Optimization of an operational forecast system for extreme tidal events in Santos estuary (Brazil)

• Main Author: Mendes, Joana Alexandra Assunção
• Publication Date: 2018
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10773/27065
• topic: Santos estuary; Hydrodynamic; Numerical modelling; Operational oceanography; Boundary conditions

Forecasting estuarine circulation is in high demand, especially in regions of high population density like Santos region. Present work intends to optimize the performance of the water level forecast for Santos Estuary, particularly, the physical forcing determining the residual tide. In extreme cases, predicted and observed water level can significantly differ, increasing significantly the predicting errors, which highlights the need to understand the factors that affect the residual tide to propose corrective measures. Furthermore, the extreme events have potentially hazardous consequences not only for navigation purposes, as also to the population that live nearby the channels. This dissertation analyzes the water level and significant wave height dataset covering the period of 2016 to 2017. Datasets comprehend 5 tide gauge stations in Santos channel and were obtained from the Pilotos da Barra and Praticagem. MOHID 2D hydrodynamic model (www.mohid.com) was used, implemented with a nested downscaling approach, being linked to AQUASAFE software that provides high-resolution forecast to give support to navigation in real-time. The model was validated for the period of 2016-2017, being minimum RMSE found of 12.5 cm and practically all stations present an excellent SKILL. The most recent astronomical global solution (FES2014) was implemented as oceanic boundary condition. Additionally, the meteorological boundary condition (CMEMS) was altered from daily to hourly data. It was also researched the possible influence of wave height on the forecast of water level (particularly the residual tide). Due to the great correlation found among these parameters, a linear regression method was developed to correct in post-processing stage the residual tide under specific wave height conditions. The appliance of distinct boundary conditions on forecasting models (meteorological and astronomical) decreased errors when compared to observations, evidencing the improvement of forecast capacity. On the other hand, the use of FES2014 shows improvements at the bay entrance, however, results get worst in the inner stations. This portrays the need of reliable bathymetric data due to increasing errors on the astronomical tidal components towards the end of the estuary. Residual tide errors remain practically constant along the estuary, however, they increase under extreme conditions. The results evidence that model modifications improve the accuracy in reproducing the water level evolution, comparing to the previous version, particularly under extreme events