Hydrodynamic Hybrid Deep-Learning for Flood Modeling HDL-FM

This dataset supports the research presented in the paper "Spatiotemporal flood depth and velocity dynamics using a convolutional neural network within a sequential Deep-Learning framework" (Environmental Modelling & Software, https://doi.org/10.1016/j.envsoft.2024.106307). 

It contains preprocessed inputs and outputs for training and testing a hybrid deep learning framework that integrates Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) models to simulate hydrodynamic flood dynamics. The complete source code for the hydrodynamic deep learning framework is available on GitHub: https://github.com/m-fathi-said/Flood-Modeling-HDL-FM

Key Features

• Format: Four .pt files (PyTorch-compatible tensors)

  • train_x.pt: Model inputs (e.g., topography, discharge)

  • train_y.pt: Model targets (water depth, velocity magnitude, and flow direction)

  • test_x.pt: Same inputs for testing part

  • test_y.pt: Same targets for testing part

• Scope: Captures river-floodplain interactions and flow properties

Recommended Citation

Fathi, M.M., Liu, Z., Fernandes, A.M., Hren, M.T., Terry, D.O., Nataraj, C. and Smith, V., 2025. Spatiotemporal flood depth and velocity dynamics using a convolutional neural network within a sequential Deep-Learning framework. Environmental Modelling & Software, 185, p.106307.