Skill assessment of different quadruplet wave-wave interaction formulations in the WAVEWATCH-III model with application to the Gulf of Mexico

Abstract

In this study, the Gaussian Quadrature Method (GQM) method was implemented in the WAVEWATCH-III model and used along with embedded formulations for estimating the nonlinear wave interactions. The results of DIA, GMD, GQM and WRT methods were compared for ideal test cases. The GQM was in good agreement with the exact WRT method while its medium resolution conjuration was ∼10 times faster than WRT. The GMD was more than 50 times faster than GQM but it could not reproduce WRT results for slanting fetch condition. Two different packages named ST3 and ST6, were employed for calculating the wind input and energy dissipation formulations over the Gulf of Mexico. The high quality ERA5 wind data from August to September in 2017 blended with Holland parametric model were used to run the wave model. The simulation period includes both fair weather condition and extreme events of Hurricanes Harvey and Irma. The performance of wave model using different nonlinear wave interaction terms was assessed against bulk wave parameters measured by in-situ buoys as well as altimeter-derived data by introducing a new error index. The general features of in-situ directional wave spectra were well captured by all DIA, GMD, GQM and WRT methods; however, the spectra produced by GMD and GQM were closer to the results by WRT method. Interestingly, the DIA method with calibrated whitecap dissipation term outperformed other methods in reproducing in-situ data (i.e. wave bulk parameters and wave spectra) during both fair weather and extreme events; indicating overfitting in the white capping or/and wind input terms in the wave model.