As coastal and urban communities face increasingly complex threats from extreme weather, traditional static flood modeling is no longer sufficient for dynamic decision-making. This talk presents FlowsDT, a hyper-local geospatial digital twin framework designed to simulate and navigate the cascading effects of fluvial and pluvial flooding. Using Galveston City, Texas, as a case study, this methodology integrates geospatial artificial intelligence (GeoAI), CyberGIS, and 3D urban modeling to evaluate varied rainfall scenarios. By providing a 4D environment for real-time flood forecasting and testing infrastructural interventions (such as "Sponge City" concepts), this approach shifts disaster management from reactive recovery to proactive, multiscale community resilience.