Hello,
I would say no. Actually your advection velocity field must be divergence free, otherwise you will have a mass loss of tracer. In the transformed mesh where the advection is done (to include relocalization), the vertical velocity (W* given per layer) is computed so that the continuity is ensured, given the horizontal velocities. This is the most complicated part of Telemac-3D. If you are to do a vertical advection, you would have to do it as we do the settling velocity, in an extra step or coupled with diffusion, in the real mesh. I would then suggest a 1D vertical advection per vertical, ensuring the conservation of mass (but monotonicity will not be ensured... what happens e.g. when sediment is advected on the bottom, the concentration can have increasing values, higher than the previous time step).
With best regards,
Jean-Michel Hervouet