open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear members of the Telemac community,

I have some troubles setting up a test case with mud consolidation.

My current set up is in attachment. Basically, my computational domain is a rectangle of 100m by 3200m with one open boundary on one of the little edges. The bathymetry is -4m everywhere, except in the 500m close to the open boundary where the bottom elevation drops down progressively to -10m. A sinusoidal water elevation is imposed at the open boundary (amplitude 2.5m). Only cohesive sediment is considered, with a certain concentration at the open boundary. I only consider two layers.

The idea is that sediments deposit on the bottom, and I want to see the effect of mud consolidation (bottom elevation rate should be slowed down with mud consolidation).

The results I got are a bit surprising.

zh2c.png shows accurate behavior without consolidation. However, consolidation seems to work way too fast (about 4m/year).

What is even more surprising is that this effect is not present in layer thicknesses. First layer (zh2c_layer1.png) varies from a few tens of centimeters (to be expected), while second layer does not change at all (99.9m, not shown).

So it's like there was no transfer from layer 1 to 2 (otherwise layer 2 should have increased) and I don't understand what's happening with the bathymetry, as the total thickness (layer 1 + 2) barely changed.

I am probably misunderstanding something. I would be very pleased if somebody could explain me what.

So, I've been looking further to my problem, and I think I might have found a bug...

In sisyphe.F, the call of TASSEMENT updates "ES", but not "ES_VASE". So, at the next time step, when calling SUSPENSION_MAIN, the line 445

ES(I,J) = ES_VASE(I,J)

brings back the initial values to each layer thickness. So, at each time step, the same amount of mud is transferred to the beneath layer, overestimating consolidation.

Adding the lines

DO I = 1, NPOIN
DO J = 1, NOMBLAY
ES_VASE(I, J) = ES(I, J)
ENDDO
ENDDO

after the call to TASSEMENT in sisyphe.F seems to solve the problem. But I guess it is not the optimal solution.

I would be happy if someone could cross check my solution.

Could someone please confirm me that this is a bug and that my solution is good? Or Am I doing something wrong here?