Abstract
The formation of biofilms in drinking water distribution networks is a significant technical, aesthetic and hygienic problem. In this study, the effects of assimilable organic carbon, microbially available phosphorus (MAP), residual chlorine, temperature and corrosion products on the formation of biofilms were studied in two full-scale water supply systems in Finland and Latvia. Biofilm collectors consisting of polyvinyl chloride pipes were installed in several waterworks and distribution networks, which were supplied with chemically precipitated surface waters and groundwater from different sources. During a 1-year study, the biofilm density was measured by heterotrophic plate counts on R2A-agar, acridine orange direct counting and ATP-analyses. A moderate level of residual chorine decreased biofilm density, whereas an increase of MAP in water and accumulated cast iron corrosion products significantly increased biofilm density. This work confirms, in a full-scale distribution system in Finland and Latvia, our earlier in vitro finding that biofilm formation is affected by the availability of phosphorus in drinking water.
Introduction
In many industrial systems, including drinking water distribution networks, the bacterial biomass is associated with biofilm growing on the surfaces of the pipelines [13]. Biofilms reduce the aesthetic quality of distributed water, may increase the corrosion rate of pipes, and increase the survival of pathogenic microbes in drinking water distribution systems [22, 27]. Thus, if we are to find means to combat biofilms, more knowledge about the factors that influence biofilm formation is needed. These factors include microbial nutrients (e.g. the concentration of carbon, nitrogen and phosphorus) in water, temperature, pipe materials, disinfectants, bacteria in water and the hydraulic regime in drinking water supply systems [4, 5, 11, 17, 21, 31]. To restrict biofilm growth, final chlorination is most frequently applied to drinking water. However, due to the risk of elevating the levels of harmful disinfection by-products of chlorine in drinking water, other approaches to reduce biofilm formation have been considered. One such solution would be to remove the nutrients during water treatment, thus causing nutrient limitation for microbial growth [12].
In many countries, organic carbon limits microbial growth in drinking water and in drinking water biofilms [1, 10, 24]. There are several regions where the concentration of organic carbon in drinking water is high and the concentration of phosphorus low; in these waters bacterial growth is often limited by phosphorus. This kind of water occurs in Finland, Japan, Latvia and China [18, 25, 30]. Usually, efficient removal of phosphorus by chemical precipitation is the reason for phosphorus limitation [15]. It could be anticipated that in this kind of drinking water, the availability of phosphorus would also affect bacterial growth in biofilms. Therefore, phosphorus removal from drinking water could be an effective way to control biofilm formation in the distribution net. However, bacterial growth is not directly coupled to phosphorus uptake because bacteria can store phosphorus and use it when phosphorus in the water becomes depleted [26, 29]. Therefore, the relationship between phosphorus and biofilm growth is not straightforward, and it is by no means certain that phosphorus removal from water would restrict biofilm development in a drinking water distribution system.
In carbon-limiting water it has been shown that phosphorus does not affect bacterial density in biofilms [1, 3, 24]. However, in in vitro experiments with water where phosphorus was limiting microbial growth of planktonic bacteria, the addition of phosphorus increased bacterial density in the biofilms [16]. Phosphorus is also known to affect biofilm growth in some natural aquatic ecosystems [20].
In this study, we installed identical biofilm collectors within the distribution networks and in the waterworks of two cities, Kuopio in Finland and the Latvian capital Riga. During the study, drinking water quality was monitored, especially for nutrients (carbon and phosphorus), and related to the development of bacterial numbers in biofilms. The aim was to identify factors affecting formation of biofilms in full-scale drinking water processing and distribution, with special attention being paid to phosphorus.
Materials and methods
Sampling points
In Kuopio, the biofilm collectors were installed in two waterworks and at three points in the drinking water distribution system. The distribution system collectors were installed at the pumping stations, where collectors were connected to the drinking water pipeline in a toilet. One waterworks used surface water as its raw water (Surface plant). The water was first bank-filtered then chemically coagulated. The other waterworks (Recharged plant) recharged groundwater (bank filtration, dry and wet contact filters, and slow sand filtration). The collectors were installed in the waterworks before the chlorination unit because we especially wanted to study the effects of nutrients on biofilm growth, and after chlorination the formation of biofilms is controlled mainly by chlorine. In the distribution networks, one collector was in an area distributing recharged groundwater (approximately 4 km from the waterworks—Recharged net), and two collectors were installed at distribution networks distributing water from surface waterworks (5 and 10 km from the waterworks—Surface net 5 and 10 km, respectively). At Surface net 5 km there was some mixing of surface and recharged waters in distributing networks. The sampling points were chosen to represent, as well as possible, the network and two different waters produced in two different waterworks.
In Riga, the biofilm collectors were installed in three waterworks and at two points in the drinking water distribution networks. One waterworks produced drinking water from surface water (Surface plant) by coagulation-precipitation and biological treatment methods (ozonation followed by biologically active carbon filtration). The collector was installed at the plant before chlorination at the outlet from one of the filters used for the biological treatment. The second waterworks extracted groundwater from which about one-half was artificially recharged. The collector was installed in the waterworks (Groundwater plant) after chlorination (0.5–1.0 mg l−1). The third waterworks extracted water from a deep confined aquifer (Artesian plant). The collector was installed at the plant before chlorination. In the distribution networks, one collector was installed in the area supplied with a mixture of water from Groundwater plant and Artesian plant (Groundwater net). Another collector was installed in an area of the distribution networks (Surface net) supplied from the Surface plant. The biofilm collectors were connected directly to the water mains in the networks.
Biofilms
The pipe systems (biofilm collectors) used for biofilm collection consisted of 21 pieces of polyvinyl chloride (PVC)-pipe (diameter 10 mm, length 10 cm), which were connected in line with stainless steel fittings. Water flowed through the collector at a velocity of 1 l min−1 (Re=1,566, laminar flow), which was adjusted with a valve at the outlet. Before biofilm sampling, each piece of PVC-pipe was disconnected by closing both ends with ball valves. The pipes were removed with the water inside the tubes. At every sampling time, three 10-cm pieces were disconnected for analyses. Biofilms were removed by shaking with sterile 2 mm glass beads and rinsing twice with 5 ml sterile water. Before glass bead shaking, 1–2 ml water was removed from the pipes. The extraction from the 10-cm pieces, as well as the removed water, were analysed for acridine orange direct counts (AODC) [7], heterotrophic plate counts (HPC) on R2A-agar [23] and, in Kuopio, for adenosine triphosphate (ATP). When calculating the results, the amount of bacteria in drinking water was distinguished from the number of bacteria in the biofilm suspension.
For AODC, 2 ml biofilm extract was filtered on a black 0.22 μm Nuclepore membrane filter (Whatman, Clifton, N.J.) and stained with a 0.01% dilution of acridine orange. Bacteria were counted with an epifluorescence microscope (Olympus BH-2, Tokyo, Japan or Leica DMLB, Wetzlar, Germany). In heterotrophic plate count (HPC) analysis, R2A-agar plates were incubated for 7 days at 22°C before counting colony forming units (cfu). The content of ATP was determined by means of a Bio Orbit 1251 luminometer using protocols supplied by the manufacturer. The measured light output was converted to ATP concentration by using a previously determined conversion factor. ATP was analysed only from biofilms growing in Kuopio.
To determine metal concentrations in the biofilms, samples were dried and metals extracted by heating the sample in solution of nitric acid and hydrogen peroxide. Concentrations of metals were detected by flame atomic adsorption spectrometer (Perkin-Elmer 403). The values obtained were converted to micrograms per square centimetre.
Water analyses
Total organic carbon (TOC) was analysed by a high temperature combustion method with a Shimadzu 5000 TOC analyser (Kyoto, Japan). Water was acidified and purged before analysis. Assimilable organic carbon (AOC) was analysed by a modification [19] of the original method [10]. The modification included addition of inorganic nutrients to ensure that only the AOC content restricted microbial growth, i.e. AOC was measured as AOCpotential [19]. The growth of Pseudomonas fluorescens P17 (ATCC 49642) and Spirillum sp. strain NOX (ATCC 49643) in water samples was calculated to correspond to acetate equivalents. In water samples containing chlorine, residual chlorine was removed by the addition of 50 μl 0.02 M (for 100 ml) sodium thiosulphate.
Microbially available phosphorus (MAP) concentrations were analysed by a bioassay where the maximum growth of P. fluorescens P17 (ATCC 49642) in sterilised water samples is related to the phosphorus concentration [14]. Inorganic salts (except phosphorus) and sodium acetate were added to the water to ensure that growth of the test bacteria was limited solely by phosphorus. The maximum microbial cell production (cfu ml−1) was converted to a phosphorus concentration using a previously determined empirical yield factor.
HPC and total concentrations of bacteria in drinking water were analysed as described above for biofilms. Free residual chlorine was analysed in Kuopio using a Palinest Micro 1000 chlorometer (Palintest, Gateshead, UK) and total residual chlorine in Riga by the N,N-diethyl-1,4-phenylenediamine method.
Statistical analyses
Spearman rank correlations were calculated with the program SPSS 10.1 (Statsoft, Tulsa Okla.). Regression analysis was performed with the Microsoft Excel program (Microsoft, Seattle, Wash.).
Results and discussion
Water quality
The concentrations of microbial nutrients in drinking water from Kuopio and Riga differed (Table 1). In Riga, the concentration of AOC in drinking water was markedly higher than the corresponding value in Kuopio. The TOC content was also higher in all sampling points in Riga except at the outlet from the artesian plant. The high content of AOC in treated surface water is, perhaps, a result of ozonation of a source water with a high content of natural organic matter, mainly humic substances. Ozone degrades organic matter into a form that is more available to microbes [10, 9]. The reason for the high AOC content in the artesian water remains unknown. Even though the AOC content was quite high in the artesian water and in the treated surface water, only a small fraction (<10%) of it could be utilised by bacteria because of the shortage of MAP in these waters. This was proved by the increase in the AOC yield after the addition of nutrients in AOC analyses. In AOC analyses with nutrient addition, all other inorganic nutrients but no organic carbon are added to the water [19]. We compared the results analysed with and without nutrient addition and found that with nutrient addition, the test produced concentrations of AOC about ten times higher than without nutrients (results not shown). Low levels of MAP were observed in artesian water, recharged groundwater and treated surface water both in Riga and Kuopio. In Kuopio, phosphorus has been previously recognised as being the limiting nutrient for bacterial growth in drinking water [15, 16]. In Riga, the content of MAP in the outlet from the groundwater plant and in the groundwater net was high (Table 1). The high MAP content in groundwater agrees with earlier findings in untreated Finnish groundwaters [15].
Chemical and microbiological characteristics of water at the biofilm sampling points. Values are average ±SD; number of analyses is in parenthesis. T Temperature (°C), TOC total organic carbon (mg l−1), AOC potential assimilable organic carbon (μg AOC-C l−1), MAP microbially available phosphorus (μg MAP l−1), HPC heterotrophic plate counts (cfu ml−1)
| Sampling point | T | TOC | AOC | MAP | HPCa | Chlorine |
|---|---|---|---|---|---|---|
| Riga | ||||||
| Groundwater plant | 9.1±0.8 (7) | 7.3 (1) | 538±229 (9) | 3.39±2.42 (9) | 70 (8) | 0.34±0.09 (8) |
| Groundwater net | 7.4±0.7 (3) | 5.9 (1) | 664±292 (9) | 2.88±2.02 (9) | 6,900 (9) | <0.1 (6) |
| Artesian plant | 7.5±0.1 (6) | 1.0 (1) | 761±525 (10) | 0.34±0.30 (10) | 70 (10) | Before chlorination |
| Surface plant | 9.9±6.6 (11) | 8.2 (1) | 765±227 (10) | 0.17±0.17 (10) | 1,600 (9) | Before chlorination |
| Surface net | 15.4±2.2 (4) | NAb | 695±426 (9) | 0.62±0.91 (8) | 3,400 (9) | NDc |
| Kuopio | ||||||
| Recharged plant | 6.6±4.1 (6) | 1.9±0.1 (2) | 31±20 (7) | 0.56±0.66 (7) | 210 (6) | Before chlorination |
| Recharged net | 8.8±5.2 (6) | 2.4±0.8 (4) | 40±13 (9) | 0.43±0.14 (9) | 380 (11) | 0.16±0.11 (4) |
| Surface plant | 8.7±1.7 (6) | 2.9±0.3 (3) | 24±8 (8) | 0.18±0.10 (7) | 300 (8) | Before chlorination |
| Surface net 5 km | 10.7±2.8 (5) | 2.5±0.8 (7) | 39±12 (9) | 0.63±0.63 (9) | 3,300 (12) | 0.08±0.07 (5) |
| Surface net 10 km | 14.4±2.4 (6) | 2.4±0.2 (4) | 34±9 (9) | 0.58±0.71 (9) | 24,000 (12) | 0.04±0.03 (5) |
| Sampling point | T | TOC | AOC | MAP | HPCa | Chlorine |
|---|---|---|---|---|---|---|
| Riga | ||||||
| Groundwater plant | 9.1±0.8 (7) | 7.3 (1) | 538±229 (9) | 3.39±2.42 (9) | 70 (8) | 0.34±0.09 (8) |
| Groundwater net | 7.4±0.7 (3) | 5.9 (1) | 664±292 (9) | 2.88±2.02 (9) | 6,900 (9) | <0.1 (6) |
| Artesian plant | 7.5±0.1 (6) | 1.0 (1) | 761±525 (10) | 0.34±0.30 (10) | 70 (10) | Before chlorination |
| Surface plant | 9.9±6.6 (11) | 8.2 (1) | 765±227 (10) | 0.17±0.17 (10) | 1,600 (9) | Before chlorination |
| Surface net | 15.4±2.2 (4) | NAb | 695±426 (9) | 0.62±0.91 (8) | 3,400 (9) | NDc |
| Kuopio | ||||||
| Recharged plant | 6.6±4.1 (6) | 1.9±0.1 (2) | 31±20 (7) | 0.56±0.66 (7) | 210 (6) | Before chlorination |
| Recharged net | 8.8±5.2 (6) | 2.4±0.8 (4) | 40±13 (9) | 0.43±0.14 (9) | 380 (11) | 0.16±0.11 (4) |
| Surface plant | 8.7±1.7 (6) | 2.9±0.3 (3) | 24±8 (8) | 0.18±0.10 (7) | 300 (8) | Before chlorination |
| Surface net 5 km | 10.7±2.8 (5) | 2.5±0.8 (7) | 39±12 (9) | 0.63±0.63 (9) | 3,300 (12) | 0.08±0.07 (5) |
| Surface net 10 km | 14.4±2.4 (6) | 2.4±0.2 (4) | 34±9 (9) | 0.58±0.71 (9) | 24,000 (12) | 0.04±0.03 (5) |
aGeometric mean value
bNot analysed
cNot detected
Chemical and microbiological characteristics of water at the biofilm sampling points. Values are average ±SD; number of analyses is in parenthesis. T Temperature (°C), TOC total organic carbon (mg l−1), AOC potential assimilable organic carbon (μg AOC-C l−1), MAP microbially available phosphorus (μg MAP l−1), HPC heterotrophic plate counts (cfu ml−1)
| Sampling point | T | TOC | AOC | MAP | HPCa | Chlorine |
|---|---|---|---|---|---|---|
| Riga | ||||||
| Groundwater plant | 9.1±0.8 (7) | 7.3 (1) | 538±229 (9) | 3.39±2.42 (9) | 70 (8) | 0.34±0.09 (8) |
| Groundwater net | 7.4±0.7 (3) | 5.9 (1) | 664±292 (9) | 2.88±2.02 (9) | 6,900 (9) | <0.1 (6) |
| Artesian plant | 7.5±0.1 (6) | 1.0 (1) | 761±525 (10) | 0.34±0.30 (10) | 70 (10) | Before chlorination |
| Surface plant | 9.9±6.6 (11) | 8.2 (1) | 765±227 (10) | 0.17±0.17 (10) | 1,600 (9) | Before chlorination |
| Surface net | 15.4±2.2 (4) | NAb | 695±426 (9) | 0.62±0.91 (8) | 3,400 (9) | NDc |
| Kuopio | ||||||
| Recharged plant | 6.6±4.1 (6) | 1.9±0.1 (2) | 31±20 (7) | 0.56±0.66 (7) | 210 (6) | Before chlorination |
| Recharged net | 8.8±5.2 (6) | 2.4±0.8 (4) | 40±13 (9) | 0.43±0.14 (9) | 380 (11) | 0.16±0.11 (4) |
| Surface plant | 8.7±1.7 (6) | 2.9±0.3 (3) | 24±8 (8) | 0.18±0.10 (7) | 300 (8) | Before chlorination |
| Surface net 5 km | 10.7±2.8 (5) | 2.5±0.8 (7) | 39±12 (9) | 0.63±0.63 (9) | 3,300 (12) | 0.08±0.07 (5) |
| Surface net 10 km | 14.4±2.4 (6) | 2.4±0.2 (4) | 34±9 (9) | 0.58±0.71 (9) | 24,000 (12) | 0.04±0.03 (5) |
| Sampling point | T | TOC | AOC | MAP | HPCa | Chlorine |
|---|---|---|---|---|---|---|
| Riga | ||||||
| Groundwater plant | 9.1±0.8 (7) | 7.3 (1) | 538±229 (9) | 3.39±2.42 (9) | 70 (8) | 0.34±0.09 (8) |
| Groundwater net | 7.4±0.7 (3) | 5.9 (1) | 664±292 (9) | 2.88±2.02 (9) | 6,900 (9) | <0.1 (6) |
| Artesian plant | 7.5±0.1 (6) | 1.0 (1) | 761±525 (10) | 0.34±0.30 (10) | 70 (10) | Before chlorination |
| Surface plant | 9.9±6.6 (11) | 8.2 (1) | 765±227 (10) | 0.17±0.17 (10) | 1,600 (9) | Before chlorination |
| Surface net | 15.4±2.2 (4) | NAb | 695±426 (9) | 0.62±0.91 (8) | 3,400 (9) | NDc |
| Kuopio | ||||||
| Recharged plant | 6.6±4.1 (6) | 1.9±0.1 (2) | 31±20 (7) | 0.56±0.66 (7) | 210 (6) | Before chlorination |
| Recharged net | 8.8±5.2 (6) | 2.4±0.8 (4) | 40±13 (9) | 0.43±0.14 (9) | 380 (11) | 0.16±0.11 (4) |
| Surface plant | 8.7±1.7 (6) | 2.9±0.3 (3) | 24±8 (8) | 0.18±0.10 (7) | 300 (8) | Before chlorination |
| Surface net 5 km | 10.7±2.8 (5) | 2.5±0.8 (7) | 39±12 (9) | 0.63±0.63 (9) | 3,300 (12) | 0.08±0.07 (5) |
| Surface net 10 km | 14.4±2.4 (6) | 2.4±0.2 (4) | 34±9 (9) | 0.58±0.71 (9) | 24,000 (12) | 0.04±0.03 (5) |
aGeometric mean value
bNot analysed
cNot detected
Drinking water temperature varied between 6.6 and 15.4°C and was highest in networks distributing surface water (Table 1), especially during summertime.
At all sampling sites, except in the groundwater plant in Riga, the total residual chlorine level was lower than 0.3 mg l−1. The highest number of heterotrophic bacteria (by HPC) was found in the Kuopio surface network, even though the AOC and MAP concentrations were relatively low (Table 1). The reason was probably the high temperature and long residence time of the water (3–4 days).
Biofilm formation
Biofilm formation on polyvinyl chloride (PVC) pipes in a Kuopio and b Riga, determined as numbers of heterotrophic microbes (cfu) per square centimetre. Error bars SD (n=3)
Biofilm formation in a Kuopio and b Riga on PVC pipes, determined as numbers of total bacteria per square centimetre. Error bars SD (n=3)
Biofilm formation in Kuopio on PVC pipes determined as ATP content in biofilms per square centimetre. Error bars SD (n=3)
Although the MAP level was relatively low (Table 1), the bacterial number in the biofilm of the distributed surface water in Riga was very high; in fact it was higher than that in biofilms of distributed water in Kuopio (Figs. 1, 2). This could be explained by the high water temperature over the study period and especially by the poor condition of the pipe where the biofilm collector was installed. The pipe was severely corroded and thus released corrosion products into the PVC collector. The high iron content was confirmed by metal analyses of the deposits (Table 2). Corrosion products increase the number of attachment sites for bacteria and decrease the disinfection efficiency, which results in more biofilm. Corroded pipes are known to adsorb organic matter from water and to make it more available for bacteria [6]. This causes an increase in the biofilm density in carbon-limiting waters. In the data from both cities, the concentration of iron correlated with HPC in biofilms with age over 100 days (r=0.73, P=0.016). Previous studies have shown that phosphate anions can be adsorbed on iron oxyhydroxide [2]. Thus we propose that corrosion products increase biofilm growth also in phosphorus-limiting waters because they may accumulate phosphorus from water needed for biofilm bacteria growth. Hence, corrosion should be eliminated in order to decrease biofilm growth both in carbon- and phosphorus-limiting drinking waters.
Average concentration (μg cm−2) of Ca, Mg, Mn, and some heavy metals in biofilms of Kuopio and Riga at the end of the experiment
| Ca | Mg | Fe | Mn | Zn | Pb | Cu | |
|---|---|---|---|---|---|---|---|
| Riga | |||||||
| Groundwater plant | 39.89 | 12.50 | 277.43 | 5.72 | 19.05 | 7.38 | 4.29 |
| Groundwater net | 10.72 | 7.26 | 122.79 | 28.28 | 3.13 | 2.01 | 2.38 |
| Artesian plant | 13.84 | 10.94 | 4.84 | 0.04 | 6.18 | 0.36 | 20.84 |
| Surface plant | 16.79 | 7.14 | 3.93 | 0.38 | 1.31 | 0.12 | 0.81 |
| Surface net | 18.46 | 8.33 | 402.46 | 0.24 | 5.12 | 0.94 | 3.93 |
| Kuopio | |||||||
| Recharged plant | 5.06 | 4.04 | 7.22 | 0.02 | 1.26 | 0.23 | 8.78 |
| Recharged net | 6.21 | 4.73 | 19.17 | 0.57 | 2.18 | 0.83 | 2.83 |
| Surface plant | 1.82 | 1.24 | 8.51 | 0.52 | 0.75 | 0.24 | 3.34 |
| Surface net 5 km | 5.82 | 4.04 | 10.19 | 0.13 | 0.29 | 0.16 | 1.30 |
| Surface net 10 km | 7.57 | 1.83 | 5.88 | 1.80 | 1.08 | 0.24 | 1.47 |
| Ca | Mg | Fe | Mn | Zn | Pb | Cu | |
|---|---|---|---|---|---|---|---|
| Riga | |||||||
| Groundwater plant | 39.89 | 12.50 | 277.43 | 5.72 | 19.05 | 7.38 | 4.29 |
| Groundwater net | 10.72 | 7.26 | 122.79 | 28.28 | 3.13 | 2.01 | 2.38 |
| Artesian plant | 13.84 | 10.94 | 4.84 | 0.04 | 6.18 | 0.36 | 20.84 |
| Surface plant | 16.79 | 7.14 | 3.93 | 0.38 | 1.31 | 0.12 | 0.81 |
| Surface net | 18.46 | 8.33 | 402.46 | 0.24 | 5.12 | 0.94 | 3.93 |
| Kuopio | |||||||
| Recharged plant | 5.06 | 4.04 | 7.22 | 0.02 | 1.26 | 0.23 | 8.78 |
| Recharged net | 6.21 | 4.73 | 19.17 | 0.57 | 2.18 | 0.83 | 2.83 |
| Surface plant | 1.82 | 1.24 | 8.51 | 0.52 | 0.75 | 0.24 | 3.34 |
| Surface net 5 km | 5.82 | 4.04 | 10.19 | 0.13 | 0.29 | 0.16 | 1.30 |
| Surface net 10 km | 7.57 | 1.83 | 5.88 | 1.80 | 1.08 | 0.24 | 1.47 |
Average concentration (μg cm−2) of Ca, Mg, Mn, and some heavy metals in biofilms of Kuopio and Riga at the end of the experiment
| Ca | Mg | Fe | Mn | Zn | Pb | Cu | |
|---|---|---|---|---|---|---|---|
| Riga | |||||||
| Groundwater plant | 39.89 | 12.50 | 277.43 | 5.72 | 19.05 | 7.38 | 4.29 |
| Groundwater net | 10.72 | 7.26 | 122.79 | 28.28 | 3.13 | 2.01 | 2.38 |
| Artesian plant | 13.84 | 10.94 | 4.84 | 0.04 | 6.18 | 0.36 | 20.84 |
| Surface plant | 16.79 | 7.14 | 3.93 | 0.38 | 1.31 | 0.12 | 0.81 |
| Surface net | 18.46 | 8.33 | 402.46 | 0.24 | 5.12 | 0.94 | 3.93 |
| Kuopio | |||||||
| Recharged plant | 5.06 | 4.04 | 7.22 | 0.02 | 1.26 | 0.23 | 8.78 |
| Recharged net | 6.21 | 4.73 | 19.17 | 0.57 | 2.18 | 0.83 | 2.83 |
| Surface plant | 1.82 | 1.24 | 8.51 | 0.52 | 0.75 | 0.24 | 3.34 |
| Surface net 5 km | 5.82 | 4.04 | 10.19 | 0.13 | 0.29 | 0.16 | 1.30 |
| Surface net 10 km | 7.57 | 1.83 | 5.88 | 1.80 | 1.08 | 0.24 | 1.47 |
| Ca | Mg | Fe | Mn | Zn | Pb | Cu | |
|---|---|---|---|---|---|---|---|
| Riga | |||||||
| Groundwater plant | 39.89 | 12.50 | 277.43 | 5.72 | 19.05 | 7.38 | 4.29 |
| Groundwater net | 10.72 | 7.26 | 122.79 | 28.28 | 3.13 | 2.01 | 2.38 |
| Artesian plant | 13.84 | 10.94 | 4.84 | 0.04 | 6.18 | 0.36 | 20.84 |
| Surface plant | 16.79 | 7.14 | 3.93 | 0.38 | 1.31 | 0.12 | 0.81 |
| Surface net | 18.46 | 8.33 | 402.46 | 0.24 | 5.12 | 0.94 | 3.93 |
| Kuopio | |||||||
| Recharged plant | 5.06 | 4.04 | 7.22 | 0.02 | 1.26 | 0.23 | 8.78 |
| Recharged net | 6.21 | 4.73 | 19.17 | 0.57 | 2.18 | 0.83 | 2.83 |
| Surface plant | 1.82 | 1.24 | 8.51 | 0.52 | 0.75 | 0.24 | 3.34 |
| Surface net 5 km | 5.82 | 4.04 | 10.19 | 0.13 | 0.29 | 0.16 | 1.30 |
| Surface net 10 km | 7.57 | 1.83 | 5.88 | 1.80 | 1.08 | 0.24 | 1.47 |
Other metals may also affect the growth of biofilms—copper especially is known to be toxic for several bacterial species (e.g. Legionella) [8]. Manganese and copper may also inhibit microbial growth in biofilms [32]. However, bacteria (P. aeruginosa) in biofilms are still more resistant to copper than planktonic bacteria in water [28]. In this study, the concentrations of metals in biofilms were usually higher in Riga than in Kuopio (Table 2). This is probably attributable to the more corroded condition of the pipes in the distribution networks and the higher content of iron in the groundwater. In Kuopio, only in the case of copper were concentrations were slightly higher than in Riga. The copper probably originates from copper pipes that were fitted before the biofilm collectors. There were no significant correlations between the concentration of any metal other than iron and biofilm growth.
We calculated the Spearman rank correlations for biofilms over 100 days old for both cities. HPC in water correlated with HPC in biofilms (r=0.60, P=0.002). The content of ATP in biofilms correlated with HPC in biofilms (r=0.52, P=0.022) and temperature correlated with HPC (r=0.60, P=0.005) and ATP in biofilms (r=0.62, P=0.017). Temperature also correlated with HPC in drinking water (r=0.70, P=0.001). However, there was no correlation with HPC in water and microbial concentration in biofilms aged under 50 days, i.e. the concentration of bacteria affected the steady state level of biofilms but not the biofilm formation rate.
Relationship between heterotrophic bacterial number in biofilm (cfu cm−2) and microbially available phosphorus (MAP) in the water supply systems of Kuopio (filled squares) and Riga (filled circles) 100 days after the start of the experiment. Data points are mean values; bars ±SE
In conclusion, this study confirmed our earlier finding that biofilm formation in drinking water distribution networks is controlled by many factors, including temperature, corrosion products and the limiting nutrient in the water. In this study we show for the first time in two independent large-scale systems that a high content of phosphorus in the water may enhance the formation of biofilms within distribution networks. This knowledge can be used by the water industry to find solutions on how best to control biofilm formation within their drinking water distribution networks.
Acknowledgements
This study was supported by the Academy of Finland (project 52459) and the Ulla Tuominen foundation. We acknowledge the staff of Kuopio Water and staff of Baltezers, Zakumuiza and Daugava waterworks of Municipal enterprise Riga Water for assisting in installation and sampling of biofilm collectors.
