Evaluation of uncertainties in settling velocities of particles in urban stormwater runoff

Field experiments were carried out to contribute to the assessment of the VICAS protocol aiming to measure settling velocities of particles. Samples of deposited sediments have been taken in the Django Reinhardt stormwater detention and settling tank in Chassieu, France, using sediment traps located on the tank bottom. The first set of experiments was designed to assess the VICAS protocol in terms of mass balance and repeatability. A bias in the measurement of settling velocities distributions of deposited sediments (i.e. particles with high settling velocities) was suspected and confirmed by specific tests. Uncertainties in the final distribution curves have been evaluated by using Monte Carlo simulations and the law of propagation of uncertainties. All uncertainty calculations were implemented in a MatLab code named UVICAS used for each experiment. This code allows analysing the main sources of uncertainties and their evolution during experiments. Uncertainties in the final distribution curves decrease with increasing values of settling velocities and are lower than 1%.