Investigating Melbourne’s drinking water fountains – a study of the microbial community at the point of use

In 2010, the World Health Organisation (WHO) stated that access to potable water is a basic human right. In a modern urban society, it is critical to maintain high quality potable water for the protection of the public. Water authorities are responsible for ensuring that drinking water is microbiologically- and chemically-safe in accordance with regulatory guidelines. They use a combination of sophisticated treatments to produce high quality water coupled with rigorous testing of water quality across distribution systems and networks. In Greater Melbourne, Australia, drinking water for the Western, Northern suburbs and for the City is sourced from reservoirs and undergoes treatment including chlorination before it enters the drinking water distribution system. Currently, the quality of drinking water is maintained by ensuring that physico-chemical and microbiological parameters remain within and/or below (maximum) guideline values. Water authorities typically utilise culture-dependent methods such as heterotrophic plate counts (HPC) and enzyme substrate tests (Colilert®) to measure microbial loads and to test for the presence of known pathogens or indicators. Whilst HPC methods provide basic data regarding the concentration of viable organisms it is not a true representation of actual numbers of microorganisms present and does not inform water companies about which organisms are there. In contrast, the water research community has regularly utilised culture-independent technologies to investigate microbial communities associated with many environments including source waters, treatment plants and the distribution system. However, detailed research investigating physico-chemical and microbiological aspects of public drinking water fountains (bubblers), where the product meets the customer, is lacking in the literature.

This research was conducted across 12 point-of-use public drinking water fountains located within the Melbourne Central Business District, comprising drinking water fountains of three different types. Fountain types A and B were installed from 2014, while type C fountains have been in use since 1996. Physico-chemical, microbiological and bacterial community analyses were undertaken on bulk water from each fountain, while the analysis of bacterial communities associated with biofilm samples taken from the fountain nozzle were also performed. Samples of both water and biofilm were taken at three independent sampling events during Dec '16 (summer), Apr '17 (autumn) and Sep '17 (spring).

The overall aim of this research was to investigate spatial and temporal variation in physico-chemical parameters and in the abundance and composition of bacterial communities within water and biofilm sampled from drinking water fountains across Melbourne.

The first aim was to determine variations in the physico-chemical and microbiological quality of drinking water samples, including comparison to data from the local water authority. Measurements of water quality parameters and ion concentrations were within guideline values and did not vary significantly from typical values reported by City West Water during the same period as sampling. The mean dissolved oxygen concentration and pH were negatively correlated with temperature while conductivity was positively correlated with temperature. There was no significant variation in any physico-chemical parameter between the three different types of fountain. Variation between individual fountains was limited, except for fountain 12, where conductivity and dissolved oxygen, sodium, potassium, magnesium, calcium, chloride, nitrate and sulphate concentrations in water were all significantly lower than in the other fountains. Across all fountains, both viable and total microbial counts were higher during the summer sampling event than in autumn or spring. Mean total microbial counts in Type B fountains were lower than in both Type A (p = 0.037) and Type C (p < 0.001) fountains while the mean total count in Type A fountains was lower than in Type C (p = 0.0031). Finally, the Colilert® most probable number (MPN) assay was used to detect and enumerate coliform and Escherichia coli numbers in drinking water from fountains. There were 5 replicates out of 48, in Dec '16, where the MPN was greater than 200 coliforms 100 mL-1. However, these values were within guideline values. Additionally, E. coli were not detected in any the 140 water samples taken across the three sampling events.

The second aim was to investigate and describe the variation in the structure, diversity and composition of bacterial communities within and between bulk water and biofilms with respect to individual sampling sites and fountain types. Bacterial communities within drinking water were richer and more diverse than those in biofilms, with significant differences between the compositions of the water and biofilm communities. Sphingomonads were most prevalent in bulk water, while Methylobacterium sp. tended to have the highest mean relative abundance in biofilms. Significant variation between the composition of bacterial communities was also observed between individual fountains (spatial variation) and between fountain types. The bacterial communities associated with both the water and biofilms within type C fountains were found to be richer, more diverse and exhibited greater variation in composition than those corresponding communities within type A or type B fountains. For example, in water, the mean Chao1 estimated number of OTUs within communities associated with type C fountains was 535 compared to 112 and 130 within fountain type A and type B, respectively. Similarly, in biofilm samples, the estimated number of OTUs were 80, 60 and 157 for communities within fountain types A, B and C, respectively. In both water and biofilm, the bacterial communities within fountain types A and B were dominated by Alphaproteobacteria and Betaproteobacteria. In contrast, the communities within water samples from type C fountains included taxa such as Deltaproteobacteria, Thermoleophilia and Planctomyces in higher proportions, while in biofilms, Actinobacteria (class), Cytophagia, Deinococci, Gemmatimonadetes and Saccharibacteria made significant contributions to the overall relative abundance of taxa.

The final aim of this research was to investigate and describe the variation and composition of bacterial communities within bulk water and biofilm samples with regards to sampling event (temporal variation). Analysis of bacterial community composition showed that the variations observed between (i) water and biofilm, (ii) individual fountains and (iii) different fountain types identified in Chapter 4 were also evident across the three sampling events (season). However, there was less variation in richness, diversity and composition of bacterial communities between sampling events. Core-microbiota were detected across all fountain types and during all sampling events. In water, these core taxa comprised 8.9% of all taxa identified but contributed 29.1-99.6% of the total relative abundance of all operational taxonomic units (OTUs) identified. Similarly, only 11.3% of the OTUs in biofilms were defined as core-microbiota, which contributed 63.8-99.9% of the total relative abundance. In agreement with earlier research, these outcomes show that a small number of taxa often contribute to a large proportion of the total relative abundance. Finally, the bacterial communities within water and biofilms from type C fountains were consistently richer and more diverse than those in type A or type B fountains regardless of season. Furthermore, nearly one third of the taxa in some type C fountains were found to be unique to that fountain type, with a significant proportion of these taxa identified as belonging to recently promoted candidate phyla including Armatimonadetes, Parcubacteria and Saccharibacteria.

This research provides valuable insights into the understanding of variation in bacterial numbers and in the structure, diversity and composition of bacterial communities within point-of-use drinking water fountains across the Melbourne Central Business District. Most notably, the existence of a temporally-conserved core microbiota coupled with variation in bacterial communities between individual fountains and fountain types may inform future developments and practices by water companies in water quality testing and management to ensure ongoing provision of high quality water in public drinking water fountains.