Stfs-Urban : Spatio-Temporal Flood Simulation Model for Urban Areas

Urbanization and the increasing frequency of extreme climate events have required cities around the world to improve their flood resilience. In this study, we present a spatiotemporal flood simulation–urban model that integrates refined and multi-temporal scales and includes multi-scale radar rainfall inversion (MRI), refined coupled flood simulation (FGCFS), and transformer–convolutional-neural-network flood prediction (TCFP) modules. This is the first urban flood model to implement radar rainfall data inversion. The model takes the Nanjing urban area as the test area, and the advantages of refinement coupling and boundary conditions using both FGCFS and TCFP are discussed. This study provides a mature framework for integrating multi-scale data, refinement coupling, and dynamic allocation; thus, it provides new ideas for future flood simulation models