Water distribution system flushing is one way to get rid of contamination. In conventional flushing, all the contaminated water is discharged to the environment, thereby harming it. A new method is proposed here as an alternative solution, in which a containment pond lined with impermeable material will be constructed in a suitable place within the municipality. Network modelling was performed to investigate the feasibility of the new method. It was found that (1) the proposed flushing method can successfully reduce environmental impacts compared to hydrant flushing only, (2) a containment pond cannot clear the system periphery away from the containment pond, (3) the best location of a containment pond is not always at the furthest location from the source reservoir, and (4) for some systems, some pond locations might be better from an economic perspective, while other locations will be better environmentally.

  • Containment ponds for water system flushing can successfully reduce environmental impacts by up to 100%.

  • A pond cannot clear an entire system with numerous dead ends.

  • The best location of a pond is not always at the furthest location from the source reservoir.

  • A pond location tradeoff between cost and environment can exists for some systems.

A water distribution system (WDS) is the physical network of pipes that delivers water from the water source to the intended users. Typically, this is achieved by pumping water from the source or reservoir, transporting it through the watermains and service pipes, and storing it in the elevated storage tank/s used for storage to handle fluctuating user demand. Valves and hydrants are secondary components of a WDS, where valves are used for controlling water flow and hydrants are for emergency situations like firefighting or contamination flushing (Viessman et al. 1998; Mays 2000; Wu 2015). The main purpose of a WDS is to deliver safe and uninterrupted drinking water to the users (Rasekh & Brumbelow 2013). However, contamination intrusion in the WDS is not an uncommon event (Hrudey & Hrudey 2004; Craun et al. 2006; Hrudey & Hrudey 2007; Seth et al. 2016). Any WDS can get contaminated both by intentionally introduced chemicals like chloride or fluoride or undesirable introduction of salt, lead, arsenic, or emerging contaminants from surface or ground water. Physical deterioration of the WDS components or equipment can also induce contamination (Rasekh & Brumbelow 2013).

In addition to accidental or intentional WDS contamination, saltwater intrusion is important in the coastal areas as a result of overextraction of water from freshwater aquifers. Overextraction of freshwater initiates sea water to seep through the aquifer (Edwards et al. 2009; Gleeson et al. 2012; Doell et al. 2014; Spellman 2017). This is now considered to be a significant threat to the drinking water quality of many coastal areas including Los Angeles (Edwards & Evans 2002; Spellman 2017), Georgia, northeastern Florida (Spechler 2001; Czajkowski et al. 2018), southwestern Nigeria (Ayolabi et al. 2013; Yusuf & Abiye 2019), Tamilnadu of India (Gopinath et al. 2016), and coastal areas of Bangladesh (Faneca Sànchez et al. 2015; National University of Singapore 2020). Saltwater is not removed in the water treatment process and then is pumped into the distribution system.

Although contamination intrusion in a WDS is not a frequent incident, quick detection and proper management are important to protect community health (Poulin et al. 2010; Seth et al. 2016). Most contaminants spread rapidly through the distribution networks. At any contamination event, contamination detection is very difficult and nearly impossible to treat (Khanal et al. 2006; Seth et al. 2016). So, when a system becomes contaminated and treatment is not convenient, system flushing is a common way to get rid of contamination (Friedman et al. 2002; Shafiee & Berglund 2017). Usually, two techniques for flushing are used for WDS decontamination: conventional and unidirectional. In conventional flushing, opening of fire hydrants (usually one by one) can be sequential or non-sequential with no valve operation and eventually can fail to remove the contamination entirely from the system due to lack of adequate velocity within the pipes. In unidirectional flushing, fire hydrants are opened in a sequential manner, while the pressure valves on the consumers’ sides are kept closed to ensure sufficient velocity. The recommended range of velocity for this flushing technique is 0.8 m/s (2.5 ft/s) to 3 m/s (10 ft/s) depending on the size of the utility and the type of the contaminant to be removed (Antoun et al. 1999; Shah et al. 2001; Hasit et al. 2004; Walski et al. 2008; Martin & Ries 2014; Wu 2015; Xie et al. 2015). Compared to conventional flushing, the use of water can be reduced by up to 40% using unidirectional flushing, however this technique also involves some constraints e.g., labor, time allocation, and proper management. Continuous blow-off is another flushing technique especially for stagnant areas of a WDS e.g., dead ends or large-sized pipes. Unfortunately, this method is not considered to be durable in general as typical velocity values obtained in this flushing technique are less than 0.3 m/s (1 ft/s) which is not sufficient to remove all types of contaminants (Oberoi 1994; Antoun et al. 1999; Hasit et al. 2004; Barbeau et al. 2005; Rebolledo et al. 2020). A routine flushing program is also followed in some water systems to maintain the water quality where the frequency of flushing can be monthly, quarterly, semiannually, annually, or seasonally. The frequency is selected based on the size of the system in addition to the susceptibility of the system to any chemicals, corrosion, high level of disinfectant residual, sediment accumulation, and/or customer complaints (Friedman et al. 2002; MELCC 2019).

Comparative studies have been performed on existing flushing methods including the use of aggressive flushing for identifying discoloration factors (Boxall et al. 2003), modelling of sediment in WDSs using unidirectional flushing (Carrière et al. 2005), evaluation of flushing to remove contamination (Polychronopolous et al. 2003; Vitanage et al. 2003; Poulin et al. 2010), modelling of a contamination events (Haxton & Walski 2009), planning and optimization of unidirectional flushing (Deuerlein et al. 2014), optimization of hydrant selection for conventional flushing (Wu 2015), mobile flushing to prevent secondary water contamination (Kowalski et al. 2015), valve management for improved water quality (Quintiliani et al. 2019), and changing of outflow to provide sufficient chlorine residuals (Avvedimento et al. 2020). However, the scope of all these studies was limited to hydrant flushing and no alternatives were introduced. Although hydrant flushing is the most effective way of WDS decontamination, it is not free from environmental problems. Whatever the technique of hydrant flushing, all the contaminated water is discharged to the environment and finally ends up in water bodies, agricultural fields, and/or wastewater treatment plants through combined sewers, which has a detrimental effect on the environment (Barbeau et al. 2005; EPA 2020).

To reduce the environmental impact, a containment pond located at the system periphery is evaluated here as an alternative solution. The pond will contain the contaminated water so that it does not spread to the environment with the impermeable liner obstructing infiltration so that the contamination does not reach the groundwater. The water can evaporate and leave behind the contaminant to be disposed of periodically. However, the pond location is critical, as its capacity will vary depending on its position in the distribution network.

The use of containment ponds is not an uncommon concept in the industrial sector. Ponds are constructed for various purposes, such as cooling, stabilization, settling, and oxidation. Cooling ponds are used to store and eventually cool down heated water from the nearby industries (Ryan et al. 1974; Mann & Mann Technology Limited Partnership 1991; Ramamoorthy et al. 2001; Barisevičiūtė et al. 2020). Stabilization ponds are used to remove or reduce turbidity, solid pollutants, and/or pathogens in many industries including wastewater treatment, mining, agriculture, aquaculture, etc. (Gray 1988; Sah et al. 2012). Settling ponds and oxidation ponds are also used for similar purposes (Elmaleh et al. 1996; Mispagel & Gray 2005; Merricks et al. 2007).

The purpose of this study was to evaluate the performance of a containment pond as a way of WDS decontamination for nine real WDSs. Here, the performance was evaluated based on the reduction of environmental impact caused during hydrant flushing alone. In addition, the best location of a containment pond based on minimizing cost and environmental contamination is examined. Capturing the contaminated water in a containment pond is more desirable than discharging to the storm sewer as the contaminant, e.g., salt, might not be removed by a conventional wastewater treatment plant and can even disrupt the biological processes used to remove pathogens. A single pond is modelled here for simplicity. The innovative aspects of this work are that it is the first such paper to address the use of using containment ponds for disposal of flushed contaminants from a WDS. No other paper has looked at that. Using containment ponds keeps the contaminant from interacting with the environment and thus polluting the environment.

Procedure

To capture the contaminated water in a containment pond as an alternative to hydrant flushing, and to find the best pond location, the contaminant transport of a conservative contaminant [in this study salt (Baird 2013) but the results will be applicable to any conservative contaminant] was modelled using the network solver EPANET (Rossman 2000), a network solver which can calculate pressure, velocity, discharge, and concentration of constituents of any WDS using a demand-driven analysis, provided that the user demands and characteristics of the WDS components are known. The proposed procedure is comprised of two phases in which the first phase is for determining the containment pond volume and the second phase is for determining the direct environmental impact, if any.

Phase I consisted of fresh water being pumped into the already fully contaminated system from the reservoir source and contaminated water discharged at the hydrant closest to the containment pond until the system was clean. If areas of the system were not cleared of contamination, then additional hydrants were opened near the contaminated area until the entire system was clean. A hydrant opening was simulated by discharging the highest hydrant flowrate that did not result in a negative pressure value anywhere in the system. Phase II consisted of measuring the volume discharged through the pond over the time to decontaminate the system. This determined the volume of excavation required for the pond. Then the amount of volume discharged from all the additional hydrants to clear the other areas of the system was added to calculate the total amount of water contaminating the environment, since it did not enter the pond.

Intuitively, it might seem that if the containment pond is placed at the furthest location from the reservoir, it should clear the system most efficiently since this is the way water has to travel through most regions of the system. To investigate the best pond location, possible pond locations examined here were at the three outer ‘corners’ of a system, away from the reservoir/s – where the second location was the furthest of all. While it is infeasible to model all possible pond locations, the ‘corner’ is representative of the main possible locations. In every case, all the valves at the consumers’ ends were kept closed, like in unidirectional flushing or, in other words, there were no used demands (Antoun et al. 1999; Shah et al. 2001; Hasit et al. 2004; Walski et al. 2008; Wu 2015). Three simulations were run to determine the capacity of the containment pond at the selected locations. For each simulation, discharge was increased until either the pressure was zero at any time or location or any part of the system experienced a reduction in contamination level to zero. Pond volume was found by that value of discharge over the entire simulation period. In case/s where draining the system into a pond was not able to remove all the contaminant, fire hydrants at the dead ends were modelled as being open to clear the rest of the system. Opening fire hydrants do directly discharge the contaminated water to the environment, however this would have happened even more in conventional flushing procedures.

The optimal location of the containment pond for a system was determined by complete enumeration of the three ‘corner’ pond locations, with the objective of minimizing both cost and amount of contaminated water discharged into the environment. This was a multi-objective optimization problem in which cost was minimized due to typical budget constraints (Equation (1)) and the amount of contaminated water put into the environment was minimized (Equation (2)) since any contaminant is assumed to be detrimental (Barbeau et al. 2005):
(1)
(2)
subject to all pressure values being positive at all locations and times*.

*It is assumed here that any positive pressure is acceptable, although some references (Mays 2000) suggest a margin of safety of pressure = 20psi.

Pond total cost was comprised of the pond excavation and lining to prevent infiltration from the pond into the soil or groundwater aquifer, pumping energy costs, and alternative water source costs. Table 1 represents the unit cost per item for pond total cost. Pond lining area was determined from the pond volume assuming a pond depth of 2.4 m (8 ft) based on USDA (2020), and both the pumping energy information and the number of water bottles required were obtained/determined from the EPANET output volume and the volume of water in a single water bottle. For provision of an alternative water source, it was assumed that only domestic water demand would be met during the flushing which is 55.6% of the total demand (Shammas & Wang 2011), and 0.5 L of bottled water would be supplied until decontamination was complete.

Table 1

Unit cost per item for determining total cost

ItemUnit costRemark
Pond excavation $131.0/m3 ($3.7/ft3This value was taken from Home Advisor (2019)  
Pond lining $74.0/m2 ($6.9/ft2This value was taken from Home Advisor (2019)  
Pumping cost Different for different systems in cost/day. Unit pumping cost was obtained directly from EPANET output. This value was also different for different pond locations because of the change in energy used. Total pumping cost was determined by multiplying unit cost with time to clear the system from contamination. 
Water bottles $0.54/0.5 L bottle Number of water bottles required was determined assuming 55.6% of the total water demand to be domestic water demand (Shammas & Wang 2011). The unit cost per 0.5 L bottle was determined based on the average cost of 15 different suppliers 
ItemUnit costRemark
Pond excavation $131.0/m3 ($3.7/ft3This value was taken from Home Advisor (2019)  
Pond lining $74.0/m2 ($6.9/ft2This value was taken from Home Advisor (2019)  
Pumping cost Different for different systems in cost/day. Unit pumping cost was obtained directly from EPANET output. This value was also different for different pond locations because of the change in energy used. Total pumping cost was determined by multiplying unit cost with time to clear the system from contamination. 
Water bottles $0.54/0.5 L bottle Number of water bottles required was determined assuming 55.6% of the total water demand to be domestic water demand (Shammas & Wang 2011). The unit cost per 0.5 L bottle was determined based on the average cost of 15 different suppliers 

For comparison, the base case i.e., hydrant flushing was modelled using the same network solver EPANET (Rossman 2000). Though all the valves at the consumers’ ends were kept closed as is done in unidirectional flushing (Antoun et al. 1999; Shah et al. 2001; Hasit et al. 2004; Walski et al. 2008; Wu 2015), no sequential order was maintained for opening the fire hydrants. Since the purpose of this study was to compare the environmental impact due to hydrant flushing and the use of containment pond, all the fire hydrants at the dead ends were modelled as simultaneously open.

Method application on real WDSs

In this study, nine looped real WDSs were used (Table 2) with systems ranging from 8 to 958 pipes, 6 to 874 junctions, 0 to 7 tanks, different tank positions as per Wang & Barkdoll (2017), 1 to 4 groundwater reservoirs, 0 to 3 surface-water reservoirs, and various range of water demand and pressure. All systems had diurnally fluctuating demand patterns and, in addition, pump controls that activated pumps at low tank water levels and deactivated pumps at high tank water levels. Most of the systems had residential water demand except System B and System C which had some industrial areas in addition to residential. All systems are based on real systems and no data were changed except the variable being studied here, containment ponds. The basis for choosing a pond location was the lowest cost and environmental impact.

Table 2

Primary data of the water distribution systems

SystemNo. of pipesNo. of junctionsNo. of tanksPosition of tank/saNo. of GW reservoirsNo. of SW reservoirsType of the consumersRange of average demand (LPS)Pressure range (m)
FS Residential 9.5–12.6 24–60 
62 44 NS Residential+Industrial 0.06–0.7 29–40 
168 126 FS Residential+Industrial 0.002–12.5 4–357 
135 118 FS Residential 0.06–3.2 14–107 
394 347 FS/MS Residential 0.001–1.4 16–276 
958 874 ND Residential 0.03–9.1 10–354 
39 19 ND Residential 12.6–63.1 19–50 
551 504 Residential 0.02–3.05 10–144 
429 388 FS/MS Residential 0.0004–4.2 7–104 
SystemNo. of pipesNo. of junctionsNo. of tanksPosition of tank/saNo. of GW reservoirsNo. of SW reservoirsType of the consumersRange of average demand (LPS)Pressure range (m)
FS Residential 9.5–12.6 24–60 
62 44 NS Residential+Industrial 0.06–0.7 29–40 
168 126 FS Residential+Industrial 0.002–12.5 4–357 
135 118 FS Residential 0.06–3.2 14–107 
394 347 FS/MS Residential 0.001–1.4 16–276 
958 874 ND Residential 0.03–9.1 10–354 
39 19 ND Residential 12.6–63.1 19–50 
551 504 Residential 0.02–3.05 10–144 
429 388 FS/MS Residential 0.0004–4.2 7–104 

aNear-Direct (ND), Near-System (NS), Far-System (FS), Mid-System (MS) as per Wang & Barkdoll (2017). GW, groundwater; SW, saltwater.

It was found that adding a pond could successfully reduce the system contaminant concentration by storing contaminated water in the pond and not letting it enter the environment (Table 3, Figures S1–S9). In comparatively smaller and completely looped System G, the reduction was 100% for all the pond locations. Two of the pond locations of Systems A, B, C, and H reduced the environmental impact by more than 80% and at least one or more pond location/s of Systems E, F, and I reduced the environmental impact by more than 50%. However, in System D the highest reduction percentage was significantly lower. Figures S1–S9 show the network condition of each system having ponds at their maximum capacities determined from Phase I. Here, the ponds’ maximum capacities are the pond volumes represented in Table 4. The time to clear the system varied from system to system and is reflected in the volume of flow discharged. Using a pond cannot clear areas of the system away from the path from the source to the pond (See Figure S1 as an example). Table 4 presents the pond volume at each location for each system and the time required to clear the entire system with the help of hydrants, when needed. The pond volumes ranged from 613 m3 (21,656 ft3) (System D) to 400,194 m3 (14,132,755 ft3) (System C). Time to clear the systems ranged from 10 (System H) to 999 hours (System C). Table 5 shows the total cost required for constructing the pond at different locations along with the associated pumping costs and bottled water costs, and the concomitant volume of contaminated water discharge to the environment, Vte. Pond total costs ranged from $0.10 M (System D) to $64.59 M (System C). This cost will be termed ‘pond total cost’ hereafter.

Table 3

Reduction of contaminated water to environment (%) after adding pond with hydrants opened to clear the remainder of the system

SystemPond Location #1Pond Location #2Pond Location #3
100 87 75 
71 98 99 
80 15 98 
20 12 16 
32 63 62 
64 71 65 
100 100 100 
22 84 84 
11 52 22 
SystemPond Location #1Pond Location #2Pond Location #3
100 87 75 
71 98 99 
80 15 98 
20 12 16 
32 63 62 
64 71 65 
100 100 100 
22 84 84 
11 52 22 
Table 4

Volume of pond and time to clear the entire system with hydrants opened when needed

SystemPond Volume (m3)
Time to clear the system (hr)
Pond Location #1Pond Location #2Pond Location #3Pond Location #1Pond Location #2Pond Location #3
3,434 2,539 4,315 28 17 32 
4,932 7,682 14,862 89 165 119 
400,194 3,936 74,570 999 178 540 
2,044 613 3,352 66 73 82 
4,614 7,524 8,458 77 163 95 
12,185 7,012 5,223 98 91 58 
29,299 19,760 40,201 43 29 59 
2,592 3,564 4,176 17 10 30 
1,080 37,908 11,664 130 456 136 
SystemPond Volume (m3)
Time to clear the system (hr)
Pond Location #1Pond Location #2Pond Location #3Pond Location #1Pond Location #2Pond Location #3
3,434 2,539 4,315 28 17 32 
4,932 7,682 14,862 89 165 119 
400,194 3,936 74,570 999 178 540 
2,044 613 3,352 66 73 82 
4,614 7,524 8,458 77 163 95 
12,185 7,012 5,223 98 91 58 
29,299 19,760 40,201 43 29 59 
2,592 3,564 4,176 17 10 30 
1,080 37,908 11,664 130 456 136 
Table 5

Pond total cost and environmental discharge for different pond locations

SystemPond Total Costa ($M)
Discharge to the Environment, Vte (M Liter)
Pond Location 1Pond Location 2Pond Location 3Pond Location 1Pond Location 2Pond Location 3
0.56 0.41 0.70 0.00 0.45 0.86 
0.80 1.24 2.40 2.45 0.16 0.06 
64.59 0.66 12.07 17.93 76.23 2.16 
0.33 0.10 0.54 10.22 11.26 10.76 
0.75 1.22 1.37 9.06 4.96 5.05 
1.97 1.14 0.85 8.16 6.72 8.06 
4.75 3.20 6.51 0.00 0.00 0.00 
0.42 0.57 0.67 2.72 0.55 0.55 
0.20 6.15 1.91 83.62 44.79 72.60 
SystemPond Total Costa ($M)
Discharge to the Environment, Vte (M Liter)
Pond Location 1Pond Location 2Pond Location 3Pond Location 1Pond Location 2Pond Location 3
0.56 0.41 0.70 0.00 0.45 0.86 
0.80 1.24 2.40 2.45 0.16 0.06 
64.59 0.66 12.07 17.93 76.23 2.16 
0.33 0.10 0.54 10.22 11.26 10.76 
0.75 1.22 1.37 9.06 4.96 5.05 
1.97 1.14 0.85 8.16 6.72 8.06 
4.75 3.20 6.51 0.00 0.00 0.00 
0.42 0.57 0.67 2.72 0.55 0.55 
0.20 6.15 1.91 83.62 44.79 72.60 

aPond total cost includes pond construction cost, pumping cost, and bottled water cost as alternative water source.

To analyze the results, pond total cost vs Vte has been plotted for each system (Figure 1) and all the systems’ plots have been combined for comparison (Figure 2). It was seen that the range of both the axes were different for different systems. Hence, all the outcomes were compared with the base case and normalized by dividing both the cost and Vte by the maximum value of the same series, thereby making the scale from zero to one. Here, the base case is typical hydrant flushing. The base case cost is the pumping cost and the water bottle costs during hydrant flushing alone and base case Vte is the volume of contaminated water discharged to the environment from those hydrants (Table 6). The base case cost ranged from $0.22 K (System H) to $48.56 K (System C) and the base case Vte ranged from 3.43 M Liter (System A) to 93.44 M Liter (System I). Both the normalized total costs and normalized Vte for different systems were plotted in the same graph (Figure 3). To compare the results with the base case, normalized total cost and Vte associated with the typical hydrant flushing were also plotted in the same graph. It is observed that some systems have a wide range of results e.g., Systems C, H, and I, while other systems have a narrow range e.g., Systems A, D, F, and G. However, Systems B and E have an intermediate range of results. Results having a wide range of values indicate that selecting a pond location has a tradeoff, in which some pond locations might be better from an economic point of view, while others will be better from an environmental perspective.

Table 6

Associated cost and volume of contaminated water to environment during typical hydrant flushing

SystemCost ($K)Volume of contaminated water to environment, Vte (M Liter)
1.77 3.43 
1.58 8.46 
48.56 90.07 
3.03 12.74 
3.77 13.34 
9.42 22.95 
8.65 16.70 
0.22 3.47 
30.74 93.44 
SystemCost ($K)Volume of contaminated water to environment, Vte (M Liter)
1.77 3.43 
1.58 8.46 
48.56 90.07 
3.03 12.74 
3.77 13.34 
9.42 22.95 
8.65 16.70 
0.22 3.47 
30.74 93.44 

Cost includes pumping cost and bottled water cost during hydrant flushing alone. Vte is the volume of contaminated water discharged to the environment from those hydrants.

Figure 1

Pond total cost vs volume of contaminated water to environment corresponding to (a) System A, (b) System B, (c) System C, (d) System D, (e) System E, (f) System F, (g) System G, (h) System H, and (i) System I.

Figure 1

Pond total cost vs volume of contaminated water to environment corresponding to (a) System A, (b) System B, (c) System C, (d) System D, (e) System E, (f) System F, (g) System G, (h) System H, and (i) System I.

Close modal
Figure 2

Pond total cost vs volume of contaminated water to the environment for different pond locations showing that results are system dependent.

Figure 2

Pond total cost vs volume of contaminated water to the environment for different pond locations showing that results are system dependent.

Close modal
Figure 3

Normalized pond total cost vs normalized volume of contaminated water to the environment for different pond locations showing that results are system dependent.

Figure 3

Normalized pond total cost vs normalized volume of contaminated water to the environment for different pond locations showing that results are system dependent.

Close modal

To choose the best location of the pond for each system from the analyzed locations, Figure 3 was utilized. From Figure 1, it is clear that a Pareto front exists for each system, which means contaminated water discharge to the environment cannot be reduced unless the pond total cost is increased. Hence, for each system, the pond location nearest to the origin in figure are optimal. Though the preliminary assumption was that Pond Location 2, being the furthest one from the reservoir, and, therefore, the location for which water would have to travel through the greatest portion of the system, would give the best result by clearing out more of the system, it was not always true. Pond Location 1 was the best solution for Systems B and E and Pond Location 3 was the best for Systems C, F, and I. This happened perhaps for the complex hydraulic characteristics of the systems. For the rest of the systems i.e., Systems A, D, G, and H, Pond Location 2 was preferable, as expected.

If a pond already exists on the system periphery, then it could be used and would avoid excavation and lining costs and thereby improve the feasibility of using ponds as a flushing option. Therefore, to examine all non-pond costs, normalized values of all costs i.e., pumping cost and bottled water cost vs normalized Vte have also been plotted for ponds at different locations and hydrant flushing alone to analyze the results based on other parameters (Figure 4). The best pond location based on non-pond costs was determined following the previous procedure i.e., for each system, the pond location nearest to the origin in Figure 4 was determined to be the best option. Results were negligibly different compared to total cost analysis (Figure 3) except for System D and System E. If pond construction cost is ignored and optimal pond location is selected based on non-pond costs and contaminated water to the environment, the best pond location for System D and System E were Pond Location 1 and Pond Location 3 instead of Location 2 and Location 1, respectively, when pond construction cost was considered.

Figure 4

Normalized non-pond costs vs normalized volume of contaminated water to environment for different pond locations showing that results are system dependent.

Figure 4

Normalized non-pond costs vs normalized volume of contaminated water to environment for different pond locations showing that results are system dependent.

Close modal

The purpose of this study was to explore an alternative to hydrant flushing for a contamination event. It was found that the containment pond method can work as an alternative. The major findings were that the method can reduce the environmental impact caused by hydrant flushing alone and optimal location of the containment pond is system dependent.

It is unclear if the results are generalizable since the pond feasibility is system dependent. The results could be somewhat more complete if all possible pond locations are studied, but the results may not change significantly, since the main pond locations were studied in this study. In addition, some cases call for large excavation amounts in the thousands of cubic meters and hundreds of hours for flushing. The feasibility of this must be decided by the water system managers.

Future improvement

In this study, optimal pond location was selected based on the pond total cost and the volume of the direct discharge of contaminated water. Here, pond total cost comprises the construction cost, pumping cost, and cost of buying water bottles. However, other parameters also exist e.g., production and transportation of the water bottles, production of the bottled water, type of harm caused to the environment from the contaminated water discharge, public perception of the decontamination procedures, peoples’ willingness to spend for the land of the containment pond, etc. The inclusion of the above-mentioned parameters can be considered as future research. Using a pressure-driven analysis may have simplified the modeling, since the authors had to keep adjusting the outflow from the system into the pond until the pressure at the outflow junction was zero, which took some time.

The following points can be concluded from this study:

  • 1.

    The proposed method for WDS flushing can be a better option than hydrant flushing since this method can successfully reduce environmental impacts due to hydrant flushing by up to 100%.

  • 2.

    The method might not be able to reduce the environmental impact by 100% for areas away from the containment pond.

  • 3.

    The best location of a containment pond is not always at the furthest location from the reservoir. So, before selecting the pond location all the outer corners of the system should be studied.

  • 4.

    For some systems, containment pond location varies since a tradeoff exists in which some locations might be better from an economic point of view, while others will be better from an environmental perspective.

Data cannot be made publicly available; readers should contact the corresponding author for details.

Antoun
E. N.
,
Dyksen
J. E.
&
Hiltebrand
D. J.
1999
Unidirectional flushing: a powerful tool
.
Journal-American Water Works Association
91
(
7
),
62
72
.
Avvedimento
S.
,
Todeschini
S.
,
Giudicianni
C.
,
Di Nardo
A.
,
Walski
T.
&
Creaco
E.
2020
Modulating nodal outflows to guarantee sufficient disinfectant residuals in water distribution networks
.
Journal of Water Resources Planning and Management
146
(
8
),
04020066
.
Baird
C. S.
2013
Science Questions with Surprising Answers
.
Barbeau
B.
,
Julienne
K.
,
Carriere
A
&
Gauthier
V.
2005
Dead-end flushing of a distribution system: Short and long-term effects on water quality
.
Journal of Water Supply: Research and Technology - AQUA
54
(
6
),
371
383
.
Barisevičiūtė
R.
,
Maceika
E.
,
Ežerinskis
Ž
,
Šapolaitė
J.
,
Butkus
L.
,
Mažeika
J.
,
Rakauskas
V.
,
Juodis
L.
,
Steponėnas
A.
,
Druteikienė
R.
&
Remeikis
V.
2020
Distribution of radiocarbon in sediments of the cooling pond of RBMK type Ignalina Nuclear Power Plant in Lithuania
.
PloS one
15
(
8
),
e0237605
.
Boxall
J. B.
,
Skipworth
P. J.
&
Saul
A. J.
2003
Aggressive flushing for discolouration event mitigation in water distribution networks
.
Water Supply
3
(
1–2
),
179
186
.
https://doi.org/10.2166/ws.2003.0101
.
Carrière
A.
,
Gauthier
V.
,
Desjardins
R.
&
Barbeau
B.
2005
Evaluation of loose deposits in distribution systems through: unidirectional flushing
.
Journal-American Water Works Association
97
(
9
),
82
92
.
Craun
M. F.
,
Craun
G. F.
,
Calderon
R. L.
&
Beach
M. J.
2006
Waterborne outbreaks reported in the United States
.
Journal of Water and Health
4
(
S2
),
19
30
.
Czajkowski
J.
,
Engel
V.
,
Martinez
C.
,
Mirchi
A.
,
Watkins
D.
,
Sukop
M. C.
&
Hughes
J. D.
2018
Economic impacts of urban flooding in South Florida: potential consequences of managing groundwater to prevent salt water intrusion
.
Science of the Total Environment
621
,
465
478
.
Deuerlein
J.
,
Simpson
A. R.
&
Korth
A.
2014
Flushing planner: a tool for planning and optimization of unidirectional flushing
.
Procedia Engineering
70
,
497
506
.
Edwards
B. D.
&
Evans
K. R.
2002
Saltwater Intrusion in Los Angeles Area Coastal Aquifers: the Marine Connection
.
US Department of the Interior, US Geological Survey
,
Washington DC
.
Edwards
B. D.
,
Ehman
K. D.
,
Ponti
D. J.
,
Reichard
E. G.
,
Tinsley
J. C.
,
Rosenbauer
R. J.
,
Land
M.
,
Lee
H. J.
&
Normark
W. R.
2009
Stratigraphic controls on saltwater intrusion in the Dominguez Gap area in coastal Los Angeles. Earth science in the urban ocean: the southern California continental borderland
.
Geol. Soc. Am. Spec. Pap.
454
,
375
395
.
EPA
2020
Featured Story: Stormwater Runoff
.
Available from: https://www3.epa.gov/region9/water/npdes/stormwater-feature.html (accessed 6 October 2021).
Faneca Sànchez
M.
,
Bashar
K.
,
Janssen
G.
,
Vogels
M.
,
Snel
J.
,
Zhou
Y.
,
Stuurman
R. J.
&
Oude Essink
G.
2015
SWIBANGLA: Managing Saltwater Intrusion Impacts in Coastal Groundwater Systems of Bangladesh
.
Friedman
M.
,
Kirmeyer
G. J.
&
Antoun
E.
2002
Developing and implementing a distribution system flushing program
.
Journal-American Water Works Association
94
(
7
),
48
56
.
Gleeson
T.
,
Wada
Y.
,
Bierkens
M. F.
&
Van Beek
L. P.
2012
Water balance of global aquifers revealed by groundwater footprint
.
Nature
488
(
7410
),
197
200
.
Gopinath
S.
,
Srinivasamoorthy
K.
,
Saravanan
K.
,
Suma
C.
,
Prakash
R.
,
Senthilnathan
D.
,
Chandrasekaran
N.
,
Srinivas
Y.
&
Sarma
V.
2016
Modeling saline water intrusion in Nagapattinam coastal aquifers, Tamilnadu, India
.
Modeling Earth Systems and Environment
2
(
2016
),
2
.
Gray
L. T.
, &
ConocoPhillips Co
1988
Settling pond separation using permeable fabric and weighting. U.S. Patent 4,752,402
.
Hasit
Y. J.
,
DeNadai
A. J.
&
Raucher
R. S.
2004
Cost and Benefit Analysis of Flushing
.
American Water Works Association
.
Haxton
T. B.
&
Walski
T. M.
2009
Modeling a hydraulic response to a contamination event
. In:
World Environmental and Water Resources Congress 2009: Great Rivers
. pp.
1
9
.
Home Advisor
2019
Landscape Cost Guides
.
Available from: https://www.homeadvisor.com/cost/landscape/ (accessed 6 October 2021).
Hrudey
S. E.
&
Hrudey
E. J.
2004
Safe Drinking Water: Lessons From Recent Outbreaks in Affluent Nations
.
IWA Publishing
,
London
.
Khanal
N.
,
Buchberger
S. G.
&
McKenna
S. A.
2006
Distribution system contamination events: exposure, influence, and sensitivity
.
Journal of Water Resources Planning and Management
132
(
4
),
283
292
.
Kowalski
D.
,
Kowalska
B.
,
Hołota
E.
&
Choma
A.
2015
Water quality correction within water distribution system
.
Ecological Chemistry and Engineering S
22
(
3
),
401
410
.
Mann
D. O.
, &
Mann Technology Limited Partnership
1991
Cooling Pond Enhancement. U.S. Patent 5,029,633
.
Mays
L. W.
2000
Water Distribution Systems Handbook
.
McGraw-Hill
,
New York, NY
.
MELCC
2019
Guidelines on Drinking Water Distribution Systems Best Practices
.
Ministère de l'Environnement et de la Lutte contre les changements climatiques, Direction de l'eau potable et des eaux souterraines
.
Merricks
T. C.
,
Cherry
D. S.
,
Zipper
C. E.
,
Currie
R. J.
&
Valenti
T. W.
2007
Coal-mine hollow fill and settling pond influences on headwater streams in southern West Virginia, USA
.
Environmental Monitoring and Assessment
129
(
1–3
),
359
378
.
Mispagel
H.
&
Gray
J. T.
2005
Antibiotic resistance from wastewater oxidation ponds
.
Water Environment Research
77
(
7
),
2996
3002
.
National University of Singapore
2020
Salty Drinking Water May Take A Toll on Education
.
Futurity
.
Oberoi
K.
1994
Distribution flushing program: the benefits and results
. In
Proc. 1994 AWWA Ann. Conf.
,
New York
.
Polychronopolous
M.
,
Dudley
K.
,
Ryan
G.
&
Hearn
J.
2003
Investigation of factors contributing to dirty water events in reticulation systems and evaluation of flushing methods to remove deposited particles
.
Water Supply
3
(
1–2
),
295
306
.
https://doi.org/10.2166/ws.2003.0117
.
Poulin
A.
,
Mailhot
A.
,
Periche
N.
,
Delorme
L.
&
Villeneuve
J. P.
2010
Planning unidirectional flushing operations as a response to drinking water distribution system contamination
.
Journal of Water Resources Planning and Management
136
(
6
),
647
657
.
Quintiliani
C.
,
Marquez-Calvo
O.
,
Alfonso
L.
,
Di Cristo
C.
,
Leopardi
A.
,
Solomatine
D.
&
de Marinis
G.
2019
Multi-objective valve management optimization formulations for water quality enhancement in WDNs
.
Journal of Water Resources Planning and Management
145
(
12
). Available from: https://doi.org/10.1061/(ASCE)WR.1943-5452.0001133
Ramamoorthy
M.
,
Jin
H.
,
Chaisson
A. D.
&
Spitler
J. D.
2001
Optimal sizing of hybrid ground-source heat pump systems that use a cooling pond as a supplemental heat rejecter-A system simulation approach
.
Transactions-American Society of Heating Refrigerating and Air Conditioning Engineers
107
(
1
),
26
38
.
Rasekh
A.
&
Brumbelow
K.
2013
Probabilistic analysis and optimization to characterize critical water distribution system contamination scenarios
.
Journal of Water Resources Planning and Management
139
(
2
),
191
199
.
Rebolledo
M.
,
Chandrasekaran
S.
&
Bartz-Beielstein
T.
2020
Technical Report: Flushing Strategies in Drinking Water Systems
.
arXiv preprint arXiv:2012.13574
.
Rossman
L.
2000
EPANET 2 User's Manual
.
Environmental Protection Agency
.
Ryan
P. J.
,
Harleman
D. R.
&
Stolzenbach
K. D.
1974
Surface heat loss from cooling ponds
.
Water Resources Research
10
(
5
),
930
938
.
Sah
L.
,
Rousseau
D. P.
&
Hooijmans
C. M.
2012
Numerical modelling of waste stabilization ponds: where do we stand?
Water, Air, & Soil Pollution
223
(
6
),
3155
3171
.
Seth
A.
,
Klise
K. A.
,
Siirola
J. D.
,
Haxton
T.
&
Laird
C. D.
2016
Testing contamination source identification methods for water distribution networks
.
Journal of Water Resources Planning and Management
142
(
4
),
04016001
.
Shah
J. B.
,
Lakin
D.
,
Singh
S. P.
,
Raval
S.
&
Grimes
M.
2001
Hydrants flushing improves water quality
.
Water Engineering & Management
. Available from https://www.wqpmag.com/sites/wqp/files/Hydrant%20Flushing%206_01.pdf
Shammas
N. K.
&
Wang
L. K.
2011
Water Supply and Wastewater Removal
.
Wiley, Inc
,
New York, NY
.
Spechler
R. M.
2001
The relation between structure and saltwater intrusion in the Floridian aquifer system, northeastern Florida
.
US Geological Survey Karst Interest Group Proceedings
,
25
29
. Available from: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.569.1089&rep=rep1&type=pdf
Spellman
F. R.
2017
The Drinking Water Handbook
.
CRC Press
,
Boca Raton, FL
.
USDA
.
2020
Ponds – Planning, Design, Construction
.
Viessman
W.
,
Hammer
M. J.
,
Perez
E. M.
&
Chadik
P. A.
1998
Water Supply and Pollution Control
. Pearson.
Vitanage
D.
,
Copelin
C.
,
Karunatilake
N.
&
Vourtsanis
T.
2003
Assessment of the impact of mains cleaning to improve distribution water quality
.
Water Supply
3
(
1–2
),
71
78
.
https://doi.org/10.2166/ws.2003.0088
.
Walski
T.
,
Wu
Z. Y.
,
Hartell
W.
&
Culin
K.
2008
Determining the best way to model distribution flushing
. In:
World Environmental and Water Resources Congress 2008: Ahupua'A
. pp.
1
10
.
Wu
Z. Y.
2015
Optimizing hydrant selections for conventional flushing of water distribution systems
. In:
World Environmental and Water Resources Congress 2015
. pp.
829
837
.
Xie
X.
,
Nachabe
M.
&
Zeng
B.
2015
Optimal scheduling of automatic flushing devices in water distribution system
.
Journal of Water Resources Planning and Management
141
(
6
),
04014081
.
Yusuf
M. A.
&
Abiye
T. A.
2019
Risks of groundwater pollution in the coastal areas of Lagos, southwestern Nigeria
.
Groundwater for Sustainable Development
9
,
100222
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Supplementary data