Accepted Manuscript

Analyzing virtual water pollution transfer embodied in economic activities based on

Gray Water Footprint: A case study

Hui Li, Zhifeng Yang, Gengyuan Liu, Marco Casazza, Xinan Yin

PII: S0959-6526(17)31098-3
DOI: 10.1016/j.jclepro.2017.05.155
Reference: JCLP 9695

To appear in: Journal of Cleaner Production

Received Date: 12 January 2017

Revised Date: 17 March 2017
Accepted Date: 25 May 2017

Please cite this article as: Li H, Yang Z, Liu G, Casazza M, Yin X, Analyzing virtual water pollution
transfer embodied in economic activities based on Gray Water Footprint: A case study, Journal of
Cleaner Production (2017), doi: 10.1016/j.jclepro.2017.05.155.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
1 Analyzing virtual water pollution transfer embodied in economic

2 activities based on Gray Water Footprint: A case study

,
3 Hui Li1, Zhifeng Yang1,2,*, Gengyuan Liu1,2 *, Marco Casazza3, Xinan Yin1
1
4 State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Beijing

5 Normal University, Beijing 100875, China

PT
2
6 Beijing Engineering Research Center for Watershed Environmental Restoration & Integrated Ecological

7 Regulation, Beijing 100875, China

RI
3
8 University of Naples ‘Parthenope’, Department of Science and Technology, Centro Direzionale, Isola C4, 80143,

9 Naples, Italy

SC
10

11 Abstract

U
12 Water pollution can be transferred either through natural water bodies or, indirectly, through
AN
13 economic activities. The former is a physical transfer, which has been extensively investigated and

14 just used for controlling pollution of transboundary river basin. The latter is a virtual transfer, as
M

15 shown by some studies, where it is demonstrated that the quantity of pollutants embodied in

16 economic activities is much larger and worthy attention. However, understanding the mechanisms
D

17 of virtual water pollution transfer through economic activities and relating them to domestic water
TE

18 resources remains a challenge. This study approaches this subject focusing on the use of the Gray

19 Water Footprint (GWF) to account for the virtual water pollution. Based on natural background
EP

20 concentration, GWF is defined as the essential quantity of fresh water in order to assimilate the

21 load of pollutants discharged into water. GWF reflects the impact of water pollution on the
C

22 available quantity of water resources. GWF assessment has been tested quantifying and tracking
AC

23 Beijing’s water pollutants emission flows embodied in economic activities. Based also on an

24 Economic Input-Output Life Cycle assessment, we found that: 1) the material exchange between

25 economic sectors cause the water pollution transferred virtually from down-stream industry to

26 up-stream industry. For example, food and tobacco sector transfers 23.28×107 m3 virtual water

27 pollution to the agriculture sector. Consequently, the sectors which receive intermediate products

*
Corresponding author:
E-mail address: zfyang@bnu.edu.cn (Z.F. YANG); liugengyuan@bnu.edu.cn (G.Y. LIU)
 ACCEPTED MANUSCRIPT
28 from others, should share responsibility for water pollution; 2) 39.19×107 m3 of virtual water

29 pollution, embodied in the exports of other regions, are induced by Beijing’s consumption pattern.

30 It is suggested that measures should be introduced to reduce water pollution by integrating

31 cross-regional production and consumption patterns.

32 Keywords: Virtual water pollution transfer; Input-Output Analysis; Life Cycle assessment; Gray

PT
33 Water Footprint

34 Nomenclatures

RI
COD Chemical oxygen demand. The amount of water-dissolved oxygen required to
oxidize all organic substances in water.

SC
DDWR Direct dilution water requirement. The volume of freshwater required for each sector
to dilute the waste water until it meets the standard concentration.

U
EGWF Export gray water footprint. The gray water footprint embodied in export products
and services.
AN
EIO-LCA Economic Input-Output Life Cycle assessment. A methodology to evaluate the
environmental impacts of a product or process over its entire life cycle based on
input-output analysis.
M

GR The gross runoff for both surface and ground water.

GWF Gray water footprint. The essential quantity of fresh water in order to assimilate the
D

load of pollutants discharged into water based on natural ground concentration.

GWFI The intensity of gray water footprint, it reflects the gray water footprint required
TE

when produce per final product for each sector

IGWF Import gray water footprint. The gray water footprint embodied in
EP

import products and services.

IOA Input-Output analysis. A quantitative economic analysis based on the
interdependencies between different sectors of a regional economy or different
C

regional economies
AC

NEGWF Net export gray water footprint, which is equals to the export gray water footprint
(EGWF) minus import gray water footprint (IGWF)
NH3-N Ammonia nitrogen
TGWF The total GWF of a region, which includes both 17 economic sectors and
households.
TN Total nitrogen.
TP Total phosphorus.
WPL Water pollution level. An indicator to measure the degree of pollution of a region,
which is obtained as the total gray water footprint divided by the gross runoff (GR).
 ACCEPTED MANUSCRIPT
35

36 1. Introduction

37 Water pollution transfer indicates the phenomenon referred to the transfer of environmental

38 damage, caused by water pollution, from a given nation, region or place (e.g.: a company) to

39 another, consequently shifting the adverse effects to another water body (Lucas et al., 1992; Mani

PT
40 et al., 1998). Pollutants are transferred either directly, through water bodies, or indirectly, through

41 a given production chain (Shi et al., 2016b). Previous studies focused on the former transfer (e.g.:

RI
42 Funfgelt and Schulze, 2016). In this case, based on different water quality models, they either

43 predicted and interpreted water pollution or guided the implementation of emission control

SC
44 measures (Zhao et al., 2012; Shi et al., 2016a). Nonetheless, water pollution transfer not only

45 occurs physically in transboundary river basins, but also virtually, embodied in traded goods

46
U
(Ghertner and Fripp, 2007; Burke et al., 2009; Duan et al., 2011; Duan et al., 2013; He et al., 2015;
AN
47 Kinna, 2016). For example, if a given commodity is traded between two regions, its associated

48 water pollution is ‘virtually’ transported from the importer to the exporter (Liu et al., 2007;
M

49 Chapagain and Hoekstra, 2008; Hoekstra and Chapagain, 2008; Liu and Savenije, 2008).

50 Monitoring the physical transport is not enough to understand the drivers of water pollution,
D

51 giving that the water pollutants emissions impact along the whole production chain (Flach et al.,
TE

52 2016; Sun et al., 2016b). In fact, due to the rapid economic development, the amount of pollutants

53 embodied in traded products are growing fast (Hertwich and Peters, 2009; Wan et al., 2016). This
EP

54 case is referred as virtual water pollution transfer, which should receive more attention both by

55 researchers and industry as well as sustainability practitioners.

C

56 Water pollution transfer assessment makes visible how water pollution is “reallocated” along
AC

57 production chains. Due to the low degree of connectivity of water systems, few studies have been

58 carried out on this subject. Considering a similar case, virtual air pollution transfer has been

59 analyzed on multiple scales regarding climate change (Shui and Harriss, 2006; Hertwich and

60 Peters, 2009; Davis and Caldeira, 2010). In this case, large scale trade is of crucial importance to

61 influence the global air quality, through integrating the input–output model with the emission

62 inventory and air quality model (Wyckoff and Roop, 1994; Su and Ang, 2014; Islam et al., 2016;

63 Rafiq et al., 2016). Furthermore, about 20-25% CO2 emissions are from the production of
 ACCEPTED MANUSCRIPT
64 interregional traded products (Peters and Hertwich, 2008; Davis and Caldeira, 2010). These

65 emissions are growing, on average, by 3.4% per year, considering the timespan from year 1990 to

66 2008 (Peters et al., 2011).

67 The pollution haven hypothesis1, namely, that polluting industries will relocate to areas of the

68 world where environmental regulation and standards are being dumped, demonstrates that

PT
69 developed regions are shifting the environment damage caused by CO2 and non-CO2 greenhouse

70 gas to the less developed regions, via importing emission-intensive goods (Kohler, 2013; Su and

RI
71 Ang, 2014; Wu et al., 2014). At the sectoral level, life cycle assessment has been developed to

72 investigate the total amount of CO2 emissions (Fu et al., 2014; O'Born et al., 2016). Both direct

SC
73 and indirect emissions generated by producing raw materials should be considered (Jiang et al.,

74 2013; Zhang et al., 2014). For instance, the suppliers of steel, glass, rubber, and plastics are

U
75 seriously important to the car manufacturers, as they are all heavily affected by the requirements
AN
76 and regulations of emission (Lee, 2011). In such a case, to achieve a sound production chain

77 management of carbon risk mitigation, it is essential to concentrate on the management of the

M

78 downstream results of the utilization of its products and the upstream inputs suppliers (Lenzen et

79 al., 2007; Chang et al., 2010). Some studies focused on the virtual transfer of water pollutants
D

80 related to economic activities. MacDonald et al. (2012) researched the phosphorus emissions
TE

81 caused by international trade. The results demonstrated that one quarter of American phosphorus

82 emissions are linked to export production. Zhi et al. (2015) calculated the Chemical Oxygen
EP

83 Demand (COD) transfer among economic sectors and point out accounting for sectoral wastewater

84 discharge should consider both direct and indirect sources induced by other sectors. Lassaletta et
C

85 al. (2014) observed that during the 1961 2009, the amount of nitrogen (N) traded between
AC

86 countries has increased from 3 to 24 TgN 1×1012 g/year and now concerns one-third of the total

87 N in world crop production. These studies showed that the quantities of pollutants, embodied in

88 economic activities, are large and worthy of attention. However, the quantification of the

89 pollutants amount is not directly related to domestic water resources.

90 The Gray Water Footprint (GWF) represents an integrated monitoring metrics for indicating

1
The pollution haven hypothesis posits that, when large industrialized nations seek to set up factories or offices
abroad, they will often look for the cheapest option in terms of resources and labor that offers the land and material
access they require.
 ACCEPTED MANUSCRIPT
91 virtual water pollution. GWF is a part of the total water footprint, which is consisted of colors in

92 green, blue and grey (Hoekstra et al., 2011). The green water footprint is the volume of rainwater

93 consumed during the productive process (Hoekstra and Chapagain, 2007a; Zhuo et al., 2016). The

94 blue water footprint refers to the volume of surface and groundwater consumed as a result of the

95 production of a good or service (Ayala et al., 2016). Based on the natural background

PT
96 concentration, GWF is defined as the essential quantity of fresh water in order to assimilate the

97 pollutants discharged into water. GWF was used first as a water research management indicator,

RI
98 which reflects the impact of water pollution on the available quantity of water resources (Cazcarro

99 et al., 2016; Wu et al., 2016). The advantages of GWF applied to virtual water pollution analysis

SC
100 are that it converts all embodied pollutants into a homogeneous unit: freshwater consumption.

101 Furthermore, it reflects the economic activities impacts on water quality, by quantifying the fresh

U
102 water required to assimilate the emissions (Liu et al., 2012; Wu et al., 2016).
AN
103 Previous GWF studies drew a picture of international virtual water pollution transfer and

104 analyzed the induced water scarcity faced by exporters on the global scale (Mekonnen et al., 2016;
M

105 Sun et al., 2016a; Wan et al., 2016). Conversely, at local scale, how the industrial structure and the

106 consumption patterns affect the GWF transfer is unknown.

D

107 In the interest of developing strategies to alleviate domestic water scarcity, it is essential to
TE

108 investigate the motivations of GWF transfer. TheGWF transfer within a region is conducted by

109 introducing the Economic Input-Output Life Cycle assessment (EIO-LCA) in this study. This
EP

110 method facilitates the study as it avoids the possible problems to approximate such as

111 overestimation or double counting (Leontief, 1941). In addition, it performs a life cycle
C

112 assessment, evaluating the environmental impacts of a product or process over its entire life cycle
AC

113 (Hawkins et al., 2007). Even though the method has already been used in the departments of

114 energy consumption and carbon emissions (Ji and Chen, 2016; Junnila, 2008), it has uncommonly

115 been introduced in the study of GWF. Traditional Input-Output analysis (IOA) only calculates the

116 GWF of each sector, since it is difficult to further analysis the source of sectoral GWF.

117 The description of the GWF method is followed by a case study of Beijing. The choice of

118 Beijing is due to the fact that this area faces severe fresh water shortages, further enhanced by

119 water pollution. Plans relative to physical water transfer, such as the South–North Water Transfer
 ACCEPTED MANUSCRIPT
120 Project, could partially alleviate the water quantity-induced scarcity (Zhao et al., 2015).

121 Nonetheless, this work could contribute to the discussion about how to cope also with

122 water-quality induced scarcity, which will be analyzed through a GWF model.

123

124 2. Methodology

PT
125 2.1 The GWF framework

126 In this paper it is proposed a definition of virtual water pollution as the hypothetical volume

RI
127 of freshwater needed to assimilate the pollutants generated by the production of commodities.

128 Virtual water pollution transfer is enabled through two pathways along a whole production chain:

SC
129 1-) The exchange of materials between economic sectors. For instance, if the intermediate

130 products produced in sector B are used by sector A in the process of manufacturing, the pollutants

131
U
discharged by sector B in the process of production of these products should be shown in sector
AN
132 A’s indirect pollutant emissions (Zhi et al., 2015). In other words, pollutants are transferred

133 “virtually” from downstream industry (sector A) to upstream industry (sector B) (Hoekstra and
M

134 Chapagain, 2007b; Zhao et al., 2009; Zhao et al., 2010). Assessment of the sectoral emissions

135 should account for both direct and indirect emissions (Aviso et al., 2011).
D

136 2-) The interregional trade. Emissions could be embodied in products and transferred through
TE

137 trade. Thus, it is necessary to account for the emissions produced along the supply chain of the

138 traded products and to relate them to final consumption (Wiedmann, 2009).
EP

139 Then, it is incorporated virtual water pollution into GWF accounting. Thus, the framework

140 for the virtual water pollution transfer assessment, as shown in Figure 1, consists of two main
C

141 stages: Stage A is represented by the dotted box, that focus on the flow analysis of sectoral GWF
AC

142 within a region; Stage B is represented by the continuous lined box, that accounts for the GWF

143 transfer embodied in interregional trade.

144 Fig. 1. The framework for the GWF assessment based on EIO-LCA method

145 2.1.1The sectoral GWF flow analysis

146 In order to research the virtual water pollution transfer among economic sectors, the sectoral

147 GWF flow is analyzed, based on EIO-LCA method (Ji et al., 2011). It is assumed that there are
 ACCEPTED MANUSCRIPT
148 three sectors (A, B and C), which have material exchange in region D (Figure 1). The Direct

149 Dilution Water Requirement (DDWR) of sector A represents the volume of freshwater required by

150 sector A to dilute the waste water until it reaches the standard concentration. The DDWR of sector

151 A is redistributed along the production chain. This step can be regarded as the GWF embodied in

152 products. In particular, sector A provides a part of products as other sectors’ intermediate products,

PT
153 while using intermediate products from sector B and C. Thus, the GWF of sector A contains three

154 parts: direct GWF embodied in the sector’s A final products; indirect GWF embodied in sector’s

RI
155 B intermediate products; indirect GWF embodied in sector’s C intermediate products. Traditional

156 IOA method only calculates the GWF of each sector, since it is difficult to further analyze how the

SC
157 DDWR redistributed among economic sectors. In order to trace the source of sectoral GWF, an

158 EIO-LCA model has been improved and, then, applied for the flow analysis of the sectoral GWF.

U
159 In particular, following (Nansai et al., 2012):
AN
160
=  ×  ×  = 
 × 
 × λ × diag (1)

161  = 
 = [ − λ × ] (2)
M


162 λ =  ! = "
⁄#"
+ %&
' (3)
  
D

163 
= ))*+ ⁄, (4)

--. = ∑1
2 0

TE

164 (5)

165
.3 = ∑1 2 0

 (6)
EP

166 where G (m3) is the gray water matrix (n×n), while n is the amount of economic sectors covered in

167 this study. R is a diagonal matrix (n×n), which refers to the coefficients of required diluted water
C

168 considered in this analysis. Each element of R, indicated as rij, represents the quantity of
AC

169 freshwater (expressed in m3) used to assimilate the wastewater discharged directly by sector i,

170 until it reaches to a standard concentration, when sector i produces one unit of output (expressed

171 in monetary units of Chinese currency, RMB). As a Leontief inverse matrix (n×n), B eliminates

172 the influence of import products which don’t cause the pollution in researched region. Each

173 element of B, namely bij, shows the increase in output from each domestic sector caused by a unit

174 increase in final demand. λ is a diagonal matrix (n×n), that represents the proportion of local

175 production, xj is the output of sector j and IMj is the input of sector j imported from other regions.
 ACCEPTED MANUSCRIPT
176 A is the amount of input from sector i needed to increase one monetary unit output in sector j. M

177 is a diagonal matrix (n×n), that represents the final demand. DDWRi (expressed in m3) is the direct

178 dilution water requirement of sector i. This is the volume of fresh water consumed to assimilate

179 the wastewater discharged by sector i into a water body. Finally, ti is the gross output (in RMB) of

180 sector i.

PT
181 Each component of matrix G is indicated as gij. Adding gij in crosswise direction, the column

182 vector DDWR (n×1) is obtained. This accounts for the direct discharge dilution requirement for

RI
183 each sector. Instead, adding gij in vertical direction, the row vector GWF (1×n) is generated,

184 representing the GWF for each sector.

SC
185 The advantage of applying EIO-LCA method is that the result is a matrix, which can

186 discriminate direct GWF and indirect GWF. The diagonal elements of matrix G are the direct

U
187 GWF embodied in its final products. The other elements are the indirect GWF produced by the
AN
188 use of intermediate products derived from other sectors.

189 The option of including wastewater discharged by households was also considered.
M

190 Households in this study are differ from the resident consumption in Input-Output table. The GWF

191 of resident consumption refers to the amount of GWF required to assimilate the production
D

192 sewage induced by providing products for local residents, which is a part of sectoral GWF.
TE

193 Nonetheless, the GWF of households here refers to the GWF required to assimilate the living

194 sewage from life. To better reflect the impact of human activities on the local water resources,
EP

195 both production sewage and living sewage should be considered. Therefore, the GWF of

196 households, which equals its DDWRi, was calculated separately.

C

197 In order to account for the water dilution requirement of each sector and households, DDWRi ,
AC

198 the model of water quality was introduced. The model hypothesizes that the conditions for dilution

199 are to some extent as same as that for the surface or ground water on the basis of a mass-balance

200 method (Fischer, 1967; Guan and Hubacek, 2008). Once integrating the waste water discharged in

201 the economic sector into the natural water entity, the model helps to account for the pollutant

202 concentration. Even though this model only concentrates on the water quality, it facilitates the

203 estimation of the dispersal of water pollution. The model, in its concise version, has the following

204 form:
 ACCEPTED MANUSCRIPT
56 76 89 5: 7: 57;
205 ))*+4 = 7
8< 7;

206 ))*+ = =>"))*+4  8

207 DDWRm represents the requirement of dilution water for a given pollutant, m. The initial

208 runoff rate and the waste water discharged rate are expressed respectively by q0 and qp (expressed

209 in units of [m3 s-1]). Where, q (expressed in [m3 s-1]) indicates the runoff rate once the mixing

PT
210 process is completed. The pollutant concentration in the water body and the discharged pollutant

211 concentration to the water body are represented respectively by c0 and cp (g m-3). The overall

RI
212 purification rate of pollutants can be represented by k1, and k2 represents this rate before the

SC
213 pollutants discharged into the waterbody. In order to eliminate water pollution, the value of cs

214 should be below the requirement of Chinese water quality standard. Cs in this article was defined

U
215 as the maximum pollutants concentration under Grade V, which indicates that water cannot be

216 cyclically used but is to some extent harmless (National Environmental Protection Agency, 2002).
AN
217 Wastewater discharges carry more than one matter, the DDWR value is determined by the key

218 pollutant. Hence, DDWRi, which refers to the direct water dilution requirement of sector i, is equal
M

219 to the water dilution requirement caused by critical pollutants.

D

220 In this research, due to data availability regarding the concentration of pollutants in

221 wastewater, four indicators were selected for the assessment: chemical oxygen demand (COD);
TE

222 ammonia nitrogen (NH3-N); total phosphorus (TP); total nitrogen (TN). The reason of such choice

223 is due to their relation with the GWF embodied in products.

EP

224

2.1.2 The GWF embodied in interregional trade

C

225

In order to analyze the virtual pollution transfer embodied in interregional trade, IOA method
AC

226

227 was introduced to specify the GWF embodied in import and export commodities. IOA is a

228 well-established top-down method, which reflects the interdependencies between different

229 economic sectors within a region or different regional economies. This method has been widely

230 used in the aspect of water footprint accounting by attributing resource use or pollution to final

231 products in a consistent framework (Chen et al., 2012; Ha et al., 2010). The equations can be

232 expressed as:

 ACCEPTED MANUSCRIPT
233 ?@
.3 = @
.3 − 
.3 (9)

234 @
.3 =  ×  × ABC (10)

235 
.3 =  ×  × ADA (11)

236 where NEGWF is a (n×1) dimensional vector, which represents the net export GWF, EGWF

237 represents GWF export and IGWF represents the GWF import. Dimensional vector (n×1)

PT
238 associated with the import is represented by mim, while mex is a (n × 1) dimensional vector referred

239 to the export.

RI
240 The equations 9-11 allows to have the following observations. If a region’s NEGWF is more

241 than zero, it means that this region bears a disproportionate share of the costs for both pollution

SC
242 and its mitigation. On the other hand, if a region’s NEGWF is smaller than zero, it implies that

243 the given domestic consumption pattern shifts the water pollution to other regions via importing

U
244 goods. It is also important to notice that the value of R imported by a region is not the real direct
AN
245 dilution water requirement by the production site. It is, instead, the water volume that the region

246 would need to produce the same product by itself. Thus, considering such a consumption, the
M

247 value of R (direct dilution water requirement coefficient) for imported products in the above

248 formulation is regraded to be the same as the domestic R.

D

249 Households here are not economic sector, which neither export nor import products. Their
TE

250 IGWF and EGWF were set to 0 and did not need to be taken into account.

251
EP

252 2.2 Indicators

253 To further analyze the GWF of each sector and reflect the impacts of GWF on water resource
C

254 availability, some more indicators are introduced.

AC

255

256 2.2.1 The WPL

257 The water pollution level (WPL) measures the degree of pollution of a region. This is

258 obtained as the overall GWF value divided by the regional gross runoff (Zeng et al., 2013):

259 *EF = GH*I⁄H+ (12)

260 TGWF (m-3) is the total GWF of a region, which includes both 17 economic sectors and
 ACCEPTED MANUSCRIPT
261 households. GR (m-3) refers to the gross runoff for both surface and ground water. If WPL<1,

262 TGWF is smaller than the GR. Thus, there will be still enough water to dilute the wastewater to an

263 acceptable concentration under the water quality standards. In contrast, when the WPL>1, the

264 assimilation capacity of GR is not enough to take up the discharge. The final result, in the latter

265 case, will be a degradation of water quality (Mekonnen and Hoekstra, 2015).

PT
266

267 2.2.2 The intensity of GWF

RI
268 Since the final demand of each sector is quite different, the use of absolute values for GWF

269 does not make much sense. Instead, it is better to compare the relative volume of final products

SC
270 GWF per unit of currency among sectors. For such a purpose, the concept of GWF Intensity

271 (GWFI) is introduced. In analogy with the water footprint intensity (i.e.: water footprint/GDP), the

U
272 sectorial GWFI for the sector i can be written as:
AN
273 H*I% = H*I ⁄= (13)

274 where GWFIi is the GWFI for the sector i, GWFi is the GWF in sector i, and mi is the final
M

275 demand in sector i. High values of GWFIi indicate the presence of intense emissions within sector

276 i.
D

277
TE

278 2.3 Data inputs

279 The EIO-LCA referenced data (including input-output matrix, consumption, export et al for
EP

280 Beijing) are achieved from the regional input-output table (Beijing Bureau of Statistics, 2007).

281 Beijing's 17 main economic sectors are defined by the National Bureau of Statistics (2007). Data
C

282 on runoff rates, wastewater discharges, and pollutants levels of most sectors are obtained from:
AC

283 Beijing Water Resources Bulletin of 2007 (Ministry of Water Resources, 2007); Haihe River

284 Yearbook 2007 (HRWCC, 2008); China Statistical Yearbook 2007 (National Bureau of Statistics,

285 2008). Additional pollutants levels for several industrial and services sectors come from previous

286 studies (Xie, 1996; Han et al., 1998; Dong, 2000; Weng, 2002; Li, 2003; Guan and Hubacek,

287 2008). For the details, see the Table S-1, Table S-2 in the supplementary materials.

288

289 3. Results
 ACCEPTED MANUSCRIPT
290 3.1 The overall value of GWF and DDWR and related indicators in Beijing

291 GWF values, wastewater discharge quantity and DDWR for 17 economic sectors and

292 households of Beijing in 2007 are presented in Table 1. The TGWF of Beijing in 2007 is

293 305.32×107 m3, which is equal to the total DDWR. The calculation results of TGWF in Beijing

294 show that Beijing required the total amount of 305.32×107 m3 fresh water to absorb, dilute and

PT
295 process the wastewater to water quality standard V grade, which was discharged by domestic

296 production and households. However, the GR in 2007 was only about 238.1×107 m3. That means

RI
297 the WPL of Beijing in 2007 is 1.28, greater than 1. The result of WPL implies that the demand for

298 clean water to dilute might exceed the locally available resource. It is worth mentioning that,

SC
299 requiring dilution to the background concentration would demand many more water-related

300 dilution services, but such a higher demand is not taken into account in this study. However, even

301
U
if the environmental background value is not considered, dilution might still be insufficient due to
AN
302 the obstacles to the water cycle in Beijing placed by the existing small reservoirs and public water

303 facilities. This result is to some extent similar with the one about Beijing’s water quality in the
M

304 year 2007. In fact, the water pollution was serious as 40% of the overall flow with water quality

305 lower than Grade V (Ministry of Water Resources, 2007).

D

306 The indicator of WPL has been used to express the effect of TGWF on the local water quality
TE

307 in some researches (Zeng et al., 2013; Mekonnen and Hoekstra, 2015). Note that the WPL is

308 inversely proportional to the GR, which can lead high value in water- stressed areas. With the GR
EP

309 of Beijing consistently to decrease in recent years (Beijing Water Resources Bulletin, 2015) , it is

310 more crucial to analyze the source of TGWF, further propose measures to control local water
C

311 pollution.
AC

312 The critical factor for agriculture is TN, while COD is more relevant for the other sectors.

313 The top three sectors with largest DDWRs are agriculture (98.80×107 m3), real estate and renting

314 (31.06×107 m3) and other services (29.92×107 m3), which account for 33.8%, 10.63%, 10.34% of

315 total, respectively. The rij for agriculture as primary industry is the highest among the considered

316 17 economic sectors (361.74×10-3 m3RMB-1). This means that, for the agriculture sector,

317 361.74×10-3 m3 of freshwater are required to assimilate the wastewater directly discharged for
 ACCEPTED MANUSCRIPT
318 producing one unit of output (in RMB). Water consumption for agriculture, causes a large quantity

319 of wastewater discharge (25.92 ×107 m3), which leads to the highest DDWR. The rij for real estate

320 and renting and for other services are, respectively, 12.55×10-4 m3RMB-1 and 5.98×10-4 m3RMB-1.

321 These values are much lower than the one for agriculture. Nonetheless, the wastewater discharge

322 of these sectors is increasing along with the production scale of Beijing’s tertiary industry growth.

PT
323 In other words, the gross outputs of real estate and renting (14.94×1010 RMB) and other services

324 (42.69×1010 RMB) are contributing to a larger final DDWR value. For the sector of wood, paper

RI
325 and others, although the rij is the second largest, induced by high COD concentrations, the DDWR

326 accounts only for 5.10% of the total, due to the their relatively low production scale (2.72×1010

SC
327 RMB).

328 The top three sectors with largest GWFs are: other services (57.85×107 m3); agriculture

329
U
(42.19×107 m3); food and tobacco (35.45 ×107 m3). They respectively account for 19.17%, 14.44%,
AN
330 and 12.13% of the total.

331 Due to the big gap among each sectoral final products, the absolute value of GWF doesn’t
M

332 reflect each specific emission intensity. Instead, the relative volume of GWF can be expressed by
D

333 GWFI, as shown in Table 1. The top three sectors with the largest GWFIs are: agriculture (214.40

334 ×10-4 m3 RMB-1); food and tobacco (47.25×

×10-4 m3 RMB-1); wood, paper and other (30.90×
×10-4
TE

335 m3 RMB-1). The GWFI reflects the GWF intensity of these three sectors. For example, considering

×10-4 m3 of freshwater are required for producing

EP

336 food and tobacco, its GWFI indicates that 47.25×

337 one unit (RMB) of final product, in order to assimilate the discharged wastewater.
C

338 Table 1. Results for 17 economic sectors and households in Beijing during year 2007.
AC

339 3.2 The sectoral GWF flow analysis

340 GWF and DDWR are compared for 17 economic sectors. Some sectoral differences,

341 evidenced by comparing the data of each specific GWF and DDWR, are due to material exchange

342 among economic sectors. GWF accounting should consider both Direct (embodied in final

343 products) and Indirect (embodied in intermediate products) GWF provided by other sectors. Thus,

344 in order to trace the source of each sectoral GWF, the flow analysis among 17 economic sectors

345 are defined and represented in Fig.2 ( the inputs data are shown in Table S-3 ). In this graphical
 ACCEPTED MANUSCRIPT
346 representation, it is possible to discern the amount of the flow between sectors according to their

347 width. The directions of flows indicate the contribution for each sector of the GWF transfer

348 network. The data show that the highest total amount of transferred flow is associated with

349 agriculture. In particular, the directions of the flows indicate that food and tobacco and other

350 services are more dependent on agriculture.

Fig 2. Flow analysis among 17 economic sectors (unit: 107 m3)

PT
351

352 The DDWR of agriculture is 98.80×107 m3. This includes 42.41% (41.90×107 m3) of the total

RI
353 caused by final products. 23.56% (23.28×107 m3) are due to intermediate products provided for

354 food and tobacco sector as raw material. 11.94% (11.80×107 m3) are caused by intermediate

SC
355 products provided for other services as raw material and 22.09% (21.82×107 m3) induced by the

356 other sectors. In Beijing municipality, agriculture provides a large part of GWF embodied in

357
U
intermediate products for other sectors, which is much more than it is received. In other words,
AN
358 virtual water pollution is transferred from the other sectors to agriculture.

359 Meanwhile, for food and tobacco, direct GWF related to its final products accounts for
M

360 29.43% (10.43×107 m3) of the sectoral GWF, while receiving 65.68% (23.28×107 m3) of indirect

361 GWF from the agriculture sector. For other services, direct GWF for final products accounts for
D

362 45.08% (57.84×107 m3) of the total. Instead, indirect GWF embodied in raw material provided by
TE

363 agriculture accounts for 20.39% (11.97×107 m3). Figure 2 shows that the virtual water pollution

364 flow is opposite direction to the GWF flow. That means the sector providing the indirect GWF is
EP

365 real pollution risk-taker.

366 These GWF transfers also exist among other sectors, but the quantities embodied are much
C

367 less with respect to above three sectors. For wood, paper and other, the DDWR is as much as
AC

368 14.89 ×107 m3, while its GWF decreases to 3.25 ×107 m3. The DDWR related to its final products

369 only accounts for 21.96% ( 3.27×107 m3). 30.89% (4.60×107 m3) of the total are due to provide

370 intermediate products for the other services as raw material. 10.95 % (1.63×107 m3) are induced

371 by intermediate products provided for construction and 7.99% (1.19×107 m3) are caused by

372 financial. In other words, other services, construction and financial, these down-stream sectors

373 receive 4.60×107 m3, 1.63×107 m3, 1.19×107 m3 indirect GWF from the wood, paper and other
 ACCEPTED MANUSCRIPT
374 sector, respectively. While these three sectors just provides 0.94×107 m3, 0.78×107 m3, 0.12×107

375 m3 indirect GWF for wood, paper and other, respectively. Wood, paper and other provides indirect

376 GWF more than it received, which cause virtual pollution is transferred from tertiary industry to

377 itself.

378 For one sector, a part of DDWR can be regarded as direct GWF embodied in its final

PT
379 products, another part of DDWR would redistribute among other sectors as indirect GWF

380 embodied in intermediate products. Even in some cases, the part of DDWR used for other sectors

RI
381 is much larger than for its own.

382

SC
383 3.3 Analyzing GWF transfer among economic regions

384 Beijing as China's political and cultural center, with advanced tertiary industry and other

385

U
commodities, heavily rely on imports. GWF transfer embodied in trade are presented in Figure 3.
AN
386 The total EGWF is 132.62 ×107 m³, while the total IGWF is 171.81 ×107 m³. Then, it was obtained

387 the total NEGWF is -39.19×107 m3. This means that Beijing imported products and services to
M

388 satisfy local consumption. Furthermore, in order to achieve self-sufficiency, 39.19×107 m3 of

389 additional fresh water would be required to assimilate the pollution.

D

390 Fig 3. Comparison of EGWF and IGWF of 17 economic sectors

TE

391 The IGWFs in 12 of 17 sectors exceed the EGWF, with the exception of the tertiary industry

392 sector (which includes transportation, postal services, information transmission, computer services
EP

393 and software, wholesale, retail trade, accommodation, and catering, real estate and renting,

394 financial sector and other services). The sector with the lowest NEGWF is agriculture (-53.60
C

395 ×107 m3), in which the value of IGWF is about 9 times higher than EGWF. It is followed by the
AC

396 sector of wood, paper and other (-8.78 ×107 m3), and food and tobacco (-6.62 ×107 m3). The values

397 of rij (the direct dilution water requirement of sector i per unit of output) of the three sectors, as

398 Table 1 shown, are 361.74 ×10-4 m3 RMB-1, 54.80×10-4 m3 RMB-1, 23.90×10-4 m3 RMB-1,

399 respectively, which are ranked in the top three. While the top three sectors with the largest

400 NEGWF are sector of transportation, postal services, information transmission, computer services

401 and software (15.32 ×107 m3), other services (15.24 ×107 m3), and financial (13.13 ×107 m3), in
 ACCEPTED MANUSCRIPT
402 which the values of rij are quite low (7.48×10-4 m3 RMB-1, 5.98×10-4 m3 RMB-1, 8.24×10-4 m3

403 RMB-1, respectively). The results imply that Beijing imported large amounts of high

404 pollution-added industrial products, while exported low pollution-added services. This trade

405 pattern because of the economic structure of Beijing, where high-tech services companies, such as

406 information transmission, soft-ware and financial are encouraged, while high-pollution enterprises

PT
407 as mentioned above were limited.

408

RI
409 4 Discussion

410 TGWF in Beijing is 305.32×107 m³. It was also calculated the blue water footprint based on

SC
411 EIO-LCA method, GWF is even a little bigger than the blue water footprint: 302.47×107 m³ (the

412 details are presented in the Table S-4). The result of domestic blue water footprint is consistent

U
413 with Zhao et al. (2016). Most previous studies focus on the blue and green water required. But,
AN
414 these studies do not account for the gray water required to assimilate pollutants during

415 manufacturing or consumption process. The results of this research emphasize that the GWF in the
M

416 total water footprint is of crucial importance. Moreover, the TGWF in this paper refers to the fresh

417 water required to assimilate the pollutants to water quality standard V grade, which is the lowest
D

418 quality level in the Chinese national water quality regulations. If the wastewater after dilution
TE

419 should achieve higher levels, the amounts of TGWFs would be much more (Table 2).

420 Table 2. Comparison of the TGWFs under different water quality levels
EP

421 The results of GWF matrix show the virtual water pollution transfer among economic sectors

422 via material exchange. The sectors which receive intermediate products from others means that the
C

423 virtual water pollution transferred from them to up-stream sectors. These down-stream sectors
AC

424 should share responsibility for both direct and indirect wastewater discharge. Traditional

425 management of water pollution mostly focus on the control of direct wastewater discharged by

426 industries and improving the wastewater treatment technology. Nonetheless, it is often to not

427 quantify the transfer in the indirect impacts brought about by the virtual water pollution transfer.

428 This quantification could help to explain questions concerning the domestic economy and the

429 management of water pollution. This can be reflected in the discussion about how the adjustment

430 of economic pattern influences the location of water pollution, and also the explanation about the
 ACCEPTED MANUSCRIPT
431 key sectors which lead to either direct or indirect wastewater discharge through the supply chain.

432 For example, the sector of other services receives a large quantity of indirect GWF from

433 agriculture, thus, saving raw material from agriculture and promoting the raw material efficiency

434 are indirect effective measures to alleviate local water pollution. One policy of the economic

435 pattern adjustments in Beijing is transforming the agricultural industry and the second industry

PT
436 into the service based industry. The “Outline of the Thirteenth Five-Year Plan of Beijing” (2016–

437 2020) concentrates on the growth of the service sector (Beijing Municipal Government, 2016).

RI
438 Nonetheless, one should admit that if this policy fails to realize that the service sector will

439 promote the agricultural outputs through the input of intermediate products indirectly from other

SC
440 pollution intensive industries (Suh, 2006), the economic pattern adjustment might be increasing

441 indirect wastewater discharge.

U
442 To replace the local products through imported trade can be viewed as a potential method to
AN
443 reduce TGWF. Although Beijing alleviates domestic water pollution through interregional virtual

444 water pollution transfer, this might have an adverse effect on the total water pollutants emission.
M

445 In order to enhance water quality, the Chinese government also introduced strict laws and

446 regulations to reduce wastewater discharge, both in the mega-city like Beijing and Shanghai, and
D

447 in some developed parts. However, if the policy only concentrates on the transformation of such
TE

448 industries out of these parts, but no changes in consumption structure occur, this might lead to a

449 growth in the overall water pollutants emission. In a given supply chain, the water pollutants
EP

450 emission through along geographically extended occur in regions where there is low efficiency

451 production and in regions with less regulation. From this point of view, this reallocation of water
C

452 pollutants emissions may have potentially important affects on the quality of regional water.
AC

453 This study offers a basic insight into the virtual water pollution transfer based on GWF model.

454 Several limitations still exist. Initially, the DDWR accounting model presented here could be

455 improved, more water quality parameters could be extended, besides NH3-N, TP, TN, COD.

456 Furthermore, the DDWRi (each sector’s DDWR), determined by a critical pollutant, does not

457 consider the combined effects of multiple pollutants. Such a limitation could be overcame through

458 the use of a comprehensive evaluation model. Other studies are necessary to know the synergy and

459 inhibition effects of the interactions of pollutants. Finally, this study researches the drives of
 ACCEPTED MANUSCRIPT
460 virtual water pollution transfer qualitatively, but the drivers and scenario analysis should be

461 deepened, allowing to obtain more detailed quantifications.

462

463 5 Conclusion

464 In this study, it was incorporated GWF to indicate virtual water pollution and to show the

PT
465 way in which the on-site water pollution is being reallocated by the domestic economic structure,

466 based on establishing GWF accounting models. The case study of Beijing in 2007 is analyzed and

RI
467 the conclusions are as follows:

468 (1) The sectoral GWF flow analysis reveals the quantities of virtual water pollution transfer

SC
469 among sectors, which links the increased water pollution to the final products with sectoral details.

470 The direction of inter-sector virtual water pollution transfer is opposite to the indirect GWF

U
471 flow.Besides reducing the direct wastewater discharge, promoting the raw material efficiency of
AN
472 down-stream sectors is an indirect measure to alleviate local water pollution.

473 (2) Accounting the IGWF and EGWF demonstrate the virtual water pollution transfer
M

474 between trade regions. Interregional virtual water transfer and the traded goods have adverse

475 directions. NEGWF is -39.19×107 m3 , which means Beijing transfer such amount of virtual
D

476 water pollution to exporters. Beijing as a developed city externalized its water pollution, which
TE

477 shows that it is necessary to quantify the pollutants embodied in traded products and to relate them

478 to final consumers.

EP

479 Analyzing virtual water pollution transfer highlights the importance of economic structure

480 and consumption pattern in local water resource management. The conclusion facilitates the
C

481 process of policy making and implementation, as it balances the trade-offs between economic
AC

482 growth and increased water quality induced-scarcity. It should be noted that the problem of policy

483 making is also a multi-criteria one that should be faced through integrating some criteria into the

484 involvement of other stakeholders.

485

486 Acknowledgements

487 This paper is supported by the Fund for Innovative Research Group of the National Natural
 ACCEPTED MANUSCRIPT
488 Science Foundation of China (Grant No. 51421065), the Projects of Sino-America International

489 Cooperation and Exchanges of NSFC (No. 51661125010), the National Key Research and

490 Development Program of China (No. 2016YFC0503005), National Natural Science Foundation of

491 China (Grant No. 41471466, 71673029) and the 111 Project (No. B17005).

492

PT
493 References

494 Aviso, K.B., Tan, R.R., Culaba, A.B., Cruz, J.B., 2011. Fuzzy input-output model for optimizing

RI
495 eco-industrial supply chains under water footprint constraints. J. Clean. Prod. 19, 187-196.

496 Ayala, L.M., van Eupen, M., Zhang, G.P., Perez-Soba, M., Martorano, L.G., Lisboa, L.S., Beltrao, N.E.,

SC
497 2016. Impact of agricultural expansion on water footprint in the Amazon under climate change

498 scenarios. Sci. Total. Environ. 569, 1159-1173.

U
499 Burke, M., Oleson, K., McCullough, E., Gaskell, J., 2009. A global model tracking water, nitrogen, and
AN
500 land inputs and virtual transfers from industrialized meat production and trade. Environ. Model.

501 Assess. 14, 179-193.

M

502 Beijing Water Resources Bulletin, 2015http://202.112.81.10/files/22110000005547C2/www.bjwater.

503 gov.cn/pub/bjwater/zfgk/tjxx/ndbg/201611/P020161117339356272712.pdf
D

504 Cazcarro, I., Duarte, R., Sanchez-Choliz, J., 2016. Downscaling the grey water footprints of production
TE

505 and consumption. J. Clean. Prod. 132, 171-183.

506 Chang, Y.A., Ries, R.J., Wang, Y.W., 2010. The embodied energy and environmental emissions of
EP

507 construction projects in China: An economic input-output LCA model. Energ. Policy. 38,

508 6597-6603.

509 Chen, Z. M., Chen, G. Q., 2012. Virtual water accounting for the globalized world economy: National
C

510 water footprint and international virtual water trade. Ecol. Indic. 28, 142-149.
AC

511 Davis, S.J., Caldeira, K., 2010. Consumption-based accounting of CO2 emissions. Proc. Natl. Acad.

512 Sci. U. S. A. 107, 5687-5692.

513 Duan, W.L., Chen, G.H., Ye, Q., Chen, Q.G., 2011. The situation of hazardous chemical accidents in

514 China between 2000 and 2006. J. Hazard. Mater. 186, 1489-1494.

515 Duan, W.L., He, B., Takara, K., Luo, P.P., Nover, D., Sahu, N., Yamashiki, Y., 2013. Spatiotemporal

516 evaluation of water quality incidents in Japan between 1996 and 2007. Chemosphere. 93, 946-953.

517 Fischer, H.B., 1967. The mechanics of dispersion in natural streams. J. Hydraul. Div. 93, 187-216.
 ACCEPTED MANUSCRIPT
518 Flach, R., Ran, Y., Godar, J., Karlberg, L., Suavet, C., 2016. Towards more spatially explicit

519 assessments of virtual water flows: linking local water use and scarcity to global demand of

520 Brazilian farming commodities. Environ. Res. Lett. 11, 10.

521 Fu, F.F., Luo, H.B., Zhong, H., Hill, A., 2014. Development of a carbon emission calculations system

522 for optimizing building plan based on the LCA framework. Math. Probl. Eng., 13.

523 Funfgelt, J., Schulze, G.G., 2016. Endogenous environmental policy for small open economies with

PT
524 transboundary pollution. Econ. Model. 57, 294-310.

525 Ghertner, D.A., Fripp, M., 2007. Trading away damage: Quantifying environmental leakage through

RI
526 consumption-based, life-cycle analysis. Ecol. Econ. 63, 563-577.

SC
527 Guan, D., Hubacek, K., 2008. A new and integrated hydro-economic accounting and analytical

528 framework for water resources: a case study for North China. J. Environ. Manage. 88, 1300-1313.

529 Haihe River Water Conservancy Commission (HRWCC), 2008. Haihe River Yearbook 2007. Fang Zhi

530 Press, Beijing, China.

U
AN
531 Han, M.Y., Chen, G.Q., Mustafa, M.T., Hayat, T., Shao, L., Li, J.S., Xia, X.H., Ji, X., 2015. Embodied

532 water for urban economy: A three-scale input-output analysis for Beijing 2010. Ecol. Model. 318,
M

533 19-25.

534 Hawkins, T., Hendrickson, C., Higgins, C., Matthews, H.S., Suh, S., 2007. A mixed-unit input-output
D

535 model for environmental life-cycle assessment and material flow analysis. Environ. Sci. Technol.
TE

536 41, 1024-1031.

537 He, J., Huang, A.P., Xu, L.D., 2015. Spatial heterogeneity and transboundary pollution: A contingent
EP

538 valuation (CV) study on the Xijiang River drainage basin in south China. China. Econ. Rev. 36,

539 101-130.
C

540 Hertwich, E.G., Peters, G.P., 2009. Carbon Footprint of Nations: A Global, Trade-Linked Analysis.
AC

541 Environ. Sci. Technol. 43, 6414-6420.

542 Hoekstra, A.Y., Chapagain, A.K., 2007a. Water footprints of nations: Water use by people as a function

543 of their consumption pattern. Water. Resour. Mange. 21, 35-48.

544 Hoekstra, A.Y., Chapagain, A.K., 2007b. Water footprints of nations: Water use by people as a function

545 of their consumption pattern. Springer, Dordrecht.

546 Islam, M., Kanemoto, K., Managi, S., 2016. Impact of trade openness and sector trade on embodied

547 greenhouse gases emissions and air pollutants. J. Ind. Ecol. 20, 494-505.
 ACCEPTED MANUSCRIPT
548 Ji, S.Y., Chen, B., 2016. Carbon footprint accounting of a typical wind farm in China. Appl. Energ. 180,

549 416-423.

550 Jiang, J.J., Ye, B., Ma, X.M., Ieee, 2013. The Impact of International Greenhouse Gas Emission

551 Constraints on Coal-fired Power Plant in China-Based on LCA Model, 2013 Fourth International

552 Conference on Digital Manufacturing and Automation. Ieee, New York, pp. 1468-1472.

553 Junnila, S., 2008. Life cycle management of energy-consuming products in companies using IO-LCA.

PT
554 Int. J. Life. Cycle. Ass-. 13, 432-439.

555 Ji, J. P., Liu, L., Ma, X. M., 2011. Greenhouse gas emissions by Chinese economy:An assessment

RI
556 based on EIO-LCA model. Acta Scientiarum Naturalium Universitatis Pekinensis. 47, 741-749.

SC
557 Kinna, R., 2016. Non-discrimination and liability for transboundary acid mine drainage pollution of

558 South Africa's rivers: could the UN watercourses convention open Pandora's mine? Water. Int. 41,

559 371-391.

560
U
Kohler, M., 2013. CO2 emissions, energy consumption, income and foreign trade: A South African
AN
561 perspective. Energ. Policy. 63, 1042-1050.

562 Lassaletta, L., Billen, G., Grizzetti, B., Garnier, J., Leach, A.M., Galloway, J.N., 2014. Food and feed
M

563 trade as a driver in the global nitrogen cycle: 50-year trends. Biogeochemistry. 118, 225-241.

564 Lee, K.H., 2011. Integrating carbon footprint into supply chain management: the case of Hyundai
D

565 Motor Company (HMC) in the automobile industry. J. Clean. Prod. 19, 1216-1223.
TE

566 Lenzen, M., Murray, J., Sack, F., Wiedmann, T., 2007. Shared producer and consumer responsibility -

567 Theory and practice. Ecol. Econ. 61, 27-42.

EP

568 Li, Y., 2003. Industry water use in northern China. Chinese Environmental Science Press, Beijing,

569 China.
C

570 Liu, C., Kroeze, C., Hoekstra, A.Y., Gerbens-Leenes, W., 2012. Past and future trends in grey water
AC

571 footprints of anthropogenic nitrogen and phosphorus inputs to major world rivers. Ecol. Indic. 18,

572 42-49.

573 MacDonald, G.K., Bennett, E.M., Carpenter, S.R., 2012. Embodied phosphorus and the global

574 connections of United States agriculture. Environ. Res. Lett. 7, 13.

575 Mekonnen, M.M., Hoekstra, A.Y., 2015. Global gray water footprint and water pollution levels related

576 to anthropogenic nitrogen loads to fresh water. Environ. Sci. Technol. 49, 12860-12868.

577 Mekonnen, M.M., Lutter, S., Martinez, A., 2016. Anthropogenic nitrogen and phosphorus emissions
 ACCEPTED MANUSCRIPT
578 and related grey water footprints caused by EU-27's crop production and consumption. Water. 8,

579 14.

580 Ministry of Water Resources, 2007. China Water Resources Bulletin 2007. China Water and Power

581 Press, Beijing, China.

582 Nansai, K., Kondo, Y., Kagawa, S., Suh, S., Nakajima, K., Inaba, R., Tohno, S., 2012. Estimates of

583 embodied global energy and air-emission intensities of Japanese products for building a Japanese

PT
584 Input-Output Life Cycle assessment database with a global system boundary. Environ. Sci.

585 Technol. 46, 9146-9154.

RI
586 National Bureau of Statistics, 2008. China Statistical Yearbook 2007. China Statistics Press, Beijing,

SC
587 China.

588 National Bureau of Statistics, 2009. Input-output Tables of China 2007. China Statistics Press, Beijing,

589 China.

590
U
National Environmental Protection Agency, 2002. Environment Quality Standard for Surface Ground
AN
591 Water Resources. Standards Press of China, Beijing, China.

592 O'Born, R., Brattebo, H., Iversen, O.M.K., Miliutenko, S., Potting, J., 2016. Quantifying energy
M

593 demand and greenhouse gas emissions of road infrastructure projects: An LCA case study of the

594 Oslo fjord crossing in Norway. Eur. J. Transport. Infrastruct. Res. 16, 445-466.
D

595 Peters, G.P., Hertwich, E.G., 2008. CO2 embodied in international trade with implications for global
TE

596 climate policy. Environ. Sci. Technol. 42, 1401-1407.

597 Peters, G.P., Minx, J.C., Weber, C.L., Edenhofer, O., 2011. Growth in emission transfers via
EP

598 international trade from 1990 to 2008. Proc. Natl. Acad. Sci. U. S. A. 108, 8903-8908.

599 Rafiq, S., Salim, R., Apergis, N., 2016. Agriculture, trade openness and emissions: an empirical
C

600 analysis and policy options. Aust. J. Agr. Resour. Econ. 60, 348-365.
AC

601 Shi, G.M., Wang, J.N., Zhang, B., Zhang, Z., Zhang, Y.L., 2016a. Pollution control costs of a

602 transboundary river basin: Empirical tests of the fairness and stability of cost allocation

603 mechanisms using game theory. J. Environ. Manage. 177, 145-152.

604 Shi, Y.X., Wu, S.H., Zhou, S.L., Wang, C.H., Chen, H., 2016b. International food trade reduces

605 environmental effects of nitrogen pollution in China. Environ. Sci. Pollut. Res. 23, 17370-17379.

606 Shui, B., Harriss, R.C., 2006. The role of CO2 embodiment in US-China trade. Energ. Policy. 34,

607 4063-4068.
 ACCEPTED MANUSCRIPT
608 Su, B., Ang, B.W., 2014. Input-output analysis of CO2 emissions embodied in trade: A multi-region

609 model for China. Appl. Energ. 114, 377-384.

610 Sun, S.K., Liu, J., Wu, P., Wang, Y.B., Zhao, X.N., Zhang, X.H., 2016a. Comprehensive evaluation of

611 water use in agricultural production: a case study in Hetao Irrigation District, China. J. Clean.

612 Prod. 112, 4569-4575.

613 Sun, S.K., Wang, Y.B., Engel, B.A., Wu, P., 2016b. Effects of virtual water flow on regional water

PT
614 resources stress: A case study of grain in China. Sci. Total Environ. 550, 871-879.

615 Wan, L.Y., Cai, W.J., Jiang, Y.K., Wang, C., 2016. Impacts on quality-induced water scarcity: drivers of

RI
616 nitrogen-related water pollution transfer under globalization from 1995 to 2009. Environ. Res.

SC
617 Lett. 11, 10.

618 Wiedmann, T., 2009. A review of recent multi-region input–output models used for consumption-based

619 emission and resource accounting. Ecol. Econ. 69, 211-222.

620
U
Wu, B., Zeng, W., Chen, H., Zhao, Y., 2016. Grey water footprint combined with ecological network
AN
621 analysis for assessing regional water quality metabolism. J. Clean. Prod. 112, 3138-3151.

622 Wu, P.I., Chen, C.T., Cheng, P.C., Liou, J.L., 2014. Climate game analyses for CO2 emission trading
M

623 among various world organizations. Econ. Model. 36, 441-446.

624 Wyckoff, A.W., Roop, J.M., 1994. The embodiment of carbon in imports of manufactured products –
D

625 implications for international agreements. Energ. Policy. 22, 187-194.

TE

626 Xie, Y.M., 1996. Environmental Water Quality Model. China Science and Technology Press, Beijing,

627 China.
EP

628 Yuan, X. H., Ji, X., Chen, H., Chen, B., Chen, G. Q., 2008. Urban dynamics and multiple-objective

629 programming: A case study of Beijing. Commun. Nonlinear. Sci. 13(9), 1998-2017.
C

630 Zeng, Z., Liu, J.G., Savenije, H.H.G., 2013. A simple approach to assess water scarcity integrating
AC

631 water quantity and quality. Ecol. Indic. 34, 441-449.

632 Zhang, J.R., Liu, G.Y., Chen, B., Song, D., Qi, J., Liu, X.Y., 2014. Analysis of CO2 Emission for the

633 cement manufacturing with alternative raw materials: A LCA-based framework, in: Yan, J., Lee,

634 D.J., Chou, S.K., Desideri, U., Li, H. (Eds.), International Conference on Applied Energy,

635 Icae2014. Elsevier Science Bv, Amsterdam, pp. 2541-2545.

636 Zhao, L.J., Qian, Y., Huang, R.B., Li, C.M., Xue, J., Hu, Y., 2012. Model of transfer tax on

637 transboundary water pollution in China's river basin. Oper. Res. Lett. 40, 218-222.
 ACCEPTED MANUSCRIPT
638 Zhao, X., Chen, B., Yang, Z.F., 2009. National water footprint in an input-output framework-A case

639 study of China 2002. Ecol. Model. 220, 245-253.

640 Zhao, X., Liu, J., Liu, Q., Tillotson, M.R., Guan, D., Hubacek, K., 2015. Physical and virtual water

641 transfers for regional water stress alleviation in China. Proc. Natl. Acad. Sci. U. S. A. 112,

642 1031-1035.

643 Zhao, X., Tillotson, M., Yang, Z., Yang, H., Liu, J., 2016. Reduction and reallocation of water use of

PT
644 products in Beijing. Ecol. Indic. 61, 893-898.

645 Zhao, X., Yang, H., Yang, Z.F., Chen, B., Qin, Y., 2010. Applying the Input-Output Method to account

RI
646 for water footprint and virtual water trade in the Haihe River Basin in China. Environ. Sci.

SC
647 Technol. 44, 9150-9156.

648 Zhi, Y., Yang, Z., Yin, X., Hamilton, P.B., Zhang, L., 2015. Using gray water footprint to verify

649 economic sectors' consumption of assimilative capacity in a river basin: model and a case study in

650
U
the Haihe River Basin, China. J. Clean. Prod. 92, 267-273.
AN
651 Zhuo, L., Mekonnen, M.M., Hoekstra, A.Y., 2016. Consumptive water footprint and virtual water trade

652 scenarios for China - With a focus on crop production, consumption and trade. Environ. Int. 94,
M

653 211-223.
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
1 Table 1. Results for the major economic sectors in Beijing during year 2007.

2 Table 2. Comparison of the GWFs under different water quality levels

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
5 Table 1. Results for 17 economic sectors and households in Beijing during year 2007.

Quantity of
rij GWFI
Sectors Wastewater DDWRi GWF
(×10-4 m3 (×10-4 m3
Discharge (×107 m3) (107 m3)
RMB-1) RMB-1)
(×107 m3)

1 Agriculture 56.76 98.80 361.74 42.19 214.40

PT
2 Mining industry 0.79 2.27 14.02 0.45 4.83

3 Food and tobacco 3.20 12.41 23.90 35.45 47.25

RI
4 Textile, clothing and leather products 1.14 3.27 17.38 3.82 10.14

5 Wood, paper and other 3.97 14.88 54.80 3.65 30.90

SC
6 Electricity, heat and hot water 5.31 3.49 3.83 0.65 9.85

7 Coke, gas and oil processing 1.52 3.65 9.31 2.72 6.08

8 Chemical industry 5.58 9.50 10.24 7.71 12.07

9 Non-metallic mineral products

U
0.63 2.23 7.91 0.91 6.72
AN
10 Metal products 2.08 3.20 4.33 1.24 3.33

Machinery and equipment

11 2.92 4.45 0.93 12.85 3.19
M

manufacturing

12 Construction 7.49 14.17 7.33 24.17 10.32

D

Transportation, postal services,

13 information transmission, computer 10.10 28.30 7.48 30.94 10.65

TE

services and software

Wholesale, retail trade,

EP

14 5.23 15.37 6.98 20.80 13.92

accommodation, and catering

15 Real estate and renting 11.14 31.06 12.55 29.08 18.10

C

16 Financial sector 5.21 15.31 8.24 17.80 12.85

17 Other services 10.71 29.92 5.98 57.85 13.55

AC

18. Household ∗ 4.79 13.04 13.04

Total 305.32 305.32

∗
6 Notes: The 18th sector, households,are not listed as one of the economic sectors. The GWF of households, which
7 equals DDWRi, can be obtained by inputting the quantity and contaminant concentration of living sewage based on
8 equation 7-8.

9
 ACCEPTED MANUSCRIPT
10 Table 2. Comparison of the TGWFs under different water quality levels

Grade I II III IV V

GWF 1665.84 830.61 622.95 415.30 305.32

11

12

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
1 Fig. 1. The framework for the GWF assessment based on EIO-LCA method

2 Fig 2. Flow analysis among 17 economic sectors (unit: 107 m3)

3 Fig 3. Comparison of EGWF and IGWF of 17 economic sectors

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
5 Fig. 1. The framework for the GWF assessment based on EIO-LCA method
AN
6

7
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
8

PT
RI
U SC
AN
M
D
TE

10 Fig 2. Flow analysis among 17 economic sectors (unit: 107 m3)

EP

11

12
C
AC
 ACCEPTED MANUSCRIPT
13

70
GWF ( 107 m3) 60 EGWF IGWF
50

40

PT
30

20

RI
10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

SC
Sector
14

15 Fig 3. Comparison of EGWF and IGWF of 17 economic sectors

U
16
AN
17
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
Fig 3. Comparison of EGWF and IGWF of 17 economic sectors

U
AN
M
D
TE
EP
C
AC