Diffusion‐Based Probabilistic Modeling for Hourly Streamflow Prediction and Assimilation

Abstract Hourly predictions are critical for issuing flood warnings as the hourly flood peaks can be distinctly higher than the corresponding daily ones. Currently a popular hourly data‐driven prediction scheme is multi‐time‐scale long short‐term memory (MTS‐LSTM), yet such models face challenges in probabilistic forecasts or integrating observations when available. Diffusion artificial intelligence (AI) models represent a promising method to predict high‐resolution information, for example, hourly streamflow. Here we develop a denoising diffusion probabilistic model (h‐Diffusion) for hourly streamflow prediction that conditions on either observed or simulated daily discharge from hydrologic models to generate hourly hydrographs. The model is benchmarked on the CAMELS hourly data set against record‐holding MTS‐LSTM and multi‐frequency LSTM (MF‐LSTM) baselines. Results show that h‐Diffusion outperforms baselines in terms of general performance as well as extreme metrics. Furthermore, the h‐Diffusion model can utilize the inpainting technique and recent observations to accomplish data assimilation that largely improves flood forecasting performance, or even inversely estimate rainfall. These advances can greatly reduce flood forecasting uncertainty and provide a unified probabilistic framework for downscaling, prediction, and data assimilation at the hourly scale, revealing risks that daily models cannot express.