Elsevier

Water Research

Volume 36, Issue 18, November 2002, Pages 4477-4486
Water Research

Bulk water phase and biofilm growth in drinking water at low nutrient conditions

https://doi.org/10.1016/S0043-1354(02)00191-4Get rights and content

Abstract

In this study, the bacterial growth dynamics of a drinking water distribution system at low nutrient conditions was studied in order to determine bacterial growth rates by a range of methods, and to compare growth rates in the bulk water phase and the biofilm. A model distribution system was used to quantify the effect of retention times at hydraulic conditions similar to those in drinking water distribution networks. Water and pipe wall samples were taken and examined during the experiment. The pipes had been exposed to drinking water at approximately 13°C, for at least 385 days to allow the formation of a mature quasi-stationary biofilm. At retention times of 12 h, total bacterial counts increased equivalent to a net bacterial growth rate of 0.048 day−1. The bulk water phase bacteria exhibited a higher activity than the biofilm bacteria in terms of culturability, cell-specific ATP content, and cell-specific leucine incorporation rate. Bacteria in the bulk water phase incubated without the presence of biofilm exhibited a bacterial growth rate of 0.30 day−1. The biofilm was radioactively labelled by the addition of 14C-benzoic acid. Subsequently, a biofilm detachment rate of 0.013 day−1 was determined by measuring the release of 14C-labelled bacteria of the biofilm. For the quasi-stationary phase biofilm, the detachment rate was equivalent to the net growth rate. The growth rates determined in this study by different independent experimental approaches were comparable and within the range of values reported in the literature.

Introduction

Numerous studies have shown that bacterial growth within drinking water distribution networks can seriously affect the hygienic and aesthetic quality of drinking water, which has been studied in pilot scale distribution systems [1], [2], [3]. In order to predict the water quality changes, the development of conceptual models and the determination of kinetic constants is essential.

The classical approach to estimate biofilm growth in drinking water distribution systems is to measure the biofilm formation rate [4], [17], [18]. However, the properties of the biofilm as well as the bacterial activity will change during the formation of the biofilm, which means that the growth rate of the mature biofilm may be very different from the initial colonisation rate. Therefore, environmentally realistic growth rates can only be determined when the biofilm has matured. The measurement of biofilm formation is laborious because frequent sampling is necessary throughout the colonisation period. An easier approach to determine biofilm growth rates might be to estimate the release rate of bacteria from the biofilm [5], [6]. This method demands model systems with mature biofilm where retention times are well defined and preferably controllable. Yet another approach could be to determine the growth rate of the bacteria in terms of uptake of specific compounds like e.g. leucine [7].

Biofilm formed on the inner surfaces of the drinking water pipes are generally believed to be responsible for the deterioration of the drinking water quality, during the transport from waterworks to the consumer. This view is supported by the fact that the biomass on the surfaces is normally larger than the biomass in bulk water phase, and that the bulk water phase bacteria will be subject to a rapid washout due to the plug flow conditions of the pipes. However, large metabolic differences between attached and free-living bacteria can occur [8] which emphasises the importance of studying the bacteria in both habitats.

The purpose of this study was to quantify bacterial growth rates by different approaches in a model drinking water distribution system at low nutrient conditions and to compare the properties of the biofilm and the bulk water phase bacteria in order to determine differences in activity.

Section snippets

Test system

The experiments were performed in a model drinking water distribution system consisting of two loops made of square stainless-steel pipes (grade 316) where the hydraulic conditions in terms of retention time and flow rate were controlled (Fig. 1). The two loops were identical and connected in series and hydraulically they behaved like two mixed reactors. Each loop contained 70 biofilm sampling points where a removable stainless-steel test plug was mounted even with the inner surface of the

Water quality changes during batch mode

During the first batch experiment, water samples were taken regularly and analysed by AODC (Fig. 2). The bacterial numbers increased steadily by 5.7×104 cells/mL during the first 12 h, while the increase only constituted 0.5×104 cells/mL during the following 12 h. The initial AOC concentration was 4.6 μg ac-C/L, which decreased to 0.5 μg ac-C/L after 24 h of recycling.

Bulk phase

In the second batch experiment, samples were taken before and after 12 h of recycling (Table 3). During the recycling period, the AOC

Net microbial production

The initial AOC content in the drinking water of the two batch mode experiments (4.6 and 5.3 μg ac-C/L) was below the 10 μg ac-C/L proposed by Van der Kooij [12] as a practical limit for biologically stable water. Nonetheless, a significant bacterial growth was observed in both experiments. The bacterial numbers in the bulk phase increased by 5.7×104 and 5.0×104 cells/mL and AOC was depleted after 12 h in the two separate batch experiments. During the batch experiments, the bulk phase ATP

Conclusion

Despite the low amount of AOC (<5 μg ac-C/L) present, a significant growth was observed. The growth rates determined by different independent experimental approaches were comparable and within the range of values reported in the literature.

The growth properties of bulk phase and biofilm bacteria were different. The culturability, the cell-specific ATP content and cell-specific leucine incorporation was generally higher for the bulk phase than for the biofilm. This was confirmed by measurements of

Acknowledgements

This study was partly financed by the Danish Academy of the Technical Sciences. The COWI foundation has funded the experimental model distribution system. The authors wish to thank Søren Lind for his constructive comments and Torben Dolin, who made the drawings for this article.

References (19)

There are more references available in the full text version of this article.

Cited by (0)

View full text