Hydrodynamic Simulation of an Irregularly Meandering Gravel-Bed River: Comparison of MIKE 21 FM and Delft3D Flow models

Abstract

This study aims at hydrodynamic modelling of Bow River, which passes through the City of Calgary, Canada. Bow River has a mobile gravel bed. Erosion and deposition processes were exacerbated by a catastrophic flood in 2013. Channel banks were eroded at various locations, and large gravel bars formed, which could lead to water level changes and accordingly more flooding. This study investigates the performance of Delft3D-Flow and MIKE 21 FM to simulate the hydrodynamics of the river during the 2013 flood. MIKE 21FM employs unstructured triangular mesh while Delft3D-Flow model uses curvilinear structured grids. Performance of each model was evaluated by the available historical water levels. The results of this study demonstrated that, with approximately the same averaged grid resolution, MIKE 21 FM resulted in more accurate results with a higher computational cost compared to the Delft3DFlow model. It was shown that Delft3D-Flow model may require higher grid cell resolution to result in comparably same depth-averaged velocities throughout the study area. However, considering the balance between the computational cost and the accuracy of the results, both models were capable to adequately replicate the hydrodynamics of the river during the 2013 flood. Results of statistical KS and ANOVA test analysis showed that the model predictions were sensitive to the horizontal eddy viscosity and the Manning roughness. This confirms the necessity of an appropriate calibration of the generated numerical models. The findings of this study shed light on the Bow River flood modelling, which can guide flood management.

Assessment of Sensitivity and Validity Hydrodynamic Model in Cisadane Using Delft3d Flow Model

“…Hence, a smaller grid size indicating high spatial resolution and a grid size of 50 m (10 km 2 domain) produces better results than other grid sizes. This condition is also achieved by Parsapour-moghaddam et al (2018).…”

“…Hence, a smaller grid size indicating high spatial resolution, a grid size of 50 m (10 km 2 domain), produces best results than other grid sizes (Figure 4.). This condition is also achieved by Parsapour-moghaddam et al (2018). Based on the validation results of TPXO 8.0, which is higher than TPXO 7.2, the 3D model uses TPXO 8.0 as the input boundary condition.…”

Study of the impact of the Îlot Tison weir on bedload transport of the Clain River (Poitiers, France) using BASEMENT software

“…Without the weir, the highest depth average velocities were located on the outer bank side of each meander bend. Levelling the weir restored the meander river flow (Lotsari, Tarsa, Kämäri, Alho, & Kasvi, 2019; Parsapour‐moghaddam, Rennie, & Slaney, 2018).…”

The effect of a local mesh refinement on hydraulic modelling of river meanders

“…The results corresponded well with our study as it was found that for triangular meshes, a higher resolution was required in order to obtain the same level of accuracy as the curvilinear meshes. In Parsapour‐Moghaddam, Rennie, and Slaney (2018), both D‐Flow FM and MIKE21 FM were used for hydraulic simulations of a meandering section of the Bow River, Canada. They found that lower depth‐averaged flow velocities and higher water levels were obtained by the triangular mesh, which is in line with our findings for the Grensmaas.…”