open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear Users,

I am trying to implement the source point into simulation. Technically, it is working fine, I think, in terms of velocities at least. However, I am not sure if Salinity and Temperature values, at the source, are applied properly. I have assigned constant Sal and Temp at a source to be 0 (psu) and 0 (degreeC). No source file, just in a steering file. The source is there, it is clearly seen especially at velocity profiles, but temperature and salinity values are already mixed with the ambient waters that have salinity 31 (psu) and temp 2.76 (degree) at a source layer. The total number of layers is 30, strongly stratified with various values of Temp and Sal at each layer. I have not amended the diffusion coefficients and let it be as per default. The horizontal mesh is too coarse for the source implementation, 1km, but my goal is too see the overall source impact on the whole domain, which has full length 130km. Even element size is too big, but I thought that temp and salinity values at the source node should be much lower then what I see. From the result files, at a source point, the temp and sal values are too high at any time step: 2 degree and 28.4 psu, instead of 0 or little bit higher due to double diffusion processes. It is mixing at the very node point too fast.

The velocities at the source are: X = 4.6*10^(-4) and Y=0, the water discharge = 500m^3/s. The total area of all 3 elements involved (the source is right at the wall) is 1084230m^2.

Please see attached temperature and salinity results. The images show just a small area where source is placed, near the closed wall on the left.

I am probably doing something wrong, or maybe it is not possible to achieve the desired values of Temp and Sal at a source with such course mesh. Let me know if you need more details. Thank you in advance, your help is much appreciated.

Kind Regards!

Violet

Hello Violeta,

The program guarantees that if you prescribe a discharge Q and give a value of tracer at source T, the quantity entered in the domain during a time step is Q*T*DT. If you have large elements, think that any value appearing at a source is immediately transmitted on all the test function (with linear interpolation). Then to have the correct flux the values at the source increase very slowly. This is what you should see, but your statement :

"From the result files, at a source point, the temp and sal values are too high at any time step: 2 degree and 28.4 psu, instead of 0 or little bit higher due to double diffusion processes. It is mixing at the very node point too fast."

is not too clear for me, I do not understand if the values computed are too high or too low... What you could check is the mass balance, if the quantity entered in the doamin is what you expect.

With best regards,

Jean-Michel Hervouet

Thank you very much Jean-Michel for your reply.

I hoped to see, at a source cross section, constant plume with 0.0 temperature and 0.0 salinity, but with warmer and saltier water everywhere else in domain. If I look at the results, the temperature and salinity values at a plume is just little bit fresher and colder then in domain (28.4 and 2.2 instead of 0.0 and 0.0). Plume is there, but too warm and salty.

In past, when I modeled a case with the source in a closed domain, the mass balance of tracers was clear to me, but in this case I have an open boundary on one side and a source on another. The open boundary has different tracers values then in domain. The result file shows the total mass balance per time step for boundary and source together. I am not sure how I can extract the mass exiting at source only, as figure includes the mass exiting at the open boundary as well. Since I have tracers difference between the open boundary and domain, then due to the horizontal pressure gradients and Coriolis force, there is a constant flux at the open boundary on one side and I have a source on another. It is probably working fine, but I just need to be sure that it takes right source tracers values I prescribed, as my goal is to look at the total mixing of stratification due to the source input on a long run. Maybe there is another way to double check the tracers values at the source? Thank you for advice. Your help is much appreciated.

Kind Regards!

Violeta

Hello,

OK no I understand that the tracer is fresh water, so that too big a value means too much diffusion! So my explanation is probably correct, with large elements you have an instantaneous dilution, and you will never see smaller values at the source point, unless you fill all the elements around this point with fresh water.

With best regards,

Jean-Michel Hervouet

Thank you very much Jean-Michel for your patience and clear explanations. I might reduce the horizontal mesh, or fill all surrounding elements with the source tracers values as per your advice. Thank you.

Kind Regards!

Violeta