Hydraulic Modeling and Evaluation Equations for the Incipient Motion of Sandbags for Levee Breach Closure Operations

1

This   manuscript concerns the analysis of laboratory modelling for the evaluation   of the empirical equations for the assessment of critical velocity formulae   of incipient motion of sandbags for levee breach closure. The experimental   dataset is compared with other studies within the same topic.

The   research herein presented is certainly within the scope of the Journal of   Water. However, the manuscript organization, the incompleteness and the lack   of clarity/novelty in some parts of the text should be amended prior to   publication. My current decision is major revision and I reckon the author to   take into consideration all the queries listed below before the manuscript   could be considered for publication.

We thank the reviewer for his time and efforts and constructive   comments which all have been incorporated in the update manuscript. The   authors hope that it will meet the reviewer expectations.

2

In   the abstract it is said that 3D velocity measurements were performed.   However, the results did not show such measurements. Could the authors really   show the velocities in the three directions?

Thank you. The reviewer’s comment was addressed accordingly. (Please   see Figures 2a and 2b).

The authors used ADV for velocity measurements and the machine   measures 3D velocity at any given point, in x-, y- and z- directions. The   authors have presented velocity in two planar surfaces, i.e., in x-y plane   and in y-z plan. This describes the velocity in the three directions. This has been   replaced by velocity in x, y and z plans in updated manuscript. We hope this   answers the reviewer comment.

3

The introduction is well written   however nothing is explained about the morphological collapse of the dikes.   Please, take a look at the following reference and expand the explanation   about the eroding process and the different numerical strategies for   modelling such phenomenon.

A 2D   weakly-coupled and efficient numerical model for transient shallow flow and   movable bed. C Juez, J   Murillo, P García-Navarro. Advances in Water Resources 71, 93-109

The authors would like to thank the reviewer for comment. They have   considered the proposed reference and cited in manuscript in a brief   discussion about the importance of modeling collapse of dikes and dam break   flows.

However, we would like to point out that in our study, we focused on   immobile solid bed, with no eroding material, to investigate the effect of   shape and size of sandbags on the threshold of motion, we did not detail   morphological collapse of dikes. The main objective of our study was to   calibrate existing equation for incipient motion of sandbags and to develop   new ones as presented in Tables 3-6.

4

Following   with this idea of the geomorphological collapse: the authors have designed   the experiments with a steady flow travelling through the lateral breach   located in the channel. However, the levee failure is combination of a   continuous eroding process and geomorphological collapse, when large chunks   of material fall down and the water is flowing as in a dam break. Could the   authors elaborate about this issue taken into account their lab setup?

We would like to thank the reviewer for this comment. The reviewer is   absolutely correct, failure of side protection levees in channels is a   gradual process with continuous erosion of bed and widening of levee until a   final stable condition is reached, where levee has widened to maximum width   and water flow is maximum. This is unlike dam break flow which is a   continuous interaction process between erosion; dike morphology and dam break   flow.

This is confirmed by Hurrican Katrina levees in New Orleans, which all   reached to a stable final condition. At this final condition (which is   reached shortly after breach), efforts to close the breach with sandbags are   attempted.

In our work, we are making experiments for incipient motion of   sandbags in side breach flows to calibrate existing and develop new empirical   equations for linking sandbag size and critical flow velocity.

To include breach failure, downstream bed morphological changes, and associated   change in breach flow/velocity; this would be an extremely complex situation   with various parameters and time would be a factor. A simple model would not   be feasible. However this can be considered in future research.

5

Quality of Figure 1 should be improved

The requested revision was done accordingly. (Please see Figure 1).

6

A table with the hydraulic parameters is missing before the   results section.

Thank   you for the comment. The   reviewer’s comment was responded to   accordingly. A table was added in the manuscript. (Please see Table 2).

7

In table 4 the authors should justified the two breach widths   chosen. How could the authors scaled up these results in reality?

We thank the reviewer for the comment. Two different breach   widths are chosen in the experimental setup, one where breach width is around   60% from channel width and the other breach width is 100% from channel width.   Authors chose them to represent two possible breaches, a small one and a   large one. In many events, e.g., in hurricane Katrina, none of the channel   breaches’ width was larger than channel width, they were smaller. This   manuscript is an experimental study with measurements done on a small scale   prototype, it is not scaled from a real situation. However, its tests   situations that are similar to what occurred during this event.

Equations produced link the incipient motion of sandbags with   breach hydraulics and sandbag hydrodynamic characteristics and have been   further tested on real large scale case of 17th street canal breach.

8

Figure 2 should be completely redone. Information given there is   really poor. Increase the resolution of the results since you are using an   ADV equipment.

The requested revision was done accordingly. (Please see Figure 2).

9

Conclusions section should be re-written. The novelty and real   conclusion don’t really arise from those lines

The requested revision was done accordingly. (Please see Conclusion).

1

First   of all I want to congratulate the authors for such a magnificent work in   applied research, always useful to work on such important issues as the   effects of floods. I think that the article is well structured and that it   meets the requirements of the journal, with a good introduction to the   subject, the search of previous cases.

We would like to thank the reviewer for   his positive opinion about our work and indeed on the time taken to review   the work and comments to enhance it.

2

A simple drawing   would have been good explaining the dimensions, a, b, c, in each case. For   example in Table 1

We thank the reviewer   for this comment. The requested revision was done accordingly. We have   defined dimensions and in case it is not clear, we would be glad to make an   explanatory drawing (Please see Section 3).

3

When is talk about   parameters b and c in Figure 5a, I understand that it´s refers to the height   of the sandbag and its length, c.

4

It could also have   been accompanied by a small explanatory diagram of the first paragraph of   section 5.1., Which describes the protocol followed in the experiments. The   experiments were sufficient and yielded a series of results to work with when   deciding the V_cr.

If after updates in manuscript, the reviewer still   feels that this figure is important, we will place it in paper.

5

In the uncertainty   analysis, it is not specified with which tools or programs has been done to   obtain the errors

Thank you for the   comment. The uncertainty analysis defines the   individual prediction error as                                                                                   . The calculated prediction errors for the entire dataset   are used to calculate the mean and standard deviation of the prediction   errors as  and , respectively. A negative mean value indicates that the   prediction model underestimated the observed values, and a positive value   indicates that the equation overestimated the observed values. Using the   values of  and S_e, a   confidence band can be defined around the predicted values of an error using   Wilson score method without continuity correction; the use of ±1.96   S_e yields an approximately 95% confidence band. These equations   are implemented in the excel file and uncertainty bounds were calculated.

1

The manuscript analyzes the problem of levee breach   closure by sandbags with an hydraulic approach considering the prediction of   critical velocity for the incipient motion of a sandbag in the breach   channel. Experiments are conducted on a physical model in order to study the   flow field around the breach channel and to detect the incipient motion   velocity of prism and spherical sandbags. A new formulation for the critical   velocity is proposed, it is compared with some existing formulae and it is   applied to the real case of 17th street canal during Katrina hurricane.

The issues analyzed are of interest either for   scientific research and for practical applications. However, my opinion is   that the manuscript needs to be improved regarding i) the existing literature   review, ii) the experimental tests presentation and results’ discussion, and   iii) the application to the real case. In fact, the paper is often not clear   in presenting the focus of the study, the experiments presentation and   discussion are not complete and refer to findings that are not well supported   by the figures, and the application to the real case needs to be more   specific and organized in a more logical order. Specific comments are given   to the authors.

Finally, the paper may be accepted for publication   after a major revision is performed in order to address all the following   issues:

Reviewer   four time and effort in reviewing the manuscript is appreciated and comments   are all instructive and have been incorporated in the updated manuscript.

2

Title   is not clear in the aim of “protecting the morphology of the river channel   and valley”. Such aims are not dealt in the paper, which is focused on the   levee breach closure. I suggest to modify the title as it follows: “Hydraulic   modeling and evaluation equations for the incipient motion of sandbags for   levee breach closure operations”

Done   as per reviewer request. The title is now; Hydraulic modeling and evaluation equations for the incipient motion of   sandbags for levee breach closure operations.

3

Lines   38-40: references should be given for levee breach events;

Done   as per reviewer request.

4

Line   41: reference should be added to support the similarity between side weir   flow and levee breach (e.g., Kamrath et al., 2006 and Michelazzo et al.,   2015)

Done   as per reviewer request.

5

Lines   113-114: the morphological changes of river channel and valley are not   addressed in the paper. For instance, the experimental tests do not deal with   movable bed or sediment transport

The   reviewer is absolutely correct and we thank him for this comment. We have   omitted this part in the updated manuscript.

6

Section   2: units of measure should be added to every variable

As   per reviewer comment, we have added units beside notations in list of   notations.

7

Sections   3-4. Experimental tests should be presented and discussed in a more detailed   way

We   have added some clarifications in the updated manuscript. We hope the   reviewer finds the updated description more clearer.

8

The   authors should discuss if the physical model is a scale model, the influence   of scale effects, the reference to a prototype and how the results scale up   to the prototype.

We   thank reviewer for this comment. The developed model is not a scale model. We   have clarified this in the updated manuscript that we are modeling a general   case of levee breach and thus there are no scale effects.

9

Line   155-156: how much was the Manning coefficient? What was the flume bed   material? In which zones of the flume were the 0.01 (unit of measure is   missing) water levels measured? For which discharge?

In   the updated manuscript, Manning coefficient is 0.01, flume bed and walls are   made from immobile material, from rough concrete.

10

The   same level of breach and main channel may not be the worst configuration for   shear stresses in the breach; were other levels taken into account or could   the authors provide some comments about this?

As   a general comment, the authors should explain the choices made for the   experimental setup and justify why other configurations were not taken into   account (e.g., different bed elevations, roughness of flume and inundation   area, movable bed conditions, curved channel, …)

We   thank reviewer for this comment.

We   were attempting to model a general basic case for levee failure and side   channel flow and to test motion of sandbags when used for breach closure.   Thus, we needed bags to be carried away by moving water and we needed to   record the onset of motion, having a breach bed lower than channel would not   allow us to capture bag motion as done in same level beds. And sandbag would   fall in breach depression.

However,   this is an important comment and we have added this in conclusion section   when discussing model assumptions.

11

Provide   a clear table explaining the performed tests: how many tests? For which Qu,   Qr, Lbr, sandbags dimensions, downstream boundary condition?

Sandbag   dimensions are given in Table 1. Number of run tests are given in Table1. For   Qr, and Lb, Table 2 is added in the updated manuscript.

12

Define   the locations where the ADV measures were taken

If   the reviewer wants that the authors place following figure, we would be glad   to do so.

13

Figure   1: should not the coordinate system origin be at the beginning of the breach?

Done   as per requested by reviewer.

14

Table   1: define variables that have not been introduced previously

Done   as per requested by reviewer.

15

In   which part of the breach the sandbags were placed?

Sandbags   are placed in mid-section of breach. This has been placed in the updated   manuscript.

Then   a sandbag with a selected dimension is placed at the bottom of the flume at   the mid-breach section.

16

Section   5. The results’ presentation should be improved and the discussion should   refer to the existing literature. In particular, flow zones around breach   channels have been analyzed in detailed by Michelazzo et al. (2015 and 2018)   for the specific case of zero-height side-weir and breach (that are the same   of the present paper).

The   requested revision was done.

17

Provide   figures to understand the flow zones in more details and to visualize the   water surface around the breach and along the channel

As   per the request of the reviewer, we have added Figure 2. We hope that it is   clear now and in satisfaction for reviewer.

18

Line   214: to which test are “38 and 42 cm” referred?

We   thank reviewer for this comment. We have indicated case in the updated   manuscript.

For   Lb=1000px, and Qr=1.

19

Provide   explanation for the separation and stagnation zones having the same water   level

We   thank reviewer for this comment, we have added a figure showing the extent of   the zones that are formed in open channel side breach. We hope that they are   clear for reviewer and readers, especially that we have presented the two   studied cases, Qr=1 (downstream channel closed end) and Qr=0.8 (downstream   channel opened end).

20

line   230: influence of breach width on 3D flow zone have been discussed by   Michelazzo et al. (2015 and 2018)

-   the 3D flow zones should be detected more precisely and supported by figures   and quantitative analyses

We   thank reviewer for this comment. Proper citation for these two important   papers has been added in updated manuscript.

21

Figure   2a: provide a scale vector velocity (e.g., a 1m/s vector)

Done   as per request of reviewer.

22

Figures   2a and 2b should be renamed in Figures 2 and 3 respectively, in order to   avoid confusion with the subplots a) and b)

Done   as per request of reviewer.

23

Plot   water surface in figure 2b

Water depth along   channel for L_b=1000px, and Q_r=1 shown in Figure 2.

24

Figure   2b: “X” or “x”? Use consistent symbols throughout the text.

X   is used in the manuscript.

25

Define   Vcr* in figure 3

As   per reviewer request, this item is defined before relevant figure and in list   of notations.

26

Is   eq.7 defined on dataset of both smooth and rough beds? Can the authors   provide justification and implications of deriving an equation on different   conditions?

This   equation has been described by the literature of Novak and Nalluri and not by   the authors.

27

What’s   the x-label of figure 4a?

We   thank the reviewer for this comment

28

Lines   320 and 360: where is USACE (2007) in figure 4b and in table 3?

The   figure shows a picture for attempts by the USACE to close breach. The table   contains the results of critical velocity using the real sandbag weights used   by the ASCE.

29

Eq.   11: should not the exponent of the lower bound be “-e - 2*Se”?

We   thank the reviewer for this comment. We have corrected this in the updated   manuscript.

30

Section   5.4. As a general comment, the text is not clear when the authors speak about   “order of magnitude” in comparing the critical velocity prediction. For   instance, at line 405, the difference between the actual velocity (2.5 m/s)   and the predicted velocity (1.67 m/s and 1.81 m/s for 3000 lbs sandbag) is   not of an order of magnitude but much less. Moreover, the reviewer suggests   to organize better this section according to the aim of the paper and to keep   only the comparisons that are focused on the aim of the paper. For instance,   is table 6 real necessary or could it be deleted? Could tables 4 and 5 be   jointed?

As   per request of reviewer, we have removed the term ‘order of magnitude’ from   the updated manuscript.

31

Line   425: is “another developed equation” eq. “Present study-1” in table 5? How   was this equation derived?

This   equation was developed using the experimental results of Zhu et al. [43] and   the present study experimental measurements between critical velocity for   sandbag threshold motion and the sandbag height to length ratio with R2 of   0.83. This is in the manuscript in description of equation 9.

32

Finally,   the authors should clarify how the “calibrated” versions of the eqs in tables   3,4,6 have been derived

This   has been done by simple multiple regression.

33

Section   6.

Lines   480-483: this sentences are not clear and refer to issued that have not been   analyzed in the paper. The reviewer suggests to delete it or to substitute   with conclusions more consistent with the analyses performed in the paper.

Done   as per request of reviewer.

34

Add   units of measure in the “List of notations”

Done   as per request of reviewer.

35

Reference:

Kamrath,   P., Disse, M., Hammer, M., & Koengeter, J. (2006). Assessment of   discharge through a dike breach and simulation of flood wave propagation.   Natural Hazards, 38(1–2), 63–78. https://doi.org/10.1007/s11069-005-8600-x

Michelazzo,   G., Oumeraci, H., & Paris, E. (2015). Laboratory study on 3D flow   structures induced by zero-height side weir and implications for 1D modelling.   Journal of Hydraulic Engineering, 141(10), 0401502.   https://doi.org/10.1061/(ASCE)HY.1943-7900.0001027

Michelazzo,   G., Oumeraci, H., & Paris, E. (2018). New hypothesis for the final   equilibrium stage of a river levee breach due to overflow. Water Resources   Research, 54, 4277–4293. https://doi.org/10.1029/2017WR021378

1

The reviewer wants to   thank the authors for their paper about the motion of sandbags, which they   have experimentally investigated. He/she has some   suggestions/comments/questions, which should be addressed by the authors:

The authors thank the reviewer for his effort in reviewing the   manuscript and detailed comments, which are incorporated in the updated   manuscript.

2

The citation style is   not according the regulation of the water. Please correct this in the paper   as well as in the References.

Done as per the request of the reviewer.

3

Line (l)135-154: Are   the test really conducted at two Universities with the same experimental   set-up. This should be clarified. 

We would like to   thank the reviewer for this comment. Initial   runs were made at the University of South Carolina, but the actual experimental   setup and measurements were made at Cairo University. In order to avoid   confusion, we have removed the part of South Carolina and updated this part. (Please see Section 3).

4

 L156: discharge   measured by only one manometer? It shows the differential pressure. But why   is in the Fig.1 one electromagnetic flow meter presented. Why is this one not   used?

We would like to   thank the reviewer for this comment. The reviewer is absolutely correct, the   authors have updated the manuscript so that it is clear that the discharge is   measured by the electromagnetic flow meter installed on discharge pipe. (Please see Section 3).

5

 L158: include   this 3 m in the Figure 1, so that the location is clear. 

Done as per the request of the reviewer.

6

Please give the   minimum and maximum water depth in the channel behind the breach as well as a   relation between the two values (water height and discharge remaining) to   have a feeling for the flow speed behind the breach. … more information than   given in line 210-217 and also for both breach width. 

We thank reviewer for this comment. We have added following figure for   clarification.

7

Please also include   the Froude numbers before and the range after the breach.

Done as per the request of the reviewer.

8

How far is the gate   downstream of the breach? Does it have an influence of the flow in the breach   because of a backflow in a high position? How height was the max position in   relation to the remaining water depth over the gate?

The authors thank the   reviewer for this comment. The authors have updated figure 1 as per request   of reviewer to be clearer, the downstream gate is at 3m from breach location.   When all channel flow is directed through breach (80% from channel flow), the   water is stagnant and a recirculation zone is developed at the downstream as   shown in below figure. There is no effect on outlet boundary on breach flow   as it is an open boundary.

9

Fig.1 the presented   measurements should concentrate on the experimental set-up and not only the   outsides.

Done as per the request of the reviewer.

10

L164: the 60 mm   honeycomb is a good solution to reduce secondary flow speeds but seem to be   not very long. But it doesn’t help in case of the incorrect distribution of the   main flow. Was the flow after the honeycomb checked as well as upstream of   the breach so that a fully developed velocity profile?

The authors thank the   reviewer for this comment. Yes, the authors have made checks on velocity   profile and taken velocity measurements before breach by 500px and by 1m and   found a satisfying developed vertical flow profile.

11

Fig.1 the printed   coordinate system doesn’t fit the description in the text. If the reviewer is   correct, the positive z-axis would point downwards. Please show it in the   correct position of the origin.

Done as per request of reviewer.

12

How does the water   flow back into the storage after it flows out of the breach? Is there an   additional step indicated by the rectangle with the arrows? How is this outflow   boundary condition checked and modelled? 

Ans.   Thank you for the comment. The outflow is a   free boundary where water flowing out of the breach is allowed to flow freely   without steps or any obstacles. It does not affect the water out of the   breach by any means. There are no additional steps in water way when it exits   breach; it is allowed to leave freely.

Thus, there is no   need to model any outflow boundary.

13

Fig.1 a photo of the   actual set-up would help. 

Ans. Thank you for   the comment. The quality of figure 1 was improved.

14

L187   and ongoing: discharge is measured incoming (it is assumed that two devices   are used) and at least one more measurement is needed. Is the discharge   measured over the slice gate on the downstream side? 

We thank the reviewer   for this comment. The discharge is measured at the upstream using the   electromagnetic flow meter, this gives us the total flow entering the   channel. When all flow is allowed through breach, no other measurement is   needed. When only a portion of flow directed to breach, a downstream gate is   used to do this job. Gate is calibrated through independent experiments to   determine flow throw it.

15

L191: this initial   one is specified by which condition?

We thank the reviewer   for this comment. The initial conditions prior to each test are described in   manuscript by steady conditions, when the flow is stabilized in the whole   experimental setup. This means when flow is developed before the breach and   when several repeated experimental measurements at same location yield same   result for water depth and velocities.

How many bags are inserted? A closed line or only   one? How is motion defined in this relation? Is it enough, when it changes   the shape or is a small movement enough?

In our experiments, we used single solitary   sandbag with pre-set shapes. Only one was tested at a time for incipient   motion. Motion is defined when bag starts to move under the effect of flow   shear stress. The sandbag does not change in shape, it remains with the pre-set   shape during the whole experiment.

Why is it reduced in three steps? 

In order to define   the incipient motion threshold, which is the velocity at which sandbag starts   to move. Authors chose a procedure mentioned in literature, which is adopted   by Gogus and Defne (2005) and Gulcu (2009). We have added this in updated   manuscript.

Gogus M, Defne Z.   Effect of shape on incipient motion of large solitary particles. Journal of   Hydraulic Engineering. 2005; 131 (1), 38–45.

Gulcu B. Incipient   motion of coarse solitary particles, MSc. Thesis, The Graduate School of   Natural and Applied Sciences of Middle East Technical University, Ankara,   Turkey, 2009.

Does it make a difference if the filled bags are   filled with water or not? How did the authors consider this? 

Bags were made from   cloth similar to that used in real sandbags, in both when bag is placed in   water, water shall penetrate bag to the sand inside. This automatically   occurs in the experiments. So the experiments imitate real case.

L198: the bag(s)   is/are inserted again and the critical conditions are reach again to measure   the velocity profile. Where was this conducted in the breach?

Bag is placed in mid-section   of breach. This is placed in the updated manuscript.

General: the reviewer   doesn’t understand why this analysis has to be conducted in this complex   situation of the breach. It could be easily done in the flume upstream with   only longitudinal velocities and then tested in the breach, where there is a   3D-veloiticy distribution to check the conducted investigation.

We thank reviewer for   this comment. We wanted to model the same case and in previous work by the   authors, it was found that the distribution of velocity in breach flow (weir   side flow) is different from that in developed open channel flow as shown in   below figure.

L188 as breach width   0.25 and 0.4 m are used. … please check the whole paper that all given values   are correctly associated with the correct width and given for both. 

Done as per the request of the reviewer.

Figure 2a and 2 b   should be 2 and 3. The axis are not clear and furthermore those cuts should   be indicated in Fig. 1 … the reviewer will check these figures after this was   clarified.

Done as per the request of the reviewer.   We hope that they are satisfactory in the updated manuscript.

L236 vectors and no   streamlines. Please provide a reference vector with a defined length in the   Figures. 

Done as per the request of the reviewer.   Streamlines are replaced by vectors in the updated manuscript. A reference   vector is added in figure.

Sect 5.2 why is the   full discharge through the breach case chosen to be shown? 

We chose results of   this case to be presented since they give higher flow velocity through the   breach and thus plots of velocity vectors would show more clearer the direction   of flow in 3D through the breach.

Figure 3 wrong   citation style and only include the circular bags? Please clarify this also   in the caption of the Figure. 

Figure 3 and 4b   please don’t use logarithmic axis. 

As per reviewer comment, the authors have updated   the figure caption to show that it includes circular sandbags.

Regarding changing the figures axes to be normal   instead of logarithmic, this would not be good in figure presentation.

General a 22 cm sphere should fill nearly half of   the breach width … are those sandbags fully filled so that they remain their shape?   In this case wouldn’t they roll way? So a certain deformation should be   considered but how does the authors include this?

We would like to   thank the reviewer for this comment. Sandbags with different shapes are fully   filled with sand prior to being woven so that shape of sandbag does not   deform. No deformation was observed during experiments.

Figure 5 b: did the authors capture this picture?   Source? 

This figure is from   Sattar et al. (2008). This has been placed in the updated manuscript.

Sattar AMA, Kassem A, Chaudhry MH. Case study: 17th   street canal breach closure procedures. Journal of Hydraulic Engineering.   2008; 134 (11), 1547–1558.

Please   include in the discussion a section with the model assumption. 

Done as per requested   by the reviewer in the conclusion section.

The   influence of the shape change of the sandbag in the experiment should be   addressed as well as the missing information about the experimental set-up.

We would like to   thank the reviewer for this comment. Sandbags with different shapes are fully   filled with sand prior to being woven so that shape of sandbag does not   deform. No deformation was observed during experiments.

2

Moreover,   there are some problems with the files provided by the authors. In   particular, the new version of the manuscript (water-406055-peer-review-v2.pdf)   does not include some of the responses to some issues raised by the reviewer,   even if the authors declare that they’ve addressed the comments in the   response-to-reviewer file. Further, the supplementary file “water-406055-supplementary.docx”   is different from the revised manuscript and, therefore, it is not clear its   connection with the manuscript. Therefore, I recommend a minor revision   before the paper publication.

Updated   revision will eliminate this problem.

3

-   it is not clear which references for flood events of Section 1 have been   added and if reference [1] includes all the levee breach events cited by the   authors

This has been updated in the revised manuscript.

4

-   no specific comments about scale effects can be found in the revised version   of the manuscript. However, even if the model is not a scale model, it is   evident that the laboratory scale is quite different from the real case scale   and that scale effects could arise.

Since   the results are used for real cases prediction, the authors should give some   comments to support the applicability of their experimental findings to the   real scale

We have added a discussion about scale effects in updated manuscript   as per request of reviewer.

5

-   the useful comments made by the authors about the experiments’ assumptions   should be added also in the Section that introduces the experimental setup   and not only in the Conclusions

Done as per request of reviewer

6

-   why Frd was not defined in Table 2 when Qr=1? How can Fru be constant   regardless variations of breach width and Qr?

The authors thank the reviewer for this comment. In case of Qr=1, all   upstream channel flow is directed towards breach and the downstream end of   the channel after the breach has stagnant water and thus no Fr. The upstream   flow to the experiment is kept constant in all breach widths and this gives similar   velocity, thus Fr does not change.

7

-   the figure proposed by the authors to locate ADV measurements is useful, and   the reviewer agrees to add it to the revised paper.

Anyway,   a reference to the locations depicted in the proposed figure should be added   in the text. Moreover, the following corrections should be taken into   account:

i)   add “X=0.25”

ii)   place the coordinate system exactly where X=0.0 and Y=0.0
  iii) remove the background squared gray grid

Done as per request of reviewer

8

-   the above comments about the placing of coordinate system and the removing of   the background squared gray grid should be applied to the other similar   figures

Done as per request of reviewer

9

-   the authors should comment further the choice of the sandbag location, since   the mid-section could not be that one with the maximum outflow velocity

We thank the reviewer for this comment.

Based on previous experience with work in breach closure with   sandbags, Sattar (2009) extensively studied various procedures for closure of   levee breaches and found that staring by dumping sandbags in mid breach   sections would be the most effective way for closure. This has been clarified   in the revised manuscript.

10

-   reference Michelazzo et al. (2018) should be added to the References Section.   “Michelazzo, G., Oumeraci, H., & Paris, E. (2018). New hypothesis for the   final equilibrium stage of a river levee breach due to overflow. Water Resources   Research, 54, 4277–4293. https://doi.org/10.1029/2017WR021378”

Done as per request of reviewer

11

 -   regarding the figures’ naming, please use letters “a”, “b”, etc only in case   of subplots of the same figure; otherwise, use progressive numbers

Done as per request of reviewer

12

-   figure 2 is a good representation of water surface variation. However, the   reviewer intended to plot water surface in figure showing y-z velocity   profiles (figure 4b in the revised manuscript) in order to visualize the flow   toward the breach

Done as per request of reviewer

13

-   it is still not clear where USACE model is reported when predicting the   critical velocity.

 For instance, it should be in table 4   according to line 435, but it can’t be found.

The authors thank reviewer for this comment. In the updated   manuscript, this has been corrected to Gulcu (2009).

In   the same way, where can the Stelczer velocity, that is cited at line 487, be   found in table 5? The authors are invited to solve this issue or to explain   it better

The authors thank reviewer for this comment. In the updated   manuscript, this has been corrected to Novak and Nalluri 1984.

14

-   no replies are given by the authors to reviewer’s comments about organization   of Section 5.4 and presentation of tables 4-5-6

Section 5.4. As a general comment, the text is not clear when the   authors speak about “order of magnitude” in comparing the critical velocity   prediction.

For instance, at line 405, the difference between the actual velocity   (2.5 m/s) and the predicted velocity (1.67 m/s and 1.81 m/s for 3000 lbs   sandbag) is not of an order of magnitude but much less.

Moreover, the reviewer suggests to organize better this section   according to the aim of the paper and to keep only the comparisons that are focused   on the aim of the paper.

The authors apologize for not addressing this comment. Done as per   request of reviewer. Discussion updated and many parts are omitted in the   updated manuscript.

15

For instance, is table 6 real necessary or could it be deleted? Could   tables 4 and 5 be jointed?

The authors apologize for not addressing this comment. In the updated   manuscript, tables 4 and 5 are deleted and information about critical   velocities has been placed in text.

16

-   the calibration methodology of the equations should be explained in the   manuscript

Done as per request of reviewer

17

-   references should be checked in order to be consistent with the citations   made in the main text. Moreover, the same style for references should be used   throughout the paper

Done as per request of reviewer

18

-   check language at lines 20-21, 67, 113-114, 115-116, 533. In general, check   for typos and grammatical errors

Done as per request of reviewer

2

-      The   coordinate system in the figure is not correct described. Please correct   this. Line 172 x in positive main direction ok but y point AWAY from the side   with the breach and currently the z points downwards. Please check this very   very carefully and look up Right -hand rule. 

Done as per request of reviewer

3

-      Figure   1 delete the gate direct after the honey comb. This one is not used and only   makes it confusing + add a cutting through the section of the breach, so that   it is clear how the ground is model in the breach case. 

Done as per request of reviewer

4

-      The   photo is missing and was not added. Why? It would be very easy and would help   to understand the set-up. 

Done as per request of reviewer

5

-      The   reviewer agrees with the authors, that the flow in the breach is more complex   than in a straight direction and that is exactly the reason, why he/she would   have suggested, that the investigations are first conducted in a simple flow   and in a second step reproduced in the breach.

This   would reduce the uncertainties of the investigation significantly.

Please   add this at least in the discussion part and argue why it was done in this   way.

Done as per request of reviewer

6

There   are a lot of assumptions made and those have to be discussed in a separate   section and not only in two sentence in the conclusion.

We thank reviewer for this comment. We have added assumptions in   experiment description.

7

-      There   should not be a figure 4a (a)!

Done as per request of reviewer

8

-      The   current Fig 4a und 4b should show the similar design and axis. Please also   make walls visible. 

Done as per request of reviewer

9

-      Reference   vector in 4a is missing. 

Done as per request of reviewer

10

-      22   cm ball and no deformation? If this is a perfect ball, why should this bag   stand still for only one second and not roll away.

The   reviewer should think carefully about the formulation of the answers and   clarify this also in the text.

The   work in this study deals with sandbags, rather than perfect geometric prisms,   i.e. the circular sandbag is not a perfect sphere/ball; or else, it would not   imitate the actual sandbag used in real sites. This work followed procedure   performed by Zu et al (2004) and Gulcu (2009), Gogus and Defne (2005), and   Defne (2002) by defining the geometry of a sandbag. Where a, b, and c are   dimensions of a prism enfolding sandbag. This has been added to the updated   manuscript.

11

-      Do   the authors have the permission of the original author of the picture in   Figure 7b? 

We thank reviewer for this comment, this picture is referenced/ originally   used by the first author in Sattar (2009) paper and cited here with   permissions from the journal.