Summary
Machine learning applications in coastal morphodynamics have emerged as powerful tools for predicting a variety of coastal processes ranging from shoreline change to wave setup, from ripple spacing to sediment resuspension. Machine Learning algorithms, including convolutional neural networks, genetic programming, and artificial neural networks, are increasingly being employed to analyze vast datasets of coastal morphodynamic variables, offering insights into the underlying patterns and drivers of coastal evolution. By integrating data from remote sensing, numerical models, and field observations, Machine Learning models can capture nonlinear relationships and spatio-temporal dynamics that traditional approaches may overlook or struggle with. Despite their computational complexity and data requirements, Machine Learning-based approaches hold great promise for advancing our understanding of coastal morphodynamics and informing coastal management strategies in an era of rapid environmental change. The presentation will describe some recent advances, address some typical criticism to Machine Learning, and even show applications from (un)related fields of science.
Brief biography
Giovanni Coco obtained a PhD in nearshore oceanography at the University of Plymouth (UK). After 3 years at the Scripps Institute of Oceanography (USA) and 8 years at the National Institute of Water and Atmospheric Research (NZ), he joined the University of Cantabria (Spain) with an Excellence fellowship. In 2015 he returned to New Zealand at the University of Auckland where he is currently Professor in the Faculty of Science. His research focuses on coastal processes using a variety of approaches that include numerical and data-driven modeling informed by field and laboratory observations. He currently works on projects dealing with the hydro- and morphodynamics of the nearshore under climatic changes.