An inquiry into water transfer network of the Yangtze River Economic Belt in China

Journal of Cleaner Production 176 (2018) 288e297

Contents lists available at ScienceDirect

Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro

An inquiry into water transfer network of the Yangtze River Economic Belt in China Feifei Tan a, b, Jun Bi a, * a b

State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing 210046, PR China Jiangsu Industry Development Research Institute, Nanjing University of Finance & Economics, Jiangsu, Nanjing 210023, PR China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 September 2017 Received in revised form 28 November 2017 Accepted 14 December 2017 Available online 18 December 2017

As a typical region moving forward on green development demonstration belt, Yangtze River Economic Belt (YREB) is sensitive to profound conflict between coordination development and unbalanced water allocation. Uncovering water transfer network can help improving water resource coordination development from the perspective of water-trade nexus. An integrated framework by uniting the interregional input-output (IRIO) analysis and social network analysis (SNA) in this study can estimate not only the scale and structure of water trade pattern, but also the topological characteristic of water transfer network of YREB. Results show that the rank of production-based virtual water volume was roughly in the order of middle, lower and upper reaches except Jiangsu, while the rank of consumptionbased virtual water volume was in the order of lower, middle and upper reaches. The provincial virtual water external dependence in YREB was not well consistent with the water resource endowment. Water transfer network was constructed on basis of the inter-provincial virtual water transfers between every two provinces inside YREB. Furthermore, the water transfer network assessment results illustrated the network structural form and revealed the network property and characteristics. Inside YREB, Shanghai, Jiangsu, Zhejiang, Anhui and Jiangxi were predominant in the water transfer network (from degree centrality), while Jiangxi, Hubei, Hunan and Anhui acted as the important medium and bridge (from betweenness centrality). The AFAF sector (Agriculture, Forestry, Animal husbandry and Fishery) and Industry sector of Jiangxi, and the Industry sector of Hubei and Jiangsu were the critical exporters in the network. Jiangsu received large amounts of virtual water from many good hubs and also transferred much to many good authorities when acted as bagmen in the network. Some provinces received large virtual water flow but provided few to others, such as Shanghai, and the AFAF sectors of most provinces transferred much to others. Thus, the research results would help understand the regional responsibility transfer in the hidden network linkages of interprovincial and intersectoral virtual water flows. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Water transfer network Virtual water flows Inter-regional input-output Social network analysis Yangtze River Economic Belt (YREB)

1. Introduction In our epoch, as it continues along the path of rapid industrialization and urbanization, the sheer quantity of water resources consumed by the world economy is recognized as the overarching threat to planetary health (Oki and Kanae, 2006). The dual-stresses of ever-increasing freshwater scarcity and water demand will place a heavy burden on future development (Lambooy, 2011; Murray et al., 2012; Hubacek et al., 2009). In terms of China, the

* Corresponding author. School of the Environment, Nanjing University, Xianlin Avenue 163#, Nanjing 210023, PR China E-mail addresses: [email protected] (F. Tan), [email protected] (J. Bi). https://doi.org/10.1016/j.jclepro.2017.12.129 0959-6526/© 2017 Elsevier Ltd. All rights reserved.

interregional water-trade nexus shows a complex and multithreaded network pattern rather than the nearest neighborhood relationship (Dong et al., 2013; Chen et al., 2017). China’s water transfers expresses as a complex scene due to its continuous advances of the marketization process and water conservation projects in recent years (Cai, 2008). Specifically, due to the spatial peculiarity of water resources and the interregional trades of goods and services, their supply and demand distributions have the potential to evolve as a picture of complex interactive information covering interprovincial and intersectoral water flows (Burkhard et al., 2012). In this study, water transfer network is exactly manifested as a network form of the aggregated interactive information, and it is the interactive information that can well reflect the impact on water flows from the interregional trades.

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

The water policy relevant to the solutions to water scarcity is often regional or river basin issues (Zhao et al., 2010), which is characterized by a mismatch between water resources distributions, economic development and other primary factors of production and consumption (Zhang and Anadon, 2014). Each region bears some degree of responsibility to water conservation and sustainable utilization (Chapagain and Hoekstra, 2008; Zhang et al., 2017). Yangtze River Economic Belt (YREB), China's emerging economic belt relied on the gold channel of Yangtze River, is another important growth engine after Yangtze River Delta, Pearl River Delta and Bohai Rim (Zhang et al., 2015). As a typical region moving forward on green development demonstration belt, YREB is a valuable subject to assess the water transfer network due to the profound conflict between coordination development and unbalanced water allocation. However, the full picture of inter-provincial and inter-sectoral virtual water transfer of YREB is still missing, especially the detailed structure characteristics from the network perspective. Thereby, water transfer network analysis should be one of significant complementary tools to help address the issues of water responsibility transfer in YREB. Tracking inter-regional virtual water flows is the first necessary step for constructing water transfer network. A reliable approach should be virtual water accounting (Allan, 1993; Hoekstra and Chapagain, 2007a; Zhang and Anadon, 2014), which can not only reveal the water consumption scale but also reflect the virtual water flows embedded in the interregional trades (Jiang et al., 2015; Zhang et al., 2017). Currently, a growing body of literature has catalogued valuable results related to virtual water accounting, and various research frameworks have been used in previous studies (Feng et al., 2011; Zhang et al., 2017). From methods point of view, a strand of researches relied on the bottom-up (e.g., product tree) framework, and the effects are calculated through arithmetic representations of all products and services and their corresponding virtual water consumption (Antonelli et al., 2012; Vanham and Bidoglio, 2013). Its shortcomings are mainly expressed in neglecting the association and dependency among different economic sectors for each region and giving a difficult practice on the secondary and tertiary industries (Hoekstra and Chapagain, 2007b; Chen et al., 2017). Another strand of researches concentrated on the top-down framework, such as the input-output model (Munksgaard et al., 2005; Zhao et al., 2009), which is capable of tracking the virtual water concealed in sectors nexus and distinguishing production-based virtual water from consumptionbased virtual water (Feng et al., 2014). The related researches prevailingly focused on the sectoral level (Zhao and Chen, 2014; Zhang and Anadon, 2014; Zhuo et al., 2016), provincial and basin level (Feng et al., 2012; Zhang et al., 2012; Wang et al., 2013; Dong et al., 2013) and national level (Guan and Hubacek, 2007; Wang et al., 2014; Chen et al., 2017), as well as some researchers focused on a global level (Ercin and Hoekstra, 2014; Lutter et al., 2016). Existing studies have enlarged the research fields related to virtual water trades. Among them, most results realized that waterdeficient regions were always driven by the consumption from water-abundant regions in China, leading to a more unsustainable direction (Jiang et al., 2015). The outline of YREB’s development program requires this region must develop the comparative advantages of the upper, middle and lower reaches to achieve a coordinated development belt, although it has obvious interprovincial development differences from the aspects of resources, environment, transportation and industry basis (Chen et al., 2017). Reshaping the regional water-trade nexus from the perspective of water responsibility transfer is helpful for improving water resource coordination development. To understand water responsibility transfer, however, it should be more informative to offer a scientific inquiry into the whole water transfer network of

289

YREB than to only investigate the interregional virtual water flows (Wichelns, 2010). In the light of water transfer network construction, the gravity model and vector auto-regression model often contribute to determine whether there is a link between any two locations or sectors (nodes), like in the existing cases of economic development network and energy network analysis (Li et al., 2014; Liu et al., 2015), as well as the ecological network analysis (Zhang et al., 2014; Chen and Chen, 2016). But it is with regret that both the estimation methods, which usually acts as the first necessary step of network construction in prior studies, have weak objectivity relative to the input-output analysis that can present more direct quantitative relations. Given these considerations, understanding the complex water transfer network of YREB is vital for reshaping and reallocating the water resource allowance across different areas. That should be helpful for environmental regulations on trans-regional water resources management system, providing the scientific evidence of optimal distribution of limited water resources. It is also of paramount importance to conduct a complete research framework through uniting the inter-regional input-output (IRIO) analysis and social network analysis (SNA), which can estimate not only the scale and structure of water trade pattern, but also the topological characteristic of water transfer network of YREB. Summarizing the extant literature, the contributions of this study are reflected in two ways. Firstly, it focuses on estimating the water trade pattern through understanding the provincial water trade with the area both inside and outside YREB, which seems to be more informative for water management. Secondly, compared with the interregional virtual water accounting in the prior studies, the topological characteristic of water transfer network structure inside YREB were assessed, which can help identify the relative importance of each province (sector) in the whole interprovincial (inter-industrial) water network and the interactions between different provinces (sectors). The coupled research framework makes the simulations of the whole process from water-trade intervention to water transfer network analysis. Overall, the research outcomes can help giving the practical policy implications on responding YREB's water crisis by considering water-trade nexus, as well as providing a theory framework for identifying water transfer responsibility. The rest of the article is structured as follows. In the following sections, after a short introduction of study area and data source, the research methodology is described. This is followed by detailed explanations of the specific methods and techniques. A detailed case study of YREB is presented in the next section. Finally, the fourth section summarizes the main findings and conclusions from this research. 2. Study area and data description 2.1. Study area The Yangtze River Economic Belt has developed to be one of regions who harbor the greatest comprehensive strength and the biggest strategic support in China since the reform and opening-up. YREB has a surface area of 2.05 million square kilometers, covers 11 provincial areas including Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, Hunan, Chongqing, Sichuan, Yunnan and Guizhou from east to west, and contributes over 40% of both the population and GDP of China. As the related guidance suggests, it is well known that the YREB should be built into a “green ecological corridor” as a demonstration zone for achieving ecological civilization in China. Although the whole economic belt moves forward into the common goal, there are obvious interprovincial development differences from the aspects of resources, environment, transportation and industry basis. In particular, spatial arrangements are the

290

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

carrier of implementing the functional orientation and the other plans of YREB, so the structure and pattern of water transfer network must be estimated and explored. 2.2. Data description and treatment The foremost data resources of this study are obtained from the “China 2010 inter-regional input-output table for 30 provinces” (Liu et al., 2014) for its public availability. It covers 30 provincial areas and 6 sectors. The six sectors include “AFAF”, “Industry”, “Construction”, “TS”, “WR” and “OS”, respectively (Table 1). To meet the demand in this study, a new inter-regional input-output (IRIO) table for YREB in 2010 is established by adjusting the original table when considering the rest 19 provinces in China as a whole. But even then, the difference with the original one is only that there are 12 regional units, including 11 provinces and the rest 19 provinces of China (abbreviating to R19). In accordance with the IRIO table of YREB, we need obtain the detailed inter-regional monetary value transfers in the year of 2010. The China’s IRIO table of 2007 has been applied in the previous studies (Dong et al., 2014; Feng et al., 2014), so the table of YREB in 2010 should be enough to practicality and honesty. Due to the roughly similar water-use intensities for each sector’s different compositions, the IRIO table of YREB in 2010 should be helpful for understanding the water transfer network to a large extent. Moreover, the water withdrawal data for more segmented sectors, such as sectors, is a challenge, and the evaluation results of water-use intensities will lead to even larger deviations. Another indispensable part of data sources includes the added values and water withdrawal of the six sectors, which are used for calculating the water use intensities of various sectors for each province. To ensure the consistency of the classification of sectors (Table 1) between economic and water withdrawal data, some adjustments should be made for some sectors. For instance, the added value of Mail business has been removed from the “Transportation, Storage and Mail business” sector to the OS sector on basis of the original classification and statistics in China Statistical Yearbook of 2011 (recording the related data of 2010), and the remainder parts constitute the TS sector. In the case of Construction, the added value of Construction sector is isolated from the whole secondary industry. As for the water withdrawal data, it is important to first clarify that this research mainly focuses on direct freshwater withdrawal, i.e. blue water, a better indicator on available water resource and ecosystems (Dong et al., 2014). To be specific, the accessible way for us is to obtain the water withdrawal data of “AFAF” for each province from China Statistical Yearbook of 2011, however, some pretreatments must be taken to obtain the data of other sectors. Firstly, the secondary industry includes Industry and Construction sectors, but it is difficult to directly obtain the water withdrawal of the two separate sectors. Because the national water consumption of various sectors can be found in China Economic Census Yearbook, water-use intensities of Industry and Construction sectors of China can be calculated. In accordance with the related studies (Dong

Table 1 The classification of the sectors. Sectors (Serial numbers) Definition AFAF(1) Industry(2) Construction(3) TS(4) WR(5) OS(6)

Agriculture, Forestry, Animal husbandry and Fishery All industries except Construction Transportation and Storage Wholesale and Retail trade Other services

et al., 2014), the ratio of China’s water intensity of Industry sector to China’s water intensity of Construction sector is supposed as the equivalent value with the ratio of each province, which does not significantly affect the results. Secondly, the water withdrawal data of the tertiary industry, including TS, WR and OS sectors, also need some similar assumptions and treatments. It's important to note that, however, in order to get the total water withdrawal of the three sectors of the tertiary industry in each province, we must not only remove the water withdrawal of primary and secondary industries from the total water use in each province, but also the household water use in each province (obtained from China UrbaneRural Construction Statistical Bulletin). In general, the water withdrawal data of some sectors for each provincial unit in previous studies has been usually estimated by combing the available data and some reasonable assumptions (Dong et al., 2014; Zhang and Anadon, 2014; Zhang et al., 2017; Chen et al., 2017), so this study adopts this estimation method for the data of 2010. Although collecting some data of different sectors from various sources, the water withdrawal data of each sector must be maintained as consistent as possible. 3. Methodology 3.1. Water transfer accounting framework The accounting of interprovincial virtual water flows requires combining the water-use intensities of various sectors in each province and the inter-provincial monetary value flows, when the latter are calculated through applying an organizational interregional input-output (IRIO) table of YREB. As a matter of fact, there are various methods to obtain the inter-provincial monetary value flows, such as life cycle analysis, single-regional inputoutput analysis, multi-regional input-output analysis, and inter-regional input-output model. Their advantages and disadvantages were discussed in many literatures (Jiang et al., 2015; Zhang et al., 2017). Among these, IRIO analysis is more suitable to perform the interprovincial virtual water transfers of YREB in this study, and is conductive to the water transfer network construction. The corresponding equilibrium equation of the IRIO table of YREB is following:

Xi;r ¼

m X n X s¼1 j¼1

xrs ij þ

m X

Firs þ Eir ¼

s¼1

m X n X s¼1 j¼1

s ars ij Xj þ

m X

Firs þ Eir

s¼1

(1) where Xi;r represents the total output of sector i in province r; xrs ij represents the intermediate input of sector i in province r; Firs represents the final demand of province s from sector i in province r; Eir denotes the foreign exports of sector i in province r, and the direct input coefficient ðars ij Þ is defined to reflect that the direct input from the sector i in province r once require to improve one monetary unit output of the sector j in province s. In addition, the total studied provinces and sectors are m and n, respectively. As shown in Table 2, the basic form of IRIO table conformed to YREB is established in this study. In terms of this table, we can present the applicable equilibrium equation when considering “m ¼ n ¼ 12” (11 provinces and the last one is the rest 19 provinces of China (R19)) and “i ¼ j ¼ 6” in the empirical study of YREB.

Xir ¼

12 X 6 X s¼1 j¼1

s ars ij Xj þ

12 X s¼1

Firs þ Eir

(2)

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

291

Table 2 The basic form of IRIO table for YREB. Intermediate input P1

…

P11

Final demand P s Pi

Export

Total export

x1;12 1 …

F111 þ F112 þ ::: þ F11;12 …

E11

X11

…

…

x1;12 6 …

F611 þ F612 þ … þ F61;12 …

E61

X61

…

…

x11;12 1 …

F111;1 þ F111;2 þ … þ F111;12 …

E111

X111

…

…

x11;12 6

F611;1 þ F611;2 þ … þ F611;12

E611

X611

x12;12 6

F 12;1 þ F 12;2 þ … þ F 12;12

E12

X 12

R19

s

S1 … S6 Intermediate input P1 … P11

S1 … S6

S1

11 x11 11 …x16

…

1;11 x1;11 11 …x16

… S6

……… 11 x11 61 …x66

… …

……… 1;11 x1;11 61 …x66

… S1

……… 11;1 x11;1 11 …x16

… …

……… x11;11 …x11;11 11 16

… S6

……… 11 x11;1 61 …x66

… …

……… x11;11 …x11;11 61 66

R19

x12;1 …x12;1 1 6

x12;11 …x12;11 1 6

Import

Im11 …Im16 Av11 …Av16 X11 …X61

11 Im11 1 …Im6 11 Av11 …Av 1 6 X111 …X611

Added value Total input

… … …

Im

12

1 FIm

þ

2 FIm

þ…þ

12 FIm

Av12 X 12

Note: P and S are abbreviated of Province and Sector, respectively; the superscript/subscript represents a value from one province/sector to another province/sector (or itself).

Then it can be presented as the matrix forms:

X ¼ ðI  AÞ

1

½FjE

WC ¼ W r/s þ W s/s þ ImW s (3)

Among theses, X is the composite matrix of virtual water transfers, both A and F are composite matrixes, and I represents the identity matrix. To be specific, A is a (12*12) matrix of direct intermediate input coefficients with the elements of Ars , and the last column and row of A are the intermediate input coefficients of the rest 19 provinces of China (R19). Further, besides the last column and row, each sub-matrix of Ars is consisted of the (6*6) matrix with elements of ars ij when considering the classification of sectors (Table 1). Thus, on the whole, A is a (67*67) matrix from more fundamental level. Further, by integrating b f (the diagonal matrix of the vector of water-use intensity (WI)) with Eq. (3), the matrix of virtual water transfers can be presented as following:

W¼b f ðI  AÞ1 ½FjE

(4)

In the results of the matrix manipulation, the specific representation formulas of interprovincial virtual water transfers are involved, such as Eqs. (5)e(6). Firstly, we can obtain each province’s virtual water volume exported to China’s another province (including R19) by Eq. (5) and each province imported from China’s another province (including R19) by Eq. (6). Secondly, when applying the equations, the virtual water transfers of each sector can be simultaneously obtained.

W r/s ¼

12 X

6 X 6  X

WIir  Xijrs



(5)

r¼1;rss i¼1 j¼1

W s/r ¼

12 X

6 X 6  X

WIjs  Xjisr



(6)

s¼1;ssr i¼1 j¼1

where W r/s denotes that the water transfer from the rth provincial unit to the sth provincial unit, and W s/r denotes the water transfer in the opposite direct; WIir and WIjs denote the water-use intensity of the sector i in province r and of the sector j in province s, respectively. Thirdly, WC (consumption-based virtual water) of sth provincial unit should cover W r/s, W s/s and ImW s (foreign imported virtual water volume), which can be expressed as below:

(7)

Similarity, WP (production-based virtual water) of sth provincial unit, which cover W s/r , W s/s and ExW s (foreign exported virtual water volume) are following:

WP ¼ W s/r þ W s/s þ ExW s

(8)

Finally, the net amount of virtual water transfers from rth provincial unit to the sth provincial unit can be derived by Eq. (9), and it can also represent the water trade balance of province r. r/s Wnet ¼ W r/s  W s/r

(9)

All the above constitutes the basic accounting framework for evaluating virtual water transfers, which is the basis of water transfer network construction.

3.2. Water transfer network construction and analysis On basis of the detailed virtual water transfers, the water transfer network both at provincial and sectoral network can be formed through connecting all of the provinces or sectors by their pairwise virtual water flows, and can be further explored through some techniques in social network analysis. Social network analysis (SNA) is an important analysis method that can depict the form, feature and structure of a complete network (Scott, 2000; Li et al., 2014), which can be applied in this study when considering each province or sector as a network node. Its advantage especially shows on the expression of correlative relationship, such as, revealing the conformity and hierarchy of the overall network, explaining the compactness of network connection, identifying the dominance of different nodes, and so on. In this study, the topological characteristics of the whole water transfer network is presented by the techniques including network density and network centrality, such degree centrality, betweenness centrality and closeness centrality. Furthermore, estimating the control index (CI) and dependent index (DI) of all network nodes can help clarify the dominance of the provinces and sectors in the whole water transfer network (Wang et al., 2017). The pair of indices are primarily quantified by the ratio of pairwise integral flows between different components (nodes), indicating the control one component exerts over the others. Moreover, the hubs index (HI) and authorities index (AI), the coefficient associated with a link analysis algorithm, are helpful for rank the importance of the

292

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

provinces and sectors in water transfer network. To calculate the indices of hubs and authorities, each network node is initially assigned a rank, and the hub/authority scores of each node are updated by repeating iterations as a link analysis algorithm. All the operational programs and procedures are completed in the UCINET 6.0. How the four indices (CI, DI, HI, AI) are calculated in details, as well as the related technique methods, can be found in the previous literature (Chen and Chen, 2016; Wang et al., 2017) in spite of applying in the field of energy consumption or carbon emission reduction for most researches, as well as the researches associated with social network analysis (Fath, 2004; Schall, 2013). Compared with the previous relevant studies, this research combines the inter-regional virtual water transfers accounting and social network analysis to explore the whole water transfer network in a comprehensive, objective, and concise way, when using the available interprovincial inputeoutput tables and detailed datasets of sector-wise water withdrawals. 4. Results and discussion 4.1. Scale and structure of virtual water trade of YREB YREB’s provincial production and consumption structures differ tremendously due to the very uneven economic development levels and endowments of various inherent factors, leading to the water-trade nexus from active exchange of goods and services. Applying the virtual water transfers accounting framework of YREB, the detailed virtual water trade situations of YREB in 2010 are presented, which are treated as the first necessary step for estimating the water transfer network. 4.1.1. Production-based and consumption-based virtual water trade of YREB Fig. 1 shows the detailed comparison of virtual water trade from two different perspectives (production and consumption) in YREB at 2010. The production-based (denoted by P) virtual water represents the total volume of local water consumption required for the production activities of all local sectors for each province, and the consumption-based (denoted by C) virtual water represents the total volume of local water consumption required for producing

final goods and services that are used by all locals. From Fig. 1, the virtual water trade volume of each province can be disaggregated into the internal water consumption, foreign export/import and domestic export/import. It is noteworthy that the domestic exported/imported virtual water includes the exported/imported virtual water into/from the provinces inside YREB and outside YREB. On the whole, the virtual water scale of YREB has a significant disparity at the provincial level, which should be attributed to the separations of production and consumption, and large volume of trade of raw and processed commodities. From Fig. 1, Jiangsu is the province that had the largest virtual water trade in the YREB at 2010 no matter from which perspective, with 172.8 billion m3and 125.5 billion m3. Many researches focused on China’s water consumption also demonstrated that Jiangsu had the largest scale in whole nation (Dong et al., 2014; Jiang et al., 2015; Chen et al., 2017), despite in different study periods. It was generally considered that the large economic scale and the property of manufacturing industries of Jiangsu led to its high values of production-based and consumption-based virtual water volume. On one hand, from the production side, the virtual water trade of Anhui, Hunan, Hubei, and Jiangxi according to the sequence of values, all of which locate at the middle reaches of YREB, take the second place. Shanghai and Zhejiang’s virtual water trade volume follow them very closely, and so does Sichuan. In terms of the remaining areas, Chongqing, Yunnan and Guizhou, have relative low scale as the upper reaches of YREB. The rank of productionbased virtual water trade volume was roughly in the order of middle, upper and lower reaches except Jiangsu. On the other hand, from the consumption side, the rank of consumption-based virtual water trade volume roughly abides by the principle of a gradual decrease from the upper reaches to lower reaches. The results of the provincial water trade are also similar to the previous researches associated with China’s water footprint in 2007 (Dong et al., 2014; Jiang et al., 2015). The main reason for the relative large consumption-based virtual water volume of Jiangsu, Zhejiang and Shanghai is that they had the largest gross domestic products and they were the most developed regions with dense populations in YREB and even in China. Moreover, Fig. 1 also demonstrates the structural composition of virtual water trade for each province from two perspectives. The

Fig. 1. Comparison between production-based and consumption-based virtual water trade of all provinces in YREB, 2010. Note: P and C represent the production-based and consumption-based.

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

difference between every two pillars for each province in Fig. 1 manifests as its water trade balance. A negative trade balance value means that one province of YREB imports virtual water from China’s other provinces or abroad, while a positive value means that one province exports to them. In 2010, it is clear that some provinces of YREB are large net virtual water exporters, including Jiangsu, Anhui, Jiangxi, Hubei, Hunan and Sichuan. Their water trade balances were 47.4 billion m3, 24.1 billion m3, 19.7 billion m3, 10.5 billion m3, 11.7 billion m3 and 3.9 billion m3, respectively. Unlike other more-developed provinces, Jiangsu is a net exporter, exporting 27.4% of its total virtual water from the perspective of production in 2010. Shanghai and Zhejiang can be called as the main water importers, with the water trade balances of 5.7 billion m3and 9.4 billion m3, while other three provinces are slightly water balanced areas. 4.1.2. Domestic and foreign water trade of YREB A negative domestic or foreign water trade value suggests that one province imports the virtual water from domestic other areas or from abroad, while a positive one suggests that the province exports the virtual water. Fig. 2 gives a clear presentation on the comparison between domestic and foreign water trade of the different provinces in YREB. As shown in the left chart of Fig. 2, except the middle reaches (Anhui, Jiangxi, Hubei and Hunan), the other provinces had the positive values, suggesting that they were heavily dependent on importing domestic virtual water. In general, the east areas with developed secondary and tertiary industries have been usually imported the virtual water from the water scarcity northern areas (Feng et al., 2014), aggravating water scarcity in the inland regions. In terms of foreign water trade, all provinces of YREB are the net foreign exporters of virtual water despite in varying degrees, while Jiangsu is the largest foreign exporters. It also shows that the most of foreign export mainly occurs at the lower reaches. The eastern areas account for a large proportion in water transfer embodied in the international trade. In addition to the geographic location, many common features are important drivers of such huge foreign water export, e.g., they are all affluent areas with large populations and major manufacturing hubs. Further investigations show that, by dividing the total virtual water by the monetary values of exported/imported goods and services, the foreign exported virtual water per unit of exported value in the lower reaches (218.6 m3/104 CNY, 321.6 m3/104 CNY and 152.3 m3/104 CNY for Shanghai, Jiangsu and Zhejiang, respectively) is far less than the average level of YREB (634.3 m3/104 CNY). The significant causes of the huge distinctions include the uneven technology level and different industrial structure.

293

4.1.3. Virtual water external dependence and water resource endowment of YREB In general, the virtual water strategy which advocates importing water intensive products and exporting products with low water intensity is gradually accepted as one of the options for solving water crisis in severely water scarce regions (Zhao et al., 2010), as well as balancing economic development and environmental protection. Therefore, it is necessary to assess whether the provincial imported virtual water is consistent with the water resource endowment. According to the previous studies (Liu et al., 2011; Jiang et al., 2015), China’s provincial comprehensive index of water resource carrying capacity (also called water scarcity index), which can reflect the water resource endowment, was divided into three levels: water scarcity (a value over 1.0), water tense (a value between 0.8 and 1.0) and water abundant (a value under 0.8). China’s provincial comprehensive index of water resource carrying capacity is an aggregative indicator that can show the water shortages to support the sustainable development of society, the economy and the environment. Based on the comprehensive evaluation indicator system covering regional social, economic, ecological and water resources, the index is assessed through integrating the carrying pressure index and coordination index of water resource complex system, and the population pressure index and economic pressure index on water resource. The details of model construction and calculation can be found in the related literature. Meanwhile, the percentage of net imported virtual water compared to water withdrawals is used to estimate the virtual water external dependence. Although all the provinces of YREB do not locate at the range of “water scarcity”, as demonstrated in Fig. 3, they hold distinct characteristics in the virtual water trade. During the range of water tense (a value between 0.8 and 1.0), some provinces in YREB exported a large amount of virtual water to the other areas, such as Anhui, Jiangsu, Hunan, Hubei and Sichuan. Since they acted as the exporters when simultaneously facing some water scarcities, they can hardly support their own development when considering the expected future economic development and population growth. Conversely, Shanghai and Zhejiang make use of virtual water as an important supplements due to such high virtual water external dependence. Jiangxi, Guizhou and Yunan, divided as the water abundant provinces, had a negative net virtual water external dependence, implying that they could export a large amount virtual water to other areas by the agricultural products (Deng et al., 2016). Unlike other water abundant province, Chongqing is an net importer, importing 8.22% of its water withdrawals.

Fig. 2. Domestic and foreign water trade of different provinces of YREB, 2010.

294

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

Fig. 3. The virtual water external dependence under different water scarcity indices.

4.2. Comprehensive assessment of water transfer network inside YREB It is noteworthy that the domestic exported/imported virtual water for each province includes not only into/from the provinces inside YREB, but also the provinces outside YREB. However, to estimate the water transfer network inside YREB, we have extracted the virtual water flows between every two provinces only inside YREB. Please find the detailed virtual water transfer between every two provinces in Table S1 of the Supporting information, which also gives further information associated with that. In order to illustrate the structural form of water transfer network inside YREB, the province-wise and sector-wise graphs in 2010 are plotted by the visualization tools of UCINET 6.0, as shown in Fig. S2 of the Supporting information. Similar to the researches associated with interprovincial carbon flows (Feng et al., 2013; Meng et al., 2013), the more-developed areas were found net receivers of water flows and addressed part of their water demand from less-developed areas. Moreover, further investigations on water transfer network of YREB have been finished in virtue of some technique methods of social network analysis.

4.2.1. Topological characteristics of the water transfer networks On basis of the constructed networks aimed at spatial and sectoral water transfer，the indicators of network density and network centrality can help depict the overall network property and the nodes' network characteristics in the social network analysis framework. The indicators of network centrality generally include degree centrality, betweenness centrality and closeness centrality. The province-wise network density was 0.4909 and the sector-wise network density was 0.4855 in 2010, revealing that it still remains a difficult problem of how to improve the spatial correlation degree of water trade inside YREB, especially for the upper reaches, and how to make spatial optimum allocation by reducing some redundant lines and increasing some necessary lines in the water transfer network. From Fig. 4, the average degree centrality value of all provinces was 49.47, while the provinces that had higher values than the average level include Shanghai, Jiangsu, Zhejiang, Anhui and Jiangxi from the order of high to low. It tells that they had more connections with other areas inside YREB. Conversely, Chongqing, Yunnan and Guizhou had smallest connections with others, with the degree centrality value of 18.05,16.96 and 12.31. Their remote geographic locations and small economic scales led to the weak water correlations with others. In terms of betweenness centrality, the provinces that had significant high values include Jiangxi, Hubei, Hunan and Anhui, revealing that they acted as the important medium and bridge in the water transfer network. Their relative abundant water resources and the specific locations (the middle reaches) brought about the strong water transfers and the key roles in the water transfer network. Meanwhile, Shanghai and Zhejiang had the highest closeness centrality values, demonstrating that they had strong attraction ability to virtual water from other provinces and had much direct correlations with others in the network.

4.2.2. Dominance of provinces and sectors in the water transfer networks A preliminary judgment on the dominance of provinces and sectors in the water transfer network can be found in Fig. S2 of Supporting information, but it is insufficient only by the initial qualitative analysis. Estimating the control index (CI) and dependence index (DI) can help identify and quantify the dominance of provinces and sectors in the water transfer network of YREB, and

Fig. 4. Network centrality characteristics of nodes (provinces) in water transfer network, 2010. SH, JS, ZJ, AH, JX, HB, HN, CQ, SC, YN and GZ are abbreviated for Shanghai, Jiangsu, Zhejiang, Anhui, Jiangxi, Hubei, Hunan, Chongqing, Sichuan, Yunnan and Guizhou, respectively. 1,2,3,4,5,6 denotes the six sectors.

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

Fig. 5. Graph of control index (CI) and dependence index (DI) of provinces and sectors inside YREB, 2010. The instruction of abbreviations can be found in Fig. 6.

can further explore the magnitudes and directions of virtual water flows across different provinces/sectors. As shown in Fig. 5, Shanghai had the highest dependence index (DI), with the value of 36.93%. Shanghai and Zhejiang attracted the products from the other provinces as well as their embedded water volumes. Anhui and Jiangxi had relative high values of CI, and they are the only two provinces with the values larger than 20%. The regional virtual water trade pattern of control and dependence in the water transfer network have been tied to the economic outputs and structure of YREB. From the sectoral perspective, Fig. 5 shows that the CI was more widely dispersed than the DI, ranging from 0 to 4.06%, implying that the virtual water trade was controlled by part of sectors with very large CI, compared to the evenly distributed DI. Several sectors that had high CI values exerted a great impact on the other sectors of all provinces. HB-2, JX-2, JX-1, JS-2 and SH-2 were the five largest controlling sectors, which were either high water-intensive sectors, and each sector controlled >5.5% of the total virtual water trade. The sector 3 (Construction sector) in all provinces had the DI larger than the average value and had the CI smaller than the average value, indicating that the consumed water of Construction sector highly dependent on the rest of the economy. Because the source of the sector’s materials and target markets were located outside itself. As long as some appropriate controls and regulations are taken in the dominant provinces and sectors, the pattern and structure of virtual water transfers of YREB have been optimized. 4.2.3. Roles of the provinces and sectors in the water transfer networks The hub index (HI) and authority index (AI) are the important indicators to facilitate the rank the importance of different nodes when each province or sector is regarded as a node in constructing water transfer network. A node's hub or authority score is higher if it has more neighbors, and if its neighbors have high scores (Wang et al., 2017). Fig. 6 shows the distribution for all nodes. For instance, both indices of Jiangsu were over the average level when receiving large amounts of virtual water from many good hubs and also transferred much to many good authorities when acted as bagmen in the network. Shanghai was located in the low HI and high AI quadrant when receiving large virtual water flow but provided few to others. As expected, Anhui and Jiangxi was located in the low AI and high HI quadrant when acted as the main exporters. From the sectoral perspective, the overall picture is like a parabola, showing that most sectors of the provinces were located

295

Fig. 6. Graph of Hub index (HI) and authority index (AI) of provinces and sectors inside YREB.

in the high HI and low AI quadrant, or the high AI and low HI quadrant. The integration of provincial and sectoral distinctions demonstrates that different sectors in the same province (or the same sectors of different provinces) attracted or provided virtual water flows in varying degree. For example, even if Shanghai had very low HI value, its AFAF and industry sectors had the HI value that was close to the average level, suggesting that more refined investigations are required when spread to sectors. Meanwhile, the same sector in different provinces always had the approximate values of AI or HI, which should be brought by the similar industrial features and structure. The control and regulation of some targeted sectors might be helpful for reshaping water-trade nexus of YREB to promote the development of sectoral collaboration system.

5. Conclusions 5.1. Main achievements As a typical region moving forward on green development demonstration belt, Yangtze River Economic Belt (YREB) is a valuable subject to estimate the water transfer network due to the profound conflict between coordination development and unbalanced water allocation. Uncovering water transfer network can help improving water resource coordination development from the perspective of water-trade nexus. This study investigated the virtual water trade and water transfer network through uniting the inter-regional input-output (IRIO) analysis and social network analysis (SNA) on basis of the inter-regional monetary trade and water withdrawal of YREB. It estimated the scale and structure of water trade pattern, and the topological characteristic of water transfer network for the Yangtze River Economic Belt. The research framework can provide the practical policy implications for the study area, as well as a theory framework for clarifying water transfer responsibility. Our research findings indicate that the rank of production-based virtual water trade volume was roughly in the order of middle, upper and lower reaches except Jiangsu, while the rank of consumption-based virtual water volume was in the order of lower, middle and upper reaches. Specifically, Jiangsu is the province that had the largest virtual water trade in YREB both from the perspective of production and consumption. At the same time, some provinces, such as Jiangsu, Anhui, Jiangxi, Hubei and Hunan,

296

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

exported net virtual water to other areas in the domestic and foreign trade, while Shanghai and Zhejiang acted as the main virtual water importers. On the whole, more-developed areas, except Jiangsu, tended to import net virtual water, mainly by the path of domestic import. Most of the foreign virtual water exports occur at the lower reaches, especially from Jiangsu. The results of provincial water trade are also similar to the previous researches associated with (Dong et al., 2014; Jiang et al., 2015; Chen et al., 2017) despite in different study periods. Moreover, further investigation of comparing virtual water external dependence and water resource endowment suggest that the water tense provinces, such as Anhui, Jiangsu, Hubei, Hunan and Sichuan still exported a large amount of virtual water to the other areas. Such a trade pattern may further aggravate their water tense situation and impede their long term sustainable development. The provincial virtual water external dependence in YREB was not well consistent with the water resource endowment. The technique methods of network analysis, such as network density, network centrality, control and dependent index, and hub index and authority index, are applied. The results illustrated the network structural form inside YREB, revealed the overall network property and the nodes network characteristics, clarified the dominance of provinces and sectors in water transfer network, and identified the roles of provinces and sectors in water transfer network. More specifically, the network density of water correlation remains to be improved by taking spatial water optimum allocations. The centrality degree measurements tell that Shanghai, Jiangsu, Zhejiang, Anhui and Jiangxi had more connections with other areas inside YREB, and the middle reaches has often acted as medium and bridge, and Shanghai and Zhejiang had strong attraction ability to virtual water in the network. At the same time, the regional pattern of control and dependence level in the water transfer network have been tied to the economic outputs and water-trade structure of different provinces in YREB. Shanghai and Anhui had the highest dependence index and control index, respectively. From the sectoral perspective, virtual water trade inside YREB was controlled by part of sectors with very large CI, compared to the evenly distributed DI. HB-2, JX-2, JX-1, JS-2 and SH2 were the five largest controlling sectors, and Construction sector of all provinces were highly dependent on the rest of the economy. 5.2. Policy implications and limitations Yangtze River Economic Belt (YREB) is experiencing significant interprovincial development differences from the aspects of resources, environment, transportation and industry basis, as well as uneven virtual water trade. To achieve the water resource coordination development of YREB, it is very challenging to reshape water-trade nexus due to increasing concerns on water scarcity and water demand. An integrated effort should be made on basis of some policy implications on responding water crisis of YREB. Firstly, from production perspective, in consideration of provincial water resource endowments in YREB, it is necessary to relocate the water-intensive sectors into the areas where has relative rich water resource (such as Yunnan, Guizhou and Jiangxi), especially relocating the agricultural sectors (Zhao and Chen, 2014; Deng et al., 2016). For instance, the provinces with relative abundant water resources and relative low virtual water external dependence (Jiangxi, Yunnan and Guizhou) can expand their agricultural production as well as export virtual water to other water tense or scarcity provinces, which can promote local economic development while making full use of their local water resources. At the same time, the more-developed provinces should facilitate the application of water saving technologies and equipment through technology transfer efforts. From consumption

perspective, the lower reaches of YREB, are the most developed regions with dense populations in YREB and even in China, had the highest virtual water consumption but during the range of water tense. For these areas, it is critical to upgrade industrial structure and improve water reuse/recycling, such as limiting the development of water consuming industries, and developing higher-valueadded products. In this regard, industrial development and trade policies should be based upon local water resource development, such as the development of eco-industrial parks (Geng and Zhao, 2009). Therefore, it is appropriate for governments at both regional and local levels, by establishing special trans-provincial agencies or mechanism, to work together so that more rational trade structure among different provinces can be established. Secondly, inside YREB, the virtual water transfers present a trend that the middle reaches provinces including water tense and water abundant region with less-developed economy exported the virtual water to the more-developed lower reaches provinces. However, although water shortage in the middle reaches are not as serious as in the northern provinces of China, the increasing demand from the other areas for water-intensive goods produced in these provinces may cause considerable impacts on local water resource and ecosystems. Thus, environmental policies should aim not only to reduce water consumption in water scarce regions but also to prevent water rich or tense areas becoming water scarce by maintaining a sustainable level. Moreover, the upper reaches provinces presented the relative slight water-trade nexus with the other areas inside YRED due to the relative low values of degree centrality, control or dependence index and hub and authority index in the water transfer network. Thereby, more targeted and focused regulatory measures on the upper reaches should be implemented. For example, some marginal sectors of lower reaches, which may always present the comparative advantages for the upper reaches, should be transferred to the upper reaches for developing the industrial connected effect. The government should pay attention on the allocation optimizations of water resources from the perspective of inter-provincial trade and industrial structure adjustment. No matter from provincial or sectoral side, the water trade ties between the upper reaches and other two reaches should be expanded and strengthened in order to develop the interregional comparative advantages in future. Developing water market, as well as establishing interprovincial virtual water compensation scheme, are useful platform to respond water crisis when considering management, legal, administrative, cultural and fiscal barriers. Thirdly, for the provinces with the highest values of dependence index (such as Shanghai and Zhejiang), we can reduce the interprovincial imports of food, which is produced by sectors with large water footprints and provide more food locally. At the same time, the AFAF sector (Agriculture, Forestry, Animal husbandry and Fishery) and Industry sector of Jiangxi, and the Industry sector of Hubei and Jiangsu were the critical exporters in the water transfer network of YREB. Therefore, the correlation between the critical importers and the water abundant areas should be strengthened by optimizing the spatial water trade-trade pattern and giving full play to the function of important medium and bridge. In addition to the dominant sectors in the water transfer network of YREB, we should also focus on the sectors where more ‘hidden water’ is consumed, such as Construction sector. The water consumption of Construction sector was highly dependent on the rest of economy. Considering that the construction sector was actually the pillar of China’s economy in recent years (Cai et al., 2015), there is an urgent need to advocate more efficient water use and higher water recycling rate of intermediate products for this sector. Overall, it is expected to develop the water-trade pattern that has the complementary advantages and synergic interactive relationship among the lower,

F. Tan, J. Bi / Journal of Cleaner Production 176 (2018) 288e297

middle and upper reaches, leading to narrow the development gap and promote balanced development in YREB. Only maintaining a well-structural and high-efficient water transfer network can help achieve the coordinated allocation and sustainable use of water resource. YREB’s water transfer network analysis can tell the water responsibility transfer and allocation issues of provinces and sectors in the water-trade nexus, and help policy makers to identify and rethink prior areas of local water problems. However, we still have some limitations of this research including the absence of historical trends, frequently updated input-output table, and the simple evaluation of water-use intensity. Further studies will address these limitations, as well as the more refined information, such as the comparative advantages involving resources, environment and development in YREB. Acknowledgments We are grateful for support from the fund projects: National Natural Science Foundation Programs in China (No.71603111), State Key Program of National Natural Science Foundation in China (No.71433007), General program of the Natural Science Research in Colleges and Universities of Jiangsu province in China (16KJB610007), and China Postdoctoral Science Foudation (2017M620207) . Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.jclepro.2017.12.129. References Allan, J.A., 1993. Fortunately there are substitutes for water otherwise our hydropolitical futures would be impossible. In: ODA, Priorities for Water Resources Allocation and Management. ODA, London. Antonelli, M., Roson, R., Sartori, M., 2012. Systemic input-output computation of green and blue virtual water flows with an illustration for the Mediterranean region. Water Resour. Manag. 26, 4133e4146. Burkhard, B., Kroll, F., Nedkov, S., Müller, F., 2012. Mapping ecosystem service supply, demand and budgets. Ecol. Indicat. 21, 17e29. Cai, X., 2008. Water stress, water transfer and social equity in Northern Chinadimplications for policy reforms. J. Environ. Manag. 87, 14e25. Cai, W., Wan, L., Jiang, Y., Wang, C., Lin, L., 2015. Short-lived buildings in China: impacts on water, energy, and carbon emissions. Environ. Sci. Technol. 49, 13921e13928. Chapagain, A.K., Hoekstra, A.Y., 2008. The global component of freshwater demand and supply: an assessment of virtual water flows between nations as a result of trade in agricultural and industrial products. Water Int. 33, 19e32. Chen, S., Chen, B., 2016. Tracking inter-regional carbon flows: a hybrid network model. Environ. Sci. Technol. 50, 4731e4741. Chen, W., Wu, S., Lei, Y., Li, S., 2017. China's water footprint by province, and interprovincial transfer of virtual water. Ecol. Indicat. 74, 321e333. Deng, G., Ma, Y., Li, X., 2016. Regional water footprint evaluation and trend analysis of Chinadbased on interregional inputeoutput model. J. Clean. Prod. 112, 4674e4682. 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 water flows. Sci. Total Environ. 500e501, 120e130. Dong, H., Geng, Y., Sarkis, J., Fujita, T., Okadera, T., Xue, B., 2013. Regional water footprint evaluation in China: a case of Liaoning. Sci. Total Environ. 442, 215e224. Ercin, A.E., Hoekstra, A.Y., 2014. Water footprint scenarios for 2050: a global analysis. Environ. Int. 64, 71e82. Fath, B.D., 2004. Distributed control in ecological networks. Ecol. Model. 179, 235e246. Feng, K., Chapagain, A., Suh, S., Pfister, S., Hubacek, K., 2011. Comparison of bottomup and top-down approaches to calculating the water footprints of nations. Econ. Syst. Res. 23, 371e385. 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. Unit.States Am. 110, 11654e11659. Feng, K., Hubacek, K., Pfister, S., Yu, Y., Sun, L., 2014. Virtual scarce water in China. Environ. Sci. Technol. 48, 7704e7713.

297

Feng, K., Siu, Y.L., Guan, D., Hubacek, K., 2012. Assessing regional virtual water flows and water footprints in the Yellow River Basin, China: a consumption based approach. Appl. Geogr. 32, 691e701. Geng, Y., Zhao, H.X., 2009. Industrial park management in the Chinese environment. J. Clean. Prod. 17, 1289e1294. Guan, D., Hubacek, K., 2007. Assessment of regional trade and virtual water flows in China. Ecol. Econ. 61, 159e170. Hoekstra, A.Y., Chapagain, A.K., 2007a. Water footprints of nations: water use by people as a function of their consumption pattern. Water Resour. Manag. 21, 35e48. Hoekstra, A.Y., Chapagain, A.K., 2007b. The water footprints of Morocco and The Netherlands: global water use as a result of domestic consumption of agricultural commodities. Ecol. Econ. 64, 143e151. Hubacek, K., Guan, D., Barrett, J., Wiedmannc, J., 2009. Environmental implications of urbanization and lifestyle change in China: ecological and Water Footprints. J. Clean. Prod. 17, 1241e1248. Jiang, Y., Cai, W., Du, P., Pan, W., Wang, C., 2015. Virtual water in interprovincial trade with implications for China's water policy. J. Clean. Prod. 87, 655e665. Lambooy, T., 2011. Corporate social responsibility: sustainable water use. J. Clean. Prod. 19, 852e866. Li, J., Chen, S., Wan, G., Fu, C., 2014. Study on the spatial correlation and explanation of regional economic growth in Chinadbased on analysis network process. Econ. Res. J. 11, 4e16 (in Chinese). Liu, H., Liu, C., Sun, Y., 2015. Spatial correlation network structure of energy consumption and its effect in China. China Ind. Econ. 5, 83e95 (in Chinese). Liu, J.J., Dong, S.C., Li, Z.H., 2011. Comprehensive evaluation of China's water resources carrying capacity. J. Nat. Resour. 26, 258e269 (in Chinese). Liu, W., Tang, Z., Chen, J., Yang, B., 2014. Input-output Table of China's 30 Provinces and Cities in 2010. China Statistics Press (in Chinese). Lutter, S., Pfister, S., Giljum, S., Mutal, C., 2016. Spatially explicit assessment of water embodied in European trade: a product-level multi-regional input-output analysis. Global Environ. Change 38, 171e182. Meng, B., Xue, J., Feng, K., Guan, D., Fu, X., 2013. China's inter-regional spillover of carbon emissions and domestic supply chains. Energy Pol. 61, 1305e1321. Munksgaard, J., Wier, M., Lenzen, M., Dey, C., 2005. Using input-output analysis to measure the environmental pressure of consumption at different spatial levels. J. Ind. Ecol. 9, 169e185. Murray, S.J., Foster, P.N., Prentice, I.C., 2012. Future global water resources withrespect to climate change and water withdrawals as estimated by a dynamic global vegetation model. J. Hydrol. 448d449, 14e29. Oki, T., Kanae, S., 2006. Global hydrological cycles and world water resources. science 313, 1068e1072. Schall, D., 2013. Measuring contextual partner importance in scientific collaboration networks. J. Inf. 7, 730e736. Scott, J., 2000. Social network analysis: a handbook. Contemp. Sociol. 22, 128. Vanham, D., Bidoglio, G., 2013. A review on the indicator water footprint for theEU28. Ecol. Indicat. 26, 61e75. Wang, W., Gao, L., Liu, P., Hailu, A., 2014. Relationships between regional economic sectors and water use in a water-scarce area in China: a quantitative analysis. J. Hydrol. 515, 180e190. Wang, Z., Huang, K., Yang, S., Yu, Y., 2013. An inputeoutput approach to evaluate the water footprint and virtual water trade of Beijing, China. J. Clean. Prod. 42, 172e179. Wang, Z., Xiao, C., Niu, B., Deng, L., Liu, Y., 2017. Identify sectors' role on the embedded CO 2 transfer networks through China's regional trade. Ecol. Indicat. 80, 114e123. Wichelns, D., 2010. Virtual water: a helpful perspective, but not a sufficient policy criterion. Water Resour. Manag. 24, 2203e2219. Zhang, C., Anadon, L.D., 2014. A multi-regional inputeoutput analysis of domestic virtual water trade and provincial water footprint in China. Ecol. Econ. 100, 159e172. Zhang, C., Wang, C., Lv, Y.Q., Shen, T., 2015. Research on city system spatial structure of the Yangtze River economic belt: based on DMSP/OLS night time light data. Urban Dev. Stud. 22, 19e27 (in Chinese). Zhang, Y., Huang, K., Yu, Y., Yang, B., 2017. Mapping of water footprint research: a bibliometric analysis during 2006-2015. J. Clean. Prod. 149, 70e79. Zhang, Y., Zheng, H., Fath, B.D., Yang, Z., Liu, G., Su, M., 2014. Ecological network analysis of an urban metabolic system based on input-output tables: model development and case study for Beijing. Sci. Total Environ. 468e469, 642e653. Zhang, Z., Shi, M., Yang, H., 2012. Understanding Beijing's water Challenge: a decomposition analysis of changes in Beijing's water footprint between 1997and 2007. Environ. Sci. Technol. 46, 12373e12380. Zhao, C., Chen, B., 2014. Driving force analysis of the agricultural water footprint in China based on the LMDI method. Environ. Sci. Technol. 48, 12723e12731. Zhao, X., Chen, B., Yang, Z.F., 2009. National water footprint in an inputeoutput frameworkda case study of China 2002. Ecol. Model. 220, 245e253. Zhao, X., Yang, H., Yang, Z., Chen, B., Qin, Y., 2010. Applying the input-output method to account for water footprint andvirtual water trade in the Haihe River basin in China. Environ. Sci. Technol. 44, 9150e9156. Zhuo, L., Mekonnen, M.M., Hoekstra, A.Y., 2016. The effect of inter-annual variability of consumption, production, trade and climate on crop-related green and blue water footprints and inter-regional virtual water trade: a study for China (19782008). Water Res. 94, 73e85.