open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I would like to let SISYPHE run based on hydro (T2D) and wave (TOM) files resulted from a previous T2D-TOM coupling run. SISYPHE is then supposed to calculate at each time step the wave-current transport and solve the Exner equation only. So, there is not coupling at all, just chaining.
Is that possible?

I see both T2D and TOM files can be included into the computation (without coupling) but only the last time step for waves is considered. Ideally, I would like all time steps to be included, exactly like for the current file.

Thank you,
JBS.

Hi JBS,

I suggest that you seriously reconsider this approach. Especially with waves included the bed changes will lead to unrealistic situations.
It is possible (I have done it in the past), but it would require quite a bit of Fortran coding. Moreover, it becomes almost impossible to get an accurate estimate of the wave orbital velocities, which are critical for the sediment transport.
Setting up the 3 way coupling is much easier to set up.

Thank you Michiel,

I forgot to mention the following: my runs are of very short duration (a couple of days) as they should each represent a given single storm event. The morphological changes are neglected accordingly. That is why I am opting for such a solution.

What do you think?

JBS.

According to the main subroutine (sisyphe.f), a simple solution coulb be to add wave results (HM0, DMOY, TPR5) into the T2D output variables. Then SISYPHE reads them when calling 'READ_DATASET' in the hydro file and checks whether there are present or not. Then the problem boils down to including these variables into T2D output...
Would this work?
JBS.

That would be an option, but be aware that SISYPHE will produce an estimate of the orbital velocities based on Hunt's method for monochromatic waves.

If the time period is short, the 3 way coupled model will not take too long to run. It is much easier to set up the 3 way coupled system and it will be more accurate. The only down side will be the computation time, which will be longer than uncoupled computations.

If I do not go for the 3-way coupling, I guess this is ok if I output the variable for orbital velocity imported in T2D (ORBVEL) directly [It seems the WAC subroutine imports FXWAVE, FYWAVE, HM0, DIRMOY, TPR5 and ORBVEL into T2D].
Then I could copy ORBVEL into a PRIVE array (whose name would be modified in NOMVAR_TELEMAC2D).
Would it work this way?
JBS.

That would be the best way to do it. But keep in mind that by uncoupling any lowering (raising) of the bed will not automatically lead to a reduction (increase) in the orbital velocities and therefore over (under) estimate the transport rates.

If you are not interested in bed changes, consider using
STATIONARY MODE = YES
which calculates the evolution without actually updating the bed. If it still works that is. I haven't tried it in the most recent versions of SISYPHE.