Flood Damage Nowcasting

Flood damage nowcasting aims to provide near real-time estimations of flood impact on buildings and infrastructure, enabling faster emergency response and resource allocation. Current research focuses on developing machine learning models, including graph convolutional networks and physics-informed neural networks, that integrate diverse data sources such as satellite imagery, rainfall data, and even social media reports to improve prediction accuracy and spatial resolution. These advancements offer significant potential for enhancing disaster preparedness and response by providing timely and accurate information to emergency managers and decision-makers, ultimately minimizing damage and improving community resilience.