A Mechanism-Data Fusion Model for Algal Growth in Hydrologically Constrained River Segments

The control of algal blooms remains a significant challenge in global water environment management. Addressing the lack of hydrodynamic data and small-scale constrained river segments, this study proposes a novel framework for an algal growth dynamics model that incorporates meteorological, water quality, and hydrodynamic factors. The model parameters were optimized using a genetic algorithm for construction. Water quality in the Jiuqu River, located in the upper reaches of the Yangtze River, was monitored through field investigations. By combining Gradient Boosting Decision Tree analysis with SHapley Additive exPlanations, the key environmental factors affecting algal growth in the watershed were identified, and the model’s effectiveness was validated. This method combines mechanistic approaches with data-driven optimization, reliance on large data sets is reduced while model generalization is enhanced. It provides an early warning system for algal bloom outbreaks in the Jiuqu River Basin. The model, optimized through time-lag correction, demonstrated a 7.2% improvement in predictive performance for algal biomass compared to the non-optimized model. Furthermore, its performance significantly surpassed that of traditional data-driven models, showcasing robust adaptability and scalability. Finally, using the Copula function, a joint distribution model of key environmental factors and algal biomass was constructed. The relationships between environmental factors and algal bloom outbreaks were explored, and the environmental thresholds for algal bloom outbreaks in the Jiuqu River Basin were determined (COD _Cr : 18.78 mg/L, COD _Mn : 5.2 mg/L, TP: 0.09 mg/L, T: 11.48°C, Runoff: 0.08 mm). These findings provide guidance for the timely and effective prevention and control of algal bloom pollution.