Abstract：
The occurrence and removal of 38 antibiotics from nine classes in two drinking water treatment plants (WTPs) were monitored monthly over one year to evaluate the efficiency of typical treatment processes, track the source of antibiotics in tap water and assess their potential risks to ecosystem and human health. In both source waters, 18 antibiotics were detected at least once, with average total antibiotic concentrations of 538.5 ng/L in WTP1 and 569.3 ng/L in WTP2. The coagulation/flocculation and sedimentation, sand filtration and granular activated carbon processes demonstrated limited removal efficiencies. Chlorination, on the other hand, effectively eliminated antibiotics by 48.7 ± 11.9%. Interestingly, negative removal was observed along the distribution system, resulting in a significant antibiotic presence in tap water, with average concentrations of 131.5 ng/L in WTP1 and 362.8 ng/L in WTP2. Source tracking analysis indicates that most antibiotics in tap water may originate from distribution system. The presence of antibiotics in raw water and tap water posed risks to the aquatic ecosystem. Untreated or partially treated raw water could pose a medium risk to infants under six months. Water parameters, for example, temperature, total nitrogen and total organic carbon, can serve as indicators to estimate antibiotic occurrence and associated risks. Furthermore, machine learning models were developed that successfully predicted risk levels using water quality parameters. Our study provides valuable insights into the occurrence, removal and risk of antibiotics in urban WTPs, contributing to the broader understanding of antibiotic pollution in water treatment systems.
摘要：
在一年中每月监测两个饮用水处理厂（WTP）中9类38种抗生素的发生和去除情况，以评估典型处理过程的效率。追踪自来水中抗生素的来源，并评估其对生态系统和人类健康的潜在风险。在两个水源中，至少一次检测到18种抗生素，WTP1中平均总抗生素浓度为538.5ng/L，WTP2中平均总抗生素浓度为569.3ng/L。混凝/絮凝沉淀,砂滤和颗粒活性炭工艺显示出有限的去除效率。氯化,另一方面,有效消除抗生素48.7±11.9%。有趣的是,沿分配系统观察到负去除，导致自来水中大量存在抗生素，WTP1中的平均浓度为131.5ng/L，WTP2中的平均浓度为362.8ng/L。来源追踪分析表明，自来水中的大多数抗生素可能来自分配系统。原水和自来水中抗生素的存在对水生生态系统构成了风险。未经处理或部分处理的原水可能对6个月以下的婴儿构成中等风险。水参数,例如,温度,总氮和总有机碳，可以作为评估抗生素发生和相关风险的指标。此外，开发了机器学习模型，该模型使用水质参数成功预测了风险水平。我们的研究提供了对这种情况的宝贵见解，城市WTP中抗生素的去除和风险，有助于更广泛地了解水处理系统中的抗生素污染。
