Search

• A holistic water quality assessment was conducted in the Erdao Songhua River Basin. • The IWPI showed "Good" water quality with spatiotemporal variations in studied basin. • PI, DO, COD cr , and BOD 5 are the most important parameters in this basin. • The potential risks of point source and non-point source pollution in middle-downstream should be of concern. The Erdao Songhua River provides drinking water for 16.82 million people; therefore, ensuring that the water quality adequate is of particular concern. In this study, an improved water pollution index (IWPI) and multiple statistical methods were employed to assess the overall water pollution situation and investigate spatiotemporal variations of seven physicochemical parameters such as the permanganate index (PI), dissolved oxygen (DO), chemical oxygen demand (COD cr), five-day biochemical oxygen demand (BOD 5), total nitrogen (TN), total phosphorus (TP) and ammonia nitrogen (NH 3 -N), which were collected monthly at 20 sites within the mainstream and major tributaries of the Erdao Songhua River Basin (ESRB) from 2015 to 2020. Stepwise regression analysis was conducted to build a minimum improved water pollution index (IWPImin) model consisted of four key elements (PI, DO, COD cr , and BOD 5) proposed from seven parameters. The results demonstrated water quality within the ESRB was considered to be "Good" with the mean IWPI values <40. However, the water quality deteriorated from the upstream to the downstream within the basin, manifested as the mean value of IWPI of the downstream is 1.7 times of upstream. Seasonally, an improving trend of water quality was observed during the monitoring period and the mean IWPI value decreased by 23%. Furthermore, Seasonal variation in the IWPI value was evident, and the best water quality was found in winter (lowest IWPI value of 14.1) and the worst in summer (highest IWPI value of 21.7). The proposed IWPImin model uses the selected four crucial parameters and the weights of those parameters has exhibited excellent performance in the water quality assessment, with the highest coefficient of determination (R2) and lowest Root Mean Square Error (RMSE) values of 0.996 and 0.51, respectively, which can be used to optimize water quality assessment strategies at a lower cost. For future management, the water quality of middle and downstream should be carefully inspected, and strictly control the effects of point source and non-point source pollution in the ESRB. [ABSTRACT FROM AUTHOR]

• Provides a protocol for transferring hydrological knowledge between basins. • Proposes a TL-Transformer model for flood simulation and forecasting in data-sparse basins. • Explores the impact of data quality on transfer learning effectiveness. • Explains the manifestation of hydrological similarity in transfer learning through topographic index. • Compares the performance of models including physical and neural network models. There exists a substantial disparity in the distribution of streamflow gauge and basin characteristic information, with a majority of flood observations being recorded from a limited number of well-monitored locations. Transferring hydrological knowledge from data-rich to data-sparse basins has been a persistent issue. While artificial intelligence methods, such as Long-Short-Term Memory Neural Networks (LSTMs) have been attempted for simulating across different basins, the problem of gradient vanishing still inevitably lead to difficulties in responding to new basin datasets. A novel solution is presented in the form of the Transfer Learning Framework based on Transformer (TL-Transformer), which can accurately predict flooding in data-sparse basins (targets) by using models from data-rich basins (sources) without requiring extensive basin attributes at the target location. The framework was demonstrated in the middle reaches of the Yellow River, and performance was evaluated using the Nash Sutcliffe efficiency coefficient, root mean square error, and bias. The results show that TL-Transformer outperforms other models, including TOPMODEL, MLP, TL-MLP, LSTM, TL-LSTM and Transformer in all target basin stations. Compared with those models that without transfer learning, TL-Transformer showed significantly improved performance, where NSE increased by 0.2 and reached above 0.75, RMSE and bias were also controlled within 60 and 20, respectively. Furthermore, pre-training on basins with hydrological similarities increases the benefits of Transfer Learning, and we explained this response phenomenon through the topographic index. These results suggest that deep learning can tap into the commonalities in hydrological data across basins to improve the accuracy of flood forecast applications in areas with limited observations. [ABSTRACT FROM AUTHOR]