Tsunami modeled with SWAN

[Benjamín]

Dear all,

First of all, I'd like to thank all the people that have worked in
developing this truly useful tool of Mirone, wich we're starting to
use in our bureau.

Now my question. We're modelling a tsunami, we created the initial
conditions with the Okada method and want to propagate it with SWAN.
I'm wondering, is there any way to change the boundary conditions at
the sea, so they can act like an open boundary? Because it seems that
by default, they act like walls, and therefore reflect the tsunami
towards the coast.

If anybody have any idea (beside modeling with a super gigantic grid,
so the tsunami won't touch the sea boundaries) I'd very ver thankful.

Thanks in advance.
Best regards,

Benjamín Carrión
Hydraulic Engineer

[jluis]

Dear Benjamín,

I tried to play with the boundary conditions long time ago when first
translated the original SWAN fortran code to C, but with no success. I
think Rajaraman did it as well and with the same result.
The only option I know to those annoying false reflection on grid
borders is to use the "sponge" technique.
Use the "Grid Tools -> Hammer grid -> Pad to zero" option to cosine
taper the grid to zero (both the bathymetry and the source grid). The
idea is that the wave will slow down to nearly zero speed when it goes
up the very shallow depths and won't never has time to reflect.
However, I have no tool to crop out the padded region. You will have
to do it externally with GMT's grdcut program.

Joaquim Luis

[Mr patel vikram]

dear sir,

do you have equations of mansinha code for initial vertical jump of waver wave. the reference paper i need for calculation.

regards,

vikram patel

Your Mail works best with the New Yahoo Optimized IE8. Get it NOW! http://downloads.yahoo.com/in/internetexplorer/

[Benjamín]

Thanks Professor,

I'll try what you say (I didn't know that option in the program)
Anyway, I don't understand why so many of this coastal models don't
have implemented (as far as I know) effective open sea boundary
conditions, since there are available many proved techniques that are,
in fact, not difficult to implement, so we're forced to try this
numericals approximations, like sponge layers, that many time have
their own parameters to be tuned, which may affect the results.

Anyway, thanks again for your help, and for developing this very
interesting tool.
I'll try it that way and post the results.

Regards,
Benjamín Carrión
Hydraulic Engineer

[jluis]

Benjamín,

Please drop that "Professor" thing. A good thing of internet is that
it's informal.

You may want to try COMCOT. It has good boundary conditions, but on
the other hand the last version I tried had bugs, which I reported to
authors but had no feed-back. Well, actually the mail bounced back
some days later.

You can also try your luck with SWAN. The swan.par parameters
indl,...,indt are supposed to control BC

Joaquim

[jluis]

Vikram,

You already asked this question 4 months ago
http://groups.google.com/group/mirone/browse_thread/thread/51f6b9d4aa01763f?hl=en

Also, please do not mix subjects on the same thread. It makes it very
confusing for later references.

Thanks

Joaquim

[Benjamín]

All right then, Joaquim

I've tried with the "continuative" boundary conditions on SWAN, but
anyway it appeared the spurious oscillations in the borders.
Anyway, the grid is large enough, and I'm interested in the first wave
arrival, so I can't stop the simulation after that.

Which leads me to another question. I'd like to run the Tsun2 module,
but I need to feed it with a parameter file .par, which is mentioned
in the manual, but I can't find it, or a copy, like the swan.par. I
searched on the internet but didn't find anything.

What is the format of this file? What should I write on it?
Thanks again

[jluis]

Like I told recently to Vikram, the Tsun2 approach I have is now
outdated. The big problem with it is that there is no transmission of
momentum (only wave heights) between SWAN and Tsun2. Anyway, I try to
explain how to proceed in the Tsun2 section of the manual. the .par
file has to be generated for each case.
I have this old (figures we a bit different back than) little tutorial
that you can use to practice for other cases

http://w3.ualg.pt/~jluis/mirone/Tsunami_inundation_tutorial.zip

For inundation studies my recommendation is to learn and use ANUGA.
Mirone can create initial condition in Anuga format from the SWAN
runs. Furthermore, and unfortunately a non documented feature, if
there is one rectangle delimiting the region of interest for the
inundation area, Mirone will create the .sww file (the Anuga format)
for that region only. This mkes it much easier to set up Anuga.

Good luck

Joaquim

[Benjamín]

Joaquim,

I've downloaded ANUGA and I'm starting to learn how to use it, but for
sake of time (always in a hurry) I ran the Tsun2 module, since it
would be a good approximation, anyway, nevertheless the lack of
transmission of momentum, but when I try to write files with the
components of velocity, the program simply doesn't do it.

I assume that tsun2 does compute the velocity in all the grid,
independently of boundary condition... or, am I wrong?
The alternative is to learn ANUGA, but I'm afraid it could take me a
lot of time.

Thanks again with any advise respect this problem with tsun2.

Saludos

[jluis]

Benjamin,

I converted Tsun2 some 7 years ago and at that time I had no need for
the velocity grids so, and since the original code didn't do it
either, it doesn't provide and way to save the velocity grids.
Actually neither SWAN used to do it. I added that option because I
needed to set up the initial conditions to Anuga. If you really need
it, I'm afraid you'll have to learn how to make Anuga work. I kinda of
forgot it but with Mirone's help it's not very difficult.

Cumprimentos

Joaquim