Implementation of a multi-grid operational wave forecast in the South Atlantic Ocean

This paper presents the development of an operational wave forecast in the South Atlantic Ocean. The system adopts the WAVEWATCH-III model version 6.07 using ST4 physics and its implementation follows four steps: (1) assessment of wind forecasts products (GFS/NOAA and ICON/DWD); (2) investigation of source-terms in the wave model; (3) sensitivity tests involving spectral and spatial resolutions combined with analysis of different wind inputs and calibration parameters; (4) definition of the best mosaic of multiple grids and fine-tuning of WAVEWATCH-III for each grid and wind input, using 6-month simulations assessed against altimeter data. The forecast evaluation and discussion are performed in terms of wind speed (U10) and significant wave height (Hs). The balance between model accuracy and computational cost is investigated. The results highlight the importance of the multi-grid scheme in WAVEWATCH-III, even at large domains in deep waters. It enabled the South Atlantic, with very heterogeneous metocean conditions, to be divided into two grids where the best resolution and source-term parameters could be obtained. Furthermore, it reduced the computational cost, which allowed the implementation of two independent wave forecasts per cycle, forced by the high-quality winds GFS/NOAA and ICON/DWD. The validation for Hs showed correlation coefficients around 0.96, scatter errors of 11%, and RMSE between 0.32 and 0.34 m. This performance is compromised at longer forecast leads associated with strong winds.

Automated summary

This paper presents the development of an operational wave forecast system in the South Atlantic Ocean, which achieved a balance between model accuracy and computational cost by using a multi-grid scheme and fine-tuning the WAVEWATCH-III model. The forecast evaluation was performed based on wind speed and significant wave height, and the results showed high correlation coefficients and low errors.