Modeling of coastal infragravity waves using the spectral model WAVEWATCH Ⅲ

Abstract

Forecasting coastal infragravity (IG) waves is crucial in preventing coastal disasters caused by IG waves. In this study, the performance of the unstructured WAVEWATCH Ⅲ (WW3) in predicting coastal IG waves is comparatively assessed using in-situ wave measurements, considering the effects of numerical schemes, time steps, and IG-source parameterization methods. The wave measurements are performed using three acoustic sensors installed at the coast of south Sri Lanka, with two being near or inside a harbor.Sensitivity analysis of the numerical schemes and time steps shows that the explicit scheme combined with a small time step leads to unexpected oscillations of the IG wave height. However, the time step has a negligible influence on the performance of the implicit scheme combined with the domain decomposition method. Generally, IG sources estimated from the empirical formula method (source M1) and the second-order spectrum method (source M2) perform similarly in reproducing the IG wave height. Regarding the IG frequency spectrum, both the IG sources properly predict the magnitude of the spectral density, but the spectral shapes are not adequately captured. M2 may provide a better directional spread of the IG wave energy. Attention is also paid to the IG waves inside the harbor. The IG waves within the harbor are underestimated, owing to the lack of some mechanisms related to IG waves in the simulation. M1 should be adopted when crudely predicting the magnitude of IG waves inside a harbor. Overall, the unstructured WW3 can effectively predict the coastal IG wave height in southern Sri Lanka, with significant correlation and acceptable errors, when compared with in-situ measurements. Finally, the IG source M1 is refined to remove the erroneous spectral gap between the IG and wind-wave frequency bands in the WW3 spectra when compared to the in-situ observed spectra.