This research focuses on the implementation of the Building with Nature (BwN) approach in the design of artificial coral reefs. The goal is to explore environmental preferences and design tools for coral development, considering both coral habitat potential and coastal protection services. The research also aims to optimize artificial reef design for a case study in Addu City, Maldives. A conceptual model is developed based on an extensive literature review, incorporating critical engineering and ecological variables relevant to coral habitat potential and coastal protection services. This conceptual model serves as a practical design tool, fulfilling three primary purposes: identifying variables that require further examination to explore coral habitat potential, identifying variables that require analysis to explore coastal protection services, and identifying design variables that influence physical-chemical and biological variables for potential integral solutions. The research also utilizes OpenFOAM, a numerical modeling tool commonly used in coastal engineering, to design artificial reefs. OpenFOAM accurately models flow and turbulence regimes, crucial for the ecological and biological functioning of coral reefs. An OpenFOAM numerical model is set up, calibrated, and validated for the case study in Addu City. Three design alternatives with different slopes are evaluated using a multi-criteria analysis (MCA), considering criteria such as coastal protection, coral habitat potential, and costs. The numerical model is applied to assess flow regimes and wave transmission for different design variables. Based on the MCA results and the evaluation of criteria, an artificial reef with a mild slope of 1:3 is selected as the optimal integral solution for the case study. This design performs well in terms of coastal protection, coral habitat potential, and costs. The research demonstrates that the conceptual model and OpenFOAM numerical model are valuable design tools for assessing and optimizing artificial reef design, adhering to the principles of the BwN approach. In conclusion, this research contributes to expanding knowledge and developing tools for implementing the BwN approach in artificial reef design. By considering both coral habitat potential and coastal protection services, the design of artificial reefs can be optimized to provide multiple benefits, including environmental and socio-economic values. The conceptual model and numerical modeling tool offer practical assistance in the design process, promoting the integration of nature and engineering for sustainable coastal solutions.

Climate change puts under pressure existing and future coastal interventions. Growing threats like sea-level rise and intensity of storms require solutions to be adaptable and resilient. Nature-based solutions have shown to tackle these challenges while providing social, environmental, and economic benefits. The role of vegetation in coastal protection is increasingly recognized. Aquatic vegetation reduces erosion, storm surge, and incoming wave height. Large-scale modeling of waves with spectral wave models such as SWAN is indispensable for the design of coastal structures and the assessment of flood risk. Wave dissipation due to vegetation can be modeled in SWAN as increased bottom friction (implicit modeling) or as an additional dissipation function (explicit modeling). The second assumes that vegetation can be represented as rigid cylinders or plates (canopies) with different properties. While some studies concluded that implicit modeling reproduces the spectral evolution of field measurements more closely, others concluded the opposite. Within the BE-SAFE project, field campaigns measured the spectral energy distribution over salt marshes in the Dutch Wadden Sea during several winter storms. The vegetated foreshore in front of the coastal dike got submerged over 2 m of water during high tide and storm surge. The measurements deployed wave gauges over the study transect, which was defined between the pioneer zone marsh edge and the near-dike location (300 m behind the salt marsh). Calibrating the implicit and explicit models in SWAN brought the modeled total wave energy decay closer to the measurement. Nevertheless, the spectral shape, which describes the energy distribution over frequencies, still showed significant and not yet understood differences near the dike. A methodology was executed to investigate the mechanisms that could reduce the spectral mismatch between the SWAN wave model and measurements over vegetation. First, the literature highlighted possible mechanisms that could be incorporated for this purpose. Next, a new frequency-distributed explicit dissipation model of Jacobsen et al. (2019) was implemented in SWAN and compared to implicit and explicit models using lab and field measurements. The results showed that the newly implemented model accurately captures the physics and the change of spectral shapes for all experimentally tested wave conditions and submergences. In contrast, the existing implicit and explicit dissipation models in SWAN reproduce the spectral evolution only under certain circumstances. In the validation and comparison to the field measurements with a much larger water depth than the vegetation height, the model of Jacobsen et al. (2019) correctly captured the vegetation's physical representation and the dissipation on the wind-sea frequencies. Nevertheless, the amount of energy on low frequencies was largely underpredicted by all frequency-distributed models. Therefore, the model of Jacobsen et al. (2019) was modified to include flexibility in a frequency-dependent reduction factor that reproduced the energy decay of the measurements in all frequency regions. Other mechanisms that could be responsible for the mismatch before and over the marsh are the redistribution of energy by non-linear triad interactions, generation of infra-gravity waves, and near-shore currents caused by horizontal variations on the vegetation properties. The present research provides the range of conditions in which the tested explicit and implicit energy dissipation functions in SWAN are able to simulate the spectral evolution over rigid canopies and flexible salt-marsh vegetation. A new version of SWAN includes a new frequency-distributed explicit model that performed more accurately than existing models for rigid canopies. The physical insights from the research contributed to developing additional versions of SWAN, which performed closely to the energy distribution of the measurements over deeply submerged and flexible salt marsh vegetation species. References: Jacobsen, McFall, Van der A (2019). A frequency distributed dissipation model for canopies. Coastal Engineering, 150, 135-146.

Recent analysis on measurements in the North Sea has shown a large amount of wave energy ranging in infragravity bands in the North Sea during storm events. To better understand the generation and propagation of free infragravity waves, the SWAN model is used to inspect the infragravity wave pattern, the infragravity wave origins, and the corresponding contribution from different coastlines in the North Sea. The local equilibrium between sea swell wave and infragravity wavebands is applied to parameterize the source infragravity wave along the coastlines bordering the North Sea following the approach of Arduin et. al. (2014). The empirical parameter α1 representing the radiating energy level from each coastline and bottom friction coefficient χ representing the dissipation strength are combined to calibrate the model. 10 storm events with an observed infragravity wave height over 10cm are chosen to gain detailed insight into the evolution of free infragravity waves under extreme conditions in SWAN. A correlation between α1 and χ is found. A relatively good explanation between measured and predicted infragravity wave height can be obtained as long as a large α1 and a large χ are implemented in the model or vice versa, showing an accuracy of up to 75% in the open sea. Based on a proper combination of two variables, the contribution of infragravity energy from each coastline varies spatially in predicted results. The Danish coastline is the dominant region radiating infragravity waves to the open sea and the reflected free infragravity waves could reach the adjacent coastlines such as Netherlands and UK. Behind the peak of an extreme storm event, the excitation of free infragravity energy is observed along the south of the UK coast originating from Denmark. The north of Holland coasts also radiate large amounts of free IG energy but are unable to have a prominent impact in the center of the North Sea due to energy dissipation. These results imply that the radiated infragravity waves can be influential in distant coastlines in the North Sea during storm events and the changes of local coastal dynamics can be attributed to a different generation region due to the storm characteristics.

Tidal energy has a large potential to contribute to achieving the sustainability goals in the Netherlands, due to the long coastline and many estuaries. The investment costs for the implementation of offshore tidal energy are however, still a drawback. A possibility to lower the investment costs for tidal energy is to implement tidal energy extraction into existing coastal infrastructure. An example of this is the current pilot project in the Eastern Scheldt barrier where five Tocardo turbines are attached to the barrier's geometry. The flow passing the turbines in the barrier is contracted by the barrier's geometry and the weir of the foundation of the barrier. Deltares executed an environmental impact assessment for the pilot project with a measurement campaign and a computationally expensive but accurate blade resolved numerical model. In this thesis, a cheaper computational model is added to the these assessment tools. To do so, a turbine parameterization is introduced in the finite element model FinLab. The relatively cheaper computational model should make it possible to execute more extensive sets of simulations of the pilot project in the future. FinLab solves the non-hydrostatic three-dimensional incompressible Navies-Stokes equations in an unstructured mesh with an optional moving free-surface. The used meshes are refined in certain areas of complex flow, such as the turbine parameterization, near the bed, the recirculation zone of the weir and the wake of the turbine. An extra source term is added to the Navier-Stokes equations on a chosen number of turbine integration points, to implement the turbine into the mesh. The placement of these integration points in a two-dimensional mesh is rather simple, the integration points are evenly distributed over the diameter of the turbine. In a three-dimensional mesh, a Fibonacci series is used to implement an actuator disc (AD) into the model domain. A benefit of the Fibonacci circle is that each turbine integration point represents a similar area. The force in the turbine integration points is determined by three methods: an actuator disc with a uniform thrust distribution (uniform AD), a rotational averaged actuator disc blade element momentum method (AD-BEM) and a non-rational averaged actuator line blade element momentum method (AL-BEM). In the uniform AD method, a measured thrust force is divided equally over all turbine integration points. In the BEM methods axial and tangential forces are determined for each turbine integration point based on the local flow speed and experimental lift and drag coefficients. With these methods, it is possible to estimate the generated power and thrust by the turbine. A tip-correction is added to the AD-BEM method to achieve accurate results of the simulations. Furthermore, in both the AD-BEM and AL-BEM methods a small actuator disc is applied in the nacelle region to reproduce the nacelle's blockage of the flow. The uniform AD, AD-BEM and AL-BEM parameterizations are validated with measurement data of flume experiments of a single turbine, flume experiments with different turbine-weir geometries and flume experiments with multiple turbines. The combination of these last two sets of experiments does very well represent the situation in the Eastern Scheldt barrier. As stated above, the AD-BEM and AL-BEM methods are able to estimate the turbine performance. The results of the AD-BEM simulations show a relative error of the turbine's generated power of within 20% in comparison with the flume experiments. The accuracy of the thrust fluctuates more. The AD-BEM method shows to be able to accurately estimate the location of optimum power harvesting in a combined turbine-weir geometry. Both the power and thrust accuracy depend on the mesh resolution and the turbine's rotational speed. Accurate results can be obtained by fine-tuning these settings. The AL-BEM method should be able to produce at least similar similar accuracy to the AD-BEM method when it comes to power and thrust. The AL-BEM method in this thesis is however, not yet able to do so. The results of the simulations with the three methods are also compared with time-averaged velocity measurements of the flume experiments. In the near wake, the uniform AD method does not imply any wake rotation. The rotation of the wake is represented accurately in a time-averaged sense by the AD-BEM method. The AL-BEM method adds transient features such as the downstream trailing tip-vortices to the flow. The velocity shear and turbulent eddies in the near wake as caused by the different methods influence the TKE in the near wake and the recovery of the far wake. The uniform AD under-estimates the wake recovery. The extra velocity shear in the AD-BEM method improves the accuracy but still under-estimates the recovery rate. Resolving the transient features of the near wake in the AL-BEM method further improves the accuracy of the wake recovery. In the simulations with the weir, the geometry seems to be dominant and the wake recovery rate is more accurate than in the simulations without the weir. The AD-BEM method is applied to the Eastern Scheldt field case. Five turbines are introduced in a mesh which represents Roompot 7 to 9 in a simplified manner. Due to the simplification of the mesh, the resistance of the barrier on the flow appeared to be too low. Therefore, the simulations are executed with an upstream discharge boundary. This discharge boundary makes the model less useful to estimate the environmental impact of the turbines in the barrier and at the moment not yet an alternative for the blade resolved model by Deltares when it comes to estimating these effects. With the discharge boundary, the AD-BEM model predicts the average thrust over the five turbines with a relative error of 3% and the average power with a relative error of 12%. This is only slightly less accurate than the results by the blade resolved model by Deltares, while the computational costs of the Eastern Scheldt field case simulations are only a fraction of the earlier executed blade resolved simulations by Deltares.

On the east coast of Romania, at Eforie, coastal erosion manifests. To strengthen the coastal area a large coastal protection project was setup involving beach nourishment combined with the construction of breakwaters. The breakwaters are designed with the well-known modified Van der Meer formulas. To ensure confidence in the breakwaters stability the designs were tested during a 3D physical model test. The designs were stable, however the measured damage to the breakwater was larger than was expected. Investigating the physical model test resulted in three important aspects, that need to be investigated. First, a broad list of configurations, that could responsible for the damage, could be identified. Secondly, excessive wave breaking on the foreshore during the model test resulted in a shift in wave energy from higher to lower frequency waves. These low frequency waves, or so-called infragravity waves became increasingly dominant near the coast. Their effect on the stability of the structure is however unknown. Thirdly, the incoming waves were very oblique and on a shallow foreshore, resulting in a different failure mechanism than is incorporated in the modified Van der Meer formulas. A new stability method has to be formulated to investigate breakwater stability for this physical model test. Due to the broad list of configurations and the possible influence of infragravity waves, real scale or physical model tests are not feasible and a numerical model needs to be used. A reliable and widely applicable method to link breakwater stability to a numerical model has however not yet been composed. Therefore, the goal of this thesis is to propose a method that can link breakwater stability to a numerical model and to assess whether the identified configurations resulted in the damage along the statically stable rubble mound breakwater, measured during the Eforie physical model test. First, an investigation is performed on the physical model test set-up and observations, resulting in a final list of 5 configurations, that are investigated in this research: The applied offshore transitional slope (1), the assumption of uni-directional waves (2), the slope of the lower foreshore (3), the depth-contour lines inducing wave focusing (4) and the very oblique wave angle on a shallow foreshore (5). Secondly, a method is proposed linking breakwater stability to a velocity signal from the numerical model SWASH. An equation is formulated, based on the theory of Izbash (1935), with a slope factor included, and scaled with the theory of Shields (1936). It requires a velocity signal, that can be obtained from SWASH, to calculate a stone size required for stability. Thirdly, a numerical model is set up in SWASH, with grid dimensions 3m x 2m, resembling the physical model test. The breakwater is modelled as an impermeable core with a permeable porosity layer placed on top. The thickness of the porosity layer is based on the thickness of the outer armour layer of the original breakwater. The numerical model is validated by comparing the wave characteristics, at several locations along the breakwater, to wave data available from the physical model test. The numerical model shows accurate resemblance of the wave characteristics. Since the wave velocity is linked to the wave height, it is assumed that the wave velocity on the breakwater is also correctly modelled. The model is therefore found valid for the modelling study. In the numerical model along the still waterline measurement points are indicated that provide the velocity and waterlevel signal during a simulation. In the numerical model two layers in the vertical are assumed and tested to be sufficient. The velocity of the top layer resembles the velocity that flows just over the stones. Therefore from the velocity signal of the top layer the governing u_0.2% along the waterline at the breakwater is obtained and from the waterlevel signal the wave spectrum is derived. In the study simulations are performed in which the configurations are tested one by one, and all simulations are assessed on two parameters: the u_0.2% and the wave spectral transformation along the breakwater. The results from the different simulations are compared relatively to identify the relative effect the configurations have on the velocity and wave characteristics. The results of this research show that breakwater stability can be predicted reasonably well from a velocity signal obtained from SWASH. The velocity signal, obtained from SWASH, results in reliable stone sizes. The configurations could be investigated with the proposed method and the results provide reliable and useful insights. In addition, the proposed method is able to identify the effect of infragravity wave energy on the stability of a breakwater. The method is also tested by calculating the relative obliqueness factors for different incoming wave angles, which shows promising results. It is important to reproduce the breakwater porosity well in the numerical model as it can significantly influence the velocity signal. A decrease/increase of the porosity thickness with 30% or 0.6m can result in an increase/reduction of 20-26% in velocity respectively. The five discussed configurations provide partial explanations to (in)directly induce the higher breakwater damage in the physical model test. Both the applied offshore transitional slope (1) as the assumption of uni-directional waves (2) result in a slight underestimation of the breakwater stability and therefore a somewhat conservative design along the entire length of the breakwater. The combined effect resulted in a reduction of 0-6% around the head of the breakwater, h/Hs = 2.5-4.8, and a reduction of 16-24% near the shore, h/Hs = 1.1-1.8. Especially near the shore the breakwater is conservatively designed, due to the fact that both the transitional slope as the assumption of unidirectional waves increases the infragravity wave energy in the system. It is found that incoming waves break around h/Hs = 1.7 after which the infragravity waves induce a temporary increase in waterlevel, around h/Hs = 1.1-1.8. This allows the depth-limited short waves to become bigger resulting in higher velocities and more damage on the breakwater. This affects the breakwater stability closer to shore and needs to be taken into account when designing a breakwater in these conditions. The lower foreshore (3) induces the generation of infragravity waves, which affect the velocity closer to shore as described above. The depth-contour lines (4) result in a wave focusing effect increasing the velocity around h/Hs = 1.1-1.8 with 8-11%. Based on the results of this thesis the very oblique wave angle on a shallow foreshore (5) does not induce higher velocities and breakwater instability. It is however assumed that the effect of a breaking plunging wave, inducing acceleration and pressure difference effects on the stones on a slope, is not sufficiently into account, due to the grid dimensions used in the model. As other plausible causes of the increased damage are disproven, it seems likely that the different oblique wave breaking process that is not modelled in detail leads to the increased damage.

All over the world, coastal protection measures are taken, which can be soft (e.g. sand nourishments on sandy beaches) or hard (e.g. seadikes or seawalls). Special care has to be taken to design the transition between these hard and soft flood defences, as they are often vulnerable components of a coastal defence system. This study therefore aims to give more insight into the morphological processes around hard-soft transitions, in order to make recommendations on design and nourishment plans for hard-soft transitions in the future. To this end, the hard-soft transition of Maasvlakte 2 (MV2) was studied. For the case of MV2 the limited knowledge of hard-soft transitions resulted in a conflict of the assessment method of the flood defence and the nourishment strategy according to the client requirements. The most important factors governing the morphological behaviour were studied using different numerical models (UNIBEST-LT/CL , SWAN and XBeach. From the results of the case study of Maasvlakte 2, it was found that the most severe erosion at the hard-soft transition is found during periods with strong southward longshore transports, which causes positive longshore transport gradients at the transition. Cross-shore processes such as storm erosion also indirectly contribute to this by transporting sediment from higher in the profile to lower in the profile. Regarding the influence of wave reflection, preliminary results were found. Using SWAN and UNIBEST, the effect of short waves was investigated, which showed that reflection of short waves is larger at the deeper located part of the hard flood defence. The preliminary results from XBeach showed that bound long wave reflection is larger closer to the hard soft-transition, and that these reflected waves also reach quite far offshore. As regards to the role of the tide, it was found that the dominant northward directed flood tidal current along the soft flood defence strengthens the supply of sediment from south during northward transport. On the other hand, the dominant southward directed ebb current along the hard flood defence hardly picks up any sediment and therefore it does not contribute to the morphological changes at the hard-soft transition. Some generic observations in the morphological behaviour of MV2 are likely to be found at other hard-soft transitions: - A strong dynamic variability is usually present at the hard-soft transition, in the form of coastline retreat which gives a rotated coastline shape. - The reflection of short waves will likely be larger at the deeper part of the hard flood defence. Closer to the transition zone, where sediment from the soft flood defence is deposited, short wave reflection will be smaller, but long wave reflection will be larger. - In general the highest erosion will be observed in periods with oblique wave incidence where sediment is lost due to longshore sediment transport gradients. Based on these findings, several recommendations are proposed to design a hard-soft transition. Moreover this study presents some other design examples to minimise the nourishment frequency. Finally, recommendations are presented for further research.

Shore-normal breakwaters are constructed in coastal zones both for beach protection (erosion reduction) and port development (wave sheltering). These breakwaters have an effect on the waves, the (wave-driven) currents and hence, the sediment transport along the coast. The waves and tide force an equilibrium sediment transport along the coast in a natural unaltered coast. When breakwaters are constructed this sediment transport in the alongshore direction is (partially) blocked. This results in accretion at the updrift side of the breakwater and coastline retreat at the lee-side. Coastal erosion behind a breakwater can result in floods, destroy property or simply narrow the beach. Not only the short term effects of these structures (what happens during and short after construction) is important to know, also the long term effects need to be known; because the breakwaters are build for decades coastal influence of these breakwaters needs to be known for these time-scales as well. Therefore it is relevant to investigate the scale (in time and space) of these adverse effects beforehand. In coastal engineering, numerical models are used to predict the impact of coastal constructions like breakwaters. High detailed models which take for all physical processes into account will result in accurate predictions but result in large computation times. Simplified models have smaller computation times and are more suitable for coastal impact predictions on larger spatial (10 - 100 km) and temporal (decades) scale. The objective of the thesis is to improve the coastline change predictions of models at decadal scale by reviewing the common practice and a new, very fast, module for lee-side wave computations and investigating different wave processes that can improve the wave modelling. This research will focus on (1) understanding of what wave processes are important for the coastline change and (2) advise on what models to use for which conditions and how to the improve model predictions. The approach of this thesis is triple. First five different modelling approaches for wave modelling are compared, varying in computation time and usability, to a very accurate model that will be used as ground truth. The wave conditions will vary in direction and both wind waves and swell waves will be modelled. From this step de accuracy of these models for different wave conditions can be analyzed. The second step is to get an better understanding of the processes that are involved with the wave modelling. This information can be used to improve the accuracy of model approaches with high computation times. The last step is to see how this relates to sediment transport and thus the coastline change at the lee-side of a breakwater. This can be concluded after the investigation of 3 wave model (SWASH, SWAN and the Kamphuis module) SWAN model approach is for cases with wind waves a good model approach: the wave height, wave direction and the sediment transport are well representative. Also the setup differences are very similar to the ground truth model. SWAN is not very accurate for cases with a small directional spreading. For these cases the diffraction is more important and there is not enough wave energy in the sheltered area nor is the wave direction well represented. Kamphuis with Snellius (refraction) is for the case with a wide directional spreading, given the computation speed, pretty good. For the cases with a small directional spreading the wave height is not accurate. The conclusions related to the influence of the wave processes are: The refraction is investigated by comparing the Kamphuis module (which does not incorporate refraction) with the ground truth model. When Snell's law was applied to the kamphuis module the model results did show good results. Therefore the refraction is (especially outside the sheltered zone) very important to the wave direction. There is a large difference for cases with small and large directional spreading. The SWAN model gives results can can be expected when no diffraction is present. Directional spreading both influences the wave height directly behind the breakwater as well as the influence lengt of the breakwater. Diffraction does not play an important role for wind waves. The large directional spreading results in much smaller energy differences between the sheltered and the non-sheltered zone. Even SWAN without diffraction gives a good representation for wind waves. For swell waves however this is very different.Then diffraction is very important. There is a much larger difference in wave energy between the shelterd and the non-shelterd zone and for a good representation a model that can coop with diffraction is a must. Current induced refraction of the waves has influences for waves of an incoming agle 0 to 30 degrees because this will result in rip currents along the breakwater that turn the waves up to an extra 10 to 15 degrees. The relative sediment transport of the five modelling aproaches was also analyzed. The sediment transport proxy for the five modelling approaches showed that SWAN computes the wave height and direction well for wind waves. For swell waves however the model did not show good results. The reason is that there is no diffraction in SWAN, the diffraction computations with SWAN did not show much improvements either. The sediment transportation proxy for the Kamphuis module with Snellius where expected to be good for wind waves, because the wave height and wave direction where well represented. However the length of the erosion pit was much smaller and also the shape of the erosion pit is very different from the other two models.

Over the recent years, flood risks and losses have been increasing for coastal cities due to climate change, subsidence, population and economic growth. Hong Kong and Macau are two cities located in the Pearl River Delta that experience a significant flood risk due to storm surges. The increased losses and risks has sparked interest around the world for efficient and accurate flood forecasting. At the moment coastal flooding events are often simulated with difficult hydrodynamic models that reproduce the physical phenoms. Over the last decades there has been more interest in other methods to forecast storm surges, namely neural networks. Other than hydrodynamic models a neural network is capable is making predictions in seconds, while the model can take hours to finish simulation. fast and accurate storm surge forecasting is of importance for disaster and evacuations management strategies and will only become more important in the future. During this research the main goal is to develop a neural network capable of prediction maximum water levels due to storm surges in case of an approaching tropical cyclone. The neural network is trained with data that is obtained from hydrodynamic simulations. A synthetic storm database is used to provide the necessary data to conduct 1000 simulations of which the results are used in the neural network. The first focusses on developing a hydrodynamic model capable of accurately simulation tropical cyclone induced storm surges in Hong Kong and Macau. The model is calibrated in a way to reproduce the real world as close as possible. It accounts for the real life bathymetry, topography, astronomical tides and wind forcing. The storm surge model is validated extensively based on three historical tropical cyclones. The errors between the observed and simulated water level are below 20 cm, after calibrations of different physical parameters within the model. The second step of this research is to use the validated storm surge model to run synthetic storm simulations. Instead of using historical storms who only have been recorded for 40 years, a synthetic storm database is used containing 10000 years worth of data. From that storm data, 1000 synthetic tropical cyclones are selected that come close to Hong Kong and Macau. With the data obtained from the previous two steps, it is finally possible to training the neural network. In this network a total of seven input parameters (tropical cyclone track parameters) are used to estimate the maximum water level that will occur during the tropical cyclone. The input parameters considered are: latitude, longitude of TC eye, maximum wind speed, minimum eye pressure, radius of maximum winds, forward speed, forward propagation direction. During the development of the network, three different types of configurations are tested. Extensive validation and calibration shows that neural networks are capable of making maximum water level predictions for a large number of cases. However, variety in the quality of the prediction is observed. Improvements still can be made for more accurate predictions.

For the design of coastal structures, the hydraulic loads that act on the coast should be known. These are often based on extrapolation of measurements. However, if the physical relation is unknown, large errors could be made. Therefore, a numerical model can be set up to take the most important processes into account.Yet, uncertainties are present in these numerical models as well. Taking for example the spectral wave model SWAN, it is hypothesised that the wave bottom refraction is not always properly modelled. Previous studies concluded that SWAN overestimates this refraction process near tidal channels, leading to an underestimation of waves entering the channel.This study uses the model SWAN in which different sensitivities are tested to assess the effect on refraction. The focus will be on the spatial and directional resolution as well as the wave propagation scheme that SWAN uses to discretise the propagation terms, which are the default SORDUP scheme and the optional BSBT scheme. From previous studies it is known that also physical processes could impact the bottom refraction, however these are only addressed shortly in this study. The aim of this study is to analyse the current performance of SWAN with respect to refraction and to assess how this performance can be improved.First, an analysis of a simple schematic channel case is made in SWAN, where waves propagate from shallow to deeper water. Physical source terms and wind wave-growth were deactivated in this part. It is found that a coarser spatial resolution can lead to a weaker wave refraction. This is caused by the way SWAN determines the turning rate, which is underestimated by SWAN compared to the theory of Snellius. On the other hand, an overestimation of wave refraction was observed for cases were waves should enter the channel. This is partly caused by the increased diffusion. These findings were taken into practice in a study of the Eastern Wadden Sea, specifically along the Ems channel, where waves propagated from deeper to shallower water. A coarser spatial resolution and the BSBT scheme showed less wave penetration from offshore into the Ems channel. The storm of January 2017 was simulated in SWAN after which a comparison was made to measurements. It was found that SWAN underestimated the wave energy for frequencies<0.15Hz. However, this did not improve by refining the resolution in both spatial and directional domain or by applying a different propagation scheme. SWAN thus overestimates the wave refraction due to bottom gradients if waves propagate to shallower water. In case waves travel to deeper water, refraction can be underestimated. If all physical source terms are deactivated, the effects of resolution on refraction are clearly visible. Including the source terms diminishes this effect. Along the channel edges, effects of increased spatial resolution are notable giving a significant wave height of 12-14cm (6-9%) closer to the measured condition. However, the effects at the coast near Uithuizerwad are negligible.

To calculate tsunami forces on coastal structures, the wave type in front of the coast is of great importance. Hence this paper aims to find ways to predict the type of tsunami wave breaking. Based on literature review, video footage, analytical reasoning and numerical modelling (SWASH) it can be concluded that both the continental shelf slope (alpha_2) and the bay geometry (beta) have a significant influence on the transformation of a tsunami wave near the coastline. After conducting 1D and 2DH wave simulations, a distinction is made in three types of tsunami waves; a non-breaking front (surging), a breaking front (plunging) and an undular bore breaking front (spilling). Tsunami waves transform into these three wave types for a steep continental shelf, an intermediate sloped continental shelf, and a gentle sloped continental shelf respectively. A new tsunami breaker parameter (xi_tsunami) is proposed to predict the type of wave at the coastline in a quantitative way.

This research describes a new method called SWDD (Klopman, Witteveen+Bos, 2018b), which can obtain information on wave propagation directions from the surface elevations at a set of positions. The primary intention of the method is to separate multiple incoming wave components, i.e. wave heights, phases and directions. The goal is to obtain the incoming wave conditions, that can among others be used in the design of (coastal) structures and assessment of moored ship response. The novelty of this method is that a large number of incoming wave directions is prescribed, equally distributed around a circle. For each of these many incoming wave conditions, the wave amplitude and phase are the unknowns (while the directions are known). The main advantage is that this makes the problem linear and in that aspect easier to solve. The disadvantage is that, most often, the resulting system becomes ill-posed (having more unknowns than equations). This problem is solved by using Tikhonov regularization (Tikhonov and Arsenin, 1977) together with the L-curve method (Hansen, 1992; 2000). The main differences with other common deterministic directional wave-analysis methods are: the SWDD method is free of user-checks after each analysis, the directional resolution is higher, the computation time is faster and a wave field reconstruction after the directional wave-analysis is possible. The applicability of the SWDD method has been tested using synthetic wave signals for (the sum of) monochromatic long-crested waves, prescribed wave patterns – containing wave-crest curvature and wave amplitude variation – and model results of a mild-slope wave model (WIHA). Multiple sensitivity analyses have been applied to check the sensitivity of the SWDD method to: various physical phenomena (e.g. diffraction and wave amplitude variation), domain variations (e.g. slopes) and input parameter variation. The study shows that the SWDD method is able to analyse irregular wave-fields using an array configuration containing a low number of gauges and a dense grid containing many gauges. Based on the findings an advisory flowchart is presented on how to determine the optimum radius of the array setup for using the SWDD method in practice, both for the analysis of data from phase-resolving numerical wave models and from measurements. The study shows that the SWDD method is a robust and reliable method to analyse (complex) wave fields on a (near) homogeneous bathymetry. The incoming wave direction(s) and associated wave height(s) are graphically depicted in a polar plot or a directional spectrum.

The Tohoku Tsunami of 2011 in Japan flooded a large part of the coastal area of Japan. The tsunami was caused by an earthquake with a magnitude of 9.0 just of the coast of Tohoku. The inundation height of the tsunami exceeded the design height of the tsunami barriers. This event led to thousands of fatalities. The aim of this study is to find the characteristics of the incoming tsunami waves for the design of coastal defence structures. This incoming tsunami wave close to the shore or on the shore is influenced by a lot of offshore factors that will change the wave and its behaviour. The tsunami wave will either develop into a bore or just run up the coast. This has large influence on the forces on the barrier. Potential influencing factors are examined on if and how they influence the tsunami wave when it travels to the coast. A numerical one-dimensional SWASH model is used throughout this study to simulate the tsunami wave. The tsunami factors and the factors that influence the wave were studied in several steps. The factors that have the most influence on the wave are used to simulate bores. From all these bores the important characteristics for the design of a barrier are investigated. These are the bore height, the bore velocity and the corresponding Froude number. With the simulations a new definition of the bore height is introduced. This is the height at the maximum velocity of the bore. The bore characteristics are also tested with an existing formula for the impact forces on a structure. The behaviour of the breaking wave is studied and a breaker parameter [ξtsunami] for tsunami waves is made. This breaker parameter defines if the tsunami wave develops into a bore before it reaches the coastline or that the wave runs up the coast without breaking. This is important for the location of the coastal structure. This breaker parameter and the Froude number of the bore give a relationship between the important parameters that influence the development of a bore and the characteristics of the incoming tsunami bore. Finally, physical tests were performed at the Waseda University in Tokyo, Japan, to simulate the bore attack on a coastal defence structure with a dam-break. The bore of the tests is compared to the bore from the SWASH simulations. This resulted that the velocities of the tests seem too high. However, with a new method to find the bore front characteristics is a Froude number constructed. This Froude number matches very well for the tests and SWASH simulations. The Froude numbers of the test represent a bore at the coastline.

Tropical Cyclones (TCs) are in many regions responsible for severe damages and a great number of casualties, resulting from the severe wind speeds, rainfall, wave heights and storm surges. The recent hurricanes Harvey and Irma are examples of the great devastation that can be caused by a single event, and of the insufficient level of preparedness that is currently present to withstand the adverse effects of such an event. Even in the U.S., which suffers relatively frequently from TCs, and has an extensive track record of historical events, the consequences of Harvey were unforeseen and destructive. Difficulties in determining extreme TC conditions, whether these are rain, wind, wave or storm surge conditions, mostly come from the fact that severe adverse effects caused by TCs are very local. This is the case because the exact track, intensity and size of the storm determine to a large extent which area is affected the most, and what the consequences are for the hydraulic conditions. Small variations in any of these parameters can greatly influence the conditions at any location. This is already the case in regions which suffer relatively often from TCs. In regions that do not suffer as regularly from TCs, data scarcity makes it even more difficult to anticipate adverse consequences. In these regions TCs are nevertheless often responsible for the most severe conditions, and the effects of such events should therefore be quantified in order be able to prepare for such conditions. For extreme cyclone wave conditions specifically, there is an additional problem of feasibly determining extreme wave conditions from cyclone wind conditions. This research presents the Tropical Cyclone Wind Statistical Estimation Tool (TCWiSE) to determine extreme TC wind speeds and focuses on the determination of its accuracy in data scarce regions. The tool was developed by applying it in a case study in the Gulf of Mexico. Moreover a brief qualitative assessment of the available TC wave models has been performed in order to identify an adequate method to determine extreme TC wave conditions from TC wind conditions.

The growth of the offshore wind industry results in intensive usage of the sandy seabed in the North Sea, currently and in the coming decades. Large-scale bed forms are present in shallow seas with sandy beds such as the North Sea. The most dynamic bed forms are sand waves. Due to their dynamic behaviour, sand waves can interact within offshore human developments and together with their dimensions pose a threat; e.g. decrease in navigation depth, exposure of submarine cables, interaction with foundations of offshore wind turbines and destabilization of bed protections. A thorough understanding of the dynamics can result in less risks for the offshore wind sector and therefore bring down the levelized cost of electricity from offshore wind. Currently, sand wave field dynamics are investigated by data-driven analyses. These analyses are based on seabed surveys over preferable more than 10 years and are considered most reliable at the moment. However, these surveys are very costly and/or often not available. Complex numerical models may provide an approach to analyse sand wave dynamics in a cost and time efficient way, though two aspects have to be considered. Not all relevant processes regarding sand wave dynamics are yet understood. Furthermore, due to the large scale of sand wave fields in combination with the fine grid resolution required to model sand waves, large computational efforts form a difficulty for numerical modelling of sand wave fields. Previous numerical studies focused on reproducing the length and height of sand waves. The migration direction is the next step towards the full prediction of sand wave fields and the subject of this research. Recent data-driven analyses showed migration directions of sand waves in opposite direction over small spatial scale, possibly related to underlying seabed topography. Understanding the governing processes of the migration direction of sand waves including underlying seabed topography is the focus of this research using the numerical process-based model Delft3D. To this end, an idealized model is used in which underlying seabed topography (tidal sand bank) is included. For a symmetrical tidal velocity signal, it is shown that the presence of the tidal sand bank influences the hydrodynamics on the scale of sand waves. Horizontal tide-averaged flow towards the top of the tidal sand bank on both flanks is observed. This results in sediment transports and migration directions on both flanks of the tidal sand bank towards the top of the tidal sand bank. The horizontal tide-averaged flow pattern around the tidal sand bank is disturbed by the inclusion of a residual current. Sand wave migration on both flanks in the direction of the residual current is the result. Including the S4-tide constituent does not disturb the tide-averaged horizontal flow pattern around the tidal sand bank. However, the asymmetry of the tidal velocity signal enhances migration in the direction of the asymmetry. Finally, it is shown that also for a more realistic model the transition in migration direction can be explained due to the presence of the tidal sand bank. The tidal sand bank influences the hydrodynamics by creating areas in which the tide-averaged sediment transport in the ebb direction are enhanced and areas in which the tide-averaged sediment transports in the flood direction are enhanced. In this way a transition in the migration direction over the tidal sand bank is observed. The migration directions from the model results and migration direction from data-driven analyses show a comparable transition over the tidal sand bank.

Large-scale measures within the river programmes such as Room for the River, Natura2000 and Water Framework Directive have increased the biodiversity in Dutch rivers and have achieved a more natural landscape. However, recent river programmes have shown conflicts between safety against flooding and riverine nature rehabilitation with maintaining navigational water depths. Therefore, an integrated approach is called out upon to improve the river management within these programmes. Conventional groynes are typically used to improve functions within the river programmes, whose primary function is to maintain navigational water depth. Groynes are transverse structures in which large turbulent structures are observed around the groynes. Consequently, significant bed shear stresses are developed, leading to significant local scour. Investigations for improving alternative groyne configurations are often studied for optimizing the groyne structure. The flexible groyne is considered an optimization of the conventional groyne structure. The flexible groyne consists of steep slopes and has a permeable characteristic. Although a fair amount of research has already been carried out on various groyne configurations, detailed flow characteristics around and through permeable, sloped groynes are limited. Hence, this research aims to numerically quantify the effects of permeability and head steepness of groynes. A literature study is executed to investigate the critical processes required for capturing within the numerical model to answer the research question. From the literature study, it appears that the most important processes are the non-hydrostatic effects, adapting large turbulent structures and implementing the porous zone using a non-linear function. Multiple software packages have been analyzed. Fluent is chosen, which is analyzed to capture the required processes adequately. A new numerical model has been set up for investigating the research question by simulating flow around groynes. The model is validated against three experimental studies for various characteristics. The important mean flow characteristics have been validated within an acceptable range. Still, it appears that the numerical model tends to underestimate the mean streamwise flow velocities, overestimate the Reynolds shear stresses and shift the peak values of the Reynolds shear stresses downstream. Four configurations are identified for the simulations of varying head steepness and porosity. It appears that the increase of the porosity reduces the large turbulent structures and bed shear stresses close to the groyne and shifts the peak values of the Reynolds shear stresses, and bed shear stresses further downstream. The porosity reduces the maximum Reynolds shear stresses and bed shear stresses compared with the Reynolds shear stresses, and bed shear stresses for an impermeable groyne. These reductions are because of the momentum exchange between the free flow region and the flow through the porous structure, which reduces the mean flow velocity in the free flow region. For decreasing steepness, the large turbulent structures and bed shear stresses are observed close to the groyne due to increasing deflection. The flow appears to follow the geometry of the sloped groyne more smoothly. This research is seen as a first approach for modelling a porous, sloped groyne. Further improving and analyzing numerical modelling for porous, head sloped groynes are advised to increase the model's accuracy. Furthermore, more simulations for varying the head steepness and porosity is suggested to improve the relations between the increase of porosity and head steepness for the flow characteristics. In addition, including sediment transport models within the model is expected to increase the understanding of the hydrodynamics and morphology around these specific groynes.

Mineral oil spills at sea can have many negative consequences. For both preventive and responsive purposes, it is essential to accurately forecast oil spill evolution. Shear diffusion (in this context, i.e. the combined effect of vertical mixing and differentiated horizontal advection of mass) determines for a significant part the evolution of an oil spill. These processes are partially forced by waves. A viscous fluid layer attenuates the waves throughout the area the layer covers. In this thesis, surface oil slicks are modeled as a continuous viscous fluid layer. It is investigated to what extent the wave-forced shear diffusion of the oil is affected by the oil-induced attenuation of the waves. For this purpose, the spectral wave model SWAN is extended with a module for energy dissipation due to a viscous fluid layer. The stationary, 1D wave energy balance is solved for uniformly forced waves in deep water. A high cutoff frequency (5 Hz) is employed to include the wave frequencies at which the dissipation is active. Also, special attention is paid to the choice of the wind and whitecapping formulation. Simulations are performed in full factorial setup, varying wind speed, oil layer thickness and oil viscosity. The results are compared to a no-oil case. Based on the difference, functions are fitted for the reduction of two key wave properties: the whitecapping dissipation rate and the surface Stokes drift velocity. The reduction functions are included in the oil spill module of the particle tracking model OpenDrift, which is subsequently used to calculate oil spill evolution due to shear diffusion for 2DV cases. The results of oil spill simulations with and without the implemented reduction functions are compared. Idealized cases (only wave-forced) show that for sufficiently thick layers (h+ ≥ O{10^-3} m) of sufficiently viscous (ν+ ≥ O{10^-3} m^2/s) oil, the Stokes drift reduction can significantly affect the wave-driven evolution of an oil spill in two ways: the average forward transport is reduced and the skewness of the oil mass distribution is increased to ‘less negative’ or even positive values. If simple sheared wind drift and ambient vertical turbulence are added, however, the relative importance of these effects becomes smaller. In none of the cases, a difference is found for the distribution of the oil mass between surface and subsurface, which implies that the whitecapping reduction hardly affects the results. It is recommended that further effort is put into obtaining a detailed understanding of the (differentiated) forward transport of the surface and near-surface oil, so that wave and wind effects can be distinguished, and modeled independently, more accurately.

The Noordstrand of the Marker Wadden has been subject to much more erosion than expected. From our current understanding of this low-energy system we cannot explain what has been the cause of this. The goal of this thesis is to create a better understanding of the different processes responsible for the development of this non-equilibrium beach in a low-energy lake environment. A better understanding of the link between hydrodynamic and morphodynamic processes can be used for the maintenance of the existing Marker Wadden beaches and the design of new sandy protections in this kind of systems. By combining the results of a hydrodynamic and morphodynamic data analysis with the results of the Delft3D Marker Wadden model, it is possible to create a conceptual system description. This can be used to improve our understanding of the Noordstrand development and the responsible processes. In this system description the influence of the general Markermeer system cannot be neglected as the different systems (Markermeer, Marker Wadden and Noordstrand) cannot be regarded separately as the different hydrodynamic processes are linked.

Across the world, the presence of humans in coastal regions is always increasing. Hydraulic forcing from extreme events is a large risk around the global coastlines, the risk being complicated by the increased human presence along the coasts. The knowledge that vegetation can be used as a coastal protection measure is not something new. The benefits of vegetation have been seen throughout history and are being researched on till date. If we know to a certain confidence what the wave heights are at the shoreline, coastal defence structures, used as a hybrid protection measure, can be designed accordingly. Not much research has been done on the wave behaviour on shallow foreshores. Localised studies have been previously done on tropical vegetated coasts, but there is a lack of efficient and accurate analysis on a large scale. As we bring more clarity on how waves transform and attenuate in a typically vegetated coast, some questions get answered, and some more questions arise, which also happened during the research done for this thesis. Observed data, be it laboratory or field, is very crucial in validating numerical models. A laboratory experiment was done in the TU Delft Laboratory of Fluid Mechanics flume, for a complete vegetation-free profile, where the surface elevations were observed for different wavemaker input conditions. A lot of numerical models have been developed that predict wave transformation and dissipation through vegetated foreshores. However, these models lack validation from observed data. This thesis first focuses on understanding the wave transformation for two unique (and mainly theoretical) wave conditions: a regular sinusoidal wave and a bichromatic wave. It was checked if the transformation is reflected in the models – SWAN and SWASH, which they did. The research proceeded on to validating the models by comparing the wave heights observed in the laboratory experiment versus when the models were inputted with the same conditions, including inputting the observed data into the models. When the laboratory conditions were replicated, the SWASH results obtained correlated quite well with what was observed in the laboratory. The same was not true in the case of SWAN. When a spectral analysis was done for the observed data, a presence of very low frequencies (VLF) as well as some minor higher frequencies was noticed. To check its effect, if any, on the model results, they were filtered out. Both the original and filtered data was inputted into the models. The difference in the foreshore region was more distinct in the filtered case, i.e., making a bichromatic elevation input purer resulted in more pronounced undulations in the wave heights than what was predicted in the unfiltered data. This result does not fit well with the existing knowledge on wave dissipation processes. It is widely known that the presence of VLFs and higher frequencies are the driving mechanisms that result in the undulations in the foreshore region, but the predicted results were exactly opposite to this knowledge. What can be thought of from the anomaly is that the presence of various frequencies (that is, waves with different periods) counteract each other’s effects and make the undulating wave heights milder, but when the signal is made purely bichromatic, it leads to more distinct undulations. This proposition is also backed by the similar SWASH results for the laboratory condition-replicating theoretical inputs. This anomaly needs further investigation. Another interesting observation was that the changes happening in the offshore region did not affect the results in the foreshore region, for varying parameters in SWAN. SWASH can be concluded as a better model for predicting wave heights, especially in the foreshore region. SWAN could not predict the fluctuations in the wave heights. Obtaining the wave heights at the shoreline with SWAN, and designing a dike with those results, for example, will lead to disastrous consequences, as SWAN underestimates the wave heights. The study is limited by the consideration of hydrodynamics only, and by the many simplifications made to simulate the conditions. One of the recommendations formulated is to obtain field data and to make a similar comparison with the models to corroborate (or correct) the observations made. This study tried to see the correlation between the models and observed data in the laboratory, for simple (and somewhat purely theoretical) cases. It is, nonetheless, a starting point for more complicated cases, the basis for which can be laid on this study.

In the past years, climate change has caused an increasing number of extreme weather events. In the Netherlands this is, for instance, reflected by the growing amount of extreme rainfall events. This creates big issues for water management authorities like water boards. Water boards are often responsible for local streams, that cannot handle the peak discharges caused by extreme rainfall. This raises the question whether measures have to be taken, for instance by creating additional storage areas to decrease discharge peaks. For questions like these computational model studies can be used to gain insight in the distribution of discharges and water levels. During these model studies often one-dimensional models are used. Recent developments made two-dimensional modelling, including for instance the simulation of inundation, possible for regular users like water boards. Two-dimensional modelling is proved to be more accurate for rivers, compared to one-dimensional modelling. Water boards, however, often deal with water systems of smaller scale. This raises the question whether it is worth for authorities like water boards to invest their resources in this new modelling approach. Several physical system properties are of influence on the answer to this question. Therefore, the main question of this research is: Which physical processes are of influence to the choice of a modelling approach, one or two dimensional, for the modelling of stream systems in particular situations? The methodology of answering this question consists of three major parts: 1) identify possible differences in the model properties of 1D versus 2D modelling; 2) analyze the output of test models to see whether the differences in model properties actually influence the effects of physical processes; and 3) analyze a case area to see whether these processes are important in practice. The research scope consists of the hydrodynamics of systems with a length scale of 10^3 – 10^5 m and a temporal scale of hours to a few days. This range corresponds with a range in discharges of a few cubic meters per second up to a few thousands of cubic meters per second. The choice for this range has been made to be able to analyze the effects between the scales of streams and rivers, as until now mostly rivers have been modelled using 2D modelling approaches. The starting point of this research has been the analysis of differences in model properties. The used flow equations, numerical methods, computational grids and bathymetry implementation are analyzed and the differences are discussed. Concluding, it can be stated that there are differences, which may be of influence to the model output. The main differences are the number of dimensions on which the flow equations are calculated and the way the bathymetry is schematized by the computational grid. Using these differences, physical parameters have been specified that were varied in a range of test models. These are mainly the parameters that cause 2D effects, like meandering and varying roughness coefficients. The modelling of flood waves could also cause differences due to the fact that the 1D model is not able to model storage with the used default settings. Lastly, the way the bathymetry is used in the simulations is of great importance, due to the way the bathymetry is discretized in 2D models. The goal of varying these physical properties is to analyze whether these variations would influence the quantity of the effects the differences in model properties are causing. Comparing the output of both a 1D and 2D model, it appeared that there are significant differences in the observed water levels. Generally speaking, the following observations were made: 1. The smaller the simulated spatial scale, the larger the 1D-2D differences are 2. The higher the meandering intensity, the larger the 1D-2D differences are 3. The lower the summer bed roughness, the larger the 1D-2D differences are 4. The larger the 2D grid cell size, the larger the 1D-2D differences are 5. The smaller the flood wave duration, the larger the 1D-2D differences are The underlying physical processes have been analyzed to see how their influence varies in different situations and for different spatial scales. In the end, six processes were found to be influencing the 1D-2D differences: 1. The effects of grid size on bathymetry discretization 2. The water level and velocity difference in cross-sectional direction for varying roughness 3. The water level and velocity difference in cross-sectional direction for varying meandering intensities 4. The interaction between summer bed and floodplains 5. The Boussinesq approach and consequences for the summer bed – floodplain flow area ratio 6. The effects of storage When combining the findings of the test models and the analysis of a small scale stream system, it was found that the above processes are of the most influence on the following situations: 1. In situations with a complex bathymetry the process of grid discretization is having the most influence. 2. In situations with a high meandering intensity two physical processes are playing a role: high water level and velocity gradients in lateral direction, and a high summer bed – floodplain interaction. 3. In situations with a very low roughness compared to the surrounding area two physical processes are playing a role: high water level and velocity gradients in lateral direction, and a high summer bed – floodplain interaction. 4. In situations with a lot of elevation changes and open water bodies, a lot of potential storage is present. In these four situations a 2D approach is preferred, as the physical processes that are present in these situations are all processes that are poorly implemented or represented by a 1D model. Of course, there are also some more practical considerations to be made about the modelling choice. The most important one is the research goal; whether inundation modelling is desired and what the required model output is. Another important consideration is the available time for running the simulations; 2D models have a significantly longer simulation time compared to 1D models. All things considered, this research provides an overview of the differences in model properties between 1D and 2D modelling approaches, an analysis of the important physical processes that determine the choice of a modelling approach and some recommendations about practical issues one could encounter when performing a modelling study.

To counteract the effect of climate change, a global agreement was put into force in 2016, aiming to limit the increase in average global warming and reduce greenhouse gas emissions. In that respect, European countries are investing in cleaner sources of energy and predominantly offshore wind, which has seen a rapid growth over the last years. The North Sea is a region suitable for developing offshore wind energy, given its strong wind climate and the relatively shallow waters. Many large-scale offshore wind farm (OWF) developments are currently ongoing in that region, while many more are planned and consented for the coming decades. The concept of large-scale OWF developments has spurred many discussions addressing its potential effect on the greater North Sea region. However, the long-term effect on the surrounding coastal areas has never been studied in detail. Especially for a low-lying country as the Netherlands, assessing the impact of large-scale OWFs is of great importance. This study aims at exploring the effect of future large-scale OWFs in the North Sea, focusing on the wave climate and the coastline response of the Dutch coast. Based on the roadmap for developing offshore wind energy until 2050, existing and future designated OWF areas are accounted in the North Sea region. The effect of OWFs on wind is introduced in a schematized way, with a constant decrease in wind speed of 20% inside the OWF areas, based on literature knowledge. Supplementary, based on the vision of creating an artificial energy island for storing and redistributing the wind farm generated electricity, a 5 km2 island is introduced, approximately 30 km away from the Dutch coast. The effect on the nearshore wave climate is studied using the numerical model SWAN, while the resulting effect on the alongshore morphology is assessed using coastline model Unibest-CL+. The impact of future OWFs on the nearshore wave climate is found to be dependent on the size, shape, orientation and distance from the coast of the individual wind farms. Results show a mean decrease in significant wave height in the order of 1 – 2%. In addition, slight changes in wave direction are observed. The effect on wave climate reduces the alongshore sediment transport at the Dutch coast, by an order of 10% with respect to present values. This results in net-induced erosion, which requires nourishment. The study shows that the areas north of Zandvoort and Petten need the greatest nourishment volumes, in the range of 1.5 – 2.5 m3/m/year. This is an additional 1% on the current annual nourishment volumes supplied along the Dutch coast. The underlying study has proved to be effective in quantifying the chain of effects of OWFs and identifying potential hot-spots along the Dutch coast. The knowledge acquired from these effects can be used to optimize future OWF planning in relation to coastline maintenance policies.