the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Mar 2025
Nationwide Application of a Rapid Tsunami Inundation Model to New Zealand: Assessing Infrastructure and Land Exposure
Abstract. This study presents a comprehensive nationwide tsunami inundation assessment of New Zealand. It utilises wave amplitude data from a recent probabilistic tsunami hazard assessment (covering seven return periods at the 50th and 84th percentiles) and a computationally efficient tsunami inundation model. This modelling approach enhances the inundation accuracy of previous inundation exposure assessments, which have been limited by inconsistent and simplified inundation techniques. This study produces comprehensive nationwide inundation maps for 14 return periods along the entire New Zealand coastline. These inundation maps have been integrated with land cover and infrastructure asset data to assess New Zealand's potential exposure and identify the most vulnerable locations. The findings reveal that while commonly studied cities like Christchurch, Gisborne, and Napier are indeed highly vulnerable, significant exposure is also evident in many provincial areas such as the Buller, Westland and Thames-Coromandel Districts. In the Thames-Coromandel District, over 13,400 buildings are exposed to inundation at the 500-year return period tsunami amplitudes, representing 26.8 % of the buildings in the district. Considering the land classification, the 'Low Vegetation' land cover group is the most exposed in every region, while determining the exposure of the 'Built-Up Area' is a good indicator of overall asset exposure for a region. The inundation model’s efficiency, simplicity, and accuracy make it well-suited for international application, particularly in regions with similar topographic, land use, and hazard data. This versatility, combined with its speed, supports its application in further large-scale probabilistic hazard assessments and smaller, targeted scenario modelling, as well as in uncertainty analyses across diverse geographic and hazard contexts.
- Preprint
(24603 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2024-3724', Anonymous Referee #1, 02 Apr 2025
-
RC2: 'Comment on egusphere-2024-3724', Anonymous Referee #2, 25 Sep 2026
The paper “Nationwide Application of a Rapid Tsunami Inundation Model to New Zealand: Assessing Infrastructure and Land Exposure” by Kimpton et al., provides a novel alternative to hydrodynamic inundation modelling of tsunami. Overall, the paper is well written and provides a lot of detailed analysis on the approach that is undertaken by the authors. However, I have some reservations about the content that is provided in the paper and the lack of comparison to other approaches and models which have previously been peer reviewed and published. I think there is potential in the approach, but additional analysis is needed to provide support.
Specific Issues:
- There is no model validation provided by the authors to compare their approach to other approaches that have previously been peer reviewed and validated. I can find no evidence that this approach has been compared to hydrodynamic models. This is specifically important to areas with complex coastlines, as the details of how the tsunami behaves in bays/inlets will not be captured by a single value. However, the way the results are presented it makes it appear that these tsunami inundation values can be used by end-users. This is misleading. Before these results can be used/published they need to compared to other methods that have been validated, any differences explained, and any limitations clearly stated.
- In the discussion there is a lack of critical analysis of the results, and while the advantages are stated there is not discussion about potential limitations or future work. These points are critical for the reader to make informed decisions about the results and if they can be reasonably applied to other regions or making advanced decisions. Further critical discussion is needed to support the ideas presented here so that the reader understands not just the advantages of this study.
- Several points have been stated in the Conclusions that have not been analysed or supported in the rest of the paper. For example, on line 557 “The modelling methods improves the accuracy of previous inundation modelling methods used in such attempts”, however there is no evidence in the previous sections which support this statement because these results are never explicitly compared to other approaches. Again, at the start of Line 567, this statement has not been fully supported with any analysis throughout the other sections and finally at the end of line 274 to 575 comparisons are needed to support these comments.
Issues per Section:
- Throughout please change “tsunamis” to “tsunami”.
- Watch the placement of figures in the text – some are appearing well before they are referenced in the text.
- In some figure captions you write 50p – please write this out fully (50th Percentile) so that the reader is not having to search what this means.
- Abstract:
- Line 5: you mention in line 2 you are looking at 7 return periods, but then in line 5 you mention 14 return periods. This also occurs several times throughout the paper. This is confusing and at all points you mention 14 return periods this should be changed to clarify that it is 7 return periods and 2 percentiles.
- Line 13: you state particularly “in regions with similar topographic, land use, and hazard data”. Please clarify if this is other similar to NZ or what you mean by this interpretation.
- Introduction:
- Line 18: Please clarify how you are defining local, region and distant source tsunamis. Are you classifying based on distance to NZ or travel time to NZ?
- Lines 44 – 47: These two sentences need have supporting references.
- Line 47: “Tsunami hazard assessments that have been performed” – clarify if these are to the coast or inundation assessments.
- Line 55: States that “internationally do not consider inundation due to the computational cost of running hydrodynamic…” there are also other reasons why like the lack of underlying data to generate DEM’s and uncertainties in underlying hazard models. This needs to be carefully explained as well.
- Line 79: Clarify if the two studies are for NZ or elsewhere. Also, for Paulik et al. (2020a) I don’t know which reference in the list this is referring to because there is no 2020a.
- Line 96: there needs to be a reference for COMCOT.
- Line 97: clarify what an F1 score is.
- Model Overview:
- How do you account for earthquake deformation in your model. Some to the coast tsunami models don’t account for deformation in the simulations so that the coastline remains fixed for analysis. As you are then taking this value through to inundation, which is strongly controlled by deformation, how is this accounted for.
- Why have you not included the Chatham Islands in any of the work?
- Line 167: Provide a reference for where to find the MHWS values for different parts of NZ.
- Lines 201 – 203: Reference is needed for the sentence about the Darcy friction factor.
- Figure 2: panels (e) and (f) look the same. Change the colour scheme to show that they are different and/or zoom into one specific segment to show that these plots are different. There also needs to be a reference figure to show readers unfamiliar with NZ where this is located.
- Results:
- Throughout the results sections you often talk about specific places (like in Figure 9), but there is no main reference location for the reader to refer to. Provide a map which highlights all the regions which are specifically mentioned so that the reader can reference for clarity.
- The discussion about Manawatu-Whanganui comes up multiple times throughout the results section. Have you considered that this is the only region to span the east and west coast of the lower North Island and how this might influence the analysis?
- Line 381: three of the coastlines that you mention here are extremely complex and as you increase the return period the single wave height value at the coast is going to be picking up larger tsunami heights on coastal cliffs. You need to acknowledge this, and state how is this considered in the model as this is likely to be skewing the results.
- Figure 7: Split (a) into two – airports one and hospitals another. There does not seem to be a reason why these are considered tother when the others are split. You also need to clarify in the caption what is being plotted. There are more airports/hospitals in NZ than are plotted in the figure and it is unknown to the reader why some are not included. State: if asset is not exposed at any return period, then it has not been plotted.
- Figure 8 and 9: it is difficult for the reader to distinguish between the 50th and 84th percentile colours. There need to be changed so that it is easier to visualise. Either change the colours or out black outlines to one of the sets of data.
- Discussion and Conclusions:
- See comments under “Specific Issues”.
Citation: https://doi.org/10.5194/egusphere-2024-3724-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 5,227 | 2,447 | 181 | 7,855 | 169 | 257 |
- HTML: 5,227
- PDF: 2,447
- XML: 181
- Total: 7,855
- BibTeX: 169
- EndNote: 257
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This is a well-written paper focusing on the Nationwide Application of a Rapid Tsunami Inundation Model to New Zealand emphasizing on the Assessment of Infrastructure and Land Exposure. The methodology and the data sets used are well described, and the results are reasonable. Certainly this is a new contribution that deserves publication. However, before its final acceptance, it would be useful to extend the discussion on the possible influence on results of the several model inputs and more specifically of the Topographic Data, Land Cover, Shoreline Tsunami Amplitude and Built Environment Assets. The authors may feel free to organize such a more detailed discussion, but some ideas can be inspired from relevant approaches/findings published by Behrens et al. (2021), a review paper that is already cited by the authors.