open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,
I computed my study area with a constant viscosity and different values of the VELOCITY DIFFUSIVITY (2. , 1. , 0.1 , 0.000001). As described in the users manual this parameter has a definite effect on the extent and shape of recirculations. The location and distribution of recirculations seems realistic to me. So far I'm satisfied with the computation.
Because I'm new to 2D-Simulations and especially to TELEMAC-2D it's hard to me to determine which value of the VELOCITY DIFFUSIVITY leads to the most realistic results. So maybe someone of you who is more sophisticated can help me and leave an assesment.
I attached a pdf-file containing screenshots of the entire study area and the certain recirculation areas using different values of VELOCITY DIFFUSIVITY.

Another and actually more important question to me is if the recirculations (and their extent) at the southernmost boundary are realistic or if the boundary condition is causing them.
Therefor I explain the flow situation shortly:
At this boundary a power house shall be simulated extracting 3900 m³/s (Boundary condition defined as "Liquid Boundary with prescribed flowrate"). In reality there are chambers just before the boundary leading to the particular turbines. To simplify the geometry just in front of the boundary I forwented designing them and allocated the outflow on the entire width.
The water is dammed up and so the water depth is at the point of lowest Bottom elevation more than 30 m.
Just before the outflow there is a steep slope followed by a flat area. The slope is limited with vertical walls on the sides up to a height as at the beginning of the steep slope. Above there are embankments (The water depth at the transition from the vertical wall to the embankments is still more than 10 m).
I attached the geometry file, so you can have a look by yourself!

Now in the results there are large recirculation areas in between the vertical walls restricting the steep slope (see screenshots in *.pdf). I'm really not sure if these results are realistic or if the boundary condition is setted wrong or is too close to the observed area!?
Is it possible that these recirculations are caused by the boundary condition and wouldn't appear if I choose another Boundary condition or something like that?

I would appreciate any assesment.

Cheers
Sarah