Was there a systematic procedure in the literature review you conducted? This is a gap that should be listed in the discussion section as different scientists would choose different criteria leading to completely different outcomes. Also, what is this “local knowledge”? Did you consult experts in the field? Or it was based on the author’s opinion. This should be clarified.
- Line 239: You need to explain how these weights were defined. How many stakeholders were involved? How were they selected? Where do they work and what is their expertise? If they were based only on the opinion of the authors, this should be stated. Furthermore, this should be added as a limitation in the discussion section.
- Line 256: What do you mean by “exploratory geospatial review”?
- Line 260: By consolidated, do you mean you aggregated several criteria into one? If yes, which and how? You should be clearer on the method used to combine these criteria.
- Line 270: Doesn’t the SoVI includes already population density? Wouldn’t there be then a redundancy? Ideally you should conduct a PCA or other data reduction techniques. See this article, it may be helpful for the discussion section: https://nhess.copernicus.org/articles/21/1513/2021/
- Line 272: It was assumed based on what? On the information provided by Klotzbach et al?
- Line 355: The validation against the stream gauge heights is not mentioned in the methods section. Also, why have you conducted validation for some alternatives and not for some? The validation procedure should be described in the methods.
Technical corrections
- Line 9: As a non-native speaker, I had to google what “community buy-in” means. It may be my ignorance, but perhaps you could just frame it as “community acceptance and support” or something similar? Still regarding to this, I do not understand why buy-in and resilience are social impacts. For me they are actually the opposite. I would keep "vulnerability" and use other examples here.
- Line 12: remove the (8)
- Line 84: Remove the word “qualitative”
- Line 154: please write “third reservoir (A2, Table 1), so the readers can understand that this is one of the 8 alternatives.
- Table 2: you should add the spatial resolution of these data.
- Line 255: What do you understand by “comprehensive risk dataset” and “ancillary risk datasets”? The difference between the two should be introduced.
- Line 290: I am not sure, but perhaps you can make a table with this information? Right now the text is too dense and difficult to have an overview of the many assumptions.
- Figure 2: the figures have a very low resolution. For the final version please use a pdf or similar graphs.
- Line 300: the information regarding the weighting should come before.
- Line 301: remove “general”
- Line 301: How exactly were these “discussions”. How many stakeholders? How did you achieved consensus between these stakeholders? Was one weighting derived for each participant and then you made an average? The procedure should be clarified.
- Line 305-307: If I understood correctly, you have not done this. Hence, it should be removed from the methods section. I would add this to the discussion, saying what future research could do/limitations in your study.
- Line 307-308: This should be in the discussion section.
- Line 310-312: this is literature review, not methods… I would remove all together.
- Figure 3: the color of the high risk easement should be changed as it is now the same as the color used for the study area border.
- Line 339: the normalization is mentioned 2 times in this paragraph.
- Line 354: Please provide this information in a table format. This way it is easier for the readers to compare the different alternatives.
- Line 354 to 364: The text reads too long and should be cut.
- Table 3: Please add to the legend of the figure what Ci, Ai, CBi, etc mean. It is easier for the readers not to need to search back in the text.
- Figure 5: I like the figure as it summarizes the outcomes and is easy to understand. However, I do not understand why some alternatives are in orange and some in blue. Please add this information to the legend.
- Figure 6: very important figure, but difficult to read because is twisted. Please use portrait orientation. Also, add the legend to the y axis. What do high and low z scores represent? Low z scores represent low social risk, for instance?
- Line 469-486: This is discussion, not results.
- Line 506: Please add a section called conclusion and add the text from here there.
Citation: https://doi.org/10.5194/nhess-2021-144-RC1 - AC1: 'Reply on RC1', Hanadi Rifai, 03 Jul 2021
-
RC2: 'Comment on nhess-2021-144', Anonymous Referee #2, 06 Aug 2021
The manuscript presents a MCA analysis of different flood mitigation options for a complex basin in Texas, USA. Topic is interesting and fits the scope of the journal. The idea itself is (obviously) not new in the literature, nor relevant methodological issues are here developed, but the case study is interesting and potentially deserves for publication, once a series of points are modified and/or clarified.
MAJOR POINTS
Equation (1) – Composite Risk.
Factors R (Environmental, Social) are weighted averages of evaluation scores. Concept is clear, but we have no information on how such scores are given, we only know about their general meaning (sources of contaminants, soil erodibility, medical facilities …). Lines 256-261 provide a long list of items to be considered in the risk evaluation. However, it is not clear which of them has been really considered in the environmental and social criteria, and how.
In different words: what do factors “ej” exactly represent? Are they binary quantities (e.g.: presence of a source of contaminant in a cell), extensive quantities (e.g.: length of inundated road in a cell), intensive quantities (percentage of flood insurances among residents)? How the scoring 0-100 is attributed to each factor for each cell? Do cells have a uniform extension?
In particular “Stream samples were obtained from field campaigns following Hurricane Harvey, which were used in this study to validate the areas of environmental burdens associated with contamination in local waterways.” I have not understood how such data were used to define values for the environmental factors in each cell.
Finally, I strongly suggest providing a short description of the SoVI (Social Vulnerability Index) and variables involved in the index.
Equation (2) – Impact Functions.
1) what are the “zonal statistics for the composite risk and the modeled inundation area of each alternative”?
2) what are the “zones”? the inundated areas for each scenario? I understand that “ai” are the corresponding inundated surfaces, is this correct? Or do they also comprehend the areas impacted by the “ancillary risk” (see below)?
3) what is the summation index in equation (2)?
At the end of the story, I cannot understand IF. If “Rbar” is an average (I guess, spatial average) over the zone and “a” is the area of the zone, than Rbar is constant over the zone and IF=Rbar+ancillary risk. But this has no sense, therefore I conclude that I was not able to understand equation (2).
Section 3.2.2 - Ancillary risk
I see conceptual inconsistencies here. Soil use / buyouts due to mitigation measures are not risks, are deterministic impacts; they have 100% probability along the lifetime of the mitigation measure, and they last for all such time. On the other hand, the damage from a flood scenario has a probability of exceedance less than 100% in the given period. What is here called “ancillary risk” is not a risk at all, and should not simply be added to the flood risk by assigning certain values of R-scoring for the areas impacted by the measures, as here done.
On the other hand, the extra-flooding expected along the Cypress Creek is an additional/ancillary risk, and it is correct to handle it as such. Why is such area simply accounted by a RS=100 scoring? It should be added to the flooded area, and evaluated with respect to CB and IFs.
Section 3.2.3 – Weight determination
This is THE key point of any MCA, making the difference between an exercise and a relevant field case. I cannot really understand how the authors determined the weights. I read about “discussions with Houston-area flood risk stakeholders, including governmental entities, interest groups, and specialized consulting firms”; the description continues with principles derived from the literature (lines 302-307); authors conclude by saying that “As participatory modelling is inherently qualitative, individual criterion weights will differ according to local conditions and stakeholder goals”. All this is true, but what did they really do? From lines 308-312 I may understand that weights in table 2 are somehow just a reasonable proposal, not yet evaluated by stakeholders?
It is also very important to clarify that weights are not general, but linked to the definition (and consequent variation ranges) for the indicators to be weighted. All this information set should be part of the discussion (with stakeholders) devoted to fixing the weights.
Again, weights are the key point. Selection criteria should be discussed. If such criteria are not robust, an extensive sensitivity analysis should be provided in order to give real value to the MCA.
Section 3.3 – CBA
Benefits are evaluated in terms of a fixed damage/hectare, without considering any specification for the soil use / exposed elements (residential, agriculture, industrial, …). Please, add some consideration about the accuracy and robustness of the used fixed value (=0.478 M$/he).
Section 3.4 – Integrated CBA + MCA
“Since the unique indicators contained different units of measurement ($/hectares, 0-100 risk) we used z-score normalization to transform the values to equivalent scales”: this is not fully true. CB is non-dimensional (ranging 0-1 if benefits exceed costs, but later I understand expressed as 0-100); R and IF are also non-dimensional (ranging 0-100). Why was then the z-normalization used? This point should be clarified.
Moreover, the choice of the uniform weighting of the three indicators (CB, IFE, IFS) is an important part of the weight determination: what is the rationale under such choice? was it discussed with stakeholders? Was any sensitivity analysis performed?
I have a wider doubt here, about the consistency of T, presumably as a consequence of not having understood how factors IFs (or Rs) are formed. Let us take two mitigation alternatives (A1, A2), with different costs C, different benefits B, different impacts IF which, for simplicity, we here limit to the number of people affected by the flood (we call it N). Let us imagine that C1>C2 but also B1>B2 (A1 reduces the flooded area more than A2) so that CB1=CB2. Thus, the CB component of T will be equal for the two cases, thus suggesting that T is an indicator of efficiency (intensiveness) rather than of efficacy (extensiveness). Let us imagine that N2>N1 (as A1 reduces the flooded are more than A2): what would happen to IF1 and IF2? Is their contribution to T consistent with an indicator of efficiency? What for all the other components of IFE and IFS?
MINOR POINTS
Line 25: “Hurricane Harvey”: add date of the event.
Line 150: “Wealthy and middle-income populations face higher risks when located outside of federally-designated floodplains where flood insurance is voluntary”. I cannot understand the reason of this higher risk.
Analysed mitigation strategies (tab. 1) are those proposed in the USACE 2020 report. In particular, costs of the mitigation actions are derived from such reports. The reader would expect that also hydrological / hydraulic scenario are derived from the report but, apparently, it is not so, as modelling of such scenarios is discussed along the manuscript. Authors should clarify this point and discuss consistency of scenarios with associated costs. In particular, for scenario A7 I read (line 400) that the authors used a wider channel extension with respect to the USACE proposal: what about costs, were they modified accordingly?
Line 235: “To standardize the point and polyline feature classes into spatially varied datasets, the Euclidean Distance method was applied. Euclidean distances convert feature layers into gridded datasets by assigning a value to each cell that indicates the distance of that cell to the nearest criterion, thus standardizing space and creating hotspots in multi-criteria decision making”. Meaning and impact of this procedure is not clear to me.
Line 271: “The spatial risk associated with flood insurance was derived from national flood hazard zones and a repository of damaged structures in the community. It was assumed that residents within the FEMA 1% and 0.1% flood zones carried flood insurance, while 20% of all other residents had purchased voluntary insurance”. Please, provide comment about soundness of such assumption.
Table 3: I cannot reconstruct values for CB. Let us take alternative A3 for Addicks as an example. Reduction of flooded area is 466 he; when multiplied by 0.478 M$/he we obtain a damage reduction B = 223 M$; cost is here C = 5000 M$ (please, make indication of units coherent in tables 1 and 3) with a consequent CB = C/B = 2200% … not comparable with values in table 3 and with common sense. This clearly means that there is something I have not understood (I honestly tried) … please, clarify.
Figure 6: I cannot understand spider graphs (c) and (d); I expect the same values as in plots (a) and (b) to be represented, but there is no coherence.
SOME SUGGESTIONS ON THE ORGANIZATION OF THE MANUSCRIPT
Section 2.1: this relatively long paragraph appears as a continuation/specification of the literature review provided in the Introduction rather than a description in the methodology here used. Moreover, the scheme for “integrated flood management decision-making” in fig. 1 is here presented but not really used (or, perhaps, not clearly explained). Consider better focusing all information within lines 18-104 with respect to the specific aim of the paper (case study).
Along the manuscript discussions are alternated between the hydrological / hydraulic scenarios (sections 2.2.1, 3.1, 4.1) and the impact/damage/cost scenarios (remaining sections). I suggest to re-organize the material so that the sections for the two groups are presented all together, and the flux of information should become more coherent. This would also avoid some repetitions.
Lines 314-319: here a general discussion about principles of a CBA is provided. However, here a simplified version of benefits evaluation is used. The general discussion could be omitted or moved to some introductory section.
Citation: https://doi.org/10.5194/nhess-2021-144-RC2 - AC2: 'Reply to RC2', Hanadi Rifai, 28 Aug 2021
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
786 | 408 | 36 | 1,230 | 111 | 36 | 45 |
- HTML: 786
- PDF: 408
- XML: 36
- Total: 1,230
- Supplement: 111
- BibTeX: 36
- EndNote: 45
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
Cyndi V. Castro
This preprint has been withdrawn.
- Preprint
(2979 KB) - Metadata XML
-
Supplement
(712 KB) - BibTeX
- EndNote