A Citizen Science Approach for Mapping Total Microbial Communities and Waterborne Pathogens in Household Drinking Water in China

2.1.1 Sites and Participants

The sampling sites (Figure 1A) were determined based on the coverage area (i.e., to cover as large an area as possible) and the possibility to recruit undergraduate student volunteers from the Duke Kunshan University (DKU) community. Due to the distribution of volunteers, samples mostly came from central and eastern China (latitude: ∼22°N - 40°N; longitude: ∼100°E - 122°E), including Beijing, Shandong Province, Jiangsu Province, Guangdong Province, etc.

• Download figure
• Open in new tab

Figure 1 Geographical locations of sampling sites, sampling regions, sampling kit, and procedure.

(A) Regions 1 to 7 represent Beijing-Tianjin region, North (Shandong and Hebei), Northwest (Gansu and Shanxi), Central (Henan, Anhui, Jiangxi, and Hunan), Central Coast (Jiangsu, Shanghai, and Zhejiang), South Coast (Fujian, Guangdong and Macau), and Southwest (Sichuan, Chongqing, and Yunnan) respectively (Software: Datawrapper) (left). (B) A screenshot from the demonstration video showing the sampling kit (upper right). (C) A screenshot from the demonstration video showing the filtering process (lower right).

All student volunteers were recruited after a simple screening process. First, their relevant experiences (e.g., majors) were considered and those with natural science major and lab experience were preferred. Besides, ideal volunteers would go home or go on a trip to any place(s) in China during the sampling period. Volunteers who met those two criteria or at least one of them were recruited through one-to-one conversation on WeChat (a Chinese social media app). Communication between volunteers and investigators is a key component of this study to ensure the quality of the samples as much as possible. This is the main reason why volunteers were recruited from the DKU community – they can easily reach out to us either in person or via WeChat once any problem emerges.

2.1.2 Sample Collection

2.1.3 DNA Extraction, PCR Amplification, and Bacterial 16S Metabarcoding

At the DKU Environmental Research Lab, all samples were stored in a −80L freezer prior to processing. For each sample, the filter membrane was removed from the sealed syringe filter unit and cut into eight strips using a sterile razor blade on a disposable petri dish. DNA extraction was then conducted using the Qiagen DNeasy Plant Mini Kit following the manufacturer’s instructions, with the exception of an additional bead-beating step implemented to enhance extraction efficiency. In brief, the filter strips were placed into a 2 mL microcentrifuge tube where 0.2 g of 0.1 mm Zr bead and 400 microliters of lysis buffer AP-1 were added for beat beating at 2000 rpm for 5 minutes. The final extracted DNA of each sample from the kit was dissolved in 40 μL Qiagen elution buffer and stored in a −80L freezer.

For PCR amplification, the V4 region of 16S rRNA gene was amplified by the universal primer pairs 515F (5’-GTGCCAGCMGCCGCGGTAA-3’) and 805R (5’-GACTACNVGGGTATCTAAT-3’) with dual barcode index and heterogenous spacers (Lin et al., 2019, Kozich et al., 2013). KAPA HiFi PCR Kit and the manufacturer’s protocol was adopted. All PCR reactions were performed in triplicates with 25 μL of each reaction mixture. Agarose gel electrophoresis was then performed to visualize amplicon fragments and PCR products were purified using QIAquick PCR Purification Kit (Qiagen) following the manufacturer’s instructions. The purified PCR amplicons of each sample was dissolved in 30 μL elution buffer and stored in a −80L freezer.

Finally, the concentration of the purified PCR amplicons was measured by Qubit^™ 4 Fluorometer. Equimolar amounts of purified amplicons were pooled together and sent to Genewiz in Suzhou for an Illumina Miseq (250 PE) sequencing run.

2.1.4 Auxiliary Data Collection

Previous studies found that drinking water microbiome exhibited seasonality which was correlated with temperature (Ley et al., 2020, Pinto et al., 2014). Given the difficulties in measuring water and indoor temperature with the citizen science approach, outdoor temperature was collected as the indicator of seasonal environmental temperature change. The outdoor temperatures when collecting valid samples (n = 40) were retrieved from Weather Underground (2022), an online portal that provides local weather data on an hourly basis.

In addition, to assess the potential risks of harmful bacteria detected in sampled tap water, information such as drinking water standards and microbial fact sheets was retrieved from environmental and health authorities including The United States Environmental Protection Agency (EPA) (EPA, 2023, EPA, 2022), and World Health Organization (WHO) (WHO, 2017).

2.2.1 Sequence Processing

To profile the microbiome in sampled tap water, 16S rRNA gene sequences were processed by a series of bioinformatics tools. First, paired-end sequencing reads with dual indices were demultiplexed and then trimmed to remove barcodes and primers using Cutadapt (Martin, 2011). The resulting reads were then further processed following the DADA2(v1.16) pipeline (Callahan et al., 2016). Specifically, using the embedded functions in DADA2, quality filtering was performed before merging paired-end sequencing reads. After chimera checking, Amplicon Sequence Variant (ASV) was identified, and the abundance table for each sample was constructed. Finally, the “assignTaxonomy” function in DADA2 was used to assign taxonomy to each ASV based on Silva r138 reference database (DOI 10.5281/zenodo.4587955).

2.2.2 Assessment of Total Microbial Communities

The relative abundance (RA) of all ASV in each sample was calculated and the following analyses of microbial compositions were based on RA. Apart for the samples collected after Typhoon In-Fa and 2021 Henan Floods (“post-weather samples”), the remaining samples (“normal samples”) were categorized in two ways to examine the spatial and temporal patterns of microbial community compositions. Firstly, due to the vast geographical scope of the sampling sites, the “normal samples” were divided into seven regions for region-based analysis (Yang et al., 2021, Xie et al., 2022). These regions, designated as Regions 1-7, were determined based on the geolocation feature of China, which are Beijing-Tianjin region, North, Northwest, Central, Central Coast, South Coast, and Southwest respectively (Figure 1). Secondly, to account for potential impacts of climate and seasonality, the “normal samples” were divided into three nearly equal groups based on the outdoor temperatures: (a) High (H): T > 25 °C; (b) Medium (M): 15 °C < T ≤ 25 °C; (c) Low (L): T ≤ 15 °C (Table 1). The mean RA of each ASV in the seven geographic regions and three temperature categories was calculated for the class- and ASV-level analyses.

To investigate the alpha diversity of tap water microbial communities, each library was resampled in equal depth, and Chao1, Fisher, Shannon, and Simpson diversity indices were then calculated from observed read counts of ASVs using the “Phyloseq” package (version 1.38.0) in R (McMurdie and Holmes, 2013). The Shannon and Simpson indices including both ASV richness and evenness were computed because of their reduced sensitivity to differences in sample depth (Haegeman et al., 2013, Preheim et al., 2013).

Beta diversity between drinking water microbiomes was calculated based on the ASV table. Beta diversity was assessed by calculating Bray-Curtis dissimilarity with respect to the RA of each ASV using “Phyloseq”.

2.2.3 Assessment of Potential Waterborne Bacterial Pathogens

The bacteria genera in the dataset that containing pathogenic species were selected based on Aquatic Pollution: An Introductory Text (Laws, 2017) and “Guidelines for Drinking-water Quality (4^th Edition)” by WHO (2017). Subsequently, the taxonomy of the resulting ASVs was further validated using the BLAST+ tool (Camacho et al., 2009) against the NCBI database during February 2022. Based on our criteria, only BLAST results with percent identity (p-ident) > 97% (Johnson et al., 2019) and expect value (E-value) < 10e-100 (Vej, 2007) were considered reliable results.

Subsequently, all confirmed pathogen genera/species were grouped into two categories based on their occurrence in the drinking water samples. Those detected in more than 30% of all samples (i.e., 7 samples) were categorized as “common pathogens” while the rest were referred to as “rare pathogens.”

2.2.4 Statistical Analyses

All statistical analyses were performed in R and a p-value threshold of 0.05 was considered significant. Fisher’s exact test (Upton, 1992) was conducted to determine whether there is a significant relationship between PCR signals (i.e., positive or negative) and average annual rainfall (AAR) (humid regions: > 800 mm, non-humid regions: < 800 mm). The grouping of sampling sites was based on Ma et al.’s (Ma et al., 2022) study which obtained AAR data from National Meteorological Information Center.

To analyze alpha diversity of total microbial communities, Dunn’s test for a post-hoc Kruskal−Wallis pairwise comparison (McKight and Najab, 2010) was conducted to identify significant differences between the seven geographical regions and the three temperature categories. For beta diversity analysis, a principal coordinates analysis (PCoA) was performed based on the distance matrix of Bray-Curtis dissimilarity to visualize the ordination among the samples by plot_ordination. To test the significant differences among groups of tap water samples, permutational multivariate analysis of variance (PERMANOVA) and a post-hoc pairwise Adonis test were performed with adonis2 function (Oksanen et al., 2015) and pairwise.adonis() function (Arbizu, 2020) respectively. To check if within-group variation is confused with among-group variation (Anderson, 2001), a permutation test for homogeneity of multivariate dispersions (PERMDISP) (Anderson, 2006) was performed with the betadisper() function. In PERMANOVA and PERMDISP, the number of permutations was set to 999.

To analyze the compositions and influencing factors of waterborne pathogens in the tap water, several correlation tests based on Spearman’s rank correlation coefficients were conducted. First, Mantel test (Smouse et al., 1986) was conducted to examine correlations between total microbial communities (top 100 ASVs), potential pathogens, geographic locations, and outdoor temperatures based on Bray-Curtis dissimilarity, Euclidean Distance, and Haversine Distance matrices (permutations = 9999). Second, linear regression fits between outdoor temperatures and RAs of total potential pathogens were examined and visualized with ggscatter() in the “ggpubr” package. Third, linear correlations between RAs of potential pathogens, RAs of total microbial communities (Top10 ASVs), and alpha diversity of total microbial communities were assessed with cor(). The corresponding p values and confidence intervals were computed with cor.mtest(). The matrix with significance level codes was visualized with corrplot.mixed().

3.4.1 Compositions of Total Microbial Communities

3.4.1.1 Phylum Level Pattern by Sample

52 microbial phyla including 46 bacterial phyla and 6 archaeal phyla were detected in the household drinking water samples (n = 22). In all samples, 9 bacterial phyla and 1 archaeal phylum accounted for over 90.7% of total taxonomically assigned reads at phylum level (Table 2). All the ten phyla are both abundant and prevalent (i.e., occur in 91-100% of all samples), indicating a relatively even microbial distribution pattern in the samples at the Phylum level. The five most dominant bacterial phyla were Proteobacteria (mean percentage + SD: 55.0% + 19.8%), Planctomycetota (10.5% + 7.9%), Acidobacteriota (7.0% + 6.0%), Actinobacteria (5.9% + 6.6%), and Cyanobacteria (4.7% + 3.5%), comparable to a previous study in China (Han et al., 2020). Three of them were reported tolerant to drinking water treatment and distribution processes except for Planctomycetota and Acidobacteriota (Han et al., 2020). Interestingly, this bacterial composition differs from the primary bacterial assemblages revealed by several highly-cited studies conducted in other countries including the U.S. and Portugal, which predominantly feature Proteobacteria, Actinobacteria, and Bacteroidetes (Hull et al., 2017, Pinto et al., 2012, Ivone et al., 2013). This implies a unique microbiome intrinsic to China’s drinking water systems.

View this table:

• View inline
• View popup
• Download powerpoint

Table 2 Number of sequences, ASVs, and genera for the top 10 phyla in the water samples.

Those three represent the coverage, diversity, and the genus spectrum of microbial community in the water samples respectively. ^a The percentage of sequences was the phylum sequences in the total of assigned sequences at the phylum level which was 1,333,322.

The phylum-level taxonomic composition for each sample is detailed in Figure 3 and Table S3. In this study, Proteobacteria (class α- and γ-Proteobacteria) was the most predominant phylum in 20 samples and the second most in others. It accounts for 29.9 – 99.0% of the reads in each sample, with the highest relative abundance detected in Huizhou (HZ_0219). Among all the samples, Sphingomonas was the most abundant genus of Proteobacteria. This result was consistent with the previous study (Han et al., 2020) that found Proteobacteria to be dominant in tap water collected mostly from central and eastern China, and the genus Sphingomonas grew during chlorination (Jia et al., 2015) or monochloramine treatment (Chiao et al., 2014).

• Download figure
• Open in new tab

Figure 3

Taxonomic composition and relative abundance of microbiota in household drinking water in China at the phylum level. “*” denotes the samples collected after extreme rainfall events. Normal samples are grouped based on the outdoor temperature. Within each temperature category, samples are in the order of Regions 1-7. Only the top 10 phyla (1-10: bottom to top in the legend) and not assigned (NA) ones are shown.

However, two samples CZ_0220 (Changzhou, Feb. 20) and NC_0210 (Nanchang, Feb.10) were dominated by Crenarchaeota, a common archaeal phylum. Specifically, Crenarchaeota accounted for 33.3% and 35.0% with the genus Candidatus Nitrosotenuis (32.0%) and Candidatus Nitrosotalea (34.9%) being most abundant in samples CZ_0220 and NC_0210 respectively. This showed that in addition to a variety of bacteria, archaea can also grow in tap water, which could be supported by the studies which detected the archaeal phylum Crenarchaeota in drinking water distribution systems or drinking water-related environments (Roeselers et al., 2015, Inkinen et al., 2021, Dai et al., 2020, Franca et al., 2015, Bautista-de los Santos et al., 2016). In particular, Inkinen et al. (Inkinen et al., 2021) found a high abundance of archaeal reads from the genus Candidatus Nitrosotenuis and Candidatus Nitrosotalea in drinking water distribution systems supplying non-disinfected waters. This suggests that the disinfection processes of samples CZ_0220 and NC_0210 may be less effective compared to others.

Interestingly, compared to CZ_0220, the pre- and post-typhoon samples collected from the same city household in Changzhou in July were much more similar in composition at the phylum level. However, the post-typhoon sample (CZ_0728) was richer in Actinobacteria (increased from 7.1% to 2.2%) and Cyanobacteria (increased from 5.2% to 14.9%) which are the phyla containing potential waterborne pathogens and the species that produce cyanotoxins respectively. Elevated levels of the pathogen Mycobacterium spp. (more details in Section 3.5 and Supplementary Material: SM1.1) as well as toxin producing Cyanobacteria spp. were identified in CZ_0728. Specifically for cyanobacteria, Microcystis spp. were higher in RA while Cylindrospermopsis sp. and Dolichospermum sp. appeared after the typhoon event. Moreover, other toxic species of Cyanobacteria including Aphanizomenon sp. and Anabaena sp. were detected in other tap samples collected from Shanghai, Lanzhou, Xi’an, etc. Many Cyanobacteria spp. from those genera can produce a variety of cyanotoxins such as Microcystins and Cylindrospermopsin which can cause damage to the liver and kidney with potential carcinogenicity (EPA, 2022). Similarly, a substantial rise in the RA of Cyanobacteria occurred in treated water samples collected from a drinking water treatment plant in Jiangsu was reported after Typhoon Lekima in August 2019 (P < 0.05) (Tang et al., 2021). In this study, although remained detectable, the proportion of Cyanobacteria decreased to the pre-typhoon level by the third day after the typhoon event.

3.4.1.2 Class Level Taxonomy

The geographical and temporal distribution of microbial classes in household drinking water in China is illustrated in Figure 4. The mean RAs of the top 10 microbial classes in all the normal samples, seven regions, and three temperature categories are summarized in Table S4. The analysis of the top 20 ASVs (mean RA) in each geographical region and temperature group is shown in Figure S2 & SM 1.1 and the detailed taxonomy classification of each ASV is provided in Table S5.

• Download figure
• Open in new tab

• Download figure
• Open in new tab

Figure 4

Taxonomic composition and relative abundance of microbiota in household drinking water in China at the class level. The pie charts show the mean composition of top 10 bacterial and archaeal classes in (A) seven geographic regions and (B) three temperature categories.

Among all the regular samples, the 10 most frequently detected classes were Alphaproteobacteria (mean RA: 37.4%), Gammaproteobacteria (16.3%), Planctomycetes (8.0%), Blastocatellia (5.7%), Actinobacteria (4.2%), Nitrososphaeria (3.6%), Bacteroidia (2.6%), Verrucomicrobiae (2.2%), Vampirivibrionia (2.1%), and Cyanobacteriia (1.8%). While many classes were shared across various regions, their prevalence varied (Figure 4A). Specifically, classes such as Alphaproteobacteria, Gammaproteobacteria, Blastocatellia, and Planctomycetes dominated all the seven regions, yet their RA differed. Notably, Alphaproteobacteria exhibited much higher RA in samples from R1: Beijing-Tianjin region (59.7%) and R6: South Coast (50.0%) than in those from R3: Northwest (25.6%) and R5: Central Coast (28.2%). On the other hand, some classes only predominated a few regions (Figure 4A). For example, Methylomirabilia and Babeliae were among the top 10 classes only in R4: Central (1.6%) and R6: Central (1.0%), respectively.

Similar to the region-based analysis, the three temperature categories shared many dominant classes (top 2: α-& γ-proteobacteria), but a few classes were found to be abundant in only certain categories (Figure 4B). For instance, Nitrososphaeria was the fourth abundant class in Medium T and Low T, but not dominated High T. Besides that, Cyanobacteriia (2.3%) and Acidimicrobiia (1.6%) were only among the top 10 classes of High T whereas Phycisphaerae (2.1%) and OM190 (1.8%) were only abundant in Low T.

Multiple factors could account for the above differences in the ultimate microbial compositions at the tap. For example, local geology, the source water properties, water treatment & distribution processes, (Ji et al., 2015) etc. In addition, historical events that happened in the sampling region, such as agricultural runoff, could also explain specific patterns of microbial communities (Holinger et al., 2014).

3.4.2 Alpha-and Beta-Diversities

Alpha-and beta-diversities have been calculated at the ASV level to further describe the microbial communities in the samples. Alpha diversity refers to diversity (richness or evenness) within an ecosystem at the local small scale, while beta diversity measures the amount of differentiation between different ecosystems or local species communities (Andermann et al., 2022). The alpha diversity indices among different geographical regions and outdoor temperature categories as well as the weather samples were calculated (Figure 5 and Table S6). Tap water samples from R3: Northwest & R7: Southwest and High T category (Shannon and Simpson) constituted the most diverse microbial community while samples from R6: South Coast & R1: Beijing-Tianjin Region and Medium T category had the lowest alpha diversity. Though all the diversity indices showed some differences, none of the differences were statistically significant (P values > 0.05, adjusted by multiple methods in Dunn’s test). Overall, the alpha diversity of the tap water microbial communities was not largely region-dependent, and the outdoor temperature was not an influential environmental driver of the alpha diversity. This may be largely due to the small sample size and the long sampling period (from late January to late August in 2021, but the sampling events concentrated in February and July).

• Download figure
• Open in new tab

• Download figure
• Open in new tab

• Download figure
• Open in new tab

Figure 5

The α diversity (ASV-based). (A) geographic regions, (B) outdoor temperature categories, and (C) samples related to extreme rainfall events.

The extreme rainfall events had a clear impact on the alpha diversity of the microbiome in tap water samples collected from Changzhou and Zhengzhou, the two cities impacted by extreme rainfall events in 2021 summer. Compared to the two samples collected in winter (CZ_0220 & ZZ_0219), higher alpha diversity was observed in the post-weather summer samples in both cities (Figure 5C). Among the three samples collected from the same household in Changzhou, however, the highest diversity was observed in the pre-typhoon summer sample CZ_0712, indicating that outdoor temperatures might be a stronger driver of the water microbiome diversity than the typhoon.

Outdoor temperature plays a significant role in shaping the community composition and structure at the ASV level. The beta diversity of the total microbial communities in sampled tap water is shown in Figure 6. Principal Axis 1, 2, and 3 for PCoA (Bray-Curtis dissimilarity) represent 20.4%, 14.5%, and 8.1% of the variation among the samples respectively. The PCoA ordination illustrates that despite the different sampling regions, the samples in High T category clustered closely except BJ_0620 which was greatly dominated by Proteobacteria (87.5%). In contrast, samples in the other two categories, especially Medium T exhibited dispersive distributions. According to PERMANOVA and PERMDISP (Table S7), the influence of outdoor temperatures was statistically significant (R² = 0.24, adjusted P = 0.012 between High T and Low T), and the differentiation was not due to differences in group dispersions (P = 0.17). However, when making comparisons among normal samples (Figure S3 and SM 1.2), samples from different regions were greatly mixed and scattered except for the two samples in R7: Southwest (CQ_0715 & LJ_0722), which corresponds to the results of top 20 ASV analysis (Figure S2). A Mantel test (Table S8) indicate a strong correlation between microbial community structures and outdoor temperatures (Spearman’s r = 0.898, P = 0.0001), whereas geographic locations had minimal impact (Spearman’s r = −0.179, P = 0.9792).

• Download figure
• Open in new tab

Figure 6

Principal coordinates analysis (PCoA). Based on the distance matrix of Bray-Curtis dissimilarity of microbial profiles (ASV-based) among all samples (n = 22). HW: the weather samples collected in summer; CZ: samples collected in the same household in Changzhou; ZZ: samples collected in two nearby households in Zhengzhou.

Research spanning China, the Netherlands, and the U.S. has consistently found that temporal temperature variation, primarily driven by seasonal changes, is a controlling factor for the tap water microbiome (Zhang et al., 2021b, Zlatanović et al., 2017, Ley et al., 2020). Complementing this, another study in China revealed that increasing air temperature through indoor heating could lead to dramatic changes in the composition of the bacterial community in overnight stagnant tap water (Zhang et al., 2021a). These findings challenge the previous discovery by Han et al. that only a weak correlation was found between air temperature and tap water bacterial community in China (Spearman’s r = 0.088, P = 0.092) (Han et al., 2020). The inconsistency might be caused by differences in sampling year, sites, and geography coverage. Notable, the source water microbiome could exert a substantive influence on tap water microbiome (Han et al., 2020). On the one hand, tap water in the same city may need different treatment processes depending on the source water. On the other hand, adequate source water protection and control strategies are necessary for the prevention of microbial contamination from source water carried over to tap water.

The pre-typhoon summer sample CZ_0712 was distinct from the other two Changzhou samples but much like other samples in High T category, which further confirmed the impact of high outdoor temperature. Comparing the samples collected from Changzhou and Zhengzhou, the pre-weather winter samples differed moderately in microbial community composition and structure, however, the post-weather samples from the two cities (purple circle and square in Figure 6) were comparable to a large extent. This indicates a potential contamination mechanism – extreme rainfall events may cause the same or highly similar shift in tap water microbiomes by favoring certain fast-growing bacteria, increasing the similarity between dissimilar tap water microbial communities.

The above patterns observed in the drinking water microbiome suggests that outdoor temperature and extreme precipitation are important environmental factors for microbial community compositions and structures. Given the significant impact of temperature, in situ water temperature measurement should be incorporated in future citizen-based studies (Ley et al., 2020). This could be done by adding a portable thermometer to each sampling kit. Beyond this citizen science approach, other drivers should be systematically examined in the future for effective drinking water management. In addition, the division of Regions 1-7 in this study is subjective to some extent. A potentially more insightful categorization might focus on drinking water source types: river, lake and reservoir, and groundwater (Zhang et al., 2022). In a recent study on nationwide biogeography of bacterial communities in household drinking water in China, they found that source rivers (i.e., major river systems such as Yangtze River) and precipitation (i.e., local averaged annual rainfall) can also drive microbial dissimilarity of drinking water (Ma et al., 2022).

3.5.1 Representative Pathogen Species

Since only the V4 region of the 16S rRNA gene was sequenced in this study, most amplicon sequences were not able to provide adequate phylogenetic resolution for classification at the species and strain level. Therefore, some ASVs are not 100% confirmed pathogenic species/strains, but highly possible (see Table S10). A summary of representative potential pathogens in drinking water systems and their related diseases is provided in Table S11 (Ramirez-Castillo et al., 2015). A discussion of the following three as well as two other representative pathogen species that is primarily for the science communication purpose as provided in SM 1.3. E. coli, a common fecal indicator bacteria (WHO, 2017), was detected by this study. The most polluted tap water samples are CZ_0728 (RA: 0.61%), BJ_0620 (0.51%), ZZ_0802 (0.05%), CZ_0712 (0.03%), and JN_0219 (0.02%). The elevated RA of E. coli in the two post-weather samples suggests that the contamination might be related to the extreme rainfall events. Notably, the proportion of E. coli in BJ_0620 was dramatically higher than that of other normal samples. It is worth mentioning that the traditional E. coli tests can be a generally reliable indicator of enteropathogenic serotypes in drinking water, but potential viable while non-culturable E. coli cells could result in underestimations of the actual water contamination (Liu et al., 2008). Therefore, it is recommended to use PCR or quantitative PCR (qPCR) methods for the monitoring of E. coli (Wolf-Baca and Siedlecka, 2019).

Additionally, the RA of total Legionella spp. was most abundant in ZZ_0802 (1.32%), CZ_0220 (1.25%), and ZB_0502 (1.09%). Almost all species in the genera Legionella are thought to be potential human pathogens, but L. pneumophila (on Contaminant Candidate List 5 - CCL 5) is the main cause of Legionnaires’ disease (pneumonia) and Pontiac fever (a milder infection) (WHO, 2017, EPA, 2023). Potential L. pneumophila ASVs (Blastn best hit results) was detected in 22.7% (5) of the tap water samples (3 from Low T, 1 from Median and High T respectively). Among those samples, the highest L. pneumophila RA was observed in ZZ_0219 (RA: 0.237%), followed by MA_0524 (0.035%), NC_0210 (0.019%), XA_0217 (0.011%), and SH_0411 (0.010%). Moreover, some other pathogenic species such as L. oakridgensis and L. maceachernii occurred in TJ_0715, LZ_0822, XA_0217, LJ_0722, and ZZ_0802.

Salmonella spp., a highly pathogenic bacteria, was detected in one sample despite its known susceptibility to disinfection. Notably, Salmonella enterica (ASV 3082, RA: 0.39%) was observed in the post-flood tap water microbiome from Zhengzhou (ZZ_0802). However, the potential health risks remains uncertain, as the severity of the disease depends on the serotype and host factors of Salmonella (WHO, 2017). Nonetheless, the presence of Salmonella enterica after 2021 Henan Floods indicates potential fecal contamination in the household drinking water after an extreme weather event.

3.5.2 Potential Environmental Drivers: Extreme Precipitation and Hot Weather

The detection of a wide range of pathogens in tap water indicates the great impacts of extreme rainfall events on urban water management. Abnormal redistribution of precipitation patterns or extreme rainfall events could cause more rainfall to be concentrated in a short period, resulting in flooding and a large amount of effluent. These events can lead to huge challenges to urban water safety and pose severe threats on public and environmental health. On the one hand, excessive effluent load on urban wastewater treatment can overwhelm wastewater treatment plants, leading to untreated sewage being released into natural waterways, and resulting in freshwater resource pollution and aquatic ecosystem degradation (Langeveld et al., 2013). On the other hand, destructive rainfall could pollute drinking water sources and cause damage to buildings and infrastructure, such as leaks and breakage of tap water distribution systems. This will eventually lead to drinking water contamination and increased infection risks, which has been supported by the analysis of microbial community and pathogen changes in tap water in this research. As shown in Table S9, the typhoon and flood events generally altered the pathogen profiles in tap water samples by increasing the relative abundance of certain species and the number of species. Specifically, the RA of the total potential pathogens in the two post-weather samples ranked second (CZ_0728, 20.7%) and third (ZZ_0802, 10.7%) among all the samples (Figure 7). In addition, the number of pathogen species in Changzhou samples collected in July increased after Typhoon In-Fa. A similar pattern was observed in Zhengzhou samples and ZZ_0802 was the one with the highest number of pathogen species. More importantly, some devastating pathogen species (Salmonella spp. and Aeromonas hydrophila) were only detected in the post-weather samples as discussed above. In terms of the elevated RA of specific genera/species, most of the pathogens in the post-typhoon sample CZ_0728 had higher proportions compared to the pre-typhoon summer sample CZ_0712, except for Legionella spp. and Brevundimonas spp. which saw a decrease in RA, and Bacillus sp. being not detected in CZ_0728. Therefore, the rising risk of waterborne pathogens in post-flush tap water samples indicates that extreme rainfall events are a prominent risk factor in public health in terms of water (and food) safety (Phan and Sherchan, 2020). To maintain the standard of the treated discharge water during extreme rainfall events in the future, the government needs to invest more to upgrade the water processing capacity of wastewater treatment plants to be prepared for unpredictable effluent load. Meanwhile, since drinking water resource pollution is expected to happen following flooding, drinking water treatment plants and distribution systems should be designed, monitored, and maintained involving emergency response planning to secure their capacity for resisting the negative impact of extreme weather events.

Hot weather (increase of temperature) is another critical climatic factor that will strongly influence the hydrological regime and facilitate the spreading and proliferation of waterborne pathogen in freshwater resources. According to IPCC Technical Paper VI: Climate Change and Water (IPCC, 2008), the increase of surface water temperature has been associated with the atmospheric warming across Asia, North America, and Europe since 1960s. Lipp et al. has discovered that increasing temperature could create conditions that are advantageous for the proliferation of bacteria species less susceptible to temperature variations (Lipp et al., 2002). This could directly trigger the growth and expansion of some pathogenic species, such as cholera (V. cholerae), Vibrio parahaemolyticus, V. vulnificus and V. alginolyticus (Funari et al., 2012). Our results confirmed the above connections between climate change and increased waterborne infection risk by showing that hot weather is potentially another important environmental driver of pathogen profiles of tap water. The outdoor temperatures higher or equal than 17°C were positively correlated with RAs of total potential pathogens but the correlation is insignificant relative to the alpha level of 0.05 probably due to the small sample size (n = 12, R = 0.46, P = 0.13) (Figure S4). Besides that, higher RAs and a larger number of pathogen species were detected in the summer Changzhou sample CZ_0712 (20.7%, 7 genera, outdoor T: 33.3°C) compared to the winter sample CZ_0220 (1.80%, 4 genera, outdoor T: 17.0°C).

An important aspect to highlight here is that, apart from the separate effects of extreme rainfall and hot extremes on waterborne disease spreading, these extreme weather events are also intercorrelated in a way that creates a detrimental cycle, exacerbating the overall situation. IPCC AR6 report has confirmed that regardless of the emissions scenarios, the Earth’s surface temperature will keep increasing until at least the mid-21st century. Unless substantial reductions in CO2 and other greenhouse gas emissions occur in the coming decades, the global warming limits of 1.5°C and 2°C will be exceeded during the 21st century. Most importantly, the report underlined the fact with high confidence that with every additional increment of global warming, the impact on extreme weather events becomes more pronounced. For instance, for each additional 0.5°C rise in global temperatures, there are noticeable and undeniable rises intensity and frequency of extreme heat events, including heatwaves and intense rainfall (Masson-Delmotte et al., 2021). As a result, ensuring tap water safety in the future will become increasingly challenging and require stricter control and maintenance as the effects of climate change continue to intensify.

3.5.3 Potential Monitoring and Control Strategy

Spearman’s correlations between potential pathogens and alpha diversity indexes are shown in Figure 8. When the impacts of extreme rainfall events were not considered (Figure 8A, n = 22), the RA of Brevundimonas spp. and that of Mycobacterium spp. was positively correlated (r = 0.84, P < 0.001), and both were positively correlated with the total RA of potential pathogens (r = 0.60, 0.67, P < 0.001). Notably, a significant correlation between the low alpha diversity of the microbial community and the samples positive for the presence of E. coli (Chao1: r = −0.24, P < 0.05; Shannon: r = −0.21, P < 0.05) was observed. Similarly, Chopyk et al. (Chopyk et al., 2016) found the same pattern in their pre-harvest cattle hide samples. Therefore, it is worth understanding the relationship between indigenous water microbiomes and pathogenic E. coli, which might contribute to the control/prevention strategies for EHEC in tap water as well as indicators for risk assessment.

• Download figure
• Open in new tab

Figure 8

Spearman’s correlation. (A) alpha diversity indexes and potential pathogens detected in normal samples; (B) potential pathogens detected in all the samples. 1-3 asterisk(s) denote the significance level of 0.05, 0.01, and 0.001 respectively.

The incorporation of the two post-weather samples (Figure 8B, n = 24) altered the correlations between different pathogens and revealed new relationships due to the occurrence of other species in the samples. The positive correlation between the RA of Brevundimonas spp. and that of Mycobacterium spp. became weaker (r = 0.68, P < 0.05) while the RA of Aeromonas hydrophila was significantly correlated with S. enterica (r = 0.72, P < 0.01). Interestingly, S. enterica and B. diminuta were only detected in the post-flood sample ZZ_0802, suggesting potential co-occurring relationships between them. Future studies could further investigate the above potential relationships among waterborne pathogens in household drinking water.