This thesis discusses the morphological modelling of the gravel revetment on artificial composite beaches. A composite beach is a combination of a sandy lower beach and a gravel upper beach. These beaches are found in nature, but can also be created artificially by placement of a gravel structure on a sandy beach. Artificial composite beaches are widely acknowledged as an effective type of shore protection against runup and overtopping (Blenkinsopp, 2016), but an assessment of such a beach with a numerical model has not been attempted yet.

Therefore this work explores to what extent the morphological developments of these artificial beaches can be modelled with a numerical model. The numerical models that were used in this work are XBeach and XBeach-G. The aim is to reproduce the short term (3 to 6 hours) cross-shore morphological developments of the DynaRev experiments (a series of physical model experiments to investigate the stability of an artificial composite beach under sea-level rise), with a focus on the beach section on which the gravel revetment is placed. The first step was to find the relevant processes that are missing in the currently available numerical models. The missing processes were identified by combining knowledge acquired through the modelling of composite beaches with the currently available models, substantiated by analysis of the DynaRev experiments and literature research.

It was concluded that the numerical models were lacking two important processes: namely the absence of a gravel transport formulation in XBeach and transport of sand in the gravel revetment. The missing processes were implemented into XBeach and the updatedmodel’s performances were verified with the DynaRev experiments as benchmark. The implementation of the XBeach-G gravel transport formulation in coherence with the existing XBeach code for sand transport required under-the-hood adaptations to improve the switches that are already in place for multiple sediment fractions. It was found that a combination of sand transport and gravel transport is possible, but this combination presents difficulties regarding suspended sediment transport in combination with the hydrodynamics and the groundwater dynamics.

The second process was transport of sand in the gravel revetment. This was first analysed by looking at the initial transport rates of sand inside the revetment with a newly introduced equation for transport of sandinside a filter layer (Jacobsen et al., 2017). It was shown that transport fluxes of sand inside the revetment are likely to occur as results of the groundwater dynamics. To see bed level changes due to these transport fluxes this transport equation needs to be implemented into the XBeach code. This was not possible with the current architecture of XBeach and the way it accounts for multiple fractions. Therefore a new accounting system for multiple fractions was introduced in this work and implemented in XBeach, named the two-line model. In this experimental model sand transport gradients in the revetment are translated into visible sinking of the revetment (erosion) and settling of sand within the gravel revetment (deposition). In the model’s current state the erosive locations appear to match with the DynaRev observations, whereas locations where deposition occurs are no match. The two-line model was sensitive and showed instabilities, mainly due to high groundwater velocities that were caused by wave breaking in themodel. The results of the experimental model can possibly be improved by including vertical groundwater velocities to model sand transports, possibly also in combination with the addition of suspended sediment transport of sand inside the revetment.