Water Science and Technology最新文献

Olive mill wastewater (OMW) poses a serious environmental challenge, specifically in the Mediterranean region, due to its high content of phenolic compounds (PCs). In this study, eco-friendly nanocomposites made of chitosan, kaolinite, and iron oxide nanoparticles were prepared, characterized, and tested for their removal efficiency (RE) of PCs from OMW. The removal efficiencies of seven targeted PCs and the overall removal for the total phenolic content were evaluated. The nanocomposite powder cross-linked with glutaraldehyde exhibited the highest RE of 91% for the sum of the seven target PCs (Σ7PCs) using a 10 g/L of adsorbent dose, pH = 4.8, at a temperature of 25 °C within 2 h. Desorption studies showed that up to 85% of the adsorbed PCs were desorbed, allowing the efficient regeneration of the adsorbent for at least four cycles with RE exceeding 50%. These promising results suggest the potential of the large-scale utilization of the developed process for large-scale remediation of OMW.

Combined sewer overflow (CSO) screens are critical components of sewer and drainage networks, separating sewer solids from overflow spills before they reach receiving waters. Selecting suitable and sustainable CSO screening devices, however, remains a complex task. This process has traditionally depended on conventional design calculations, technical guidance from screen manufacturers and precedents from past projects. Inappropriate screen selections have led to adverse effects on water quality and public health, due to insufficient screening capacity, the unpredictable behaviour of sewer solids of varying densities, low trapping efficiency, frequent screen blinding or high equipment failure rates, particularly at unmanned or remote sites. This paper presents a design methodology for screen selection and formulates an input-output relationship model. Using 50 screen project data, a framework has been proposed to construct a predictive model that integrates sustainability criteria, lessons learnt from historical applications and artificial neural network (ANN) techniques. A Levenberg-Marquardt-based ANN was developed and trained to identify optimal selection between 2 categories of screen solutions, encompassing 12 screen types - 3 within non-powered self-cleaning and 9 within the powered screen category. The framework aims to provide an initial proof-of-concept evidence with a supplementary decision-support tool, enabling design engineers to make intelligent, resilient and sustainable choices in screen application.

Using opportunistic sensors, such as commercial microwave links from mobile networks, to estimate precipitation is an innovative and promising approach to improving hydrometeorological monitoring in urban areas. As part of the TOPRAINCELL project in Burkina Faso, a real-time system for collecting transmitted and received power data was deployed in collaboration with the national operator Telecel Faso. This study is based on data acquired in 2022 in the country's two main cities, Ouagadougou and Bobo-Dioulasso. These cities have different climatic contexts, yet they both have limited conventional rainfall coverage. Cross-analyzing data from opportunistic sensors, ground-based rain gauges, and satellites reveals a strong correlation between microwave link estimates and reference measurements, with Pearson coefficients reaching 0.97 in Ouagadougou and 0.94 in Bobo-Dioulasso. Spatial precipitation maps have been produced to demonstrate the ability of this multi-source approach to reproduce the spatial variability of urban rainfall. These results confirm the potential of opportunistic sensors as a complementary and adaptable solution for rainfall monitoring in West Africa.

The evaluation of water quality constitutes a critical aspect of water management strategies, particularly in arid and semi-arid environments, where the use and protection of sustainable resources are crucial. This study focuses on assessing and predicting water quality in three Jordanian dams using advanced data mining techniques. Physical, chemical, and biological water quality parameters were collected and analyzed over a four-year period. The Weighted Arithmetic Water Quality Index (WA-WQI) was used to evaluate the overall water quality. Various data mining algorithms, including generalized linear models, decision trees, random forests, gradient-boosted trees, and support vector machine (SVM), were employed to predict WQI and understand the seasonal and annual variations. Key findings highlight significant fluctuations in water quality, influenced by parameters such as pH, conductivity, nutrients, and microbial contamination. The study emphasizes the importance of continuous monitoring and predictive modeling for effective water resource management. It also demonstrates the effectiveness of using SVM for water quality prediction in arid regions. The models were evaluated using different performance metrics. The SVM outperformed other employed models. This study provides a critical benchmark and a robust predictive framework for water resource management in Jordan and semi-arid areas, addressing a significant gap in regional environmental monitoring.

This research paper presents an advanced water demand forecasting model through CaffeNet deep-learning architecture as well as a developed student psychology-based optimizer (DSPBO), aiming to improve the predictability of water consumption for the domestic, industrial, and agricultural sectors. The combined CaffeNet-DSPBO model has performed well in the performance evaluation to capture the complex nonlinear relationships caused by weather conditions, seasonality, and sector-specific patterns, and is trained using real data from the Yangtze River Delta of China. The main findings show a model with low RMSE values of 0.25 (domestic), 0.40 (industrial), and 0.58 (agricultural) and high correlation coefficients of 0.87, 0.75, and 0.62, respectively. This indicates that the domestic consumption sector, in particular, can be considered a reliable and accurate forecasting model. Also, the model demonstrated superior performance compared to other meta-heuristic algorithms in terms of convergence stability and solution accuracy. Another performance advantage is the training time of less than an hour and the inference latency of less than 10 ms. The results show how important this can be in combining deep-learning and better optimization techniques for predicting multi-sector water needs, paving the way for sustainable yet efficient management of this precious resource.

Digital twin models offer great potential for process improvements in wastewater treatment plants (WWTPs). Such models require a constant real-time input data feed from the physical process. Collecting these data is challenging, especially in the harsh conditions in the headworks of the process. In this study, data-driven models and process and sewer system expertise were combined to design soft-sensors for primary effluent COD and NH₄-N prediction. Ordinary least squares regression and the seasonal autoregressive integrated moving average model with exogenous variables were tested using flow rate and suspended solids concentration as model input. An excellent NH₄-N prediction was achieved, and the prediction accuracy was further improved by implementing process-insight-driven weights. The tested models were able to achieve either good COD estimation accuracy or effectively capture the variability in the target data. However, achieving both simultaneously remained challenging, with or without weights. Simulation tests using the calibrated process model demonstrated that the developed soft-sensors were able to provide real-time predictions leading to goodness-of-fit in simulations comparable to or better than that achieved using laboratory data influent quality.

To develop a more cost-effective nitrogen removal strategy, this study investigated the impact of low-temperature storage methods on nitrifying bacterial activity. Sludge was stored under laboratory-scale static batch conditions in three media: (1) distilled water, (2) nutrient solution, and (3) nutrient solution supplemented with hydroxylamine (NH₂OH). Ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) activity, sludge properties, and microbial characteristics were examined. Results revealed that all storage methods inhibited both AOB and NOB activity. Notably, nutrient solution storage demonstrated the most significant effect: it suppressed NOB activity by 86.6% and reduced its relative abundance by 20%, while maintaining high extracellular polymeric substance content (43.5 mg/g VSS) and AOB relative abundance (0.18%). This method substantially shortened the required storage duration (from 8 months to 60 days) and better preserved AOB activity and sludge stability. Metagenomic analysis indicated strong inhibition of the NOB functional gene nitrite oxidoreductase across all methods, while nutrient solution storage specifically elevated the abundance of the AMO gene. Although NH₂OH supplementation exhibited inhibitory effects on microorganisms, the concurrent addition of nutrient solution effectively mitigated this impact. Consequently, sludge properties and functional microbiota abundance showed no significant difference between the NH₂OH-supplemented nutrient solution method and distilled water storage.

Ultrafiltration (UF) membranes are widely used in potable water reuse, but their virus removal capabilities can be underestimated due to operational variability and membrane damage over time. This study evaluates the log reduction values (LRVs) of a pilot-scale UF system continuously processing tertiary treated wastewater, focusing on a compromised membrane. Virus removal was assessed under various operational states, including physical backwash (PBW) and chemically enhanced backwash (CEB). Samples were collected after CEB, before PBW, and after PBW. Indigenous viruses such as AiV, NoVGII, enteric AdV, PMMoV, CGMMV, and crAssphage were quantified using (RT-)qPCR, alongside spiked MS2 bacteriophage. A laboratory-scale study examined the synergistic effects of hydraulic and chemical stresses, with deteriorated membrane fibers analyzed through field emission scanning electron microscope (FE-SEM), SEM equipped with energy-dispersive X-ray spectroscopy (SEM-EDS), and liquid-liquid displacement porometry (LLDP). Despite structural damage and fouling observed in compromised fibers, the Kruskal-Wallis test revealed no significant differences (p > 0.05) in virus removal across operational states, indicating consistent UF performance. Laboratory-scale MS2 filtration studies showed a significant effect of water quality on increasing LRV (p < 0.05) in compromised fibers. This study underscores UF systems' robustness in virus removal and highlights membrane integrity loss pathways in real-world applications.

Large Language Models (LLMs) have shown strong performance across natural language processing tasks, yet their general-purpose embeddings often fall short in domains with specialized terminology and complex syntax, such as hydrology and environmental science. This study introduces HydroEmbed, a suite of open-source sentence embedding models fine-tuned for four QA formats: multiple-choice (MCQ), true/false (TF), fill-in-the-blank (FITB), and open-ended questions. Models were trained on the HydroLLM Benchmark, a domain-aligned dataset combining textbook and scientific article content. Fine-tuning strategies included MultipleNegativesRankingLoss, CosineSimilarityLoss, and TripletLoss, selected to match each task's semantic structure. Evaluation was conducted on a held-out set of 400 textbook-derived QA pairs, using top-k similarity-based context retrieval and GPT-4o-mini for answer generation. Results show that the fine-tuned models match or exceed performance of strong proprietary and open-source baselines, particularly in FITB and open-ended tasks, where domain alignment significantly improves semantic precision. The MCQ/TF model also achieved competitive accuracy. These findings highlight the value of task- and domain-specific embedding models for building robust retrieval-augmented generation (RAG) pipelines and intelligent QA systems in scientific domains. This work represents a foundational step toward HydroLLM, a domain-specialized language model ecosystem for environmental sciences.

Among emerging contaminants, pharmaceutical compounds have garnered significant scientific attention due to their presence in the environment and potential adverse effects on aquatic ecosystems and human health. The detection of pharmaceutical compounds, their ecological threat, and water quality was evaluated at six points along the Cauca River, Colombia's second most important river. The detected compounds included diclofenac, ibuprofen, naproxen, and paracetamol, with the latter presenting maximum concentrations of up to 4.20μg/L. Domestic wastewater discharges impacted the river's water quality, increasing the frequency and concentration of pharmaceutical contaminants. Ibuprofen and paracetamol were identified as high-risk compounds for aquatic biota, with Hazard Quotient (HQ) values between 190 and 250 in areas near urban wastewater discharges. This finding also indicated a high ecological risk due to the mixture of these pharmaceuticals. No single removal technology proved completely effective, highlighting the need for complementary treatments to conventional systems to ensure safe discharge into water bodies. Moreover, given the presence of these compounds in surface waters, drinking water treatment systems must be adapted to minimize health risks in distributed water. Finally, the study underscores the need for regulatory measures and continuous wastewater monitoring to protect both aquatic ecosystems and public health.