University of Twente Student Theses


Representing spatially variable bathymetry and vegetation in a hydrodynamic model : a subgrid-based case study in the Whitianga Estuary, New Zealand

Baltus, O. (2022) Representing spatially variable bathymetry and vegetation in a hydrodynamic model : a subgrid-based case study in the Whitianga Estuary, New Zealand.

[img] PDF
Abstract:Due to climate change and sea level rise, there is an increasing need for sustainable and cost-effective coastal protection. Mangrove eco-systems have shown to be a sustainable addition to traditional coastal protection. Mangrove forests are capable to attenuate waves and lower water levels and by sediment trapping elevate their surroundings to grow with sea level rise. However, mangrove eco-systems have a limited capability to adapt to sea level rise. Therefore, it is important to learn more about the functioning and persistence of mangroves. Often hydrodynamic models form the basis of numerical studies into these dynamics. To correctly model the hydrodynamics, high resolution spatial data needs to be taken into over large areas, which leads to long computational times. This leads to interest in computationally efficient and accurate hydrodynamic models. Subgrid-based hydrodynamic models have proven to use spatial data (e.g. bathymetry) with a high spatial resolution in a computationally efficient way. But in subgrid-based hydrodynamic models, vegetation is currently only taken into account by increasing the bed roughness. Representing vegetation via bed roughness has the downside that the roughness contribution does not scale with the water depth, while drag due to vegetation does scale with the water depth. Also, when looking into morphodynamics, the bed shear stresses are over-estimated if vegetation is incorporated as a bed roughness. Therefore, this study explores two options of how to include vegetation in a subgrid-based hydrodynamic model. More specifically, how the Avicennia marina variant Australasica mangrove species can be represented correctly in the model. A mangrove study area in New Zealand is used, where an extensive measurement campaign took place in February 2017. In this period, bathymetric, vegetation, water level and velocity measurements have been done. Based on these measurements, a hydrodynamic base-case model without a vegetation representation is made. Two vegetation representations are added to the model. The first representation is using a spatially varying roughness, where the bed roughness is increased at vegetated areas to take into account the effect of vegetation. The second representation will use the Darcy equations to model the flow through vegetation as flow through a porous medium. To test the functioning of both vegetation representations, competences for model accuracy are defined, based on the key characteristics of the tidal dynamics of the area. The results of the different models demonstrate that the use of subgrid has significant potential in the modeling of intertidal area. As it allows for the use of larger computational grid cells without loss of accuracy. This results in lower computational cost of the model. Modeling a whole spring-neap cycle (12 days) took only a little more than an hour, while decreasing the computational grid to the resolution of the spatial data the computational cost of the model increased to 140 hours. Next to this, the research has shown that the subgrid modeling technique can especially be used in intertidal areas as the flow is mainly bathymetry driven, shown in the accuracy of the modeled tidal flow-stage curves. The results of the vegetation representations show that spatially varying bed roughness did improve the overall functioning of the model on the defined competences. Due to the increased bed shear stress on the forest platform, more water is routed via the creeks, increasing the flow velocity in the creeks according to the measurements. But on the other hand, the varying roughness did show to have only minor affect on the highest water level modeled in the study area, caused by the decreased influence of the bed roughness on large water depths. The flow through a porous medium, implemented via an interflow layer, showed to have no significant effect on the highest water levels in the area and showed to have less predictive capability on the low water levels in the creek due to the decreased of the bathymetry on the flow. But, the interflow layer did improve the modeled flow velocities on the forest platform. For future research, it would be recommended to look further into using the subgrid modeling technique to model intertidal area like mangroves as it has shown to decrease computational cost without massively affecting the modeled hydrodynamics. For the vegetation representations, it is recommended to use varying vegetation characteristics when extensive vegetation measurements are available. Additionally it is recommended to further develop the interflow layer, so that the interflow layer can be applied to parts of the model domain with dense vegetation. At last, it is recommended to look into other vegetation representations that scale with the water depth, as the roughness raster has shown to be less applicable for highly varying water levels.
Item Type:Essay (Master)
Faculty:ET: Engineering Technology
Programme:Civil Engineering and Management MSc (60026)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page