MOIRAI’s AI-based system will significantly enhance coastal early warning capabilities by analysing extensive data from diverse sources, and uncovering complex patterns. These systems, fed by high-fidelity coastal models, will dynamically adapt, continuously improving predictions by retraining on emerging patterns. AI’s capacity to process unstructured data and customize warnings for specific communities will improve their preparedness, aligning with broader goals like a carbon neutral blue economy or resilient coastal communities in the context of climate change.
This early warning system will be integrated into REASSHORE, allowing for its interconnection with digital frameworks such as Digital Earth and the Digital Twin of the Ocean. The proposed bidirectional-LSTM (bi-LSTM) model will predict storm surges and coastal inundation by modelling complex time series with input variables like wind speed, angle of approach, landfall locations, and translation speed. This system will be specifically tested in each of the coastal risk services use cases, ensuring end-user feedback.
Graph Neural Networks (GNNs) will forecast marine heat waves. This will be demonstrated in the use cases on the services for the blue economy, in link with fisheries and aquaculture in each of the three selected seas. These real-time predictions, efficiently stored in a database, will be visually presented in maps for better explainability and accessibility to decision-makers. This AI-based early warning offers faster responses and serves as a simplified version of high-fidelity models. Integrated into the REASSHORE hub, they play a crucial role in supporting a carbon-neutral blue economy and resilient coastal ecosystems. They empower end-users to develop the best adaptation and mitigation measures, contributing to sustainable coastal management strategies.