Unchartered waters: the unintended impacts of residual chlorine on water quality and biofilms

Abstract

Disinfection residuals in drinking water protect water quality and public heath by limiting planktonic microbial regrowth during distribution. However, we do not consider the consequences and selective pressures of such residuals on the ubiquitous biofilms that persist on the vast internal surface area of drinking water distribution systems. Using a full scale experimental facility, integrated analyses were applied to determine the physical, chemical and biological impacts of different free chlorine regimes on biofilm characteristics (composition, structure and microbiome) and water quality. Unexpectedly, higher free chlorine concentrations resulted in greater water quality degredation, observable as elevated inorganic loading and greater discolouration (a major cause of water quality complaints and a mask for other failures). High-chlorine concentrations also reduced biofilm cell concentrations but selected for a distinct biofilm bacterial community and inorganic composition, presenting unique risks. The results challenge the assumption that a measurable free chlorine residual necessarily assures drinking water safety.

Fig. 1. Drinking Water Distribution System (DWDS) experimental facility.

a The three independent high density polyethylene loops, each 203 m long, ensuring pipe effects dominate, with Pennine Water Group coupons (comprising outers and inserts) facilitating biofilm sampling^,,. Water quality was monitored using online metres and spot samples. b Schematic of one loop showing: enclosed reservoir tank, pump, type of pipelines, valves and monitoring equipment. C/T₁ denotes the location of turbidity and chlorine metres during growth, C/T₂ indicates the relocation of a set of metres (from one of the other loops) during flushing to enable extra monitoring of the loop being flushed.

Fig. 2. Discolouration responses to elevated shear stress during the flushing of the chlorine regimes.

Discolouration was determined primarily by a Turbidity (506 ≤ n ≤ 1091) with consideration of b Iron (n = 3) and c Manganese (n = 3) concentrations. Flush1 refers to the flushing phase of test 1, Flush2 indicates data from the flushing phase of test 2. Data normalised to well-mixed concentrations (0.09 Pa) of each system, mean ± standard deviation plotted. Linear regressions in each plot had R² values of a 0.82 ≤ R² ≤ 0.99, b 0.89 ≤ R² ≤ 1.00 and c 0.76 ≤ R² ≤ 0.98. High-chlorine: metal concentrations only available for final flushing step for Flush1. Chlorine regimes differed in their turbidity (ANCOVA on raw data: F ≥ 2869, p < 0.001), iron (ANCOVA on raw data: F ≥ 26, p < 0.001) and manganese (ANCOVA on raw data: F ≥ 10, p ≤ 0.003) responses.

Fig. 3. Quantification of iron in biofilms from the Pre- and Post-flush phases of the three chlorine regimes.

a Flush1, b, c Flush2. L Low, M Medium, H High; Asterisk significance determined via Wilcox-1-tailed tests (W = 9, 0.03 ≤ p ≤ 0.05), NS = not significant (W = 6.5, p = 0.24). Chlorine regimes differed at Pre-flush1 and Pre-flush2 χ² ≥ 5.65, p ≤ 0.05. N.b. different y-axis scale in c.

Fig. 4. Total and intact cell concentration within DWDS biofilms of each chlorine regime, sampled Pre- and Post-Flush.

a Test 1, b Test 2. L Low-chlorine, M Medium-chlorine, H High-chlorine; Asterisk indicates significant differences between pre- and post-flush biofilms, tested using Wilcox (0 ≤ W ≤ 9, 0.04 ≤ p ≤ 0.05); where no asterix is shown differences were not statistically significant (2 ≤ W ≤ 7, 0.20 ≤ p ≤ 0.80).

Fig. 5. Variation in bacterial and fungal communities of pre- and post-flush biofilm from Low-, Medium- and High-chlorine regimes.

nMDS plots based on Bray-Curtis similarities of pre-flush biofilm a 16S rRNA and b ITS mOTUs. Average relative abundance of c bacterial and d fungal genera, percentage similarity between bio-replicates (n = 5, or n = 4, see Methods section: Biofilm microbiome) is shown in brackets, _gx= genus unknown; c “Others” ≤1% total relative abundance (Supplementary Table 5); d UnknownA=Fungi, further taxonomic information unavailable. Pre1/Pre2 = Pre-Flush1 or Pre-Flush2, Post1/Post2 = Post-Flush1 or Post-Flush 2, L Low-chlorine, M Medium-chlorine, H High chlorine.