Urban Inundation Forecasting from Sparse Drainage-Network Hydrodynamic States: A Physics-Constrained Graph–LSTM Surrogate Model

The data uploaded this time are the foundational data used in the process of building the physics-constrained Graph-LSTM surrogate model for urban inundation forecasting from sparse drainage-network hydrodynamic states. These data provide the necessary inputs for model construction and validation, mainly including the structural attributes of the drainage network, time-series observations of hydrodynamic states, topographic information, rainfall time-series inputs, and corresponding inundation monitoring or simulation results. They support the research on achieving efficient and accurate urban inundation simulation and forecasting from a perspective that integrates data-driven methods with physical constraints.