From Flood Extent Mapping to Mechanism-Aware Flood Products: Integrating Flood Type Classification into Satellite-Based Flood Monitoring​

Flood type information is critical for effective flood risk management, as different flood-generating mechanisms are associated with distinct hydrodynamic behaviour, contamination pathways, and recovery trajectories. However, most operational flood mapping products provide only binary inundation extent, offering limited insight into underlying flood processes. Existing flood type classification approaches rely predominantly on hydrometeorological observations and modelling, which are often unavailable in data-scarce regions and can be unstable in mechanism-complex environments such as estuarine deltas, river-urban corridors, and coastal cities. To address these limitations, this research proposes a multi-convolutional neural network (CNN) framework that integrates flood type classification directly into satellite-based flood mapping. A U-Net model is first trained for flood extent segmentation, followed by a CNN-based architecture for event-level flood type identification. Several CNN architectures were compared for flood type classification, with Inception-ResNet selected from these based on performance–complexity trade-offs. The framework is evaluated using a multi-event dataset spanning six continents and 15 countries. Results indicate strong performance in both inundation segmentation (96.5% overall accuracy) and flood type classification (98.9% overall accuracy under dominant-mechanism conditions). These findings demonstrate the feasibility of deriving mechanism-aware flood products directly from remote sensing imagery, advancing conventional extent-based flood mapping toward more decision-relevant flood intelligence for emergency response and post-disaster assessment.