Assessing wind datasets and boundary conditions for wave hindcasting in the southern Brazil nearshore

This work investigates the sensitivity of the SWAN model with respect to state-of-the-art wave and wind datasets. The wave model was implemented and validated for the Southern Brazilian nearshore environment and simulated for 2015. Nine combinations of wave and wind datasets have been considered, aiming at establishing the boundary conditions and atmospheric forcing which best reproduce windsea and swell systems. The wave model results have been validated comparing against two buoys deployed in deep water and one buoy in shallow water. Wave parameters such as significant wave height Hs, mean wave direction Dm and peak wave period Tp have been computed and compared. Overall, the performance of the SWAN model satisfactorily reproduce the distinct waves fields for all wind and wave databases. However, the combination of datasets with ERA5 winds and CAWCR wave boundary conditions slightly outperforms the other combinations tested. -The use of curvilinear mesh in a long-term wave hindcast. -Model validation with a state-of-the-art wave and wind databases. -The most accurate combination of wind forcing and wave boundary conditions are established. -Assessment of using whole waves spectrum vs bulk waves parameters as boundary conditions.