Deep Neural Networks for Coastal Climate Risk Prediction Using Flood and Climate Variables

Coastal areas are more vulnerable to climaterelated threats, such as floods, storm surges, and rising sea levels, requiring precise risk assessment for effective mitigation and planning. This paper introduces a deep neural network (DNN) methodology for predicting coastal flood risk, using the Coastal Climate Risk Dataset, which consists of 1,653 geospatial grid records including environmental, topographical, and climatic characteristics. Essential characteristics include elevation, land use, soil composition, drainage capability, proximity to the sea, mean precipitation, and historical flood occurrence. The DNN model was developed to categorize each grid unit as Low, Moderate, or High flood risk. The experimental findings indicate that the proposed model attains an overall accuracy of 98.55 %. Comparisons with baseline models, such as Random Forest and Support Vector Machines, demonstrate that the DNN outperforms conventional machine learning methods by4-7 \%in accuracy, underscoring its enhanced capacity to grasp nonlinear interactions among factors impacting floods. The findings indicate that deep learning methodologies may provide dependable, detailed forecasts for coastal flood risk, facilitating data-informed decision-making and proactive climate resilience initiatives. This approach may be adapted to other climate-sensitive areas to improve environmental planning and catastrophe readiness.