Validation of Hindcasting Results of a Wave Model—FVWAM using CFOSAT/SWIM Wave Data

A Finite Volume WAve model (FVWAM) which adopted a novel spherical centroid voronoi tessellation computational grid and an advanced GPU acceleration algorithm was developed by National Marine Environmental Forecasting Center, China. The hindcasting performance of this model was validated using CFOSAT (China France Oceanography SATellite)/SWIM (Surface Wave Investigation and Monitoring) wave data, in which significant waveheight and wave period generated by FVWAM were compared with corresponding wave information retrieved from SWIM measurements onboard CFOSAT. Specifically, wave period was retrieved using an AI method extending the validation range with satellite data. These data were compared monthly and for a whole year. Results showed that model simulations and satellite measurements were in good agreements with average bias, RMSE and correlation coefficient of -0.11 m, 0.29 m and 0.96 for significant waveheight and average bias, RMSE and correlation coefficient of -0.34 s, 0.86 s and 0.86 for mean wave period.