Wave setup is a critical component of coastal flooding. Consequently, understanding and being able to predict wave setup is vital to protect coastal resources and people living near the shore. Here, we applied machine learning to improve the accuracy of present predictors of wave setup. The results show that the new predictors outperform existing formulas demonstrating the capability of machine learning models to provide physically sound descriptions of the processes modelled.

For predicting flooding events at the coast, bathymetric data (or depth/elevation information) are essential. However, elevation data can be unavailable. To tackle this issue, recent efforts have centred on the use of satellite images to estimate bathymetry (SDB). This work is aimed at evaluate SDB’s accuracy and its use for water level prediction of estuaries. Results show that the use of SDB in numerical modelling can produces similar predictions when compared to the surveyed bathymetry data.

The barrier tidal basin is increasingly altered by human activity and sea-level rise. These environmental changes probably lead to the emergence or disappearance of islands, yet the effect of rocky islands on the evolution of tidal basins remains poorly investigated. Using numerical experiments, we explore the evolution of tidal basins under varying numbers and locations of islands. This work provides insights for predicting the response of barrier tidal basins in a changing environment.

Sandbars are ubiquitous features of the surf zone. They are rarely straight and often develop crescentic shapes. Double sandbar systems are also common, but the possibility of feedback between inner and outer sandbars has not been fully explored. The presence of double sandbar systems affects wave transformation and can result in a variety of spatial patterns. Here we model the conditions, waves and initial bathymetry that lead to the emergence of different patterns.