Deep learning for streamflow simulation and forecasting
State-space models and diffusion-based generative models for streamflow simulation and probabilistic forecasting.
Accurate streamflow simulation and forecasting underpin flood preparedness and water-resources management. This line of work advances modern deep learning architectures for hydrology, including:
- State-space models for simulating streamflow over long hydrologic records.
- Diffusion-based generative models for probabilistic streamflow forecasting.