Water pollution loads and shifting within China's inter-province trade

Water pollution loads and shifting within China's inter-province trade

Zhaodan Wu

, Quanliang Ye

a_{Business School, Hohai University, Nanjing, 211100, China}

b_{Jiangsu Provincial Collaborative Innovation Center of World Water Valley, Water Ecological Civilization, Nanjing, 211100, China} c_{One Belt& One Road African Research Center, Hohai University, Changzhou, 213022, China}

d_{Faculty of Engineering Technology, University of Twente, Enschede, the Netherlands}

a r t i c l e i n f o

Article history: Received 13 August 2019 Received in revised form 1 February 2020 Accepted 2 March 2020 Available online 6 March 2020

Handling Editor: Dr. Govindan Kannan

Keywords: Gray water footprint Pollution loads and shifting Multi-regional input-output model

a b s t r a c t

Inter-province trade of commodities and services involves not only the virtual waterflows, but also the water pollution shifting which influences the water quality stress in both import and export provinces. This study estimated the production-based and consumption-based gray water footprints (GWFs) of chemical oxygen demand (COD) and ammoniacal nitrogen (NH3eN) of 30 provinces in China, and further elaborated the pollution shifting due to inter-province trade of commodities and services. The national GWF was 1586 billion m3_{in 2012, mainly distributed in agriculture-dominated provinces, e.g.,} Guang-dong, Hunan, Henan, Jiangsu, ShanGuang-dong, Sichuan and Hubei. Our results also found that the national net water pollution shifting volumes were 2.8 million tons of COD and 0.2 million tons of NH3eN. Relatively wealthy provinces, such as Beijing, Shanghai and Tianjin, played significant roles in the national net pollution shifting. In accordance with the shifting of COD and NH3eN, 525 billion m3of net GWFs were shifted from the importers to the exporters. Hebei, Shandong and Guangdong were the major outsourcing provinces of GWFs, whereas Hunan, Henan, and Heilongjiang were the major receiving provinces. We also found that most of the major water pollution receiving provinces present severe water quality stresses, particularly for Jiangsu, Henan and Anhui. Our results present the comprehensive water pollution consequences (internal and external) of daily economic activities in each province, and provide scientific instructions for inter-regional environment protection and policy making in China and far beyond.

© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Apart from the socioeconomic implications, trade (both inter-regional and intra-inter-regional) has had broad and significant im-pacts on natural environments, e.g. by affecting carbon dioxide (CO2) emissions, air pollution, and biodiversity in importing and exporting regions (Feng et al., 2013;Guan et al., 2008;Moran et al., 2016; Yang and Suh, 2011). The exchange of the water flows embedded in goods/services traded among regions is commonly referred to as virtual water trade (Allen, 1992;Hoekstra and Hung, 2002). Previously, researchers have focused on analyzing the water quantify implications of virtual water trade. A few of them explic-itly assessed the impacts of virtual water trade on water quantity stress, revealing various (i.e. relieving or aggravating) effects at the global (Hoekstra and Mekonnen, 2012;Lenzen et al., 2013;Wang

and Zimmerman, 2016), national (Dalin et al. 2014, 2015), and river basin (Zhao et al., 2010) scales. However, little is known about how water pollution is shifted from one place to another through commodity trade. A prior study showed that fast-growing econo-mies with liberal trade (such as Chile) experienced less water quality stress growth than closed economies (such as Bolivia and El Salvador), because the former imported pollution-intensive goods and services from the latter (Low and Yeats, 1992). To further study the pollution aspect of virtual water trade, detailed data of water pollution discharge by producing sectors are essential but are just becoming available.

Gray water footprint (GWF) is always used to quantify the pollution burden in term of water volume, which is able to express water pollution in the same unit with water quantity. GWF is measured as the volume of water required to assimilate pollution based on existing ambient water quality standards (Chapagain et al., 2006;Postel et al., 1996). Liu et al. (2012) calculated past and future trends in GWFs related to anthropogenic nitrogen (N) and phosphorus (P) inputs into major rivers around the world, and * Corresponding author.

E-mail address:q.ye@utwente.nl(Q. Ye).

Contents lists available atScienceDirect

Journal of Cleaner Production

https://doi.org/10.1016/j.jclepro.2020.120879

found that about two-thirds of the rivers had severe water pollu-tion, and their pollution assimilation capacity has been fully consumed. Expanding the scope from major rivers to global rivers, Mekonnen and Hoekstra (2015) and (Mekonnen and Hoekstra, 2018) estimated N-related and P-related loads per sector and crop, and found China contributed about half of the global N loads as well as one third of P loads, respectively. The large water pollution discharges exacerbate China's water scarcity, with a se-vere consequence that over half of the population are affected by water scarcity (Ma et al., 2020). Although these studies have made great contributions for water pollution loads estimation for a global scope, few studies focused on the impacts of “virtual” water pollutionflows occurring separately and/or interdependently on the water-quality stress in import and export regions. The only one undertaken byZhao et al. (2016)drew water resources in Shanghai from all over China and outsourced the pollution through virtual quality water flows associated with trade. However, the regions facing severe water environmental problems are always those with net commodity and service export, instead of these developed cities like Shanghai in China. Thus, a systematic and comprehensive evaluation of virtualflows of water pollution is needed.

This study aims to elaborate the water pollution shifting due to inter-province trade of commodities and services within China. We estimated both production-based and consumption-based GWFs of water pollution indicators, chemical oxygen demand (COD) and ammoniacal nitrogen (NH3eN) in 30 provinces (except Tibet, Hong Kong, Macao and Taiwan). The production-based GWF is the direct GWF of water pollution released by different producing sectors. Particularly, we highlighted six sectors with high COD and NH3eN loads for further analysis, i.e.‘Agriculture’, ‘Paper and paper prod-ucts’, ‘Food and tobacco manufacturing’, ‘Chemical raw materials and chemical products manufacturing’, ‘Textile industry’, and ‘Services’. The consumption-based GWF is the total GWFs stem-ming from thefinal demand. A multi-regional input-output (MRIO) model was used to estimate the consumption-based water pollu-tion loads, and the associated GWFs. The advantage of using the MIRO model to analyze consumption-based environment issues was that distinguishing production structure and technology for each region. Finally, the virtual water pollution flows among 30 provinces were elaborated, and the environmental consequences were analyzed. This study could present the comprehensive water pollution consequences (internal and external) of daily economic activities in each province, and provide scienti_{fic instructions for} inter-regional environment protection and policy making.

2. Method

2.1. Consumption-based pollution footprint accounting and pollution embodied in inter-province trade

2.1.1. Multi-regional input-output modelling

A multi-regional input-output (MRIO) model is used to calculate the consumption-based water pollution offinal demand in each province, and the associated water pollution embodied in the inter-province trade. First, the national MRIO table should be established and described (Table 1). The basic framework of a MRIO table separates the economic outputs as intermediate inputs and final demands. We have: Xr i¼ Xm s¼1 Xn j¼1 xrs ij þ Xm s¼1 Frs i þ Eri (1) where Xr

i is the total output of sector i in region r; xrsij is the inter-mediate input of sector i in region r to sector j in region s; Frs

i is the

final demand of region s from sector i in region r; Er

i is the inter-national export of sector i in region r; m and n is the total numbers of regions and sectors, respectively.

The direct input coefficient (ars

ij) is defined as the direct input from sector i in region r for increasing one monetary unit output of sector j in region s (Dong et al., 2014), calculated by Eq.(2):

ars_ij¼x rs ij Xs j (2)

Then Eq.(1)could be rewritten as:

Xr_i¼Xm s¼1 Xn j¼1 ars_ijX_jsþXm s¼1 F_irsþ Er i (3)

Eq.(3)could also be transferred into matrix formats as Eq.(4)or Eq.(5):

X¼ AX þ Fe þ E (4)

X¼ ðI  AÞ1ðFe þ EÞ ¼h

a

i

ðFe þ EÞ (5)

where X is the mn 1 total output matrix, with elements of Xr i; A is the mn mn matrix of direct intermediate input coefficients, with elements of ars

ij; I is the identity matrix, with 1 in the diagonal cells; F is the mnm finial demand matrix, with elements of Frs

i ; e is a m 1 matrix with ones; E is the mn  1 export matrix, with ele-ments of Er

i;½

a

a

ij defined as the total input from sector i in region r for increasing one monetary unit output of sector j in region s.

When we extend the MRIO table with a satellite environmental account table, here we use water pollution loads from each sector in each province (PLr

i), the direct pollution load coefficient (dplri) could be calculated by the same way as ars

ij, representing the direct pollution load volume for increasing one monetary unit output of sector i in region r:

dplr_i¼PLri

Xr i

(6)

The total pollution load coefficient (tplr

i) is defined as the total pollution load volume for increasing one monetary unit output of sector i in region r, shown as:

tplr_i¼Xm s Xn j dpls_j,

a

The total water pollution intensity tplr

i is considered both the direct water pollution loads and indirect pollution loads for a sys-temic input-output evaluation.

Thus, the consumption-based pollution loads of thefinal de-mand in region r could be calculated as:

TPLr¼Xm s Xn j tpls_j,Fsr j (8)

2.1.2. Inter-province pollution shifting calculation

Inter-province virtual pollution shifting is caused by the trade of commodities and services from export regions to import regions. Like virtual water trade (water scarce regions through importing water-intensive commodities from water-abundant regions to relieve the pressure on their own water resources), commodity import regions have shifted the burden of water pollution stress during commodity production periods to commodity export

regions through trade activities. Thus, in this study, the inter-province pollution shifting volumes are calculated by:

RPrs¼Xn i tplr_i,Frs i (9) OPrs¼Xn j tpls_j,Fsr j (10)

where RPrswas the total receiving pollution volume of region r from region s; OPrswas the total outsourcing pollution volume of region r to region s; The net pollution outsourcing volume could be shown as:

NOPr_{¼ OP}rs_{ RP}rs ₍₁₁₎

2.2. Gray water footprint (GWF) accounting

GWF is calculated by dividing the COD and NH3eN load by the difference between the ambient water quality standard for COD and NH3eN (the maximal acceptable concentration, cmax) and the natural concentration of COD and NH3eN in the receiving water body (cnat), respectively.

GWF¼ L

cmax cnat (12)

where L is the total load of pollutants, for production-based GWF accounting, L ¼ Pn

i PLr

i, while for consumption-based GWF ac-counting, L¼ TPLr_{; c}

maxis derived from the Environmental Quality Standards for Surface Water (MEP, 2002). According to this stan-dard, the maximum concentration (cmax) of COD and NH3eN are 20 mg/L and 1 mg/L, respectively, which represent the minimum quality of water that is suitable forfishery, aquaculture, and rec-reational use. cnatis usually set as zero due to data limitations (Zeng et al., 2013). The total GWF of a region is selected as the maximal GWF between COD and NH3eN (Sun et al., 2016).

GWFr_{¼ max}_GWFr

COD; GWFrNH3N



(13)

2.3. Data

The sectoral COD and NH3eN discharge data at a provincial level were collected from China's Environmental Statistics Yearbook 2012 (NBS, 2012). Annual COD and NH3eN discharges are reported for agricultural, industrial and residential sources in each province. Particularly, the national COD and NH3eN discharges from four high-discharging industrial sectors are provides, i.e. ‘Paper and paper products’, ‘Food and tobacco manufacturing’, ‘Chemical raw materials and chemical products manufacturing’, and ‘Textile in-dustry’. We allocate the national COD and NH3eN discharges from industrial sectors to each province based on the water withdrawal data of associated sectors. We assume the more water withdrawn for sectoral production, the higher discharges of water pollution. The withdrawal data of industrial sectors were collected from Chinese Economic Census Yearbook in 2008 (CEC, 2008), and scaled up according to provincial industrial water withdrawal in 2012 (NBSC, 2013). The residential COD and NH3eN discharges were not reported separately between household and service sectors. We also utilized the water withdrawal proportion in residential water withdrawal between households and service sectors as a proxy to estimate the COD and NH3eN discharges in service sectors (Guan et al., 2014).

The national MRIO table of 2012 fromMi et al. (2017)was used for consumption-based pollution footprint accounting. In order to match the resolution of sectors of MRIO table and water pollution discharges, we aggregated the 41 industrial sectors with water pollution data into the 22 industrial sectors of MRIO tables, which are listed in Table S1. The Chinese MRIO table includes several categories of_{final demand: rural and urban households,} govern-ment expenditure, gross capital formation and changes in stocks. Here the focus of this study was the impact of inter-province commodity trade on water pollution loads to other provinces within China, and thus, the virtual COD and NH3eN flows embedded in international trade were not considered.

3. Results

3.1. Gray water footprints in China

The national GWF was 1586 billion m3in 2012 (Fig. 1). The top three provinces for production-based GWF were Guangdong, Hu-nan and Sichuan, contributing 8.3%, 8.0% and 6.6% in the national Table 1

Format of a national multi-region input-output (MRIO) table.

Intermediate Input Final Demand Export Total Output Province 1 … Province m Province 1 … Province m

S1 … Sn … S1 … Sn

Intermediate Input Province 1 S1 x11₁₁ … x11_1n … x1m₁₁ … x1m_1n F₁11 … F₁1m E1₁ X₁1

… … … … … Sn x11 n1 … x11nn … x1mn1 … x1mnn Fn11 … Fn1m E1n Xn1 … … … … Province m S1 xm1 11 … xm11n … x mm 11 … xmm1n F1m1 … F mm 1 Em1 X1m … … … … Sn xm1 n1 … xm1nn … x mm n1 … xmmnn Fnm1 … F mm n Emn Xnm Import IM1 1 … IM1n … IMm1 … IMmn FM1 … FMm Added Value V1 1 … Vn1 … V m 1 … Vnm Total Input X1 1 … Xn1 … X m 1 … Xnm Pollution Load PL1 1 … PL1n … PLm1 … PLmn

total, respectively; while for consumption-based GWF were Guangdong, Sichuan and Shandong, accounting for 9.4%, 6.4% and 6.1% in the national total. The GWFs showed an obvious distinction among 30 provinces, and the ratio between the maximal and the minimal GWF was up to 24 times for production-based (maximal 132 billion m3 in Guangdong whereas minimal 5.4 billion m3 in Qinghai) and 20 times for consumption-based (maximal 150 billion m3 in Guangdong whereas minimal 7.3 billion m3 in Qinghai). When the provincial production-based GWF is smaller than the consumption-based GWF, it means that the water required to assimilate the water pollution discharged for the production of commodities and services in that province is smaller than that for the commodities and services consumed in that province. In another words, the province partly outsources the local water pollution to other regions. In our study, relative wealthy provinces, such as Beijing, Tianjin, Zhejiang, Guangdong and Shandong, were notable for this outsource, which could be explained by the considerable import of goods and services from other provinces to satisfy thefinal demand of local population.

The major GWFs, both production-based and consumption-based, distributed in agriculture-dominated provinces, e.g., Guangdong, Hunan, Henan, Jiangsu, Shandong, Sichuan and Hubei (Fig. S1). Further analyzing the production-based GWFs of specific sectors (Fig. S2): ‘Agriculture’ sector highlighting in Shandong, Henan and Hunan;‘Food and tobacco manufacturing’ sector high-lighting in Guangxi, Hunan and Inner Mongolia;_{‘Paper and paper} products’ sector highlighting in Hunan and Ningxia two provinces; ‘Textile industry’ sector highlighting in Jiangsu, Zhejiang and Guangdong; ‘Chemical raw materials and chemical products manufacturing’ sector highlighting in Hunan and Hubei; ‘Services’ sectors dominant in eastern China. The rationalization of sectoral structure from the water pollution perspective, i.e. the degree of the equilibrium of sectoral pollution loads, was also assessed through the standardized information entropy (details could be found in Supporting Information). The degree of the equilibrium of sectoral pollution loads in the China's agriculture oriented provinces, such as Shandong, Henan, Heilongjiang, Hebei and Sichuan, were overall lower than that other provinces. As agriculture for the dominant sector in these provinces, an obvious distinction of pollution loads has exited among recently emerging sectors (e.g. internet

technology, and services) and traditional historic sectors.

The GWF only provided the pollution loads from sectors, which could not reflect the water pollution stress in a region. The regional water pollution stress was determined by the water pollution loads and local available water resources. If the GWF less than local available water resources, it meant that local available water re-sources could assimilate the pollution loads under the local water quality standard, i.e. no water pollution stress. Water pollution level (WPL, the ratio of the total GWF to the water availability volume) was always used to measure the degree of water pollution stress in previous studies (Liu et al., 2012;Mekonnen and Hoekstra, 2015, 2018; Zhao et al., 2016). Fig. 1 illustrated the GWFs and local available water resources in 30 provinces. Hence, half provinces had no water pollution stress. One thing should be concerned is that the GWFs in Anhui, Chongqing, Shaanxi, and Gansu were almost equal to local available water resources. When we estimate both water quantity and quality stress (considering both water consumption and GWF) in these provinces, the water stress would be severe. For the provinces with severe water pollution stress, they could be sorted into two categories: one were provinces with limited water resources, like Shanghai, Tianjin and Beijing; another were provinces with high water pollution loads, such as Shandong, Henan, and Jiangsu.

3.2. Water pollution shifting among 30 provinces

The trade of commodities and services outsourced the water pollution from the import regions to the export. The national net pollution shifting volumes among 30 provinces were 2.8 million tons of COD and 0.2 million tons of NH3eN in 2012. Half of 30 provinces showed a net COD outsourcing whereas fourteen prov-inces showed a net NH3eN outsourcing. It implied that these provinces would shift the responsibility, at least some, of water pollution loads for their ownfinal demand to other provinces. The major water pollution were shifted from the east and the southeast to the north, the middle and the northwest (Fig. 2). The shifting of COD illustrated broader and more complex connections between provinces than that of NH3eN. Shandong, Shanghai, Beijing, and Hebei net outsources the most responsibilities of COD loads to other provinces, meanwhile Heilongjiang, Henan and Jilin were the Fig. 1. Production-based and consumption-based gray water footprints (GWFs) and the available water resources of 2012 in 30 provinces of China.

Fig. 2. Net outsourcingflows of embodied COD and NH3eN in the trade of commodity and services from the importing provinces (with bubbles) to the exporting provinces. Top seven were shown.

main net receiving provinces from other provinces (details could be found inFig. S5). In addition, Shandong, Guangdong, Hebei, Beijing and Zhejiang net outsourced the most responsibility of NH3eN loads to other provinces, meanwhile Hunan, Henan and Anhui were the main net receiving provinces of NH3eN (details could be found inFig. S5). Beijing, Chongqing, and Zhejiang showed a relative low net receiving of COD and NH3eN, or even zero net receiving in Shanghai and Shanxi. These relatively wealthy regions in China significantly outsourced their water pollution loads for local de-mand to poorer central and western provinces. Particularly in Beijing and Shanghai, more than 70% of COD and NH3eN pollution derived by product consumption in Beijing and Shanghai occurred in other regions, mostly for agricultural products and services, together accounting for>90% of the total net shifting volume.

The associated outsourcingflows of GWFs embodied in inter-province trade were illustrated in Fig. 3. The main net outsourcing flows of GWF were consistent to the agricultural dominated provinces, from Shandong to Hebei (4.5 billion m3), from Guangdong to Hunan (4.0 billion m3), from Shandong to Heilongjiang (3.8 billion m3), and from Hebei to Heilongjiang (3.5 billion m3), accounting for 9% of the national total. Jiangsu and Liaoning were not highlighted in the analysis of net shifting of water pollution, which could be explained by the almost same

volumes of water pollution outsourcing and receiving (GWF outsourcing V.S. receiving: Jiangsu 30.4 billion m3V.S. 34.4 billion m3; Liaoning 22.4 billion m3V.S. 21.8 billion m3). It did not mean that the roles these provinces played in national GWF shifting were not significant. In contrast, actually, these provinces act actively in national commodity markets, and seek more and more bilateral and multilateral corporations with other provinces. Comparing with the local water resources to estimate the impact of commodity trade in local water quality stress, the topfive GWF receivers, i.e. Hunan, Henan, Heilongjiang, Anhui and Jiangsu, three of them suffered from severe water quality stresses in local water envi-ronmental (Fig. 1). Particularly for Anhui, the water quality stress shifted from sever to extreme due to the inter-regional commodity trade. For the future water environment protection, the mecha-nisms, regulations, and decision-making systems should be con-structed to limit water pollution loads and shifting within and among provinces, and to instruct the managers to deal with the water quantity and quality stress in China.

4. Discussions

This study estimated the both production-based and consumption-based gray water footprints of water pollution Fig. 3. Outsourcingflows of embodied gray water footprints in the trade of commodity and services between 30 provinces in China.

discharges in China, and further elaborated the virtualflows of water pollution within inter-province trade. The national GWF was 1586 billion m3 in 2012, mainly distributed in agriculture-dominated provinces, e.g., Guangdong, Hunan, Henan, Jiangsu, Shandong, Sichuan and Hubei. Agricultural and service sectors accounted for the main pollution discharges in China as well as in most provinces. Because of the commodity trade, 525 billion m3of GWFs were shifted from the exporters to the importers. Hebei, Shandong and Guangdong were the major outsourcing provinces of GWFs, whereas Hunan, Henan, and Heilongjiang were the major receiving provinces. When we focus on the net shifting of water pollution, relatively wealthy provinces, such as Beijing, Shanghai and Tianjin, played significant roles in the national net water pollution shifting. We also found that most of the major water pollution receiving provinces present severe water quality stresses, particularly for Jiangsu, Henan and Anhui.

The most interesting results were existing in Guangdong prov-ince. The GWF of each sector in Guangdong showed a large volume in previous analysis. In this context, Guangdong still outsourced a considerable volume of its water pollution to other provinces, especially to Hunan, Guangxi and Jiangsu. It was largely for the high requirements of intermediate inputs to local manufacturing sec-tors. Guangdong contributes more than a quarter of made-in-China clothes, leather and furs. Guangdong's fibre, textile and dyeing sectors account significantly for China's water pollution (Oita et al., 2016). Although the available water resource in Guangdong is suf_{ficient to assimilate the current water pollution loads, it did not} put Guangdong in a sustainable development direction yet. When we consider the both water demand for consumption and pollu-tion, the water stress in Guangdong will still be considerable (Ma et al., 2020). In order to increase the water supplement, seawater desalination and wastewater reclamation have been paid great attention these years. The recent seawater utilization volume is more than 39 billion m3every year in Guangdong. Moreover, the construction of wastewater treatment plants and the investments for municipal pipe networks of reclaimed wastewater are all considered in the Five-Year Plan of Guangdong province.

The significant but uneven distribution within China is not only natural resources, but also economic activities, level of technolog-ical development, and levels of consumption and pollution. While water quantity stress in northern China has received much atten-tion, the water quality stress in the south and across the country has been largely neglected. A widespread pollution, mostly from the sectors we selected in this paper for further analysis, has resulted in 75% of lakes and rivers and 50% of groundwater supplies contaminated (Zhang et al., 2014). This situation has caused great concerns in southern China for the lack of available water even for irrigation purpose. This is mainly due to the high cost of industrial wastewater treatments and weak awareness of environment pro-tection. It should also be noted that to seek sustainable urbaniza-tion, a series of policies and instruments have been carried out to protect water resources, such as the well-known“Three Red Lines” water policy,‘Sponge City’ construction, as well as the substantial investments in constructing wastewater treatment plants in large cities (Liu et al., 2013). Therefore, some suggestion could be pointed out for policy making or decision making in the future: 1) updating technologies especially for highly pollution-intensive sectors, like chemical manufacturing and textile production (Levinson, 2009; Quere et al., 2019;Zhao and Chen, 2014); 2) more investments in unconventional water supply systems (e.g. seawater desalination, rainwater harvesting) and wastewater treatment plants (Thacker et al., 2019); 3) optimizing economic structure, regions with suf fi-cient water and other resources should fully take the advantages to produce cheaper and greener productions, which could make them more competitive within national even international markets; 4)

seeking multilateral corporation and coordinated development. However, there are still some limitations in our study. Future studies will need to compile a more systemic and comprehensive dataset of water pollution, with a higher resolution in sectors, a longer time series, and involving more provinces. Lack of water pollution discharge data requires more assumptions to formulate the satellite accounts, which will more or less reduce the reliability of our results. Long time-series water pollution discharge data could provide robust empirical analysis to examine the trends and drivers of the virtualflows of water consumption and pollution, to implement factor-specific measurements to reduce water con-sumption and pollution. Furthermore, the unprecedented interna-tional trade also highlights the importance to analyze the influence of globalization in the global water pollution discharges, and one step further, the nitrogen cycle in natural ecosystems.

5. Conclusion

This study estimated the both production-based and consumption-based gray water footprints (GWFs) of 30 provinces in China, and further elaborated the pollution shifting due to inter-province trade of commodities and services. The national GWF was 1586 billion m3, mainly distributed in agriculture-dominated provinces, e.g., Guangdong, Hunan, Henan, Jiangsu, Shandong, Sichuan and Hubei. Our results also indicated that the national net water pollution shifting volumes were 2.8 million tons of COD and 0.2 million tons of NH3eN. Relatively wealthy provinces, such as Beijing, Shanghai and Tianjin, played significant roles in the na-tional net pollution shifting. In accordance with the shifting of COD and NH3eN, 525 billion m3 of net GWFs were shifted from the exporters to the importers. Hebei, Shandong and Guangdong were the major outsourcing provinces of GWFs, whereas Hunan, Henan, and Heilongjiang were the major receiving provinces. We also found that most of the major water pollution receiving provinces present severe water quality stresses, particularly for Jiangsu, Henan and Anhui. Our results present the comprehensive water pollution consequences (internal and external) of daily economic activities in each province, and provide scientific instructions for inter-regional environment protection and policy making in China and far beyond.

Declaration of competing interest

The authors claim no conflicts of interest.

CRediT authorship contribution statement

Zhaodan Wu: Data curation, Methodology, Software, Pro-gramming, Writing- Reviewing and Editing; Quanliang Ye: Soft-ware, Writing- Original draft preparation, Writing- Reviewing and Editing, Supervision.

Acknowledgements

This work was supported by the Youth Foundation for Hu-manities and Social Sciences of the Ministry of Education of China “Study on the Coordination Mechanism of Dynamic Benefit in the Water Environment Treatment Integration in the Yangtze Delta”; the Fundamental Research Funds for the Central Universities “Economic Efficiency of Broad-sense Water Use for Crop Production and Its Influencing Factors in the Main Grain-producing Areas of China” (2017B19014). We also appreciate the initial help of Q Zhou for the framework of this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jclepro.2020.120879.

References

Allen, T., 1992. Fortunately There Are Substitutes for Water Otherwise Our Hydro-political Futures Would Be Impossible in Priorities for Water Resources Allo-cation and Management. ODA, London, pp. 13e26.

CEC, 2008. Chinese Economic Census Yearbook. The State Council Leading Group Office of Second China Economic Census. China Stat. Press, Beijing, China. Chapagain, A.K., Hoekstra, A.Y., Savenije, H.H.G., Gautam, R., 2006. The water

footprint of cotton consumption: an assessment of the impact of worldwide consumption of cotton products on the water resources in the cotton producing countries. Ecol. Econ. 60 (1), 186e203.

Dalin, C., Hanasaki, N., Qiu, H., Mauzerall, D.L., Rodriguez-Iturbe, I., 2014. Water resources transfers through Chinese interprovincial and foreign food trade. Proc. Natl. Acad. Sci. U.S.A. 111, 9774e9779.

Dalin, C., Qiu, H., Hanasaki, N., Mauzerall, D.L., Rodriguez-Iturbe, I., 2015. Balancing water resource conservation and food security in China. Proc. Natl. Acad. Sci. U.S.A. 112 (15), 4588e4593.

Dong, H., Geng, Y., Fujita, T., Fujii, M., Hao, D., Yu, X., 2014. Uncovering regional disparity of China's water footprint and inter-provincial virtual waterflows. Sci. Total Environ. 500e501, 120e130.

Feng, K., Davis, S.J., Sun, L., Li, X., Guan, D., Liu, W., Liu, Z., Hubacek, K., 2013. Outsourcing CO2 within China. Proc. Natl. Acad. Sci. U.S.A. 110 (28), 11654e11659.

Guan, D., Hubacek, K., Tillotson, M., Zhao, H., Liu, W., Liu, Z., Liang, S., 2014. Lifting China's water spell. Environ. Sci. Technol. 48, 11048e11056.

Guan, D., Hubacek, K., Weber, C.L., Peters, G.P., Reiner, D.M., 2008. The drivers of Chinese CO2 emissions from 1980 to 2030. Global Environ. Change 18, 626e634.

Hoekstra, A.Y., Hung, P.Q., 2002. Virtual Water Trade: A Quantification of Virtual Water Flows between Nations in Relation to International Crop Trade. IHE Delft. Hoekstra, A.Y., Mekonnen, M.M., 2012. The water footprint of humanity. Proc. Natl.

Acad. Sci. U.S.A. 109 (9), 3232e3237.

Lenzen, M., Moran, D., Bhaduri, A., Kanemoto, K., Bekchanov, M., Geschke, A., Foran, B., 2013. International trade of scarce water. Ecol. Econ. 2013, 78e85. Levinson, A., 2009. Technology, international trade, and pollution from US

manufacturing. Am. Econ. Rev. 99 (5), 2177e2192.

Liu, C., Kroeze, C., Hoekstra, A.Y., Gerbens-Leenes, W., 2012. Past and future trends in grey water footprints of anthropogenic nitrogen and phosphorus inputs to major world rivers. Ecol. Indicat. 18, 42e49.

Liu, J., Zang, C., Tian, S., Liu, J., Yang, H., Jia, S., You, L., Liu, B., Zhang, M., 2013. Water conservancy projects in China: achievements, challenges and way forward. Global Environ. Change 23, 633e643.

Low, P., Yeats, A., 1992. International Trade and the Environment. World Bank Discussion Papers. World Bank, Washington, DC, pp. 89e104.

Ma, T., Sun, S., Fu, G., Hall, J.W., Ni, Y., He, L., Yi, J., Zhao, N., Du, Y., Pei, T., Cheng, W., Song, C., Fang, C., Zhou, C., 2020. Pollution exacerbates China's water scarcity

and its regional inequality. Nat. Commun. 11 (1).

Mekonnen, M.M., Hoekstra, A.Y., 2015. Global gray water footprint and water pollution levels related to anthropogenic nitrogen loads to fresh water. Environ. Sci. Technol. 49, 12860e12868.

Mekonnen, M.M., Hoekstra, A.Y., 2018. Water Resour. Res.https://doi.org/10.1002/ 2017WR020448. Accepted Article.

MEP, 2002. Ministry of Environment Protection of China) Environmental Qual-ityStandards for Surface Water [GB3838-2002]. Stand. Press of China, Beijing, China.

Mi, Z., Meng, J., Guan, D., Shan, Y., Song, M., Wei, Y.M., Liu, Z., Hubacek, K., 2017. Chinese CO2 emissionflows have reversed since the global financial crisis. Nat. Commun. 8 (1), 1712.

Moran, D., Petersone, M., Verones, F., 2016. On the suitability of inputeoutput analysis for calculating product-specific biodiversity footprints. Ecol. Indicat. 60, 192e201.

NBS, 2012. In: o, N.B. (Ed.), China Environmental Statistical Yearbook 2012. Statis-tics. China Statistics Press, Beijing.

NBSC, 2013. In: National Statistic Yearbook (2013). National Bureau of Statistics of the People's Republic of China, Beijing, China.

Oita, A., Malik, A., Kanemoto, K., Geschke, A., Nishijima, S., Lenzen, M., 2016. Sub-stantial nitrogen pollution embedded in international trade. Nat. Geosci. 6, 111e115.

Postel, S.L., Daily, G.C., Ehrlich, P.R., 1996. Human appropriation of renewable fresh water. Science 271 (5250), 785e788.

Quere, C.L., Korsbakken, J.I., Wilson, C., Tosun, J., Andrew, R., Andres, R.J., Canadell, J.G., Jordan, A., Peters, G.P., Vuuren, D.P.v., 2019. Drivers of declining CO2 emissions in 18 developed countries. Nat. Clim. Change 9, 213e217. Sun, C.Z., Han, Q., Zheng, D.F., 2016. The spatial correlation of the provincial grey

water footprint and its loading coefficient in China. Acta Ecol. Sin. 36 (1), 86e97 (in Chinese).

Thacker, S., Adshead, D., Fay, M., Hallegatte, S., Harvey, M., Meller, H., O'Regan, N., Rozenberg, J., Watkins, G., Hall, J.W., 2019. Infrastructure for sustainable development. Nat. Sustain. 2, 324e331.

Wang, R., Zimmerman, J., 2016. Hybrid analysis of blue water consumption and water scarcity implications at the global, national, and basin levels in an increasingly globalized world. Environ. Sci. Technol. 50 (10), 5143e5153. Yang, Y., Suh, S., 2011. Environmental impacts of products in China. Environ. Sci.

Technol. 45, 4102e4109.

Zeng, Z., Liu, J., Savenije, H.H.G., 2013. A simple approach to assess water scarcity integrating water quantity and quality. Ecol. Indicat. 34, 441e449.

Zhang, W., Wang, C., Li, Y., Wang, P., Wang, Q., Wang, D., 2014. Seeking sustain-ability: multiobjective evolutionary optimization for urban wastewater reuse in China. Environ. Sci. Technol. 48 (2), 1094e1102.

Zhao, C., Chen, B., 2014. Driving force analysis of the agricultural water footprint in China based on the LMDI method. Environ. Sci. Technol. 48 (21), 12723e12731. Zhao, X., Liu, J., Yang, H., Duarte, R., Tillotson, M.R., Hubacek, K., 2016. Burden shifting of water quantity and quality stress from megacity Shanghai. Water Resour. Res. 52 (9), 6916e6927.

Zhao, X., Yang, H., Yang, Z., Chen, B., Qin, Y., 2010. Applying the input-output method to account for water footprint and virtual water trade in the haihe river basin in China. Environ. Sci. Technol. 44, 9150e9156.