Multi-risk assessment in coastal areas through the implementation of Artificial Intelligence methods

Coastal areas are increasingly threatened by sea-level rise and shifting pattern of extreme events. In response to these complex dynamics, it is crucial to adopt integrated approaches that assess risks posed by both natural and anthropogenic stressors. The complexity of interactions between risk factors, combined with the availability of heterogeneous data, makes Artificial Intelligence (AI) and Machine Learning (ML) algorithms ideal tools to tackle multi-hazard and multi-risk assessments. This thesis goes beyond traditional methods by exploring the advantages of AI-based tools, leveraging the potential of big data from open-source databases and local authorities through a series of articles focused on the coastal area of Veneto. In the first study, different ML algorithms were implemented to estimate daily coastal risks and identify the main factors driving impacts in the coastal municipalities of the Veneto region in the 2009-2020 timeframe. The most effective algorithm was enhanced in the second chapter with a two-tier approach to estimate impacts under different climate change scenarios up to 2100. The final paper introduces COAST-AId, a custom language model designed to analyse vast sets of information and facilitate coastal risk management, adapting EUCRA approach to identify intervention priorities for the Veneto littoral. The results provide an assessment of the risks linked to extreme events in the Veneto coastal area, suggesting improvements to adaptation strategies.