Application of SWASH to Compute Wave Overtopping in Ericeira Harbour for Operational Purposes

Round 1

Reviewer 1 Report

Manuscript ID:jmse-2016654

Title: Application of SWASH to compute wave overtopping in Ericeira harbour for operational purposes

The manuscript is focused on the use of the SWASH numerical model to simulate wave overtopping phenomenon concerning the breakwater in the Ericeira harbour (Portugal).

The model has been implemented in 1D, considering two different cross-section characterized by different  rmour artificial stones: antifer and tetrapods.

Because of the covered topics, the article may be of interest to readers of JMSE, nevertheless some major and minor points should be addressed before considering it for publication.

MAJOR POINTS

-Within the Abstract, the Introduction and the Discussion, it should be declared more effectively the need to carry out simulations by using SWASH instead of NN_OVERTOPPING2 (neuronal network system) to predict the overtopping, especially in light of the obtained numerical results, which sometimes tend to underestimate the overtopping discharge (see lines 17-18 and 464).

- Why have you decided to simulate only 500 waves? Previous studies (e.g. Suzuki et al., 2017) suggest simulating at least 1000 waves to get reliable overtopping results.

In this regard, the EurOtop manual suggests simulating a range between 500 and 1000 waves, specifying that for a Nonlinear Shallow Water Equation model (like SWASH) 1000 waves could give more realistic results.

- Since no validation has been carried out, at least a grid convergence study should be performed. As reported by Suzuki et al., 2017, "the performance of the mean overtopping estimation is somewhat sensitive to the non-dimensional parameter L/dx".

- at line 155 the authors state that simulating the porosity may lead to an excessive wave dissipation. Have you tried to modify the laminar and turbulent frictional coefficients? Moreover, even when a porous structure is simulated, the effects of roughness should be taken into account (line 158).

- lines 161-163 and 213: Actually, in this way, the authors have calibrated the Manning coefficient by taking into account even the effect of the porosity. This should be declared clearly.

-line 172: It is not SWASH that does not take into account the wave direction in a 1D mode. This is a feature of all 1D models!

-line 221-224: the authors state that Equation 2 is only applicable to oblique wave attack with incident angles greater than 15°, but even Equations 3 and 4 are valid for incident angles greater than 15°. A clarification is required.

- Results: How have you computed the wave overtopping? It is not declared in the text.

- Line 246: Even (or expecially) case 13 underestimates the overtopping (-42%).

- lines 317-328: It is not clear when the proposed equations should be used. 

- lines 411-414. The authors state that according to their results it is not possible to assess if SWASH tends to underestimate the overtopping discharge or NN_OVERTOPPING2 tends to overestimate it. Then, the need to implement SWASH models within HIDRALERTA should be declared more effectively. 

- lines 460-461. Actually, this result comes from your setting! It is clear the role of the roughness, but, it this work, the manning coefficient takes into account even the effect of the porosity.

- line 463: What about the computational effort of NN_OVERTOPPING2?

- line 465: The authors state that "the simulation revealed that Rc/Hs and Sop are important variables defining the Manning coefficient...", but actually are the only variables considered.

 MINOR POINTS

-line 12: after NN_OVERTOPPING2 the term "results" can be added.

-line 32: please spell out the name of the acronym upon the first mention and then provide the abbreviation (e.g. UN)

-line 42: what do you mean by "wave disturbance"? Please, try to be more specific.

-line 58-59: The authors could be interested in the works of Liang et al., Celli et al., etc which used SWASH to evaluate different fluid-structures interactions (even with porous structures).

References:

Liang, B., Wu, G., Liu, F., Fan, H., & Li, H. (2015). Numerical study of wave transmission over double submerged breakwaters using non-hydrostatic wave model. Oceanologia, 57(4), 308-317.

Celli, D., Pasquali, D., De Girolamo, P., & Di Risio, M. (2018). Effects of submerged berms on the stability of conventional rubble mound breakwaters. Coastal Engineering, 136, 16-25.

-line 63: the term "prototype" is superfluous.

-line 68: the verb "represented" could be replaced be "represent" since it represents the friction of the armour layer.

-line 69: "... breakwater and to adjust it to the..." could be replaced by "breakwater to get the best agreement..."

-line 155: what do you mean by "wave climbing"? Do you mean "wave run-up"?

- Figure 2: Even "q" is reported. Please, specify its meaning in the caption and in the text.

- Table "3" caption: please, amend "Pofile".

-line 379: please, amend.

- line 483: What do you mean by "real data"? Are you intending experimental data?

Author Response

Please see file attached.

Author Response File: Author Response.pdf

Reviewer 2 Report

Summary

This article discusses the potential of implementing a well-validated general-purpose numerical model, SWASH, into HIDRALERTA, an overtopping and flooding forecast system. About 38 numerical simulations were run at two different profiles at Ericeira harbor, and the overtopping discharge results were compared to that from the validated neural network tool, NN_OVER-11 TOPPING2. The authors have represented the friction and rugosity of the breakwater by varying the Manning coefficient, and a set of four equations are presented to calculate it. Errors between the calculated and calibrated Manning coefficients are small, and the comparison results with the NN_OVER-11 TOPPING2 are also in agreement.

 Comments: The manuscript is clear and relevant. The experimental design is appropriate. Some specific comments and suggestions are the following.

• It seems that Equation 1 is missing two terms: Sop^3 and Sop^2 x (Rc/Hs). The coefficients a9 and a10 are provided. This needs to be clarified and corrected.
• RMSE of Equation 2 seems to be a bit high, around 13% (0.02/0.15).
• When the number of calibration data and number of coefficients in the manning coefficient expression is similar, then the relevance of RMSE error estimation reduces since the expressions with more degrees of freedom can match data well and produce low RMSE error. It would be good if authors could justify this or add more calculated coefficients using the same calibration.
• Authors may suggest possible scientific reasons behind the findings, such as the dependency of the Manning coefficient on the angle of incidence.
• A convergence check of empirical expressions at the boundaries would increase the reliability of the four expressions. 
• Figure 4 does not show values for Sop = 0.02, yet the legend is shown.

Comments for author File: Comments.docx

Author Response

Please see file attached.

Author Response File: Author Response.pdf

Reviewer 3 Report

See attached file

Comments for author File: Comments.pdf

Author Response

Please see file attached.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

This is my second review of the manuscript. 

The quality of the manuscript has been improved. Nevertheless, some minor points should be still addressed.

1) To estimate the flooded area, a 2D numerical model should be implemented.

2) As reported in the USER MANUAL, 

the command DISCH represent discharge per unit width (vector; in m2/s).

It is not clear the need to divide the sum of discharges by the computational time.

Moreover, as suggested by Sukuzi et al.,2011

the mean wave overtopping discharge q can be estimated by extracting results for layer thickness h(t) and layer speed u(t) at the end of the dike.

Suzuki, T., Verwaest, T., Hassan, W., Veale, W., Reyns, J., Trouw, K., ... & Zijlema, M. (2011, November). The applicability of SWASH model for wave transformation and wave overtopping: A case study for the Flemish coast. In Proc. 5th Int. Conf. Advanced Computational Methods Engineering (ACOMEN 2011), Liège, Belgium (pp. 14-17).

3) Continues to be unclear the need to propose 3 equations for the T profile. 

The reader could get into confusion. Why not directly propose the best between 2 and 3; 2 and 4?

4) Figure C1: q seems not to be a ratio.

Author Response

Please see file attached.

Author Response File: Author Response.pdf

Reviewer 3 Report

The Authors responded satisfactorily to my comments. Therefore, the manuscript can be accepted for publication.

Author Response

The authors thank Reviewer #3 for his comments and suggestions and are glad to hear that the manuscript can now be accepted for publication.