Deep Vision-Based Framework for Coastal Flood Prediction Under Climate Change Impacts and Shoreline Adaptations

Climate change and its consequences, such as sea-level rise (SLR), pose escalating threats to coastal regions, intensifying the need for efficient and accurate flood prediction models. While traditional physics-based hydrodynamic simulators offer high fidelity, they are computationally prohibitive for integration into planning or optimization workflows. On the other hand, Deep Learning (DL) techniques can produce substantially faster models, yet, their development typically requires a large number of training samples. To remove this barrier, we present a systematic framework for training high-fidelity and high-resolution coastal flood prediction models in low-data settings. We test the proposed framework on multiple DL architectures, including a fully Transformer-based model and a Convolutional Neural Network (CNN) with additive attention gates. Additionally, we introduce a deep CNN architecture tailored specifically to the studied climate adaptation-aware coastal flood prediction problem. The model was designed with a particular focus on its compactness so as to cater to resource-constrained scenarios and accessibility aspects. The performance of the DL models developed through the proposed framework is validated against state-of-the-practice methods as well as traditional Machine Learning approaches. The results demonstrate substantial improvement in prediction quality, ranging from 100% to 400% across key metrics. Lastly, we round up the contributions by providing a meticulously curated dataset of simulated inundation maps for the coast of Abu Dhabi, which can serve as a benchmark for evaluating future coastal flood prediction models. The complete source code of the proposed framework, including the trained models, data and the evaluation scripts, can be accessed at https://anonymous.4open.science/anonymize/paper93-FC15.