Evaluation and prediction of water quality in the dammed estuaries and rivers of Taihu Lake.

Proper evaluation of water quality is pertinent to estuarine habitat restoration. Identifying the degrading factors of the water environment and predicting the trend of eutrophication are key to restore the habitat. Through trophic level index (TLI), water quality index (WQI), modified Nemerow pollution index (NPI), and the Random Forest (RF) model, water samples collected from various estuaries of Taihu Lake from 2017 to 2019 were evaluated. To predict the water quality development, four scenarios were set viz. S1: add or remove an ecological buffer, S2: increase or reduce the external nutrients, S3: open or close the dam/gate, and S4: increase or decrease the internal release. In Wuli Lake, the nutrient concentrations in the river regions were higher than in the lake regions, while a contrary trend was observed in Gonghu Bay. The estuarine water quality in the dry season (WQI = 40.91, NPI = 1.73) was merely worse than that in the wet season (WQI = 47.27, NPI = 1.67). On the other hand, the eutrophic status in the wet season (TLIWet = 57.93) was worse than that in the dry season (TLIDry = 57.23). The estuarine water quality of Taihu Lake has improved from 2017 to 2019 but still belongs to medium level. The principal component analysis (PCA) revealed that dam construction, land use types, unstable hydrodynamic conditions, and trumpet-shaped estuary were the main factors that aggravated the water quality degradation. The RF model has strong forecasting capabilities for estuarine water quality. When the estuaries are close to residential and industrial districts, controlling the surface runoff and improving sewage treatment efficiency are the most effective measures to improve the water quality. In the estuaries, the sediments are usually disturbed by the wind-waves. Conclusively, reducing sediment disturbance and internal contamination accumulation via biological and engineering measures is the key to estuarine restoration.