I have tried to use Thompson conditions for the liquid boundary but I didn't work. The simulation ran a long time, the time step was reducing to 10e-10 second so I stopped it.
It tried to adjust the bathymetry of the estuary to the boundary conditions but it didn't work yet.
I finally succeed to have a simulation without free surface perturbation reducing the maximum current number from 1 to 0.5.
I know what maximum current number means but I really don't know why it delete the free surface perturbations???