Abstract The increasing frequency of extreme hydrological events highlights the critical importance of reliable dam safety assessment and sediment management in large river systems. Three-dimensional numerical modeling of overflow and sediment transport processes is essential for ensuring the reliability of hydraulic structures and understanding the dynamics of reservoir systems. This study focuses on the Shardara Reservoir, located in the Turkestan region of Kazakhstan. The primary objective is to develop a high-resolution 3D numerical model to assess potential variations in water discharge and to analyze the hydrodynamic behavior of the reservoir, including the evaluation of dam breach risks. To simulate multiphase flows, the Volume of Fluid (VOF) method was employed in conjunction with the Pressure-Implicit with Splitting of Operators (PISO) algorithm to solve the Navier–Stokes equations, ensuring computational stability and accuracy under complex flow conditions. Unmanned Aerial Vehicles (UAV) were used to acquire high-resolution topographic data and conduct detailed monitoring of the hydraulic infrastructure. The integration of real-world topography significantly improved the accuracy of boundary conditions and hydrological parameters, enabling precise identification of high-risk zones. The proposed modeling approach is novel in its incorporation of actual terrain features and engineered structures for simulating overflow and dam breach scenarios within this regional context. Numerical experiments demonstrated how changes in discharge affect flow velocity and sediment transport dynamics. The assessment of erosion and sediment transport processes was conducted using a weakly coupled approach, in which flow velocity fields and hydrodynamic regime characteristics were first computed, followed by the analysis of potential erosion and deposition zones based on the Froude criteria. The simulation results showed good agreement with experimental and observational data, confirming the reliability and predictive capability of the developed model.
Omarova et al. (Thu,) studied this question.