Hydropower plants (HPP) constitute a major obstacle to fish migration. To insure their safety during downstream migration, bypass channels are installed in HPP vicinities, deviating the fishes into a reception area through a waterfall. However, unoptimized reception conditions (low water level, high velocities or pressure…) can lead these bypass facilities to be harmful. Therefore, HPP stakeholders may want to enhance the fish-friendliness of the concerned power plants, with effective and easily adaptable technical solutions. For this, an accurate description of the downstream passage hydrodynamics is necessary., especially inside the waterfall region. Here, the outlet flow of a downstream passage device, related to the Las Rives HPP (Ariège River), is investigated. A 1,35 m³/s flow rate goes through a nearly 30 m long water channel finishing into a 3,35 m height waterfall. The reception basin has a specific topography, with a 0,68 m average water level. This site has been investigated by Bercovitz et al. (2022), evidencing some fish injuries. Unsteady Reynolds Averaged Navier-Stokes (URANS) numerical simulations are performed with the OpenFOAM software. Initially, the downstream channel is reproduced. A 1,35 m³/s flow rate is imposed, with a Volume of Fluid (VoF) method and a k-ω_SST turbulence model. It provides a realistic distribution of the fluid phase, with the related velocities, over the channel outlet boundary. Then, this phase / velocity distribution is used as the waterfall inlet. To define the more appropriate approach, five cases are compared, with a VoF (laminar, k-ω_SSTor k-ε turbulence model) or a Two Phase method (laminar or k-ε). In the same time, in situ measurements are performed in the reception area. Topographic data are provided into the simulation, to fit the experimental conditions for the bottom wall. A differential pressure sensor is placed in various locations below the waterfall. The dynamic pressure is acquired and an equivalent velocity is derived. Comparisons between observations and TP cases indicate some major issues in the description of the waterfall. TP solver seems to overestimate the air / water exchange, the waterfall being way too dispersed than the observed one. A significant drop for pressure and velocities, compared to other simulations, is also noted. VoF approach compares accurately to experimental measurements. Qualitatively, the waterfall shape matches in situ observations. Averaged dynamic pressures are similar to experimental measurements (30~70 mbars depending on the location); with a good agreement falling on the k-ε model. Standard deviations appear lower than the measured ones, due to the URANS numerical schemes. The numerical computations bring a valuable tool to investigate the inside of the waterfall, tracking specific locations potentially harmful for fishes. Therefore, the appropriate approach (VoF - k-ε) can be employed to propose and evaluate technical solutions to improve the fish-friendliness of the downstream equipment.