- Email: [email protected]
Biofouling potential and material reactivity in a simulated water distribution network supplied with stormwater recycled via managed aquifer recharge
Water Research 105 (2016) 110e118
Contents lists available at ScienceDirect
Water Research journal homepage: www.elsevier.com/locate/watres
Biofouling potential and material reactivity in a simulated water distribution network supplied with stormwater recycled via managed aquifer recharge Dennis Gonzalez a, *, Grace Tjandraatmadja b, Karen Barry a, Joanne Vanderzalm a, Anna H. Kaksonen c, Peter Dillon a, Geoff J. Puzon c, Jatinder Sidhu d, Jason Wylie c, Nigel Goodman b, Jason Low b a
CSIRO Land and Water, Private Bag 2, Glen Osmond, SA, 5064, Australia CSIRO Land and Water, CSIRO, Private Bag 10, Clayton South, Vic, 3169, Australia c CSIRO Land and Water, CSIRO, Private Bag 5, Wembley, WA, 6913, Australia d CSIRO Land and Water, CSIRO, GPO Box 2583, Brisbane, Qld, 4001, Australia b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 April 2016 Received in revised form 29 July 2016 Accepted 30 August 2016 Available online 31 August 2016
The injection of stormwater into aquifers for storage and recovery during high water demand periods is a promising technology for augmenting conventional water reserves. Limited information exists regarding the potential impact of aquifer treated stormwater in distribution system infrastructure. This study describes a one year pilot distribution pipe network trial to determine the biofouling potential for cement, copper and polyvinyl chloride pipe materials exposed to stormwater stored in a limestone aquifer compared to an identical drinking water rig. Median alkalinity (123 mg/L) and colour (12 HU) in stormwater was significantly higher than in drinking water (82 mg/L and 1 HU) and pipe discolouration was more evident for stormwater samples. X-ray Diffraction and Fluorescence analyses confirmed this was driven by the presence of iron rich amorphous compounds in more thickly deposited sediments also consistent with significantly higher median levels of iron (~0.56 mg/L) in stormwater compared to drinking water (~0.17 mg/L). Water type did not influence biofilm development as determined by microbial density but faecal indicators were significantly higher for polyvinyl chloride and cement exposed to stormwater. Treatment to remove iron through aeration and filtration would reduce the potential for sediment accumulation. Operational and verification monitoring parameters to manage scaling, corrosion, colour, turbidity and microbial growth in recycled stormwater distribution networks are discussed. Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.
Keywords: Aquifer storage and recovery Robbins test device Biofilm Discoloured water Corrosion Scaling
1. Introduction Increasing pressure on water resources due to urbanisation and climate variability has resulted in growing interest to harvest and use alternative sources of water supply, such as urban stormwater. Aquifers can be used to improve water quality (Page et al., 2010; Vanderzalm et al., 2013), store seasonally available stormwater via managed aquifer recharge (MAR) to balance supply with demand, and potentially provide a reliable alternative water source under urbanisation and a changing climate (Clark et al., 2015) with better economic and energy efficiency than conventional surface
* Corresponding author. E-mail address: [email protected] (D. Gonzalez). http://dx.doi.org/10.1016/j.watres.2016.08.066 0043-1354/Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.
storage solutions (Dillon et al., 2010). Biofilm growth and sediment deposition are common processes that take place in water distribution networks and are influenced by the chemical and microbial characteristics of the water supply, pipe material and flow regimen (Ginige et al., 2011; Lehtola et al., 2004, 2006). Understanding material reactivity and biofilm growth in distribution networks using recycled stormwater is critical for managing operational risks associated with infrastructure and water aesthetics. This study presents the results of an experiment analysing the biofouling potential of stormwater harvested and recycled via a constructed wetland and limestone aquifer on different pipe materials using a pipe rig simulating a distribution network. Polyvinyl chloride (PVC) and mortar lined ductile cast iron are commonly used in distribution networks and copper is commonly used in household plumbing. Surface accumulation of particulate
D. Gonzalez et al. / Water Research 105 (2016) 110e118
iron and manganese through attachment to biofilm and subsequent release through biofilm inactivation and sloughing has been shown to increase turbidity of reticulated water (e.g. by a factor of 5) (Ginige et al., 2011). Microbial growth can affect corrosion of distribution pipe networks (Lee et al., 1980) and biofilm can provide an environment where microbes including pathogenic organisms (e.g. Pseudomonas, Legionella, Campylobacter jejuni, Salmonella sp.) can grow protected from disinfection chemicals (Percival and Walker, 1999; Le Chevallier, 2003; Miller et al., 2015). Currently in Australia, recycled water sources (e.g. stormwater and treated wastewater) are distributed though dedicated third pipe networks. Future adoption of a common distribution network to transport treated water from various origins or blended water from various sources could become more widespread. To date, the impact of stormwater on biofouling, sedimentation, corrosion and scaling potential of different materials used in distribution systems has not been addressed in the literature. The risk profile of stormwater stored and recovered from a limestone aquifer and distributed with no further treatment or disinfection could potentially differ from conventional drinking water supplies in a number of ways. Untreated stormwater quality is characteristically variable and can contain high concentrations of suspended solids, metals (NRMMCeEPHCeNHMRC, 2009), and bioavailable organic carbon (Vanderzalm et al., 2010, 2014). Without disinfection (e.g. chlorination) the potential for biofouling and pathogen occurrence increases and higher levels of bioavailable organic carbon (e.g. > 0.15 mg/L) can promote biofilm growth (Percival and Walker, 1999). Elevated concentrations of dissolved and particulate matter increases potential for sediment deposition and resuspension causing aesthetic quality issues (colour and turbidity). Storage in a limestone aquifer while acting to dampen stormwater quality variability can raise hardness and alkalinity thereby reducing corrosion potential but increasing likelihood of scaling. Quantification of the potential operational, aesthetic and public health risks associated with reticulation of stormwater harvested via MAR is required for effective scheme planning and operational management. This paper describes an experimental study that explored the effects of stormwater recycled via an aquifer in a simulated distribution network to evaluate the impact of the water source on biofilm formation in different pipe materials. The experimental pipe rig used stormwater harvested from an urban catchment in Salisbury, South Australia, that was stored and recovered from a confined limestone aquifer. An identical pipe rig was installed using conventional drinking water supply as the control (hereafter referred to as mains water). Mains water was derived from surface water reservoirs in protected rural catchments and had been treated by coagulation and filtration followed by chlorination. The objectives of this paper were to: (1) Describe differences in water quality, material colour change, sediment composition, and biofilm development (as microbial cell counts) between different water sources. (2) Discuss implications for material reactivity, water aesthetics, biofouling and operational and verification monitoring.
2. Materials and methods 2.1. Rig set up and operation To simulate conditions in water distribution, a buried experimental pipe loop system was constructed at Mawson Lakes, South Australia (34.8047, 138.6024) and operated from July 2012 to
111
January 2014. The installation consisted of two identical pipe rigs; one supplied with aquifer recovered stormwater hereafter referred to as the ‘stormwater rig’, and one supplied with conventionally treated drinking water hereafter referred to as the ‘mains water rig’. Fig. 1 shows a schematic diagram of the rigs. Stormwater was sourced from the storage tank from the Parafield Stormwater Harvesting Scheme in South Australia. This scheme captures stormwater runoff from a 1590 Ha mixed urban residential, commercial and industrial use catchment. The stormwater undergoes detention in two surface storage basins, followed by passage through a horizontal-flow constructed wetland and injection into a confined limestone aquifer for storage, and subsequent recovery into 600 m3 header tanks from where the recycled stormwater is distributed for non-drinking uses (e.g. irrigation, industrial process water) via a dedicated reticulation pipe network (Gonzalez et al., 2015). The Parafield stormwater harvesting scheme uses aquifer storage and recovery (ASR) wells targeting a ~200 m deep confined limestone aquifer with brackish salinity of ~2000 mg/L total dissolved solids (TDS) that is characteristically low in nutrients (e.g. nitrogen <0.05 mg/L and phosphorus <0.02 mg/L) and oxygen (<0.1 mg/L) (Vanderzalm et al., 2010). Recycled stormwater undergoes no further treatment (e.g. filtration or disinfection) following aquifer recovery. Drinking water was sourced from the drinking water supply distribution network which in this part of the metropolitan Adelaide distribution system originates mainly from the Little Para Reservoir surface water catchment and water treatment plant (SA Water, 2014). In each rig, the water supply entered into a 200 L header tank via a float valve and was circulated through a pipe loop made of 36 m of PVC pipe (DN100, PN12) and 6 m of ductile iron cement lined (DICL) pipe (DN100, PN12). The pipe loops were buried 600 mm below ground. Each rig also had a 3.2 m coupon testing section made of PVC (DN150), based on a modified Robbins test device (McCoy et al., 1981), where coupons of selected pipe materials (cement-lined, copper and PVC) were installed. These are shown in Fig. 1 as sections (V3eV4 and V8eV9). Each section could hold 117 coupons measuring 4 cm by 2 cm in size. Coupons were mounted onto plastic, screw-fit holders and oriented in the direction of the flow. This section was housed in a covered pit 900 mm deep. Water in each rig was circulated using a suction pump (Lowara ITT CEA 370/3/A-V 240/380e415 50) controlled by a Vasco variable speed drive coupled to an Omni P1600-100 pressure sensor and set to a constant pressure of 200 kPa (maximum achievable with rig configuration). The flow rate was measured using an Endress Hauser Proline Promag 10 W flow meter (pulse output: þ0.5% of reading þ 2 mm/s). Running the pumps at maximum speed (50 Hz) the flow rate was 140e145 L/min, which equates to a mean flow velocity of 0.14 m/s. Total rig volume (including header tank) was determined experimentally as 561 L by running a slug test using a salt (NaCl) tracer (Tjandraatmadja et al., 2014). Turnover of water within each rig was controlled using a solenoid activated irrigation valve on the pipe return line (V5 and V10 in Fig. 1) and an irrigation timer. The valve was set at ~200 kPa back pressure with a bleed rate of ~1 L/ min. It would open for 16 h and shut for 8 h every day, meaning that 960 L was bled (and replaced with fresh water) from each rig every 24 h. Water temperature, pH, electrical conductivity (EC) and oxidation-reduction potential (ORP) were continuously monitored in the pipe lines using sensors (T-type mineral insulated metal sheathed stainless steel thermocouples 3 mm e TC Measurement and Control (±1%), SensoreX pH electrodes e flat surface CPVC (±10%), Thermo Fisher Scientific Alpha Cond 500 connected to a Thermo Fisher Scientific stainless steel conductivity electrode (±5%) and SensoreX ORP electrode sensors (±10%), respectively).
112
D. Gonzalez et al. / Water Research 105 (2016) 110e118
Fig. 1. Schematic representation of experimental rig for the evaluation of biofouling potential and material reactivity due to reticulation of recycled stormwater (from Tjandraatmadja et al., 2014).
The pH and ORP sensor signals were amplified using Heaston Electronics self-powered pre amplifiers (PHAMP-A30). The sensors were installed in the section preceding the D150 mm pipe length (above V3 and V8 on Fig. 1). The sensor and flow rate data were recorded at 5 min intervals using a DT85 DataTaker logger with remote connection via a wireless modem router interfaced with deX DataTaker software and a dynamic domain name system hosting service.
2.2. Sample collection Water and pipe coupons were sampled from the rigs according to Table 1. Coupons were sampled on 10 instances between May 2012 and January 2014. Their exposure times ranged from 5 to 48 weeks and covered all four seasons as shown in Table 1. On some occasions only one type of coupon was sampled and there was some variance of exposure times sampled on some occasions.
D. Gonzalez et al. / Water Research 105 (2016) 110e118
113
Table 1 Coupon sampling events. Sample event
Date in
Date out
Exposure time (weeks)
Exposure season/s
Comments
1 2 3 4
23/05/2012 20/08/2012 25/09/2012 25/09/2012 21/08/2012 21/08/2012 25/09/2012 25/09/2012 21/08/2012 e
20/08/2012 25/09/2012 29/10/2012 30/01/2013 30/01/2013 13/03/2013 13/03/2013 30/04/2013 30/04/2013 e
13 5 5 18 23 29 24 31 36 e
AutumneWinter WintereSpring Spring SpringeSummer WintereSummer WintereAutumn SpringeAutumn SpringeAutumn WintereAutumn e
PVC only cement, PVC, Cu cement only cement PVC, Cu PVC, Cu cement cement PVC, Cu water quality 22/5/2013
25/09/2012 21/08/2012 25/09/2012 21/08/2012 30/01/2013 1/10/2013
11/06/2013 11/06/2013 22/07/2013 22/07/2013 1/10/2013 8/01/2014
37 42 43 48 35 14
SpringeWinter WintereWinter SpringeWinter WintereWinter SummereSpring SpringeSummer
cement PVC, Cu cement PVC, Cu cement, PVC, Cu cement, PVC, Cu
5 6 e 7 8 9 10
At each sampling event, duplicate coupons of each pipe material were removed for surface characterisation and microbial analysis. Coupons were stored in 50 mL centrifuge tubes with 25 mL of source water and kept below 4 C for transportation to the laboratory, together with samples of respective source waters (mains and stormwater). Microbial analyses were conducted within 24 h of sampling.
Averages of 5 readings across coupon surfaces were taken. The elemental composition of sediment on coupons was determined using micro X-ray Diffraction (micro-XRD) using a Bruker GADDS micro diffractometer, and X-ray Fluorescence (XRF) using a Key Master Tracer III-V portable light-elemental analyser.
2.3. Water quality analysis
Each of the coupons was rinsed (in turn) with 1 mL of filter sterilised (0.2 mm syringe filter) source water using a 1 mL pipette and sterile cotton buds were used to swab coupons surfaces. The cotton buds were placed in individual sterile tubes containing 5 mL of sterile filtered source water, sonicated (Branson 1200) for 5 min and then twisted vigorously to release attached biofilm. Samples were put on ice for transportation to the laboratory for analysis along with the 2 500 mL source water samples. The swabbed surface areas of each coupon where also measured. The resulting suspensions from the coupons were mixed well prior to dilution (1:10), then stained (SYBR green). Total microbial cell counts were determined using a Cell Lab QuantaTM SC Beckman Coulter flow cytometer (equipped with an air-cooled 15 mW argon ion laser, emitting at a fixed wavelength of 488 nm). Logarithmic signals were collected (Hoefel et al., 2003) and data analysed using Cell Lab Quanta® SC MPL Analysis software. Thermotolerant coliforms were assessed by plating onto chromocult agar (Merck, catalogue number 1.10426.0500) and enumerating new colonies formed following incubation at 45 C for 24 h after initial incubation at 37 C for 24 h. Statistics were calculated using SigmaPlot 13.0 (Systat Software Inc.).
At each coupon sampling event (and on two other occasions, see Table 1) water from the return lines of the rigs was sampled for analyses. Water samples were maintained below 4 C and delivered to the laboratory within 24 h according to procedures and storage times according to methods recommended in the Standard Methods for the Examination of Wastewater (APHA, 2005). Alkalinity (as HCO 3 ) was determined by automated acidimetric titration; chloride, total nitrogen (TN) and total phosphorus (TP) were determined by colorimetric flow analysis; determination of metals by ICP spectrometry by ICP2; elemental analysis by ICP/MS; turbidity by Nephelometric measurement; true colour by spectrophotometric measurement; faecal coliforms/Escherichia coli by membrane filtration; total organic carbon (TOC) by flame ionisation; and biodegradable dissolved organic carbon (BDOC) was calculated from the change in DOC during incubation with a bacterial inoculum. Analyses were conducted by the Australian Water Quality Centre laboratory (ISO 9001 and ISO17025 certified and NATA accredited). Additionally, conductivity, temperature, pH and dissolved oxygen of rig water were also measured in situ at each sample collection event using a 90FL-mV field analyser and probes in a flow-through cell and residual chlorine was measured using a HACH DR/890 portable colorimeter. Mineral saturation indices were calculated with PHREEQC (Parkhurst and Appelo, 1999). 2.4. Surface morphology and sediment elemental composition Surface morphology was determined using an Olympus SZX10 optical microscope (magnification 126) with Dyno Capture 2.0 software, and a 3D confocal laser microscope (Olympus OLS4000) with MPFLN10, 20 and 50 objective lens under XYZ scanning mode. Colour was determined using the CIE L*a*b* colour scale with a Konica Minolta CR-300 Chroma meter colorimeter and a white tile reference standard. Total colour difference (DE) was calculated using the CIE 1976 Colour-Difference Formula (Robertson, 1977).
2.5. Biofilm analysis
3. Results and discussion 3.1. Operating conditions A constant flow rate of 145 L/min at a pressure of 200 kPa for both rigs was the maximum that could be maintained throughout the experiment using the pumps specified. This translated to a water velocity in the coupon section (DN150) of 0.14 m/s, which was 30e70% less than typical velocities (0.2e0.5 m/s) for drinking water distribution systems (Manuel et al., 2007). Various studies documented by Liu and Tay (2002) suggest the effect of higher shear force results in higher density, smoother and thinner biofilm layers whereas when formed under lower shear stress environments biofilms are thicker, less dense and more prone to sloughing. The flow conditions under which the rigs were operated in the current study represent conditions under which biofilms (and
114
D. Gonzalez et al. / Water Research 105 (2016) 110e118
deposited sediments) could be unstable enough to allow sloughing once a critical thickness was reached. Water temperature in the rigs was seasonally and diurnally affected. Summer mean daytime temperature was 30 ± 3 C while winter daytime mean temperature was 19 ± 2 C. Ambient air temperatures in 2013 were 1e45 C so burying the pipework provided a degree of insulation although maxima varied above typical values of <20 C for Adelaide mains distribution system (Tjandraatmadja et al., 2014). Improvements to the design could include insulation on the pit lid for operation within a narrower and lower temperature range. 3.2. Discolouration There was a general rapid increase in discolouration (determined as total colour difference, DE) for all pipe materials in both rigs within the first 5 weeks (Fig. 2). Further discolouration of cement coupons exposed to stormwater occurred after 20 weeks and copper and PVC coupons after 30 weeks. Colour differences are reflected in photographic time series of coupons where darkening of coupons exposed to stormwater is more visible than for those exposed to mains water (Fig. 3). The variability in discolouration of cement and PVC coupons in mains and stormwater over time was indicative of build-up and subsequent sloughing of surface deposits (Fig. 2). Notable differences were observed in copper exposed to stormwater where a transition from yellow to green was observed (reduction in b*) particularly after 36 weeks and for PVC and cement exposed to stormwater where marked darkening (reduction in L*) was observed after 35 weeks (Fig. 4). This suggests that corrosion mechanisms, biofilm and sediment composition differed according to source water. Cement coupons in mains water retained a red hue similar to the original colour for up to 35 weeks before discolouring towards the green coordinate (from a* ¼ þ5.28 ± 0.13 to a* ¼ þ1.36 ± 0.66). Copper coupons in mains and particularly stormwater also discoloured towards a green hue (reduction in a*, Fig. 4) due to oxidation and formation of the characteristic green patina (Fig. 3). PVC coupons discoloured markedly more in stormwater than in mains water (Fig. 2) and the hues also differed notably after 35 weeks (Fig. 4). In stormwater PVC coupons discoloured toward red (increase in a*) whereas in mains water coupons discoloured toward green (reduction in a*). The total colour difference of PVC coupons exposed in mains water was small (DE 18 units) and slight yellowing was observed, due to sediment attachment. In stormwater, coupons displayed a larger colour difference (DE 61
Fig. 2. Total colour difference (DE) of copper, cement and PVC coupons exposed to stormwater (sw) and mains water (mains) over time.
units) with reduced lightness (L*), and a shift towards red and yellow coordinates. This confirms that cement and copper are more reactive than PVC and more prone to biofouling (Niquette et al., 2000). 3.3. Elemental surface and sediment analyses XRF analysis indicated sediment deposition on PVC and cement coupons exposed to both source waters consisted of mainly iron rich amorphous compounds. There were greater intensities of spectra associated with PVC manufacturing additives on coupons exposed to mains water indicating the surface layer of cement and PVC coupons exposed to stormwater was substantially thicker. This is consistent with significantly higher median concentration of total Fe (Kruskal-Wallis test, P ¼ 0.04) in the stormwater (0.45 mg/ L) compared to mains water (0.09 mg/L). XRF analysis also showed high peaks for copper on cement coupons exposed to mains water, but not for stormwater coupons. This was attributed to the significantly higher (Kruskal-Wallis test, P ¼ 0.02) median concentration of total copper in the mains water (0.18 mg/L) compared to stormwater (0.09 mg/L). XRD traces indicated the presence of copper oxidation products, cuprite and copper oxide on surface layers of copper coupons exposed to both source waters with thin deposits comprising Cu and Fe also indicated through fluorescence analysis. 3.4. Biofilm microbial density Biofilm microbial density (live and dead organisms) determined through total cell counts were in the order of 105e107 cells/cm2 in both rigs and did not appear to increase continuously as a function of exposure time (Fig. 5). With samples pooled across exposure times, there were no significant differences in log10 transformed mean cell numbers between mains water and stormwater for the different pipe materials (ANOVA test, P ¼ 0.74). There were however differences in the number of thermotolerant coliforms between source waters. For copper coupons exposed to mains water, thermotolerant coliforms were not detected but in stormwater were detected twice (~36 cells/cm2) from 14 samples. For stormwater, median thermotolerant coliforms counts for both cement (88 cells/cm2) and PVC (85 cell/cm2) were significantly higher than for mains water (medians of zero) (Kruskal-Wallis test, P < 0.001). There was no chlorination treatment of stormwater so the residual in the stormwater rig throughout the experiment was zero. The mean chlorine residual in the mains water rig was 0.02 mg/L (±0.02) which was an order of magnitude lower than averages across the Adelaide mains distribution network (SA Water, 2014). This low level of disinfection was unlikely to limit biofilm growth especially at the temperatures observed in the rigs (frequently >20 C). Biodegradable dissolved organic carbon (BDOC) is a microbial food source and can stimulate biological growth. Mean BDOC concentrations in mains and stormwater samples were comparable (0.9 and 0.7 mg/L respectively, Table 2) and both were 5e6 times higher than the recommended maximum of 0.15 mg/L to limit biofilm growth (Escobar et al., 2001). Mean total cell counts for each coupon type and source water were variable with respect to exposure time (Fig. 5). This was most evident for PVC exposed to stormwater where the difference between lowest and highest mean counts was factor of 132 (3 104 and 4 106 cell/cm2 respectively) with the minimum observed at the longest exposure time (35 weeks). The lowest overall variance (factor of 5) was for mains water PVC where counts remained with the range of 2e7 105 cells/cm2 (lowest recorded at 18 weeks exposure). Differences between minimum and maximum cell counts for other material/water types ranged from a factor of
D. Gonzalez et al. / Water Research 105 (2016) 110e118
115
Fig. 3. Appearance of cement (top), PVC (middle) and copper (bottom) coupons after exposure to mains water and stormwater (image from Tjandraatmadja et al., 2014).
Fig. 4. Mean colour changes (from 5 readings of a single coupon at each time step) of cement, copper and PVC coupons exposed to stormwater and mains water (error bars represent one standard deviation, blank is unexposed coupon).
116
D. Gonzalez et al. / Water Research 105 (2016) 110e118
1.E+08
Total cell count (cells/cm 2)
1.E+07 Cu (MW) Cu (SW)
1.E+06
Cement (MW) Cement (SW)
1.E+05
PVC (MW) PVC (SW)
1.E+04
1.E+03 0
5
10
15
20
25
30
35
40
45
50
Exposure (weeks) Fig. 5. Mean total cell counts (n ¼ 2) on copper, cement and PVC coupons exposed to mains water (MW) and stormwater (SW) determined by flow cytometry. ©Copyright CSIRO Australia, 2016 CSIRO.
Table 2 The quality of water sampled from rigs during coupons collection during this study and quality of treated drinking water from Little Para water treatment plant and recovered stormwater based on previously published literature. Parameter
Mains rig
(mg/L unless otherwise indicated)
Mean (SD)
Stormwater rig
pH (pH units) DO TDS Turbidity (NTU) True colour (HU) Alkalinity SICalcite Cl residual free Ca Mg SO4 Na K TN TP TOC BDOC As (total) Fe (total) Mn (total) Cu (total) Escherichia coli (cfu/100 mL) Faecal coliforms (cfu/100 mL)
7.5 (0.4) 12 6.3 (2.1) 8 287 (69.7) 12 2.5 (3.8) 10 1.6 (0.8) 12 74 (16) 12 0.7 (0.3) 12 0.03 (0.007) 5 26 (2.5) 12 11 (3.7) 12 40 (7.5) 12 51 (11) 12 4.9 (1.6) 12 0.449 (0.096) 11 0.013 (0.007) 12 3.2 (1.3) 12 0.7 (0.4) 12 0.0005 (0.0001) 10 0.165 (0.21) 12 0.004 (0.003) 12 0.288 (0.227) 12 0 (0) 8 0 (0) 8
n
Mean (SD)
n
7.4 (0.4) 11 5.7 (1.9) 8 326 (166) 11 9.7 (14.9) 11 18 (26) 11 136 (45) 11 0.5 (0.7) 12 0 (0) 7 38 (12) 11 12 (6.5) 11 34 (26) 11 60 (50) 11 4.4 (1.2) 11 0.335 (0.15) 11 0.059 (0.08) 11 2.6 (1.7) 11 0.9 (1.2) 11 0.004 (0.006) 11 0.562 (0.41) 10 0.036 (0.033) 10 0.115 (0.10) 10 2 (2) 9 2 (2) 9
Little para WTP Mean
na
7.4 265 e 356 60 0.16 265 1 265 69 60 0.8 60 e 25 60 16 60 47 60 75 60 4.1 60 0.398 60 0.005 59 e e 0.0009 60 0.0083 134 0.00027 134 0.0167 59 e e
Recovered stormwater Median
nb
7.3 15 0.14 15 290 15 1.5 15 9 15 148 15 0.3 15 e 37 15 15 15 29 15 45 15 4.0 15 0.220 15 0.023 15 2.3 15 0.8 15 0.002 4 0.376 15 0.019 15 0.0006 15 14 15 14 15
SD ¼ standard deviation; n ¼ number of samples. a Little Para Water Treatment Plant (WTP) treated drinking water quality summary 2005e2010 (Page et al., 2013). b Parafield aquifer recovered stormwater (Page et al., 2013).
15e27. Cell counts were therefore not dependent on exposure time and could be driven by cycles of build-up and subsequent sloughing of biofilm through flow induced friction once a certain level of biofilm was reached (Ginige et al., 2011). 3.5. Implications for material reactivity and biofouling High alkalinity (e.g. >200 mg/L) is conducive to increased scaling (NHMRCeNRMMC, 2011). However, mean alkalinity in the stormwater rig was 136 ± 45 mg/L (Table 2) and therefore scaling would generally be unlikely. pH for both mains and stormwater rigs also remained within target criteria (6.5e8.5) for prevention of corrosion and scaling (NHMRCeNRMMC, 2011).
Compared to mains water, stormwater presented greater risks in terms of water aesthetics. On average, water in the stormwater rig was a factor of 4 higher in turbidity, a factor of 3 higher in Fe and a factor of 11 higher in colour than water in the mains rig and exceeded water aesthetic criteria (NHMRCeNRMMC, 2011) of 5 NTU, 0.3 mg/L, and 15 HU respectively (Table 2). The presence of high Fe and turbidity can also affect the efficacy of disinfectants e.g. chlorine. Iron precipitation can occur when soluble iron in anoxic water, such as deep groundwater, is subjected to aeration e.g. through reticulation or splash entry to header tanks. Insoluble iron (oxy) hydroxides or oxides can contribute to sedimentation and pipe scaling. Iron removal through aeration and filtration prior to reticulation may therefore be warranted to mitigate these risks.
D. Gonzalez et al. / Water Research 105 (2016) 110e118
Total microbial cells counts did not differ significantly between the two water sources and this was ascribed mainly to the low chlorine residual (0.03 ± 0.007 mg/L) of the mains water circulating in the rig but also the relatively high mean levels of BDOC (>0.7 mg/ L) in both rigs. However, thermotolerant coliform numbers indicated biofilms on coupons exposed to stormwater were subject to greater levels of faecal contamination. There is evidence of the potential for BDOC removal in aquifers as the median for injected stormwater at the Parafield ASR site was 2.1 mg/L and on recovery was about half (0.8 mg/L) (Page et al., 2013). Sufficient aquifer storage time may reduce labile organic carbon concentrations to levels that would reduce biofilm growth potential in the distribution pipelines (Vanderzalm et al., 2010) but further investigation to determine rates and required storage times would be required. Australian health-based log-removal targets for open space irrigation with stormwater can be achieved through simple exposure control e.g. restricted public access while irrigating such that no disinfection is required (NRMMC-EPHC-AHMC, 2006). This study highlights the importance of monitoring of key parameters to manage operational and aesthetic risks for stormwater recycling and distribution. Verification monitoring should include Fe and Mn, BDOC and faecal indicators (e.g. E. coli, thermotolerant coliforms). Operational monitoring should include easily or continuously monitored surrogate parameters including colour and turbidity (appearance and sedimentation), EC (indicator for any solute), pH (important for scaling and metal corrosion), and UV/Vis for total organic carbon (to indicate potential levels of BDOC). 4. Conclusions This study examined the impacts of recycled stormwater on cement, PVC and copper distribution pipes using recirculating pilot rigs. Conventional drinking water supply was used as the control. (1) Following storage in a carbonate aquifer, stormwater had about twice the alkalinity, and more than three times the turbidity and iron compared to the control rig running treated drinking (mains) water. Pipe discolouration differed between stormwater and mains water indicating differences in corrosion rates, sediment and biofilm composition and this was supported by X-ray Fluorescence and Diffraction analyses of pipe surfaces. Biofilm determined by total microbial cell densities were similar between rigs for the different pipe materials. However, faecal indicators were higher in stormwater. (2) Distribution of recycled stormwater has the potential to result in greater corrosion and sediment deposition than distribution of conventional drinking water. Operational monitoring should include colour and turbidity to manage water aesthetics and sedimentation potential, electrical conductivity as an indicator for changes in solute concentration, pH to manage scaling and metal corrosion, and UV/ Vis for total organic carbon as an indicator for labile organic carbon important for limiting microbial growth. Acknowledgements The authors acknowledge support of the partners to the Managed Aquifer Recharge and Stormwater Use Options research project. These are the National Water Commission through the Raising National Water Standards Program, the SA Government through the Goyder Institute for Water Research, CSIRO Water for a Healthy Country Flagship, City of Salisbury, the Adelaide and Mt Lofty Ranges Natural Resources Management Board and former United Water International.
117
References APHA, 2005. Standard Methods for the Examination of Water and Wastewater. APHA-WEF-AWWA, Washington. Clark, R., Gonzalez, D., Dillon, P., Charles, S., Cresswell, D., Naumann, B., 2015. Reliability of water supply from stormwater harvesting and managed aquifer recharge with a brackish aquifer in an urbanising catchment and changing climate. Environ. Model. Softw. 72, 117e125. Dillon, P., Toze, S., Page, D., Vanderzalm, J., Bekele, E., Sidhu, J., Rinck-Pfeiffer, S., 2010. Managed aquifer recharge: rediscovering nature as a leading edge technology. Water Sci. Technol. 62 (10), 2338. Escobar, I.C., Randall, A.A., Taylor, J.S., 2001. Bacterial growth in distribution systems: effect of assimilable organic carbon and biodegradable dissolved organic carbon. Environ. Sci. Technol. 35 (17), 3442e3447. Ginige, M.P., Wylie, J., Plumb, J., 2011. Influence of biofilms on iron and manganese deposition in drinking water distribution systems. Biofouling 27 (2), 151e163. Gonzalez, D., Page, D., Vanderzalm, J., Dillon, P., 2015. Setting water quality trigger levels for the operation and management of a MAR system in Parafield, South Australia. J. Hydrologic Eng. 20 (3), B5014001. Hoefel, D., Grooby, W.L., Monis, P.T., Andrews, S., Saint, C.P., 2003. Enumeration of water-borne bacteria using viability assays and flow cytometry: a comparison to culture-based techniques. J. Microbiol. Methods 55 (3), 585e597. Le Chevallier, M.W., 2003. Conditions favouring coliform and HPC bacterial growth in drinking-water and on water contact surfaces. In: Bartram, J., Cotruvo, J., Exner, M., Fricker, C., Glasmacher, A. (Eds.), Heterotrophic Plate Counts and Drinking-water Safety. World Health Organisation, London, pp. 177e196. Lee, S.H., O'Connor, J.T., Banerji, S.K., 1980. Biologically mediated corrosion and its effects on water quality in distribution systems. J. Am. Water Works Assoc. 82, 636e645. €nen, M.M., Kekki, T.K., Laine, O., Hirvonen, A., Lehtola, M.J., Miettinen, I.T., Keina Vartiainen, T., Martikainen, P.J., 2004. Microbiology, chemistry and biofilm development in a pilot drinking water distribution system with copper and plastic pipes. Water Res. 38 (17), 3769e3779. Lehtola, M.J., Laxander, M., Miettinen, I.T., Hirvonen, A., Vartiainen, T., Martikainen, P.J., 2006. The effects of changing water flow velocity on the formation of biofilms and water quality in pilot distribution system consisting of copper or polyethylene pipes. Water Res. 40 (11), 2151e2160. Liu, Y., Tay, J.H., 2002. The essential role of hydrodynamic shear force in the formation of biofilm and granular sludge. Water Res. 36 (7), 1653e1665. Manuel, C.M., Nunes, O.C., Melo, L.F., 2007. Dynamics of drinking water biofilm in flow/non-flow conditions. Water Res. 41 (3), 551e562. McCoy, W.F., Bryers, J.D., Robbins, J., Costerton, J.W., 1981. Observations of fouling biofilm formation. Can. J. Microbiol. 27 (9), 910e917. Miller, H.C., Wylie, J., Dejean, G., Kaksonen, A.H., Sutton, D., Braun, K., Puzon, G.J., 2015. Reduced efficiency of chlorine disinfection of Naegleria fowleri in a drinking water distribution biofilm. Environ. Sci. Technol. 49 (18), 11125e11131. Niquette, P., Servais, P., Savoir, R., 2000. Impacts of pipe materials on densities of fixed bacterial biomass in a drinking water distribution system. Water Res. 34, 1952e1956. NHMRCeNRMMC, 2011. Australian Drinking Water Guidelines. National Health and Medical Research Council and Natural Resource Management Ministerial Council, Canberra. NRMMCeEPHCeAHMC, 2006. Australian Guidelines for Water Recycling: Managing Health and Environmental Risks (Phase 1). NRMMCeEPHCeAHMC, Australia. NRMMCeEPHCeNHMRC, 2009. Australian Guidelines for Water Recycling: Managing Health and Environmental Risks (Phase 2) Stormwater Harvesting and Reuse. NRMMCeEPHCeNHMRC, Australia. Page, D., Dillon, P., Toze, S., Sidhu, J.P.S., 2010. Characterising aquifer treatment for pathogens in managed aquifer recharge. Water Sci. Technol. 62 (9), 2009. Page, D., Gonzalez, D., Dillon, P., Vanderzalm, J., Vadakattu, G., Toze, S., Sidhu, J., Miotlinski, K., Torkzaban, S., Barry, K., 2013. Managed Aquifer Recharge Stormwater Use Options: Public Health and Environmental Risk Assessment Final Report. Goyder Institute for Water Research Technical Report 13/17. http:// goyderinstitute.org/index.php?id¼20. Parkhurst and Appelo, 1999. User's Guide to PHREEQC (Version 2) A Computer Program for Speciation, Reaction Path, Advective Transport, and Inverse Geochemical Calculations. Report 99e4259. U.S. Geological Survey Water Resources Investigation. Percival, S.L., Walker, J.T., 1999. Potable water and biofilm: a review of the public health implications. Biofouling 14 (2), 99e115. Robertson, A.R., 1977. The CIE 1976 color-difference formulae. Color Res. Appl. 2 (1), 7e11. SA Water, 2014. SA Water Drinking Water Quality Report 2013-14. SA Water. Government of South Australia. Accessed [Sep 2015]. https://www.sawater.com.au. Tjandraatmadja, G., Kaksonen, A.H., Gonzalez, D., Barry, K., Vanderzalm, J.V., Puzon, G., Sidhu, J., Wylie, J., Goodman, N., 2014. Managed Aquifer Recharge and Stormwater Use Options: Investigation of Stormwater Impact on Water Quality and Distribution Infrastructure. Goyder Institute for Water Research Technical Report Series No. 14/8. Vanderzalm, J.L., Page, D.W., Barry, K.E., Dillon, P.J., 2010. A comparison of the geochemical response to different managed aquifer recharge operations for injection of urban stormwater in a carbonate aquifer. Appl. Geochem. 25 (9), 1350e1360. Vanderzalm, J.L., Page, D.W., Barry, K.E., Dillon, P.J., 2013. Application of a
118
D. Gonzalez et al. / Water Research 105 (2016) 110e118
probabilistic modelling approach for evaluation of nitrogen, phosphorus and organic carbon removal efficiency during four successive cycles of aquifer storage and recovery (ASR) in an anoxic carbonate aquifer. Water Res. 47 (7), 2177e2189. Vanderzalm, J., Page, D., Gonzalez, D., Barry, K., Toze, S., Bartak, R., Shisong, Q.,
Weiping, W., Dillon, P., Lim, M.H., 2014. Managed Aquifer Recharge and Stormwater Use Options: Satellite Sites Stormwater Quality Monitoring and Treatment Requirements Report. Goyder Institute for Water Research Technical Report 14/10.
Contents lists available at ScienceDirect
Water Research journal homepage: www.elsevier.com/locate/watres
Biofouling potential and material reactivity in a simulated water distribution network supplied with stormwater recycled via managed aquifer recharge Dennis Gonzalez a, *, Grace Tjandraatmadja b, Karen Barry a, Joanne Vanderzalm a, Anna H. Kaksonen c, Peter Dillon a, Geoff J. Puzon c, Jatinder Sidhu d, Jason Wylie c, Nigel Goodman b, Jason Low b a
CSIRO Land and Water, Private Bag 2, Glen Osmond, SA, 5064, Australia CSIRO Land and Water, CSIRO, Private Bag 10, Clayton South, Vic, 3169, Australia c CSIRO Land and Water, CSIRO, Private Bag 5, Wembley, WA, 6913, Australia d CSIRO Land and Water, CSIRO, GPO Box 2583, Brisbane, Qld, 4001, Australia b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 April 2016 Received in revised form 29 July 2016 Accepted 30 August 2016 Available online 31 August 2016
The injection of stormwater into aquifers for storage and recovery during high water demand periods is a promising technology for augmenting conventional water reserves. Limited information exists regarding the potential impact of aquifer treated stormwater in distribution system infrastructure. This study describes a one year pilot distribution pipe network trial to determine the biofouling potential for cement, copper and polyvinyl chloride pipe materials exposed to stormwater stored in a limestone aquifer compared to an identical drinking water rig. Median alkalinity (123 mg/L) and colour (12 HU) in stormwater was significantly higher than in drinking water (82 mg/L and 1 HU) and pipe discolouration was more evident for stormwater samples. X-ray Diffraction and Fluorescence analyses confirmed this was driven by the presence of iron rich amorphous compounds in more thickly deposited sediments also consistent with significantly higher median levels of iron (~0.56 mg/L) in stormwater compared to drinking water (~0.17 mg/L). Water type did not influence biofilm development as determined by microbial density but faecal indicators were significantly higher for polyvinyl chloride and cement exposed to stormwater. Treatment to remove iron through aeration and filtration would reduce the potential for sediment accumulation. Operational and verification monitoring parameters to manage scaling, corrosion, colour, turbidity and microbial growth in recycled stormwater distribution networks are discussed. Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.
Keywords: Aquifer storage and recovery Robbins test device Biofilm Discoloured water Corrosion Scaling
1. Introduction Increasing pressure on water resources due to urbanisation and climate variability has resulted in growing interest to harvest and use alternative sources of water supply, such as urban stormwater. Aquifers can be used to improve water quality (Page et al., 2010; Vanderzalm et al., 2013), store seasonally available stormwater via managed aquifer recharge (MAR) to balance supply with demand, and potentially provide a reliable alternative water source under urbanisation and a changing climate (Clark et al., 2015) with better economic and energy efficiency than conventional surface
* Corresponding author. E-mail address: [email protected] (D. Gonzalez). http://dx.doi.org/10.1016/j.watres.2016.08.066 0043-1354/Crown Copyright © 2016 Published by Elsevier Ltd. All rights reserved.
storage solutions (Dillon et al., 2010). Biofilm growth and sediment deposition are common processes that take place in water distribution networks and are influenced by the chemical and microbial characteristics of the water supply, pipe material and flow regimen (Ginige et al., 2011; Lehtola et al., 2004, 2006). Understanding material reactivity and biofilm growth in distribution networks using recycled stormwater is critical for managing operational risks associated with infrastructure and water aesthetics. This study presents the results of an experiment analysing the biofouling potential of stormwater harvested and recycled via a constructed wetland and limestone aquifer on different pipe materials using a pipe rig simulating a distribution network. Polyvinyl chloride (PVC) and mortar lined ductile cast iron are commonly used in distribution networks and copper is commonly used in household plumbing. Surface accumulation of particulate
D. Gonzalez et al. / Water Research 105 (2016) 110e118
iron and manganese through attachment to biofilm and subsequent release through biofilm inactivation and sloughing has been shown to increase turbidity of reticulated water (e.g. by a factor of 5) (Ginige et al., 2011). Microbial growth can affect corrosion of distribution pipe networks (Lee et al., 1980) and biofilm can provide an environment where microbes including pathogenic organisms (e.g. Pseudomonas, Legionella, Campylobacter jejuni, Salmonella sp.) can grow protected from disinfection chemicals (Percival and Walker, 1999; Le Chevallier, 2003; Miller et al., 2015). Currently in Australia, recycled water sources (e.g. stormwater and treated wastewater) are distributed though dedicated third pipe networks. Future adoption of a common distribution network to transport treated water from various origins or blended water from various sources could become more widespread. To date, the impact of stormwater on biofouling, sedimentation, corrosion and scaling potential of different materials used in distribution systems has not been addressed in the literature. The risk profile of stormwater stored and recovered from a limestone aquifer and distributed with no further treatment or disinfection could potentially differ from conventional drinking water supplies in a number of ways. Untreated stormwater quality is characteristically variable and can contain high concentrations of suspended solids, metals (NRMMCeEPHCeNHMRC, 2009), and bioavailable organic carbon (Vanderzalm et al., 2010, 2014). Without disinfection (e.g. chlorination) the potential for biofouling and pathogen occurrence increases and higher levels of bioavailable organic carbon (e.g. > 0.15 mg/L) can promote biofilm growth (Percival and Walker, 1999). Elevated concentrations of dissolved and particulate matter increases potential for sediment deposition and resuspension causing aesthetic quality issues (colour and turbidity). Storage in a limestone aquifer while acting to dampen stormwater quality variability can raise hardness and alkalinity thereby reducing corrosion potential but increasing likelihood of scaling. Quantification of the potential operational, aesthetic and public health risks associated with reticulation of stormwater harvested via MAR is required for effective scheme planning and operational management. This paper describes an experimental study that explored the effects of stormwater recycled via an aquifer in a simulated distribution network to evaluate the impact of the water source on biofilm formation in different pipe materials. The experimental pipe rig used stormwater harvested from an urban catchment in Salisbury, South Australia, that was stored and recovered from a confined limestone aquifer. An identical pipe rig was installed using conventional drinking water supply as the control (hereafter referred to as mains water). Mains water was derived from surface water reservoirs in protected rural catchments and had been treated by coagulation and filtration followed by chlorination. The objectives of this paper were to: (1) Describe differences in water quality, material colour change, sediment composition, and biofilm development (as microbial cell counts) between different water sources. (2) Discuss implications for material reactivity, water aesthetics, biofouling and operational and verification monitoring.
2. Materials and methods 2.1. Rig set up and operation To simulate conditions in water distribution, a buried experimental pipe loop system was constructed at Mawson Lakes, South Australia (34.8047, 138.6024) and operated from July 2012 to
111
January 2014. The installation consisted of two identical pipe rigs; one supplied with aquifer recovered stormwater hereafter referred to as the ‘stormwater rig’, and one supplied with conventionally treated drinking water hereafter referred to as the ‘mains water rig’. Fig. 1 shows a schematic diagram of the rigs. Stormwater was sourced from the storage tank from the Parafield Stormwater Harvesting Scheme in South Australia. This scheme captures stormwater runoff from a 1590 Ha mixed urban residential, commercial and industrial use catchment. The stormwater undergoes detention in two surface storage basins, followed by passage through a horizontal-flow constructed wetland and injection into a confined limestone aquifer for storage, and subsequent recovery into 600 m3 header tanks from where the recycled stormwater is distributed for non-drinking uses (e.g. irrigation, industrial process water) via a dedicated reticulation pipe network (Gonzalez et al., 2015). The Parafield stormwater harvesting scheme uses aquifer storage and recovery (ASR) wells targeting a ~200 m deep confined limestone aquifer with brackish salinity of ~2000 mg/L total dissolved solids (TDS) that is characteristically low in nutrients (e.g. nitrogen <0.05 mg/L and phosphorus <0.02 mg/L) and oxygen (<0.1 mg/L) (Vanderzalm et al., 2010). Recycled stormwater undergoes no further treatment (e.g. filtration or disinfection) following aquifer recovery. Drinking water was sourced from the drinking water supply distribution network which in this part of the metropolitan Adelaide distribution system originates mainly from the Little Para Reservoir surface water catchment and water treatment plant (SA Water, 2014). In each rig, the water supply entered into a 200 L header tank via a float valve and was circulated through a pipe loop made of 36 m of PVC pipe (DN100, PN12) and 6 m of ductile iron cement lined (DICL) pipe (DN100, PN12). The pipe loops were buried 600 mm below ground. Each rig also had a 3.2 m coupon testing section made of PVC (DN150), based on a modified Robbins test device (McCoy et al., 1981), where coupons of selected pipe materials (cement-lined, copper and PVC) were installed. These are shown in Fig. 1 as sections (V3eV4 and V8eV9). Each section could hold 117 coupons measuring 4 cm by 2 cm in size. Coupons were mounted onto plastic, screw-fit holders and oriented in the direction of the flow. This section was housed in a covered pit 900 mm deep. Water in each rig was circulated using a suction pump (Lowara ITT CEA 370/3/A-V 240/380e415 50) controlled by a Vasco variable speed drive coupled to an Omni P1600-100 pressure sensor and set to a constant pressure of 200 kPa (maximum achievable with rig configuration). The flow rate was measured using an Endress Hauser Proline Promag 10 W flow meter (pulse output: þ0.5% of reading þ 2 mm/s). Running the pumps at maximum speed (50 Hz) the flow rate was 140e145 L/min, which equates to a mean flow velocity of 0.14 m/s. Total rig volume (including header tank) was determined experimentally as 561 L by running a slug test using a salt (NaCl) tracer (Tjandraatmadja et al., 2014). Turnover of water within each rig was controlled using a solenoid activated irrigation valve on the pipe return line (V5 and V10 in Fig. 1) and an irrigation timer. The valve was set at ~200 kPa back pressure with a bleed rate of ~1 L/ min. It would open for 16 h and shut for 8 h every day, meaning that 960 L was bled (and replaced with fresh water) from each rig every 24 h. Water temperature, pH, electrical conductivity (EC) and oxidation-reduction potential (ORP) were continuously monitored in the pipe lines using sensors (T-type mineral insulated metal sheathed stainless steel thermocouples 3 mm e TC Measurement and Control (±1%), SensoreX pH electrodes e flat surface CPVC (±10%), Thermo Fisher Scientific Alpha Cond 500 connected to a Thermo Fisher Scientific stainless steel conductivity electrode (±5%) and SensoreX ORP electrode sensors (±10%), respectively).
112
D. Gonzalez et al. / Water Research 105 (2016) 110e118
Fig. 1. Schematic representation of experimental rig for the evaluation of biofouling potential and material reactivity due to reticulation of recycled stormwater (from Tjandraatmadja et al., 2014).
The pH and ORP sensor signals were amplified using Heaston Electronics self-powered pre amplifiers (PHAMP-A30). The sensors were installed in the section preceding the D150 mm pipe length (above V3 and V8 on Fig. 1). The sensor and flow rate data were recorded at 5 min intervals using a DT85 DataTaker logger with remote connection via a wireless modem router interfaced with deX DataTaker software and a dynamic domain name system hosting service.
2.2. Sample collection Water and pipe coupons were sampled from the rigs according to Table 1. Coupons were sampled on 10 instances between May 2012 and January 2014. Their exposure times ranged from 5 to 48 weeks and covered all four seasons as shown in Table 1. On some occasions only one type of coupon was sampled and there was some variance of exposure times sampled on some occasions.
D. Gonzalez et al. / Water Research 105 (2016) 110e118
113
Table 1 Coupon sampling events. Sample event
Date in
Date out
Exposure time (weeks)
Exposure season/s
Comments
1 2 3 4
23/05/2012 20/08/2012 25/09/2012 25/09/2012 21/08/2012 21/08/2012 25/09/2012 25/09/2012 21/08/2012 e
20/08/2012 25/09/2012 29/10/2012 30/01/2013 30/01/2013 13/03/2013 13/03/2013 30/04/2013 30/04/2013 e
13 5 5 18 23 29 24 31 36 e
AutumneWinter WintereSpring Spring SpringeSummer WintereSummer WintereAutumn SpringeAutumn SpringeAutumn WintereAutumn e
PVC only cement, PVC, Cu cement only cement PVC, Cu PVC, Cu cement cement PVC, Cu water quality 22/5/2013
25/09/2012 21/08/2012 25/09/2012 21/08/2012 30/01/2013 1/10/2013
11/06/2013 11/06/2013 22/07/2013 22/07/2013 1/10/2013 8/01/2014
37 42 43 48 35 14
SpringeWinter WintereWinter SpringeWinter WintereWinter SummereSpring SpringeSummer
cement PVC, Cu cement PVC, Cu cement, PVC, Cu cement, PVC, Cu
5 6 e 7 8 9 10
At each sampling event, duplicate coupons of each pipe material were removed for surface characterisation and microbial analysis. Coupons were stored in 50 mL centrifuge tubes with 25 mL of source water and kept below 4 C for transportation to the laboratory, together with samples of respective source waters (mains and stormwater). Microbial analyses were conducted within 24 h of sampling.
Averages of 5 readings across coupon surfaces were taken. The elemental composition of sediment on coupons was determined using micro X-ray Diffraction (micro-XRD) using a Bruker GADDS micro diffractometer, and X-ray Fluorescence (XRF) using a Key Master Tracer III-V portable light-elemental analyser.
2.3. Water quality analysis
Each of the coupons was rinsed (in turn) with 1 mL of filter sterilised (0.2 mm syringe filter) source water using a 1 mL pipette and sterile cotton buds were used to swab coupons surfaces. The cotton buds were placed in individual sterile tubes containing 5 mL of sterile filtered source water, sonicated (Branson 1200) for 5 min and then twisted vigorously to release attached biofilm. Samples were put on ice for transportation to the laboratory for analysis along with the 2 500 mL source water samples. The swabbed surface areas of each coupon where also measured. The resulting suspensions from the coupons were mixed well prior to dilution (1:10), then stained (SYBR green). Total microbial cell counts were determined using a Cell Lab QuantaTM SC Beckman Coulter flow cytometer (equipped with an air-cooled 15 mW argon ion laser, emitting at a fixed wavelength of 488 nm). Logarithmic signals were collected (Hoefel et al., 2003) and data analysed using Cell Lab Quanta® SC MPL Analysis software. Thermotolerant coliforms were assessed by plating onto chromocult agar (Merck, catalogue number 1.10426.0500) and enumerating new colonies formed following incubation at 45 C for 24 h after initial incubation at 37 C for 24 h. Statistics were calculated using SigmaPlot 13.0 (Systat Software Inc.).
At each coupon sampling event (and on two other occasions, see Table 1) water from the return lines of the rigs was sampled for analyses. Water samples were maintained below 4 C and delivered to the laboratory within 24 h according to procedures and storage times according to methods recommended in the Standard Methods for the Examination of Wastewater (APHA, 2005). Alkalinity (as HCO 3 ) was determined by automated acidimetric titration; chloride, total nitrogen (TN) and total phosphorus (TP) were determined by colorimetric flow analysis; determination of metals by ICP spectrometry by ICP2; elemental analysis by ICP/MS; turbidity by Nephelometric measurement; true colour by spectrophotometric measurement; faecal coliforms/Escherichia coli by membrane filtration; total organic carbon (TOC) by flame ionisation; and biodegradable dissolved organic carbon (BDOC) was calculated from the change in DOC during incubation with a bacterial inoculum. Analyses were conducted by the Australian Water Quality Centre laboratory (ISO 9001 and ISO17025 certified and NATA accredited). Additionally, conductivity, temperature, pH and dissolved oxygen of rig water were also measured in situ at each sample collection event using a 90FL-mV field analyser and probes in a flow-through cell and residual chlorine was measured using a HACH DR/890 portable colorimeter. Mineral saturation indices were calculated with PHREEQC (Parkhurst and Appelo, 1999). 2.4. Surface morphology and sediment elemental composition Surface morphology was determined using an Olympus SZX10 optical microscope (magnification 126) with Dyno Capture 2.0 software, and a 3D confocal laser microscope (Olympus OLS4000) with MPFLN10, 20 and 50 objective lens under XYZ scanning mode. Colour was determined using the CIE L*a*b* colour scale with a Konica Minolta CR-300 Chroma meter colorimeter and a white tile reference standard. Total colour difference (DE) was calculated using the CIE 1976 Colour-Difference Formula (Robertson, 1977).
2.5. Biofilm analysis
3. Results and discussion 3.1. Operating conditions A constant flow rate of 145 L/min at a pressure of 200 kPa for both rigs was the maximum that could be maintained throughout the experiment using the pumps specified. This translated to a water velocity in the coupon section (DN150) of 0.14 m/s, which was 30e70% less than typical velocities (0.2e0.5 m/s) for drinking water distribution systems (Manuel et al., 2007). Various studies documented by Liu and Tay (2002) suggest the effect of higher shear force results in higher density, smoother and thinner biofilm layers whereas when formed under lower shear stress environments biofilms are thicker, less dense and more prone to sloughing. The flow conditions under which the rigs were operated in the current study represent conditions under which biofilms (and
114
D. Gonzalez et al. / Water Research 105 (2016) 110e118
deposited sediments) could be unstable enough to allow sloughing once a critical thickness was reached. Water temperature in the rigs was seasonally and diurnally affected. Summer mean daytime temperature was 30 ± 3 C while winter daytime mean temperature was 19 ± 2 C. Ambient air temperatures in 2013 were 1e45 C so burying the pipework provided a degree of insulation although maxima varied above typical values of <20 C for Adelaide mains distribution system (Tjandraatmadja et al., 2014). Improvements to the design could include insulation on the pit lid for operation within a narrower and lower temperature range. 3.2. Discolouration There was a general rapid increase in discolouration (determined as total colour difference, DE) for all pipe materials in both rigs within the first 5 weeks (Fig. 2). Further discolouration of cement coupons exposed to stormwater occurred after 20 weeks and copper and PVC coupons after 30 weeks. Colour differences are reflected in photographic time series of coupons where darkening of coupons exposed to stormwater is more visible than for those exposed to mains water (Fig. 3). The variability in discolouration of cement and PVC coupons in mains and stormwater over time was indicative of build-up and subsequent sloughing of surface deposits (Fig. 2). Notable differences were observed in copper exposed to stormwater where a transition from yellow to green was observed (reduction in b*) particularly after 36 weeks and for PVC and cement exposed to stormwater where marked darkening (reduction in L*) was observed after 35 weeks (Fig. 4). This suggests that corrosion mechanisms, biofilm and sediment composition differed according to source water. Cement coupons in mains water retained a red hue similar to the original colour for up to 35 weeks before discolouring towards the green coordinate (from a* ¼ þ5.28 ± 0.13 to a* ¼ þ1.36 ± 0.66). Copper coupons in mains and particularly stormwater also discoloured towards a green hue (reduction in a*, Fig. 4) due to oxidation and formation of the characteristic green patina (Fig. 3). PVC coupons discoloured markedly more in stormwater than in mains water (Fig. 2) and the hues also differed notably after 35 weeks (Fig. 4). In stormwater PVC coupons discoloured toward red (increase in a*) whereas in mains water coupons discoloured toward green (reduction in a*). The total colour difference of PVC coupons exposed in mains water was small (DE 18 units) and slight yellowing was observed, due to sediment attachment. In stormwater, coupons displayed a larger colour difference (DE 61
Fig. 2. Total colour difference (DE) of copper, cement and PVC coupons exposed to stormwater (sw) and mains water (mains) over time.
units) with reduced lightness (L*), and a shift towards red and yellow coordinates. This confirms that cement and copper are more reactive than PVC and more prone to biofouling (Niquette et al., 2000). 3.3. Elemental surface and sediment analyses XRF analysis indicated sediment deposition on PVC and cement coupons exposed to both source waters consisted of mainly iron rich amorphous compounds. There were greater intensities of spectra associated with PVC manufacturing additives on coupons exposed to mains water indicating the surface layer of cement and PVC coupons exposed to stormwater was substantially thicker. This is consistent with significantly higher median concentration of total Fe (Kruskal-Wallis test, P ¼ 0.04) in the stormwater (0.45 mg/ L) compared to mains water (0.09 mg/L). XRF analysis also showed high peaks for copper on cement coupons exposed to mains water, but not for stormwater coupons. This was attributed to the significantly higher (Kruskal-Wallis test, P ¼ 0.02) median concentration of total copper in the mains water (0.18 mg/L) compared to stormwater (0.09 mg/L). XRD traces indicated the presence of copper oxidation products, cuprite and copper oxide on surface layers of copper coupons exposed to both source waters with thin deposits comprising Cu and Fe also indicated through fluorescence analysis. 3.4. Biofilm microbial density Biofilm microbial density (live and dead organisms) determined through total cell counts were in the order of 105e107 cells/cm2 in both rigs and did not appear to increase continuously as a function of exposure time (Fig. 5). With samples pooled across exposure times, there were no significant differences in log10 transformed mean cell numbers between mains water and stormwater for the different pipe materials (ANOVA test, P ¼ 0.74). There were however differences in the number of thermotolerant coliforms between source waters. For copper coupons exposed to mains water, thermotolerant coliforms were not detected but in stormwater were detected twice (~36 cells/cm2) from 14 samples. For stormwater, median thermotolerant coliforms counts for both cement (88 cells/cm2) and PVC (85 cell/cm2) were significantly higher than for mains water (medians of zero) (Kruskal-Wallis test, P < 0.001). There was no chlorination treatment of stormwater so the residual in the stormwater rig throughout the experiment was zero. The mean chlorine residual in the mains water rig was 0.02 mg/L (±0.02) which was an order of magnitude lower than averages across the Adelaide mains distribution network (SA Water, 2014). This low level of disinfection was unlikely to limit biofilm growth especially at the temperatures observed in the rigs (frequently >20 C). Biodegradable dissolved organic carbon (BDOC) is a microbial food source and can stimulate biological growth. Mean BDOC concentrations in mains and stormwater samples were comparable (0.9 and 0.7 mg/L respectively, Table 2) and both were 5e6 times higher than the recommended maximum of 0.15 mg/L to limit biofilm growth (Escobar et al., 2001). Mean total cell counts for each coupon type and source water were variable with respect to exposure time (Fig. 5). This was most evident for PVC exposed to stormwater where the difference between lowest and highest mean counts was factor of 132 (3 104 and 4 106 cell/cm2 respectively) with the minimum observed at the longest exposure time (35 weeks). The lowest overall variance (factor of 5) was for mains water PVC where counts remained with the range of 2e7 105 cells/cm2 (lowest recorded at 18 weeks exposure). Differences between minimum and maximum cell counts for other material/water types ranged from a factor of
D. Gonzalez et al. / Water Research 105 (2016) 110e118
115
Fig. 3. Appearance of cement (top), PVC (middle) and copper (bottom) coupons after exposure to mains water and stormwater (image from Tjandraatmadja et al., 2014).
Fig. 4. Mean colour changes (from 5 readings of a single coupon at each time step) of cement, copper and PVC coupons exposed to stormwater and mains water (error bars represent one standard deviation, blank is unexposed coupon).
116
D. Gonzalez et al. / Water Research 105 (2016) 110e118
1.E+08
Total cell count (cells/cm 2)
1.E+07 Cu (MW) Cu (SW)
1.E+06
Cement (MW) Cement (SW)
1.E+05
PVC (MW) PVC (SW)
1.E+04
1.E+03 0
5
10
15
20
25
30
35
40
45
50
Exposure (weeks) Fig. 5. Mean total cell counts (n ¼ 2) on copper, cement and PVC coupons exposed to mains water (MW) and stormwater (SW) determined by flow cytometry. ©Copyright CSIRO Australia, 2016 CSIRO.
Table 2 The quality of water sampled from rigs during coupons collection during this study and quality of treated drinking water from Little Para water treatment plant and recovered stormwater based on previously published literature. Parameter
Mains rig
(mg/L unless otherwise indicated)
Mean (SD)
Stormwater rig
pH (pH units) DO TDS Turbidity (NTU) True colour (HU) Alkalinity SICalcite Cl residual free Ca Mg SO4 Na K TN TP TOC BDOC As (total) Fe (total) Mn (total) Cu (total) Escherichia coli (cfu/100 mL) Faecal coliforms (cfu/100 mL)
7.5 (0.4) 12 6.3 (2.1) 8 287 (69.7) 12 2.5 (3.8) 10 1.6 (0.8) 12 74 (16) 12 0.7 (0.3) 12 0.03 (0.007) 5 26 (2.5) 12 11 (3.7) 12 40 (7.5) 12 51 (11) 12 4.9 (1.6) 12 0.449 (0.096) 11 0.013 (0.007) 12 3.2 (1.3) 12 0.7 (0.4) 12 0.0005 (0.0001) 10 0.165 (0.21) 12 0.004 (0.003) 12 0.288 (0.227) 12 0 (0) 8 0 (0) 8
n
Mean (SD)
n
7.4 (0.4) 11 5.7 (1.9) 8 326 (166) 11 9.7 (14.9) 11 18 (26) 11 136 (45) 11 0.5 (0.7) 12 0 (0) 7 38 (12) 11 12 (6.5) 11 34 (26) 11 60 (50) 11 4.4 (1.2) 11 0.335 (0.15) 11 0.059 (0.08) 11 2.6 (1.7) 11 0.9 (1.2) 11 0.004 (0.006) 11 0.562 (0.41) 10 0.036 (0.033) 10 0.115 (0.10) 10 2 (2) 9 2 (2) 9
Little para WTP Mean
na
7.4 265 e 356 60 0.16 265 1 265 69 60 0.8 60 e 25 60 16 60 47 60 75 60 4.1 60 0.398 60 0.005 59 e e 0.0009 60 0.0083 134 0.00027 134 0.0167 59 e e
Recovered stormwater Median
nb
7.3 15 0.14 15 290 15 1.5 15 9 15 148 15 0.3 15 e 37 15 15 15 29 15 45 15 4.0 15 0.220 15 0.023 15 2.3 15 0.8 15 0.002 4 0.376 15 0.019 15 0.0006 15 14 15 14 15
SD ¼ standard deviation; n ¼ number of samples. a Little Para Water Treatment Plant (WTP) treated drinking water quality summary 2005e2010 (Page et al., 2013). b Parafield aquifer recovered stormwater (Page et al., 2013).
15e27. Cell counts were therefore not dependent on exposure time and could be driven by cycles of build-up and subsequent sloughing of biofilm through flow induced friction once a certain level of biofilm was reached (Ginige et al., 2011). 3.5. Implications for material reactivity and biofouling High alkalinity (e.g. >200 mg/L) is conducive to increased scaling (NHMRCeNRMMC, 2011). However, mean alkalinity in the stormwater rig was 136 ± 45 mg/L (Table 2) and therefore scaling would generally be unlikely. pH for both mains and stormwater rigs also remained within target criteria (6.5e8.5) for prevention of corrosion and scaling (NHMRCeNRMMC, 2011).
Compared to mains water, stormwater presented greater risks in terms of water aesthetics. On average, water in the stormwater rig was a factor of 4 higher in turbidity, a factor of 3 higher in Fe and a factor of 11 higher in colour than water in the mains rig and exceeded water aesthetic criteria (NHMRCeNRMMC, 2011) of 5 NTU, 0.3 mg/L, and 15 HU respectively (Table 2). The presence of high Fe and turbidity can also affect the efficacy of disinfectants e.g. chlorine. Iron precipitation can occur when soluble iron in anoxic water, such as deep groundwater, is subjected to aeration e.g. through reticulation or splash entry to header tanks. Insoluble iron (oxy) hydroxides or oxides can contribute to sedimentation and pipe scaling. Iron removal through aeration and filtration prior to reticulation may therefore be warranted to mitigate these risks.
D. Gonzalez et al. / Water Research 105 (2016) 110e118
Total microbial cells counts did not differ significantly between the two water sources and this was ascribed mainly to the low chlorine residual (0.03 ± 0.007 mg/L) of the mains water circulating in the rig but also the relatively high mean levels of BDOC (>0.7 mg/ L) in both rigs. However, thermotolerant coliform numbers indicated biofilms on coupons exposed to stormwater were subject to greater levels of faecal contamination. There is evidence of the potential for BDOC removal in aquifers as the median for injected stormwater at the Parafield ASR site was 2.1 mg/L and on recovery was about half (0.8 mg/L) (Page et al., 2013). Sufficient aquifer storage time may reduce labile organic carbon concentrations to levels that would reduce biofilm growth potential in the distribution pipelines (Vanderzalm et al., 2010) but further investigation to determine rates and required storage times would be required. Australian health-based log-removal targets for open space irrigation with stormwater can be achieved through simple exposure control e.g. restricted public access while irrigating such that no disinfection is required (NRMMC-EPHC-AHMC, 2006). This study highlights the importance of monitoring of key parameters to manage operational and aesthetic risks for stormwater recycling and distribution. Verification monitoring should include Fe and Mn, BDOC and faecal indicators (e.g. E. coli, thermotolerant coliforms). Operational monitoring should include easily or continuously monitored surrogate parameters including colour and turbidity (appearance and sedimentation), EC (indicator for any solute), pH (important for scaling and metal corrosion), and UV/Vis for total organic carbon (to indicate potential levels of BDOC). 4. Conclusions This study examined the impacts of recycled stormwater on cement, PVC and copper distribution pipes using recirculating pilot rigs. Conventional drinking water supply was used as the control. (1) Following storage in a carbonate aquifer, stormwater had about twice the alkalinity, and more than three times the turbidity and iron compared to the control rig running treated drinking (mains) water. Pipe discolouration differed between stormwater and mains water indicating differences in corrosion rates, sediment and biofilm composition and this was supported by X-ray Fluorescence and Diffraction analyses of pipe surfaces. Biofilm determined by total microbial cell densities were similar between rigs for the different pipe materials. However, faecal indicators were higher in stormwater. (2) Distribution of recycled stormwater has the potential to result in greater corrosion and sediment deposition than distribution of conventional drinking water. Operational monitoring should include colour and turbidity to manage water aesthetics and sedimentation potential, electrical conductivity as an indicator for changes in solute concentration, pH to manage scaling and metal corrosion, and UV/ Vis for total organic carbon as an indicator for labile organic carbon important for limiting microbial growth. Acknowledgements The authors acknowledge support of the partners to the Managed Aquifer Recharge and Stormwater Use Options research project. These are the National Water Commission through the Raising National Water Standards Program, the SA Government through the Goyder Institute for Water Research, CSIRO Water for a Healthy Country Flagship, City of Salisbury, the Adelaide and Mt Lofty Ranges Natural Resources Management Board and former United Water International.
117
References APHA, 2005. Standard Methods for the Examination of Water and Wastewater. APHA-WEF-AWWA, Washington. Clark, R., Gonzalez, D., Dillon, P., Charles, S., Cresswell, D., Naumann, B., 2015. Reliability of water supply from stormwater harvesting and managed aquifer recharge with a brackish aquifer in an urbanising catchment and changing climate. Environ. Model. Softw. 72, 117e125. Dillon, P., Toze, S., Page, D., Vanderzalm, J., Bekele, E., Sidhu, J., Rinck-Pfeiffer, S., 2010. Managed aquifer recharge: rediscovering nature as a leading edge technology. Water Sci. Technol. 62 (10), 2338. Escobar, I.C., Randall, A.A., Taylor, J.S., 2001. Bacterial growth in distribution systems: effect of assimilable organic carbon and biodegradable dissolved organic carbon. Environ. Sci. Technol. 35 (17), 3442e3447. Ginige, M.P., Wylie, J., Plumb, J., 2011. Influence of biofilms on iron and manganese deposition in drinking water distribution systems. Biofouling 27 (2), 151e163. Gonzalez, D., Page, D., Vanderzalm, J., Dillon, P., 2015. Setting water quality trigger levels for the operation and management of a MAR system in Parafield, South Australia. J. Hydrologic Eng. 20 (3), B5014001. Hoefel, D., Grooby, W.L., Monis, P.T., Andrews, S., Saint, C.P., 2003. Enumeration of water-borne bacteria using viability assays and flow cytometry: a comparison to culture-based techniques. J. Microbiol. Methods 55 (3), 585e597. Le Chevallier, M.W., 2003. Conditions favouring coliform and HPC bacterial growth in drinking-water and on water contact surfaces. In: Bartram, J., Cotruvo, J., Exner, M., Fricker, C., Glasmacher, A. (Eds.), Heterotrophic Plate Counts and Drinking-water Safety. World Health Organisation, London, pp. 177e196. Lee, S.H., O'Connor, J.T., Banerji, S.K., 1980. Biologically mediated corrosion and its effects on water quality in distribution systems. J. Am. Water Works Assoc. 82, 636e645. €nen, M.M., Kekki, T.K., Laine, O., Hirvonen, A., Lehtola, M.J., Miettinen, I.T., Keina Vartiainen, T., Martikainen, P.J., 2004. Microbiology, chemistry and biofilm development in a pilot drinking water distribution system with copper and plastic pipes. Water Res. 38 (17), 3769e3779. Lehtola, M.J., Laxander, M., Miettinen, I.T., Hirvonen, A., Vartiainen, T., Martikainen, P.J., 2006. The effects of changing water flow velocity on the formation of biofilms and water quality in pilot distribution system consisting of copper or polyethylene pipes. Water Res. 40 (11), 2151e2160. Liu, Y., Tay, J.H., 2002. The essential role of hydrodynamic shear force in the formation of biofilm and granular sludge. Water Res. 36 (7), 1653e1665. Manuel, C.M., Nunes, O.C., Melo, L.F., 2007. Dynamics of drinking water biofilm in flow/non-flow conditions. Water Res. 41 (3), 551e562. McCoy, W.F., Bryers, J.D., Robbins, J., Costerton, J.W., 1981. Observations of fouling biofilm formation. Can. J. Microbiol. 27 (9), 910e917. Miller, H.C., Wylie, J., Dejean, G., Kaksonen, A.H., Sutton, D., Braun, K., Puzon, G.J., 2015. Reduced efficiency of chlorine disinfection of Naegleria fowleri in a drinking water distribution biofilm. Environ. Sci. Technol. 49 (18), 11125e11131. Niquette, P., Servais, P., Savoir, R., 2000. Impacts of pipe materials on densities of fixed bacterial biomass in a drinking water distribution system. Water Res. 34, 1952e1956. NHMRCeNRMMC, 2011. Australian Drinking Water Guidelines. National Health and Medical Research Council and Natural Resource Management Ministerial Council, Canberra. NRMMCeEPHCeAHMC, 2006. Australian Guidelines for Water Recycling: Managing Health and Environmental Risks (Phase 1). NRMMCeEPHCeAHMC, Australia. NRMMCeEPHCeNHMRC, 2009. Australian Guidelines for Water Recycling: Managing Health and Environmental Risks (Phase 2) Stormwater Harvesting and Reuse. NRMMCeEPHCeNHMRC, Australia. Page, D., Dillon, P., Toze, S., Sidhu, J.P.S., 2010. Characterising aquifer treatment for pathogens in managed aquifer recharge. Water Sci. Technol. 62 (9), 2009. Page, D., Gonzalez, D., Dillon, P., Vanderzalm, J., Vadakattu, G., Toze, S., Sidhu, J., Miotlinski, K., Torkzaban, S., Barry, K., 2013. Managed Aquifer Recharge Stormwater Use Options: Public Health and Environmental Risk Assessment Final Report. Goyder Institute for Water Research Technical Report 13/17. http:// goyderinstitute.org/index.php?id¼20. Parkhurst and Appelo, 1999. User's Guide to PHREEQC (Version 2) A Computer Program for Speciation, Reaction Path, Advective Transport, and Inverse Geochemical Calculations. Report 99e4259. U.S. Geological Survey Water Resources Investigation. Percival, S.L., Walker, J.T., 1999. Potable water and biofilm: a review of the public health implications. Biofouling 14 (2), 99e115. Robertson, A.R., 1977. The CIE 1976 color-difference formulae. Color Res. Appl. 2 (1), 7e11. SA Water, 2014. SA Water Drinking Water Quality Report 2013-14. SA Water. Government of South Australia. Accessed [Sep 2015]. https://www.sawater.com.au. Tjandraatmadja, G., Kaksonen, A.H., Gonzalez, D., Barry, K., Vanderzalm, J.V., Puzon, G., Sidhu, J., Wylie, J., Goodman, N., 2014. Managed Aquifer Recharge and Stormwater Use Options: Investigation of Stormwater Impact on Water Quality and Distribution Infrastructure. Goyder Institute for Water Research Technical Report Series No. 14/8. Vanderzalm, J.L., Page, D.W., Barry, K.E., Dillon, P.J., 2010. A comparison of the geochemical response to different managed aquifer recharge operations for injection of urban stormwater in a carbonate aquifer. Appl. Geochem. 25 (9), 1350e1360. Vanderzalm, J.L., Page, D.W., Barry, K.E., Dillon, P.J., 2013. Application of a
118
D. Gonzalez et al. / Water Research 105 (2016) 110e118
probabilistic modelling approach for evaluation of nitrogen, phosphorus and organic carbon removal efficiency during four successive cycles of aquifer storage and recovery (ASR) in an anoxic carbonate aquifer. Water Res. 47 (7), 2177e2189. Vanderzalm, J., Page, D., Gonzalez, D., Barry, K., Toze, S., Bartak, R., Shisong, Q.,
Weiping, W., Dillon, P., Lim, M.H., 2014. Managed Aquifer Recharge and Stormwater Use Options: Satellite Sites Stormwater Quality Monitoring and Treatment Requirements Report. Goyder Institute for Water Research Technical Report 14/10.