Dear Mathieu,
using a fixed Qs at the liquid boundary is always a little tricky because the chosen/measured values of the water discharge which trigger the shear stresses and the chosen/measured bedload discharge are seldom in equilibrium. This leads to unwanted erosion or deposition at the inlet.
One possibility you have is to fix the bottom at the first metres downstream the inlet boundary. This will avoid erosion but not sedimentation which is your main problem.
In your case you need to increase the shear stress at the inlet boundary. You can do this e.g. using another bedload transport formula, change the critical shear stress (if using a threshold formula), change the roughness or modify the inlet boundary conditions.
I would first calculate the needed velocity to transport your input bedload discharge and compare this to the velocities you have now at the inlet. You can play around with different bedload transport formula, critical shear stresses, roughness, ... to see which configuration would be better.
Concerning the inlet boundary conditions:
With VELOCITY PROFILES = 4;1 the distribution of the velocities will be calculated according your water depth at the inlet. This would lead to higher velocities at the deeper parts (=higher shear stresses = less sedimentation). In case of a natural (non-rectangular) cross-section you could open the boundary for the bedload only for the deeper parts (in the cli-file).
If you start from a rectangular cross section at the inlet you could decrease it in order to increase your shear stress.
Best regards,
Rebekka
