Wave hindcast under tropical cyclone conditions in the South China Sea: sensitivity to wind fields

Reliable wave information is critical for marine engineering. Numerical wave models are useful tools to obtain wave information with continuous spatiotemporal distributions. However, the accuracy of model results highly depends on the quality of wind forcing. In this study, we utilize observations from five buoys deployed in the northern South China Sea from August to September 2017. Notably, these buoys successfully recorded wind field and wave information during the passage of five tropical cyclones of different intensities without sustaining any damage. Based on these unique observations, we evaluated the quality of four widely used wind products, namely CFSv2, ERA5, CCMP, and ERAI. Our analysis showed that in the northern South China Sea, ERA5 performed best compared to buoy observations, especially in terms of maximum wind speed values at 10 m height ( U 10), extreme U 10 occurrence time, and overall statistical indicators. CFSv2 tended to overestimate non-extreme U 10 values. CCMP showed favorable statistical performance at only three of the five buoys, but underestimated extreme U 10 values at all buoys. ERAI had the worst performance under both normal and tropical cyclone conditions. In terms of wave hindcast accuracy, ERA5 outperformed the other reanalysis products, with CFSv2 and CCMP following closely. ERAI showed poor performance especially in the upper significant wave heights. Furthermore, we found that the wave hindcasts did not improve with increasing spatiotemporal resolution, with spatial resolution up to 0.5°. These findings would help in improving wave hindcasts under extreme conditions.