What will the water quality of the Yangtze River be in the future?

Rapid COD Sensing in Complex Surface Water Using Physicochemical-Informed Spectral Transformer with UV-Vis-SWNIR Spectroscopy

Water, as a finite and vital resource, necessitates water quality monitoring to ensure its sustainable use. A key aspect of this process is the accurate measurement of critical parameters such as chemical oxygen demand (COD). However, current spectroscopic methods struggle with accurately and consistently measuring COD in large-scale, complex water environments due to an insufficient understanding of water spectra and limited generalizability. To address these limitations, we introduce the physicochemical-informed spectral Transformer (PIST) model, combined with ultraviolet-visible-shortwave-near-infrared (UV-vis-SWNIR) spectroscopy for water quality sensing. To the best of our knowledge, this is the first approach to combine Transformer with spectroscopy for water quality sensing. PIST integrates a physicochemical-informed block to incorporate existing physical and chemical information into the spectral encoding for domain adaptation, along with a feature embedding block for comprehensive spectral features extraction. We validated PIST using an actual surface water spectral data set with extensive geographic coverage including the Yangtze River and Poyang Lake. PIST demonstrated notable performance in COD sensing within complex water environments, achieving an impressive R2 value of 0.9008 and reducing root mean squared error (RMSE) by 45.20% and 29.38% compared to benchmark models such as support vector regression (SVR) and convolutional neural network (CNN). These results emphasize PIST's accuracy and generalizability, marking a significant advancement in multidisciplinary approaches that combine spectroscopy with deep learning for rapid water quality sensing.

Effect of riverfront utilization transitions on riparian water quality in the middle-lower Yangtze River

The Yangtze River Protection Strategy aims to enhance water quality and mitigate environmental risks associated with riverfront utilization. This study targets a 2000-km stretch of the middle and lower Yangtze River, where we conducted field investigations and sampled at 76 locations. We investigated the relationship between riverfront utilization and riparian water quality, and proposed zoned management strategies. Our findings reveal that the riverfront development utilization rate decreased by 0.5 % from 2017 to 2022, with the transitions towards living-oriented utilization from production-oriented uses. The impact of riverfront utilization on water quality differs by type: living-oriented riverfronts significantly elevate NH3-N levels, while production-oriented riverfronts notably increase TP levels. Regarding heavy metal pollutants, life-oriented riverfronts have a minimal impact, whereas production-oriented riverfronts significantly affect Zn concentrations. A strong positive correlation between changes in riverfront utilization rate and Zn levels suggests a primary influence from industrial sources. Based on these results, we have delineated four types of environmental management zoning for riverfront, pinpointed 29.0 % of critical riverfront segments, and outlined corresponding management measures. This research provides practical insights for advancing the scientific management of riverfront utilization along the Yangtze River and enhancing riparian water quality.

Decoupling the effects of climate, topography, land use, revegetation, and dam construction on streamflow, sediment, total nitrogen and phosphorus in the Yangtze River Basin

Evaluating changes in streamflow, sediment, and nutrient fluxes, as well as quantifying their influencing factors, is crucial for regional water resource protection. While the relationships between major influencing factors and these indicators have been widely studied, the quantitative contributions of the separate and interactive effects of these influencing factors have not been fully explored. This study quantitatively evaluated the changing characteristics of streamflow, sediment discharge, total nitrogen (TN) and total phosphorus (TP), as well as the separate and interactive effects of various major influencing factors such as-rainfall, temperature, evapotranspiration (ET), revegetation, dam construction, and land use change-by applying the GeoDetector method to account for their spatial heterogeneity and contributions. Our findings reveal that the influence of these factors has gradually intensified over time, with dam construction and land use change emerging as the most significant contributors to changes in sediment discharge and TN, respectively. Notably, the interactive effects between dam capacity and vegetation cover on streamflow and sediment discharge was twice as strong as their separate impacts, highlighting the effectiveness of integrating dam construction with reforestation to control erosion and sediment transport. Similarly, the interaction of dam capacity and land use change had a 1.5 times greater impact on TN and TP than their separate effects, indicating that reducing fertilizer application at the source and in the meantime implementing direct interception measures are more effective ways to control water pollution. These findings provide a solid foundation for policymakers to develop integrated water management strategies targeting multiple factors simultaneously, that address both water quantity and quality concerns in the Yangtze River Basin and similar regions.

Multi-machine learning methods to predict spatial variation characteristics of total nitrogen at watershed scale: Evidences from the largest watershed (Yangtze River Watershed), Asian

Nitrogen pollution has emerged as a significant threat to the health of global river systems, garnering considerable attention. However, numerous challenges persist in understanding the characteristics and predicting the spatial changes of total nitrogen (TN) at the catchment scale. We leveraged data from 530 monitoring sections to calculate a land-use composite index and perform statistical analyses to explore the primary factors influencing nitrogen enrichment in the Yangtze River Watershed. We developed three machine learning models to forecast future TN concentrations at monitoring points. Our results showed that agricultural activities and rainfall were the primary drivers of monthly variations in TN concentrations. The upstream region of the watershed exhibited larger variations in TN concentrations (0.097 to 11.099 mg/L), significantly higher than the middle and downstream areas (0.348 to 6.844 mg/L). Microbial-mediated organic matter decomposition in sediment and changes in land-use were identified as key contributors to regional differences in nitrogen enrichment. Potential nitrogen sources include sediment release, urban sewage, and agricultural fertilization. Random Forest model achieved a prediction accuracy of 77.6 %, surpassing the BP and LSTM models. We identified 37 high-risk areas of nitrogen enrichment, concentrated in the Chengdu-Chongqing, Yunnan-Central urban cluster, and the Chaohu Lake sub-watershed. Increased urban land-use and industrial inputs primarily influenced nitrogen enrichment in the upstream area, while agricultural inputs were the main drivers in the middle and downstream regions. Our multi-machine learning models identified the relationship between TN and influencing factors, providing a reliable method for assessing nitrogen enrichment risk in the watershed.

A Copula-based spatiotemporal probabilistic model for heavy metal pollution incidents in drinking water sources

Water pollution incidents pose a significant threat to the safety of drinking water supplies and directly impact the quality of life of the residents when multiple pollutants contaminate drinking water sources. The majority of drinking water sources in China are derived from rivers and lakes that are often significantly impacted by water pollution incidents. To tackle the internal mechanisms between water quality and quantity, in this study, a Copula-based spatiotemporal probabilistic model for drinking water sources at the watershed scale is proposed. A spatiotemporal distribution simulation model was constructed to predict the spatiotemporal variations for water discharge and pollution to each drinking water source. This method was then applied to the joint probabilistic assessment for the entire Yangtze River downstream watershed in Nanjing City. The results demonstrated a significant negative correlation between water discharge and pollutant concentration following a water emergency. The water quantity-quality joint probability distribution reached the highest value (0.8523) after 14 hours of exposure during the flood season, much higher than it was (0.4460) during the dry season. As for the Yangtze River downstream watershed, five key risk sources (N1-N5) and two high-exposure drinking water sources (W3-W4; AEI=1) should be paid more attention. Overall, this research highlights a comprehensive mode between water quantity and quality for drinking water sources to cope with accidental water pollution.

Which riverine water quality parameters can be predicted by meteorologically-driven deep learning?

River water quality has been significantly impacted by climate change and extreme weather events worldwide. Despite increasing studies on deep learning techniques for river water quality management, understanding which riverine water quality parameters can be well predicted by meteorologically-driven deep learning still requires further investigation. Here we explored the prediction performance of a traditional Recurrent Neural Network, a Long Short-Term Memory network (LSTM), and a Gated Recurrent Unit (GRU) using meteorological conditions as inputs in the Dahei River basin. We found that deep learning models (i.e., LSTM and GRU) demonstrated remarkable effectiveness in predicting multiple water quality parameters at daily scale, including water temperature, dissolved oxygen, electrical conductivity, chemical oxygen demand, ammonia nitrogen, total phosphorous, and total nitrogen, but not turbidity. The GRU model performed best with an average determination coefficient of 0.94. Compared to the daily-average prediction, the GRU model exhibited limited error increment of 10-40 % for most water quality parameters when predicting daily extreme values (i.e., the maximum and minimum). Moreover, deep learning showed superior performance in collective prediction for multiple water quality parameters than individual ones, enabling a more comprehensive understanding of the river water quality dynamics from meteorological data. This study holds the promise of applying meteorologically-driven deep learning techniques for water quality prediction to a broader range of watersheds, particularly in chemically ungauged areas.

Pollution loads in the middle-lower Yangtze river by coupling water quality models with machine learning

Pollution control and environmental protection of the Yangtze River have received major attention in China. However, modeling the river's pollution load remains challenging due to limited monitoring and unclear spatiotemporal distribution of pollution sources. Specifically, anthropogenic activities' contribution to the pollution have been underestimated in previous research. Here, we coupled a hydrodynamic-based water quality (HWQ) model with a machine learning (ML) model, namely attention-based Gated Recurrent Unit, to decipher the daily pollution loads (i.e., chemical oxygen demand, COD; total phosphorus, TP) and their sources in the Middle-Lower Yangtze River from 2014 to 2018. The coupled HWQ-ML model outperformed the standalone ML model with KGE values ranging 0.77-0.91 for COD and 0.47-0.64 for TP, while also reducing parameter uncertainty. When examining the relative contributions at the Middle Yangtze River Hankou cross-section, we observed that the main stream and tributaries, lateral anthropogenic discharges, and parameter uncertainty contributed 15, 66, and 19% to COD, and 58, 35, and 7% to TP, respectively. For the Lower Yangtze River Datong cross-section, the contributions were 6, 69, and 25% for COD and 41, 42, and 17% for TP. According to the attention weights of the coupled model, the primary drivers of lateral anthropogenic pollution sources, in descending order of importance, were temperature, date, and precipitation, reflecting seasonal pollution discharge, industrial effluent, and first flush effect and combined sewer overflows, respectively. This study emphasizes the synergy between physical modeling and machine learning, offering new insights into pollution load dynamics in the Yangtze River.

Application of multi-criteria group decision-making for water quality management

As waste discharge into numerous river systems escalates, the pollution of water bodies typically rises. Given the limited capacity of rivers to withstand pollution and their constrained self-cleaning capabilities, treated pollutants from waste discharge must be released into the river. Despite numerous models and algorithms proposed for managing river water quality to meet standards, literature, to our awareness, lacks the utilization of a comprehensive multi-criteria group decision-making approach for water quality management, particularly in river systems. Therefore, this research introduces a new, comprehensive multi-criteria group decision-making for the management of water quality in the Haraz River basin, located in Iran. To do so, the water quality of the basin, a one-dimensional water quality model, QUAL2Kw, was employed to simulate and calibrate the water quality along the river. The simulation results revealed that the downstream water quality violates the water quality standards. To mitigate this issue, various scenarios for waste load allocation (WLA) were evaluated, including no wastewater treatment, primary wastewater treatment, advanced secondary wastewater treatment utilizing the activated sludge (AS) method, and advanced wastewater treatment via the membrane bioreactor (MBR) method. Utilizing the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) and Fuzzy TOPSIS group decision-making model, it was determined that the optimal solution was the implementation of secondary wastewater treatment utilizing the activated sludge method for the 11 PS of pollution, while still adhering to Iranian water quality standard. In addition, the findings of the present study indicate that the implementation of primary wastewater treatment, advanced secondary wastewater treatment utilizing AS, and advanced wastewater treatment through MBR within the study area led to a significant enhancement in water quality. This enhancement ranged from 35 to 105% across various scenarios when compared to conditions where no actions were taken to the treatment of water.

Water quality prediction based on sparse dataset using enhanced machine learning

Water quality in surface bodies remains a pressing issue worldwide. While some regions have rich water quality data, less attention is given to areas that lack sufficient data. Therefore, it is crucial to explore novel ways of managing source-oriented surface water pollution in scenarios with infrequent data collection such as weekly or monthly. Here we showed sparse-dataset-based prediction of water pollution using machine learning. We investigated the efficacy of a traditional Recurrent Neural Network alongside three Long Short-Term Memory (LSTM) models, integrated with the Load Estimator (LOADEST). The research was conducted at a river-lake confluence, an area with intricate hydrological patterns. We found that the Self-Attentive LSTM (SA-LSTM) model outperformed the other three machine learning models in predicting water quality, achieving Nash-Sutcliffe Efficiency (NSE) scores of 0.71 for CODMn and 0.57 for NH3N when utilizing LOADEST-augmented water quality data (referred to as the SA-LSTM-LOADEST model). The SA-LSTM-LOADEST model improved upon the standalone SA-LSTM model by reducing the Root Mean Square Error (RMSE) by 24.6% for CODMn and 21.3% for NH3N. Furthermore, the model maintained its predictive accuracy when data collection intervals were extended from weekly to monthly. Additionally, the SA-LSTM-LOADEST model demonstrated the capability to forecast pollution loads up to ten days in advance. This study shows promise for improving water quality modeling in regions with limited monitoring capabilities.

Research progress in water quality prediction based on deep learning technology: a review

Water, an invaluable and non-renewable resource, plays an indispensable role in human survival and societal development. Accurate forecasting of water quality involves early identification of future pollutant concentrations and water quality indices, enabling evidence-based decision-making and targeted environmental interventions. The emergence of advanced computational technologies, particularly deep learning, has garnered considerable interest among researchers for applications in water quality prediction because of its robust data analytics capabilities. This article comprehensively reviews the deployment of deep learning methodologies in water quality forecasting, encompassing single-model and mixed-model approaches. Additionally, we delineate optimization strategies, data fusion techniques, and other factors influencing the efficacy of deep learning-based water quality prediction models, because understanding and mastering these factors are crucial for accurate water quality prediction. Although challenges such as data scarcity, long-term prediction accuracy, and limited deployments of large-scale models persist, future research aims to address these limitations by refining prediction algorithms, leveraging high-dimensional datasets, evaluating model performance, and broadening large-scale model application. These efforts contribute to precise water resource management and environmental conservation.