the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Implementation of additional spectral wave field exchanges in a 3D wave-current coupled WAVEWATCH-III (version 6.07) – CROCO (version 1.2) configuration and assessment of their implications for macro-tidal coastal hydrodynamics
Gaetano Porcile
Anne-Claire Bennis
Martial Boutet
Sophie Le Bot
Franck Dumas
Swen Jullien
Abstract. An advanced coupling between a three-dimensional ocean circulation model (CROCO) and a spectral wave model (WAVEWATCH-III) is presented to better represent wave-current interactions in coastal areas. In the previous implementation of the coupled interface between these two models, some of the wave-induced terms in the ocean dynamic equations were computed from their monochromatic approximations (e.g., Stokes drift, Bernoulli head, near-bottom wave orbital velocity, wave-to-ocean energy flux). In the present study the exchanges of these fields computed from the spectral wave model are implemented and evaluated. A set of numerical experiments for a coastal configuration of the circulation near the Bay of Somme (France) is designed. The impact of the spectral versus monochromatic computation of wave-induced terms significantly affects the hydrodynamics at coastal scale in the case of storm waves and winds opposed to tidal flows, reducing the wave-induced deceleration of the vertical profile of tidal currents. This new implementation provides current magnitudes closer to measurements than those predicted using their monochromatic formulations, particularly at the free surface. The spectral surface Stokes drift and the near-bottom wave orbital velocity are found to be the most impacting spectral fields, respectively increasing advection towards the free surface and shifting the profile close to the seabed. In the particular case of the Bay of Somme, the approximation of these spectral terms with their monochromatic counterparts ultimately results in an underestimation of ocean surface currents. Our model developments thus provide a better description of the competing effects of tides, winds, and waves on the circulation of coastal seas with implications to the study of air-sea interactions and sediment transport processes.
- Preprint
(17128 KB) - Metadata XML
- BibTeX
- EndNote
Gaetano Porcile et al.
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-715', Anonymous Referee #1, 20 Jul 2023
General comments:
This paper presents a new coupling procedure between the 3D ocean circulation model CROCO and the spectral wave model WAVEWATCH-III, handled through the OASIS-MCT coupler, in order to improve the representation of wave-current interactions in the coastal region. More precisely, efforts are made to incorporate wave quantities derived from the full (bi-dimensional) wave spectrum that better represents “real” directionally and frequency-broad wave fields. This is in contrast with “monochromatic” approximations that use a representative wavenumber (e.g. mean or peak) as originally used and implemented in ROMS (Uchiyama et al., 2009, 2010).
After a broad overview of the different approaches for simulating wave-current interaction in the coastal context, the paper describes the new forcing terms. The rest of the paper focusses on a series of numerical experiments performed in the geographical setting of the Bay of Somme, France, a macrotidal site located in the English Channel. In situ data (vertical current profiles at two locations) are used to assess the model performances as well as evaluate the added value of the newly-proposed coupling procedure.
I found the manuscript relatively well organised, though not particularly well written. At least, it could have used a few more readings by the authors: e.g. some internal notes made by the authors remain in the core of the manuscript (see at lines 102-107), which is not acceptable in my opinion. The efforts to incorporate full spectral representation of wave quantities are welcome, as it should help to get a better and more realistic representation of wave forces while modelling wave-current interactions with this modelling system. The (scientific) novelty is not obvious, though, since similar efforts made by other authors in ROMS are overlooked, and such spectral representation already exists in other modelling systems. So it is not clear to me whether this contribution justifies a publication or not. I have several major comments on the present work that, I think, should be addressed before a resubmission:
- The manuscript completely ignores many relevant and recent works that have looked into these aspects, some within the frame of the same modelling system. Within ROMS, from which CROCO is derived, the works by L. Romero and the UCLA team are completely overlooked (in particular Romero et al., 2021; Hypolite et al., 2022). This also includes the work initiated by N. Kumar (Kumar et al., 2017; Liu et al., 2021) that has evaluated the impact of the definition used for the Stokes drift velocities on the modelling of wave-current interactions at regional scales. The present manuscript also ignores modelling systems already incorporating full spectral wave quantities such as MOHID (Delpey et al., 2014) or SCHISM, which has been used to simulate and analyse the wave-induced nearshore circulation in numerous realistic settings at both regional (Guérin et al., 2017; Lavaud et al., 2020; Pezerat et al., 2022) and local (Martins et al., 2022) scales.
- The present numerical setup is far from being ideal to discuss the added value of (full) spectral representation of wave quantities, for several reasons. The hydrodynamics simulated here is overly controlled by the forcing used at boundaries, both in terms of water levels, currents and waves, to the point that the reader asks himself what is the added value of using a local model. For instance, the wave results at the measurement site show no improvements over the large scale model used as forcing. Furthermore, water levels and currents are (to my understanding) forced from a 2D hydrodynamic model. Considering the importance given to the vertical shear, how can we be sure that the boundary is taken sufficiently far away from the location where measurements were obtained and vertical velocity profiles discussed? I suspect that the boundary is too close to the measurement stations in order to discuss differences of the order of 1 cm/s at the top layer of the water column.
- The discussion of wave processes, particularly in shallow water depths, is in my opinion weak throughout the paper. Discussion on depth-induced refraction (short waves in relatively deep water depth) is not supported by the numerical results, while the section dealing with wave breaking in shallow water (set-down and setup) is really weak, to say the least. Reading that wave effects are “removed” in depths less than 2 m (why is that so?), how can we then expect a wave setup to form near the shoreline? The influence of using a spectral representation is expected to be much more important in wave-dominated shallow water environments, so we expect this study to also consider such environments where this model has already been applied (e.g. Duck N.C. in Uchiyama et al., 2010; or the Biscarosse site in Marchesiello et al., 2015).
- Instead of a lengthy description of the field site, and given the chosen journal, we expect much more description on the coupling procedure: where there files modified in the wave or hydrodynamic models or can new versions of these models be pulled directly? Can the new coupling procedure be reproduced elsewhere easily? What about interpolation of the different fields and where is it handled?
In conclusion, I would advise the manuscript in its current form against publication and would suggest a future resubmission that address all four major points listed above. A list of minor comments or suggestions are also provided below and should help the authors to improve their manuscripts or make it clearer.
More specific comments:
Abstract
Lines 7-8: given the presented results, and the comments above, I think that the “significant” currently needs to be toned down. Considering nearshore situation with adequate numerical experiments could surely better highlight the impact of using (full) spectral wave quantities.
Line 15-16: what about nearshore environments, where Bernoulli heads and breaking acceleration terms will be the dominant forcing of the nearshore circulation?
Line 21: “wave-induced currents and setup” are not forcing terms.
Line 26: reformulate with “wave-dominated environments”?
Lines 34-35: “ three-dimensional modelling of the wave-induced flow was requested”. I think that the authors can find much better context and needs for resolving 3D currents in the nearshore context (e.g. the transport of sediment, particles or tracers).
Line 35: specify what these theories are used for.
Line 40: “with simplifications”. The end of this section feels too light, and considering the scope of the paper (the added-value of spectral estimates), more details should be provided with a more precise view on the state of the art regarding the modelling of wave-current interaction in 3D (cf. my general comment above).
Line 46: I do not think that “initiated” is correct, given that most of the work come from that of Uchiyama et al. (2009, 2010). The study by Marchesiello et al. (2015) brings no new development on the wave-current interaction part, and the authors do not even mention the CROCO modelling system.
Line 52: again, what do the authors mean? This is literally the model developed in Uchiyama et al. (2010) and used in CROCO.
Lines 52-55: if the model used in Marchesiello et al. (2015) is so accurate, why not using their configuration to test the added-value of the new developments in the context of the nearshore region?
Section 2
This section feels quite tedious to read, with too many and sometimes irrelevant details given the scope of the present study. I think that the authors need to reshape this section and keep the most relevant information only.
Line 86: “[…] a not fully developed wind sea. The energetic swell are generated on limited fetches with a maximum length of 400 km”. What do the authors mean? Swell developping over 400 km-long fetches?
Lines 102-107: a careful read is the minimum to provide before submission and review.
Line 110: “2 meters above the bottom”, so not over the whole water column as stated?
Description of WAVEWATCH-III: to my understanding, this model solves the evolution of directional wave spectra in the wavenumber space. Why is this equation different from that given in the model manual?
Line 140-141: please clarify what wetting and drying means for spectral wave modelling.
There seems to be a typo in Tm01 in much of the manuscript, even the schematic of Fig. 2.
Line 287: similar to earlier comment, what do the authors mean here? the wave setup or set-down are not forcing terms but result from the wave forces.
Section 3.5
Why using such a small scale model, which depend so strongly on wave and tidal forcing used at the boundaries? In particular, the tidal forcing is derived from a 2D hydrodynamic model, how does that affect the vertical shear observed at the instrument positions? The discussion of the results presented lated in the paper depends too heavily on this aspect for being ignored and not investigated.
Lines 310-311: why such drastic choices? Can this model be reliably used in a nearshore context? Or is it suffering from instabilities? In any case, given this information the authors simply cannot discuss shallow water processes like wave setup.
The impact of surface waves on the wind drag coefficient is not described. Have the authors investigated this? How did they make sure that their tests are all consistent with this regard?
Line 331: why is “fresh” relevant here? I might be missing some specific term here, though.
Lines 333: why switching to Hrms, is that for reducing the errors of the model? Though the authors consider their model “validated” (line 337), at the peak of the event they consider, half the wave energy is missing in the model. In my opinion, this is far from being satisfactory, given that the authors then discuss added-value of the order of 1 cm/s in terms of currents, without proper sensitivity analysis of other relevant parameters (wind drag coefficient, mixing length, amount of tke injected in the water column and so on).
Line 343: what is a wavenumber velocity?
Line 358: this statement is clearly not supported by the material presented in the manuscript.
Section 4.2
Line 360-363: to me, it simply shows how dependent the local model is to the forcing, and that there is a systematic low bias at every high tide (half a meter). Then, the authors provide some explanation to the observed phase shift, while it clearly comes from the forcing. How does the original hydrodynamic model (that served to derive the forcing) compare with the data?
Line 382-383: not really true since a better match is systematically obtained with the previous configuration during ebbs.
Line 390: “[…] due to the presence of sub-gris scale bedforms”. Again not supported. Please check the forcing.
Section 4.3
Line 408: how is wind drag coefficient computed, and how did the authors make sure all simulations are consistent with this regard? i.e. is the wind contribution similar when waves are accounted for?
Lines 413-414: please be more specific. What about the impact of the mixing length and its parametrisation (end of Section 3.3.5) on the current magnitude at the top of the water column?
Line 433: I rather see a 1 and 3 cm/s increase dependening on the configuration.
Line 435: “[…] improves the accuracy of the results”. Given the results, this is all relative, i.e. whether taken at the surface or bottom. What are the rmsd for depth-averaged values and how do they compare?
Line 437: why jumping over 3 Figures? Also, this specific graph shows velocities (m/s) and heights (m) over the same axis, a different one should be used.
Line 447: given that the boundary remains close to the measurements location, too much uncertainty exists on whether or not the model develops a realistic 3D profile (i.e. taking into account the response to winds etc). In this context, the statement is speculative and should be investigated further.
Line 443-444: what do the authors mean here? Please develop.
Line 451-452: so why choosing this limited experimental dataset or focussing on the upper section of the water column?
Line 465: not supported. Using more carefully designed numerical experiments, including synthetic cases would be extremely useful here.
Line 472-473: here, I think it is misleading to compare the wave breaking terms for bulk and full spectral quantities. The bulk representation has so far only been derived and used for the case of depth-induced wave breaking (Uchiyama et al., 2010; Kumar et al., 2012) and not for whitecapping, which is the dominant dissipative process for waves in the present case. Whitecapping affects mostly high-frequency components, so the comparison is not relevant in my opinion. This made me realise that the authors did not describe the full spectral representation of the wave forces (not that of dissipation).
Line 490: this is all relative, we speak of 0.5 cm/s differences while wave model performances are poor and RMSE for currents is ~10 cm/s, with much larger errors locally.
Fig. 17: please clarify the representation of water levels as it is not clear. Also, in some Figures, the vertical datum for depth is not clear.
Line 497: is that so? Unless there are inconsistencies in ROMS/CROCO, the peak wavenumber used in the monochromatic formulation should be Doppler-shifted too in full-couped configurations.
Line 505: please be consistent. In some sections of the manuscript, refraction processes are well represented and in this one, they are likely misrepresented. In both cases, the analysis is poorly supported by results.
Line 506: so why not increasing the spatial resolution of the model? And why analysing processes such as depth-induced refraction and breaking?
Line 515-520: how is the surf zone identified here? Does that include region where whitecapping occurs? Like many sections and physical interpretations of results, this discussion is poorly supported.
Line 544-545: I struggle to understand and believe the key message here.
References:
Delpey, M.T., Ardhuin, F., Otheguy, P., Jouon, A., 2014. Effects of waves on coastal water dispersion in a small estuarine bay.J. Geophys. Res. Oceans 119 (1), 70–86.
Guérin, T., Bertin, X., Coulombier, T., de Bakker, A., 2018. Impacts of wave-induced circulation in the surf zone on wave setup. Ocean Model. 123, 86-97.
Hypolite, D., L. Romero, J. C. McWilliams, and D. P. Dauhajre, 2021. Surface Gravity Wave Effects on Submesoscale Currents in the Open Ocean. J. Phys. Oceanogr., 51, 3365–3383.
Kumar, N., D. L. Cahl, S. C. Crosby, and G. Voulgaris, 2017. Bulk versus Spectral Wave Parameters: Implications on Stokes Drift Estimates, Regional Wave Modeling, and HF Radars Applications. J. Phys. Oceanogr., 47, 1413–1431.
Kumar, N., Voulgaris, G., Warner, J.C., Olabarrieta, M., 2012. Implementation of the vortex force formalism in the coupled ocean-atmosphere-wave-sediment transport (COAWST) modeling system for inner shelf and surf zone applications. Ocean Model. 47, 65–95.
Lavaud, L., Bertin, X., Martins, K., Arnaud, G., Bouin, M.-N., 2020. The contribution of short-wave breaking to storm surges: The case Klaus in the southern Bay of Biscay. Ocean Model. 156, 101710.
Liu, G., Kumar, N., Harcourt, R., & Perrie, W., 2021. Bulk, spectral and deep water approximations for Stokes drift: Implications for coupled ocean circulation and surface wave models. Journal of Advances in Modeling Earth Systems 13, e2020MS002172.
Marchesiello, P., Benshila, R., Almar, R., Uchiyama, Y., McWilliams, J. C., and Shchepetkin, A., 2015. On tridimensional rip current modeling, Ocean Model., 96, 36–48.
Pezerat, M., Bertin, X., Martins, K., Lavaud, L., 2022. Cross-shore distribution of the wave-induced circulation over a dissipative beach under storm wave conditions. J. Geophys. Res. Oceans 127 (3), e2021JC018108.
Romero, L., Hypolite, D., and McWilliams, J.C., 2020. Submesoscale current effects on surface waves. Ocean Model.. 511, 135-178.
Uchiyama, Y., McWilliams, J. C., and Restrepo, J. M., 2009, Wave-current interaction in nearshore shear instability analyzed with a vortex force formalism, J. Geophys. Res., 114, C06021.
Uchiyama, Y., McWilliams, J.C., Shchepetkin, A.F., 2010. Wave–current interaction in an oceanic circulation model with a vortex-force formalism: application to the surf zone. Ocean Model. 34 (1), 16–35.
Citation: https://doi.org/10.5194/egusphere-2023-715-RC1 -
RC2: 'Comment on egusphere-2023-715', Anonymous Referee #2, 01 Aug 2023
This is an excellent manuscript. In this work, the authors discussed the implementation of WW3 and CROCO. They tested the coupled model using a representative case near the Bay of Somme and compared the simulation results against available observations. The paper is very well written. The experiments are designed appropriately. The equations used in the model look correct to me, but I did not derive the equations.
My major concern is, the simulated wave variables from the experiments are not different by a lot considering the feedback of the current velocity (Fig. 3, 7, 8, 9). Sometimes the new implementations are worse compared with the previous versions. Does it mean the fully coupled model has larger errors?
It would also be desirable if the authors can discuss the extra computational cost introduced by the new implementations.
Here are some minor suggestions from me that may be helpful for the authors:
- Line 79: why are the exceptional tides not simulated in this case?
- Line 93: is the CROCO model adequate to resolve the small dunes? I think the small ripples/dunes can be as small as a few centimeters (~150 d50).
- The text between lines 105-106 should be cleaned up.
- Line 125: it would be helpful if the author could add more about this equation. What is the unknown variable? How is the equation solved? How are c_theta and c_w represented in the equations?
- I think Eq. (2) are the equations actually solved in CROCO, why don’t the authors put this part in Section 3.2?
- Line 164: I think the monochromatic approximations are introduced in the latter sections. How about “... are computed from their monochromatic approximations, which are introduced in the latter sections”?
- Line 309: The wind forcing resolution seems too coarse compared with the model resolution. Can the authors comment on this gap?
- Line 311: I am not sure about the meaning of mean grain size. Is it d50 (mass median diameter)?
- Line 413: Why does the flow fit better with observations?
Citation: https://doi.org/10.5194/egusphere-2023-715-RC2 -
CC1: 'Comment on egusphere-2023-715', Chu-En Hsu, 02 Aug 2023
Respected colleagues,
We appreciate your attention and time. While the authors did an excellent job of presenting the wave-current coupling configuration using WAVEWATCH-III and CROCO, I would like to draw attention to Hsu et al.'s (2023; https://doi.org/10.3390/jmse11061152) relevant analyses and discussions on wave model performance using a different modeling system (SWAN-ROMS coupling under COAWST) during various tropical cyclones. During hurricanes with various storm characteristics, the evolution of wave energy spectra and model performance for wave energy distribution were documented. We identified and investigated scenarios in which model performance might not be as good as it could be.
Our research concentrated on three historical storms that had an impact on the East Coast of the United States: Matthew (2016), Dorian (2019), and Isaias (2020). It will be great to see if these conclusions can be developed upon and explored further using this new modeling framework, in our opinion. It may also be interesting to observe if specific tropical cyclone characteristics will have an impact on the model performance of wave energy distribution, even if the authors noted that the performance of 3-way coupling was already assessed for the case of Tropical Cyclone Bejisa (Pianezze et al., 2018).
Best wishes, sincerely,
Chu-En Hsu
Citation: https://doi.org/10.5194/egusphere-2023-715-CC1
Gaetano Porcile et al.
Data sets
Bay of Somme coupled CROCO (v1.2) - WAVEWATCH-III (v6.07) configuration files used in Porcile et al. (2023) Gaetano Porcile, Anne-Claire Bennis, Martial Boutet ,Sophie Le Bot, Franck Dumas, and Swen Jullien https://doi.org/10.5281/zenodo.8046629
Model code and software
Bay of Somme coupled CROCO (v1.2) - WAVEWATCH-III (v6.07) configuration files used in Porcile et al. (2023) Gaetano Porcile, Anne-Claire Bennis, Martial Boutet ,Sophie Le Bot, Franck Dumas, and Swen Jullien https://doi.org/10.5281/zenodo.8046629
Gaetano Porcile et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 217 | 64 | 14 | 295 | 6 | 7 |
- HTML: 217
- PDF: 64
- XML: 14
- Total: 295
- BibTeX: 6
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1