Spectral modeling of wave propagation in coastal areas with a harbor navigation channel
This study presents a comparison of numerical model results and laboratory experiments of wave propagation in a coastal area with a harbor navigation channel. The results of wave models HARES, SWAN and SWASH are compared with physical model results in order to investigate the performance and accuracy of these models.
The finite element model HARES is a stationary phase-resolving 2D parallel wave model based on the Mild-Slope Equation, developed in-house by Svašek Hydraulics.In recent years 3D non-hydrostatic wave models like SWASH have been increasingly used. Conceptually, such full 3D models are more accurate because they take into account all non-linear wave propagation effects; a downside however is the large amount of computational time needed in practice due to resolution requirements. Another model type often applied in near-shore areas is given by phase-averaged spectral wave-energy models like SWAN.
Compared to the laboratory experiments, it turns out that HARES (and to a lesser extent SWAN) results contain more wave energy and are generally more accurate than SWASH results; the latter show significant damping of high-frequency waves and hence a systematic underprediction of the wave energy. The amount of wave energy crossing the navigation channel as computed by the phase-resolving model HARES is in accordance with the physical experiments, whereas the phase-averaged spectral model SWAN underpredicts this wave energy flux across the channel. The average error over all wave measurement locations was found to lie between 15% and 24% for SWASH, between 18% and 20% for SWAN and between 3% and 10% for HARES results. The computational time on a 16-core cluster-computer was about 20 minutes for both SWAN and HARES, which was 0.7% of the SWASH computational time (about 50 hours).
In view of the above observations, we may state that mild-slope models combine “the best of both worlds” conceptually. On the one hand, the 2D time-independent spectral approach yields an efficiency advantage above 3D time-dependent models; on the other hand, the phase-resolving mild-slope approach yields an accuracy advantage above phase-averaged spectral models for complex geometries like harbors with approach channels.
We conclude that wave models based on the Mild-Slope Equation, in combination with efficient and accurate spectral treatment of bottom friction and wave breaking based on the entire wave spectrum, are quite competitive to numerical models that are conceptually more sophisticated. In practice, mild-slope models like HARES remain a preferable tool for the design of harbor layouts.
Read full article.