Climate change impacts on flood risk and asset damages within mapped 100-year floodplains of the contiguous United States

Wobus, Cameron; Gutmann, Ethan; Jones, Russell; Rissing, Matthew; Mizukami, Naoki; Lorie, Mark; Mahoney, Hardee; Wood, Andrew W.; Mills, David; Martinich, Jeremy

doi:https://doi.org/10.5194/nhess-17-2199-2017

Articles | Volume 17, issue 12

https://doi.org/10.5194/nhess-17-2199-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/nhess-17-2199-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 17, issue 12

Research article

|

08 Dec 2017

Research article |

| 08 Dec 2017

Climate change impacts on flood risk and asset damages within mapped 100-year floodplains of the contiguous United States

Cameron Wobus, Ethan Gutmann, Russell Jones, Matthew Rissing, Naoki Mizukami, Mark Lorie, Hardee Mahoney, Andrew W. Wood, David Mills, and Jeremy Martinich

Download

Final revised paper (published on 08 Dec 2017)
Supplement to the final revised paper
Preprint (discussion started on 24 Apr 2017)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Method estimates flood-related costs of GHG emissions across an entire continent', Anonymous Referee #1, 22 May 2017
RC2: 'An interesting paper combining several modelling approaches to deliver damage estimates from flooding with climate change', Simon Dixon, 23 May 2017
AC1: 'Author responses to reviewer comments', Cameron Wobus, 10 Jul 2017

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by Editor and Referees) (17 Jul 2017) by Paolo Tarolli

AR by Cameron Wobus on behalf of the Authors (11 Aug 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (13 Aug 2017) by Paolo Tarolli

RR by Simon Dixon (06 Sep 2017)

RR by Anonymous Referee #3 (02 Oct 2017)

Suggestions for revision or reasons for rejection

The proposed work deals with the examination of future changes in the frequency of floods across contiguous United States. A methodology is proposed to assess future flood impacts in terms of monetary damages to assets. The topic is of broad interest and the methodology described involves novel elements. While the authors admittedly have applied and described a comprehensive framework for addressing the issue of future flood risk and associated impacts, I have a number of concerns that need to be addressed in my opinion to improve the clarity of the work and make this contribution more meaningful. Below, I provide a list of comments.

1. First of all the title is somewhat misleading because it is focused on “100yr” flood. In fact the authors investigate frequency of floods exceeding a threshold considered to be the 100yr flood from the early 21st century. I think that removing the “100yr” would make the title more representative of the work.
2. A number of important aspects are simply mentioned without providing any evidence that support them. For example:
a. L15-16,p1: “Our model generates early 21st century flood damages that reasonably approximate….”. You need to provide some evidence on this
b. L4-5,p5: “…there were no systematic differences in results in the annual maximum time series…”. How did you evaluate this? Please specify and report associated metrics/results.
c. L23,p5: “we used a series of steps to calculate the depth of flooding” .This is an important methodological step. It should be clearly discussed and demonstrated in the manuscript (rather than provided as supplement).
3. A critical element of this work is the identification of the 1% AEP from the 2001-2020 period and its connection to the 100yr flood maps generated by FEMA. To be consistent in your analysis, and be able to connect the 100yr flood map with the 100yr estimate of your simulations you should use the same period that FEMA used to estimate the 100yr flood. In that case you will at least sample the same climate period. Otherwise you may be introducing a systematic error in your analysis. Grouping results from all models does not resolve this, given that all models are realizations of the same climate period. I understand that the issue of “data jump”after year 2000 prevents you from using the before 2000 simulations, however the point of uncertainty in the 100yr flood estimated by the last 20yrs could be demonstrated by observations at selected locations (e.g. USGS stations).
4. A major concern of the validity of the results is that there is no sort of evaluation of the performance of the model used. Can you provide any indication that the VIC-based hydrologic simulations can reproduce a realistic flood response at least for the period 2000-2017, where observations exist? Otherwise in the absence of any evaluation procedure findings are highly questionable.
5. Although the connection of changes in flood frequency with monetary damages is a novel element and generally welcome, I feel that it is the weak link of the study and should be minimized to a section that offers some indication rather than be the central part of the work. The authors themselves in their conclusions acknowledge a number of very important sources of uncertainty and factors that have not been taken into account in the damage estimates. Acknowledging is good but not enough to justify why we should then consider meaningful actual numbers of future flood damages when population expansion for the next 80 yrs is not even considered.
I would suggest the authors to restructure the work and have as central point of the work the analysis of changes in flood frequency under different climate scenarios. An effort for a more in-depth explanation of the patterns in Fig.6 would be mostly welcome. The association of predicted increases with potential damages should remain as a secondary element in the paper but analysis should be mostly focused on the relative differences between the two emission scenarios rather than an effort of providing actual numbers. Translation of future flood risk increases into damages is indeed a highly complex topic. I believe that the authors can benefit from a more thorough literature review on this as well. See for example

Liu, Jing, et al. "Future property damage from flooding: sensitivities to economy and climate change." Climatic change132.4 (2015): 741-749.

Bubeck, P., et al. "How reliable are projections of future flood damage?." Natural Hazards and Earth System Sciences 11.12 (2011): 3293.

Hide

ED: Reconsider after major revisions (further review by Editor and Referees) (02 Oct 2017) by Paolo Tarolli

AR by Cameron Wobus on behalf of the Authors (06 Oct 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (11 Oct 2017) by Paolo Tarolli

RR by Anonymous Referee #3 (24 Oct 2017)

Suggestions for revision or reasons for rejection

I have reviewed the revised manuscript and the author’s response. Indeed, I had accidentally reviewed the original version of the manuscript and the revised version is certainly improved. As I stated before, I consider the methodological framework presented in this work novel and as such, I wish to see the paper published. I have included some further comments below:

1. Regarding my previous comment on the connection between the modeled 1% AEP and the 100yr flood map generated by FEMA. I would like to add few more comments as a response to author’s comments. The authors stated that
“We disagree with the assessment of the Referee , for two reasons.. First,, we note that the core of our methodology is to examine changes in the frequency of modeled 1% AEP events.. These changes are based on model to model comparisons through the 21 st century. While we recognize that the time period of our “baseline”” does not coincide with the time period during which the 100 - year floodplains were mapped, this is likely to make our results conservative if one assumes that extreme hydrologic events have already been increasing through the late 20th century.”

I am fine with your methodology for assessing changes in modeled 1% AEP events, as you mention. However, your assessment of changes in frequency is based on a threshold (the one you identified at the 1%AEP from the 2001-2020 period), which means that your results are also threshold-dependent. Considering a longer period could (certainly would) lead to a different threshold that would lead to potential differences in the results. In fact, this relates to your discussion on the sampling uncertainty which you have identified to be in the order of 5-20% that you chose not to propagate in your assessment. However, presenting some evidence on the sensitivity of the results on the threshold (even just for the min/max possible values) would be beneficial for the readers. Again, due to the complexity of the problem you are addressing it is understandable that all the different sources of uncertainty cannot be realistically quantified. However, since you have already estimated a range of uncertainty for the 1% AEP I would strongly encourage you to show how this affects the results.

The authors' second argument stated

“Second, because we are using a delta method, we neither expect nor require our model to perfectly simulate 100 - year events in any one location. The mapped 100 - year floodplains are used only to calculate the damages associated with 1% AEP events in each stream reach, and our method does not require that we demonstrate the perfect correspondence between modeled and observed hydrologic extremes. We believe that the discussion of the model performance in Supplemental Information File #11 and in Mizukami et al., (2017), both of which were added to the revised manuscript, provide sufficient documentation of the model’s performance to justify its use under the delta method framework we applied. Furthermore and as described above, Figure 7 demonstrates that despite the uncertainties described above and in Section 2.4, the simulated damages in the baseline model run very closely mimic observed flood damages over the late 20th century.”

I understand the point of the authors and my point was not to require that the simulated flows matched perfectly the observed. My point connects to the previous point on the impact of uncertainty in the estimation of 1% AEP. Let’s assume that FEMA calculated the 100yr flood based on N observations available from the simulations. Take the N simulated events and calculate the 100yr flood (Q100N). If we take a sub-sample n<N of those observations and recalculate the 100yr flood we will end up with a different estimate Q100n (uncertainty due to sample size). The number of times the future (simulated) flows exceed Q100N and Q100n will be different and thus the associated estimates of future damages would be different.
Again, in my opinion, this point could be addressed (at least partially) by providing some indications on the sensitivity of the results to the estimation of the baseline 1% AEP.

2. My second comment relates to page 5, L5 where the authors define as “flood” the annual maximum flow value that exceeds the baseline 1% AEP. Why are you considering only the annual max? There can be also other events within the year that exceed the baseline 1% AEP. Accounting for these as well will have a significant impact in the future change of “flood” frequency and associated damages. Please clarify/justify your choice.

Hide

ED: Publish subject to minor revisions (review by editor) (24 Oct 2017) by Paolo Tarolli

AR by Cameron Wobus on behalf of the Authors (27 Oct 2017) Author's response Manuscript

ED: Publish as is (27 Oct 2017) by Paolo Tarolli

AR by Cameron Wobus on behalf of the Authors (30 Oct 2017) Author's response Manuscript

Download

Article (12178 KB)
Full-text XML

Short summary

We linked modeled changes in the frequency of historical 100-year flood events to a national inventory of built assets within mapped floodplains of the United States. This allowed us to project changes in inland flooding damages nationwide under two alternative greenhouse gas (GHG) emissions scenarios. Our results suggest that more aggressive GHG reductions could reduce the projected monetary damages from inland flooding, potentially saving billions of dollars annually by the end of the century.