Secondary water supply systems (SWSSs) are widely used to supply water to high-rise households in urban residential buildings. A special mode of double tanks with one used while another was spared was noted in SWSSs, which would facilitate microbial growth due to longer water stagnation in the spare tank. There are limited studies on the microbial risk of water samples in such SWSSs. In this study, the input water valves of the operational SWSSs consisting of double tanks were artificially closed and opened on time. Propidium monoazide-qPCR and high-throughput sequencing were performed to systematically investigate the microbial risks in water samples. After closing the tank input water valve, it may take several weeks to replace the bulk water in the spare tank. The residual chlorine concentration in the spare tank decreased by up to 85 % within 2-3 days compared with that in the input water. The microbial communities in the spare and used tank water samples clustered separately. High bacterial 16S rRNA gene abundance and pathogens-like sequences were detected in the spare tanks. Most antibiotic-resistant genes (11/15) in the spare tanks showed an increase in their relative abundance. Moreover, when both tanks within one SWSS were in use, the water quality of the used tank water samples deteriorated to varying degrees. Overall, running SWSSs with double tanks will reduce the replacement rate of water in one storage tank, and consumers who use taps served by the presented SWSSs may have a higher microbial risk.

Water quality deterioration often occurs in secondary water supply systems (SWSSs), and increased heavy metal concentrations can be a serious problem. In this survey, twelve residential neighborhoods were selected to investigate the influence of SWSSs on the seasonal changes in heavy metal concentrations from input water to tank and tap water. The concentrations of nine evaluated heavy metals in all groups of water samples were found to be far below the specified standard levels in China. The concentrations of Fe, Mn, and Zn increased significantly from the input water samples to the tank and tap water samples in spring and summer (p < 0.05), especially for the water samples that had been stagnant for a long time. Negative correlations were found between most of the heavy metals and residual chlorine (Fe, Cu, Zn, and As, r = -0.186 to -0.519, p < 0.05). In particular, a high negative correlation was observed between Fe and residual chlorine (r = -0.489 to -0.519, p < 0.01) in spring and summer. Fe and Mn displayed positive correlations with turbidity (r = 0.672 and 0.328, respectively; p < 0.05). In addition, Cr and As were found to be positively associated with some nutrients (NO3-, TN, and SO42-; r = 0.420-0.786, p < 0.01). The material of the storage tanks had little influence on the difference in heavy metal concentrations. Overall, this survey illustrated that SWSSs may pose a chronic threat to water quality and could provide useful information for practitioners.

Secondary water supply systems (SWSSs) are characterized by long water stagnation and low levels of chlorine residuals, which may pose a high microbial risk to terminal users. In this study, the SWSSs of 12 residential neighborhoods in a metropolitan area of 5 million people in southeastern China were seasonally investigated to assess their microbial risks by determining more than 30 physicochemical and biological parameters. Although the microbiological quality of SWSS water met the requirements of the standards for drinking water quality of China, it did deteriorate in various aspects. The heterotrophic plate counts with R2A media were high (> 100 CFU/mL) in some SWSS tank and tap water samples. Propidium monoazide (PMA)-qPCR revealed a one magnitude higher abundance of viable bacteria in the tank and tap water samples (average 103.63±1.10 and 103.65±1.25 gene copies/mL, respectively) compared with the input water samples, and Enterococcus, Acanthamoeba, and Hartmannella vermiformis were only detected in the tanks. In particular, the high detection frequency of Legionella in 35% tank and 21% tap water samples suggested it is a supplementary microbial safety indicator in SWSSs. The microbial regrowth potential was more obvious in summer, and Illumina sequencing also demonstrated distinct seasonal changes in the relative abundance of bacterial gene sequences at the genus level. Turbidity and residual chlorine were closely connected with total bacterial biomass, and the latter seemed responsible for microbial community structure alteration. The extremely low chlorine residuals associated with a high abundance of total bacteria (as high as 106.48 gene copies/mL) and Legionella (as high as 106.71 gene copies/100 mL) in the closed valve tanks highlighted the high microbial risk increased by mishandling the operation of SWSSs. This study found that SWSSs possessed a higher microbial risk than the drinking water network, which suggested that the frequency and scope of monitoring the microbial risk of SWSSs in megacities should be strengthened for the purpose of waterborne epidemic disease prevention and control.

Small-scale distributed water purifiers (SSDWPs), providing better quality drinking water, are popularly used both in homes and in the public domain. Non-continuous operation leads to water stagnation and ultimately induces microbial contamination. However, information related to such contamination in these purifiers is reported scarcely. In the present study, an SSDWP, consisting of sand filtration (SF), granular activated carbon (GAC), and ultrafiltration (UF) processes, was established to explore microbial changes induced by water stagnation, based on the aspects of bacterial count, microbial size, microbiome and pathogenic communities. Our results primary showed that: first, compared with drinking water distribution system (DWDS), bacterial counts increased more rapidly in SSDWPs, growing to > 500 cfu/mL after 2.5 h stagnation. The proportion of intact cells also increased with stagnation time. Conversely, microbial size decreased with stagnation time according to changes in forward scatter detected using flow cytometry. Second, microbiome evolution followed the isolated island model, while in stagnated DWDS, microbiome evolved according to the continent island model, and the former had higher abundance of biodiversity. Furthermore, stagnation evidently caused microbiome changes in each unit, and spatial differences contributed to microbiome dissimilarity more significantly than temporal differences. Third, Mycobacterium was the dominant pathogenic genus in the SF and GAC units while Acinetobacter was the most abundant in the UF unit. Pathogenic risks increased with water stagnation time and lower nutrients level contributed to pathogenic community richness. Therefore, terminal disinfection of SSDWPs is strongly advised.

This paper studied stagnation-induced changes of disinfectant and bacteria using an orthogonal test and kinetic analysis, and then proposed a disinfection strategy. Tap water from a drinking water distribution system and ultrafiltrated water were collected and disinfected with four disinfectants (concentrations were set 0.2-1 mg/L as Cl2. The study had several findings. First, disinfectants expanded lag phases and shortened generation times of the microbiome. Reduction in culturability, substrate responsiveness, respiratory activity, membrane potential and integrity subsequently occurred with increasing disinfection concentration. Second, the disinfectant decay rate decreased with initial disinfection concentration, and the effective disinfection phase (heterotrophic plate count (HPC) was less than 100 cfu/mL) was longer in water samples with lower organic matter. Moreover, the disinfection process was divided into an effective phase and an invalid phase (HPC>100 cfu/mL). Then a disinfection efficiency model was built and the regulation of disinfection by-products (DBPs) production was studied in chlorinated water samples, which provides a general method for other disinfectant studies. The average trihalomethanes (THMs) production during the effective phase (marked as THM/th) and THMs production during the invalid phase (marked as ΔTHM) were proposed to evaluate the DBPs production. The level of THM/th and ΔTHM were lower in ultrafiltrated water than those in tap water. THM/th were negatively correlated with initial chlorine concentration while ΔTHM were positively correlated with initial chlorine concentration. Finally, for the purpose of raising disinfection efficiency and decreasing DBPs, we propose periodic pulse disinfection.