Due to the influence of complex regional climate, water quality perturbation factors of lakes in cold regions are complicated, and the uncertainty of each factor needs further study. This study coupled two algorithms (clustering and EM) to establish a water quality uncertainty model of Chagan Lake, a typical cold region lake in China. A BN model containing nine influencing factors (including water temperature (WT), total phosphorus (TP), total nitrogen (TN), etc.) was established and optimized, and sensitivity analysis was also performed. The results indicate that the water quality status of the lake is class III and 27.47% risk of exceeding the standard. The water quality of the lake is more susceptible to disturbance during the freezing period (WT < 1 °C). TP is the most sensitive factor for water quality disturbance in the lake followed by chemical oxygen demand (COD), TN, and fluoride (F). Parameter control result displays, and the multifactor synergistic control scheme could reduce the water quality risk of the lake by 36.47%. This study demonstrates that our proposed method can be used to predict both sudden water quality events and the overall trend of water quality fluctuation, which is important for rapid water quality evaluation and management decisions.
Keywords: Bayesian network; Cold region; Risk control; Surface water; Water management; Water quality.
© 2022. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.