Dear Filippo, you are talking about very different tools and as all numerical models there is not the "medicine for all illnesses". First of all, you must realize that none of the tools available at the moment can cope with a 10 year simulation over 25 km of coastline! For sure it is not very feasbile with 2-d or 3-d computational domains, where even building a grid over 25 km would make you miss all coastal features, unless you have a 25 km straight coastline, which I doubt. Over such a space you are forced to work in 1-d mode and for this I think Xbeach is far better that other models at least for dune-beach erosion. The fact is that to run 10 years of waves will not be easy, unless you play around a lot with time-scaling, which may make the mode unstable. If the 10 years of waves means instead a wave climate representative of those 10 years, that is a different thing. I hope it helps

dear Paolo,

thank you very much for your timely and illuminating answer. Yes, I guess that I would need to stick to 1-D modeling and then use 2D only for small areas of interest. By the way, the coast is in effect almost completely straight, but I guess this does not solve my problems. We are mainly interested in dune and beach erosion in fact so that probably we will try to use Xbeach.

As for the simulation time again you are right, it has to be intended as using a climate representative of 10 years. The problem is, I actually don't know how to follow this approach using Xbeach (or the various modules of the Delft3D code which work also at 1D and 2D).

Hello Filipo,

While I generally agree with Paolo, I disagree that 10years of a 25km can not be simulated. The tool to do this is in my best opinion MIKE21-FM latest version.

As mentioned earlier, the choice of the type of model should be based on the type of problem you try to investigate/solve. While at the moment MIKE21 or MIKE3 are commercial and quite expensive, DELFT3 is now open source in respect to the hydrodynamic, waves refraction and sediment transport modules, though my past experience with it was of less satisfactory documentation, but I don't know how much it has improved since the 2005-6, so it may be OK now. X-beach developed also at Delft Hydraulics (now Deltares) by Prof. Dano Roelving is specially designed for long term sediment transport studies, but is of less detail and more schematization than the previous two mentioned. One warning about Swan - from our experience it needs to be properly calibrated for your coast (as a matter of fact also the others) and does not directly simulated directional spectrum wave refraction, and the area near a breakwater may be not properly resolved in regards to the radiation stresses obtained there. So, if you can describe the problem you try to solve, than I might be able to advice for the most suitable model in terms of use, and cost. In any case, the results would depend on the quality and extent of your calibration and validation, for which suitable data on the boundaries and for the initial conditions are imperative.

Dear Filippo, my recommendation goes to use the "right tool, for the right job". That means that you can probably use any of the above, but the answers and the modelling approach shall be different. You won't be doing 10yr simulation for a 25km coast with Delft3D (which, btw, is open-source also) - in this case you'd perhaps do several simulations with different wave conditions and "get a feel" of what would be the coastal evolution tendency associated to each of those.

From what I recall seeing until last year, XBeach has also not been used with such a goal. Again, it has been used for shorter periods.

1D shoreline models are the ones that, so far, have been used more widely for the simulations you describe. They require calibration data usually in terms of past shorelines. Good luck.

Hello again,

Just another remark. Neither Delft3D nor MIKE21 (or MIKE3) are not run in the 3D mode for sediment transport along the coasts. They are run in a quasi-3dimensional mode known as 2Dh, whereas the vertical flow profile is considered known/selected for each XY location.

Good morning, Filippo. It is a little bit old information, but you can check a paper on beach and dune erosion; http://coastalresearch.sakura.ne.jp/nishi24/hpdata/sediment/dissert/dissertrn.htm

Nishi,

Hi to all, just a quick reply as myself and my team extensively used Xbeach in the MICORE project (www.micore.eu). On my research gate you find our recent papers on the topic.

This is just a couple from my group's papers but many more were published during the project and beyond by other project members

10.1016/j.geomorph.2013.04.034

10.1016/j.coastaleng.2013.02.006

Xbeach, being specifically developed for storm erosion, has better physics compared to "commercial competitors", especially for the long period-low frequency wave processes. However, the long-term type of modelling you want to do, I believe it is hard to be done, first it is too demanding as computational requirements. Second, all profile models do a good job for erosive high energy-condition but struggle with recovery during low energy. You may realistically run simulation for 1-100 return period events, but not for years and years of hourly wave data. Infact, you admit you will try with a wave climate. Bear in mind that over 25 km you will have to do reasonable wave propagation from deep water conditions, having the buoy pretty close to your computational domain.

For what you want to do you may be better off with an area model, to be realistically run in 2D mode. A 3D mode may require a coarse grid, a number of variable grid types and difficult boundary conditions. You may do some 3D modelling but you may need to go into parallel code compilation. MIKE 21 and DELFT3D are not the only solution, we have used an alternative open source code, ROMS coupled with SWAN with reasonable results under parallel compilation. Have look at this paper, which deals with a much more complex problem than a straight beach

http://journals.tdl.org/icce/index.php/icce/article/view/6721

Besides, I do not believe you have a real 25 km long straight long beach. Strictly speaking it means that you have a 25 km long regular bathymetry so that you can dismiss grid types needed to resolve rhythmic nearshore bars and so on. If a type of beach like that exists in Italy it would make it worth entering it into this list

http://amazing-funfacts.blogspot.it/2011/04/15-of-longest-beaches-in-world.html

Finally, I close the discussion supporting Francisco Sancho's comment, use the right tool for the right job. Maybe a 1-d shoreline model is more appropriate to you case.

Good luck

Thank you again Paolo. I will try to follow your suggestions. Of course I do not need 3D modeling. I never wanted to do this in the first place. Too many data needed... And the 25 km is the extension of the wider area. The modeling probably should be run over much less.

Hello,

I use MIKE21 for Mediterranean coastline in South France. I have got good results.

V.V.Than

Mike 21 is best

- XBeach: more detailed, but for small areas.

- Delft3D + SWAN: 4th generation wave model, good for reflective beaches.

- MIKE21 SW: based on mild slope equation, good enough for reflective beaches.