EPA/600/R-20/116 | May 2020
www.epa.gov/homeland-security-research
United States
Environmental Protection
Agency
&EPA
Decontamination of Drinking
Water Distribution System
Infrastructure after
Contamination with
Untreated Source Water
Office of Research and Development
Homeland Security Research Program

-------
Disclaimer
The U.S. Environmental Protection Agency (EPA) through its Office of Research and
Development funded and managed the research described herein under Contract
68HERC19D0009, Task Order 68HERC19F0172 with Aptim and Interagency Agreement DW-
89-92381801 with the Department of Energy. It has been subjected to the Agency's review and
has been approved for publication. Note that approval does not signify that the contents
necessarily reflect the views of the Agency. Any mention of trade names, products, or services
does not imply an endorsement by the U.S. Government or EPA. The EPA does not endorse any
commercial products, services, or enterprises.
11

-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The Center for Environmental Solutions and Emergency Response (CESER) within the Office of
Research and Development (ORD) conducts applied, stakeholder-driven research and provides
responsive technical support to help solve the Nation's environmental challenges. The Center's
research focuses on innovative approaches to address environmental challenges associated with
the built environment. We develop technologies and decision-support tools to help safeguard
public water systems and groundwater, guide sustainable materials management, remediate sites
from traditional contamination sources and emerging environmental stressors, and address
potential threats from terrorism and natural disasters. CESER collaborates with both public and
private sector partners to foster technologies that improve the effectiveness and reduce the cost
of compliance, while anticipating emerging problems. We provide technical support to EPA
regions and programs, states, tribal nations, and federal partners, and serve as the interagency
liaison for EPA in homeland security research and technology. The Center is a leader in
providing scientific solutions to protect human health and the environment.
Gregory Sayles, Director
Center for Environmental Solutions and Emergency Response
111

-------
Table of Contents
Disclaimer	ii
Foreword	iii
List of Figures	iv
List of Tables	v
Abbreviations and Acronyms	vi
Acknowledgments	vii
Executive Summary	viii
1.0 Introduction	1
1.1	Background	1
1.2	WSTB System Description	1
2.0 Description of Untreated Water Contamination and Decontamination Experiments	4
2.1	Pipe Materials Tested	5
2.2	Source Waters Used	6
2.3	Contamination and Decontamination Procedures	7
2.3.1	Distribution System Pipe (450 ft)	7
2.3.2	Premise Plumbing System	9
2.3.3	Individual Pipe Sections	11
2.4	Experimental Methods	12
2.5	Quality Control and Data Quality	13
2.5.1	Quality Control	13
2.5.2	Data Quality	15
2.5.3	Deviations	15
3.0 Experimental Results	15
3.1	Decontamination of the 450-ft 8-in Diameter Distribution Pipe	15
3.2	Decontamination of the Premise Plumbing System	18
3.3	Decontamination of individual corroded iron pipe sections	22
4.0 Conclusions	24
5.0 References	26
List of Figures
Figure 1: Schematic overview of the Water Security Test Bed (WSTB)	2
Figure 2: Aerial view of the Water Security Test Bed (WSTB)	2
Figure 3: Water Security Test Bed discharge lagoon	3
Figure 4: Individual pipe sections next to the Water Security Test Bed lagoon (left) and the same
iv

-------
pipe sections with the open drainage valves shown (right)	3
Figure 5: Premise plumbing setup at the WSTB: the water meter outside the building (top left),
premise plumbing pipes (top middle), water heater (top right), dishwasher, washing machine and
refrigerator (bottom left), utility sink (bottom middle) and exterior tank (bottom right)	4
Figure 6: External and internal view of the cement-mortar lined pipe	5
Figure 7: External (left) and internal (right) view of the cast iron pipe with heavy corrosion
(tuberculation) on the interior	5
Figure 8: Mixing E. coli into lagoon water (left) and a carboy of Potomac River water (right).... 6
Figure 9: The setup used to contaminate the WSTB pipe with lagoon water: hose connected to
pump in lagoon (top left), hose from lagoon across WSTB site (top right), disconnection of fire
hose (bottom left), and connection of lagoon hose (bottom left)	8
Figure 10: Flushing contaminated water from the WSTB distribution pipe via the downstream
fire hydrant. The left image is the fire hydrant with the hose connected, and the is the hose in the
lagoon with sandbags is on the right	8
Figure 11: Example of how chlorine was introduced to WSTB distribution pipe	9
Figure 12: Flow meters downstream from the utility sink used to control flow through the
premise plumbing system	10
Figure 13: Contamination and decontamination of coli forms/A", coli in the bulk water phase of
the 450 ft cement-mortar lined iron distribution pipe	16
Figure 14: Contamination and decontamination of coli forms/A", coli from the pipe surface of the
450-ft cement-mortar-lined iron distribution pipe	17
Figure 15: Conductivity and turbidity measurements during each stage of the decontamination
experiment	17
Figure 16: TOC and pH during each stage of the decontamination experiment	18
Figure 17: Total coliform levels in the bulk water phase of the pipes and appliances in the home
plumbing system	19
Figure 18: E. coli levels in the bulk water phase of the pipes and appliances in the home
plumbing system	19
Figure 19: TOC levels in the bulk water phase of the pipes and appliances in the home plumbing
system	20
Figure 20: pH levels at the hot and cold taps in the home plumbing system	21
Figure 21: Turbidity levels at the hot and cold taps in the home plumbing system	21
Figure 22: Conductivity levels at the hot and cold taps in the home plumbing system	22
Figure 23: Total coliform levels in the bulk water phase of the individual short pipe sections... 23
Figure 24: Total coliform levels on the inner pipe wall of the individual short pipe sections	24
List of Tables
Table 1: Quality control data quality objectives	14
v

-------
Abbreviations and Acronyms
atm	atmosphere(s)
AWWA	American Water Works Association
BWS	bulk water sample
CCC	continuing calibration check
CFU	colony forming unit(s)
cm	centimeter(s)
DC	District of Columbia
DPD	N,N-di ethyl -p-pheny 1 enedi amine
E. coli	Escherichia coli
EPA	Environmental Protection Agency
ft	foot/feet
gpm	gallon(s) per minute
h	hour(s)
HOC1	hypochlorous acid
HSRP	Homeland Security Research Program
in2	square inch
INL	Idaho National Laboratory
L	liter(s)
m	meter(s)
mg	milligram(s)
min	minute(s)
mL	milliliter(s)
MPN	most probable number
NTU	nephelometric turbidity units
PEX	cross-lined polyethylene
PPD	presidential policy directive
ppm	parts per million
psi	pound(s) per square inch
PVC	polyvinyl chloride
QC	quality control
QCS	quality control sample
TOC	total organic carbon
WSTB	Water Security Test Bed
vi

-------
Acknowledgments
Contributions from the following individuals to the field work described in this report are
acknowledged: Stephen Reese and Travis McLing of the Idaho National Laboratory.
vii

-------
Executive Summary
The U.S. Environmental Protection Agency's (EPA's) Homeland Security Research Program
partnered with the Idaho National Laboratory (INL) to build the Water Security Test Bed
(WSTB) at the INL test site outside Idaho Falls, Idaho. The WSTB was built using an 8-inch
diameter cement-mortar-lined ductile iron drinking water distribution pipe that had previously
been taken out of service. The pipe was exhumed from the INL grounds and oriented in the
shape of a small (450 feet long) drinking water distribution system. The WSTB can support
drinking water distribution system research on a variety of topics, including biofilms, water
quality, sensors, and homeland security-related contaminants. Since the WSTB is constructed of
real drinking water distribution system pipes, research can be conducted under conditions that
are representative of the conditions in a drinking water distribution system (USEPA, 2016;
USEPA, 2018).
This report summarizes the results of infrastructure decontamination experiments performed at
the WSTB. These experiments focused on simulating contamination of drinking water
distribution pipes with untreated source water due to a treatment failure at a water treatment
facility. A loss of water treatment capability due to an emergency (e.g., power loss) could force
a utility to pump untreated source water into the distribution system to maintain basic sanitation
(such as toilet flushing), fire protection, and to maintain pressure in the pipes. Compared to
treated drinking water, untreated source water will likely have higher levels of various water
quality parameters such as turbidity, conductivity, pH and organic carbon, and include increased
levels of microbial contamination such as coliform bacteria (including Escherichia coli (E. coli)).
Should an event like this occur, results from this study can help water utilities understand the
effectiveness of common distribution system decontamination methods such as flushing and
chlorination and decrease the time their systems are offline.
To assess infrastructure decontamination, common processes used by water utilities to clean their
pipes were implemented. First, a fire hydrant attached to the 450-ft WSTB distribution pipe was
opened, and the water was flushed at approximately 150 gpm (0.96 ft/sec) for 20 minutes (min).
The pipe was then chlorinated at approximately 55 mg/L, with a contact time of 24 hours (h).
After chlorination, the water was flushed through the same hydrant, and uncontaminated local
chlorinated tap water was allowed to flow through the pipes. Bulk water samples and samples of
the pipe interior were taken during each phase of the experiment to determine if contamination
was removed from the distribution pipe.
In addition, a premise plumbing system, which included a hot water heater, refrigerator,
dishwasher and washing machine, was contaminated with the untreated water. The system was
flushed by opening taps or running the appliance repeatedly. When the 8-inch diameter
distribution pipe was chlorinated, this chlorine was allowed to flow into the plumbing and
appliances. Preliminary research was also performed with sections of heavily tuberculated
unlined cast iron pipe obtained from a drinking water utility. These iron pipe sections were
contaminated and decontaminated by filling and draining them, with stagnant water sitting inside
the pipe during each phase of the experiment. The goal was to collect preliminary data with
heavily tuberculated iron pipe, which may be harder to decontaminate than smoother cement-
lined ductile iron pipe.

-------
These experiments were designed to provide water utility operators with realistic expectations of
how effective standard decontamination processes like flushing and chlorination would be for
returning the distribution system to service. The following is a summary of the results:
•	After contamination of the 450-ft distribution pipe, an increase in conforms/A", coli was
observed, as well as increases in Total Organic Carbon (TOC), conductivity, pH and
turbidity. After flushing water through the fire hydrant at approximately 150 gallons per
minute (gpm), no conforms/A", coli were observed in the bulk water or on the pipe
surface. TOC, conductivity and turbidity returned to near-baseline levels (pH was more
variable). Baseline levels were maintained during chlorination and after tap water was
returned to the pipe.
•	Contamination from the WSTB distribution pipe entered the premise plumbing system
with appliances, which produced an increase in coliforins/A. coli, TOC, conductivity, pH
and turbidity. Flushing the plumbing pipes, emptying the water tank and running the
appliances for a cycle reduced coliforins/A. coli to undetectable levels and returned
turbidity and conductivity to baseline. TOC also returned to baseline after flushing
except for the dishwasher, where increased levels persisted after flushing.
•	Individual sections of heavily tuberculated iron pipe were filled with contaminated water,
emptied and rinsed (to simulate flushing), then chlorinated and rinsed again. After
rinsing (flushing), coli forms/A. coli were detected on the inner surface of the corroded
iron pipe and in the bulk water. None were detected after chlorination, but coliforms
were detected in the bulk water after the chlorine was flushed out.
In summary, flushing uncontaminated tap water for 20 min at 150 gpm (0.96 ft/sec) was
effective at removing coliform bacteria from the 450 ft cement-mortar-lined pipe. Flushing was
also effective for home plumbing and appliances. In general, water quality parameters such as
TOC, conductivity and turbidity returned to pre-contamination baseline levels, suggesting that
monitoring water quality might be an effective method of monitoring the progress of
decontamination. Coliforms were present after flushing in sections of heavily tuberculated iron
pipe, and they appeared in the bulk water after chlorination and a second round of flushing.
These data suggest that heavily tuberculated iron may be more difficult to decontaminate via
flushing and chlorination than the cement-mortar-lined iron used in other experiments. Also,
water quality parameters such as TOC, conductivity and turbidity returned to pre-contamination
baseline levels in most cases, suggesting that monitoring water quality might be an effective
method of determining the progress of decontamination.
IX

-------
1.0 Introduction
1.1	Background
The U.S. Environmental Protection Agency's (EPA's) Homeland Security Research Program
(HSRP) has partnered with the Idaho National Laboratory (INL) to build the Water Security Test
Bed (WSTB) at the INL near Idaho Falls, Idaho. The centerpiece of the WSTB is an 8-inch
diameter cement-mortar-lined ductile iron drinking water pipe that had been taken out of service.
The pipe was exhumed from the INL grounds and then oriented in the shape of a small drinking
water distribution system. The WSTB has been fitted with service connections, a premise
plumbing and appliance system, fire hydrants, and removable coupons (excised sample
materials) to collect samples from the pipe inner surface (USEPA, 2016; Szabo et al, 2017;
USEPA, 2018).
Previously, experiments focused on decontamination of various contaminants that adhered to the
inner surface of the 8-inch water pipe have been conducted at the WSTB (USEPA, 2016; Szabo
et al, 2017; USEPA, 2018). In response to a contamination event, drinking water utilities will
likely flush the distribution system using fire hydrants and possibly chlorinate as described in
American Waterworks Association (AWW A) Standard C-651-05: Disinfecting of Water Mains
(AWW A, 2005). The experiments described in this report examine different aspects of
distribution system decontamination using these methods.
These experiments focused on simulating contamination of drinking water distribution pipes
with untreated source water due to a treatment failure at a water treatment facility. A loss
treatment capability due to an emergency (e.g., power loss) could lead a utility to consider
pumping untreated source water into the distribution system to maintain basic sanitation (such as
toilet flushing), fire protection, and to maintain pressure in the pipes. Compared to treated
drinking water, untreated source water will likely have higher levels of various water quality
parameters such as turbidity, conductivity, and organic carbon and include increased levels of
microbial contamination such as coliform bacteria (including Escherichia coli [E. coli]). These
experiments examine the ability of a water utility to decontaminate water distribution pipes
contaminated with untreated source water using flushing and chlorination. This work was
extended to premise plumbing, which includes common appliances and water pipes found in
homes. The experiments were designed to provide full scale data to utility operators considering
using flushing and chlorination to return a water distribution system to service after
contamination with untreated source water.
1.2	WSTB System Description
A primary feature of the WSTB is an 8-inch (20-cm) diameter cement-mortar-lined ductile iron
drinking water pipe oriented in the shape of a small drinking water distribution system. The
WSTB contains ports for service connections and a 15-foot (ft) (5-meter [m]) removable coupon
section designed to sample the pipe interior to examine the results from
contamination/decontamination experiments on the pipe wall. Figure 1 schematically depicts the
main features of the WSTB, and Figure 2 shows an overhead view with major components
labeled.
1

-------
Figure 1: Schematic overview of the Water Security Test Bed (WSTB).
Downstream Sensor Box and Auto
Flush Hydrant	~ J
Copper Service Line
Premise Plumbing'
Pump House
mr'i"
Upstream Sensor
Box and Injection
WSTB Start
Water Security Test Bed

Figure 2: Aerial view of the Water Security Test Bed (WSTB).
2

-------
Drinking water supplied to the WSTB is chlorinated ground water that also supplies the
surrounding INL facilities. The WSTB incorporates approximately 450 ft (137 m) of 8-inch (in)
(20-centimeter [cm]) diameter cement-mortar-lined ductile iron pipe. The 8-in (20 cm) pipe
system is constructed directly over the lined drainage ditch for spill/ leak containment (as shown
in Figure 2). The total volume of the WSTB was estimated to be approximately 1,150 gallons
(4,353 liters [L]). The effluent water from the WSTB system was discharged to a lined lagoon
(Figure 3) that has a total water storage capacity of 28,000 gallons (105,980 L).
North
Figure 3: Water Security Test Bed discharge lagoon,
Three individual short sections of pipe were set up next to the lagoon. The individual pipe setup
is shown in Figure 4. One pipe section was the same cement-mortar-lined iron pipe used in the
450-ft WSTB pipe. The other two sections were unlined cast iron pipe sections with heavy
corrosion (tuberculation) on the interior. Each pipe was approximately 10 ft long. The unlined
iron pipes were obtained from the District of Columbia Water and Sewer Authority (DC Water).
All pipe surface samples taken from the individual pipes were direct scrapings of the inner
surface. Further details on the contamination, flushing, chlorination, and sampling processes are
described in Section 2.

Unlined Iron Pipe
Cement-Mortar-
Lined Iron Pipe
1
Figure 4: Individual pipe sections next to the Water Security Test Bed lagoon
(left) and the
3

-------
same pipe sections with the open drainage valves shown (right).
The WSTB also includes a premise plumbing system that is connected to the 8-in diameter
distribution main via a 1-in copper service connection (Figure 2). Water flows through the
sen-ice connection into a water meter, copper plumbing pipes, a removable pipe coupon section
with copper, polyvinyl chloride (PVC) and cross-linked polyethylene (PEX) pipes, and
appliances including a water heater, dishwasher, washing machine and refrigerator with water
dispenser. Water empties into a utility sink with hot- and cold-water taps and then drains to an
exterior tank. All water from the tank eventually flows to the lagoon. These components are
shown in Figure 5.
Figure 5: Premise plumbing setup at the WSTB: the water meter outside the building (top
left), premise plumbing pipes (top middle), water heater (top right), dishwasher, washing
machine and refrigerator (bottom left), utility sink (bottom middle) and exterior tank
(bottom right).
2.0 Description of Untreated Water Contamination and
Decontamination Experiments
The following section provides more detailed descriptions of the pipe materials used, how
untreated source water was prepared, the procedures used for forming biofilms in the pipes,
4

-------
contamination and decontamination.
2.1 Pipe Materials Tested
Most of the distribution system-sized pipe used at the WSTB for decontamination experiments is
constructed of 8-in diameter cement-mortar-lined ductile iron, which was previously excavated
from the INL property. This piping had been used as water distribution piping for over 30 years
prior to excavation and reuse on this project. Visual inspection of the cement-mortar lining
indicated that the lining was in good condition. The interi or and exterior view of this pipe is
shown in Figure 6.
Figure 6: External and internal view of the cement-mortar lined pipe.
Pipe in the premise plumbing system is mostly 1 -in diameter copper, with coupon sections made
of copper, PVC and PEX (Figure 5, top middle). The other piping used in this experiment was
heavily corroded and tuberculated 8-in diameter unlined iron pipe sections obtained from DC
Water. Individual sections of pipe (approximately 10 ft long) were set up next to the lagoon.
The individual pipe setup is shown in Figure 4, and internal and external pictures of the pipe are
shown in Figure 7. In Figure 4, two of the pipe sections are unlined iron, and one section is
cement-mortar-lined iron pipe.
Figure 7: External (left) and internal (right) view of the cast iron pipe with heavy corrosion
(tuberculation) on the interior.
5

-------
2.2 Source Waters Used
Two different simulated source waters were evaluated during this experiment. The primary
simulated source water used to contaminate the 450-ft distribution pipe, the premise plumbing,
and one of the short sections of corroded unlined iron pipe was referred to as "lagoon water".
This water was local tap water (chlorinated groundwater) that had traveled through the WSTB
pipes and emptied into the lagoon. In addition to this water, the lagoon contains dirt and organic
matter such as grasses that have blown in from the surrounding desert. Algae were also present
and non-pathogenic E. coli K-12 (cultured off site, see Section 2.4) was added to the lagoon. To
mix the contents of the lagoon, an individual walked around the lagoon in waders (see Figure 8).
Agitation of the water and sediment in the lagoon mixed the E. coli and created turbidity in the
water. Target water quality in the lagoon was based on the 2018 Annual Report of Water
Analysis for the Washington Aqueduct produced by the Army Corps of Engineers
(https://www.nab.usace.armv.mil/MissionsAVashington-AqueductAVater-Oualitv/. last accessed
April 7, 2020). In particular, the targets for key water quality parameters were as follows:
Figure 8: Mixing E. coli into lagoon water (left) and a carboy of Potomac River water
(right).
•	Turbidity: 20 Nephelometric Turbidity units (NTU) (range: 8-33 NTU)
•	Total Organic Carbon (TOC): 3.5 parts per million (ppm) (range: 1.2-2.2 ppm)
•	Coliforms: 45,000 most probable number (MPN)/100 milliliters (mL) (range: 5,335 to
97,250 MPN/100 mL)
The other water used in the short secti ons of pipe was collected from the Potomac River
downstream from the intake to the DC Water Dalecarlia treatment plant, and shipped by DC
Water to the INL WSTB site in a carboy (Figure 8). This source water was not spiked with E.
coli. This water was used to contaminate two of the individual pipe sections set up next to the
lagoon (one corroded iron and one cement-mortar-1 itied iron). The contamination procedure is
further described in Section 2.3.3.
6

-------
2.3 Contamination and Decontamination Procedures
Contamination and decontamination took place in the 450-ft distribution pipe, home plumbing
and short individual sections of pipe. The following sections describe how contamination and
decontamination took place.
2.3.1 Distribution System Pipe (450 ft)
Before contamination, biofilrn formation was accomplished by passing INL tap water through
the pipe continuously for four weeks. After initial flushing to remove any debris, the 450-ft
WSTB pipe was set to 2.5 gallons per minute (gpm) for the four-week biofilm formation period.
This procedure has been used to form biofilms in the pipe during previous experiments. Forming
biofilms using this procedure resulted in biofilm levels of 104 to I05 colony forming units
(CFU)/square centimeter (cm2) in past experiments (USEPA, 2016; USEPA, 2018). These
biofilm levels have been consistent between past experiments, and since the same biofilm
formation process was used in this study, biofilm levels were not measured.
Contamination of the pipe occurred by pumping the lagoon water (described in Section 2.2) from
the lagoon into the upstream end of the pipe, letting it flow through the 450 ft of pipe and
emptying back into the lagoon. Pumping lagoon water in a loop kept the pipe full of
contaminated water and kept the sediment and E. coli in the lagoon well mixed. Water was
recirculated in this manner for 18 h (hours). Figure 9 shows the setup used to contaminate the
pipe. A hose connected to a swimming pool pump was put into the lagoon (top left) and run the
length of the WSTB pipe (top right). On the upstream end, the fire hose that fed tap water to the
pipe was disconnected (bottom left), and the line that supplied the lagoon water was attached
(bottom right). The pipe was then filled with lagoon water and flow commenced. The pump
supplied approximately 10 pounds per square inch (psi) of pressure in the pipe.
7

-------
Figure 9: The setup used to contaminate the WSTB pipe with lagoon water: hose connected
to pump in lagoon (top left), hose from lagoon across WSTB site (top right), disconnection
of fire hose (bottom left), and connection of lagoon hose (bottom left).
After the 18-hour contamination period, the first decontamination step was to flush the system
through the downstream fire hydrant. Before the flushing began, the pool pump that supplied
lagoon water to the 8-inch line was disconnected and the fire hose that normally supplies tap
water to the 450 ft pipe was reattached. Once the fire hose with tap water was reconnected, the
downstream fire hydrant was used to flush the pipe.
Flushing is shown in Figure 10. A fire hose was attached to the downstream fire hydrant (Figure
10, left) with the outlet placed in the lagoon and braced with sandbags and cement blocks (Figure
10, right). The hydrant was opened as much as possible while keeping the fire hose stable and
against the ground. Flow through the hose was estimated to be approximately 150 gpm (0.96
ft/sec). Flushing in this manner took place for 20 min.
Figure 10: Flushing contaminated water from the WSTB distribution pipe via the
8

-------
downstream fire hydrant. The left image is the fire hydrant with the hose connected, and
the is the hose in the lagoon with sandbags is on the right.
After hydrant flushing was complete, the pipe was chlorinated. Chlorination took place by
pumping store-bought concentrated bleach (8.25% hypochlorous acid [HOC1]) into the WSTB
pipe. One gallon of bleach was diluted into five gallons of tap water in a carboy and then
pumped into the WSTB pipe over a period of one hour with flow through the pipe at 15 gpm.
This procedure effectively filled the pipe with chlorinated water. Figure 11 shows a past
example of how the disinfectant was added to the pipe.
Once chlorine was detected at the downstream fire hydrant, the flow through the pipe was shut
off, resulting in a chlorine concentration of 55 milligrams (mg)/L in the pipe. The water then sat
stagnant overnight for approximately 20 h. After the 20-h contact period, the fire hydrant was
opened again and flushed for 20 min at a rate of 150 gpm. After flushing, flow through the pipe
with tap water (i.e., return to service) was re-established at 2.5 gpm for the duration of the
experiment.
Figure 11: Example of how chlorine was introduced to WSTB distribution pipe.
2.3.2 Premise Plumbing System
Before contamination, tap water continuously flowed through the premise plumbing system for
four weeks to allow biofilm and deposits to form on the inner surfaces of the pipes and
appliances. During this pre-contamination phase, total flow though the plumbing system was set
to 138 gallons per day (0.096 gpm or 363 mL/min), which is the typical usage in many
households (DeOreo et al., 2016). Water flowed through the utility sink taps, with equal flows
through the hot and cold water taps. Flow was regulated by a set of flowmeters downstream
from the utility sink (Figure 12). The dishwasher and washing machine were operated for one
cycle once per week, and the refrigerator water dispenser was opened for 10 minutes once per
week.
9

-------
Figure 12: Flow meters downstream from the utility sink used to control flow through the
premise plumbing system.
Contamination of the premise plumbing system was accomplished by allowing lagoon water to
flow through the 1-in copper service connection during contamination of the 450-ft distribution
pipe. The utility sink's cold tap was opened to allow lagoon water to flow through the plumbing.
Once lagoon water reached the cold tap (determined by a spike in turbidity), the hot water tank
was emptied through the drain valve at the bottom of the tank, and then refilled with lagoon
water. After the hot water tank was filled with lagoon water, the utility sink hot water tap was
opened, and the dishwasher and washing machine were run for one cycle. The refrigerator water
dispenser was also opened during the contamination phase, but the inline filter quickly became
clogged with the particles from the lagoon water. A replacement filter was not available, so the
refrigerator water was not sampled during the remainder of the experiment.
Decontamination was performed based on the findings of the Water Research Foundation report
4572 titled "Flushing Guidance for Premise Plumbing and Service Lines to Avoid or Address a
Drinking Water Advisory" (WRF, 2016). In this report, suggestions for how to flush household
plumbing and appliances were derived from an expert panel of water industry professionals. The
specific suggestions summarized in the report are as follows (reproduced verbatim from the
report):
Flushing Cold Water Taps
•	Begin by nmning the cold-water faucet closest to where water enters the house. Starting
from the point closest to where water enters the house, open all the other cold water taps
and allow the water to run for 20 minutes.
•	Next, flush toilets at least once. If a bathtub has a spout and showerhead, direct flow
through the spout.
•	Flush all outside spigots for 10 minutes.
•	After flushing all cold taps, direct the flow from the bathtub spout to the showerhead, if
applicable.
Flushing Hot Water Taps and Water Heater
10

-------
•	Run the hot water tap closest to the hot water heater and proceed to open all hot water
taps.
•	If a bathtub has a spout and shower head, direct flow through the shower head first.
•	Allow the water to run for at least 75 minutes and then turn off the faucets.
•	If applicable, direct shower head flow to bathtub tap for 2 minutes.
Flushing Appliances
•	Run empty dishwasher and washing machine once on rinse cycle.
•	Replace all water filters (e.g., whole-house filter, refrigerator filter, etc.) and empty ice
from ice maker bin; run ice maker and discard 2 additional batches of ice.
Based on the suggestions given in the report, the premise plumbing system was flushed in stages.
First, the utility sink hot water tap was closed, the valve to the hot water heater was turned off,
and the cold water tap on the utility sink was fully opened. Simultaneously, the cold-water
dispenser on the refrigerator was supposed to be opened to its maximum setting. However, as
noted earlier, clogging of the refrigerator water system prevented the cold-water dispenser on the
refrigerator from being used. Therefore, only the utility sink tap was flushed for 20 min. After
the cold-water pipes were flushed, the hot water heater tank was drained and filled with tap
water, and the hot water tap in the utility sink was fully opened for 75 min. At the conclusion of
the hot water flushing, the dishwasher and washing machine were operated for one cycle.
When chlorine was introduced to the 450-ft WSTB pipe, it flowed into the plumbing system.
The hot water tank was emptied and refilled with highly chlorinated water, and elevated chlorine
level was confirmed to be present at the hot and cold utility sink taps. Subsequently, the
appliances were run for one cycle, and then the whole premise plumbing system sat stagnant for
approximately 20 h. After the 20 h stagnation period, the plumbing system was flushed as
described above, and then baseline flow was restored for the duration of the experiment (138
gallons per day, or 0.096 gpm (363 mL/min)).
2.3.3 Individual Pipe Sections
To compare data on decontamination of heavily tuberculated (corroded) unlined pipe to the
much smoother cement-lined pipe from INL, individual sections of pipe were set up next to the
lagoon. The setup is shown in Figure 4. In this setup, two pipe sections obtained from DC
Water were made of unlined iron with heavy corrosion, and one section was a piece of cement-
mortar-lined iron INL pipe. The goal of this experiment was to have decontamination data on
both types of pipe contaminated with lagoon water and both pipe types contaminated with
Potomac River water. Potomac River water was added to a short (10 ft) section of corroded iron
and a separate short section of cement-mortar-lined INL pipe. Lagoon water was used to
contaminate the other corroded section (the middle pipe section in Figure 4) of DC Water pipe.
A cement lined pipe section from INL was not contaminated with the lagoon water, since data
was already available for the prior experiment with the 450 ft of INL pipe.
Pipe sections were set up on a rack as shown in Figure 4. It was not practical to form biofilms in
the short pipes in the same way as the 450-ft distribution pipe or home plumbing system.
Therefore, before contamination, the inner pipe surfaces were wetted with local tap water from a
11

-------
garden hose. Surfaces were completely wetted over the course of approximately three minutes.
After wetting, the pipe sections were filled with either lagoon or Potomac River water. Potomac
water shipped from DC Water was poured directly into a DC Water and an INL pipe, and a
separate carboy was filled with lagoon water using and poured into a DC water pipe. Water was
poured into the pipe sections directly from the carboy or using a bucket. As seen in Figure 4, the
pipe sections had removable caps on each end that kept the pipes full and were used for drainage.
Each pipe also had a valve that was used to allow air to escape and reduce pressure on the pipe
caps. Once contaminated, the pipe sections sat stagnant for 20 h.
After contamination, the pipe caps were removed, and the pipes were drained. Flushing was
simulated by running tap water over the inner pipe surfaces for five minutes using a garden hose.
Chlorination was accomplished by filling the pipes with chlorinated water and allowing the pipes
to sit stagnant for another 20 hours. Chlorine was made using store-bought concentrated bleach,
and concentrations in the pipe sections ranged from 60 to 150 mg/L. The range in values is
likely due to variation in chlorine demand in different pipes, which is difficult to predict. After
the chlorination period, the pipes were emptied and rinsed with a garden hose. The pipes were
then filled with tap water and allowed to sit for one day before being emptied and rinsed. Bulk
water and scrape samples were collected from the pipe sections at each step in the process.
2.4 Experimental Methods
Methods used to conduct the decontamination methods at the WSTB site are described in this
section. Non-standard methods are described in detail. Standard methods or methods with
publicly available references are noted by referencing the method but are not described in detail.
Preparation and transport of E. coli stock
An E. coli suspension was produced by mixing an inoculum of E. coli cells in nutrient broth and
incubating at 37 °C for 24 h. The resulting stock concentration was lxlO11 CFU/100 mL. The
resulting prepared stock was shipped to the INL site in two separate 500-mL containers inside
coolers at 4 ± 2 °C. This suspension was dumped into the lagoon waters and mixed by having an
individual walk around in the lagoon in waders and mix the water.
Extraction of biofilm and adhered E. coli from coupon and pipe surfaces
Removable cement-mortar coupons in the WSTB pipe were sampled by shutting off water flow
to a small section of the pipe containing the coupons, draining water to relieve pressure, and
unscrewing the coupons. The coupons were scraped with a sterile scalpel (Thermo Scientific,
Waltham, MA) into a sterile sample bottle (Thermo Scientific, Waltham, MA) containing sterile
buffer (Sigma Aldrich, St. Louis, MO) while periodically rinsing the scalpel with sterile buffer.
Pipe surface samples were taken directly from the inner surface of the tuberculated pipe sections.
The biofilm, corrosion and spores were scraped from the surface using a disposable sterile
surgical scalpel (Thermo Scientific, Waltham, MA). An O-ring (Grainger, Lake Forest, IL) with
an area of 0.371 square inches (in2) (2.4 cm2) was placed on the pipe wall, and the area inside the
O-ring was scraped to ensure that the same area was scraped for each sample.
For each type of sample, the extracted material was collected in a sterile sample bottle with a
sodium thiosulfate tablet (Thermo Scientific, Waltham, MA) (for dechlorination of the water)
12

-------
and 100 mL of pre-filled carbon-filtered water. The extracted sample was transferred to a cooler
at 4 ±2 °C. The samples were shipped cooled overnight to the EPA laboratory and analyzed
upon receipt but within 24 h of sampling. As noted in the results, some samples were analyzed
in the field for coliform presence/absence.
Bulk Water Sampling
The Bulk Water Sample (BWS) for conforms/A", coli and other water quality parameters were
collected using the grab sampling technique in 100 mL sterile sample bottles with a sodium
thiosulfate tablet. The bulk water sampling port in the WSTB coupon section was opened and
the water was drained for 15 seconds prior to collection of 100 mL of water from the WSTB.
Laboratory and Field Enumeration of Coli forms//',', coli
Upon receipt in the laboratory, coliform and E. coli samples were immediately analyzed using
Colilert-18 (IDEXX Corp, Westbrook, ME), which conforms to method 9222D in Standard
Methods for the Examination of Water and Wastewater (APHA, 2005). Samples analyzed in the
laboratory were quantified for the number of coli forms/A. coli. Samples analyzed in the field
were noted for presence or absence of conforms/A", coli.
Total Organic Carbon
Upon receipt in the lab, TOC samples were analyzed via EPA Method 9060a
(https://www.epa.gov/hw-sw846/sw-846-test-method-9060a-total-organic-carbon. last accessed
April 7. 2020). Samples were preserved via addition of acid and had a holding time of 28 days,
so analysis was not always immediate upon receipt.
Free Chlorine
Free chlorine samples were analyzed immediately in the field using the Hach Method 10102
using N,N-di ethyl-/;-phenylenedi amine (DPD) (https://www.hach.com/asset-get.download-
en.isa?code=55578 . last accessed April 7. 2020). Samples were diluted in distilled water as
needed.
Turbidity. Conductivity. pH and Temperature
These parameters were analyzed immediately in the field. Turbidity was measured using a Hach
2100P Portable Turbidimeter (Hach Corp., Loveland, CO), and all analyses and calibration
followed the manufacturer's instructions. Conductivity, pH and temperature were measured by a
YSI 556 multiprobe sonde (Xylem, Rye Brook, NY). All analyses and calibrations followed the
manufacturer's instructions.
2.5 Quality Control and Data Quality
2.5.1 Quality Control
Quality control (QC) samples for the contaminant reference method included continuing
duplicate samples, controls and laboratory blanks. The data quality objectives for each of these
quality control samples are provided in Table 1. The acceptable ranges limit the error introduced
into the experimental work. All analytical methods operated within the QC requirements for
controls and laboratory blanks, and unless otherwise noted in the Deviations (Section 2.5.3), all
data quality objectives in Table 1 were met. Note that duplicate samples for E. coli refer to a
13

-------
duplicate analysis of one sample. All E. coli, free chlorine and TOC samples were collected in
duplicate.
Table 1: Quality control data qua
ity objectives
Measurement
QA/QC Check
Frequency
Acceptance Criteria
Corrective Action
Total Coliform/
E. coli
Positive Control
Negative Control
Every batch of
samples
Positive -
Total coliforms - all
wells yellow
E. coli - all wells
fluorescing
Negative -
Total coliforms - No
yellow wells
E. coli - No
fluorescent wells
Use new media vessel and
dilution buffer
Free Chlorine
Manufacturer
DPD* color
standards kit
Once per
experiment
As specified by the
color standards kit
Clean the colorimeter
measuring cell. Clean the
DPD standards vials and
recheck.
Turbidity
Check standard set
for 2100 P
Once per day
Deviation of ± 0.2
NTU
If it fails, repeat
calibration
Conductivity,
temperature, pH
Calibration
Once per
experiment
As specified by
manufacturer
If it fails, repeat
calibration
Total Organic
Carbon
Calibration Curve
(5-point minimum)
When new
standards are
made (30-day
hold time) or a
CC** has
failed
r-value > 0.993
Prepare fresh standards
and analyze again

Calibration Blank
One per batch
of 20 field
samples
[blank] <0.35 mg/L
Organic Carbon
Suggest carryover or
contamination.
Troubleshoot method or
instrument and correct.

Quality Control
Sample (QCS);
also called Initial
Calibration Check
Immediately
after
calibration
80% to 120%
recovery
Remake standard and if
that fails, recalibrate with
fresh calibration standards

Continuing
Calibration Checks
(CCCs)
After every
10th field
sample and at
the end of the
sequence
Vary concentrations
for longer sequences
(low to mid to high)
Low ± 50%
Mid ± 20%
High± 15%
Instrument response may
have drifted. Troubleshoot
and recalibrate if needed.
All field samples should
be bracketed by acceptable
Continuing Calibration
Verifications (CCVs)

Duplicate samples
(Field duplicates
when possible,
Laboratory
duplicates if not)
One pair per
batch of 20
field samples
Relative Percent
Difference (RPD) <
20%
If field duplicates fail, run
laboratory duplicates.
Failing field duplicates
could be sample collection
or matrix issues if the
laboratory duplicates pass.
14

-------
Measurement
QA/QC Check
Frequency
Acceptance Criteria
Corrective Action




If laboratory duplicates
fail, instrument
troubleshooting is needed

Matrix Spike
One per batch
of 20 field
samples
Recover 70% -130%
of spiked
concentration when
compared to a
duplicate sample
Rerun - continued failure
suggests matrix
interference
*DPD = N,N-diethyl-phenylenediamine sulfate
**CC=calibration check
2.5.2	Data Quality
At least 10% of the data acquired during the evaluation were audited. These data include the
biofilm E. coli measurements, BWS and water quality measurements. The data were traced from
the initial acquisition, through analysis, to final reporting, to ensure the integrity of the reported
results. All calculations performed on the data undergoing the audit were checked. No significant
adverse findings were noted in this audit.
2.5.3	Deviations
When conducting scrape samples from the interior of the drinking water pipes, the sampling
method calls for using an O-ring to isolate the area to be sampled. Scraping within the O-ring
area was meant to standardize the pipe surface area that was sampled. The tip of a scalpel was to
trace the sampled area inside the O-ring. It was observed in the field that the traced area was not
always an exact circle. Therefore, the area sampled may have varied between samples. In some
cases where tuberculation was heavy, the O-ring was not used at all, and the scraped area was
estimated. It was not possible to precisely quantify this variation. It should also be noted that
the level of tuberculation varied between pipes, and spatially within individual pipes. However,
it was estimated that the sampled area could have varied by 5% between samples and this
variation should be considered when interpreting the data.
Another deviation was observed with the Colilert-18 E. coli presence/absence testing in the field.
The interaction of the E. co/z'-spiked lagoon water with the heavily corroded DC Water short pipe
section and increased levels of chlorine during that phase of the experiment created a yellow
tinted water. When coliforms are present, the Colilert-18 test turns water samples yellow. The
water samples from this pipe were already yellow before they were placed in the incubator. The
sample did appear to turn a darker shade of yellow overnight, indicating the possible presence of
E. coli, but this result was a judgment call. The over-chlorination step for the DC short pipe
section (filled with lagoon water) was repeated and the yellow tint was much lower but still
yellow prior to incubation. The laboratory sample results should be used to provide the final
determination of coliform presence for this pipe section.
3.0	Experimental Results
3.1	Decontamination of the 450-ft 8-in Diameter Distribution Pipe
Figure 13 and the figures that follow in this section show the various phases of the
15

-------
decontamination experiment. Background is the baseline level in the water or on the pipe
surface before contamination. The contamination phase is when lagoon water was introduced
into the pipe. The flushing phase is when the downstream fire hydrant was opened for 20 min at
a flow of approximately 150 gpm. Chlorination is when chlorine was introduced to the pipe, and
the water in the pipe sat stagnant for roughly 20 h. Return to service occurred after the chlorine
was flushed from the pipe, and local tap water flow was reintroduced.
Figure 13 shows that coliforms/A", coli were not detected in the tap water flowing through the
pipe before contamination, but they increased upon introduction of the lagoon water. After
flushing the pipe for 20 min, the coli forms/A", coli levels in the bulk water retuned to non-
detectable. The same was true after chlorination and during the return to service phase. The data
suggest that flushing alone may remove coliforms/A. coli from the cement-mortar lined
distribution pipe. However, only one data point after flushing exists, which was followed by
chlorination. Flushing followed by chlorination should be sufficient to remove coliforms/A. coli
in untreated river water from cement-mortar lined iron pipe.
l.E+06
E l.E+05
o
o
l.E+04
R l.E+03
£ l.E+02
o
u
l.E+01
0.E+00

t










u
1





Total C.nlifnrm (MPN/lOO mil







E. coli (MPN/100 ml)


cd r
c
n



n
zr




23
ID
r-f
c

S3 5!
| 3<£
1 3


o
-1
3'
QJ
i-f




—1
3
1—~
0
in

Q_ r+
o'
6 m—


O
3




fD
-1
<
n"
a>
10 20 30 40 50 60
Time (hrs)
70
80
90
100
Figure 13: Contamination and decontamination of coliforms/ii. coli in the bulk water phase
of the 450 ft cement-mortar lined iron distribution pipe.
Figure 14 shows that the number of coliforms/A. coli adhered to the interior surface of the
cement-mortar lined iron pipe followed the same trend as in the bulk water phase. The number
of adhered coli forms/A. coli spiked during contamination but returned to non-detectable levels
after flushing. No detectable coliforms/A. coli were observed after chlorination or return to
service. The data suggest that flushing alone may remove coli forms/A. coli from the lined
distribution pipe interior surfaces. However, only one data point after flushing exists, which was
followed by chlorination. Flushing followed by chlorination should be sufficient to remove
conforms/A", coli in untreated river water from the distribution pipe interior surfaces.
16

-------
£
1_
g
o
u
l.E+04
l.E+03
l.E+02
l.E+01
O.E+OO -
0
03
CJ
o
7T
UQ
—1
o
c
3
Q-
i


f


n -n
2	c
zi. co
rj =r
3	3°


|. era
QJ
o"
3


-Total Coliform (MPN/cm2)
-E. coli (MPN/cm2)
n
o
3
CJ
o
3
10 20 30 40 50 60
Time (hrs)
70
80
90
33
fD
<¦+
c
-1
3
r+
O
LO
fD
<
o'
fD
100
Figure 14: Contamination and decontamination of coliforms/E. coli from the pipe surface
of the 450-ft cement-mortar-lined iron distribution pipe.
Figure 15 and Figure 16 show how water quality parameters changed during each phase of the
decontamination experiment. Distinct spikes in conductivity, turbidity and TOC were observed
during the contamination phase, with decreases following in the flushing, chlorination and return
to service phases. In the decontamination phases, turbidity and TOC returned to baseline levels.
After the contamination phase, conductivity was 4% lower than the baseline levels. Ftowever, it
is unclear if the decrease relative to baseline was a true decrease in the conductivity of the water
or drift in the conductivity sensor. Compared to the baseline levels, a drop in pH was observed
during contamination. However, in the decontamination phases, pFI levels were variable and
below the baseline. The data suggest that TOC, turbidity and possibly conductivity could be
parameters that could indicate that untreated water has been removed from the distribution
system.
500 i ^—i—	—r 25
450
400 •	k		 «			S	20
£ 350
=3
300 I	15 fc
250
T3
o 150
o
100
50
0

t








#	

t






S




V*
			

f











1
u


—•—Conductivity (uS/cm
> „

ro r
CD O
S- 3
' -n
c

n
zr
q
—~—Turbidity (NTU)
fD
c

era cu
3 i
3^
3*

—i
3*
QJ
o
3




3

i. "
Q_ !—*
• CO





o
CO
fD

O
p







—%
<
o"
fD






•
	4
»-• 1
200	7T rt a	o	5 10
•Q
I—
3
10 20 30 40 50 60 70 80 90 100
Time (hrs)
Figure 15: Conductivity and turbidity measurements during each stage of the
17

-------
decontamination experiment.
Figure 16: TOC and pH during each stage of the decontamination experiment.
3.2 Decontamination of the Premise Plumbing System
The figures in this section show the results from the background, contamination and
decontamination sampling in the home plumbing system. Background is baseline level of each
parameter coming out of the water taps or water in the appliances before contamination. The
contamination phase is when lagoon water was introduced into the pipes and appliances. The
flushing phase is when the taps were flushed, the hot water heater drained and refilled, and the
appliances run for one cycle (without detergent). Chlorination is when chlorine was introduced
to the pipes and appliances and allowed to sit for approximately 20 h. Return to service occurred
after the chlorine was flushed from the pipe, and local tap water flow was reintroduced into the
plumbing and appliances.
Figure 17 and Figure 18 show the coliform and E. coli levels in the plumbing and appliances
during the phases of the experiment. In both figures, no microbial contamination was detected in
the baseline phase, which was expected. Increases in coliform and E. coli were observed in all
pipes and appliances. After following the flushing procedure described in Section 2.3.2, no
coliforms or E. coli were detected in the hot- or cold-water pipes, or in any of the appliances.
The same result was found after chlorination. The data suggest that flushing the hot- and cold-
water pipes according to the procedure described in Section 2.3.2 removed coliform and E. coli
contamination that came from untreated water.
18

-------
l.E+06
.=- l.E+05
o
o
l.E+04
~ l.E+03
l—
o l.E+02
~ra
° l.E+01
O.E+OO
I Utility Sink-Cold
I Utility Sink-Hot
I Hot Water Heater
I Washing Machine
I Dishwasher
Background
During
Contamination
Post-Flushing Post-Chlorination
Figure 17: Total coliform levels in the bulk water phase of the pipes and appliances in the
home plumbing system.
l.E+06
l.E+05
l.E+04
l.E+03
8 l.E+02
l.E+01
O.E+OO
I Utility Sink-Cold
I Utility Sink-Hot
I Hot Water Heater
I Washing Machine
I Dishwasher
Background
During
Contamination
Post-Flushing Post-Chlorination
Figure 18: E. coli levels in the bulk water phase of the pipes and appliances in the home
plumbing system.
Figure 19, Figure 20, Figure 21 and Figure 22 show the change in TOC, pH, turbidity and
conductivity, respectively, during the phases of the contamination and decontamination
experiment. In Figure 19, TOC increased during the contamination phase relative to the
baseline. In the hot and cold taps and the hot-water heater, TOC levels returned to baseline
levels after flushing and running the appliances and remained at that level after chlorination. In
the washing machine, TOC levels returned to baseline after chlorination, but levels remained
elevated after the flushing phase. The reason for this result is unclear, but it is possible that some
untreated water was trapped in the washing machine after one cycle. It is possible that the
washing machine may need to be run for multiple cycles before all untreated water can be
19

-------
cleared.
TOC levels in the dishwasher were elevated relative to the pipes and other appliances in the
baseline phase and remained elevated after flushing and chlorination. This elevation of TOC
levels has been observed in previous decontamination studies conducted with the home plumbing
setup (Szabo et al., 2017). This result is likely due to organic compounds leaching from the
plastic components of the dishwasher. Still, TOC values could indicate when untreated water
has been removed from plumbing pipes, the hot-water heater, and possibly the washing machine.
If the dishwasher is primarily plastic on the inside, TOC values should be used cautiously.
¦	Utility Sink-Cold
¦	Utility Sink-Hot
¦	Hot Water Heater
¦	Washing Machine
¦	Dishwasher
¦¦¦II ¦¦¦!!
Post-Flushing Post-Chlorination
Figure 19: TOC levels in the bulk water phase of the pipes and appliances in the home
plumbing system.
Figure 20 shows pH values during each phase of the experiment. The pH value generally
decreased as contamination, flushing and chlorination took place. However, it is possible that
this decrease is due to natural variation in the pH of the source water, and it is unlikely that this
variation could be used to determine when untreated water had been removed from the plumbing
and appliances. Conversely, Figure 21 and Figure 22 show that turbidity and conductivity,
respectively, increased during contamination. Similar to the 450 ft distribution pipe, turbidity
returned to its baseline value after flushing and chlorination, but conductivity was 3-5% lower
than baseline. It is unclear if the decrease relative to baseline was a true decrease in the
conductivity of the water or drift in the conductivity sensor. The data suggest that turbidity and
possibly conductivity could be parameters that could indicate that untreated water has been
removed from the plumbing system pipes.
6.00
5.00
4.00
~5fl
3.00
u
O
2.00
1.00
0.00
llll
I
Background
During
Contamination
20

-------
Figure 20: pH levels at the hot and cold taps in the home plumbing system.
30
¦ Utility Sink-Cold
Background	During	Post-Flushing Post-Chlorination
Contamination
Figure 21: Turbidity levels at the hot and cold taps in the home plumbing system.
21

-------
¦	Utility Sink-Cold
¦	Utility Sink-Hot
II II
Post-Flushing Post-Chlorination
Figure 22: Conductivity levels at the hot and cold taps in the home plumbing system.
3.3 Decontamination of individual corroded iron pipe sections
As described in Section 2.3.3, individual short sections of pipe were set up next to the lagoon.
Two of these sections were made of cast iron and had a heavily corroded (tuberculated) inner
surface, and one section was cement-mortar lined water pipe. Lagoon water was added to one
section of corroded iron pipe, and water from the Potomac River was added to one corroded iron
and one cement-mortar pipe section. This allowed for data collection with lagoon water on both
types of pipe (one short corroded iron section, and the 450-ft cement-mortar-lined pipe) and real
source water from the Potomac River on both types of pipe sections.
TOC and other water quality parameters were not monitored in the short pipe sections. Data
collection focused exclusively on coliform analyses. Figure 23 shows the number of coliforms
in the bulk water phase in the individual pipe sections during each phase of the experiment.
Potomac River water coliform levels were between 10-20 MPN/100 mL inside both types of
pipes. After the river water was drained, and the pipe section was "flushed" with a garden hose,
no coliforms were detected in the bulk water when the pipe was refilled. No coliforms were
detected after chlorination in either pipe. The same trend was observed with coliforms adhered
to the inner pipe surface (Figure 24).
Figure 23 also shows the level of coliforms in the bulk water phase in a corroded iron pipe
section filled with lagoon water. Lagoon water was substantially more concentrated in coliforms
than Potomac River water, with approximately 104 MPN/100 mL. After draining the pipe
section filled with lagoon water, flushing with a garden hose and then refilling, coliforms were
still observed in the bulk phase. Figure 24 shows that coliforms were also detected on the pipe
surface after flushing. No coliforms were seen in either phase after chlorination. However, once
chlorination was complete, and the pipe was refilled, coliforms were observed in the bulk water
phase (2 MPN/100 mL). These data suggest that coliform remained adhered to the corroded iron
pipe surface after chlorination but detached from the pipe surface and were detected in the water.
480
Background	During
Contamination
22

-------
The results from these experiments indicate that that coliforms in the range of 10-20 MPN/mL
from untreated Potomac River water can be removed from the corroded iron (DC Water) and
cement-mortar lined (INL) iron pipe with "flushing" and chlorination (possibly with only
flushing). However, the data from these pipe sections is a small-scale representation of an actual
distribution pipe, and the results should be replicated on a larger scale. For the cement-mortar-
lined pipe, lagoon water with three orders of magnitude more coliforms were successfully
decontaminated from the 450-ft distribution pipe with flushing and chlorination. Therefore, it is
likely that coliforms from Potomac River water could also be successfully decontaminated using
these techniques on the full scale.
Compared to lagoon water results in the 450-ft pipe cement-mortar pipe, coliforms were more
persistent on the corroded iron pipe section (Figure 23). It is possible that multiple rounds of
flushing and chlorination would be needed to remove them. However, for the lagoon water tests,
the corroded iron (pipe section) and cement-mortar lined pipe (450-ft distribution pipe)
experimental setups were not the same, and caution should be used when comparing the results.
In the future, adding corroded iron pipe sections to the 450-ft distribution pipe and repeating
these experiments with untreated lagoon water would shed light on whether the results can be
replicated at a larger scale.
l.E+05
£ l.E+04
o
o
l.E+03
£
o
~o
u
¦*-'
o
l.E+02
l.E+01
0.E+00
I INL Pipe with Potomac River Water
I DC Pipe with WSTB Lagoon Water
I DC Pipe with Potomac River Water
III!
Background
Post
Contamination
Post Flushing
Post
Chlorination
Post 2nd
Flushing
Figure 23: Total coliform levels in the bulk water phase of the individual short pipe
sections.
23

-------
l.E+04
E l.E+03
IINL Pipe with Potomac River Water
I DC Pipe with WSTB Lagoon Water
I DC Pipe with Potomac River Water
E l.E+02
i—
o
o
u
l.E+01
0.E+00
Background	Post Post Flushing Post	Post 2nd
Contamination	Chlorination Flushing
Figure 24: Total coliform levels on the inner pipe wall of the individual short pipe sections.
4.0 Conclusions
The following points summarize the results of the untreated water contamination and
decontamination experiments performed at the WSTB:
•	After contamination of the 450-ft distribution pipe, an increase in conforms/A", coli was
observed, as well as increases in TOC, conductivity, pH and turbidity. After flushing
water through the fire hydrant at approximately 150 gpm, no coli forms/A", coli were
observed in the bulk water or on the inner pipe surface. The same was true after
chlorination and returning tap water flow to the pipe. However, only one sampling event
took place after flushing and before chlorination. In future experiments, it should be
confirmed that flushing alone is adequate to remove microbial contamination from the
cement-mortar lined pipe. If possible, the impact of above ground storage tanks,
distribution network complexity (e.g. pipe loops and other configurations) and various
flow velocities will be investigated.
•	In the 450-ft distribution pipe, TOC, conductivity and turbidity returned to near baseline
levels after decontamination (pH was variable). Baseline levels were maintained during
chlorination and after tap water was returned to the pipe. The data suggest that TOC,
turbidity and possibly conductivity measurements could indicate that untreated water has
been removed from the distribution system. Using ultraviolet-visible light-based on-line
TOC sensors in real time to detect changes in TOC is also a topic of interest.
• Increases in coli forms/A. coli TOC, conductivity, pH and turbidity were observed in the
plumbing pipes and appliances after contamination. Flushing the plumbing pipes,
emptying the hot water tank and running the appliances for a cycle reduced coli forms/A.
coli to undetectable levels, and returned turbidity and conductivity to baseline. Like the
450-ft distribution pipe, it might be informative to repeat these experiments using
flushing only.
24

-------
•	In the plumbing system, TOC returned to baseline after flushing and chlorination in the
hot- and cold-water pipes and the hot-water heater. A spike in TOC was observed in the
washing machine after the flushing cycle, possibly due to untreated water being trapped
in the appliance. TOC returned to baseline after the chlorination step. TOC data from
the dishwasher were elevated during all phases, likely due to organic material leaching
from the plastic surfaces. The data suggest that TOC (in the situations noted above)
could indicate that untreated water has been removed from the plumbing system.
Turbidity and possibly conductivity might also be parameters that could indicate that
untreated water has been removed from the plumbing pipes.
•	Individual sections of heavily tuberculated iron and cement-mortar-lined iron pipe were
filled with contaminated water, emptied and rinsed (to simulate flushing), then
chlorinated and flushed. Potomac River water with conform/A", coli levels in the 10-20
MPN/100 mL was successfully flushed from the pipe with no microbial contamination
remaining in the water or on the pipe surface.
•	After flushing, coli forms/A", coli from untreated lagoon water were detected on the inner
surface of the corroded iron pipe and in the bulk water. None were detected on the pipe
surfaces after chlorination, but coliforms were detected in the bulk water after the
chlorine was flushed out. This differs from the decontamination results using untreated
lagoon water in the 450-ft cement-mortar-lined pipe. However, the experimental systems
used for the pipe types were different, and it would be informative to repeat these
experiments on the full scale with corroded iron pipe built into the 450-ft distribution
pipe system.
In summary, flushing and chlorination were effective at removing coliform bacteria from
cement-mortar-lined infrastructure and home plumbing and appliances. However, it would be
informative to repeat the contamination experiments in the 450-ft distribution pipe and home
plumbing to confirm that flushing alone is an effective decontamination technique without
chlorination. Water quality parameters such as TOC, conductivity and turbidity returned to pre-
contamination baseline levels in most cases, suggesting that monitoring water quality might be
an effective method of determining the progress of decontamination. This information would be
useful to utility responders since parameters like TOC, conductivity, and turbidity are easier and
faster to obtain with field instruments than E. coli, which requires lengthy incubation times (18
to 24 h). Coliforms were present after simulated flushing in one section of heavily tuberculated
iron pipe, and they reappeared in the bulk water after flushing and chlorination. These data
suggest that heavily tuberculated iron may be more difficult to decontaminate than the cement-
mortar lined iron used in other experiments. It would be informative to repeat these experiments
at full scale with corroded iron pipe built into the 450-ft distribution pipe.
25

-------
5.0 References
APHA (American Public Health Association) (2005). Standard Methods for the Examination of
Water and Waste Water, Method 9222D, 21st edition. Washington, DC: American Public
Health Association.
AWWA (American Water Works Association) (2005). Standard C651-05: Disinfection of Water
Mains. Denver, CO: American Water Works Association.
DeOreo, W.B., Mayer, P.W., Dziegielewski, B., and Kiefer, J.C. (2016). Residential Uses of
Water 2016. Water Research Foundation. Denver, CO.
Szabo, J.G., Hall, J., Goodrich, J. and Ernst, H. (2017). Full scale drinking water system
decontamination at the Water Security Test Bed. Journal of the American Water Works
Association, 109 (12), E535-E547.
USEPA (U.S. Environmental Protection Agency), 2016. Water Security Test Bed Experiments at
the Idaho National Laboratory. EPA/600/R-15/146, Washington DC: U.S. Environmental
Protection Agency.
USEPA (U.S. Environmental Protection Agency), 2018. Decontamination of Bacillus Spores
from Drinking Water Infrastructure with Physical Removal (Pigging). EPA/600/R-
18/078, Washington DC: U.S. Environmental Protection Agency.
WRF (Water Research Foundation), 2016. Flushing Guidance for Premise Plumbing and
Service Lines to Avoid or Address a Drinking Water Advisory. Report #4572. Denver,
CO.
26

-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300

-------