Modelling flow through vegetation is of great practical and scientific importance, especially when one considers the modelling of wetlands and vegetated floodplains, which are essential in wetland restoration and flood risk assessment projects. SOBEK 1D-2D is particularly useful for such modelling, since it combines 1D modelling of rivers and channels with 2D overland flow modelling. However, the resistance caused by vegetation is difficult to model: it is included in the bed friction term through a constant 'equivalent resistance (or roughness)' coefficient which is incapable of representing the resistance at all water depths.
Three different approaches were considered to obtain such coefficients, namely: 1DV model, theorybased formulae and empirical formulae using genetic programming (GP). It was found that the combination of the theory-based formula for un-submerged vegetation and the GP-obtained formula for submerged vegetation has great potential, given its simplicity, ease of use and accuracy, especially when vertically uniform vegetation is considered.
With the implementation of water-depth dependent resistance coefficients, SOBEK 1D-2D acquires some of the benefits of 3D-computation through the relationship between water depth and resistance coefficients, while maintaining the advantages of 2D simulations. In this way, a typical closure coefficient, usually determined by comparison with values that worked well in previous similar situations, is replaced by a theory-based coefficient, which can be directly related to the floodplain vegetation characteristics.
This project was executed in cooperation with UNESCO IHE Delft (MSc study). |
