A Multi-Scenario Synthetic River Network Dataset Based on Hydrological Characteristics of Pingwu, Sichuan, China
收藏资源简介：
This dataset provides a multi-scenario synthetic river-network benchmark for spatio-temporal flood forecasting, generated based on representative hydrological and hydraulic characteristics of Pingwu, Sichuan, China. It contains three scenarios representing plain (SF1), hilly (SF2), and mountainous (SF3) environments with progressively increasing hydrological and topological complexity. The dataset includes hourly rainfall and discharge time series, directed river-network topology, inter-node distance matrices, and hydraulic attributes, including channel width, bed slope, and Manning's roughness coefficient. It is designed for studies on multi-step flood forecasting, directed spatio-temporal graph learning, and physics-informed hydrological modeling. Data are provided in both CSV and NumPy formats to support reproducible machine-learning experiments.
