1 The following document is supplementary material for the paper: 2 Paton, Maier and Dandy, Including adaptation and mitigation responses to climate change in a multi- 3 objective evolutionary algorithm framework for urban water supply systems incorporating GHG 4 emissions 5 Submitted to Water Resources Research, December 19, 2013 6 1 7 As the local catchment reservoirs and their associated water treatment plants (WTPs) were existing 8 features of Adelaide’s southern water supply system, only ongoing GHG emissions were accounted 9 for, namely the energy required for treatment at the WTPs and those associated with chemical use. 10 GHG emissions associated with the replacement of Happy Valley WTP in 2041 and Myponga WTP 11 in 2043 (see section 1 of auxiliary material on economic costs) could not be accounted for due to the 12 absence of data. 13 The power required at Happy Valley WTP was assumed to be 100 kWh/ML, while for Myponga 14 WTP it was assumed to be 140 kWh/ML (see section 1 of auxiliary material on economic costs). 15 While there is an absence of case-specific data for the GHG emissions associated with chemicals for 16 Adelaide’s WTPs, Racoviceanu et al. [2007] found that 94% of energy usage is attributable to 17 operation, 5% is attributable to chemical manufacturing and 1% to chemical transportation, which was 18 similar to a study by Tarantini and Ferri [2001] that estimated chemical manufacturing accounted for 19 10% of energy usage and chemical transportation was insignificant. Consequently, the power rates of 20 100 kWh/ML and 140 kWh/ML for each of the water treatment plants were assumed to account for 21 94% of energy, with 6% attributed to chemical manufacturing and transport. Consequently, for Happy 22 Valley and Myponga WTP, GHG emission estimates were based on energy requirements of 106.4 23 kWh/ML and 148.9 kWhh/ML, respectively. These energy intensities were multiplied by the GHG 24 emissions factor of 0.73 kgCO2-e/kWh and the volume of water treated at each WTP to calculate the 25 total emissions attributable to treatment energy for the local catchment reservoirs. Local Catchment Reservoirs 26 2 27 The River Murray was an existing Adelaide southern system supply source, so it was not necessary to 28 account for capital GHG emissions associated with the Murray Bridge-Onkaparinga (MBO) pipeline. 29 Consequently, only ongoing GHG emissions of River Murray supply were accounted for, which 30 included GHG emissions of treating the water at Happy Valley WTP (see section 1) and GHG 31 emissions due to pumping. As per the estimation of costs, GHG emissions due to pumping were 32 derived from first principles, using the Darcy-Weisbach head loss equation and the pump power 33 equation (see Eq. 2 and Eq. 3 in section 2 of auxiliary material on economic costs). The pump power 34 was then converted to kWh, based on 24-hour operation, before being converted to GHG emissions 35 using the GHG emission rate of 0.73 kgCO2-e/kWh, which is the latest full fuel cycle emission factor 36 estimate for purchased electricity by South Australian end users [Department of Industry Innovation 37 Climate Change Science Research and Tertiary Education, 2013]. For example, if the River Murray 38 supplied 50GL/yr, the GHG emissions associated with electricity to pump River Murray water would 39 be 1.1 kgCO2-e/m3 or 55,000 tCO2e-/yr. 40 3 41 3.1 Capital GHG Emissions – Desalination Plant 42 The capital GHG emissions for the desalination plant were attributed to the materials, electricity, and 43 diesel used to construct the main plant and onsite power facilities. SA Water [2009] estimated 44 quantities and resulting GHG emissions for these resources for the 50 GL/yr proposed Adelaide 45 desalination plant and proposed that there would be minimal additional construction energy for the 46 100 GL/yr plant because the 50 GL/yr plant was designed for an upgrade to 100 GL/yr capacity. 47 Consequently, an indicative 10% of capital GHG emissions was added to the GHG emissions reported 48 by SA Water [2009] (Table 1), to represent the additional buildings for process equipment that were 49 required for the 100 GL/yr capacity plant. 50 River Murray Desalination 51 52 Table 1: Estimated GHG emissions for a 100GL/yr desalination plant based on values reported by SA Water [2009] for the Adelaide desalination plant GHG emissions (tCO2-e) Resource Electricity Diesel Concrete Steel Stainless Steel Aluminium Copper Significant Plastics Minor and Unknown GHG emissions Total 3,289 12,148 20,306 23,058 1,760 431 1,181 1,071 8,017 71,262 53 Using the ongoing GHG emissions derived for desalination plants (see section 3.3), a 100 GL/yr 54 (≈274 ML/day) plant would have operational GHG emissions of about 394,000 tCO2-e/yr. 55 Consequently, as capital GHG emissions are estimated at approximately 71,000 tCO2-e/yr (Table 1), 56 capital GHG emissions would equate to about 0.9% of total GHG emissions over the plant’s lifetime 57 (assuming a lifetime of 20 years [SA Water, 2009]). This percentage is smaller than that reported by 58 GHD Fitchner [2005], as they estimated that 5% of total GHG emissions were attributable to 59 materials and the construction phase for an RO plant. However, the 5% estimate by GHD Fitchner 60 [2005] was for desalination plants in Sydney, so the value for Adelaide of 0.9% was used to estimate 61 the capital GHG emissions for each desalination capacity (Table 2). For ongoing GHG emission 62 calculations of the desalination plant, see section 3.3. 63 Table 2: Capital GHG emissions for the different desalination plant capacities Plant Capacity (ML/day) 100 150 200 250 300 350 400 450 500 Capital GHG emissions (tonnesCO2e-) 25,887 38,830 51,773 64,717 77,660 90,603 103,547 116,490 129,443 64 3.2 Capital GHG Emissions – Transfer Pipeline 65 Capital GHG emissions associated with materials for the mild steel cement lined (MSCL) pipeline 66 and construction of the pipeline were accounted for; however, GHG emissions associated with the 67 materials for the pumps were considered insignificant and thus were excluded from the inventory 68 based on the fact that for the Murrumbidgee to Googong MSCL pipeline (12km, 1.0m diameter), the 69 materials and transport of the pumps only constituted 35 tCO2-e or 0.1% of total capital GHG 70 emissions [ACTEW Corporation, 2009]. 71 For the GHG emissions associated with the cement and steel of the pipeline, nominal diameters were 72 assumed with the cement and steel thicknesses detailed in Table 3 to estimate the volume of cement 73 and steel for each pipeline design (Table 3). Assuming a density of 2,300 kg/m3 for cement and 7,850 74 kg/m3 for steel, and applying the emission factors of 3.5 GJ/t for cement and 32 GJ/t for steel 75 [Investor Group on Climate Change, 2007], the resulting GHG emissions for the cement and steel of 76 the pipeline were derived (Table 3). 77 78 Table 3: Pipe diameter, cement and steel thickness and volumes of materials and resulting GHG emissions for the 12km MSCL transfer pipeline designed for the different desalination plant capacities Plant Capacity (ML/day) 100 150 200 250 300 350 400 450 500 Internal Pipe Diameter (m) Cement thickness (mm) Mild steel thickness (mm) Cement Volume (m3) 0.90 1.13 1.20 1.30 1.45 1.50 1.60 1.75 1.75 19 25 25 25 25 25 25 25 25 8 9.5 10 11 11 13 13 13 13 285 424 475 564 627 766 815 889 889 Mild GHG steel emissions volume for steel (m3) and cement (tCO2e-) 658 1,084 1,155 1,249 1,390 1,437 1,532 1,673 1,673 15,607 23,378 26,081 30,787 34,187 41,365 44,015 47,991 47,991 Total GHG emissions including diesel and vegetation clearance (tCO2e-) 23,107 30,878 33,581 38,287 41,687 48,865 51,515 55,491 55,491 79 For the construction of the Adelaide desalination plant transfer pipeline, SA Water [2009] estimated 80 that 1650 m3 of diesel would be required, resulting in GHG emissions of 4,698 tCO2e-. In addition to 81 diesel and steel, other major GHG emissions were expected to be from vegetation clearance and 82 concrete, which were expected to contribute 1,932 tCO2e- and 954 tCO2e-, respectively [SA Water, 83 2009]. Consequently, an additional 7,500 tCO2e- were added to the GHG emissions for steel and 84 cement to account for other GHG emissions involved in the construction of the pipeline. Due to 85 insufficient data, these GHG emissions were not scaled for the different sized pipelines. 86 3.3 Ongoing GHG Emissions – Desalination Plant 87 GHG emissions of the operating desalination plant were based on GHG emissions associated with 88 electricity required for treatment; chemicals; membrane and plant replacement. Assuming the plant 89 was running at full capacity, the electricity required for power was assumed to be 5.0 kWh/m3 (see 90 Section 3.3 of auxiliary material on economic costs). GHG emissions for chemicals, membranes and 91 diesel were based on those for the proposed 50 GL/year Adelaide desalination plant, which were 92 estimated to be 26,771 tonnesCO2e-/yr, 1,773 tonnesCO2e-/yr and 41 tonnesCO2e-/yr, respectively [SA 93 Water, 2009]. The GHG emissions of these three components were relatively small compared with 94 those associated with power, which contributed to 92.7% of the ongoing GHG emissions (Table 4). 95 The percentage breakdowns are very similar to other published percentage breakdowns (Table 4) and 96 were consequently applied to all desalination plant capacities to determine total ongoing GHG 97 emissions (Table 5). 98 Table 4: Ongoing GHG emission breakdown for the major desalination operating processes GHG emissions (% of Total Ongoing GHG emissions) Power Chemicals Membranes Other 92.7 6.8 0.5 negligible 92.1 7.0 <0.1 95.0 99 100 4.0 1.0 Source [SA Water, 2009] [Biswas, 2009] [Mrayed and Leslie, 2009] Table 5: GHG emissions attributed to power, chemicals, and membranes for each desalination plant capacity (assuming the plant operated at full capacity) Desalination Plant Capacity (ML/day) 100 150 200 250 300 350 400 450 500 Power (tCO2e-/yr) Chemicals (tCO2e/yr) Membranes (tCO2e-/5yrs) Total (tCO2e-/yr) 133,316 199,974 266,633 333,291 399,949 466,607 533,265 599,923 666,581 9,779 14,669 19,559 24,449 29,338 34,228 39,118 44,007 48,897 3,595 5,393 7,191 8,988 10,786 12,584 14,381 16,179 17,977 143,815 215,722 287,629 359,537 431,444 503,352 575,259 647,166 719,074 101 The GHG emissions associated with membrane replacement were applied every five years (Table 4) 102 and were independent of the amount of water the desalination plant produced. Furthermore, the GHG 103 emissions associated with plant replacement were accounted for once in 2030 and were estimated to 104 be the same as the initial capital GHG emissions (see Table 2, section 3.1). However, GHG emissions 105 associated with power and chemicals depended on the amount of water supplied by the desalination 106 plant. Energy consumption was computed based on the 5.0 kWh/m3 rate, resulting in a linear scaling 107 of GHG emissions for power with volume of water supplied. Similarly, the GHG emissions associated 108 with chemicals were scaled linearly, for example, if 25 ML of water were produced per day by the 109 250 ML/day plant, 10% of the chemical GHG emissions would be accounted for, while 50% of the 110 chemical GHG emissions would be accounted for if water was produced at the rate of 125 ML/day. 111 While accredited green power has been assigned to power the Adelaide Desalination Plant, in this 112 case study renewable energy was not assumed to run the desalination plant. This was appropriate 113 because (1) the renewable energy for the desalination plant is being sourced from South Australia’s 114 electricity grid, so technically renewable energy is not directly powering the desalination plant; (2) SA 115 Water have elected to purchase renewable energy to cover the electricity requirements of the 116 desalination plant but have not chosen to do so for other water sources, thus creating bias for one 117 water source over another from a GHG emissions point of view; and (3) by using the same factors, the 118 unit GHG emissions derived for each water source truly reflect the electricity required to power the 119 water sources, rather than reflecting the source of electricity for each water source. Consequently, the 120 same GHG emission rate of 0.73 kgCO2-e/kWh was applied to all water sources for the purposes of 121 this case study, including the desalination plant. 122 3.4 Ongoing GHG Emissions – Transfer Pipeline 123 The operating GHG emissions for the transfer pipeline were attributed to the electricity required to 124 power the transfer of water. The power required to pump the water was estimated using first 125 principles, as illustrated in Section 2 of auxiliary material on economic costs. The power requirements 126 were then multiplied by the electricity GHG emission factor of 0.73 kgCO2-e/kWh and assuming 24- 127 hour continuous pumping. The maximum annual GHG emissions associated with pumping water from 128 Port Stanvac to Happy Valley for each desalination plant capacity are reported in Table 6. GHG 129 emissions associated with pump replacement were excluded, as they were negligible compared with 130 GHG emissions associated with pumping: the materials for pumps were estimated to be in the order of 131 35 tCO2-e (see section 3.2), equating to 0.04-0.21% of the maximum annual GHG emissions required 132 for pumping (Table 6). Maintenance for the pipeline was also assumed to have negligible GHG 133 emissions and was thus not included in estimating GHG emissions. 134 Table 6: Maximum Pump Power and Maximum Annual GHG Emissions for the transfer pipeline Desalination Plant Capacity (ML/day) 100 150 200 250 300 350 400 450 500 Maximum Pump Power (Watts) 2620 3679 5175 6540 7533 9064 10,227 11,011 12,779 Maximum Annual GHG emissions (tCO2e-) 16,765 23,541 33,116 41,851 48,205 58,000 65,445 70,461 81,774 135 4 Rainwater Tanks 136 4.1 Capital GHG Emissions 137 Capital GHG emissions were attributed to the rainwater tank, pump, pipes, fixtures, and installation of 138 materials to site, while the transport of the rainwater tank system to site was not accounted for, as this 139 was expected to vary considerably for each house. When greater than 50% roof connectivity was 140 selected, capital GHG emissions for indicative lengths of PVC pipe to connect the extra roof area 141 were also accounted for. However, there was no differentiation in capital GHG emissions when water 142 was used for different end uses because there was insufficient data to accurately determine the 143 materials required to plumb in different end uses. This was also difficult to estimate given that house 144 layouts vary considerably. 145 The rainwater tanks were assumed to be made from high density polyethylene (HDPE). To estimate 146 the GHG emissions associated with using HDPE for the tanks, the volume of HDPE for each tank 147 (Table 7) was first estimated for each tank size using middle-of-the-range diameters sourced from a 148 number of rainwater tank suppliers in Adelaide (Table 7), their corresponding heights (Table 7) and 149 assuming that the thickness of the base was 10mm and the thickness of the sides and tops were either 150 5mm (1.0-7.5 m3 tank) or 7.5mm (10.0-27.0 m3). These values were then converted to GHG 151 emissions (Table 7) assuming a density of 950 kg/m3, an embodied energy factor of 75.2 MJ/kg (20.9 152 kWh/kg) [Piratla et al., 2012] and the GHG electricity conversion rate of 0.73 kgCO2e-/kWh. The 153 resulting GHG emissions for the southern system in 2010 (Table 7) were therefore estimated by 154 multiplying these GHG emissions by the number of houses in 2010 (260,261 houses). 155 Table 7: Diameter, height, volume of HDPE and GHG emissions for different sized rainwater tanks Size (m3) Diameter (m) Height (m) 1.0 2.0 3.0 4.0 5.0 7.0 10.0 15.0 22.5 27.0 0.93 1.30 1.50 1.83 1.84 2.00 2.55 2.50 3.70 3.85 1.72 2.00 2.00 1.75 2.15 2.42 2.35 3.10 2.50 3.10 Volume of HDPE for single tank (m3) 0.035 0.061 0.074 0.090 0.102 0.123 0.243 0.282 0.426 0.508 GHG emissions for single tank (kgCO2e-) 512 880 1067 1300 1478 1784 3513 4091 6171 7360 Total GHG emissions for southern system in 2010 (tCO2e) 133,150 229,045 277,608 338,411 384,665 464,308 914,418 1,064,695 1,606,188 1,915,500 156 Parkes et al. [2010] concluded that there is a lack of good quality data for estimating the embodied 157 energy for components of rainwater tank systems. However, they report that a generic electro- 158 mechanical pump for rainwater harvesting once installed has GHG emissions of 184 kgCO2e-. 159 Furthermore, they estimated that excluding the tank and pump, all other initial rainwater tank 160 components (e.g. pipes and fixtures) once installed contributed another 143 kgCO2-e. The applicability 161 of these data to this case study is questionable, considering the study is from the United Kingdom and 162 applies to a single household with a header tank, rather than a direct feed (as is assumed for this case 163 study). However, in lieu of any other data regarding embodied energy estimates for components of 164 household rainwater tank systems, these data were used with 327 kgCO2e- being added to the capital 165 GHG emissions of the rainwater tanks in Table 7. 166 To estimate the GHG emissions associated with redirecting more roof runoff into the rainwater tank, 167 indicative lengths of PVC pipe were assumed, with a diameter of 90mm and wall thickness of 1.9mm. 168 Similar to estimating the GHG emissions from using HDPE for the rainwater tanks, the volume of 169 PVC was thus estimated, and then converted to GHG emissions assuming a density of 1390 kg/m3, an 170 embodied energy factor of 74.9 MJ/kg (20.8 kWh/kg) [Piratla et al., 2012] and the GHG electricity 171 conversion rate of 0.73 kgCO2e-/kWh (Table 8). While for larger tanks these additional GHG 172 emissions only constituted a small proportion of the capital GHG emissions (e.g. 0.4%-1.5% for the 173 27.0 m3 tank), for the smaller tanks they added considerably to the capital GHG emissions (e.g. 5.1%- 174 17.8% for the 1.0 m3 tank). 175 Table 8: Length, volume of, and GHG emissions associated with the PVC pipe for the different roof connectivities Roof Connectivity 0.6 0.7 0.8 0.9 Length (m) 2.5 5.0 7.5 10.0 Volume (m3) 0.0013 0.0026 0.0039 0.0053 GHG emissions for single tank (kgCO2e-) 28 56 83 111 Total GHG emissions for southern system in 2010 (tCO2e) 7,223 14,447 21,670 28,894 176 4.2 Ongoing GHG Emissions 177 Ongoing GHG emissions for rainwater tanks were attributed to electricity use of the pump and 178 replacement of the tank and pumps. However, no replacement GHG emissions associated with pipes 179 nor fixtures were accounted for, as these were assumed to have a lifetime greater than the planning 180 horizon. Similarly, no GHG emissions associated with maintenance were accounted for due to a lack 181 of data. The electricity use of the pump was determined from the average energy intensity of the 182 pump (Table 9), which, as explained in section 4.2 of the auxiliary material on economic costs, was 183 dependent on the end use (a decision variable in the case study). To transfer these values to GHG 184 emissions, the energy intensities were multiplied by the volume of rainwater supplied and the GHG 185 emission rate of 0.73 kgCO2-e/kWh. For example, if 50 m3 was supplied to the garden and laundry in a 186 year, the annual GHG emissions associated with pumping would be about 45 kgCO2e- per household 187 and about 11,700 tCO2e- for the whole of the southern system in 2010. As the rainwater tanks were 188 assumed to have a lifetime of 25 years, replacement GHG emissions for the tanks were accounted for 189 and were assumed to be the same as the initial capital GHG emissions (see section 4.1). Similarly, 184 190 kgCO2e- was accounted for when the pump was replaced every 10 years (see section 4.1). 191 Table 9: Energy intensities of the pump for different harvested rainwater end uses Harvested Rainwater End Use Garden only Garden & Toilet Garden, Toilet, & Laundry Garden & Laundry Garden, Laundry, & Hot Water Garden & Hot Water Garden, Hot Water, & Toilet Garden, Toilet, Laundry, & Hot Water Energy Intensity (kWh/m3) 1.20 1.50 1.45 1.23 1.25 1.24 1.44 1.41 192 5 Stormwater Schemes 193 5.1 Capital GHG Emissions 194 The capital material and construction GHG emissions for the stormwater schemes were attributed to 195 the materials and construction of the wetland, Aquifer Storage and Recovery (ASR) wells, and the 196 distribution network. The GHG emissions associated with pumps were not considered, given that 197 GHG emissions for pumps of large pipelines were found to be negligible compared to other capital 198 GHG emissions and the operating GHG emissions of pumps (see section 3.4). 199 For the wetlands and ASR wells, GHG emissions were attributed to the concrete and steel used in 200 their construction and the excavation of soil required to create the wetlands and wells. The amount of 201 concrete and steel used in a wetland’s outlet structure was estimated from data supplied in 202 supplementary material by Moore and Hunt [2013]. Specifically, values for the mass of concrete and 203 mass of steel per square metre (1.90 kg/m2 and 0.02 kg/m2, respectively) were derived by dividing the 204 mass of concrete and steel estimated by Moore and Hunt [2013] for their case study wetland by its 205 area. For this case study, these rates were then applied to the total wetland area for the stormwater 206 schemes and converted to GHG emissions using the emission factors of 1.3 GJ/t and 32 GJ/t for 207 concrete (30 MPa) and steel, respectively (Table 10) [Investor Group on Climate Change, 2007]. In 208 the study by Moore and Hunt [2013], GHG emissions associated with the PET plastic trays used to 209 package wetland seedlings were also accounted for; however, these were found to constitute less than 210 2% of the wetland material GHG emissions, so they were neglected. 211 Table 10: Wetland area, mass of concrete and steel estimated for each wetland, and corresponding GHG emissions Stormwater Scheme Brownhill-Keswick Sturt Creek Field River Pedler Creek Total Wetland Area (ha) 49.72 50.4 25.26 40.03 Mass of Concrete (kg) 944,051 956,962 479,620 760,063 Mass of Steel (kg) 10,070 10,208 5,116 8,107 GHG Emissions (tCO2e-) 265 268 135 213 212 The volume of concrete used for the ASR wells was calculated by multiplying the number of wells for 213 each stormwater harvesting scheme (Table 11) [Wallbridge and Gilbert, 2009] by the volume of 214 concrete used in each well, which was estimated by assuming a well height of 150 m, an outside 215 diameter of 500 mm, and a wall thickness of 50 mm. Using the emission factor for concrete (30MPa) 216 of 3.2 GJ/m3 [Investor Group on Climate Change, 2007], the GHG emissions for the concrete used for 217 the wells was subsequently estimated (Table 11). 218 Table 11: Materials and construction GHG emissions for the ASR wells Stormwater Scheme Brownhill-Keswick Sturt Creek Field River Pedler Creek Number of wells 73 80 5 30 Volume of Concrete (m3) 409 448 28 168 Material GHG Emissions (tCO2e-) 213 233 15 87 219 For excavation energy (Table 12), the volumes of the wetlands and ASR wells were multiplied by the 220 assumed density of soil of 1250 kg/m3, and the energy required for excavation of 0.1 MJ/kg [Alcorn, 221 2003]. This energy was then multiplied by the emissions associated with the diesel used in excavation, 222 which was estimated at 69.2 kgCO2-e/GJ [Department of Industry Innovation Climate Change Science 223 Research and Tertiary Education, 2013], to obtain the final estimates of GHG emissions for wetland 224 construction (Table 12). The relatively greater number of GHG emissions associated with 225 construction (Table 5.3) compared with those associated with materials (Tables 5.1 and 5.2) for the 226 wetlands was expected, considering Moore and Hunt [2013] found that construction of stormwater 227 wetlands accounted for the majority of capital GHG emissions. 228 Table 12: Wetland and ASR construction GHG emissions for each stormwater scheme Stormwater Scheme Brownhill-Keswick Sturt Creek Field River Pedler Creek Total Wetland & ASR Well Volume (m3) 729,430 739,686 365,927 573,464 Total Energy for Excavation (GJ) Construction GHG Emissions for Wetland (tCO2e-) 91,179 92,461 45,741 71,683 6,310 6,398 3,165 4,960 229 For the distribution network, pipes were assumed to be made from high density polyethylene (HDPE), 230 have a pressure rating of 600 kPa, an outside diameter of 280mm, and a wall thickness of 10.8 mm. 231 The length of pipeline required for the distribution of harvested stormwater to non-potable industrial 232 and commercial users (Table 13) was estimated from a number of similar stormwater schemes in 233 Adelaide. Specifically, the following equation was derived to estimate pipe length based on the 234 potential yield of the schemes: 235 L 18.108 * Y 0.5681 (1) 236 where L is the length of pipeline (in km) and Y is the potential yield of the stormwater scheme (in 237 GL/yr). Consequently, the volume of HDPE for each of the stormwater scheme distribution networks 238 was estimated and converted to GHG emissions (Table 13) by multiplying the volume of HDPE by its 239 density (950 kg/m3), an embodied energy factor of 75.2 MJ/kg (20.9kWh/kg) [Piratla et al., 2012] 240 and the GHG electricity conversion rate of 0.73 kgCO2e-/kWh. Piratla et al. [2012] estimated that 241 GHG emissions of 2,830.4 kgCO2e- were incurred for installation of a 152.4 m-long, 200 mm- 242 diameter section of pipe buried 1.22 m below the surface. This value converts to an average 18.6 243 kgCO2e-/m of pipe installed. Considering the pipelines for the stormwater distribution in this case 244 study were almost a third larger in diameter than that reported by Piratla et al. [2012], an indicative 245 value for installation of 26 kgCO2e-/m was applied to this case study to estimate GHG emissions for 246 the distribution network installation (Table 13). 247 248 Table 13: Estimated length of pipelines for the stormwater schemes’ distribution network, the corresponding volume of HDPE required and the resulting GHG emissions associated with the HDPE and installation of the pipeline Stormwater Scheme BrownhillKeswick Sturt Creek Field River Pedler Creek Length of Pipeline (km) Volume of HDPE (m3) GHG Emissions for HDPE (tCO2e-) GHG Emissions for Pipe Installation (tCO2e-) 58.5 51.0 31.3 45.5 529.8 461.7 283.1 412.3 7,675 6,688 4,101 5,973 1,522 1,326 813 1,184 249 Total capital GHG emissions for the stormwater harvesting schemes are summarised in Table 14. 250 Table 14: Summary of capital GHG emissions for the stormwater harvesting schemes Stormwater Scheme BrownhillKeswick Sturt Creek Field River Pedler Creek Wetland and Well Materials 478 501 150 300 Capital GHG Emissions (tCO2e-) Wetland and Well Distribution Network Excavation 6,310 6,398 3,165 4,960 9,197 8,014 4,914 7,158 Total 15,985 14,913 8,229 12,418 251 5.2 Ongoing GHG Emissions 252 Ongoing GHG emissions for the stormwater harvesting schemes were based on operating GHG 253 emissions, namely those associated with pumping the stormwater. The wetlands, wells and 254 distribution network were assumed to have lifetimes greater than the planning horizon of 40 years 255 considered, so no GHG emissions associated with replacement were attributed to stormwater 256 harvesting. GHG emissions for pump replacement were also ignored, as these were assumed to be 257 negligible compared with those associated with the electricity required to pump the harvested 258 stormwater. 259 Dandy et al. [2013] estimated that the GHG emissions associated with pumping for a stormwater 260 harvesting scheme (the scheme included a wetland, ASR and distribution to open spaces for 261 irrigation) were approximately 554 MWh for 0.88 GL of stormwater. This translates to a rate of 0.63 262 kWh/m3 or 0.50 kgCO2e-/m3 using the electricity emissions factor of 0.73 kgCO2e-/kWh, which was 263 applied to the yield of the stormwater harvesting schemes to estimate ongoing GHG emissions. For 264 example, if the Brownhill-Keswick scheme produced its maximum yield of 7.9 GL/yr, annual 265 operating GHG emissions for the Brownhill-Keswick scheme would be 3,950 tCO2e-. 266 6 267 268 References ACTEW Corporation (2009), Appendix N: Greenhouse Gas Assessment, in Murrumbidgee to Googong Water Transfer Project - Environmental Impact Statement, Canberra. 269 Alcorn, A. (2003), Embodied Energy and CO2 Coefficients for NZ Building Materials, 31 pp, 270 Centre for Building Performance Research, Victoria University of Wellington, Wellington, New 271 Zealand. 272 273 Biswas, W. K. (2009), Life Cycle Assessment of Seawater Desalinization in Western Australia, World Academy of Science, Engineering & Technology(32), 369-375. 274 Dandy, G. C., A. Ganji, J. Kandulu, D. Hatton MacDonald, A. Marchi, H. R. Maier, A. Mankad, 275 and C. E. Schmidt (2013), Managed Aquifer Recharge and Stormwater Use Options: Net Benefits 276 Report, 180 pp, Goyder Institute for Water Research - Marsuo Project Final Report, Adelaide. 277 Department of Industry Innovation Climate Change Science Research and Tertiary Education 278 (2013), Australian National Greenhouse Accounts, 84 pp, Department of Industry, Innovation, 279 Climate Change, Science, Research and Tertiary Education; Australian Government, Canberra, 280 Australia. 281 282 GHD Fitchner (2005), Planning for Desalination, 47 pp, GHD Fitchner & Sydney Water, Sydney, Australia. 283 Investor Group on Climate Change (2007), Potential Earnings Impacts from Climate Change: 284 Construction Materials, 29 pp, Investor Group on Climate Change, Goldman Sachs JBWere, and 285 Monash Sustainability Enterprises Melbourne. 286 287 Moore, T. L. C., and W. F. Hunt (2013), Predicting the carbon footprint of urban stormwater infrastructure, Ecological Engineering, 58(0), 44-51, doi:10.1016/j.ecoleng.2013.06.021. 288 Mrayed, S., and G. Leslie (2009), Examination of Greenhouse footprint for both Desalination 289 and Water recycling processes, paper presented at Australian Water Association (AWA) Ozwater' 09 290 Conference, Australian Water Association, Melbourne, Australia. 291 Parkes, C., H. Kershaw, J. Hart, R. Sibille, and Z. Grant (2010), Energy and carbon implications 292 of rainwater harvesting and greywater recycling Rep. SC090018, 93 pp, Environment Agency, Bristol. 293 Piratla, K., S. Ariaratnam, and A. Cohen (2012), Estimation of Emissions from the Life Cycle of 294 a Potable Water Pipeline Project, Journal of Management in Engineering, 28(1), 22-30, 295 doi:10.1061/(ASCE)ME.1943-5479.0000069. 296 Racoviceanu, A., B. Karney, C. Kennedy, and A. Colombo (2007), Life-Cycle Energy Use and 297 Greenhouse Gas Emissions Inventory for Water Treatment Systems, Journal of Infrastructure 298 Systems, 13(4), 261-270, doi:10.1061/(ASCE)1076-0342(2007)13:4(261). 299 300 SA Water (2009), Proposed Adelaide Desalination Plant, Environmental Impact Statement Response Document, 170 pp, Adelaide. 301 Tarantini, M., and F. Ferri (2001), LCA of drinking and wastewater treatment systems of 302 Bologna City: Final Results, in 4th Inter-Regional Conference on Environmental Water: Competitive 303 use and conservation strategies for water and natural resources, Fotaleza, Brazil. 304 305 Wallbridge and Gilbert (2009), Urban Stormwater Harvesting Option Study, 93pp, Wallbridge and Gilbert, Adelaide.