Numerical modelling of the evolution of a river reach with a complex morphology to help define future sustainable restoration decisions

Yassine, Rabab; Cassan, Ludovic; Roux, Hélène; Frysou, Olivier; Pérès, François

doi:https://doi.org/10.5194/esurf-2021-91

Preprints

https://doi.org/10.5194/esurf-2021-91

Preprints

26 Jan 2022

| 26 Jan 2022

Status: a revised version of this preprint was accepted for the journal ESurf and is expected to appear here in due course.

Numerical modelling of the evolution of a river reach with a complex morphology to help define future sustainable restoration decisions

Rabab Yassine, Ludovic Cassan, Hélène Roux, Olivier Frysou, and François Pérès

Abstract. The evolution of river morphology is very complicated to predict, especially in the case of mountain and Piedmont rivers with complex morphologies, steep slopes, and heterogeneous grain sizes such as the "Lac des Gaves" (LDG) reach, located within the Gave de Pau river in the Hautes-Pyrénées department, France, has precisely the complex morphological characteristics mentioned above. This reach has gone through severe sediment extractions for over 50 years, leading to the construction of two weirs for riverbed stabilisation. Two large floods resulted in changes in the LDG's hydromorphological characteristics as it went from a single channel river section to a braided river reach. In this study, a 2D hydromorphological model is developed with the TELEMAC-MASCARET system to reproduce the evolution of the channel following a flood that occurred in 2018. The model's validity is assessed by comparing the simulated topographic evolution to the observed one. The results reveal the challenge of choosing well-fitted sediment transport equations and friction laws that would make it possible to reproduce such complex morphology. Even if the exact localisation of the multiple channels forming the braided nature of the LDG was challenging to reproduce, our model could provide reliable volumetric predictions as it reproduces the filling of the LDG correctly. The influence of the two weirs on the river's current and future morphology is also studied. The aim is to provide decision-makers with more reliable predictions to design suitable restoration measures for the LDG reach.

Received: 22 Nov 2021 – Discussion started: 26 Jan 2022

Download & links

Rabab Yassine et al.

Status: closed

RC1:
'Comment on esurf-2021-91', Damien KUSS, 18 Mar 2022

General comments

This article is original and particularly interesting with the use of a 2D hydromorphological numerical model in a torrential context. The task is challenging as the study concerns the modeling of a reach of braided river with wandering flows which are by nature random and are therefore particularly difficult to model in a deterministic way.

From a methodological point of view, the numerical aspects are very well described. On the other hand, the study site and the modeled domain deserve to be better described: length, slopes... A longitudinal profile, encompassing the modeling domain, would make it possible to better understand the problem of deposition during floods in link with the weirs and with the decrease in the longitudinal slope. The influence of the solid volume taken inot account, linked with the slope, could be discussed in the results section.

The hydromorphological modeling is carried out taking into account only bedload (with Meyer Peter and Recking formulas). But you mention, by exploiting the data from dredging, that the fraction of the volume transported by bedload would represent only 8 to 16% of total transport (lines 417 to 417). There is an inconsistency:

• The altitudinal evolutions observed by DoD are compared to the modeled bed evolutions, which only include sediment transport by bedload;

• On the other hand, the modeled transported volume is compared to the total volume observed minus the fine fraction (volume transported by suspension).

The performance of each modeling scenario is evaluated with the BSS score. In a braided river context, the scores obtained are not good (BSS = 0.06). You discuss the representativeness of this metric for such river morphologies. If the metric can be criticized because of the random nature of the wanderings, it would also seem relevant to question the added value of a 2D model compared to a 1D model in such a context. You could precise the configurations where a 2D model is apropriate.

Specific comments

Reference Reisenbücchler et al., 2019 at the end of sentence line 19. What does it refers to ?

Line 19 : “They showed”. “They” refers to Rickenmann et al. or Reisenbuchler et al. ?

Line 38 to 40 :

- a mention could be made of more recent formulas partially based on field data (Recking, 2013 ; Lefort, 2015).

- are bedload transport formulas established in 1d narrow channels directly transposable in 2D models ?

Lines 40-41 : the slope used for the sediment transport calculation at the upstream boundary condition of the model is a key parameter to perform realistic simulations. It must be apprehended by a geomorphological analysis based on the the longitudinal profile.

Line 63 : “The TELEMAC-MASCARET modelling system has been considered well suited to perform 2D morphodynamic simulations on the LDG reach”. Why ?

Line 92-93 : you could add a reference at the end of the sentence.

a longitudinal profile would complete the description and would allow to better understand the effects of the two weirs during the floods.

Lines 101-104 : you could add references concerning the peak discharges.

Figure 3 : you could add the reference for each picture

Lines 164-166 : you could add a reference : Rickenmann and Recking (2011) ?

Lines 185-188 : please define all the terms used in equation 5

Lines 195-208 : note that the recking formula : (a) is mainly base on lab experiments for high shields numbers ; (b) is very sensitive to the choice of the shields mobility parameter. So the avantages of using this formula for intense floods where the shields number exceeds the mobility parameter could be discussed.

Lines 225-226 : it is not clear how you use the dredging data.

Lines 234-235 : “However, the recorded volumes represent both very fine sediments probably transported by suspension and very coarse sediment via bedload transport”. The model used in the study takes into account only the bedload transport, Is it right ?

Lines 269-270 : you therefore make the assumption of no downstream influence.

Line 276 : what do you mean by “instabilities” ?

Lines 276-278 : the slope of the LDG reach is never given. What is slope at the upstream boundary condition ? Is the hypothesis of stable riverbed evolution justified ? What was the observed evolution of the river bed at the upstream condition during the floods ?

Lines 286-287 : you could detail what you mean by numerical and physical parameters.

Figure 6 : how do you transform 2D results into 1D longitdunal profile ?

Lines 326-327 : “To date, numerical models cannot predict channel migration processes that occur in braided rivers. These phenomena are uncertain and random. A modeler should thus not expect the model to predict channel migration accurately during a flood.”. I agree. Butin consequence you should better justify why you have chosen a 2D hydromorphological model for this case study.

Lines 331-332 : is the 2019 LiDAR realigned ?

Figures 10-11 : I regret that the 2016 profile was not drawn with a solid line. it's hard to see the position of the fall. You could also explain how you transform 2D results with a bed level not constant over cross sections in to longitudinal profiles.

Figure 13 : how do you explain the deposits above the max water level ?

Figure 14 : the initial profile is missing. It could be a usell information.

Citation: https://doi.org/10.5194/esurf-2021-91-RC1
- AC1: 'Reply on RC1', Rabab YASSINE, 26 Jun 2022
  
  Dear reviewer,
  
  Thank you very much for this detailed and constructive review which will be very useful in improving the quality of our article. Please find attached the detailed answers. The reviewer's comments are shown in bold, and some modifications of the manuscript are emphasized in blue.
  Kind regards.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC1
RC2:
'Comment on esurf-2021-91', Clément Misset, 31 Mar 2022

The comment was uploaded in the form of a supplement: https://esurf.copernicus.org/preprints/esurf-2021-91/esurf-2021-91-RC2-supplement.pdf

Citation: https://doi.org/10.5194/esurf-2021-91-RC2
- AC2: 'Reply on RC2', Rabab YASSINE, 28 Jun 2022
  
  Dear reviewer,
  
  Thank you very much for this detailed and constructive review which will help us improve the quality of our article. Please find attached the detailed answers. The reviewer's comments are shown in bold, and some modifications of the manuscript are emphasized in blue.
  Kind regards.
  The authors.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC2
RC3:
'Comment on esurf-2021-91', Saraswati Thapa, 03 Apr 2022

Review of “Numerical modelling of the evolution of a river reach with a complex morphology to help define future sustainable restoration decisions”, submitted to Earth Surface Dynamics, January 2022.

Dear editor and dear authors,

This review has been done jointly by Saraswati Thapa and Mikael Attal. We have read this manuscript with great interest: the research topic is very valuable in its content and we like that the paper highlights the importance of the numerical modelling approach for the evolution of a river reach in response to extreme flood events. We need more studies such as this one, that do combine high resolution pre and post surveys with numerical modelling, to test and calibrate models, and to assess their ability to replicate a range of features of importance to scientists and policy-makers, e.g., volumes eroded and deposited, changes in elevation and morphology, predicted response to anthropogenic changes (land use or risk management).

We enjoyed reading this manuscript that presents a very nice set of experiments using the TELEMAC-MASCARET model to reproduce the dramatic changes that occurred in a reach of the Gave the Pau. The results are enlightening, providing answers to a series of scientific questions and directions for future work. However, there are issues that need to be addressed before publication.

One of the main issues is the weak motivation for using this model in this particular example. We feel this could be better motivated, in particular when it appears, as we go through the results, that this model is not very good at reproducing braiding or suspended sediment transport (and this is a braided reach with ~90% sediment transport in suspension!) There are many landscape evolution models considering many sediment transport equations and multiple grain sizes available. The model in this study used two bed load transport equations and neglected suspended load. The study area has very heterogeneous grain size, however, the model used single grain size D₅₀ for the MPM formula and the D₈₄ for the Recking formula rather than multiple grain size distribution (see for example Ramirez, J. A. et al. (2020) ‘Modeling the geomorphic response to early river engineering works using CAESAR-Lisflood’, Anthropocene, 32. doi:10.1016/j.ancene.2020.100266).

We made recommendations in the annotated manuscript, and one of the suggestions is that you could highlight the strengths and weaknesses of the model from the onset, highlight that the strengths make this model a good potential candidate to model the changes in the study area (it is one of a few models available that are able to model morphodynamics (erosion and deposition) during large flood events), and that here you use this well constrained example to assess the model’s ability to reproduce volumes and cross-sections, and assess its suitability as a tool to inform policy makers" (or something along these lines). Having a clear motivation for the use of the model and clear aims will strengthen the argument, as the reader will know what to expect as they progress through the paper. You can also build on these aims to justify the strategy for modelling, that is, which parameters you are planning to test (or not) and why. The information is there in the text, but we feel it would help if that were presented clearly at the onset. It is also important to build the argument on literature, and we have made suggestions throughout the text.

In general, the paper could do with more details in many sections: description of the model and parameters, description of how data were collected, description of results. There are also places where the outcomes of the model could be evaluated in a more quantitative way. This is particularly crucial for the last section where the impact of restoration scenarios is assessed through a couple of cross-sections, when the previous section demonstrated that the model was not very good at modelling cross-sections and better at modelling volumes.

Finally, the writing can be improved. We have made suggestions in the attached annotated manuscript.

This is an original study. Very few studies have attempted to apply numerical models to natural / real examples to model morphological changes on these space and time scales. We believe there is potential for a strong publication. We hope you find these suggestions useful and wish you all the best with your revisions.

Citation: https://doi.org/10.5194/esurf-2021-91-RC3
- AC3: 'Reply on RC3', Rabab YASSINE, 03 Jul 2022
  
  Dear reviewers,
  
  Thank you very much for this detailed and constructive review which will be very useful to clarify
  
  and improve our article. Please find below the detailed answers. The reviewer’s comments are
  
  shown in bold and some modifications of the manuscript are emphasized in blue.
  Kind regards,
  The authors.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC3

Status: closed

RC1:
'Comment on esurf-2021-91', Damien KUSS, 18 Mar 2022

General comments

This article is original and particularly interesting with the use of a 2D hydromorphological numerical model in a torrential context. The task is challenging as the study concerns the modeling of a reach of braided river with wandering flows which are by nature random and are therefore particularly difficult to model in a deterministic way.

From a methodological point of view, the numerical aspects are very well described. On the other hand, the study site and the modeled domain deserve to be better described: length, slopes... A longitudinal profile, encompassing the modeling domain, would make it possible to better understand the problem of deposition during floods in link with the weirs and with the decrease in the longitudinal slope. The influence of the solid volume taken inot account, linked with the slope, could be discussed in the results section.

The hydromorphological modeling is carried out taking into account only bedload (with Meyer Peter and Recking formulas). But you mention, by exploiting the data from dredging, that the fraction of the volume transported by bedload would represent only 8 to 16% of total transport (lines 417 to 417). There is an inconsistency:

• The altitudinal evolutions observed by DoD are compared to the modeled bed evolutions, which only include sediment transport by bedload;

• On the other hand, the modeled transported volume is compared to the total volume observed minus the fine fraction (volume transported by suspension).

The performance of each modeling scenario is evaluated with the BSS score. In a braided river context, the scores obtained are not good (BSS = 0.06). You discuss the representativeness of this metric for such river morphologies. If the metric can be criticized because of the random nature of the wanderings, it would also seem relevant to question the added value of a 2D model compared to a 1D model in such a context. You could precise the configurations where a 2D model is apropriate.

Specific comments

Reference Reisenbücchler et al., 2019 at the end of sentence line 19. What does it refers to ?

Line 19 : “They showed”. “They” refers to Rickenmann et al. or Reisenbuchler et al. ?

Line 38 to 40 :

- a mention could be made of more recent formulas partially based on field data (Recking, 2013 ; Lefort, 2015).

- are bedload transport formulas established in 1d narrow channels directly transposable in 2D models ?

Lines 40-41 : the slope used for the sediment transport calculation at the upstream boundary condition of the model is a key parameter to perform realistic simulations. It must be apprehended by a geomorphological analysis based on the the longitudinal profile.

Line 63 : “The TELEMAC-MASCARET modelling system has been considered well suited to perform 2D morphodynamic simulations on the LDG reach”. Why ?

Line 92-93 : you could add a reference at the end of the sentence.

a longitudinal profile would complete the description and would allow to better understand the effects of the two weirs during the floods.

Lines 101-104 : you could add references concerning the peak discharges.

Figure 3 : you could add the reference for each picture

Lines 164-166 : you could add a reference : Rickenmann and Recking (2011) ?

Lines 185-188 : please define all the terms used in equation 5

Lines 195-208 : note that the recking formula : (a) is mainly base on lab experiments for high shields numbers ; (b) is very sensitive to the choice of the shields mobility parameter. So the avantages of using this formula for intense floods where the shields number exceeds the mobility parameter could be discussed.

Lines 225-226 : it is not clear how you use the dredging data.

Lines 234-235 : “However, the recorded volumes represent both very fine sediments probably transported by suspension and very coarse sediment via bedload transport”. The model used in the study takes into account only the bedload transport, Is it right ?

Lines 269-270 : you therefore make the assumption of no downstream influence.

Line 276 : what do you mean by “instabilities” ?

Lines 276-278 : the slope of the LDG reach is never given. What is slope at the upstream boundary condition ? Is the hypothesis of stable riverbed evolution justified ? What was the observed evolution of the river bed at the upstream condition during the floods ?

Lines 286-287 : you could detail what you mean by numerical and physical parameters.

Figure 6 : how do you transform 2D results into 1D longitdunal profile ?

Lines 326-327 : “To date, numerical models cannot predict channel migration processes that occur in braided rivers. These phenomena are uncertain and random. A modeler should thus not expect the model to predict channel migration accurately during a flood.”. I agree. Butin consequence you should better justify why you have chosen a 2D hydromorphological model for this case study.

Lines 331-332 : is the 2019 LiDAR realigned ?

Figures 10-11 : I regret that the 2016 profile was not drawn with a solid line. it's hard to see the position of the fall. You could also explain how you transform 2D results with a bed level not constant over cross sections in to longitudinal profiles.

Figure 13 : how do you explain the deposits above the max water level ?

Figure 14 : the initial profile is missing. It could be a usell information.

Citation: https://doi.org/10.5194/esurf-2021-91-RC1
- AC1: 'Reply on RC1', Rabab YASSINE, 26 Jun 2022
  
  Dear reviewer,
  
  Thank you very much for this detailed and constructive review which will be very useful in improving the quality of our article. Please find attached the detailed answers. The reviewer's comments are shown in bold, and some modifications of the manuscript are emphasized in blue.
  Kind regards.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC1
RC2:
'Comment on esurf-2021-91', Clément Misset, 31 Mar 2022

The comment was uploaded in the form of a supplement: https://esurf.copernicus.org/preprints/esurf-2021-91/esurf-2021-91-RC2-supplement.pdf

Citation: https://doi.org/10.5194/esurf-2021-91-RC2
- AC2: 'Reply on RC2', Rabab YASSINE, 28 Jun 2022
  
  Dear reviewer,
  
  Thank you very much for this detailed and constructive review which will help us improve the quality of our article. Please find attached the detailed answers. The reviewer's comments are shown in bold, and some modifications of the manuscript are emphasized in blue.
  Kind regards.
  The authors.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC2
RC3:
'Comment on esurf-2021-91', Saraswati Thapa, 03 Apr 2022

Review of “Numerical modelling of the evolution of a river reach with a complex morphology to help define future sustainable restoration decisions”, submitted to Earth Surface Dynamics, January 2022.

Dear editor and dear authors,

This review has been done jointly by Saraswati Thapa and Mikael Attal. We have read this manuscript with great interest: the research topic is very valuable in its content and we like that the paper highlights the importance of the numerical modelling approach for the evolution of a river reach in response to extreme flood events. We need more studies such as this one, that do combine high resolution pre and post surveys with numerical modelling, to test and calibrate models, and to assess their ability to replicate a range of features of importance to scientists and policy-makers, e.g., volumes eroded and deposited, changes in elevation and morphology, predicted response to anthropogenic changes (land use or risk management).

We enjoyed reading this manuscript that presents a very nice set of experiments using the TELEMAC-MASCARET model to reproduce the dramatic changes that occurred in a reach of the Gave the Pau. The results are enlightening, providing answers to a series of scientific questions and directions for future work. However, there are issues that need to be addressed before publication.

One of the main issues is the weak motivation for using this model in this particular example. We feel this could be better motivated, in particular when it appears, as we go through the results, that this model is not very good at reproducing braiding or suspended sediment transport (and this is a braided reach with ~90% sediment transport in suspension!) There are many landscape evolution models considering many sediment transport equations and multiple grain sizes available. The model in this study used two bed load transport equations and neglected suspended load. The study area has very heterogeneous grain size, however, the model used single grain size D₅₀ for the MPM formula and the D₈₄ for the Recking formula rather than multiple grain size distribution (see for example Ramirez, J. A. et al. (2020) ‘Modeling the geomorphic response to early river engineering works using CAESAR-Lisflood’, Anthropocene, 32. doi:10.1016/j.ancene.2020.100266).

We made recommendations in the annotated manuscript, and one of the suggestions is that you could highlight the strengths and weaknesses of the model from the onset, highlight that the strengths make this model a good potential candidate to model the changes in the study area (it is one of a few models available that are able to model morphodynamics (erosion and deposition) during large flood events), and that here you use this well constrained example to assess the model’s ability to reproduce volumes and cross-sections, and assess its suitability as a tool to inform policy makers" (or something along these lines). Having a clear motivation for the use of the model and clear aims will strengthen the argument, as the reader will know what to expect as they progress through the paper. You can also build on these aims to justify the strategy for modelling, that is, which parameters you are planning to test (or not) and why. The information is there in the text, but we feel it would help if that were presented clearly at the onset. It is also important to build the argument on literature, and we have made suggestions throughout the text.

In general, the paper could do with more details in many sections: description of the model and parameters, description of how data were collected, description of results. There are also places where the outcomes of the model could be evaluated in a more quantitative way. This is particularly crucial for the last section where the impact of restoration scenarios is assessed through a couple of cross-sections, when the previous section demonstrated that the model was not very good at modelling cross-sections and better at modelling volumes.

Finally, the writing can be improved. We have made suggestions in the attached annotated manuscript.

This is an original study. Very few studies have attempted to apply numerical models to natural / real examples to model morphological changes on these space and time scales. We believe there is potential for a strong publication. We hope you find these suggestions useful and wish you all the best with your revisions.

Citation: https://doi.org/10.5194/esurf-2021-91-RC3
- AC3: 'Reply on RC3', Rabab YASSINE, 03 Jul 2022
  
  Dear reviewers,
  
  Thank you very much for this detailed and constructive review which will be very useful to clarify
  
  and improve our article. Please find below the detailed answers. The reviewer’s comments are
  
  shown in bold and some modifications of the manuscript are emphasized in blue.
  Kind regards,
  The authors.
  
  Citation: https://doi.org/10.5194/esurf-2021-91-AC3

Rabab Yassine et al.

Viewed

Total article views: 995 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
725	236	34	995	20	18

HTML: 725
PDF: 236
XML: 34
Total: 995
BibTeX: 20
EndNote: 18

Views and downloads (calculated since 26 Jan 2022)

Month	HTML	PDF	XML	Total
Jan 2022	119	36	3	158
Feb 2022	101	23	2	126
Mar 2022	165	22	7	194
Apr 2022	57	22	3	82
May 2022	26	14	1	41
Jun 2022	37	11	5	53
Jul 2022	38	18	2	58
Aug 2022	8	12	0	20
Sep 2022	8	5	0	13
Oct 2022	15	3	0	18
Nov 2022	19	5	1	25
Dec 2022	16	2	0	18
Jan 2023	15	5	0	20
Feb 2023	13	7	0	20
Mar 2023	22	5	0	27
Apr 2023	13	6	0	19
May 2023	14	2	4	20
Jun 2023	7	1	1	9
Jul 2023	6	15	0	21
Aug 2023	6	9	3	18
Sep 2023	20	13	2	35

Cumulative views and downloads (calculated since 26 Jan 2022)

Month	HTML	PDF	XML	Total
Jan 2022	119	36	3	158
Feb 2022	101	23	2	126
Mar 2022	165	22	7	194
Apr 2022	57	22	3	82
May 2022	26	14	1	41
Jun 2022	37	11	5	53
Jul 2022	38	18	2	58
Aug 2022	8	12	0	20
Sep 2022	8	5	0	13
Oct 2022	15	3	0	18
Nov 2022	19	5	1	25
Dec 2022	16	2	0	18
Jan 2023	15	5	0	20
Feb 2023	13	7	0	20
Mar 2023	22	5	0	27
Apr 2023	13	6	0	19
May 2023	14	2	4	20
Jun 2023	7	1	1	9
Jul 2023	6	15	0	21
Aug 2023	6	9	3	18
Sep 2023	20	13	2	35

Viewed (geographical distribution)

Total article views: 936 (including HTML, PDF, and XML) Thereof 936 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 28 Sep 2023

Short summary

The evolution of river morphology is very complicated to predict, especially for mountain rivers with complex morphologies such as the LDG reach in France. A 2D hydromorphological model was developed to reproduce the channel's evolution and provide reliable volumetric predictions while revealing the challenge of choosing adapted sediment transport and friction laws. Our model can provide decision-makers with reliable predictions to design suitable restoration measures for this reach.


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