open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi and thanks in advance for reading.

I'm currently setting up an idealized channel in which I hope to simulate internal waves, and had a few questions after failing to find anything in the documentation.

1. Is there an easy way to impose uniform stratification in the channel?

2. Has anything been done on internal waves before with TELEMAC3D?

3. Am I able to extract an x-z cross section? Or simulate in just this plane (2D)?

Let me know if you need any more information!

Hello,

This is an interesting topic. The stratification can be set up by giving initial values plane by plane, in the subroutine condim.f.

Up to now we did tests on stratifications, e.g. their tilting under the action of wind, but nothing on internal waves. The extra difficulty will be to follow the stratification with the mesh, this is an opportunity to test the option programmed at HR-Wallingford : MESH TRANSFORMATION = 5, which is a mesh adaptation after the values of a tracer.

Extracting cross-sections can be done with Tecplot and probably many other programs, but you can do it with Postel-3D, which provides cross sections as selafin format files, which can be seen by Fudaa, BlueKenue, etc. You can look at example "stratification" in Telemac-3D, where an example of Postel-3D steering file is "p3d_stratification.cas".

With best regards,

Jean-Michel Hervouet

Thanks Jean-Michel for your prompt reply.

Postel-3D was very intuitive and I've got some ouput from it. However, my research involves forcing OSCAR (Oil spill contingency and response) with an idealized internal wave field to measure oil dispersion etc. I can manually save each frame to a .xyz file and conglomerate into a netcdf file (The input to OSCAR), but this would take some time, and I was wondering if there was a more automated method?

I've since been successful in applying uniform stratification to a flat channel with evenly distributed layers (MESH TRANSFORMATION = 3), by enforcing a linear gradient of temperature. However, as you stated the problem occurs when e.g. trying to force a barotropic tide over a Witch of Agnesi curve to generate internal waves, where there is much stronger stratification at the peak due to more closely packed vertical nodes. Is this what MESH TRANSFORMATION = 5 attempts to solve (I can't see a reference in the manual)?

One final question: can density be used as a tracer, rather than indirectly forcing it with other variables?

Many thanks for your help!

Hi
About your first point, I think you could also try to use the python scripts of Telemac to produce the results file you need. Maybe also directly in netcdf...
I don't know if the latest developement made by Sébastien at HR Wallingford are available in the trunk but I know there is a lot of possibilities to create some plots of vertical sections in 3D results...
He is also the father of the AMR method (mesh transfromation=5)...

As far as I know, density cannot be use as a tracer in the current version of Telemac.

Kind regards

Hello,

So far, but it is easy to change, the first tracer is taken for refining the mesh (this is hardcoded in subroutine calcot.f). So far also density is only a byproduct of tracers (this goes with the Boussinesq approximation). When density is due to several tracers or sedimentI do not think it can have an equation initself. However taking the density for refining the mesh is a good idea.

With best regards,

Jean-Michel Hervouet