the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Nearshore Tsunami amplitudes across the Maldives archipelago due to worst case seismic scenarios in the Indian Ocean
- Department of Earth Science and Engineering, Imperial College London, UK
- Department of Earth Science and Engineering, Imperial College London, UK
Abstract. The Maldives face the threat of tsunamis from a multitude of sources. However, the limited availability of critical data, such as bathymetry (a recurrent problem for many island nations), has meant that the impact of these threats has not been studied at an island scale. Studies of tsunami propagation at the island scale but across multiple Atolls is also a challenging task due to the large domain and high resolution required for modelling. Here we use a high resolution bathymetry dataset of the Maldives archipelago, and corresponding high numerical model resolution, to carry out a scenario-based tsunami hazard assessment for the entire Maldives archipelago to investigate the potential impact of plausible far-field tsunamis across the Indian Ocean at the island scale. The results indicate that the bathymetry of the Atolls, which are characterized by very steep boundaries offshore, is extremely efficient in absorbing and redirecting incoming tsunami waves. Results also highlight the importance that local effects have in modulating tsunami amplitude nearshore, including the location of the Atoll in question, the location of a given island within the Atoll, and the distance of that island to the reef, as well as a variety of other factors. We also find that the refraction and diffraction of tsunami waves within individual Atolls contribute to the maximum tsunami amplitude patterns observed across the islands in the Atolls. The findings from this study contribute to a better understanding of tsunamis across complex Atoll systems, and will help decision and policy makers in the Maldives asses the potential impact of tsunamis across individual islands. An online tool is provided which presents users with a simple interface allowing the wider community to browse the simulation results presented here and assess the potential impact of tsunamis at the local scale.
Shuaib Rasheed et al.
Status: final response (author comments only)
-
RC1: 'Comment on nhess-2022-95', Anonymous Referee #1, 27 Apr 2022
This reviewer would like to commend the authors on their work. Efforts associated with quantifying tsunami hazards/vulnerability is a welcome addition and this work represents first of its kind tsunami simulations for the Maldives archipelago. Further, I would particularly like to commend their efforts on providing an online server to access the results, as stated by the authors this will be of great benefit to the wider community.
However, before I can recommend this paper for publication I would encourage the authors to engage with the feedback and comments provided below.
Main Comments
- Aside from the detailed feedback below my main critique of the work is related to the numerical modelling and would recommend some additional efforts in this regard. The main issue is the use of coarse (50m) `fine-resolution’ grids. The authors themselves repeatedly state the necessity of high resolution bathymetry/meshes to capture the complex wave patterns of tsunami waves around and within Atolls. They state that their high resolution mesh has a minimum mesh element size of 50m however in their referenced work [Rasheed, 2021 (a)] it appears that bathymetry data on a ~10m resolution is available. If high resolution information is key to capturing the complex tsunami wave patterns, something which this reviewer agrees with, why have the authors not used a finer resolution mesh? Is there an issue with computational resources? Please expand on this.
- The authors state that the model Thetis can capture wetting/drying using the algorithm described in Eq. 3, however they have chosen a minimum water depth of 0.1m. From this reviewer’s experience this minimum depth is overly conservative. If a higher resolution mesh is used than I would encourage the authors to reduce this value. Otherwise the overtopping of low-lying islands may not be captured accurately and thus the influence on the resultant wave pattern will be missed. Further comparisons to run-up and inundation measurements from the 2004 survey could also be made. It should be noted that despite the recognised absence of additional terms in the non-linear shallow water equations (NSWE) for capturing inundation, numerous NSWE solvers have been validated against inundation and runup tasks, [Macias 2017] is one such example.
Line by Line Comments
Section 1.
Line 25: Rephrase ``These data imply …”
Line 30 - 35: It might be worth mentioning the return periods of some of the findings
Line 38: Please expand on the ``safe island concept” for those who are unaware of it.
Line 41-42: Rephrase ``impacted less and others being impacted more”
Lines 50-55: Please make reference of [Xie 2019], where a tsunami hazard assessment of the Xisha Archipelago was carried out.
Section 2.
Line 102: Typo ``Boxing Day even” should be ``Boxing Day event”
Line 113: Typo ``Table 1” should be ``Table 2”
Line 113: Rephrase ``worst most-likely”
Section 2.2: It would be good to make reference to Figure 1, which showcases the location of the various sources
Section 2.3: It appears that this section is a direct copy of a section in the authors previous works [Rasheed, 2021 (a) and Rasheed, 2021 (b)].
Line 150: Units for kinematic viscosity
Line 152: I would suggest introducing tau and rho here instead of on Line 160. Please also make reference to that fact that the value of n will be discussed in section 2.3.4.
Line 153: Please provide a further explanation for the P^{DG}_1 - P^{DG}_1 term.
Comment: Has the Thetis model in this set up been validated against traditional tsunami benchmark problems? If not I would suggest taking a look at the problems outlined in [https://nctr.pmel.noaa.gov/benchmark/].
Line 166: ``Each simulation was run in a full simulation which spatially extended ..” It is not clear from this sentence how the nested simulations were carried out, please clarify this.
Line 166-167: Reorder the Figure numbers, ``Figures 2 (a) and 1 (a)” and ``Figures 2 (b) and 1 (b)”.
Line 171-173: The authors state that the tidal variation of 1m is ``very small” and can therefore be discounted however in later sections they state ``If combined with a high tide, tsunamis generated from scenario 2a would likely have an impact across locations predicted to have higher amplitudes”. I would agree with the later and state that tidal forcing can play a role in inundation levels. Not including tidal forcing in the study is acceptable but please correct this section and make note of the limitation in section 4.3 (Limitations of the Study and Recommendations for Improvement).
Figure 1: Please increase the size of Figure 1 (b). It is difficult to make out the high resolution features. Please also ensure that all x and y axis are labelled.
Figure 1: Typo in the caption ``13 orth Nilandhe” and ``20. ddu Atoll”.
Figure 2: Please make the subplot (b) larger and label the axis correctly.
Section 2.3.5: How are the full and higher resolution mesh merged? From reading it appears there is a mismatch in mesh resolution at the boundaries, with the full mesh having a resolution of 5km to 7.5km while the nested mesh has a resolution of 10km at the boundaries. Please provide some details on nested procedure.
Section 3.
Figure 3: Please make subplots (a) and (b) larger. There appears to be an artefact of the nesting in the bottom right of both subplots (a) and (b). The bottom right corner exhibits some high wave heights which exhibit a discontinuous drop-off when moving in a north western direction. Is this physical or an artefact of the nesting?
Section 3.1.2: Why have you chosen a Pearson correlation coefficient? Would a RMSE value be more appropriate? Please explain.
Lines 260 -266: This appears to be a one sided comparison. What about the areas of low impact? Do the simulated and observed areas of low impacts also match up? This is an equally interesting comparison.
Figure 4. It might be useful to provide a map showing where these subplots are located in the Maldives. Please provide a wave height scale and label the axis for each subplot.
Figure 6 and 7: Typo ``worst case fault case”
Section 4.
Lines 347-349: Are the refraction, diffraction and reflection patterns repeated across different simulated sources? Please comment on this.
Figure 8: Mark the Atolls (South Nilande, Mulaku and Kolhumadula) in subplot (a). Please re-plot with the wave height coloring centred on 0m. It may be interesting to see reflections etc.
Line 352: Please reference the work of [Reymond, 2012], where the role of reef systems on the amplification of tsunami waves is captured as a site specific amplification parameter.
Lines 353-357: Please rephrase this sentence.
Line 372: Can you please qualify ``with high tsunami flow velocities across the shallow and narrow channels”, as there are no plots explicitly showing this behaviour.
Section 4.2: As stated above I highly commend the authors for providing the online explorer. However the links do not work and I was unable to access the server. Please correct this.
Line 450: The following work should be cited as an additional approach for investigating uncertainties [Giles, 2021]. In that work the uncertainty on the source is propagated to maximum wave heights using cheap statistical emulators.
Section 5.
Line 456-457: The statement ``variability in the resulting tsunami amplitudes due to the development of complex refraction and reflection wave patterns” should be qualified with further results such as those shown in section 4.1
To finish I would like to reiterate my commendation of the authors efforts and appreciate that my main comments listed above may be deemed harsh. However, if the high resolution data is available I would encourage the authors to re simulate at a higher resolution.
References
- Rasheed, 2021 (a): Rasheed, S., Warder, C. S., Plancherel, Y. and Piggott, M. D: “An Improve Gridded Bathymetric Data Set and Tidal Model for the Maldives Archipelago”, Earth and Space Science, 8, 5, 2021
- Macias, 2017: Macias, J., Castro, M. J., Ortega, S., Escalante, C., Gonzalez-Vida, J. M.: “Performance Benchmarking of Tsunami-HySEA Model for NTHMP's Inundation Mapping Activities”, Pure and Applied Geophysics, 8, 3147--3183, 2017
- Xie, 2019: Xie, X., Chen, C., Li, L., Wu, S., Yuen, D. A., Wang, D.:“Tsunami hazard assessment for atoll islands inside the South China Sea: A case study of the Xisha Archipelago”, Physics of the Earth and Planetary Interiors, 290, 2019
- Rasheed, 2021 (b): Rasheed, S., Warder, C. S., Plancherel, Y. and Piggott, M. D: “Response of tidal flow regime and sediment transport in North Male Atoll, Maldives, to coastal modification and sea level rise”, Ocean Sci., 17, 319-334, 2021
- Reymond, 2012: Reymond, D. and Okal, E. A. and H{\'{e}}bert, H. and Bourdet, M.: “Rapid forecast of tsunami wave heights from a database of pre-computed simulations, and application during the 2011 Tohoku tsunami in French Polynesia”, Geophysical Research Letters, 11, 1—6, 2012
- Giles, 2021: Giles, D. and Gopinathan, D. and Guillas, S. and Dias, F.: “Faster Than Real Time Tsunami Warning with Associated Hazard Uncertainties”, Frontiers in Earth Science, 8, 2021
-
AC1: 'Reply on RC1', Shuaib Rasheed, 11 Jun 2022
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2022-95/nhess-2022-95-AC1-supplement.pdf
-
RC2: 'Comment on nhess-2022-95', Joern Behrens, 28 Apr 2022
The manuscript by Rasheed et al. describes the detailed modeling of tsunami hazard to the Maldives caused by credible worst case scenarios from various sourse locations. An emphasis is put to the ability to accurately model the wave behavior in the complex topography/bathymetry of the archipelago.The work is significant, the presentation clear and well structured and the methods mostly up to date. I still recommend some major revisions, because I think the whole potential of this work has not been harvested. However, with relatively little effort, this work could in my view be of even more relevance, so I suggest considering the following.## General comments1. A short discusion on why not a standard tsunami model was used, but a self built non-validated (at least not with the standard tsunami benchmarks according to Synolakis et al., 2008) based on Firedrake. What are the advantages compared to e.g. COMCOT or TsunaCLAW?2. What kind of criteria were used for the diverse decisions made:a. Mesh refinement - is it just proximity to coast?b. removal of islands from the large scale simulation - is it size?3. To me the local resolution mesh sizes seem still rather large. A 5000 m mesh size at the Maledives Atoll coast for the large-scale simulation yields an effective wave lenght representation of 30 km or more (given the linear P1 elements of the DG discretization). Is this a reasonable scale? Additionally, the non-uniform mesh would allow for higher local resolution without much additional effort in terms of added unknowns, since the local area of refinement would cover only fractions of the domain. The same applies to the local simulations, where a 50 m mesh size allows to represent wave lengths of approx. 500 m or a little less that that. With island sizes of only meters in size, I doubt if this is high enough a resolution for quantitatively accurate results. Some sensitivity studies would be helpful in this.4. Since you indicated in the text that you are only considering wave heights at the coast and no inundation, what are the boundary conditions at the coasts then? In Harig et al. (2008) it was found that inundation BC are necessary even if not used to realistically prepresent coastal reflection of waves.5. In order to evaluate the wave build-up it would also be valuable to consider the different wave lengths/periods in comparison to the obstacle size (atoll diameter e.g.) to have a conceptual understanding of this phenomenon. I hypothesize that a singular atoll of a size less than - say - half the deep ocean wave length will be passed by the wave without major harm, given the extremely steep bathymetry. But this would be an interesting topic of diagnostics, analysis and discussion for the different locations and angles of attack.6. You claim that such results are only possible by high resolution bathymetry data and go further to ask for even higher resolution in this respect. But you do not prove that this is really the case. It would be very instructive (and in your case probably easily possible) to actually demonstrate this claim by comparing the effect of diffraction, reflection and deflection in your large-scale and small-scale simulations. For example the results in figure 8, do they differ substantially for your large- and small-scale simulations? If so, I would buy your demand for ever higher resolution ;-) Here I assume that you use the same bathymetry data in your simulations, but that you interpolate to your unstructured mesh and therefore have different discrete bathymetries in your simulations.## Minor comments1. **Lines 205-210**: I don't understand why 2500 m mesh sizes are used at continental shore lines, while only 5000-7500 m are used at the atoll boundaries - isn't this the area of interest, which should have the highest resolution? Additionally, given the capability of non-uniform meshing, a resolution of approx. 250 m along the atoll shore should be much more useful and feasible.2. **Figure 4**: Can you explain why the wave patterns are so different in the two scenarios with respect to locality? It appears from the visuals that scenario a) produces much smoother results than scenario b), while the maximum values are somewhat similar. Have these simulations been run with different meshes?3. **Figure 5**: These figures are very small and hardly distinguishable. Lagoon boundaries are invisible, and details are additionally blurred by the different scales used in the upper and lower row. I would suggest to improve by cropping the areas sideways and lining them besides each others increasing size. Also the same scale (maybe logarithmic) should be used for better comparison.4. **Line 296**: Bengukulu -> Bengkulu.5. **Section 3.2**: You mention several times that hazard levels with 1 m wave height increase with high tide levels. However, you don't mention how high high tide actually is. It would therefore be helpful to learn about the tidal elevation for those places, you mention.6. **Line 340**: Calrseberg Ridge -> Carlsberg Ridge.7. **Line 465**: Link does not work or is incomplete.## ReferencesSynolakis et al. (2008): Validation and verification of tsunami numerical models, Pure appl. geophys. 165:2197-2228, doi:10.1007/s00024-004-0427-y.
-
AC2: 'Reply on RC2', Shuaib Rasheed, 11 Jun 2022
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2022-95/nhess-2022-95-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Shuaib Rasheed, 11 Jun 2022
Shuaib Rasheed et al.
Shuaib Rasheed et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 732 | 181 | 21 | 934 | 10 | 14 |
- HTML: 732
- PDF: 181
- XML: 21
- Total: 934
- BibTeX: 10
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1