Suitability of 2D depth-averaged simulations to support predictions of cyprinids passage through Vertical Slot Fishways

Fishways are a key measure to recover longitudinal connectivity. Among them, Vertical Slot Fishways (VSFs) have undergone extensive studies. However, very few numerical studies have included coupled hydrodynamics and ecological simulations. We investigated the impact of using three-dimensional (3D RANS) and two-dimensional (2D SWE) flow simulations in a Eulerian Lagrangian Agent-based Model (ELAM) for VSFs. Spatial distribution of turbulent kinetic energy (k), velocity magnitude (|ui|), and velocity strain rate (Sij) were evaluated. While the |ui| was consistent across simulations, k was overestimated by the 2D SWE simulations. The frequency of Sij highest values was different between 3D RANS and 2D SWE. Also, the influence of these variables on the agent's cognitive response was inconsistent. 2D SWE displayed an influence of around 0%, 35%, and 25% for k, |ui| and Sij, respectively. 3D RANS resulted in percentages of around 80% of k, 10-20% of |ui| and 0-10% of Sij. All cases were validated, demonstrating that both approaches are suitable to reproduce experimental observations. We demonstrated that the 2D SWE might serve as a good input to simulate fish movement in VSF. Nonetheless, hydrodynamic variables' influence on agents' decision making might be biased from real fish experience.