A- A+
Alt. Display

# Holistic impact assessment and cost savings of rainwater harvesting at the watershed scale

## Abstract

We evaluated the impacts of domestic and agricultural rainwater harvesting (RWH) systems in three watersheds within the Albemarle-Pamlico river basin (southeastern U.S.) using life cycle assessment (LCA) and life cycle cost assessment. Life cycle impact assessment (LCIA) categories included energy demand, fossil fuel, metals, ozone depletion, global warming, acidification, smog, blue and green water use, ecotoxicity, eutrophication, and human health effects. Building upon previous LCAs of near-optimal domestic and agricultural RWH systems in the region, we scaled functional unit LCIA scores for adoption rates of 25%, 50%, 75%, and 100% and compared these to conventional municipal water and well water systems. In addition to investigating watershed-scale impacts of RWH adoption, which few studies have addressed, potential life cycle cost savings due to reduced cumulative energy demand were scaled in each watershed for a more comprehensive analysis. The importance of managing the holistic water balance, including blue water (surface/ground water), green water (rainwater) use, and annual precipitation and their relationship to RWH are also addressed. RWH contributes to water resource sustainability by offsetting surface and ground water consumption and by reducing environmental and human health impacts compared to conventional sources. A watershed-wide RWH adoption rate of 25% has a number of ecological and human health benefits including blue water use reduction ranging from 2–39 Mm3, cumulative energy savings of 12–210 TJ, and reduced global warming potential of 600–10,100 Mg CO2 eq. Potential maximum lifetime energy cost savings were estimated at $5M and$24M corresponding to domestic RWH in Greens Mill and agricultural RWH in Back Creek watersheds.

##### Knowledge Domain: Earth & Environmental Science Sustainable Engineering Sustainability Transitions
How to Cite: Ghimire, S.R. and Johnston, J.M., 2017. Holistic impact assessment and cost savings of rainwater harvesting at the watershed scale. Elem Sci Anth, 5, p.9. DOI: http://doi.org/10.1525/elementa.135
Published on 10 Mar 2017
Accepted on 25 Jan 2017            Submitted on 09 Sep 2016
Domain Editor-in-Chief: Michael Chang; Georgia Institute of Technology, GA, United States

## Introduction

Regarding our global water footprint, Hoekstra and Mekonnen (2012) describe consumptive use of surface and groundwater as blue water, rainwater as green water, and waste assimilation water as gray water. Consumptive use can be direct use (including evaporated water) or indirect use (embedded volume of water in a product’s full supply chain); non-consumptive water use (return flow) is not included in any water footprint (WFN 2009). Worldwide from 1996–2005, rain-fed agricultural production had a 91% green and 9% gray water footprint, while irrigated agriculture had 48% green, 40% blue, and 12% gray water footprint (Hoekstra and Hung 2002; Chapagain and Hoekstra 2004; Mekonnen and Hoekstra 2011).

Rainwater harvesting (RWH) is a technique to capture and store green water for future use. Examples include rain barrels and cisterns for non-potable household use and farm ponds for crop irrigation and livestock watering. RWH is also a climate change adaptation strategy and rehabilitates ecosystems services for human well-being by reducing stormwater runoff, mitigating sewer overflows, conserving soil, and enhancing food and economic security (USEPA 2008c; UNEP-SEI 2009). By increasing the use of green water, RWH can augment conventional centralized water supplies during water shortages and may function as a decentralized water system for climate change adaptation (Sedlak 2014). Globally, 71% of irrigated areas and 47% of large cities are reported to experience periodic water shortages (Brauman et al. 2016). RWH practices widely used in Asia and Africa are of interest in the U.S., such as irrigation in North Carolina, Virginia, Texas, and Ohio. These states promoted RWH in their regulations as well (TWDB 2006; DeBusk et al. 2013; Shuster et al. 2013; ARCSA 2014; Thomas et al. 2014). In the southeastern U.S., RWH is getting attention due to recent droughts (DeBusk et al. 2013), and RWH adoption will likely increase in the face of unreliable rainfall during the growing season. While most of the Southeast received heavy downpours in recent autumns, it also observed increases in moderate-to-severe drought in spring and summer (12% and 14%) from 1970 to 2007 (USGCRP 2009; 2012; USEPA 2015).

“Decreased water availability is very likely to affect [the Southeast] region’s economy as well as its natural systems” (USGCRP 2012). This is due to climate impacts combined with increased water demand, such as a 20% increase in agricultural irrigation by surface and groundwater withdrawals in Georgia, anticipated to continue through 2050 (Hook et al. 2009). However, adoption of RWH to compensate for water needs can introduce environmental and human health impacts and their associated economic costs (USGCRP 2009; 2012). RWH impoundments can impact biodiversity and wetland and upland habitat by interrupting stream ecosystems (USEPA 1990). The number of small reservoirs in the Georgia Piedmont increased from 19 to 329, mostly due to agricultural practices from 1950–1970 and suburban growth from 1980–1990 (Ignatius and Jones 2014). Although water resource decisions are local, the scope of impact is national. Approximately 20% of 2.6 million small water bodies account for the majority of standing water areas across the conterminous U.S. (Smith et al. 2002), and approximately half of the estimated 80,000 ponds for fishing or irrigation in the Coastal Plain of Georgia are man-made (Vellidis 2015). Therefore, understanding the economic, environmental and human health impacts of RWH is crucial in decision making, especially in the context of current global climate change and water scarcity.

Life cycle impacts associated with production, installation and maintenance to point-of-use of domestic and agricultural RWH systems include greenhouse gas (GHG) generation, energy use, smog, hydrologic impacts on blue water, and human and ecological health impairment (Ghimire et al. 2014). Life cycle assessment (LCA) is a widely accepted decision-support framework that assesses environmental and human health impacts of systems/products in a cradle-to-grave approach. The International Organization for Standardization (ISO 2006b; 2006a) provides LCA’s framework which includes goal and scope definition, inventory analysis, and life cycle impact assessment (LCIA). LCIA methods like the Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts, or TRACI (USEPA 2012), characterize environmental emissions utilizing inventory data such as that compiled and recommended by the U.S. Environmental Protection Agency (EPA), National Institute of Standards and Technology, and Department of Energy (NIST 2013a; NREL 2013; USEPA 2014b) in conjunction with commercially available databases. LCA thus addresses problem-shifting effects with a systematic, holistic approach that produces a comprehensive suite of environmental and human health impact indicators of sustainability. Prior LCA modeling of domestic RWH at household level and agricultural RWH at farm level found decreased environmental and human health impacts compared to conventional municipal drinking water and well water (Ghimire et al. 2014).

Previous studies have explored the feasibility and benefits of RWH. Hydrologic feasibility of watershed-wide adoption of RWH was modeled in three watersheds within the Albemarle-Pamlico river basin in North Carolina and Virginia, where a 25% adoption rate reduced downstream water yield 6%–16% (Ghimire and Johnston 2013). Other U.S. examples include economic, environmental, and water quality evaluation of RWH in the midwestern U.S. (Anand and Apul 2011; Ghimire et al. 2012; Shuster et al. 2013) and RWH performance in the southeastern U.S. (Jones and Hunt 2010; Steffen et al. 2013). In Switzerland, Crettaz et al. (1999) performed LCA of RWH for toilet flushing and recommended a combined system of conventional and low-flushing toilets for environmental benefits. An LCA of RWH for laundry use in Spain’s Mediterranean climate compared environmental impacts of various urban densities and reported favorable performance in compact density (Angrill et al. 2012). Globally, others have focused on RWH design in Taiwan (Liaw and Tsai 2004); RWH viability in Bangladesh and India (Alam et al. 2012; Van Meter et al. 2014); water saving, cost-effectiveness potential, and water quality evaluation of RWH in Sweden, Australia, and the UK (Villarreal and Dixon 2005; Eroksuz and Rahman 2010; Tam et al. 2010; Ward et al. 2012; van der Sterren et al. 2014); runoff volume reduction in China (Zhang and Hu 2014); hydrologic assessment in Africa and India (Ngigi et al. 2005; Glendenning and Vervoort 2010); RWH contamination issues (Lye 2009); and LCA environmental and cost impacts in Europe (Roebuck et al. 2011). Other countries exploring RWH include Singapore, Japan, Germany, Thailand, Indonesia, the Philippines, and Brazil (UN 2014). The economic feasibility, hydrologic feasibility, environmental impacts, energy and water savings of RWH varies with design, water demand and rainfall availability, geographic location (topography), and pumping energy needs. Moreover, an understanding of RWH impacts at wider adoption scales (such as watershed-scale) holistically is an unknown, which is the primary driver of this study.

A watershed-wide approach is preferred to address complex water resource and environmental management challenges including water supply and recreation demands, aquatic habitat protection, and water quality regulations, as well as consideration of RWH and other green infrastructure practices (USEPA 1996; 2008b). The USEPA’s watershed ecological risk assessment process guides problem formulation, analysis, and risk characterization to improve environmental decision making (Bruins and Heberling 2004; USEPA 2008a). The goal is to provide a more comprehensive perspective on watershed-wide impacts of RWH to support sustainable use of water resources and minimize unintended consequences such as creation of new problems. We focus in this study on domestic RWH for toilet flushing and agricultural RWH for crop irrigation.

### Objective and novelty

Our objective is to comprehensively address the impacts of RWH at the watershed scale, including LCIA impact categories of energy use, GHG emissions, human and ecological health, in addition to life cycle water balance and life cycle energy cost savings. The study builds upon previous work (Ghimire et al. 2014) that compared the LCIA of household-level domestic RWH with municipal drinking water for toilet-flushing and farm-level agricultural RWH with well water for crop irrigation. We analyze scaled, functional unit (normalized to 1 m3 water supply) impacts of cumulative energy demand, fossil fuel depletion, global warming potential, metal depletion, ozone depletion, acidification, smog, blue water use, green water use, ecotoxicity-total, eutrophication-total, human health criteria pollutants, human health cancer, and human health non-cancer. This study also addresses (1) the variation in domestic and agricultural RWH adoption rates; (2) differences in holistic blue water savings; and (3) differences in cumulative energy cost savings in two types of RWH practices. While gray water use is an important consideration, it was not included due to LCA data limitations.

Net water balance between reduced life cycle blue water (surface and ground water) use and harvested rainwater (green water) use, watershed-wide is also of interest. We evaluated net blue water savings by incorporating the loss in annual water yield due to watershed-wide RWH. The loss in water yield incorporates rainfall influences combining surface runoff, lateral flow, groundwater contribution (return flow from shallow aquifer), transmission losses, and pond abstractions. The difference in potential cost savings from the life cycle’s cumulative energy demand reduction are addressed for each watershed by performing life cycle cost assessment (LCCA) of an agricultural RWH system, contrasting it with the capital costs of RWH installation.

Our approach utilizes functional unit LCIA impacts and embraces systems thinking through incorporation of environmental, economic, and social dimensions of sustainability relevant to other locations. To our knowledge, no previous study has simultaneously addressed watershed-wide environmental and human health impacts and potential energy and cost savings of RWH.

## Methods

The study area includes Back Creek (152 km2 Hydrologic Unit Code or HUC # 30101010405), Sycamore Creek (41.5 km2 HUC # 30202010802), and Greens Mill Run (33.9 km2 HUC # 30201030403) within the Albemarle-Pamlico river basin in North Carolina and Virginia (Figure 1). These watersheds span a rural to urban land use gradient from the Highlands of Virginia to the Coastal Plain of North Carolina and include a total of 5,768, 10,296, and 11,582 households, and 34, two, and three farms, respectively (Table 1). The 30-year (1980–2009) mean annual precipitation is 106 cm for Back Creek, 111 cm for Sycamore Creek, and 127 cm for Greens Mill Run (Ghimire and Johnston 2013). RWH includes domestic systems for toilet-flushing equal to the number of households and agricultural systems for corn irrigation equal to the number of farms in each watershed consistent with near-optimal RWH designs from Ghimire et al. (2014). The near-optimal RWH systems minimize infrastructure through shortened pipe lengths and maximize the use of gravity for RWH delivery.

Figure 1

Study area within the Albemarle-Pamlico basin with 25% adoption rate for domestic RWH (DRWH) systems (1 dot = 100). Figure modified from Ghimire and Johnston (2013). DOI: https://doi.org/10.1525/elementa.135.f1

Table 1

Characteristics of study watersheds and number of RWH systems with 100% adoption rate, adapted from Ghimire and Johnston (2013). DOI: https://doi.org/10.1525/elementa.135.t1

Watershed Total area (km2) Average farm area (km2) Total farm area (km2) Number of agricultural RWH systems Urban area (%) Number of domestic RWH systems

Back Creek 152 0.34 11.7 34 18 5,768
Sycamore 41.5 0.42 0.91 2 49 10,296
Greens Mill 33.9 1.6 4.6 3 62 11,582

We used adoption rates of 25%, 50%, 75% and 100%, with 25% a reasonable lower rate in light of increased interest in RWH to span the potential range of adoption, consistent with prior studies (Kahinda et al. 2009; Ghimire and Johnston 2013). Thomas et al. (2014) reported that approximately 90% of respondents used harvested rainwater for irrigation, and more than 25% of the harvested water was used for potable purposes. We also consulted water resource management researchers from various organizations regarding the potential for RWH adoption in the region, including The University of Georgia, EPA Region IV, the Natural Resources Conservation Service, the U.S. Army Corps of Engineers, and the Southeast representative of the American Rainwater Catchment Systems Association.

Total household and farm numbers reflect the actual state of the watersheds. RWH systems are not currently installed and represent potential adoption scenarios. We estimated the number of agricultural RWH systems (Nt) in each watershed based on average area of a family farm (a) and actual area (AT); see Equation 1 and Table 1 (Ghimire and Johnston 2013).

(1)

Different adoption scenarios represent scaling of RWH at the specified percentage. 100% agricultural RWH adoption equates to the ratio of all available farm areas to the designated average family farm size defined by Equation 1. 50% adoption represents half of total available farm area, 0.5 × Nt. Agricultural RWH adoption represented the whole or partial farm, depending on percentage of adoption by all farms. For example, 25% adoption of total agricultural RWH systems in Sycamore Creek means that farmers utilized agricultural RWH in 25% of the available farm area. We estimated the number of domestic RWH systems equal to the household units in each watershed, obtained from the U.S. Census Bureau’s Topologically Integrated Geographic Encoding and Referencing housing units database (Table 1) (Bureau 2011). The LCA system boundary for domestic RWH and agricultural RWH is shown in Figure 2.

Figure 2

LCA system boundary for domestic rainwater harvesting (RWH), agricultural RWH, municipal drinking water (MDW), and well water from household/farm-level to watershed wide use (adapted from Ghimire et al. 2014): Water acquisition refers to RWH, surface water for MDW, and well water for irrigation. DOI: https://doi.org/10.1525/elementa.135.f2

### Watershed-wide LCA impact calculations

Change or reduction in each impact category of domestic RWH and agricultural RWH was calculated with respect to baselines of conventional municipal drinking water and well water irrigation. Equation (2) quantified the change in watershed-scale life cycle impacts, based on functional unit impacts of 1 m3 of rainwater delivery with respect to a conventional water supply:

(2)

where

• Iw = change in watershed scale impact (Units)
• A = RWH adoption rates (0.25, 0.50, 0.75, and 1)
• Nt = total number of RWH systems (Table 1)
• Qy = annual water demand for household toilet flushing or corn crop irrigation (m3/y): the annual household toilet flushing water demand was estimated at 37 m3/y using average demand at 37.8 liter per capita per day for low-flush toilet (Ghimire et al. 2014) and crop irrigation demand estimated at 90,900 m3/y using 606 m3/day for 150 days per year irrigation (Ghimire and Johnston 2013)
• T = service life of a RWH system (50 y)
• Δi = iconirwh = difference in impact per 1 m3 of rainwater delivery, irwh, with respect to conventional water supply, icon, (Units/m3), where Units are as follows: cumulative energy demand (MJ), fossil fuel depletion (kg oil eq), global warming potential (kg CO2 eq), metal depletion (kg Fe eq), ozone depletion (kg CFC11 eq), acidification (kg SO2 eq), smog (kg O3 eq), blue water use (m3), green water use (m3), ecotoxicity-total (CTU or comparative toxic units), eutrophication-total (kg N eq), human health criteria pollutants (kg PM2.5 eq), human health cancer (CTU), and human health non-cancer (CTU).

### Life cycle assessment (LCA) methods

A prior study (Ghimire et al. 2014) briefly summarized here provided the functional unit LCIA impacts (per m3 water delivery) of domestic RWH at household level, agricultural RWH at farm level, conventional municipal drinking water for toilet flushing, and well water for irrigation. Calculations of impact per functional unit (i.e., icon and irwh in Equation 1) were performed using OpenLCA (OpenLCA 2013), a publicly available software linked to the Ecoinvent database and the EPA’s LCIA methods. LCIA methods incorporated blue and green water use from water footprint methods (WFN 2009), fossil and metal depletion methods from ReCIPE (Goedkoop et al. 2012), and the non-renewable cumulative energy demand method from Ecoinvent version 2.2 (Hischier et al. 2010) within TRACI version 2.0 (USEPA 2013). TRACI’s characterization of human toxicity and ecotoxicity was adapted from the USETox Model and adjusted to remove characterization of metal toxicity due to uncertainty (Bare 2002; Hauschild et al. 2008; Ghimire et al. 2014).

Details on each impact category in TRACI are available from Bare (2011) and the USEPA (2012). TRACI utilizes the amount of chemical emission or resource used to estimate potential impact of each category, as described by Equation 3:

(3)

where,

• for a specific impact category, j:
• Ij = the potential impact of all chemicals (k)
• CFkm.j = the characterization factor of chemical (k) emitted to media (m)
• Mkm = the mass of chemical (k) emitted to media (m)

Our method (Equation 2) is based on linear scaling of the functional unit impacts compared to conventional water supplies. It should be noted that virtually all LCA studies (and databases) utilize simplified models of linear form, and normalizing unit process flows to functional unit is a widely accepted LCA practice (Heijungs and Suh 2002). Heijungs and Suh (2002) considered this concept mathematically by describing the scaling vector s as a function of external demand (output flow) vector d and the inverse of flow matrix T (flows) in estimating unit process in a system, or s = T–1d. Economies of scale arising from wider adoption of RWH systems are beyond the scope of this study.

### Life cycle inventory

Life cycle inventory (LCI) details including data inventory sources, design parameters, and underlying assumptions were described in a prior study (Ghimire et al. 2014), with near-optimal RWH system designs minimizing infrastructure through shortened pipe lengths and maximizing gravity for delivery. The near-optimal domestic RWH systems used a polyethylene storage tank of 6.2 m3, reduced pipe length (5 m) of chlorinated polyvinyl chloride, and eliminated the pump and pumping energy. Similarly, the near-optimal agricultural RWH used polyvinyl chloride pipe of 150 m, polyethylene storage tank of 606 m3, and no pump or pumping energy. The Building for Environmental and Economic Sustainability database (NIST 2013a) and a European database (Ecoinvent 2012) were used to complete the materials inventory.

### Life cycle water balance

We evaluated net life cycle blue water (surface and ground water) savings by incorporating the loss in annual water yield due to watershed-wide RWH. The loss in water yield incorporates rainfall influences combining surface runoff, lateral flow, groundwater contribution (return flow from shallow aquifer), transmission losses, and pond abstractions. Net water balance between reduction in life cycle blue water use and green water use is described (Equation 4):

${W}_{b}=B{U}_{r}-\mathit{\text{GU}}$
(4)

where

• Wb = life cycle water balance in a watershed (Mm3)
• BUr = net reduction in life cycle blue water use (Mm3)
• GU = life cycle green water use (i.e., harvested rainwater) (Mm3)

Life cycle water balance, Wb, is estimated utilizing the watershed scale savings in blue water and green water obtained from Equation 2. Equation 4 estimates reduction in embedded blue water in all life cycle stages of RWH compared to green water harvesting over a 50 year lifetime. Blue water consumption or withdrawal does not include release to the same water body. Equation 4 also does not include rainfall influences. We integrate rainfall influence on life cycle water balance using watershed annual water yield loss, i.e., the combination of surface runoff, lateral flow, groundwater contribution, transmission losses, and pond abstractions in each watershed for 100% domestic and agricultural RWH adoption (Equation 5):

${B}_{H}=B{U}_{y}-{W}_{a}×{R}_{y}$
(5)

where,

• BH = annual blue water savings due to domestic or agricultural RWH, compared to conventional water supplies, by integrating rainfall (m3/y)
• BUy = annual net savings in life cycle blue water use due to domestic or agricultural RWH, compared to conventional water supplies (m3/y) (see Tables 2 and 3)
• Wa = watershed area contributing to total water yield (m2) (see Table 1)
• Ry= the loss in annual water yield due to domestic or agricultural RWH, compared to no-RWH annual water yield (m3/y) (Equation 6):

${R}_{y}={\Sigma }_{i=1}^{12}{Y}_{i.0}-{\Sigma }_{i=1}^{12}{Y}_{i.R}$
(6)

where,

• Yi.0 is the monthly water yield for No-RWH and Yi.R is the monthly water yield for the specified RWH adoption within the watershed (m/month). Water yield, Yi for each RWH system, is defined as the net amount of water leaving the watershed during the monthly time step (Equation 7) obtained from Ghimire and Johnston (2013):

(7)

where, for monthly time step,

• Yi = net amount of water leaving the watershed during the monthly time step (m/month)
• S = surface runoff contribution to streamflow (mm/month)
• L = lateral flow (within the soil profile) contribution to streamflow (mm/month)
• G = groundwater contribution (return flow from shallow aquifer) to streamflow (mm/month)
• R = transmission losses (becomes recharge for the shallow aquifer) from tributary channels via transmission through the bed (mm/month)
• P = pond abstractions (net change in water volume of pond) (mm/month)

Table 2

Reduction in watershed scale life cycle impacts of domestic RWH adoption, with respect to conventional municipal drinking water, in three watersheds: Back Creek = BC, Sycamore Creek = SC, and Greens Mill = GM. CTU = Comparative Toxic Units. DOI: https://doi.org/10.1525/elementa.135.t2

LCA impact category Unit Reduction in watershed-scale impacts of domestic RWH adoption in three watersheds
% Reduction of impacts
100% 75% 50% 25%

BC SC GM BC SC GM BC SC GM BC SC GM

Energy Demand TJ 85 152 171 64 114 128 42 76 85 21 38 43 57.8
Fossil Fuel Depletion Mg oil eq 1068 1906 2144 801 1429 1608 534 953 1072 267 476 536 42.8
Global Warming Mg CO2 eq 4667 8331 9372 3501 6248 7029 2334 4166 4686 1167 2083 2343 51.7
Metal Depletion Mg Fe eq 50 89 100 37 66 75 25 44 50 12 22 25 13.2
Ozone Depletion Mg CFC11 eq 0.0004 0.0008 0.0008 0.0003 0.0006 0.0006 0.0002 0.0004 0.0004 0.0001 0.0002 0.0002 81.8
Acidification Mkg H+ mole eq 62 110 124 46 83 93 31 55 62 15 28 31 90.1
Smog Mkg O3 eq 305 544 612 229 408 459 152 272 306 76 136 153 61.3
Green Water Use Mm3 –11 –19 –21 –8 –14 –16 –5 –10 –11 –3 –5 –5 --
Blue Water Use Mm3 13 23 26 10 17 19 6 11 13 3 6 6 99.9
Ecotoxicity CTU –1069 –1908 –2146 –802 –1431 –1610 –534 –954 –1073 –267 –477 –537 –15.9
Eutrophication Mkg N eq 39 70 79 30 53 59 20 35 40 10 18 20 90.3
Human Health Criteria Mkg PM2.5 eq 4 7 8 3 6 6 2 4 4 1 2 2 88.0
Human Health, Cancer CTU 0.00004 0.00008 0.00009 0.00003 0.00006 0.00007 0.00002 0.00004 0.00004 0.00001 0.00002 0.00002 13.9
Human Health, NonCancer CTU 0.0005 0.0010 0.0011 0.0004 0.0007 0.0008 0.0003 0.0005 0.0005 0.0001 0.0002 0.0003 84.6

Table 3

Reduction in watershed scale life cycle impacts of agricultural RWH adoption, with respect to conventional well water, in three watersheds: Back Creek = BC, Sycamore Creek = SC, and Greens Mill = GM. CTU = Comparative Toxic Units. DOI: https://doi.org/10.1525/elementa.135.t3

LCA impact category Unit Reduction in watershed-scale impacts of agricultural RWH adoption in three watersheds
% Reduction of impacts
100% 75% 50% 25%

BC SC GM BC SC GM BC SC GM BC SC GM

Energy Demand TJ 838 49 74 629 37 55 419 25 37 210 12 18 77.7
Fossil Fuel Depletion Mg oil eq 15032 884 1326 11274 663 995 7516 442 663 3758 221 332 77.6
Global Warming Mg CO2 eq 40369 2375 3562 30277 1781 2671 20185 1187 1781 10092 594 890 75.8
Metal Depletion Mg Fe eq 2018 119 178 1514 89 134 1009 59 89 505 30 45 20.1
Ozone Depletion Mg CFC11 eq 0.00210 0.00012 0.00019 0.00158 0.00009 0.00014 0.00105 0.00006 0.00009 0.00053 0.00003 0.00005 67.2
Acidification Mkg H+ mole eq 265 16 23 199 12 18 133 8 12 66 4 6 86.7
Smog Mkg O3 eq 2669 157 236 2002 118 177 1335 79 118 667 39 59 79.5
Green Water Use Mm3 –154.5 –9 –14 –116 –7 –10 –77 –5 –7 –39 –2 –3 ---
Blue Water Use Mm3 154.7 9 14 116 7 10 77 5 7 39 2 3 99.9
Ecotoxicity CTU 123602 7271 10906 92702 5453 8180 61801 3635 5453 30901 1818 2727 71.8
Eutrophication, total Mkg N eq 147 9 13 111 7 10 74 4 7 37 2 3 76.9
Human Health Criteria Mkg PM2.5 eq 19.2 1.1 1.7 14.4 0.8 1.3 9.6 0.6 0.8 4.8 0.3 0.4 66.8
Human Health, Cancer CTU 0.00142 0.00008 0.00013 0.00107 0.00006 0.00009 0.00071 0.00004 0.00006 0.00036 0.00002 0.00003 58.3
Human Health, NonCancer CTU 0.0062 0.0004 0.0005 0.0047 0.0003 0.0004 0.0031 0.0002 0.0003 0.0016 0.0001 0.0001 62.3

Yis (Equation 6) were obtained from Ghimire and Johnston (2013) who used the Soil and Water Assessment Tool (SWAT) model developed by the U.S. Department of Agriculture to simulate change in water yield (Yis) of domestic and agricultural RWH systems within the three watersheds (Arnold et al. 2011). Thirty-year (1980–2009) daily rainfall data (NCDC 2012) were used as input, and for details on SWAT-based RWH model development, calibration and validation, see Ghimire and Johnston (2013).

### Life cycle cost savings due to cumulative energy demand reduction

Energy cost savings from watershed-wide life cycle cumulative energy demand reduction are provided for potential cumulative energy savings (Equation 8):

${C}_{s}=\text{2.78}×{10}^{\text{5}}×{P}_{e}×{E}_{s}$
(8)

where,

• Cs = potential energy cost savings in $• Es = cumulative energy demand savings (TJ) (obtained from Equation 2) • Pe = energy price, ($0.103/kWh) (USEIA 2014)
• 2.78 × 105 = a conversion factor (i.e., 1TJ = 2.78 × 105 kWh)

The cumulative energy savings (Es) for domestic RWH and agricultural RWH systems adopted to watershed-scale were calculated using Equation 2, and corresponding potential energy cost savings Cs were estimated by Equation 8.

Combined with life cycle costs, LCIA impact reduction provides a more comprehensive perspective on economic viability of agricultural RWH systems. We discuss the economic viability of RWH by combining the initial capital cost of agricultural RWH system with the cumulative life cycle energy savings (see Supplementary Material for LCCA details). The economic viability of agricultural RWH involves combining the cumulative energy cost savings with life cycle costs of installing, replacing, operating and maintaining a near-optimal agricultural RWH system. The contributions of cumulative energy cost savings to recouping RWH system life cycle costs are discussed by comparing them to conventional irrigation water prices. Economic viability of domestic RWH was not included because it is not a primary consideration for its adoption.

We also performed sensitivity analysis of future discount rates to LCCA, from 3% to 10%, similar to the Department of Energy’s minimum and maximum real discount rates, as stated by the Code of Federal Regulations 10 CFR 436, Federal Energy Management and Planning Programs (§ 436.14, page 521) (Register 1999).

## Results and discussion

Compared to conventional water supply watershed-wide domestic RWH adoption in the three watersheds reduced the 14 LCIA impact scores, except for ecotoxicity impact of domestic RWH (Table 2). The exception in ecotoxicity impact was consistent with the greater functional unit (per household) ecotoxicity score at 7.3 × 10–04 Comparative Toxic Units/m3 (CTU/m3) for near-optimal domestic RWH than that of the conventional municipal water supply at 6.3 × 10–04 CTU/m3 as reported by Ghimire et al. (2014). The ecotoxicity factor CTU is an estimate of the potentially affected fraction of species integrated over time and volume, per unit mass of a chemical emitted. The characterization factor for human toxicity impacts is the estimated increase in morbidity in the total human population, per unit mass of a chemical emitted, assuming equal weighting between cancer and non-cancer.

Agricultural RWH adoption in the three watersheds reduced all 14 LCIA impact scores compared to conventional well water irrigation (Table 3). Impact reductions varied with RWH adoption rates and the watershed in consideration: domestic RWH impact reductions for Sycamore and Greens Mill were twice those of Back Creek, and reductions due to agricultural RWH were 17 times higher for Back Creek than Sycamore. Importantly, savings from agricultural RWH for life cycle energy demand were 10 times greater than domestic RWH in Back Creek watershed, but lower (~0.4 times domestic RWH savings) in Sycamore and Greens Mill, offering the opportunity to provide economic incentives (e.g., tax incentives) for agricultural RWH adoption in a watershed with a greater number of farms (Tables 2 and 3).

For 25% domestic RWH adoption rate, impact reductions ranged from 21–43 TJ cumulative energy demand, 267–536 Mg oil eq fossil fuel depletion, 1167–2343 Mg CO2 eq. global warming potential, 15–31 Mkg H+ mole eq. acidification, 76–153 Mkg O3 eq. smog, 3–6 Mm3 blue water use, 10–20 Mkg N eq. eutrophication, and 1–2 Mkg PM2.5 eq. human health criteria pollutants (Table 2). With a 25% adoption of agricultural RWH, the impact reductions ranged from 12–210 TJ cumulative energy demand, 594–10092 Mg CO2 eq. global warming potential, and 39–2 Mm3 blue water use (Table 3). Reductions were higher in Back Creek for agricultural RWH than domestic (from 4 – 41 times) but lower in the other two watersheds (from 10% – 60% of domestic RWH reductions), except for metal depletion, ecotoxicity, and human health-cancer impacts (Tables 2 and 3). The reduction in other environmental dimensions such as ecotoxicity impact due to 100% agricultural RWH watershed-wide over the lifetime of 50 years was substantial at 123602 CTU in Back Creek watershed, although it was negative (at –1069 CTU) due to 100% domestic RWH. Similarly, life cycle reduction in smog impact (as ground level ozone) due to 100% agricultural RWH adoption in Back Creek watershed was 2669 Mkg O3 eq., much higher than domestic RWH at 305 Mkg O3 eq. Larger agricultural RWH reductions in Back Creek are due to more farms, and larger domestic RWH reductions in Greens Mill watershed are due to more households. The reduction in watershed scale LCIA score varied by adoption rates, watersheds, and number and type of RWH systems; however, the % reduction differed only by the type of RWH as demonstrated by Equation 2. Percent reduction of watershed-scale domestic RWH impact with respect to conventional municipal water supply ranged from 13% (metal depletion) to 99.9% (blue water use) (Table 2). Similarly, % reduction of watershed-scale agricultural RWH impact with respect to conventional well water irrigation ranged from 20% (metal depletion) to 99.9% (blue water use) (Table 3).

### Differences in life cycle water balance

Agricultural RWH green water use offset blue water use 1:1 (~155 Mm3:155 Mm3). While agricultural RWH blue water use reductions were 12 times those of the domestic RWH in Back Creek, agricultural RWH reductions were lower (~0.5 times) than domestic RWH reductions in Sycamore and Greens Mill watersheds, due to more households and fewer farms (Tables 2 and 3). Life cycle water balance analysis also showed greater water balance in urban areas with more households and in the rural areas with more farms (Equation 4 and Figures 3 and 4).

Figure 3

Life cycle water balance for (a) domestic RWH adoption of 25%, 50%, 75%, and 100% and (b) agricultural RWH adoption of 25%, 50%, 75%, and 100%. DOI: https://doi.org/10.1525/elementa.135.f3

Figure 4

Captured life cycle green water per unit watershed area for (a) domestic RWH adoption of 25%, 50%, 75%, and 100% (b) agricultural RWH adoption of 25%, 50%, 75%, and 100%. DOI: https://doi.org/10.1525/elementa.135.f4

Life cycle water balance for 25% domestic RWH adoption ranged from 0.5 – 1.1 Mm3 (Figure 3a), suggesting increased surface water availability for downstream ecosystems. This is of particular interest in the light of downstream water availability reduction by 6% due to 25% RWH adoption in Back Creek watershed (Ghimire and Johnston 2013) because from a life cycle point of view, this reduction can be offset by life cycle blue water savings. Also, the water balance analysis for agricultural RWH (Figure 3b) revealed that reducing blue water use approximately equals harvested rainwater, offsetting ground water irrigation. For maximum adoption rates of 100%, the estimated life cycle green water use (or captured rainwater) per unit watershed area for domestic RWH in Greens Mill and agricultural RWH in Back Creek were at 0.63 Mm3/km2 and 1.02 Mm3/km2 (Figure 4a, b).

Potential life cycle blue water cost savings from RWH are considerable at larger geographic scales. Considering a domestic water price of $1.6/m3 (Mayer et al. 1999), potential maximum life cycle cost savings from domestic RWH blue water use reduction (26 Mm3) over a lifetime of 50 years are$41.1M for Green Mills watershed. Considering an irrigation water price of $0.1/m3 (Wichelns 2010), maximum potential life cycle blue water cost savings from agricultural RWH (reduced blue water use 155 Mm3) are$15.5M for Back Creek watershed. Savings increase with rising tiered water price structures, increasing block rates as the water use increases (USEPA 2016). In Atlanta, GA, for example, the average monthly water rate for a family of four using 189.3 liters (50 gallons) per capita per day and 567.8 liters (150 gallons) per capita per day were at $1.76/m3 and$2.05/m3 in 2015 (Walton 2015), and the life cycle blue water cost savings can be even higher for these water rates. In a 2015 survey of 30 major U.S. cities, the household water bill had increased 41% since 2010 (Walton 2015). The survey also reported that, in 2015, the highest monthly water bill for a family of four using 378.5 liters (100 gallons) per person per day was $153.78/family-month in Santa Fe, New Mexico, equating to$3.39/m3,

and the lowest rate was $0.51/m3 in Fresno, California. Note that the life cycle water balance addressed the net savings of embedded blue water by accounting for green water use in all life cycle stages of RWH systems during a service life of 50 years. Further analysis incorporating rainfall effects in life cycle water balance (Equation 5) revealed greater annual holistic blue water savings for 100% agricultural RWH for Back Creek (2.4 Mm3/y) and Sycamore Creek (0.1 Mm3/y), due to more farms and lower water yield losses. Negative savings were found in Greens Mill due to higher water yield loss and fewer farms (Figure 5a–c). For 100% domestic RWH, holistic blue water use reduction in Greens Mill watershed was higher than others (0.1 Mm3/y) due to more domestic RWH systems (Figure 5a). We need to draw an important distinction between LCA impacts and life cycle water impacts. LCA impacts can be combined; however, agricultural and domestic RWH life cycle water balance impacts were modeled separately using the SWAT model. Holistic blue water balance integrating water yields is not additive (100% DRWH and 100% ARWH cannot be combined). Figure 5 (a) annual water loss, (b) annual water yield, (c) holistic annual water balance for 100% domestic RWH and agricultural RWH adoption in three watersheds. DOI: https://doi.org/10.1525/elementa.135.f5 ### Differences in cumulative energy cost savings While domestic RWH in Greens Mill realized 57.8% cumulative energy savings when compared to municipal drinking water, agricultural RWH in Back Creek realized 77.7% cumulative energy savings compared to well water irrigation. Cumulative energy savings ranged from 43–171 TJ domestic RWH to 210–838 TJ agricultural RWH savings at adoption rates of 25% and 100% (Figure 6a). Potential energy cost savings (Equation 8) from watershed-wide life cycle cumulative energy use reduction over a lifetime of 50 years are in the millions of dollars. Potential maximum lifetime energy cost savings were estimated at$5M and $24M corresponding to domestic RWH in Greens Mill and agricultural RWH in Back Creek watershedsy (Figure 6b), which are substantial amounts that indicate potential economic viability. This savings is approximately$421/household over the lifetime of domestic RWH systems in Greens Mill. Note that this value does not reflect the potential savings realized by municipalities who may be required to upgrade or replace conventional infrastructure. In Back Creek watershed, $705,515/farm would be realized over the lifetime of agricultural RWH systems. On an annual basis, potential maximum energy coss savings for domestic RWH in Greens Mill was$8.42/household and was $14,110/farm for agricultural RWH in Back Creek. Figure 6 (a) Cumulative energy savings and (b) life cycle cost savings due to cumulative energy reductions by domestic RWH (DRWH) and agricultural RWH (ARWH) adoption rates in three watersheds: Back Creek (BC), Sycamore Creek (SC), and Greens Mill (GM). DOI: https://doi.org/10.1525/elementa.135.f6 The economic viability of RWH is further investigated by combining cumulative energy cost savings with initial capital costs and life cycle costs of installing, replacing, operating and maintaining an agricultural RWH system. We estimated initial capital costs of an agricultural RWH system at$271,314 (details in Supplemental Material), equal to $0.060 per m3 rainwater supply for irrigation of 4,545,000 m3. This price matches approximately the median conventional irrigation water price that ranges from$0.005/m3 to $0.10/m3 in the U.S. (Wichelns 2010). We estimated life cycle costs of an agricultural RWH system at$0.11/m3 (details in Supplemental Material). The life cycle energy cost savings are estimated at $0.16/m3 using$6M lifetime savings at the 25% adoption rate of agricultural RWH in Back Creek watershed, i.e., $6M/(0.25 × 34 farms × 4545000 m3/farm) (Figure 6b). The difference in estimated life cycle cost of rainwater for agricultural irrigation ($0.11/m3) and conventional irrigation ($0.10/m3) can be recouped by life cycle cumulative energy cost savings at$0.16/m3. LCCA of domestic RWH was not performed because it is not a primary consideration for its adoption.

The viability of increased RWH adoption is influenced by water requirements, system design, material and energy throughput, and water price, as well as future discount rates (Ghimire et al. 2012). The real discount rate at 3% (i = 0.03) was used in all life cycle cost calculations as suggested by the National Institute of Standards and Technology (NIST 2013b). Sensitivity analysis of future real discount rates (3% – 10%) on life cycle costs of agricultural RWH per functional unit of water delivery (1 m3) suggested lower RWH life cycle cost with higher discount rates. Agricultural RWH water price ranged from $0.08/m3 to$0.11/m3, corresponding to 10% and 3% real discount rates (Figure 7). Although cost savings evaluation for remaining impacts and other RWH uses is beyond the scope of this study, environmental and human health impact reductions contribute positively to the overall viability of RWH.

Figure 7

Sensitivity of life cycle costs of agricultural RWH to real discount rates. DOI: https://doi.org/10.1525/elementa.135.f7

### Energy use and GHG emissions

Energy supplies (coal, natural gas, oil) are the largest source of global GHG emissions (USEPA 2014a), and energy reductions result directly in reduced emission contributions. RWH indicated potential cumulative energy savings as well as reductions in GHG emissions. Total U.S. GHG emissions in 2012 were 6,526 M Metric Tons of CO2 eq., with agriculture accounting for approximately 10% (USEPA 2014b). The annual GHG reduction per farm was 24 Metric Tons of CO2 eq/Farm-Year (i.e., 40369 Metric Tons of CO2 eq/34 farms × 50 y), with substantial reductions possible watershed-wide. Maximum potential GHG emission reduction due to agricultural RWH for the three watersheds varied from 3562–40369 Metric Tons of CO2 eq. (Table 3). Life cycle emission reduction through agricultural RWH adoption may have the co-benefit of helping to fulfill U.S. commitments to the United Nations Framework Convention on Climate Change (USEPA 2014b) when considered at the watershed scale.

The methodology is generally applicable to other regions with comparable watershed characteristics, installation costs, and treatment technologies, but it requires appropriate data such as life cycle inventory of used material and energy. Watersheds with higher water yields due to greater rainfall can contribute to water resource sustainability by offsetting water yield losses over life cycle blue water savings of RWH. However, variation in land use/cover type, topography, geology, land surface evaporation, water routing, and quick flow versus impoundments and the social preferences of water sources (e.g., conventional water supplies versus RWH) influence holistic blue water savings but are beyond the scope of this study. We also note that grey and black water (i.e., municipal wastewater) reuse can also increase water resource sustainability, considering that annual coastal wastewater discharge is approximately 6% of the U.S. total estimated water use (NRC 2012). Grey and black water reuse were not considered due to data limitations.

As climate change becomes more severe and communities across the southeastern U.S. experience more frequent droughts (USGCRP 2012), interest in RWH as a climate change adaptation strategy is expected to increase. Attention must be given not only to potential benefits but also to human health and environmental impacts. This study utilized a near-optimal design that minimized material and energy inputs, consistent with Ghimire et al. (2014). Potential energy savings and impact reductions vary with water treatment processes and costs, annual rainfall and water demand, and regulatory standards and design requirements. Tradeoffs may also exist between other irrigation water sources such as rivers versus well water, primarily due to pumping energy caused by spatial variation from pumping distance and necessary pump head.

## Conclusions

Societies across the U.S. and around the world require reliable water supply in the face of population growth and climate change. RWH is being recognized not only as a viable decentralized water supply option for augmenting modern centralized water infrastructures that are costly and resource intensive but also as a green infrastructure strategy for climate change adaptation. Policies are changing to encourage RWH adoption, but RWH sustainability (including unintended environmental consequences and economic viability) should be well studied and across scales. We provided a holistic assessment of environmental and economic viability of domestic and agricultural RWH at the watershed scale in the southeastern U.S., evaluating differences in holistic water savings, life cycle energy cost savings, and GHG emission reductions of both RWH systems in addition to life cycle costs of agricultural RWH.

Watershed-scale impact reductions due to agricultural RWH were 17 times higher for Back Creek than Sycamore due to more farms, and domestic RWH impact reductions for Sycamore and Greens Mill were twice those of Back Creek due to more households. Domestic RWH reduced life cycle environmental and human health impacts compared to municipal drinking water, ranging from 57.8% cumulative energy demand to 99.9% blue water use. Agricultural RWH also reduced the environmental and human health impacts compared to well water, ranging from 20% metal depletion to 99.9% blue water use. Life cycle costs combined with impact reductions, and benefits from shifting to green water over blue water use translates directly to water resource sustainability in terms of return on investment and blue water availability by offsetting surface and ground water consumption. While the difference in estimated life cycle cost of rainwater for agricultural irrigation versus conventional irrigation cost can be recouped by life cycle cumulative energy savings at \$0.16/m3, the holistic water balance analysis revealed that reducing blue water use approximately equals harvested rainwater. Increased green water use over blue water, reduced energy demand, savings in life cycle energy costs, and decreased global warming potential indicate the potential of RWH as a sustainable water resource management strategy.

Policies encouraging RWH practices, allowing, defining, and clarifying RWH use for various purposes due to climate change are increasing across the U.S.oin Texas, Ohio, Virginia, North Carolina, Illinois, and California, and internationally to Australia, the U.K., Canada, and India (TWDB 2006; GANC 2009; EA 2010; VDH 2011; NCSL 2014). While many states allow rainwater for non-potable purposes, states such as Ohio and Texas allow the practice for potable purposes as well (NCSL 2016). This study more comprehensively evaluates potential economic and environmental benefits of RWH at a range of adoption rates, 25%, 50%, 75%, and 100% and identifies potential economic savings. RWH adoption, even at the lower end of the spectrum, has positive implications for water balance, energy savings, and GHG emission reductions. This systems approach to impact assessment also avoids unanticipated effects of cumulative impacts between different product systems (e.g., RWH components) which can lead to unintended consequences.

There are limits to our study related to inherent limitations of LCA. Attribution of ecological and human health impacts to a specific geographic scale, i.e., local versus national or regional, is a known limitation of current LCA methodology. As an example, eutrophication, although considered regional, may or may not be realized within the assessment area boundary due to the use of surrogate data from Ecoinvent. This study builds on the previous LCI database (Ghimire et al. 2014), Building for Environmental and Economic Sustainability (BEES) (NIST 2013a), U.S. LCI (NREL 2013), and a European database (Ecoinvent 2012). Although use of surrogate data in the life cycle inventory is an accepted and necessary practice, the result is that global impacts such as global warming and ozone depletion are generally considered representative of that scale, but local, regional and national impacts must be qualified by the data relied upon. Future research should focus on the data and methods needed to attribute impacts to these various scales of exposure. Other issues such as regulatory requirements on water quality, water withdrawal, and plumbing codes should also be considered. This study intended to inform decentralized RWH decision making for sustainable water management from an increase in heavy downpours to more frequent droughts.

## Abbreviations

 Acronym Description BC Back Creek BU Blue water use CFR Code of Federal Regulations CTU Comparative toxic units EPA U.S. Environmental Protection Agency GHG Greenhouse gas GM Greens Mill GU Green water use HUC Hydrologic Unit Code LCA Life cycle assessment LCCA Life cycle cost assessment LCI Life cycle inventory LCIA Life cycle impact assessment MDW Municipal drinking water NIST National Institute of Standards and Technology RWH Rainwater harvesting SC Sycamore Creek SM Supplemental Material SWAT Soil and Water Assessment Tool TRACI Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts WW Well water

## Acknowledgements

Although this document has been reviewed in accordance with EPA policy and approved for publication, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. A poster on the preliminary findings of this study was presented at the World Environmental & Water Resources Congress 2014, Portland, Oregon. We thank Fran Rauschenberg (Senior Environmental Employee) and Donnie Vineyard (ORISE Research Participant) for technical editing and review comments that improved the manuscript. We thank George Vellidis and Seth Wenger (The University of Georgia (UGA)), James Hook (UGA retired), Christine McKay (EPA Region IV), and Steve Williams (southeast representative of the American Rainwater Catchment Systems Association) for their help in obtaining RWH and impoundment information and policy insights in Georgia and the Southeast. We also thank three anonymous reviewers for their helpful comments that improved the manuscript.

## Funding information

This research was supported in part by an appointment to the Postdoctoral Research Program at the U.S. Environmental Protection Agency (EPA), Office of Research and Development, Athens, GA, USA, administered by the Oak Ridge Institute for Science and Education through Interagency Agreement Number DW8992298301 between the U.S. Department of Energy and the U.S. EPA.

## Competing interests

The authors have no competing interests to declare.

## Contributions

Santosh R. Ghimire (SRG) wrote the first draft and John M. Johnston (JMJ) contributed to the revision and development of the final manuscript. SRG performed modeling, analyzed the results, and created figures and tables with guidance from JMJ. Both authors reviewed and approved the final manuscript.

## References

1. Alam R, Munna G, Chowdhury MAI, Sarkar MSKA, Ahmed M, Rahman MT, Jesmin F and Toimoor MA 2012. Feasibility study of rainwater harvesting system in Sylhet City. Environmental Monitoring and Assessment 184: 573–580, DOI: http://dx.doi.org/10.1007/s10661-011-1989-7

2. Anand C and Apul DS 2011. Economic and environmental analysis of standard, high efficiency, rainwater flushed, and composting toilets. Journal of Environmental Management 92: 419–428, DOI: http://dx.doi.org/10.1016/j.jenvman.2010.08.005

3. Angrill S, Farreny R, Gasol CM, Gabarrell X, Vinolas B, Josa A and Rieradevall J 2012. Environmental analysis of rainwater harvesting infrastructures in diffuse and compact urban models of Mediterranean climate. International Journal of Life Cycle Assessment 17: 25–42, DOI: http://dx.doi.org/10.1007/s11367-011-0330-6

4. ARCSA 2014. The American Rainwater Catchment Systems Association [Internet] In: USA: The American Rainwater Catchment Systems Association. Available from: http://www.arcsa.org/.

5. Arnold J, Kiniry J, Srinivasan R, Williams J, Haney E and Neitsch S 2011. Soil and Water Assessment Tool Input/Output File Documentation Version 2009 In: Texas, USA: Texas Water Resources Institute, Texas A & M University.

6. Bare J 2011. TRACI 2.0: the tool for the reduction and assessment of chemical and other environmental impacts 2.0. Clean Technologies and Environmental Policy 13: 687–696, DOI: http://dx.doi.org/10.1007/s10098-010-0338-9

7. Bare JC 2002. TRACI The Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts. Journal of Industrial Ecology 6: 49–78, DOI: http://dx.doi.org/10.1162/108819802766269539

8. Brauman KA, Richter BD, Postel S, Malsy M and Flörke M 2016. Water depletion: An improved metric for incorporating seasonal and dry-year water scarcity into water risk assessments. Elementa: Science of the Anthropocene 4: 000083.DOI: http://dx.doi.org/10.12952/journal.elementa.000083

9. Bruins RJ and Heberling MT 2004. Economics and ecological risk assessment: Applications to watershed management In: CRC Press, DOI: http://dx.doi.org/10.1201/9780203486603

10. Bureau USC 2011. Census Block Shapefiles with 2010 Census Population and Housing Unit Counts, Topologically Integrated Geographic Encoding and Referencing system [Internet] In: USA: U.S. Census Bureau.

11. Chapagain AK and Hoekstra AY 2004. Water footprints of nations.

12. Crettaz P, Jolliet O, Cuanillon JM and Orlando S 1999. Life cycle assessment of drinking water and rain water for toilets flushing. Journal of Water Services Research and Technology-Aqua 48: 73–83, DOI: http://dx.doi.org/10.1046/j.1365-2087.1999.00134.x

13. DeBusk K, Hunt W and Wright J 2013. Characterizing Rainwater Harvesting Performance and Demonstrating Stormwater Management Benefits in the Humid Southeast USA. JAWRA Journal of the American Water Resources Association 49: 1398–1411, DOI: http://dx.doi.org/10.1111/jawr.12096

14. EA 2010. Harvesting rainwater for domestic uses: an information guide In: Almondsbury, Bristol: Environment Agency.

15. Ecoinvent 2012. Ecoinvent Centre In: Switzerland: Ecoinvent Centre.

16. Eroksuz E and Rahman A 2010. Rainwater tanks in multi-unit buildings: A case study for three Australian cities. Resources, Conservation and Recycling 54: 1449–1452, DOI: http://dx.doi.org/10.1016/j.resconrec.2010.06.010

17. GANC 2009. The General Assembly of North Carolina, SESSION LAW 2009–243, HOUSE BILL 749 In: NC, USA: The General Assembly of North Carolina.

18. Ghimire SR and Johnston JM 2013. Impacts of domestic and agricultural rainwater harvesting systems on watershed hydrology: A case study in the Albemarle-Pamlico river basins (USA). Ecohydrology & Hydrobiology 13: 159–171, DOI: http://dx.doi.org/10.1016/j.ecohyd.2013.03.007

19. Ghimire SR, Johnston JM, Ingwersen WW and Hawkins TR 2014. Life Cycle Assessment of Domestic and Agricultural Rainwater Harvesting Systems. Environmental Science & Technology 48: 4069–4077, DOI: http://dx.doi.org/10.1021/es500189f

20. Ghimire SR, Watkins DW Jr. and Li K 2012. Life cycle cost assessment of a rain water harvesting system for toilet flushing. Water Science & Technology: Water Supply 12: 309–320, DOI: http://dx.doi.org/10.2166/ws.2011.135

21. Glendenning CJ and Vervoort RW 2010. Hydrological impacts of rainwater harvesting (RWH) in a case study catchment: The Arvari River, Rajasthan, India. Part 1: Field-scale impacts. Agricultural Water Management 98: 331–342, DOI: http://dx.doi.org/10.1016/j.agwat.2010.09.003

22. Goedkoop MHR, Huijbregts M, Schryver A, Struijs J and Zelm R 2012. ReCiPe 2008 – A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level In: Netherlands: Ministry of Housing, Spatial Planning and Environment.

23. Hauschild MZ, Huijbregts M, Jolliet O, Macleod M, Margni M, van de Meent D, Rosenbaum RK and McKone TE 2008. Building a Model Based on Scientific Consensus for Life Cycle Impact Assessment of Chemicals: The Search for Harmony and Parsimony. Environmental Science & Technology 42: 7032–7037, DOI: http://dx.doi.org/10.1021/es703145t

24. Heijungs R and Suh S 2002. The computational structure of life cycle assessment In: Springer Science & Business Media, DOI: http://dx.doi.org/10.1007/978-94-015-9900-9

25. Hischier RAH-JB C, Büsser S, Doka G, Frischknecht R, Kleijer A, Leuenberger M, Nemecek T and Simons A 2010. Documentation of changes implemented in ecoinvent Data v2.2 In: Switzerland: Swiss Centre for LCI. No. Final report ecoinvent data v2.2 Dübendorf.

26. Hoekstra A and Hung P 2002. Virtual Water trade: a quantification of virtual water flows between nations in relation to crop trade In: Value of Water Research Report Series. 11. the Netherlands: UNESCO: IHE, Institute for water education, Delft.

27. Hoekstra AY and Mekonnen MM 2012. The water footprint of humanity. Proceedings of the National Academy of Sciences 109: 3232–3237, DOI: http://dx.doi.org/10.1073/pnas.1109936109

28. Hook JE, Hoogenboom G, Paz J, Mullen J, Bergstrom J and Risse M 2009. Agricultural Irrigation Water Demand, Georgia’s Major and Minor Crops, 2011 through 2050 In: GA, USA: The National Environmentally Sound Production Agriculture Laboratory (NESPAL), University of Georgia.

29. Ignatius AR and Jones JW 2014. Small Reservoir Distribution, Rate of Construction, and Uses in the Upper and Middle Chattahoochee Basins of the Georgia Piedmont, USA, 1950–2010. ISPRS International Journal of Geo-Information 3: 460–480.

30. ISO 2006a. Environmental management — Life cycle assessment — Principles and framework. ISO 14040 In: Switzerland: International Organization for Standardization.

31. ISO 2006b. Environmental management — Life cycle assessment — Requirements and guidelines. ISO 14044 In: Switzerland: International Organization for Standardization.

32. Jones MP and Hunt WF 2010. Performance of rainwater harvesting systems in the southeastern United States. Resources Conservation and Recycling 54: 623–629, DOI: http://dx.doi.org/10.1016/j.resconrec.2009.11.002

33. Kahinda JM, Taigbenu AE, Sejamoholo BBP, Lillie ESB and Boroto RJ 2009. A GIS-based decision support system for rainwater harvesting (RHADESS). Physics and Chemistry of the Earth 34: 767–775, DOI: http://dx.doi.org/10.1016/j.pce.2009.06.011

34. Liaw CH and Tsai YL 2004. Optimum storage volume of rooftop rain water harvesting systems for domestic use In: Wiley Online Library.

35. Lye DJ 2009. Rooftop runoff as a source of contamination: A review. Science of the total environment 407: 5429–5434, DOI: http://dx.doi.org/10.5194/hess-15-1577-2011

36. Mayer PW, DeOreo WB, Opitz EM, Kiefer JC, Davis WY, Dziegielewski B and Nelson JO 1999. Residential end uses of water In: AWWA Research Foundation and American Water Works Association Denver, CO.

37. Mekonnen MM and Hoekstra AY 2011. The green, blue and grey water footprint of crops and derived crop products. Hydrology and Earth System Sciences 15: 1577–1600, DOI: http://dx.doi.org/10.5194/hess-15-1577-2011

38. NCSL 2014. State Rainwater | Graywater Harvesting Laws and Legislation [Internet]. USA: National Conference of State Legislatures. Available from: http://www.ncsl.org/research/environment-and-natural-resources/rainwater-harvesting.aspx.

39. NCSL 2016. State Rainwater Harvesting Laws and Programs [Internet]. USA: National Conference of State Legislatures. Available from: http://www.ncsl.org/research/environment-and-natural-resources/rainwater-harvesting.aspx.

40. Ngigi SN, Savenije HHG, Rockstrom J and Gachene CK 2005. Hydro-economic evaluation of rainwater harvesting and management technologies: Farmers’ investment options and risks in semi-arid Laikipia district of Kenya. Physics and Chemistry of the Earth 30: 772–782, DOI: http://dx.doi.org/10.1016/j.pce.2005.08.020

41. NIST 2013a. Building for Environmental and Economic Sustainability (BEES) In: USA: The National Institute of Standards and Technology (NIST).

42. NIST 2013b. Energy Price Indices and Discount Factors for Life-Cycle Cost Analysis – 2013 Annual Supplement to NIST Handbook 135 and NBS Special Publication 709 In: National Institute of Standards and Technology, U.S. Department of Commerce.

43. NRC 2012. Water Reuse: Potential for Expanding the Nation’s Water Supply Through Reuse of Municipal Wastewater In: Washington, DC: National Research Council, The National Academies Press.

44. NREL 2013. U.S. Life Cycle Inventory Database, National Renewable Energy Laboratory (NREL) In: USA: National Renewable Energy Laboratory (NREL).

45. OpenLCA 2013. OpenLCA, Modular Open Source Software for Sustainability Assessment. GreenDelta.

46. Register F 1999. 10 CFR 436 – FEDERAL ENERGY MANAGEMENT AND PLANNING PROGRAMS In: House, TW ed.  USA: The White House.

47. Roebuck RM, Oltean-Dumbrava C and Tait S 2011. Whole life cost performance of domestic rainwater harvesting systems in the United Kingdom. Water and Environment Journal 25: 355–365, DOI: http://dx.doi.org/10.1111/j.1747-6593.2010.00230.x

48. Sedlak D 2014. Water 4.0: the past, present, and future of the world’s most vital resource In: Yale University Press.

49. Shuster WD, Lye D, La Cruz A, Rhea LK, O’Connell K and Kelty A 2013. Assessment of Residential Rain Barrel Water Quality and Use in Cincinnati, Ohio. JAWRA Journal of the American Water Resources Association 49: 753–765, DOI: http://dx.doi.org/10.1111/jawr.12036

50. Smith S, Renwick W, Bartley J and Buddemeier R 2002. Distribution and significance of small, artificial water bodies across the United States landscape. Science of the Total Environment 299: 21–36, DOI: http://dx.doi.org/10.1016/S0048-9697(02)00222-X

51. Steffen J, Jensen M, Pomeroy CA and Burian SJ 2013. Water Supply and Stormwater Management Benefits of Residential Rainwater Harvesting in US Cities. Journal of the American Water Resources Association 49: 810–824, DOI: http://dx.doi.org/10.1111/jawr.12038

52. Tam VWY, Tam L and Zeng SX 2010. Cost effectiveness and tradeoff on the use of rainwater tank: An empirical study in Australian residential decision-making. Resources, Conservation and Recycling 54: 178–186, DOI: http://dx.doi.org/10.1016/j.resconrec.2009.07.014

53. Thomas RB, Kirisits MJ, Lye DJ and Kinney KA 2014. Rainwater harvesting in the United States: a survey of common system practices. Journal of Cleaner Production 75: 166–173, DOI: http://dx.doi.org/10.1016/j.jclepro.2014.03.073

54. TWDB 2006. Texas Rainwater Harvesting Evaluation Committee, Rainwater Harvesting Potential and Guidelines for Texas, Report to the 80th Legislature In: TX, USA: Texas Water Development Board.

55. UN 2014. Rainwater Harvesting And Utilisation, Examples of Rainwater Harvesting and Utilisation Around the World [Internet] In: United Nations Environment Programme, Division of Technology, Industry and Economics. Available from: http://www.unep.or.jp/ietc/publications/urban/urbanenv-2/9.asp.

56. UNEP-SEI 2009. Rainwater Harvesting: A Lifeline for Human Well-being: a Report Prepared for UNEP In: Nairobi, Kenya: United Nations Environment Programme (UNEP)/Stockholm Environment Institute (SEI). No. 978-92-807-3019-7.

57. USEIA 2014. Electricity [Internet] In: USA: U.S. Energy Information Administration (EIA). Available from: http://www.eia.gov/electricity/data.cfm#sales.

58. USEPA 1990. Recommended Determination to Prohibit Construction of Two Forks Dam and Reservoir Pursuant to Section 404(c) of the Clean Water Act In: USA: U.S. Environmental Protection Agency.

59. USEPA 1996. Watershed Approach Framework [Internet].  Available from: U.S. Environmental Protection Agency, http://water.epa.gov/type/watersheds/framework.cfm.

60. USEPA 2008a. Application of Watershed Ecological Risk Assessment Methods to Watershed Management In: Washington, DC: U.S. Environmental Protection Agency.

61. USEPA 2008b. Managing Wet Weather with Green Infrastructure Action Strategy 2008 [Internet] In: USA: U.S. Environmental Protection Agency. Available from: U.S. Environmental Protection Agency, http://water.epa.gov/infrastructure/greeninfrastructure/upload/gi_action_strategy.pdf.

62. USEPA 2008c. Managing Wet Weather with Green Infrastructure Municipal Handbook Rainwater Harvesting Policies In: USA: U.S. Environmental Protection Agency.

63. USEPA 2012. Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts (TRACI) TRACI version 2.1 User’s Guide In: Cincinnati, OH, USA: U.S. Environmental Protection Agency. No. EPA/600/R-12/554.

64. USEPA 2013. Tool for the reduction and assessment of chemical and other environmental impacts (TRACI) [Internet] In: USA: U.S. Environmental Protection Agency. Available from: http://www.epa.gov/nrmrl/std/traci/traci.html.

65. USEPA 2014a. Global Greenhouse Gas Emissions Data [Internet] In: USA: U.S. Environmental Protection Agency. Available from: http://www.epa.gov/climatechange/ghgemissions/global.html.

66. USEPA 2014b. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990–2012 In: Washington, DC: U.S. Environmental Protection Agency.

67. USEPA 2015. Climate Impacts in the Southeast [Internet] In: USA: U.S. Environmental Protection Agency. Available from: http://www.epa.gov/climatechange/impacts-adaptation/southeast.html.

68. USEPA 2016. Pricing and Affordability of Water Services [Internet] In: USA: US Environmental Protection Agency. Available from: https://www.epa.gov/sustainable-water-infrastructure/pricing-and-affordability-water-services#pricing.

69. USGCRP 2009. Highlights of Global Climate Change Impacts in the United States [Internet] In: Washington, DC: U.S. Global Change Research Program. Available from: http://www.globalchange.gov.

70. USGCRP 2012. Regional Climate Impacts: Southeast, U.S. Global Change Research Program [Internet].  Available from: http://www.globalchange.gov/publications/reports/scientific-assessments/us-impacts/regional-climate-change-impacts/southeast.

71. van der Sterren M, Rahman A and Ryan G 2014. Modeling of a lot scale rainwater tank system in XP-SWMM: A case study in Western Sydney, Australia. Journal of environmental management 141: 177–189, DOI: http://dx.doi.org/10.1016/j.jenvman.2014.02.013

72. Van Meter KJ, Basu NB, Tate E and Wyckoff J 2014. Monsoon Harvests: The Living Legacies of Rainwater Harvesting Systems in South India. Environmental Science & Technology 48: 4217–4225, DOI: http://dx.doi.org/10.1021/es4040182

73. VDH 2011. Virginia Rainwater Harvesting & Use Guidelines In: USA: Virginia Department of Health.

74. Vellidis G 2015. Santosh, R and Ghimire, JMJ eds.  E-mail communication (Potential for rainwater harvesting in Southeast US).

75. Villarreal EL and Dixon A 2005. Analysis of a rainwater collection system for domestic water supply in Ringdansen, Norrköping, Sweden. Building and Environment 40: 1174–1184, DOI: http://dx.doi.org/10.1016/j.buildenv.2004.10.018

76. Walton B and Price of Water 2015. Up 6 Percent in 30 Major U.S. Cities; 41 Percent Rise Since 2010 [Internet] In: USA: Circle of Blue. Available from: http://www.circleofblue.org/2015/world/price-of-water-2015-up-6-percent-in-30-major-u-s-cities-41-percent-rise-since-2010/.

77. Ward S, Memon FA and Butler D 2012. Performance of a large building rainwater harvesting system. Water Research 46: 5127–5134, DOI: http://dx.doi.org/10.1016/j.watres.2012.06.043

78. WFN 2009. Water footprint manual Enschede In: The Netherlands: The Netherlands Water Footprint Network.

79. Wichelns D 2010. Agricultural Water Pricing: United States In: Indiana, USA: Hanover College.

80. Zhang XQ and Hu MC 2014. Effectiveness of Rainwater Harvesting in Runoff Volume Reduction in a Planned Industrial Park, China. Water Resources Management 28: 671–682, DOI: http://dx.doi.org/10.1007/s11269-013-0507-9