Wave overtopping in harbour areas is a fairly untreated topic in the context of flood risk analyses. Pragmatic approaches were undertaken to evaluate the possible influence of wave overtopping on the frequency of flooding. However no consensus was reached on how wave overtopping in a harbour area could be quantified. Therefore it is also difficult to determine whether wave overtopping should be accounted for in a flood risk assessment or not. This master thesis seeks to find an answer on the question which method is best suitable to quantify wave overtopping in harbour areas. Along the way it is studied which wave processes play a role and how available models take into account these processes. Subsequently a choice for a certain method or model can be made. Using this approach the question whether the effect wave overtopping might be significant in a flood risk analysis is answered. At first a theory study is undertaken which consists of two parts. The first part examines the theory on the wave processes which play a role in the nearshore and inside harbours. The second part explores the capabilities and limitations of empirical and numerical modelling of wave heights and wave overtopping in harbour areas. It was chosen to develop a non-numerical model within the scope of this thesis. The available numerical models are either not capable of modelling all relevant wave processes or model an entire harbour area in a sufficiently efficient way. These considerations might change by future updates of the discussed numerical models or the development of a new and suitable numerical model. Furthermore the computational speed is expected to keep growing, such that the problem of a lack of computational efficiency might decrease in the future. The developed non-numerical WGPO model considers the most relevant processes, namely: wave generation, wave penetration and wave overtopping. Wave generation is estimated using the empirical Bretschneider formula. Wave diffraction and wave transmission and the interaction between the two are the considered components to determine the contribution of penetrated waves. The wave diffraction pattern is determined using the analytically derived Goda diffraction diagrams. Finally the expected amount of wave overtopping over the quay walls of the harbour is estimated using the empirical EurOtop formulas. Using the case of the Vlissingen Buitenhaven the performance of the WGPO model was verified with respect to earlier researches. Due to the assumptions which lie at the basis of the model it is not (yet) possible to model complex harbour geometries. This would require improvements to the WGPO model or the application of a numerical wave model. By applying the model to the case of the Brittaniëhaven in the Botlek area of the harbour of Rotterdam the influence of wave overtopping on the probability of flooding was estimated. It appears that the frequency of flooding of the harbour areas may double. In conclusion, this case study of the Brittaniëhaven thus shows the importance of taking into account wave overtopping in a flood risk analysis of a harbour case.

Additional thesis - The Rhine river outflow has a major impact on the North Sea in front of the Dutch coast. It creates the Rhine ROFI (region of freshwater influence), a very complex three-dimensional volume of water with a relatively low salinity. Many researches have been conducted on this phenomenon. Now another complex factor is added: a non-straight coastline. The specific case of the Sand Engine, a sandbar-shaped peninsula in front of the Holland coast, is studied. The objective is to identify changes in the Rhine ROFI caused by the Sand Engine and their possible causes. Simpson (Simpson, et al., 1993) and De Boer (2008) have identified several mechanisms influencing the shape and size of the Rhine ROFI. The major ones are: the deflection of the fresh water jet from the river Rhine through the Coriolis force towards the north forming a ‘coastal river’ of fresh water, advection due to tidal propagation, density currents as a consequence of horizontal density gradients, tidal straining and tidal mixing. Fortnightly and semidiurnal variations of velocities and stratification can be expected within the Rhine ROFI. Signell (1989) has had major contributions to the understanding of tidal propagation around coastal headlands. In his work a categorization is made for different combinations of tidal conditions and headland shapes. Within this categorization the Sand Engine is seen as a rather small and streamlined headland. Flow separation may be expected with a stagnant lee-side eddy forming each tidal period. These researches cover the main domain of interest of this thesis. An extension is made by exploring the baroclinic effects of the perturbation of the coast to the Rhine ROFI. For this purpose a numerical model was set up. In essence the original model De Boer used in his dissertation was applied. The Sand Engine was added as a blunt rectangular shape at the same distance from the river mouth as in reality. No numerical problems were found after adapting the model. The performance of the model was successfully verified by comparing time-averaged salinity distributions to figures in De Boer’s work. The Rhine ROFI and the ‘coastal river’ remain largely unchanged, comparing outcomes of the models with and without the Sand Engine. On a more detailed scale some interesting phenomena can be distinguished. During both neap and spring tide a fresh water feature forms ahead of high tide at the location of the Sand Engine. The offset of the fresh water feature has a barotropic origin, being generated by a strong current at the southwest corner of the Sand Engine. Under neap tide conditions this fresh water feature grows in the offshore direction, whereas this does not occur for spring tide conditions. This offshore advection is a baroclinic effect as such widespread offshore velocities involved with the offshore advection of fresh water were not found under barotropic conditions. A possible explanation of the offshore velocities is the strength of the tidal straining effect, being enforced by the stronger vertical density gradient when the fresh water is located at the Sand Engine. Tidal mixing is larger under spring conditions, preventing tidal straining from happening, explaining why offshore-directed velocities and subsequent offshore fresh water advection are not found in the results. In this thesis a simplified approach has been applied. Therefore the results must be treated with care under the likely possibility that flow mechanisms have been altered, enhanced or ruled out. However, the results do show the likely importance of baroclinic effects. These effects may have large consequences on the hydrodynamics in the area surrounding the Sand Engine. Therefore it is recommended to perform additional research on this topic.