Aquatic plants –or macrophytes- are an important part of coastal, estuarine and freshwater ecosystems worldwide, both from an ecological and an engineering viewpoint. Their meadows provide a wide range of ecosystem services: forming a physical protection of the shoreline, enhancing water quality and harbouring many other organisms. Unfortunately, these vegetations such as salt marshes, seagrasses or mangroves have been on the decline as a result of anthropogenic pressure and climate change, despite costly conservation and restoration efforts. The low success rate of these efforts might partially be due to a lack of understanding of the complex bio-physical interactions between plant properties, plant growth, hydro- and morphodynamics and water quality. The capability of plants to alter their abiotic environment via these interactions is referred to as ‘ecosystem engineering’. Many experimental studies, both in the field and in laboratory flumes, have been performed to unravel these interactions. Since such experiments are always hampered by practical limitations such as flume dimensions, available time, or uncontrolled conditions, this knowledge cannot always be generically applied. Therefore, the primary objective of this study is to develop a generically applicable model for feedbacks between flexible macrophytes and their physical environment. To warrant this general applicability under the various circumstances occurring in estuaries, the model development follows a process based approach; a data-orientated approach is merely applicable to known conditions. Modelling starts out on the scale of one plant to finish at the scale of a meadow. The focus is on seagrass, as seagrasses are well studied, highly flexible, have a relatively simple shape and are among the most productive as well as threatened ecosystems. The first step was to create the numerical model called ‘Dynveg’, by combining a novel dynamic plant bending model based on a Lagrangian force balance to an existing 1DV k-? turbulence model (Chapter 2). The plant bending model is based on measurable biomechanical properties of plants: length, width, thickness, volumetric density and the elasticity modulus. Because very flexible plants can assume a position almost parallel to the flow direction, friction too needed to be incorporated rather than pressure drag alone. Flume measurements on strips of eelgrass-like proportions provided the actual values for drag- and friction coefficients, as well as validation data for predicted strip positions and forces. The effect of multiple plants on hydrodynamics was incorporated by assuming that all plants in a meadow do the same, and by defining two turbulence length scales: One for internally generated turbulence, related to the wakes behind individual stems, and one for larger eddies created in the shear layer above, penetrating the canopy depending on the space between the stems. Dynveg compared favourably with the measurements of hydrodynamic characteristics in mimicked eelgrass by Nepf & Vivoni [2000]. Next, Dynveg was combined with the large-scale hydro- and morphodynamic model Delft3D to simulate two-dimensional spatial processes in and around meadows of flexible macrophytes (Chapter 3). The leading principle for this integration is the conditional similarity between flow characteristics in flexible vegetation and those in rigid vegetation: If the rigid vegetation has i) the same height as the deflected vegetation, ii) its plant volume redistributed over the vertical accordingly and iii) a drag coefficient representative of the streamlined shape, the flow is practically analogous for a range of plant properties and hydrodynamic conditions. This modelling method was validated by comparing model results with flume experiments on two seagrass species, showing good agreement for canopy height, flow velocity profile and flow adaptation length. A field measurement campaign in a French macrotidal bay bordered by an eelgrass meadow provided validation data for application to real meadows (Chapter 5). Along with a detailed bathymetry survey by jetski, time-series of flow velocity and sediment dynamics inside a meadow and over a bare adjacent area were measured over two tidal periods. The applied sediment transport formula [van Rijn, 1993] deals with vegetation effects on sediment pick-up and transport via the effects of plants on hydrodynamics. Vegetation-specific interactions such as particle trapping by blades or flow intensification directly around shoots were not taken into account. Nevertheless, the three-dimensional numerical model was able to reproduce the main features of the observations, indicating that the processes of vegetation bending in non-stationary flow and sediment transport through vegetated areas are incorporated correctly. Thus, the objective of making a model for feedbacks between flexible macrophytes and their physical environment has been met. The model can be applied as a tool in conservation and restoration studies or in long-term biogeomorphological feedback studies. Recommended extensions are the incorporation of plant-wave interactions, more intricate plant morphologies and a vegetation-specific transport formula. The second objective of this thesis was to use the developed model(s) as a tool to learn more about biophysical interactions under different conditions. In Chapter 4, Dynveg and the two-dimensional model were used to assess the ecosystem engineering capacities of three plant species that partly co-occur in temperate intertidal areas: the stiff Spartina anglica, the short flexible seagrass Zostera noltii and the tall flexible seagrass Zostera marina. The flow velocity inside the canopy, the canopy flux and the bed shear stress were used as proxies for the species’ ability to respectively absorb hydrodynamic energy, the supply of nutrients or sediment and the ability to prevent erosion. This analysis showed that a species’ eco-engineering capacities depend on its spatial density, its size, its structural rigidity and its buoyancy, but also on environmental conditions. Therefore, biomass, leaf area index or other lumped parameters that neglect structural properties are no good generic indicators of ecosystem engineering capacities. Rigid plants have more potential to trap sediment due to a higher canopy flux than flexible plants. This canopy flux showed to be inversely related to spatial density along the entire natural range. For flexible plants, the canopy flux is only related to density in relatively sparse meadows; in denser meadows the canopy flux is constant with increasing density. Flexible plants are better at preventing erosion because they are more efficient in reducing bed shear stresses than rigid plants. For very thin plants, buoyancy is the most important determinant of position in given flow conditions. For intermediate flexible plants, the structural rigidity is the most influential parameter, whereas for (nearly) rigid plants, the spatial density is dominant. In Chapter 6, the three-dimensional model of the macrotidal bay was used to study the effects of different types of macrophytes on (residual) sediment transport and light availability. The effects of the real, relatively sparse eelgrass meadow were compared to those of a meadow with rigid plants of the same spatial density, with a dense eelgrass meadow, and with a bare bed. Though the differences between these four vegetation scenarios were small –only a few percent- the consequences on long timescales can be considerable. In deep water, sparse flexible vegetation kept more sediment inside the bay than rigid or denser plants. When vegetation only occupies a small part of the water column, plants prevent erosion rather than promote deposition and they have more effect on bed-load transport than on the transport of suspended sediment. Stiff and denser plants affect the bed-load more than sparse flexible vegetation, thereby blocking the transport from outside to inside. The presence of dense or stiff macrophytes increased the light availability at the bed over a tidal cycle up to 7% with respect to a bare bed. The increase of light availability was less pronounced for the relatively open eelgrass meadow: up to 3%. Overall, this study has resulted in a widely applicable model for the interactions between flexible aquatic plants, flow and sediment transport and in more insight in some of these interactions. Other researchers are encouraged to use this tool complementary to fieldwork and laboratory experiments, and to extend it with other functionalities, e.g. for wave attenuation or vegetation development.