Impact of flushing procedures on drinking water biostability and invasion susceptibility in distribution systems

Abstract

Ensuring high-quality drinking water remains challenging, as complaints about odors, discoloration, or contamination persist. In Belgium and beyond, traditional flushing is a common curative strategy that involves discharging large water volumes through hydrants while the network remains in use. In some cases, free chlorine (0.5 mg/L) is added, and consumers are advised not to drink the water. However, flushing can alter water biostability, potentially increasing susceptibility to microbial invasion. This study used a pilot-scale drinking water distribution system with three identical 100 m polyvinyl chloride(PVC) loops (DN 80 mm) to assess the impact of flushing with and without chlorination as practiced in chlorinated networks. Loop 1 was flushed with tap water and sodium hypochlorite (NaOCl), followed by two non-chlorinated flushes, loop 2 was unflushed, and loop 3 underwent three flushes. Biostability was assessed using online flow cytometry, and susceptibility to bacterial invasion (Aeromonas media, Pseudomonas putida, and Serratia fonticola) was evaluated in the days following flushing. The water had a 7-day residence time. Results showed that chlorinated flushing promoted microbial regrowth (3.8 × 10⁵ vs 2.0 × 10⁵ and 1.6 × 10⁵ cells/mL for loops 1, 2, and 3, respectively), primarily of resident Sphingopyxis spp. Biofilm cell densities (~4 × 10⁶ cells/cm²) remained stable across conditions. Bacterial indicators declined over time, with P. pudita and S. fonticola surviving longer (>100 hours) than A. media (13 hours). Decay rates were highest in chlorinated loops, likely due to increased microbial competition. For example, the decay constant of S. fonticola at 20°C was -0.082 h^-1, -0.042 h^-1, and -0.027 h^-1 for loops 1, 2, and 3, respectively.

Importance: Traditional flushing is used as a curative strategy to solve unwanted quality issues during distribution, yet its impact on microbial biostability remains poorly understood. This study provides critical insights into how traditional flushing, both with and without chlorination, influences microbial regrowth and susceptibility to invasion. Findings reveal that chlorinated flushing promotes the regrowth of resident drinking water bacteria while accelerating the decay of introduced unwanted bacterial indicators, emphasizing the complex trade-off between microbial control and system stability. Understanding these dynamics is essential for optimizing flushing procedures, minimizing unintended consequences, and improving distribution system resilience.

Keywords: biofilm; biostability; drinking water microbiology; flushing; invasion.

Fig 1

Cell concentration and phenotypic diversity of the bulk water community in the function of time. Average concentration (cells/mL) in the function of time (days) of the microbial bulk community in loop 1 (orange), loop 2 (green), and loop 3 (blue) at (A) 16°C and (B) 20°C. Phenotypic diversity (D₂) derived from the flow cytometric fingerprinting in the function of time for each loop is shown in (C) and (D) for 16°C and 20°C, respectively. Day 0 represents the start of the experiment, where loop 1 underwent a flush with chlorine, loop 2 received no flushing, and loop 3 was flushed without chlorine. This flushing experiment was followed by an invasion experiment. Per timepoint, biological replicates (n = 3) were taken, and corresponding error bars are shown in black.

Fig 2

Bacterial community composition in the bulk water samples. Relative abundances of the 20 most abundant families of the bulk water before the flush, during the flush, after the flush, and after 4 and 7 days. At each timepoint, one water sample was taken per loop (n  =  1).

Fig 3

Bacterial community composition in the biofilm. Relative abundances of the 18 most abundant families of the biofilm in each loop before the flush, before the invasion, and at the end of the experiment (day 7). Three biological replicates per timepoint were taken during the first experiment (16°C), while the data from the second experiment (20°C) correspond to two biological replicates each. A flush with and without chlorination or no flush had no significant influence on the biofilm community (P > 0.05, PERMANOVA).

Fig 4

(A) Non-metric multidimensional scaling (NMDS) analysis of bacterial communities in bulk and biofilm samples. NMDS analysis of the relative bacterial community composition (16S rRNA gene) based on Bray-Curtis dissimilarities at amplicon sequence variants levels of bulk (▲) and biofilm (●) samples of loop 1 (flush with chlorine, orange), loop 2 (not flush, green), loop 3 (flush, blue), tap water fed to the pilot (rose), and tap water after the garden hose (red). Timepoints are indicated above each shape. (B) Black lines divide the samples into three groups from right to left: the bulk samples of water fed to the pilot, the bulk samples in the pilot, and the biofilm samples.

Fig 5

Concentration and decay rates of the bacterial indicators. The average concentration in the function of time (hours) for each bacterial indicator at each temperature scenario in loop 1 (orange), loop 2 (green), and loop 3 (blue). The average concentration was calculated as $C_t/C_0$ where $C_t$ represents the average concentration at time t, and $C_0$ the average concentration at the initial timepoint (t = 0). Per timepoint, biological replicates (n = 3) were taken, and corresponding error bars are shown in black. First-order decay rate constants (k [h⁻¹]) were calculated using Equation 1. The model predictions, along with their respective R² values, are depicted for each loop with corresponding colors, represented by dotted lines.