Understanding flow characteristics from tsunami deposits at Odaka, Joban coast, using a DNN inverse model

Mitra, Rimali; Naruse, Hajime; Abe, Tomoya

doi:https://doi.org/10.5194/egusphere-2023-369

Preprints

https://doi.org/10.5194/egusphere-2023-369

Preprints

24 Apr 2023

| 24 Apr 2023

Understanding flow characteristics from tsunami deposits at Odaka, Joban coast, using a DNN inverse model

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Abstract. The 2011 Tohoku-oki tsunami inundated the Joban coastal area in the Odaka region of Minamisoma City, up to 2,818 m from the shoreline. In this study, the flow characteristics of the tsunami were reconstructed from deposits using the DNN (deep neural network) inverse model, suggesting that the tsunami inundation occurred in the Froude-supercritical condition. The DNN inverse model effectively estimated the tsunami flow parameters in the Odaka region, including the maximum inundation distance, flow velocity, maximum flow depth, and sediment concentration. Despite having a few topographical instabilities that caused the flow height to fluctuate greatly, the reconstructed maximum flow depth and flow velocity were reasonable and close to the values reported in the field observations. The reconstructed data around the Odaka region were characterized by an extremely high velocity (12.1 m/s). This study suggests that the large fluctuation of flow depths at the Joban coast compared with the stable flow depths at the Sendai plain can be explained by the inundation in the supercritical flow condition.

Received: 01 Mar 2023 – Discussion started: 24 Apr 2023

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 3741 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (3741 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

08 Feb 2024

Understanding flow characteristics from tsunami deposits at Odaka, Joban Coast, using a deep neural network (DNN) inverse model

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Nat. Hazards Earth Syst. Sci., 24, 429–444, https://doi.org/10.5194/nhess-24-429-2024,https://doi.org/10.5194/nhess-24-429-2024, 2024

Short summary

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-369', Anonymous Referee #1, 01 Jun 2023

The authors applied their machine learning approach of tsunami deposit inversion (Mitra et al., 2020a,b) to the data from Odaka region in Fukushima Prefecture, where tsunami flow characteristics were different from that in Sendai Plain, Miyagi Prefecture. The inversion analysis yielded an estimated flow speed of 12.1 m/s and flow depth of 2.4 m, which suggest a supercritical (Fr > 1) flow during tsunami inundation. The authors noted that the results are successful, by comparing with some known flow speeds estimated from field records. Validatation of the inverse model based on field records (deposit data and flow measurements) will improve applicability of the method and will improve our understandings on the flow characteristics during inundaiton of large-scale tsunamis.
My main question is related to the assumption of the inverse modeling and interpretation of the results. In this inverse modeling, as explained in the methodology section, flow velocity is assumed to remain constant (in time and space?). If this can be interpreted as averaged velocity over time or space, comparison to the observed values should be made carefully. In particular, I think comparison to observed local maximum velocity is not appropriate.
Iijima et al. (2021) propvided temporal and spatial changes in flow speed estimated based on forward tsunami inundation modeling along a nearby transect in Odaka region. I suggest careful reviewing and comparison of the findings from the paper.
Please find the attached supplement PDF for specific comments.

Citation: https://doi.org/10.5194/egusphere-2023-369-RC1
- CC1: 'Reply on RC1', Rimali Mitra, 29 Jun 2023
  
  Publisher’s note: this comment is a copy of AC1 and its content was therefore removed.
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-CC1
- AC1: 'Reply on RC1', Hajime Naruse, 16 Jul 2023
  
  Publisher’s note: this comment was edited on 17 July 2023. The following text is not identical to the original comment, but the adjustments were minor without effect on the scientific meaning.
  Dear Reviewer 1,
  Thank you for your thorough review of our paper. We appreciate your time and valuable feedback. In this response, we aim to address each of your major comments.
  Firstly, we would like to address the major comment concerning the flow velocity. We recognize the importance of comparing the maximum velocity reported by Ijima et al. (2021), and a review of their study and a detailed comparison with our results will be described in a revised manuscript to be submitted. It is important to note that our approach, assumptions, and model formulation differ significantly from the study mentioned. Ijima et al. (2021) primarily used the forward model with the assumption of the fault model for producing the initial waveform for their simulation, while the depositional information was not considered in their simulation. In contrast, our study employed an inverse model incorporating both tsunami hydrodynamic and sediment transport simulations. The essential difference is that their study estimates the hydrological conditions in the study area based on macroscopic information for the entire Tohoku region. In contrast, our study characterizes tsunamis based on the evidence available in the study area (i.e., tsunami deposits). Based on this difference, we intend to make a detailed comparison of the results of the two studies.
  Regarding the comparison with observed values, we understand the difference between average hydraulic conditions and maximum velocities, and we will fully consider this difference in our discussions. We acknowledge a range of reported values (10-17 m/s) from the surrounding areas. Unfortunately, we do not accurately represent the maximum observed value. Therefore, we relied on the reported ranges without specific observed data for the area under study. Based on this information, we concluded that our simulated flow velocity of 12.1±0.4 m/s falls within the reported range of values from the area.
  We hope this clarifies our position and justifies our conclusions regarding the flow velocity. We thank you once again for your thoughtful review and constructive comments. We have carefully considered your feedback and will incorporate it into the revised version of our paper. We intend to address each comment in the response letter thoroughly. Among the comments made in the PDF file, we also emphasize the importance of describing the sedimentological characteristics of tsunami deposits. In the revised manuscript, detailed columnar sections of the tsunami deposits, grain size analysis results, and outcrop photographs will be added as a supplement.
  Please do not hesitate to reach out if you have any further questions or require additional clarification. We appreciate your time and consideration.
  Regards,
  Rimali Mitra, Hajime Naruse, and Tomoya Abe
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-AC1
RC2:
'Comment on egusphere-2023-369', Masaki Yamada, 20 Jun 2023
This paper reconstructs the tsunami flow characteristics using a DNN inverse model based on the tsunami deposit survey data at the Joban Coast south of the Sendai Plain. I am not a specialist in the details of the research methodology, but I can say that the paper has a clear objective and presents important results. It is particularly important that the inverse analysis suggested the unique tsunami velocity in this region.

Major comments
Figure 9 compares measured and calculated inundation depths. Shouldn't the measured inundation depths be normalized by the elevation of each location? The elevations at which the inundation depths were measured appear to vary (Figure 1). The inundation depths should differ if the elevation is different, although Figure 9 compares them only with the distance. The measured data at the Sendai Plain is on the estimated line (Figure 9b), but isn't this related to the fact that the elevation of the plain is almost constant? It is also noted that the estimated result matches the average of the measured inundation depth data, but this average does not seem to be meaningful. Since the number of measurement points is small, the average value will change significantly as the number of measured points changes.

In the discussion section (5.2), the authors mention the sedimentary structure of the 2004 Indian Ocean tsunami and the flow conditions inferred from it. Were there any differences in the sedimentary structure of tsunami deposits between Odaka and Sendai? Please describe any differences in the sedimentary structure if possible. Many sedimentary data are available, not only data from this study.

Minor comments and corrections are noted in the PDF file. Since most of them are not critical, there is no need to respond to all of them in a response letter, but I hope they will be helpful. If you have any questions, please do not hesitate to contact me.

Masaki Yamada
Citation: https://doi.org/10.5194/egusphere-2023-369-RC2
- CC2: 'Reply on RC2', Rimali Mitra, 29 Jun 2023
  
  Publisher’s note: this comment is a copy of AC2 and its content was therefore removed.
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-CC2
- AC2: 'Reply on RC2', Hajime Naruse, 16 Jul 2023
  
  Publisher’s note: this comment was edited on 17 July 2023. The following text is not identical to the original comment, but the adjustments were minor without effect on the scientific meaning.
  Dear Yamada San,
  Thank you for your thorough review of our manuscript. We appreciate the time and effort you have invested in providing valuable feedback. In this response, we aim to address each of your major comments.
  Regarding the major comment concerning inundation depth and elevation, we have carefully considered your suggestion that addressing this issue may not be necessary. We acknowledge that the information in the PDF about supercritical flows and their in-phase relationship with the topography is relevant. To further improve clarity and eliminate any potential ambiguity, we will include an additional figure in the revised version of the manuscript. This figure will complement Figure 9 and specifically illustrate the correlation between measured watermark elevation and the distance from the shoreline. We are confident that this inclusion will enhance the comprehension of our findings and effectively address your concerns.
  Moving on to the second comment regarding sedimentary structure, we acknowledge that there are differences in sedimentary structure between the Odaka and Sendai areas. Specifically, the tsunami deposits from the Odaka area are characterized by thick mud drapes at their tops, while the Sendai area exhibits deposits with a lower mud content. We appreciate your observation and agree that providing more detailed information on these variations in sedimentary structure is important. In the response letter, we will include additional explanations and discussions highlighting the differences in sedimentary characteristics between the two areas.
  We sincerely appreciate you bringing these points to our attention. We will carefully consider and address each comment in the response letter, ensuring that our revised version comprehensively covers the raised concerns. If you have any further suggestions or questions, please do not hesitate to reach out to us. We greatly value your expertise and input.
  Regards,
  Rimali Mitra, Hajime Naruse, and Tomoya Abe
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-369', Anonymous Referee #1, 01 Jun 2023

The authors applied their machine learning approach of tsunami deposit inversion (Mitra et al., 2020a,b) to the data from Odaka region in Fukushima Prefecture, where tsunami flow characteristics were different from that in Sendai Plain, Miyagi Prefecture. The inversion analysis yielded an estimated flow speed of 12.1 m/s and flow depth of 2.4 m, which suggest a supercritical (Fr > 1) flow during tsunami inundation. The authors noted that the results are successful, by comparing with some known flow speeds estimated from field records. Validatation of the inverse model based on field records (deposit data and flow measurements) will improve applicability of the method and will improve our understandings on the flow characteristics during inundaiton of large-scale tsunamis.
My main question is related to the assumption of the inverse modeling and interpretation of the results. In this inverse modeling, as explained in the methodology section, flow velocity is assumed to remain constant (in time and space?). If this can be interpreted as averaged velocity over time or space, comparison to the observed values should be made carefully. In particular, I think comparison to observed local maximum velocity is not appropriate.
Iijima et al. (2021) propvided temporal and spatial changes in flow speed estimated based on forward tsunami inundation modeling along a nearby transect in Odaka region. I suggest careful reviewing and comparison of the findings from the paper.
Please find the attached supplement PDF for specific comments.

Citation: https://doi.org/10.5194/egusphere-2023-369-RC1
- CC1: 'Reply on RC1', Rimali Mitra, 29 Jun 2023
  
  Publisher’s note: this comment is a copy of AC1 and its content was therefore removed.
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-CC1
- AC1: 'Reply on RC1', Hajime Naruse, 16 Jul 2023
  
  Publisher’s note: this comment was edited on 17 July 2023. The following text is not identical to the original comment, but the adjustments were minor without effect on the scientific meaning.
  Dear Reviewer 1,
  Thank you for your thorough review of our paper. We appreciate your time and valuable feedback. In this response, we aim to address each of your major comments.
  Firstly, we would like to address the major comment concerning the flow velocity. We recognize the importance of comparing the maximum velocity reported by Ijima et al. (2021), and a review of their study and a detailed comparison with our results will be described in a revised manuscript to be submitted. It is important to note that our approach, assumptions, and model formulation differ significantly from the study mentioned. Ijima et al. (2021) primarily used the forward model with the assumption of the fault model for producing the initial waveform for their simulation, while the depositional information was not considered in their simulation. In contrast, our study employed an inverse model incorporating both tsunami hydrodynamic and sediment transport simulations. The essential difference is that their study estimates the hydrological conditions in the study area based on macroscopic information for the entire Tohoku region. In contrast, our study characterizes tsunamis based on the evidence available in the study area (i.e., tsunami deposits). Based on this difference, we intend to make a detailed comparison of the results of the two studies.
  Regarding the comparison with observed values, we understand the difference between average hydraulic conditions and maximum velocities, and we will fully consider this difference in our discussions. We acknowledge a range of reported values (10-17 m/s) from the surrounding areas. Unfortunately, we do not accurately represent the maximum observed value. Therefore, we relied on the reported ranges without specific observed data for the area under study. Based on this information, we concluded that our simulated flow velocity of 12.1±0.4 m/s falls within the reported range of values from the area.
  We hope this clarifies our position and justifies our conclusions regarding the flow velocity. We thank you once again for your thoughtful review and constructive comments. We have carefully considered your feedback and will incorporate it into the revised version of our paper. We intend to address each comment in the response letter thoroughly. Among the comments made in the PDF file, we also emphasize the importance of describing the sedimentological characteristics of tsunami deposits. In the revised manuscript, detailed columnar sections of the tsunami deposits, grain size analysis results, and outcrop photographs will be added as a supplement.
  Please do not hesitate to reach out if you have any further questions or require additional clarification. We appreciate your time and consideration.
  Regards,
  Rimali Mitra, Hajime Naruse, and Tomoya Abe
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-AC1
RC2:
'Comment on egusphere-2023-369', Masaki Yamada, 20 Jun 2023
This paper reconstructs the tsunami flow characteristics using a DNN inverse model based on the tsunami deposit survey data at the Joban Coast south of the Sendai Plain. I am not a specialist in the details of the research methodology, but I can say that the paper has a clear objective and presents important results. It is particularly important that the inverse analysis suggested the unique tsunami velocity in this region.

Major comments
Figure 9 compares measured and calculated inundation depths. Shouldn't the measured inundation depths be normalized by the elevation of each location? The elevations at which the inundation depths were measured appear to vary (Figure 1). The inundation depths should differ if the elevation is different, although Figure 9 compares them only with the distance. The measured data at the Sendai Plain is on the estimated line (Figure 9b), but isn't this related to the fact that the elevation of the plain is almost constant? It is also noted that the estimated result matches the average of the measured inundation depth data, but this average does not seem to be meaningful. Since the number of measurement points is small, the average value will change significantly as the number of measured points changes.

In the discussion section (5.2), the authors mention the sedimentary structure of the 2004 Indian Ocean tsunami and the flow conditions inferred from it. Were there any differences in the sedimentary structure of tsunami deposits between Odaka and Sendai? Please describe any differences in the sedimentary structure if possible. Many sedimentary data are available, not only data from this study.

Minor comments and corrections are noted in the PDF file. Since most of them are not critical, there is no need to respond to all of them in a response letter, but I hope they will be helpful. If you have any questions, please do not hesitate to contact me.

Masaki Yamada
Citation: https://doi.org/10.5194/egusphere-2023-369-RC2
- CC2: 'Reply on RC2', Rimali Mitra, 29 Jun 2023
  
  Publisher’s note: this comment is a copy of AC2 and its content was therefore removed.
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-CC2
- AC2: 'Reply on RC2', Hajime Naruse, 16 Jul 2023
  
  Publisher’s note: this comment was edited on 17 July 2023. The following text is not identical to the original comment, but the adjustments were minor without effect on the scientific meaning.
  Dear Yamada San,
  Thank you for your thorough review of our manuscript. We appreciate the time and effort you have invested in providing valuable feedback. In this response, we aim to address each of your major comments.
  Regarding the major comment concerning inundation depth and elevation, we have carefully considered your suggestion that addressing this issue may not be necessary. We acknowledge that the information in the PDF about supercritical flows and their in-phase relationship with the topography is relevant. To further improve clarity and eliminate any potential ambiguity, we will include an additional figure in the revised version of the manuscript. This figure will complement Figure 9 and specifically illustrate the correlation between measured watermark elevation and the distance from the shoreline. We are confident that this inclusion will enhance the comprehension of our findings and effectively address your concerns.
  Moving on to the second comment regarding sedimentary structure, we acknowledge that there are differences in sedimentary structure between the Odaka and Sendai areas. Specifically, the tsunami deposits from the Odaka area are characterized by thick mud drapes at their tops, while the Sendai area exhibits deposits with a lower mud content. We appreciate your observation and agree that providing more detailed information on these variations in sedimentary structure is important. In the response letter, we will include additional explanations and discussions highlighting the differences in sedimentary characteristics between the two areas.
  We sincerely appreciate you bringing these points to our attention. We will carefully consider and address each comment in the response letter, ensuring that our revised version comprehensively covers the raised concerns. If you have any further suggestions or questions, please do not hesitate to reach out to us. We greatly value your expertise and input.
  Regards,
  Rimali Mitra, Hajime Naruse, and Tomoya Abe
  
  Citation: https://doi.org/10.5194/egusphere-2023-369-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Publish subject to minor revisions (review by editor) (30 Sep 2023) by Rachid Omira

AR by Hajime Naruse on behalf of the Authors (27 Nov 2023) Author's response Author's tracked changes Manuscript

ED: Publish as is (20 Dec 2023) by Rachid Omira

AR by Hajime Naruse on behalf of the Authors (22 Dec 2023)

Journal article(s) based on this preprint

08 Feb 2024

Understanding flow characteristics from tsunami deposits at Odaka, Joban Coast, using a deep neural network (DNN) inverse model

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Nat. Hazards Earth Syst. Sci., 24, 429–444, https://doi.org/10.5194/nhess-24-429-2024,https://doi.org/10.5194/nhess-24-429-2024, 2024

Short summary

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Data sets

Odaka_DNN_inverse_2011_tsunami: DNN inverse model for 2011 Tohoku-oki tsunami at Odaka, Japan Rimali Mitra, Hajime Naruse, and Tomoya Abe https://doi.org/10.5281/zenodo.4764153

Model code and software

DNN inverse model for 2011 Tohoku-oki tsunami at Odaka, Japan Rimali Mitra, Hajime Naruse, and Tomoya Abe https://doi.org/10.5281/zenodo.4764153

Rimali Mitra, Hajime Naruse, and Tomoya Abe

Viewed

Total article views: 380 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
239	112	29	380	16	16

HTML: 239
PDF: 112
XML: 29
Total: 380
BibTeX: 16
EndNote: 16

Views and downloads (calculated since 24 Apr 2023)

Month	HTML	PDF	XML	Total
Apr 2023	55	16	3	74
May 2023	30	8	1	39
Jun 2023	41	16	8	65
Jul 2023	57	21	6	84
Aug 2023	15	6	0	21
Sep 2023	9	12	2	23
Oct 2023	12	10	2	24
Nov 2023	6	2	2	10
Dec 2023	7	6	1	14
Jan 2024	7	12	2	21
Feb 2024	3	2	5
Mar 2024	0
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0
Sep 2024	0

Cumulative views and downloads (calculated since 24 Apr 2023)

Month	HTML	PDF	XML	Total
Apr 2023	55	16	3	74
May 2023	30	8	1	39
Jun 2023	41	16	8	65
Jul 2023	57	21	6	84
Aug 2023	15	6	0	21
Sep 2023	9	12	2	23
Oct 2023	12	10	2	24
Nov 2023	6	2	2	10
Dec 2023	7	6	1	14
Jan 2024	7	12	2	21
Feb 2024	3	2	5
Mar 2024	0
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0
Sep 2024	0

Viewed (geographical distribution)

Total article views: 378 (including HTML, PDF, and XML) Thereof 378 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 06 Sep 2024

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (3741 KB)
Metadata XML

Short summary

This study estimates the behavior of the 2011 Tohoku-oki tsunami from its deposit distributed in the Joban coastal area. In this study, the flow characteristics of the tsunami were reconstructed using the DNN (deep neural network) inverse model, suggesting that the tsunami inundation occurred in the very high-velocity condition.


Total:	0
HTML:	0
PDF:	0
XML:	0