Incorporating hydrological constraints with deep learning for streamflow prediction

Yi Zhou, Yilin Duan, Hong Yao, Xinchuan Li, Shengwen Li

Abstract: Highlights•Proposed to integrates hydrological constraints for streamflow prediction.•A graph that present the spatiotemporal relationships and hydrological constraints.•A module that can adaptively learn the complex spatiotemporal dependencies.•A streamflow prediction neural network stacking the feature aggregation module.•The proposed model outperformed baselines and exhibits robust performance.