Rehabilitating the Walkerton drinking water distribution system

By Garry Palmateer, and Dr. Michele Van Dyke, GAP EnviroMicrobial Services

The objectives of this high profile project in Walkerton, Ontario, were to examine possible sources of contamination of the municipal wells and the distribution system, to ensure all sources of contamination were eliminated and to ensure that Campylobacter jejuni and Escherichia coli O157:H7 were removed from both the wells and the distribution system.

It was known that three drilled wells provided water for the town. Well #5 is 15 metres deep in limestone bedrock and has a capacity to supply 1,771 m³/day. Shortly after the investigation began, Well #5 was considered to have been the major source of contamination of the distribution system. This was based on a number of factors:

• It is a shallow well located at the southwest perimeter of town, near a farm with cattle.
• The aquifer and/or well were strongly suspected to be affected by surface water.
• The presence in the well of coliforms and E. coli was detected in late May and early June, and continued in large volume (1 – 5 litre) samples.

Well #6 is 72.2 metres deep and has a capacity to supply 1,446 m³/day. This well was considered a possible minor source of contamination late in the investigation. The aquifer was found to be connected to surface ponds near Well #6. Well #7 is 76.2 metres deep in limestone bedrock and has the capacity to supply 4,390 m³/day.

An investigation of Well #5 by hydrogeologists revealed two major facts:

• The majority of water entered the well at 5.5 to 6.4 metres, immediately below the casing.
• Approximately 15 – 20 metres from the well, water was found exiting from the ground. It was found that when the well pump operated, water exiting the ground reversed flow and re-entered the aquifer.

Surface contamination, including water and soil, was observed in Well #5 by means of a video camera situated in the well. It was concluded early in the investigation that the distribution system was affected by soil and surface water contamination.

It was speculated that during rainfall events, ponding occurred on the soil surface. Because of the limited amount of soil covering the limestone bedrock, there were many opportunities for surface contaminants, such as microbial pathogens or pesticides, to enter underground aquifers through the fissured limestone. When Well #5 was pumping, surface water containing soil microbes and nutrients was drawn into the distribution system.

Potential urban sources of contamination were also investigated including intrusion of sewage or stormwater into watermains, watermain repairs performed improperly and the improper digestion of sludge at the sewage plant. None of these potential conditions were found to have contaminated the water supply. Moreover, there were some abandoned wells which remained connected to the municipal system. There were also private wells directly connected to the municipal system.

For example, one shallow dug well that was directly connected, had a pump capable of producing water pressures significantly greater than the municipal main pressure. The surface of another well contained hundreds of earwigs and the water contained 40,000 coliforms per 100 mL and few E. coli.

The investigation revealed many surprises. Cisterns, some in basements, which were used as a source of soft water, were commonly found to be connected to the municipal water supply. Such cistern water has the potential of being highly contaminated by bacteria, viruses, and parasites.

Swab dispenser attached to fire hydrant.

Distribution System Investigation

GAP conducted microbial testing and consulting for OCWA, including:

• Tests to determine the source of contamination through environmental, distribution system, raw water and well zone samples.
• Consulting work for disinfection of the distribution system, including provision of disinfection protocols from mains and service lines, testing distribution system samples of water and biofilm, and conducting a biofilm study.

The Walkerton water distribution system was comprised of approximately 40 km of water mains. Watermain materials were primarily iron and PVC, with lesser amounts of polyethylene and copper. Some water services were galvanized iron and lead. Most pipes contained high amounts of biofilm. Rehabilitation of the Walkerton distribution system was conducted to ensure the complete removal of microbial pathogens, includng E. coli O157:H7 and Campylobacter jejuni. Rehabilitation included biofilm removal by swabbing and disinfection using high levels of chlorine. In addition, dead end elimination, pipe replacement, and disconnection of cisterns and private wells were done.

In investigating the efficacy of disinfection of the distribution system, it was decided to assess levels of coliforms, E. coli and heterotrophic bacteria (HPC) at multiple sites of the system including source wells. Aerobic spore-forming bacteria, such as Bacillus species, were also assessed. These bacteria are found naturally in soil, and will be at low concentrations in groundwater. They are very resistant to chlorine, and, as such, can serve as indicators of disinfection beyond that of coliforms and E. coli. High level recovery of spore-forming bacteria can indicate contamination by surface water.

Results showed that after swabbing and chlorination, the occurrence of adverse results in the distribution system, including coliform detection and high heterotrophic bacterial counts, decreased over time. As well, the chlorine demand decreased over time, as shown by the steady rate of increase in free chlorine residual with constant chlorine dose. The chlorine-resistant spore-forming bacteria were reduced to low levels with only a few exceptions.

Coliform bacteria were isolated from Well #6 and from the distribution system late in August 2000. These coliforms were identified as Klebsiella and Enterobacter, and were resistant to elevated levels of chlorine (greater than 1.5 mg/L free chlorine). In a comparative study, Klebsiella sp. and Enterobacter sp. isolated from the Walkerton distribution system showed increased chlorine resistance when compared to E. coli.

Six inch cast iron distribution pipes. Left – not swabbed, right – swabbed.

Biofilm

A biofilm study was initiated by the Health unit, Ontario Clean Water Agency and GAP to determine if the disinfection program could kill pathogens introduced into pipes containing biofilm. A pipe-rig apparatus was constructed comprising a 60 year old, 4-inch cast iron pipe and a new 4-inch cast iron pipe. This system was built in a maintenance shed close to Well #5. Water from Well #5 was passed through the pipes.

Experimental protocol data

• Raw water flow for two weeks – pipe conditioning.
• Inoculum recirculated through pipes for two days ­ E. coli nal^r, E. coli O157:H7, C. jejuni.
• Raw water flow for one week.
• Water containing 5 mg/L chlorine for two weeks.
• Swabbing, chlorination at 200 mg/L for 24 hours.
• Water containing 2 mg/L chlorine for four weeks.
• Biofilm and pipe effluent samples collected at each stage.

Biofilm controls­ before inoculation

• Unchlorinated raw water from Well #5 for two weeks.
• Removed 8 cm² of biofilm from inside of pipes.
• No coliforms, E. coli, E. coli O157:H7 or C. jejuni.
• Heterotrophs (HPC)
  – old pipes = 7 x 10⁶ cfu/cm²
  – new pipe = 6 x 10⁵ cfu/cm²

Pipes inoculated with bacteria

• Escherichia coli nal^r (1 x 10⁵ cfu/mL)
  – nalidixic acid resistant, non-pathogenic
• Escherichia coli O157:H7 (1 x 10⁵ cfu/mL)
  – verotoxin-negative strain
• Campylobacter jejuni (2 x 10³ cfu/mL)
• Bacteria added using sterile water from Well #5.
• Concentrations decreased by two log units after two days recirculation.

Pipe effluents ­after inoculation

• All three strains present in pipe effluent after seven days of unchlorinated raw water flow (old and new pipes).
• All three strains absent in pipe effluents after three hours of water flow containing 5 mg/L chlorine.

Biofilm samples ­after inoculation

• After swabbing and superchlorination.
  – three strains not detected in new and old pipes.
• After chlorinated water (2 mg/L) for four weeks.
  – three strains not detected in new and old pipes.

Conclusions

• Experimental results show that the protocols used to disinfect the water mains in Walkerton could remove introduced E. coli and C. jejuni from pipes containing high amounts of biofilm.
• Biofilm formation can occur rapidly in pipes without disinfectant.
• Old pipes containing high amounts of biofilm require mechanical treatment to reduce total bacterial numbers.

The team was comprised of Michele Van Dyke, Andrew Scott, Dan Van Bakel, Lily Gin, and Garry Palmateer of GAP EnviroMicrobial Services; Marc Ethier, Rick Turnbull, Carl Grimstead, Trevor McTeer, Brent Zehr, Roger Mellow, and Gord Eagles of the Ontario Clean Water Agency.