Wave propagation at coast combining numerical and statistical models

In this work, a hybrid model that combines numerical and statistical models is provided in order to study the propagation of the wave parameters from offshore areas to coastal areas. The methodology utilized involves a preliminary aggregation of 6 buoys data around an island in the Pacific Ocean in a single point. In the ocean environment studies usually the aggregated parameters of significant wave height, peak period and mean direction are utilized: in this work is performed also the separation between swell and wind sea waves. The time-series is 40 years long, thus 500 cases are selected utilizing the maximum dissimilarity algorithm, which represents different scenarios. After that, this cases are propagated utilizing the software SWAN, which is a phase-averaged resolving model. The inputs are the wave parameters of the 500 cases and the wind velocity, while the results are the wave parameters propagated in each point of a map. Three different maps have been used, with increasing resolution of the bathymetry. In the end, the output of a specific point has been extracted and, with the application of the radial basis function algorithm, the original time series has been rebuilt. The values obtained have been validated with the buoy data present. After, two sets of other simulations have been performed varying the mean sea level in order to have a qualitative analysis of the influence of the tide level on the significant wave height inside the lagoon.