Physics-guided deep learning improves canal water forecasting, reducing uncertainty in large-scale water diversion systems

A multi-institutional research team has developed a physics-guided mixture density network (PgMDN) that integrates physical hydraulic laws into a probabilistic deep-learning framework, significantly improving the prediction of lateral offtake discharges in large canal systems. The study, published in Environmental Science and Ecotechnology on May 7, 2026, addresses the challenge of unpredictable water flows that compromise water-level forecasts and operational decisions in inter-basin water transfer projects.

Lateral offtake discharges—flows diverted from main canals through side structures—often deviate from planned targets due to real-time hydraulic states and unplanned gate operations, producing multi-peaked, highly uncertain flow distributions. Traditional physics-based methods for quantifying this uncertainty are computationally expensive, while purely data-driven models struggle with complex multimodal patterns, especially when training data are scarce.

The PgMDN incorporates two physical constraints directly into its loss function. First, it promotes local mass-balance consistency by aligning predicted mean discharges with inflow-minus-outflow values from a simplified hydraulic model. Second, it imposes a consistency rule: when predicted mean flows change rapidly—indicating operational shifts or abrupt gate movements—the model's uncertainty is expected to increase accordingly, preventing overconfident predictions during unstable conditions.

Tested on real-world data from two reaches of China's South-to-North Water Diversion Project, the PgMDN reduced mean absolute error (MAE) by more than 25% and root mean square error (RMSE) by over 25% compared to standard mixture density networks (MDNs). Reliability improved from 0.45 to 0.82 at the 90% confidence level. The model maintained stable performance when training data were intentionally reduced, demonstrating strong generalization under data-scarce conditions. Using SHapley Additive exPlanations (SHAP) analysis, the team identified water level fluctuations and boundary inflows as the dominant drivers of predictive uncertainty.

"We wanted a model that doesn't just give a single number but actually tells operators how much to trust that number," the authors said. "By embedding two simple physical rules into the learning process—promoting local mass-balance consistency and linking sudden flow changes to wider uncertainty—we got much more reliable forecasts, even when data were limited."

This approach enables more adaptive water allocation in real time, allowing operators to adjust safety margins, optimize gate operations, and respond effectively to unexpected events such as unplanned withdrawals. The framework is scalable and can be integrated into existing hydrodynamic models to estimate plausible water-level ranges under different scenarios. By bridging physical understanding with data-driven learning, the PgMDN offers a practical pathway toward resilient management of large-scale water systems, especially in regions facing increasing hydrological variability. It also opens the door for similar hybrid models in other environmental infrastructure applications, from flood control to water distribution networks.

The full study is available online.