For safe design of culverts, the estimation of scour depth downstream of the culvert outlet is one of the most remarkable considerations. Poor design of culverts inevitably leads to destruction and increases expenditures on management and rehabilitation. The objective of this study is to investigate the performance of numerical models in predicting scour depth and its location downstream of box culverts in unsteady flow conditions. Culverts with unblocked and partially blocked outlets were used, and the predictions of two turbulence methods, namely renormalization group (RNG) and $k\text{-}\epsilon$ , were compared. Comprehensive experimental data of box culverts from the literature were used and the numerical models were performed in commercially available software. The downstream profile of scour development, maximum scour depth, and its location were obtained and compared with observed data. It was found that the numerical model was in good agreement with the experimental data. In both turbulence models, the scour depth increased with increase of discharge in the raising limb of the hydrograph, whereas for lower discharges in the falling limb of the hydrograph, the scour depth decreased.

This study presents a two-dimensional (2D) hydro-morphodynamic simulation of sediment dynamics in the Gemenc floodplain, a critical ecological zone along Hungary’s Danube River. The 60 km study area has a mean discharge of approximately 2300 m³/s, with peak floods exceeding 8000 m³/s. The objective was to analyze sediment transport, deposition, and flood hydrodynamics to support future floodplain restoration. The HEC-RAS 2D model was calibrated using water levels (Baja station), 2024 flood discharges, suspended sediment measurements, and visual stratigraphy surveys conducted after the event. A roughness sensitivity analysis was conducted to optimize Manning’s n values for various land covers. The hydrodynamic model showed strong agreement with observed hydrographs and discharge distributions across multiple cross-sections, capturing complex bidirectional flow between the main River and side branches. Sediment dynamics during the September 2024 Danube flood were effectively simulated, with SSC calibration showing a decreasing concentration trend, highlighting the floodplain’s function as a sediment trap. Predicted deposition patterns aligned with field-based visual stratigraphy, confirming high sediment accumulation near riverbanks and reduced deposition in distal zones. The model reproduced deposition thickness with acceptable variation, demonstrating spatial reliability and predictive strength. This study underscores the value of 2D modeling for integrating hydrodynamics and sediment transport to inform sustainable floodplain rehabilitation.