An operational forecasting system for physical processes in the Santos-Sao Vicente-Bertioga Estuarine System, Southeast Brazil

We describe the design, implementation, and performance of a fully automated Santos Operational Forecasting System (SOFS), built to monitor and predict short-term (< 3 days) sea surface elevations, currents, temperature, and salinity in the Santos-Sao Vicente-Bertioga Estuarine System (SSVBES). The SSVBES located at 24.0 ∘ S, 46.3 ∘ W is a complex estuarine system with many interconnected channels and two connections with the open sea. The system is prone to storm tides that bring coastal flooding to and interrupt ship traffic through Santos Port. The SOFS hydrodynamic module is based on the Princeton Ocean Model (POM) version POM-rain. The SSVBES model grid is forced by tides, winds, and river runoff and is nested into a coarse-resolution South Brazil Bight (SBB) grid. The SBB grid is forced by winds, density gradients, and the Brazil Current flowing offshore. Within SSVBES, SOFS works in parallel with three real-time observation stations. The model performance was tested against observed data with a best Willmott skill of 0.97 and root mean square error (RMSE) of 13.0 cm for tidal sea level (15.9% of the mean tidal range). For tidal currents, the best skill and RMSE were above 0.99 and 3.9 cm/s (4.3% of the mean tidal current range), respectively. The coupled system was able to simulate seven storm tides with average skill of 0.95 and average RMSE of 17.0 cm. The good agreement with observed data shows the potential use of the designed system to protect both human life and assets.