open TELEMAC-MASCARET The mathematically superior suite of solvers

Thank you very much Jean-Michel. I think we have created an ice-model within the Telemac:). Well, you mostly did everything. Thank you very much. I will keep calibrating this case until I get good results. I will keep you posted. Thank you very much for your time and your valuable advice. I will be back:).

Kind Regards!

Violeta

Dear Jean-Michel,

I hope you are doing well and looking forward to Christmas break! :) Thank you again for all your help with this case. Technically it is running fine, but needs more calibration.
I wonder if you could give me an advice. I have been playing around with values of ATABOL and BTABOL and reduced salinity and temperature values at the "closed wall". However, since the difference in values of these parameters between the boundary nodes and next to them in domain is huge, then this causes huge pressure gradients difference. And as a results it generates some velocities. It is probably not far from reality, but is there any other parameters that I can adjust to reduce the velocities and also some surface level jump has been observed at a closed wall. I have attached some pictures from the results. Please advise. Thanks a mil.

Kind Regards!

Violeta

Hello,

Only turbulence can reduce the high gradients of salinity and temperature, so a higher diffusion should settle the velocities. What you could also look at is the density that you give for RHO0, it should be an average between the minimum and maximum density you have, so that delta(RHO) is not too big (this will have probably little effect but well, we use the Boussinesq approximation which linearises 1/RHO around RHO0, hence the choice of RHO0 is not innocent).
Then of course your coefficients ATABOL and BTABOL are parameters, but they should stem from a theory and the theory should stem from the knowledge of what is behind the solid boundary. They are not the same for salt and temperature. If your solid wall is a glacier, you cannot get fresh water without melting the glacier, this sounds a nearly negligible long term effect, whereas the water along the glacier can be cooled rapidly by contact, is that the basics of what you do?

With best regards, and merry Christmas...

Jean-Michel

Dear Jean-Michel,

Thank you very much for valuable advice. I will play around with diffusion coefficients and look at the density value too.

I should have explained properly what I am doing from the beginning:). In current case, I am trying to "trigger" the correct circulation in stratified environment without tide, Coriolis and wind forces. Basically, at a glacier wall the less dense water flows out on the surface, and the warmer saline water flows in the lower layer. Since my mesh is huge, and I am using the full size of the fjord, then the main goal is the circulation and mixing processes in the fjord. My colleague is working with plume model and ice melting, so his mesh is tiny and his domain is very small, and he can concentrate on details. For my case, the solid wall roughly could represent the glacier, but the liquid boundary points ( the ones you helped me to code in bord3d) can be the vertical layer between the glacier and domain, so it is like melted mixed water, but still cold and fresh enough. When it is working fine, then I will add tide, wind and Coriolis. I have already run simulations with these parameters, but in some cases (just with tidal forcing) the water in domain become much warmer then at an open boundary due to mixing, so the fresher water was flowing into fjord on the surface at open boundary.
The final missing element would be the constant source of fresh water to recreate the right circulation. I can add discharge of fresh water at a glacier wall as well and that would represent the "summer scenario" , but in "winter fjord scenario" there is believed to be a quiet discharge free period. I am not a field worker, so I just follow people's observational reports.
The only my concern would be, how model would behave when I join all parameters together in one run. So far I have good cases with tide, Coriolis and wind, all in one simulation. Also I have cases with discharges (3 so far). So when "glacier wall" case is producing good results, and I join all other pieces together, then I hope the model will still produce "realistic" scenario. I know that it is preferable to keep things simple, and I am doing opposite:).
Thank you very much for your help, it is very much appreciated.

I will do as per your advice, and hopefully velocities will be reduced. I thought that it might require some time for velocities to settle down so I will run it for longer time as well.

Thanks again. I will keep you posted.

Have a lovely Christmas and Happy New Year to you and all your family!

Kind Regards!

Violeta