Transformation of wave spectra across the surf zone

Abstract

One of the most applied models for wave prediction is the spectral wave model SWAN. SWAN is not always capable of predicting wave conditions in the surf zone with sufficient accuracy. In this study the transformation of wave spectra across the surf zone and its spectral modelling was investigated. The objective of this thesis work was to improve the prediction of wave spectra and their integral parameters in the surf zone. Focus was on the improvement of the dissipation term for depth-induced wave breaking for the use in SWAN. A literature study was carried out to make an inventory of previous work on wave breaking and its spectral modelling.In order to select the most promising alterations wave data was analysed and simulated with SWAN 40.11. It appeared that SWAN consistently underestimated the significant wave height in the surf zone. Three different adaptations to the present source term for wave breaking in SWAN were proposed, implemented and tested. Firstly, the Rayleigh distribution for both breaking and non-breaking waves was implemented. Secondly, the breaker parameter r was adapted. The third adaptation concerned the distribution of energy dissipation over the frequencies. It was concluded that the prediction of the cross-shore variation of the significant wave height can be considerably improved with the implantation of the breaker parameter depending on the local wave steepness. Applying a full Rayleigh distribution improves the accuracy in the prediction of the fraction of breaking waves and the significant wave heights on a steep slope. Moreover it was concluded that the distribution of energy dissipation is frequency dependent and that distributing energy dissipation proportional to the energy density at each frequency is a wrong assumption. Dissipation of energy is higher at higher frequencies. Adding a frequency-dependency to the source term for wave breaking had an effect on predicted mean wave periods and spectral energy density levels although no unique optimal formulation could be found for this frequency dependency. Furthermore it was concluded that the effect of triad interactions is not correctly represented in SWAN. The implementation of both the Rayleigh model and the formulation for the breakerparameter depending on the local wave steepness as an optional command in SWAN is recommended. The effect of wave breaking on the spectral distribution of energy is still poorly understood. Investigation of the spectral energy balance of breaking waves in the surf zone might give more insight in the spectral representation of wave breaking. Besides wave breaking, non-linear triad interactions are very important regarding the transformation of wave spectra across the surf zone. In this study triad-interactions are almost not considered. When triad interactions are better approximated in SWAN the representation of wave breaking in SWAN can be assessed better. Further investigation of the spectral modelling of triad interactions is recommended.