In the current era， the rapid development of environmental science and computer technology complement each other. Among them， the application of machine learning in the research field of water quality purification in constructed wetlands has become increasingly prominent， showing vigorous vitality and broad prospects. This paper focuses on this aspect and systematically reviews the main research progress of machine learning in water quality purification in constructed wetlands in recent years， striving to comprehensively and deeply analyze its key threads. On the one hand， an in-depth exploration of the diverse application directions of machine learning algorithms in constructed wetlands is carried out， covering key links in water quality purification such as microbial metabolism and pollutant degradation， plant absorption and transformation processes， substrate adsorption and filtration mechanisms， and hydraulic conditions， accurately grasping the adaptation patterns of algorithms in each link. At the same time， the characteristics of constructed wetland data are deeply excavated to lay a solid foundation for algorithm optimization. Further， it is elaborated from three core dimensions， in which we： ① focus on the key processes of water purification assisted by machine learning and clarify the complex interaction mechanisms among microorganisms， plants， substrates， and hydraulic conditions； ② commit to multi-dimensional data fusion to reshape the architecture of water purification mechanisms and drive in-depth understanding with data； and ③ explore the practical applications of machine learning in the design optimization and operation control of constructed wetlands to enhance wetland efficiency. In addition， the article takes a long-term view and looks ahead to future research directions from five aspects， namely model improvement and innovation， data fusion and sharing， real-time monitoring and intelligent control， integration with other technologies， and environmental impact assessment and ecological restoration. To summarize， machine learning injects new vitality into the research on water quality purification in constructed wetlands， opens up new perspectives， and effectively promotes the efficiency of water quality purification and management levels. However， at present， continuous efforts still need to be made in the in-depth optimization of algorithms and solving practical application problems to fully release its potential.

Wastewater treatment plants (WWTPs) require precise real-time technologies to maintain effluent quality while optimizing the chemical consumption under fluctuating influent conditions. This study proposes a Bayesian optimized ensemble machine learning (BO-EML) framework to predict various contaminants, optimize chemical dosages, and analyze economic costs of chemical dosing. The model employs three ensembles and three baseline models within a structured nested cross-validation (NCV) framework to optimize hyperparameters through Bayesian prediction of chemical oxygen demand (COD), biochemical oxygen demand (BOD), total nitrogen (TN), and total phosphorus (TP). A full-scale WWTP was examined as a case study, resulting in test R2 values of COD (XGBoost R2 = 0.899), BOD (Gradient Boosting R2 = 0.975), TN (XGBoost R2 = 0.837), and TP (Gradient Boosting R2 = 0.737), significantly surpassing the performance of linear and single-tree baseline models, respectively. The SHAP findings indicated that influent indicators and chemicals displayed variable directional effects based on the dosage and connection to additional process parameters. Economic optimization showed a 32.1% cost reduction while maintaining the removal efficiency and regulatory compliance in chemical dosing. Thus, this framework shows the potential for integrating data-driven prediction, interpretation, and cost-efficient optimization, which can serve for stable WWTP operations.

Aerobic granular sludge (AGS) is widely recognized as a highly promising technology for wastewater treatment. However, enhancing the granulation rate and ensuring long-term stability in full-scale applications, particularly in continuous-flow reactors (CFRs), remains a critical challenge. Inoculation with pre-granulated sludge, prepared using natural polyelectrolytes and dewatered sludge, is an effective strategy for enhancing the rapid start-up and stable operation of AGS, while the practical application of this technology in CFRs remains unvalidated at present. In this study, dewatered sludge was conditioned separately using two natural polysaccharide polyelectrolytes (PSEs), namely sodium alginate (SA) and chitosan (CTS), to fabricate PSE-enhanced AGS. These preparations were then introduced into continuous-flow airlift reactors (CFARs), which are R1-ctrl, R2-SA and R3-CTS, to assess system enhancement. The results demonstrated that PSEs distinctly accelerated granulation in CFARs. AGS in R3-CTS exhibited improved settling performance (SVI5: 34.27 ± 5.09 mL/g) and biomass (MLSS: 5.30 ± 4.31 g/L) and maintained a relatively stable particle size growth. The TN removal capacities of PSE-AGS were both distinctly enhanced. TN removal rates for R2-SA (63.88 ± 2.45%) and R3-CTS (63.4 ± 1.42%) were 8.79% and 8.31% higher than R1-ctrl (55.09 ± 2.80%). The microbial community analysis demonstrated that both PSEs altered the microbial community structure and contributed to the selection of dominant bacterial (e.g., Acinetobacter, unclassified_f_Rhodobacteraceae and Halomonas), consequently facilitating granulation and enhancing the TN removal capacity. While SA gel provided more sites for microbial attachment and growth, and CTS neutralized the negative charge of the sludge to adsorb positively charged particles. By inoculating PSE-AGS, this study successfully facilitated the maturation of AGS in CFARs, which demonstrates that inoculating PSE to enhance AGS formation in CFRs is a promising strategy for future upgrading in wastewater treatment plants (WWTPs).

The addition of carriers or assisted granulation is one of the effective techniques to facilitate the start-up of aerobic granular sludge (AGS) technology, whereas this approach requires low-cost raw materials. To address this issue, the dewatered sludge from wastewater treatment plant was utilized as raw materials to produce pregranulated AGS (PG-AGS) and dosed into the sequencing batch reactor (SBR). The results indicate that the granulation step was skipped and the startup period was shortened to 12 days in the PG-AGS dosed reactors. The PG-AGS with the smaller particle size of 700-830 mu m appeared higher stability than larger particle size of 1180-1400 mu m. Meanwhile, the inoculated PG-AGS achieved effective efficiency for chemical oxygen demand (COD) and total nitrogen (TN) removal and exhibited good settlement ability during long-term operation. According to scanning electron microscopy (SEM), the interior of AGS showed a porous structure, and the developed pores inside facilitated the transportation of nutrients and dissolved oxygen. Microbial community analysis showed that proteobacteria (critical for microbial aggregation/granulation) increased from 50.8 % to 72.1 % (NFAGS in reactor 1) and 39.8-86.5 % (PG-AGS with particle size of 700-830 mu m in reactor 3) over 80 days. Finally, a novel structure of PG-AGS, characterized by a more evenly and diversely distributed population of ammoniaoxidizing bacteria (AOB), nitrite-oxidizing bacteria (NOB), and polyphosphate-accumulating organisms (PAO) within the inner layer, was proposed according to fluorescence in situ hybridization (FISH) analysis. This study confirmed the viability of dewatered sludge as a feedstock for granular sludge cultivation, with comprehensive characterization of its intrinsic microbial consortia, functional genes, and spatial architecture. The results revealed the unique structure of PG-AGS and proposed the mechanism of PG-AGS in accelerating the rapid startup of AGS system.

The equal and fair right to use water environmental capacity is the basic development right for regions who are with unadvanced economic or in vulnerable water environment in China. Clearly defining the responsibilities is the prerequisite to make a comprehensive knowledge of this right. This work collected emission data on 3 representative water pollutants (COD, total nitrogen (TN), and total phosphorus (TP)) at provincial level, and utilized a multi-regional input-output model and data envelopment analysis to calculate the specific water pollution emission responsibilities as well as the corresponding environmental fees payable or compensation receivable. Pollutants generated related to products which were locally produced for local consumption represented the primary source of pollution emissions. The embodied transfer of COD due to inter-provincial trade amounted to 7.03 million tons, TN to 0.97 million tons, and TP to 99,300 tons. The transfer of water pollutants between less developed provinces constituted the largest proportion (similar to 40 %) of all pollutant transfers. The main provinces paying water pollution compensation were those economically developed eastern coastal regions, along with some economically underdeveloped provinces, while the receiving provinces were mostly located in central and western regions with a certain industrial base but underdeveloped economies. This study found that the structural transformation has occurred, which was no longer the past traditional pattern of embodied transfer of water pollutants. The pattern of inter-provincial and inter-industrial transfer were intricate, and the responsibility identification should be more prudent with focus on both objectiveness and fairness of development.

To achieve higher denitrification efficiency with reduced energy consumption in aerobic granular sludge (AGS) system, a systematic evaluation of the carbon and nitrogen metabolism process for AGS under different stage is essential. Herein, this study established the prediction models via interpretable machine learning (ML) for simulating the nitrogen metabolism by using 312 sets of data collected from four reactors with different kinds of AGS. The results indicated Gradient Boosting Decision Trees (GBDT) achieved R2 values of 0.729, 0.875, and 0.807, respectively by selecting water temperature, carbon source components and particle size as input factors and NH4+-N, NO2--N, and NO3--N as prediction targets. Furthermore, Shapley Additive Explanations (SHAP) analysis was used to make global and local interpretations of the GBDT models. The global explanation revealed that particle size of AGS and carbon source of component 4 (C4) with the Ex/Em at 225/335 nm significantly influenced denitrification process. And local analysis results proved that enhanced nitrogen removal performance is attainable when the DX50 and DX10 range between 500-600 and 200-500 mu m and the Fmax value of C4 exceeds 0.2 R.U. These findings provide an effective tool for evaluating nitrogen removal performance and identifying the key factors influencing nitrogen metabolism in AGS systems.

The excessive discharge of phosphorus from rural domestic sewage is a problem that worthy of attention. If the phosphorus in the sewage were recovered, addressing this issue could significantly contribute to mitigating the global phosphorus crisis. In this study, corn straw, a common agricultural waste, was co-pyrolytically modified with eggshells, a type of food waste from university cafeterias. The resulting product, referred to as corn straw eggshell biochar (EGBC) was characterized using SEM, XRD, XPS, XRF, and other methods. Batch adsorption experiments were conducted to determine the optimal preparation conditions of EGBC and to explore its adsorption characteristics. EGBC showed strong adsorption effectiveness within a pH range of 5-12. The adsorption isotherm closely followed the Sips model (R-2 > 0.9011), and the adsorption kinetics were more consistent with the pseudo-second-order model (R-2 > 0.9899). The process was found to be both spontaneous and endothermic. Under optimal conditions, the phosphorus adsorption capacity of EGBC was measured to be 288.83 mg/g. This demonstrates the high efficiency of EGBC for phosphorus removal and illustrates an effective method of utilizing food waste for environmental remediation.

Dual-source drinking water distribution systems (DWDS) over single-source water supply systems are becoming more practical in providing water for megacities. However, the more complex water supply problems are also generated, especially at the hydraulic junction. Herein, we have sampled for a one-year and analyzed the water quality at the hydraulic junction of a dual-source DWDS. The results show that visible changes in drinking water quality, including turbidity, pH, UV254, DOC, residual chlorine, and trihalomethanes (TMHs), are observed at the sample point between 10 and 12 km to one drinking water plant. The average concentration of residual chlorine decreases from 0.74 ± 0.05 mg/L to 0.31 ± 0.11 mg/L during the water supplied from 0 to 10 km and then increases to 0.75 ± 0.05 mg/L at the end of 22 km. Whereas the THMs shows an opposite trend, the concentration reaches to a peak level at hydraulic junction area (10-12 km). According to parallel factor (PARAFAC) and high-performance size-exclusion chromatography (HPSEC) analysis, organic matters vary significantly during water distribution, and tryptophan-like substances and amino acids are closely related to the level of THMs. The hydraulic junction area is confirmed to be located at 10-12 km based on the water quality variation. Furthermore, data-driven models are established by machine learning (ML) with test R2 higher than 0.8 for THMs prediction. And the SHAP analysis explains the model results and identifies the positive (water temperature and water supply distance) and negative (residual chlorine and pH) key factors influencing the THMs formation. This study conducts a deep understanding of water quality at the hydraulic junction areas and establishes predictive models for THMs formation in dual-sources DWDS.

Elucidating the dynamics of dissolved organic matter (DOM) transport and transformation under seasonal rainfall events is essential for the conservation of riverine ecosystems, for mitigating the effects of climate change, and for crafting informed water management strategies. Therefore, this study aimed to investigate the evolutionary characteristics of organic pollution sources during consecutive rainfall events in early spring and to quantify their relative contributions to the process of surface water pollution. The results showed seasonal rainfall induces water quality exceedances in rivers due to the combined impacts of terrestrial inputs and endogenous releases. Humic acid (HA) (region V) and fulvic acid (FA) (region III) emerged as the predominant organic matter in the water column, with their fluorescence intensity altering as rainwater flushed the riverbed. Sources of pollution include agricultural and urban domestic sources (AS + DS) (72.29 %), industrial and urban domestic and microbial sources (IS + DS + MS) (37.71 %), and agricultural and industrial sources (AS + IS) (63.32 %), indicating that agricultural surface pollution discharges contribute significantly. The gas-chromatography-mass spectrometry (GC-MS) further confirmed that exogenous inputs were predominantly comprised of particulate pollutants. This study underscores the efficacy of fluorescence difference spectrometry in delineating the migration and transformation of river pollution sources during seasonal rainfall and facilitating the implementation of targeted management strategies for river ecosystems.

Understanding the transfer of embodied emissions along trade chains is crucial for promoting sustainable economic development. Here, we constructed the provincial inventory of emissions of 6 typical air/aquatic/solid pollutants (SO2, NOX, Dust, COD, NH3-N, and Solid wastes) via a multiple regional input-output (MRIO) model. We quantified the emissions responsibilities from both producer and consumer perspectives, adopting the 'Beneficiaries are responsible' principle for allocation. Further, we calculated appropriate payments or subsidies for each province's emissions using the slacks-based Data Envelope Analysis (SBM-DEA), aiming to provide a fair and transparent reference for China's current fiscal transfer payment system. Our findings revealed significant progress in emission reductions between 2012 and 2017 by 36.64%-66.49% excluding Solid wastes. Provinces within the Yangtze River Delta and Guangdong were the primary emitters at the consumption end, while provinces in Central, Western, and Northeastern China bore the brunt of embodied emissions through trade at the production end. In 2012, a dominant emission transfer was from developed to undeveloped provinces. However, by 2017, a new trend emerged, with emissions being transferred between provinces with similar levels of financial development. The secondary industry was the largest adsorber for the emission transfer, and reductions in emissions from this sector benefited the output of provinces who experienced rapid development during 2012-2017. Based on the offsetting calculated by this work, the current transfer payments system in China is not fair for some provinces who paid more or accepted less than they deserved. Some provinces should take a more active gesture in pursing their development instead of relying upon the transfer which they do not deserve. This study provides a comprehensive understanding of emissions from various perspectives, highlighting changes in inter-provincial and inter-industrial transfer pathways, which was valuable for multi-regional management efforts aimed at aligning national economic development with ecological conservation goals.

China's staple crops face heavy metal (HMs) contamination, a widespread issue lacking a national assessment. We used machine learning (ML) to assess risks of 8 HMs in rice, wheat, and maize, and estimated a financing strategy for soil remediation via linear optimization and computable general equilibrium (CGE). The accumulation of HMs in crops depends on Soil-HMs, climate, soil properties, and crop types. Cd and Hg pose major soil pollution risks, while Cr, Pb, and Cd are the most threatening in crops. High-risk zones are located at the warm temperature and subtropical zones, with wheat most vulnerable. Over a quarter (26.77 %) of the nation's croplands are classified as high-risk, with a significant 60.89 % falling into the medium-risk category, leaving merely 12.34 % of the agricultural land in a safe condition. The estimated remediation cost is 58596.73 billion RMB and the crop loss is 808.03 billion RMB in a ten-year remediation period at the context of secure crop supply. The reallocation of social investment rather than raising new taxation for the remediation is beneficial to the GDP increase and social welfare despite some loss in the household income and enterprise income. This study provides a comprehensive evaluation for Crop-HMs risk and remediation policy, crucial for national crop security.

采用亚铁(Fe2+)分别联合次氯酸钠(NaClO)、过硫酸钠(Na₂S₂O8)和过氧化氢(H2O2)3种氧化剂处理某原油浸染土壤的热脱附废水,考察了pH值、Fe2+投加量、氧化剂投加量和反应时间对处理效能的影响.Fe2+/H2O2 和Fe2+/Na₂S₂O8均能有效去除热脱附废水中的典型污染物——石油烃物质,去除率分别达到100.0%和97.8%,但Fe2+/H2O2体系对化学耗氧量(COD)的去除效果明显优于Fe2+/Na₂S₂O8.Fe2+/H2O2 可将有机物彻底氧化分解,而Fe2+/NaClO和Fe2+/Na₂S₂O8仅是将有机物氧化分解为中间物质.多元变量分析结果表明,pH值和氧化剂投加量是影响去除效果的主要因素,Fe2+/H2O2处理热脱附废水的吨水处理成本较高,但是综合效果最好;Fe2+/Na₂S₂O8处理成本次之,但无法彻底降解COD;Fe2+/NaClO成本最低,但对有机物去除效果最差.综合去除效果和经济因素考量,Fe2+/H2O2体系更适用于处理热脱附废水.

Ferrate(VI) is a green and efficient water treatment agent for drinking and wastewater. It is widely used in water treatment because it has multi-functional uses such as oxidation, algae removal, disinfection, and adsorption flocculation. It does not cause secondary pollution to the environment. This paper compares ferrate(VI) with other water treatment agents and discusses three methods of preparing ferrate(VI). The removal, adsorption, and control of organic matter, algae, disinfection by-products, and heavy metal ions in water when ferrate(VI) was used as an oxidant, disinfectant, and coagulant were summarized. Ferrate(VI) has some advantages in removing toxic, harmful, and difficult-to-degrade substances from water. Due to the disadvantages of ferrate(VI) such as oxidation selectivity and instability, it is necessary to develop the hyphenated techniques of ferrate(VI). In this review, three hyphenated techniques of ferrate(VI) are summarized: ferrate(VI)–photocatalytic synergistic coupling, ferrate(VI)–PAA synergistic coupling, and ferrate(VI)–PMS synergistic coupling.

The detection of tetracyclines (TCs) adsorption performance by specific adsorbent is a time-consuming process. Thus, developing prediction framework based on existing data to quickly evaluate adsorption performance is necessary especially in the urgent scenario. Herein, we employ machine learning to deliver the accurate prediction of TCs adsorption performance by biochar via grouping numeric/nonnumeric information over physiochemical properties of biochar and environmental pressures. We find the TCs adsorption by biochar is a purely physical behavior where porous filling is the primary mechanism. Porous structure of biochar and environmental pressures co-determine the adsorption amount, and the synergism of physiochemical properties of biochar and environmental pressures determines adsorption capacity. Adsorption kinetics is subject to physical properties and environmental pressures. Chemical properties of biochar have a limited influence on adsorption performance and its main contribution is delivered via the interactions between C-contained or O-contained functional groups and TCs. Simplifying model only with reduced variables input can also accurately predict TCs adsorption performance by biochar. This strategy enables the portable prediction in the urgent scenario since expensive instruments and complex detections can be avoided. These findings provide a comprehensive understanding of the way of physiochemical characteristics and environmental pressures on adsorption performance, and offer useful tips in designing biochar-based adsorbents for TCs removal.

The complex hydraulic environment in dual-source drinking water distribution system (DWDS) in metropolitan puts water safety in an unknown situation, which requires systematic water quality monitoring and controlling to reduce the water risk during water distribution process. Given that, this study conducted a one-year sampling of trihalomethanes (THMs) and identified the hydraulic junction area in a real dual-source DWDS. For efficient prediction of THMs, several predictive frameworks were established via mechanistic algorithms and machine learning (ML). The results indicated that Gradient Boosting Decision Tree (GBDT) algorithm achieved higher interpretability and generalizability in describing the relationship between various of features with R2 reaching 0.903, 0.908, 0.945 and 0.959 for trichloromethane (TCM), bromodichloromethane (BDCM), dibromomonochloromethane (DBCM) and tribromomethane (TBM), respectively. Shapley additive explanations (SHAP) analysis of GBDT models further identified water temperature, residual chlorine, water supply distance, and pH were key features affecting THMs formation in practical DWDS. Furthermore, the spatio-temporal distribution of THMs was simulated by GBDT models and identified the region might be exposed to high risk of the THMs. The simulation results indicated that the DBCM and TBM were more likely to form in the region nearly to DWTP-X, whereas in the region nearly to DWTP-Y, the TCM and BDCM were the dominant THMs species due to differences in raw water characteristics. Undoubtedly, this study provided novel insights into utilizing ML models to predict THMs levels in dual-source DWDS in the metropolitan.

Hydrolytic acidification integrated with denitrification (DEN) presents a promising treatment strategy for high -temperature industrial wastewater, such as oil, leather and ethanol wastewater. This study aimed to investigate the synergistic influence of temperature and the COD/NO3--N ratio on the nitrogen reduction pathway using a series of semi-continuous flow reactors. The results revealed that the optimal total nitrogen (TN) removal per-formance was achieved under thermophilic (T = 60 degrees C) and COD/NO3--N ratio of 37, with the highest TN reduction amount of 594.3 +/- 41.8 mg/L. However, under other conditions, the TN removal efficiency reduced due to NH4+-N accumulation. Higher temperature facilitated the nitrogen reduction rate and butyrate yield rate, while inhibiting the dissimilatory reduction to ammonium (DNRA) process to 20.7 %. Furthermore, the higher temperature increased the bacterial diversity and the relative abundances of Clostridium, Acidaminococcus and Ruminococcus, thus, enhancing organic matter hydrolysis and nitrogen removal efficiency. In contrast, Carno-bacterium, with percentage account of 34.4 %, was only detected under thermophilic conditions with a low COD/ NO3--N ratio. The COD/NO3--N ratio only generated a slight effect on the production of volatile fatty acids (VFAs) and nitrogen reduction pathways, but had significant influence on the microbial community. These findings provide novel insight into the implementation of hydrolysis acidification coupled with anaerobic denitrification process in thermophilic wastewater treatment.

Herein, we employ machine learning to deliver the accurate prediction of tetracyclines (TCs) adsorption performance by biochar via inputting several numeric/nonnumeric information over physiochemical properties of biochar and environmental pressures. We find the TCs adsorption by biochar is a pure physical behavior where porous filling is the primary mechanism regardless of the complex surface modification. It is the changes in C-contained and O-contained functional groups not the surface modifying materials as such that improve TCs adsorption amount. Furthermore, adsorption amount is mainly subject to the environmental pressures, and maximum adsorption amount is subject to synergism of physiochemical properties of biochar and environmental pressures, and adsorption kinetics is mainly subject to the physical properties of biochar. Simplified model only with a handful independent variables input can also accurately predict TCs adsorption performance by biochar, which avoids expensive while complex instruments operation and provides portable prediction models in urgent scenario.

针对我国东南沿海地区4个使用不同类型二次消毒剂(氯、氯胺、二氧化氯)的局部管网水样的中三卤甲烷(THMs)、卤乙酸(HAAs)、卤乙腈(HANs)的分布特征及影响因素进行了分析.结果表明,3座城市DBPs的超标风险整体较低,采集的117份水样中有10份水样THMs超过国标限值,其余DBPs均符合标准.以氯为消毒剂的管网水样中THMs浓度和HANs浓度远高于其他管网,以二氧化氯为消毒剂可以减少消毒副产物THMs的生成,但以氯胺为消毒剂的上海地区HAAs浓度较高,是其他3个系统的5～19倍,可能与原水中HAAs前体物含量较高有关.在SH-NH₂Cl系统中,消毒剂浓度与输配距离成反比(r =-0.57),同时该系统中硝酸盐、亚硝酸盐及氨氮浓度也随距离变化,说明饮用水在管网输送过程中余氯被逐渐消耗,同时管网中发生了硝化反应,氨氮被氧化为亚硝酸盐和硝酸盐.WJ-Cl2系统中DBPs与输配距离相关,HANs浓度沿程减少(r =-0.49),HAAs浓度沿程增加(r = 0.45).此外,建筑管网中长停滞时间也会影响水样理化指标.该研究揭示了不同类型二次消毒剂的管网水中DBPs浓度水平和变化规律,研究结果可为不同地区水质安全评价和风险控制提供依据.