Integrated Multi-Factor Risk Transmission of Water Transfer During Flood Season in the Nansi Lake of SNWD: A Hybrid Model Approach Driven by Process-Based and Machine Learning Model
摘要
As water scarcity intensifies, the demand for extending the operational duration of water diversion projects has been continuously increasing. However, the environmental risks associated with water transfer during the flood season remain insufficiently understood. To address this issue, this study constructs an integrated multi-factor risk transmission analysis model (MRTM), taking the Nansi Lake of the Eastern Route of the South–to–North Water Diversion (SNWD) Project as a case study. The hybrid model is calibrated and validated using flood runoff and water quality data from the Nansi Lake between 2007 and 2018, demonstrating its capability to accurately quantify the probability of flood-induced environmental risk propagation and to characterize the transmission process. The results indicate that the environmental risk propagation pattern in the Nansi Lake evolves in a phased manner, from stable transmission in the southern lakes, to enhanced transmission in the central lakes, and finally to stable transmission in the northern lakes. With an increase in water diversion flow, the risk propagation pathway shifts from the southern to the northern regions, and the overall environmental risk is mitigated. This study provides a methodological framework for understanding risk propagation in hydrologically complex water diversion systems, particularly for those where the mechanisms and sources of risk transmission are not well defined. The proposed approach offers a scientific basis for water environmental risk management and policy formulation in water transfer projects.