Interregional trade among regions of urban energy metabolism: A case study between Beijing-Tianjin-Hebei and others in China

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Resources, Conservation & Recycling journal homepage: www.elsevier.com/locate/resconrec

Full length article

Interregional trade among regions of urban energy metabolism: A case study between Beijing-Tianjin-Hebei and others in China ⁎

Hongmei Zhenga, Xinjing Wanga, Mingjing Lia, Yan Zhanga, , Yachun Fanb a State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Xinjiekouwai Street No. 19, Beijing 100875, PR China b Colloge of Information Science and Technology, Beijing Normal University, Xinjiekouwai Street No. 19, Beijing 100875, PR China

A R T I C L E I N F O

A B S T R A C T

Keywords: Urban ecology Interregional trade Ecological network analysis Multi-regional input-output table Jing-Jin-Ji urban agglomeration

Interregional trade of energy exchanges from urban metabolism perspective fills the gap of regional disparity due to economic development, physical geography, and lifestyle. The integrated development of Beijing-TianjinHebei (Jing-Jin-Ji) urban agglomeration results in actors changes of the three provinces (or municipalities) between provider or receiver from 2002 to 2010, meanwhile, to treat them as Jing-Jin-Ji region, the energy transfer with other seven regions in China also changed. This research will analyze energy flows between JingJin-Ji agglomeration and other regions in China from two perspectives: first, treat Beijing-Tianjin-Hebei as a whole, and to trace the energy processes among Jing-Jin-Ji region and other seven regions; second, see Beijing, Tianjin, and Hebei as three and to analyze the energy exchanges among them three and also the exchanges with other 27 provinces. Each perspective includes the energy exchanges through direct, indirect, and integral processes from 2002, 2007, and 2010. The results show that the flow through multiple paths from the Northern region accounts for more than 25% of total input throughflow from other seven regions, and the Eastern region receives the highest flow and takes a proportion more than 31% of total output throughflow. The flow through multiple paths among Beijing, Tianjin, and Hebei has changed since 2007 due to the industrial restructuring in 2007. After comparing the path flow through one path and multiple paths at two levels, the regional analysis shows the directions of energy flows between Jing-Jin-Ji and other regions, furthermore, the provincial level reflects the details within the region.

1. Introduction The unbalanced economic development in China results in frequent trade among regions or provinces in China (Zhang et al., 2015b). Meanwhile, as the industrialization and urbanization are increasing rapidly, urban agglomeration has become one kind of core area for regional economic development (http://city.ce.cn/news/201510/13/ t20151013_2940863.shtml). The total area of Beijing-Tianjin-Hebei (Jing-Jin-Ji) region accounted for 2.30% in China, and the proportion of population was 7.23%, but its GDP took an account of 10.2% in 2015. Furthermore, the capital transferred through trade among the three provinces (or municipalities) increased from 2002 to 2010, taking an average growth of 15% each year. However, the capital output from Beijing to Tianjin and Hebei decreased, the decreasing proportion were 4% and 6% for each year, separately. This increase or decrease trend will cause the quantity of energy consumption or carbon emission which embodied in transferring commodities also changed. Under this circumstance, it is a hot topic that how the trade within Jing-Jin-Ji

⁎

region influence the corresponding energy flows. Otherwise, Jing-Jin-Ji region is an area with limited energy resources that cannot meet its own requirement, to be specific, Beijing needs to input physical capital from other provinces outside Jing-Jin-Ji region, Tianjin is a port in the Northern area and the input and output with other regions in China, or import and export with areas outside China are its center for economy, and Hebei mainly runs goods trade with others. Therefore, not only the interregional trade within Jing-Jin-Ji regions, but these three provinces with others in China are contributed to fill the gap of regional disparity. In summary, the two aims for this research are, firstly figuring out how other regions in China support the development of Jing-Jin-Ji region; secondly, illustrating the interactions among Beijing, Tianjin, Hebei, and also with other 27 provinces in China when exchanging resources. Urban energy metabolism includes processes for exploiting, transforming, and consuming energy, as well as processes for recycling byproducts and wastes (Zhang et al., 2010). Current analysis of urban energy metabolism typically focuses on a single city or several cities in one region. Then urban energy metabolism extended to the level of

Corresponding author. E-mail address: [email protected] (Y. Zhang).

http://dx.doi.org/10.1016/j.resconrec.2017.05.010 Received 15 February 2017; Received in revised form 28 April 2017; Accepted 24 May 2017 0921-3449/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Zheng, H., Resources, Conservation & Recycling (2017), http://dx.doi.org/10.1016/j.resconrec.2017.05.010

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

structural and functional characteristics of sectors. Chen and Chen (2015) compared material flow analysis, input-output analysis and ecological network analysis when analyzing urban energy metabolic processes, and they stated that ecological network analysis can indicate the mechanism of sectors among their interaction. The combination of input-output table and ecological network analysis has already been used to study the ecological element, energy, and carbon footprint in single city, urban agglomeration, and country (Zhang et al., 2014a,b, 2015b, 2016b). Based on the traditional input-output analysis of Li et al. (2014), some research introduced ecological network analysis to account for energy-water nexus in urban systems, Chen and Chen (2016) took Beijing as an example, Wang and Chen (2016) studied Beijing-Tianjin-Hebei agglomeration. Although they used control analysis in ecological network analysis, they did not quantify the direct or indirect energy or water flows among sectors or regions. Zhang et al. (2014a) based on the resource flows between any two sectors (the input and output processes) and with external environment to establish an equilibrium equation. They converted the monetary input-output table into a physical table with resource flows among sectors in Beijing from 1997 to 2007. The calculation of direct, indirect, and embodied ecological elements helped to evaluate the resource distribution within the system. Then, Zhang et al. (2014b) discussed embodied energy consumption and carbon footprint of 28 sectors in Beijing from 2000 to 2010, and proposed energy reduction and carbon emission policies for future. This method has been extended to the levels of urban agglomeration (Zhang et al., 2016b) and provinces (Zhang et al., 2015b). However, current research are mainly on one level, such as city, region, or country, instead of combining two levels. This research chooses Jing-Jin-Ji urban agglomeration as case study, and conducts at two levels, firstly, treats Jing-Jin-Ji region as a whole, this region will be divided from the Northern region which is named by the administrative division, to analyze the energy flows between Jing-Jin-Ji region and other regions in China; secondly, sub-divides Jing-Jin-Ji region, to see Beijing, Tianjin, and Hebei as three separate city or province, traces energy flows between each of them with the other 27 provinces or cities in China. Both of the two levels will be discussed from path flow and node flow. At regional level, the Jing-JinJi is one node in the network model, and at provincial level, Beijing, Tianjin, and Hebei are three nodes. The supplementary analysis between regional and provincial levels not only indicates the energy flows between Jing-Jin-Ji urban agglomeration and other regions, but also presents the heterogeneity of the two cities and one province within the agglomeration. This paper will reflect the dependence of the urban agglomeration on other regions in China and the independence within the agglomeration comprehensively.

urban agglomeration (Zhang et al., 2016b; Baynes and Bai, 2012; Garcia-Montiel et al., 2014). Baynes and Bai (2012) analyzed Melbourne region in Australia from urban metabolism perspective, firstly they conducted calculation of primary energy, and then extended the results to a lone time span, and attribute upstream primary energy consumption to sectors based on the direct secondary energy use. Garcia-Montiel et al. (2014) illustrated how the socio-ecological processes in San Juan metropolitan area of Puerto Rico influence the energy flows within the region. Zhang et al. (2016b) chose Jing-Jin-Ji region as a case study, when tracing energy metabolic processes, they accounted for the energy flows among sectors within each city or province in Jing-Jin-Ji region, and also the flows among sectors in different provinces (or municipalities). In addition, their results explained the energy utilization structure of this sectors and cities as well. Embodied energy is the total energy consumed during the whole production processes (Lenzen, 1998), and it includes direct energy and indirect energy. Indirect energy is unphysical energy that embodied in the by-products, products, or service exchanges from upstream steps. Multi-regional input-output analysis has been used to account for embodied resources among different regions, it represents the economic flows among sectors not only within one region, but also between different regions (Zhang et al., 2013). This method quantifies exchanges among sectors from economic perspective, reflects the interactions between sectors when trading resources, and traces resource flows among regions caused by one region’s consumption activities (Wiedmann, 2009; Wiedmann et al., 2007). It was firstly used to analyze economic flows among sectors in Italy or America (Polenske, 1980). And now the multi-regional input-output analysis has been used in the analysis of materials footprint (Wiedmann et al., 2015), pollution emission related to air quality, such as black carbon (Li et al., 2016), atmospheric mercury (Liang et al., 2014), particulate matter (Yang et al., 2015), or carbon emission (Tian et al., 2014), and also energy consumption (Liang et al., 2007). In energy analysis, Liang et al. (2007) divided China into eight administrative regions based on the data in 1997, accounted for and compared the embodied energy consumption of all these regions from final consumption perspective, and also predicted their energy consumption in 2010 and 2020. Based on the same method, Cui et al. (2015) studied energy flows caused by trade between China and other counties using the data from Global Trade Analysis Project. Li et al. (2014) compared the embodied energy trade flows from the perspectives of production and consumption as the regional disparity in China. Then, some analysis combined system ecology and input-output analysis, they stated that besides direct energy in form of physical fuel consumed by countries or the sectors within the countries, the production processes need the indirect energy embodied in the exchanges among sectors. These helped establish an equilibrium describing economic or energy flows among sectors. Chen and Chen (2013) constructed a network model with 6384 nodes to account for the proportion of energy sources in global embodied energy consumption, the directions of energy flows, and finally chose five countries to discuss the contributions of all their sectors. Chen (2011) specified 29 sectors in each of the 30 provinces and cities in China, calculated the embodied energy consumption from final consumption aspect, also the input, output and net embodied energy of them. Zhang et al. (2013, 2016a) analyzed the embodied energy consumption of 30 sectors in Chinese 30 provinces or cities. However, when combining system ecology to conduct embodied energy flow accounting among regions or countries, they used the embodied energy coefficient from output region multiplying the quantity of input energy, but did not quantify the energy flows through multiple paths among sectors or regions. Flow analysis in ecological network analysis can assess the integral flows among nodes in the network model, and analyze the contribution of each node (Lu et al., 2014; Zhang et al., 2016b). Combining multiregional input-output table and ecological network analysis is effective to calculate the indirect energy embodied in the exchanges among sectors. Otherwise, ecological network analysis also fulfills the

2. Method To combine multi-regional input-output analysis with ecological network analysis can not only account for the energy consumption embodied in the exchanges of commodities through one path or multiple paths among sectors in one city, and the sum of all these indirect energy consumption is equal to the integral energy in this research (the integral energy consumption in this research plus the direct energy consumption for each sector that consumes the energy in the form of various forms of physical fuel is the embodied energy), in this study, two terms have been proposed, the path flow means the flow among sectors through one path or more than one path (multiple paths), and the node flow is the sum of the flows through one path and multiple paths, it equals to the total indirect energy flow into one sector; but also reflect the relationships or actors of different sectors in the same system (Zheng et al., 2017). These advantages could fully be used to achieve our goal that are to calculate the energy flows through one path and multiple paths among sectors, and to analyze the receiver or provider between Jing-Jin-Ji region and other regions in China or even within Jing-Jin-Ji region. To conduct our analysis, based on the monetary 2

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Combining multi-regional input-output tables with the energy consumption data of each sector in each province can help to calculate the associated direct energy consumption coefficient for each sector (Zhang et al., 2014b). Details of the equilibrium equations are provided in Section 3.1 of Zhang et al. (2014a). Based on the calculations in that paper, the following equation will be obtained:

Table 1 The division of provinces in China. Source: http://xzqh.org.cn/. Regional level

Provincial level

Regional level

Provincial level

Jing-Jin-Ji

Beijing Tianjin Hebei Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Shandong Guangdong Guangxi Hainan

Northern

Shanxi Inner Mongolia Jiangxi Henan Hubei Hunan Chongqing Sichuan Guizhou Yunnan Shaanxi Gansu Qinghai Ningxia Xinjiang

Northeastern

Eastern

Southern

Central

Southwestern

Northwestern

ˆ − H]−1 ε = E [U

(1)

Where ε represents the direct energy coefficient matrix, in which the element εki represents an energy consumption coefficient embodied in the products from sector i in region k to other sectors. E represents the energy consumption matrix that enters or flows out of the sectors in a ˆ = [uji]n×n , when i = j, uji = Xi, where Xi is the total ecogiven area. U nomic output of sector i; when i ≠ j, uji = 0. H represents the monetary value flow matrix that accounts for the intermediate exchanges among sectors on an economic basis. Using this equation, the direct energy consumption coefficients for in the 8 regions and 30 Chinese provinces can be calculated. We can then multiply the economic data (Xi) by the corresponding energy coefficient to determine the energy flow through one path among regions or provinces. In this way, the monetary input-output table containing data for the 8 regions and 30 provinces can be converted into a corresponding energy table. These flows represent the energy flows through a path of length 1 between pairs of regions or provinces (i.e., with no intermediates). Based on the flows (fij) through one path (F means the matrix including fij, F = (fij)) and the flow into node i from the environment (zi), we can define Ti (throughflow) as the sum of the flows between nodes and of the imports into node i (Li et al., 2012).

multi-regional input-output tables from 2002 to 2010, we establish equilibrium equations to help convert the monetary input-output tables into physical input-output tables describing the flows of energy flows among sectors, based on these physical tables, the embodied energy consumption of each sectors could be assessed to evaluate the relationships of sectors when exchanging resources and the resources distribution in the system based on the flow analysis or utility analysis in ecological network analysis method. The data used in this research comes from the < China Energy Statistical Yearbook > in 2003, 2008, and 2010, and China Emission Accounts & Datasets (http://www.ceads.net/). The multi-regional input-output tables are in 2002 (Shi and Zhang, 2012), 2007 (Liu et al., 2012), and 2010 (Liu, 2014). To avoid the inflation in different years, the data from 2010 and 2007 were adjusted for inflation using a mean annual inflation rate of 4.133% and 5.645% compared with the data in 2002, so that the data in the three years represented constant 2002 prices. For simplicity, we will refer to all of the areas that make up the regions of China as “provinces”, even though some (e.g., Inner Mongolia) are autonomous regions and others (e.g., Beijing) are provincialscale municipalities. The division of provinces is shown in Table 1. As the energy data in energy balance tables and in China Emission Accounts & Datasets are all based on different kinds of energy, to standardize the data for different kinds of energy, we used standard coal-equivalent coefficients (Supplemental Table 1) to convert the different units for different energy sources into consistent units (i.e., the tonnes of coal equivalent; tce), then totaled these values for all energy sources used by a sector to provide the sector’s total direct energy consumption. To calculate the path flow and node flow need to combine the multiregional input-output table and ecological network analysis methods, that is the flows through one path among nodes, and the flows through multiple paths among nodes. The network models include node (circles in Fig. 1) and path (the line segment between any two nodes). The nodes are the regions or provinces, and the paths are the flows from region i to region j (fji), and the flows from region j to region i (fij). Fig. 1 shows the network models in two levels, Fig. 1a is the model at the regional level, and Fig. 1b represents provincial model. In addition, the international trade is not considered in the multi-regional input-output tables, as the international imports in these tables are all assumed to be non-competitive. The input and output in this research means the energy between Jing-Jin-Ji region and other regions in China, and the import and export are the energy between Jing-Jin-Ji region and areas outside China, and the external environment is the areas outside China, we can define zi and yi as the import flow from the external environment to node i and as the export flow from node i to the external environment, respectively.

n

Ti =

⎛ ⎞ f + zi ⎜ ∑ ij ⎟ ⎝ j=1 ⎠

(2)

We can then calculate the nondimensional energy flux matrix N' (n'ij) and the G' matrices for flows along pathways of each possible path length (l). First, the input-oriented flows from node j to node i (g'ij) are defined as follows (Li et al., 2012):

gij '= fij / Ti

(3)

From the direct flow intensity matrix G' = (g'ij), the dimensionless integral flow intensity matrix N' = (nij) can be calculated using the following power series (Fath and Patten, 1999):

N' = (n'ij ) = (G′)0 + (G′)1 + (G′) 2 + (G′)3 + ...+(G′)l + ...=(I−G′)−1 (4) where (G') is a self-feedback matrix that reflects flows that occur within each node. (G')1 is the energy flux matrix when the flow follows a single path between two nodes (i.e., l = 1). (G')l reflects flows of length l (l ≥ 2) between nodes, and I represents the identity matrix. For this analysis to be valid, the power series must converge. We confirmed that the series converged. N' indicates the nondimensional energy flux between nodes, and we can multiply N' by the diagonal of the energy input matrix (Ti) from the local environment, diag(T), to obtain the indirect energy transfer flow matrix (Y): 0

Y = diag(T)N'

(5)

Based on the integral flow matrix, flow matrix through one path, the flows among nodes through multiply paths can be calculated:

Indirect = Y-diag(T) −F

(6)

3. Results The analysis at two levels both conducts from path flow and node flow perspectives, path flow means the energy flows through one path and more than one path (multiple paths), and node flow means the sum 3

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 1. Network models at regional and provincial levels.

were different, the growth from 2002 to 2007 was because the input from other regions to Jing-Jin-Ji increased more significantly than the import flow, the input in 2007 was 1.99 times that in 2002; and the growth from import flows resulted in the increasing trend from 2007 to 2010, the import flow in 2010 was 32.09% more than that in 2007. For each region, in three time points, the Northern and Eastern regions input most quantity of flows to Jing-Jin-Ji region, and otherwise, the energy from Southern and Southwestern were the smallest. The black arrows in Fig. 2 show the flows from Jing-Jin-Ji to other regions, also we see the sum of output flows and export flows from Jing-Jin-Ji to the external environment as the total output throughflow. The total throughflow also increased from 2002 to 2010, and it was 2.24 and 1.12 times than each previous time point. Similar to the input flows, the increase from 2002 to 2007 was caused by the growth of output flows to other regions, the output in 2007 was 2.47 times than in 2002; from 2007 to 2010, the output and export flows both changed little, only increased 12% and 10% compared with that in 2007, respectively. In three time points, Jing-Jin-Ji output most energy through one path to

of the one path and multiple path for each region or province. 3.1. Regional flow analysis To analyze the urban energy metabolic processes at regional level, it is important to account for the energy flows between Jing-Jin-Ji urban agglomeration and other regions in China through one path and multiple paths, and also assess the import and export energy with the external environment. 3.1.1. Path flow Fig. 2 shows the energy flows between Jing-Jin-Ji and other regions through one path in 2002, 2007, and 2010. The blue arrows in Fig. 2 are the flows from other regions to Jing-Jin-Ji agglomeration (input to Jing-Jin-Ji), we treat the sum of input flows and import flows from external environment as the total input throughflow. This total input throughflow increased from 2002 to 2010, it was 1.89 and 1.13 times than each previous time point. But the reasons for these two increase 4

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 2. Energy flows between Jing-Jin-Ji and others through one path (all the bar chart has the same scale, and the one at the lower right corner is as a legend and it is the input and output data from the Eastern region).

5

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 3. Energy flows between Jing-Jin-Ji and others through multiple paths.

the Eastern region (accounted for more than 45% of the total output to other regions), and the energy to the Southwestern was the smallest. Fig. 3 indicates the energy flows between Jing-Jin-Ji and other

regions through multiple paths in 2002, 2007, and 2010. Also the blue arrows in Fig. 3 are the flows from other regions to Jing-Jin-Ji agglomeration (input to Jing-Jin-Ji). The total inflow increased from 2002 6

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 4. The node flow of each region.

different time points, the inflow of the Northern region increased from 2002 to 2010, the result in 2007 increased 8.27% compared to that in 2002, and that in 2010 further increased 3.33% than that in 2007. The growing speeds of the Southern, Southwestern, and Northwestern regions were higher than others, the inflow in the last two time points all increased more than 13% compared with that in their previous time points. The flow from the Northeastern, the Central, and the Eastern regions increased from 2002 to 2007, they were 1.60, 1.48, and 1.97 times in 2007 that of their flow in 2002, but then decreased from 2007 to 2010. The Northeastern and Eastern regions decreased the most, they were only 71.41% and 38.96% of that in 2007. The Central region did not changed a lot. The decrease of these three regions also caused the inflow decreased of the Jing-Jin-Ji region. Fig. 4b shows the output flow from Jing-Jin-Ji region to others. The total outflow increased from 2002 to 2010, but it grew faster from 2002 to 2007 than that from 2007 to 2010. The result in 2007 was 1.75 times than in 2002, but in 2010, it was only 1.02 times that in 2007. In the three time point, the outflow to the Eastern region was the highest and to the Northwestern region was the smallest. In 2002, the outflow to the Eastern region accounted for 37.84%, and it was 5.08 times than from the Northwestern region. In 2007, the proportion of the outflow to the Eastern region increased to 41.40%, and it was much higher than the regions ranking second (the Central region only accounted for 15.56%). In 2010, the proportion of outflow to the Eastern region decreased and accounted for 31.63%, but that of the Central region increased to 26.35%. The outflow to other regions were all less than 11%. Comparing the outflow of the same region in different time points, the outflow of the Eastern region increased from 2002 to 2007, and then

to 2007 and then decreased from 2007 to 2010. The total inflow in 2007 was 1.42 times than in 2002, but in 2010, it was only 89% of that in 2007. In the three time points, the inflow from the Northern region was highest, otherwise, the flow from the Southern region was the smallest. The black arrows in Fig. 3 show the flows from Jing-Jin-Ji to other regions through multiple paths. The total outflow increased from 2002 to 2010, and it was 1.73 times in 2007 than that in 2002. But it did not changed a lot from 2007 to 2010, only increased less than 2%. In three time points, Jing-Jin-Ji outflowed most energy to the Eastern region, and the energy to the Northwestern was the smallest.

3.1.2. Node flow Fig. 4 shows the node flows between Jing-Jin-Ji and other regions in China. Fig. 4a means the input flow to Jing-Jin-Ji region from others, it represented the integral flows including the energy flows through one path and multiple paths. The total inflow increased from 2002 to 2007 and then decreased. The result in 2007 was 1.43 times than in 2002, but in 2010, it was only 89.61% of that in 2007. In the three time point, the inflow from the Northern region was the highest and from the Southern region was the smallest. In 2002, the inflow from the Northern region accounted for 34.25%, and it was 7.57 times than from the Southern region. In 2007, the proportion of the inflow from the Northern region decreased, and it accounted for 25.86%, this meant that the inflow from other seven regions were more balanced. Compared with the results in 2007, in 2010, the proportion of inflow from the Central region decreased a lot, from 16.35% to 7.11%, but that of the Southwestern and Northwestern increased from 7.30% to 10.84%, and from 8.14% to 10.30%, respectively. Comparing the inflow of the same region in 7

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

by Hebei and Tianjin were still high, especially before 2007. This pointed out that the energy exchanges within the urban agglomeration were more frequent than with the provinces outside. Fig. 5b is the flows from other regions to Beijing through multiple paths. The total inflow increased from 2002 to 2010, it was 1.36 and 1.17 times than each previous time point. For each province, in three time points, Shanxi input most quantity of flows to Beijing in 2002, and then changed to Hebei. The energy from Hainan was the smallest. Fig. 5b also showed the flows from Beijing to other regions through multiple paths. This total outflow decreased from 2002 to 2010. The total outflow in 2007 was 63.52% of that in 2002, and in 2010 was 64.98% of that in 2007. This was because the industrial restructured from Beijing to Tianjin and Hebei since 2007, most industries with high energy consumption moved to other two areas. And this resulted in the decrease of product production in Beijing. For each province, in three time points, Beijing output most quantity of flows to Hebei in 2002, and in 2007, changed to Shandong, Jiangsu became the city received most energy through multiple paths in 2010. The energy from Hainan and Qinghai were the smallest. Fig. 6 showed the input and output energy flows through one path and multiple paths between Tianjin and other 29 provinces in China. Fig. 6a is the flows from other regions to Tianjin. This total throughflow increased from 2002 to 2010, it was 1.67 times in 2007 that in 2002. From 2007–2010, the total throughflow did not change a lot. The reason for these increase was the growth of input from other provinces, the input in 2007 was 2.23 times that in 2002, However, the import to Tianjin in 2007 decreased, and it was 96.50% of that in 2002. Although the total throughflow did not change from 2007 to 2010, the import increased 31.79% compared with that in 2007, and the inflow from other provinces decreased 8.40%. For each province, in three time points, Hebei input most quantity of flows to Tianjin, and the energy from Hainan was the smallest. Fig. 6a also showed the flows from Tianjin to other regions. This total throughflow increased from 2002 to 2010, but it grew more significant from 2002 to 2007 than from 2007 to 2010. The total throughflow was 1.80 times in 2007 that in 2002, then only increased 5.25% in 2010. The reason for the changes was both the increase of outflow to other provinces. It was 2.18 times in 2007 that in 2002, and increased 22.13% from 2007 to 2010. The export from Tianjin first increased and then decreased. The export grew 49.01% from 2002 to 2007, and in 2010, it was only 84.53% of that in 2007. For each province, in three time points, Tianjin output most quantity of flows to Hebei in 2002, and in 2007 and 2010, Jiangsu became the province received most energy through one path. The energy from Hainan and Qinghai were the smallest. Fig. 6b is the flows from other regions to Tianjin through multiple paths. The total inflow increased from 2002 to 2007, and then decreased, it was 1.82 times in 2007 that in 2002. But in 2010, it was only 76.68% of that in 2007. For each province, in three time points, Shanxi input most quantity of flows to Tianjin in 2002, and then changed to Hebei. The energy from Hainan was the smallest. Fig. 6b also showed the flows from Tianjin to other regions through multiple paths. This total outflow increased from 2002 to 2010. The total outflow in 2007 was 1.72 that in 2002, and in 2010 increased 1.08% compared that in 2007. For each province, in three time points, Tianjin output most quantity of flows to Hebei in 2002, and in 2007 and 2010, changed to Jiangsu. The energy from Hainan, Ningxia and Qinghai were the smallest. Fig. 7 showed the input and output energy flows through one path and multiple path between Hebei and other 29 provinces in China. Fig. 7a is the flows from other regions to Hebei. This total throughflow increased from 2002 to 2010, it was 2.01 times in 2007 that in 2002. From 2007–2010, the total throughflow did not change a lot. The reason for these increase was the growth of import, the import in 2007 was 3.78 times that in 2002, meanwhile, the inflow from other provinces in 2007 also increased 79.00%. The import in 2010 was 2.04 times that in 2007, however, the inflow from others decreased, and it

decreased. The result in 2007 was 1.93 times that in 2002, and that in 2010 only 78.05% of that in 2007. The outflow to the Northern, the Southwestern, and the Northwestern regions all showed the same trend, but the flow to the Southwestern region decreased most, and this reduce was more that 14%. The outflow to the Central region increased, it was 1.64 and 1.73 times that of their corresponding precious time points. The trend of the Northeastern and Southern regions were much similar, they both increased 83.28% and 96.52% from 2002 to 2007, but did not changed after 2007. For the node flow of each region, their flow through multiple paths took an account for more than 95%, and this was much higher than the flow through one path. Therefore, the results of the energy flows through multiple paths were much similar with the regions’ node flow. From the inflow perspective, the proportion of energy flow through one path among the node flow of the Southern region were the smallest, this showed that the Southern region mostly received products or by-products from other regions for its production processes. From the outflow perspective, the proportion of energy flow through one path of the Central region were the smallest. Using the total outflow of each region to minus its total inflow, we could obtain the net outflow. The results reflected that only the Northern region were negative in the three time points, representing Jing-Jin-Ji region received net resources from it. The Eastern, the Central, and the Southern region received net resources from Jing-Jin-Ji region, the Northeastern region changed from provider in 2007 to receiver in 2010, and the Southwestern and Northwestern changed from receiver to provider. 3.2. Provincial flow analysis When analyzing path flow and node flow at provincial level, the Jing-Jin-Ji region will be separate into Beijing, Tianjin, and Hebei, to trace the energy flows among them, between them and the other 27 provinces. 3.2.1. Path flow Fig. 5 showed the input and output energy flows through one path and multiple paths between Beijing and other 29 provinces in China. Fig. 5a is the flows from other regions to Beijing, we treat the sum of input flows and import flows from external environment as the total input throughflow. This total throughflow increased from 2002 to 2010, it was 2.02 and 1.10 times than each previous time point. The reason for these increase was the growth of input from other provinces, the input in 2007 was 2.29 times that in 2002, and that in 2010 was 1.33 times that in 2007. The import flows of Beijing although increased 63% from 2002 to 2007, but decreased from 2007 to 2010, and it was only 63% of the flow in 2007. The import only accounted for 19% to 41% of Beijing’s input from other provinces, this showed the flows from other provinces contributed most to the production of Beijing. For each province, in three time points, Hebei input most quantity of flows to Beijing, and the energy from Hainan and Qinghai were the smallest. Fig. 5a also showed the flows from Beijing to other regions. This total throughflow increased from 2002 to 2007 and then decreased, it was 3077.08 × 104 tce and 1.33 times that in 2002, then it decreased to 2708.05 × 104 tce. The reason for the changes was different. The total throughflow increase was due to the increase of the export to external environment, it was 1.81 times that in 2002. Then the decrease from 2007 to 2010 resulted from the decrease of outflow to other provinces, the total outflow to other provinces in 2010 was only 75.51% of that in 2007. For each province, in three time points, Beijing output most quantity of flows to Hebei in 2002 and 2007, and in 2010, Shanghai became the city received most energy through one path. The energy from Hainan and Qinghai were the smallest. Within the Jing-Jin-Ji region, to the inflow and outflow of Tianjin and Hebei through one path, the inflow from Hebei to Beijing was the highest in three time points, and Tianjin ranked medium among all 29 provinces. Although the outflow to Hebei was not the highest in 2010, the energy received 8

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 5. The inflow and outflow through one path and multiple paths of Beijing. Note: 1 Beijing, 2 Tianjin, 3 Hebei, 4 Shanxi, 5 Inner Mongolia, 6 Liaoning, 7 Jilin, 8 Heilongjiang, 9 Shanghai, 10 Jiangsu, 11 Zhejiang, 12 Anhui, 13 Fujian, 14 Jiangxi, 15 Shandong, 16 Henan, 17 Hubei, 18 Hunan, 19 Guangdong, 20 Guangxi, 21 Hainan, 22 Chongqing, 23 Sichuan, 24 Guizhou, 25 Yunnan, 26 Shaanxi, 27 Gansu, 28 Qinghai, 29 Ningxia, 30 Xinjiang.

energy from Hainan and Qinghai were the smallest.

was only 78.24% of that in 2007. For each province, in three time points, Shanxi input most quantity of flows to Hebei, and the energy from Hainan was the smallest. Fig. 7a also showed the flows from Hebei to other regions. This total throughflow increased from 2002 to 2010, but it grew more significant from 2002 to 2007 than from 2007 to 2010. The total throughflow was 2.74 times in 2007 that in 2002, then only increased 14.94% in 2010. The reason for the changes was different. The export and outflow from 2002 to 2007 both increased, they were 2.65 and 2.75 times that in 2002, respectively. From 2007–2010, the export dominated the increase of the total throughflow, the export grew 53.96% from 2007 to 2010. For each province, in three time points, Hebei output most quantity of flows to Jiangsu in three time points. The energy from Hainan and Qinghai were the smallest. Fig. 7b is the flows from other regions to Hebei through multiple paths. The total inflow increased from 2002 to 2007, and then decreased, it was 1.48 times in 2007 that in 2002. But in 2010, it was only 57.88% of that in 2007. For each province, in three time points, Shanxi input most quantity of flows to Hebei in three time points. The energy from Hainan was the smallest. Fig. 7b also showed the flows from Hebei to other regions through multiple paths. This total outflow increased from 2002 to 2007, and then decreased. The total outflow in 2007 was 2.37 that in 2002, and in 2010 decreased to 79.06% of that in 2007. For each province, in three time points, Hebei output most quantity of flows to Jiangsu in 2002, and in 2007 and 2010, changed to Shandong. The

3.2.2. Node flow Fig. 8 reflected the node flow between Beijing and other 29 provinces in China. The total inflow increased from 2002 to 2010. The result in 2007 was 1.38 times than in 2002, and that in 2010 was 1.13 times that in 2007. This indicated that Beijing depended more on other provinces. For each province, the inflow from Shanxi was highest in 2002, and then Hebei was the largest energy provider in 2007 and 2010. And the energy from Hainan and Qinghai were the smallest. The total outflow decreased from 2002 to 2010. The result in 2007 was 64.43% of that in 2002, and in 2010, it was only 62.55% of that in 2007. In the three time point, the outflow to Hebei was the highest in 2002, and in 2007 was Shandong, Jiangsu was the largest receiver in 2010. And the energy to Hainan and Qinghai were the smallest. Fig. 9 reflected the node flow between Tianjin and other 29 provinces in China. The total inflow increased from 2002 to 2007, then decreased. The result in 2007 was 1.83 times than in 2002, and that in 2010 was 77.20% of that in 2007. For each province, the inflow from Shanxi was highest in 2002, and then Hebei was the largest energy provider in 2007 and 2010. And the energy from Hainan was the smallest. The total outflow increased from 2002 to 2010. The result in 2007 was 1.73 times that in 2002, and in 2010, it still increased 1.91% compared with that in 2007. In the three time point, the outflow to 9

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 6. The inflow and outflow through one path and multiple paths of Tianjin (the names of the nodes are the same as that in Fig. 5).

region after 2007. Similar to Beijing, in 2002, Tianjin also mostly provided resources to Hebei, and then to Jiangsu after 2007. Hebei output its resources to Shandong and Jiangsu in the Eastern region at three time points. This showed that the analysis at regional level hides some changes caused by industrial restructuring at provincial level. As at the two levels, the energy flows through multiple paths dominated the node flow, their results have much similar. The flows through one path also showed difference at regional and provincial level. The results from regional analysis indicated that the Northern and the Eastern regions were the main provider from the Jing-Jin-Ji region, but at provincial level, for Beijing and Tianjin, the main energy provider was Hebei, and also the province which input most energy to the three areas were different. The provinces which provided most energy to Beijing besides Tianjin and Hebei which were located within JingJin-Ji region, Shaanxi should not be neglected. To Tianjin, Beijing and Hebei input most resources to it. Besides Shanxi, the provinces in the Northeastern also provided large quantity of resources to Hebei. The results of energy flows through one path at provincial level reflected that the three areas exchanges resources with provinces which have the advantages on the geographic position and administrative division, but the analysis at regional hides these exchanges. In energy outflow, the Jing-Jin-Ji region mostly output enegy to the Eastern, but at provincial level, Beijing provided most resources to Tianjin and Hebei in 2002 and 2007, and then changed to Shanghai, Besides Hebei, Tianjin also transferred resources to Jiangsu and Shanghai in the Eastern region. Hebei mostly provided resources to Jiang and Zhejiang. From these analysis, the results at regional level provides directions for the

Hebei was the highest in 2002, and in 2007 and 2010 changed to Jiangsu. And the energy to Hainan, Ningxia and Qinghai were the smallest. Fig. 10 showed the node flow between Hebei and other 29 provinces in China. The total inflow increased from 2002 to 2007, then decreased. The result in 2007 was 1.49 times than in 2002, and that in 2010 was 58.33% of that in 2007. For each province, the inflow from Shanxi was highest in three time points. And the energy from Hainan was the smallest. The total outflow increased from 2002 to 2007, and then decreased. The result in 2007 was 2.37 times that in 2002, and in 2010, it was only 79.82 of that in 2007. In the three time point, the outflow to Jiangsu was the highest in 2002 and 2010, and in 2007 changed to Shandong. And the energy to Hainan and Qinghai were the smallest. 3.3. The comparison of two levels From the node flow perspective, the inflow from the Northern and Eastern were the highest, but at the provincial level, the results in 2002 was similar with that at regional level, Shanxi input the most energy to others, but it changed since 2007, this was due to the industrial restructuring from 2007, some heavy industry and heavy pollution industries have been moved to Tianjin and Hebei from Beijing, especially to Hebei. This resulted in that Hebei became the resource provider for Tianjin and Beijing. From the aspect of outflow, the regional analysis showed that the Jing-Jin-Ji region mostly provided resources to the Eastern, but at provincial level, in 2002, Beijing provided most resources to Hebei, and then to Shandong and Jiangsu in the Eastern 10

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 7. The inflow and outflow through one path and multiple paths of Hebei (the names of the nodes are the same as that in Fig. 5).

Fig. 8. The node flow between Beijing and other provinces (the names of the nodes are the same as that in Fig. 5).

provincial research.

research chose the Jing-Jin-Ji urban agglomeration as case study. The node flows of each region and province have been divided into the energy flows through one path and multiple paths. Previous analysis used input-output analysis and deduced the embodied energy coefficient of the sectors or regions in the input-output tables, when

4. Discussion Based on input-output tables and ecological network analysis, this 11

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

Fig. 9. The node flow between Hebei and other provinces (the names of the nodes are the same as that in Fig. 5).

research considered the intermediate flows in the first quartile of the input-output table. Furthermore, their research traced the embodied energy flows, and did not subdivide these into flows through one path or multiple paths. Some research developed the tradition input-output analysis methods, based on the direct energy consumption coefficient (direct energy consumption divide the total output of each sector) and the second quartile of the input-output table, to diagonalize the matrix including direct energy consumption coefficient and multiply the Leontief inverse matrix and the corresponding final consumption, capital formation, and export, and then calculate the embodied energy consumption (Li et al., 2014). Li et al. (2014) accounted for the embodied energy consumption per capita in Chinese 30 provinces, but they did not quantify the flows among provinces. Chen and Chen (2016) combined ecological network analysis and improved the functional attributes of the urban energy network model, for example, cycling index, capability, and ascendency, and used ecological network analysis on the urban energy metabolic system. Then Wang and Chen (2016) extended this method to the Jing-Jin-Ji urban agglomeration, but their research all focused on energy-water nexus, and did not quantify the energy exchanges between the Jing-Jin-Ji region and other regions in China. Our research team already combined input-output tables with ecological network analysis methods to conduct urban energy metabolic processes analysis, for a single city (Beijing), firstly, to explore the effectiveness of this combination (Zhang et al., 2014b, 2015a), then extended this method to urban agglomeration analysis, they considered

calculating the energy flows among multi regions, introducing the theory of energy trade, the embodied energy flows equals to the embodied energy coefficient multiply the corresponding import quantity (Cui et al., 2015). In addition, some analysis combined system ecology, and based on the input and output flows of each sector to establish an equilibrium equation, this can also obtain the embodied energy coefficient among sectors or regions, and then multiply the import or export quantity. Zhang et al. (2013) studied the embodied energy flows among 30 provinces in China and their research showed that in 2007, Hebei was the main area to output energy resources, it distributed resources to Beijing, Tianjin, Jiangsu, Shanghai, Zhejiang and Guangdong. Compared to Hebei, Beijing and Tianjin only output less energy resources. This is different from the results in this research. In this analysis, in 2007, Hebei provided most energy resources to Shandong, then was Zhejiang, and the next Henan and Guangdong. To the four municipalities in China (Beijing, Tianjin, Shanghai, and Chongqing), Zhang et al. (2016a) pointed out that in 2007, Tianjin provided most node flow to Beijing and then was Jiangsu, the next was Hebei; Tianjin output most energy resources to Jiangsu, then was Zhejiang and Shanghai. From the inflow perspective, Shaanxi, Inner Mongolia, and Hebei input most energy resources to Beijing and Tianjin. In this research, Beijing output most energy resources to Shandong, then was Jiangsu and Hebei. For the output flow, Tianjin provided most energy resources to Jiangsu, then was Shandong and Zhejiang. And Hebei, Shanxi, and Shandong output most energy resources to Beijing and Tianjin. The reason for this difference is because Zhang et al. (2013) only focused on the quantity of energy trade among provinces, and this

Fig. 10. The node flow between Hebei and other provinces (the names of the nodes are the same as that in Fig. 5).

12

Resources, Conservation & Recycling xxx (xxxx) xxx–xxx

H. Zheng et al.

that energy flows not only exists among the sectors within one area, but also exchanges among areas (Zheng et al., 2017). Zhang et al. (2015b) further extended this method to regional level and analyzed the embodied energy flows among 30 provinces in China. But this research only chose the time points 2007 and 2010. This research based on the previous analysis to trace the energy metabolic flows between the JingJin-Ji region and others in China at two levels from 2002 to 2010.

Simulat. 18 (7), 1757–1774. Chen, S., Chen, B., 2015. Urban energy consumption: different insights from energy flow analysis, input-output analysis and ecological network analysis. Appl. Energy 138, 99–107. Chen, S., Chen, B., 2016. Urban energy-water nexus: a network perspective. Appl. Energy 184, 905–914. Chen, H., 2011. Interregional Systems Accounting of Energy and Carbon Emissions for China. Ph.D. Thesis. Peking University, Beijing, China (in Chinese). Cui, L.B., Peng, P., Zhu, L., 2015. Embodied energy, export policy adjustment and China's sustainable development: a multi-regional input-output analysis. Energy 82, 457–467. Fath, B.D., Patten, B.C., 1999. Review of the foundations of network environ analysis. Ecosystems 2 (2), 167–179. Garcia-Montiel, D.C., Verdejo-Ortiz, J.C., Santiago-Bartolomei, R., Vila-Ruiz, C.P., Santiago, L., Melendez-Ackerman, E., 2014. Food sources and accessibility and waste disposal patterns across an urban tropical watershed: implications for the flow of materials and energy. Ecol. Soc. 19 (1), 37. Lenzen, M., 1998. Primary energy and greenhouse gases embodied in Australian final consumption: an input-output analysis. Energy Policy 26, 495–506. Li, S.S., Zhang, Y., Yang, Z.F., Liu, H., Zhang, J.Y., 2012. Ecological relationship analysis of the urban metabolic system of Beijing, China. Environ. Pollut. 170, 169–176. Li, Z., Pan, L., Fu, F., Liu, P., Ma, L., Amorelli, A., 2014. China's regional disparities in energy consumption: an input-output analysis. Energy 78, 426–438. Li, Y., Meng, J., Liu, J.F., Xu, Y., Guan, D.B., Tao, W., Huang, Y., Tao, S., 2016. Interprovincial reliance for improving air quality in China: a case study on black carbon aerosol. Environm. Sci. Technol. 50, 4118–4126. Liang, Q.M., Fan, Y., Wei, Y.M., 2007. Multi-regional input-output model for regional energy requirements and CO2 emissions in China. Energy Policy 35, 1685–1700. Liang, S., Zhang, C., Wang, Y., Xu, M., Liu, W., 2014. Virtual atmospheric mercury emission network in China. Environ. Sci. Technol. 48, 2807–2815. Liu, W.D., Chen, J., Tang, Z.P., Liu, H.G., Han, D., Li, F.Y., 2012. Compliment Theory and Practice of China’s Interregional Input-output Table for 30 Regions in 2007. China Statistics Press, Beijing (in Chinese). Liu, W.D., 2014. Compliment Theory and Practice of China’s Interregional Input-output Table for 30 Regions in 2010. China Statistics Press, Beijing (in Chinese). Lu, W., Su, M., Zhang, Y., Yang, Z., Chen, B., Liu, G., 2014. Assessment of energy security in China based on ecological network analysis: a perspective from the security of crude oil supply. Energy Policy 74, 406–413. Polenske, K.R., The U.S. Multiregional Input-Output Accounts and Model, Lexington, MA, USA, 1980. Shi, M.J., Zhang, Z.Y., 2012. Multi-regional Input-output Model and Inter-regional Economic Linkages. Science Publishing House, Beijing (in Chinese). Tian, X., Chang, M., Lin, C., Tanikawa, H., 2014. China’s carbon footprint: a regional perspective on the effect of transitions in consumption and production patterns. Appl. Energy 123, 19–28. Wang, S., Chen, B., 2016. Energy-water nexus of urban agglomeration based on multiregional input-output tables and ecological network analysis: a case study of the Beijing-Tianjin-Hebei region. Appl. Energy 178, 773–783. Wiedmann, T., Lenzen, M., Turner, K., Barrett, J., 2007. Examining the global environmental impact of regional consumption activities-Part 2: review of input-output models for the assessment of environmental impacts embodied in trade. Ecol. Econ. 61 (1), 15–26. Wiedmann, T.O., Schandl, H., Lenzen, M., Moran, D., Suh, S., West, J., Kanemoto, K., 2015. The material footprint of nations. Proc. Natl. Acad. Sci. 112 (20), 6271–6276. Wiedmann, T., 2009. A review of recent multi-region input-output models used for consumption-based emission and resource accounting. Ecol. Econ. 69 (2), 211–222. Yang, S., Chen, B., Fath, B., 2015. Trans-boundary total suspended particulate matter (TSPM) in urban ecosystems. Ecol. Modell. 318, 59–63. Zhang, Y., Yang, Z., Fath, B.D., Li, S.S., 2010. Ecological network analysis of an urban energy metabolic system: model development, and a case study of four Chinese cities. Ecol. Modell. 221 (16), 1865–1879. Zhang, B., Chen, Z.M., Xia, X.H., Xu, X.Y., Chen, Y.B., 2013. The impact of domestic trade on China's regional energy uses: a multi-regional input-output modeling. Energy Policy 63, 1169–1181. Zhang, Y., Zheng, H.M., Fath, B.D., Liu, H., Yang, Z.F., Liu, G.Y., Su, M.Y., 2014a. Ecological network analysis of an urban metabolic system based on input-output tables: model development and case study for Beijing. Sci. Total Environ. 468, 642–653. Zhang, Y., Zheng, H.M., Fath, B.D., 2014b. Analysis of the energy metabolism of urban socioeconomic sectors and the associated carbon footprints: model development and a case study for Beijing. Energy Policy 73, 540–551. Zhang, Y., Zheng, H.M., Yang, Z.F., Li, J.J., Yin, X.A., Liu, G.Y., Su, M.R., 2015a. Analysis of urban energy consumption in carbon metabolic processes and its structural attributes: a case study for Beijing. J. Clean. Prod. 103, 884–897. Zhang, Y., Zheng, H.M., Yang, Z.F., Su, M.R., Liu, G.Y., Li, Y.X., 2015b. Multi-regional input-output and ecological network analysis of embodied energy exchanges: a case study of the exchanges in China. Energy Policy 86, 651–663. Zhang, B., Qiao, H., Chen, Z.M., Chen, B., 2016a. Growth in embodied energy transfers via China’s domestic trade: evidence from multi-regional input-output analysis. Appl. Energy 184, 1093–1105. Zhang, Y., Zheng, H.M., Yang, Z.F., Li, Y.X., Liu, G.Y., Su, M.R., Yin, X.A., 2016b. Urban energy flow processes in the Beijing-Tianjin-Hebei (Jing-Jin-Ji) urban agglomeration: combining multi-regional input-output tables with ecological network analysis. J. Clean. Prod. 114, 243–256. Zheng, H.M., Fath, B., Zhang, Y., 2017. An urban metabolism and carbon footprint analysis of the Jing-Jin-Ji regional agglomeration. J. Ind. Ecol. 21 (1), 166–179.

5. Conclusions This research chose the Jing-Jin-Ji urban agglomeration as case study, to conduct urban energy metabolism analysis at the regional and provincial levels from the perspectives of path flow and node flow. The path flow through one path showed the Northern and the Eastern regions were the main provider from the Jing-Jin-Ji region at the regional level; at provincial level, for Beijing and Tianjin, the main energy provider was Hebei, for Hebei, besides Shanxi, the provinces in the Northeastern also provided large quantity of resources. The provincial results reflected that the three areas exchanges resources with provinces which have the advantages on the geographic position and administrative division, but the analysis at regional hides these exchanges. The energy flows through multiple paths dominated the node flow, the results of path flow on multiple paths were much similar with node flow at two level. The input node flow from the Northern and Eastern to Jing-Jin-Ji region were the highest, but at the provincial level, the results in 2002 was similar with that at regional level, Shanxi input the most energy to others, but it changed since 2007, this was due to the industrial restructuring from 2007, some heavy industry and heavy pollution industries have been moved to Tianjin and Hebei from Beijing, especially to Hebei. This resulted in that Hebei became the resource provider for Tianjin and Beijing. The analysis showed that the regional level can provide directions for provincial level, but hide the resource exchanges among provinces due to the industrial restructuring. These two levels research not only reflects the supplementary processes between the Jing-Jin-Ji region and other regions in China, but also highlights the heterogeneity of the two cities and one province within the agglomeration. In this research, we merged the sectors in each province or city to the provincial, and even regional level. In the next step, we can further specify the discussion at the sectoral level and illustrate the urban energy metabolic processes between the Jing-Jin-Ji region and other regions in China. Acknowledgments This work was supported by the National Key Research and Development Program of China (No. 2016YFC0503005), by the Fund for Innovative Research Group of the National Natural Science Foundation of China (no. 51421065), by Funds for International Cooperation and Exchanges of the National Natural Science Foundation of China (no. 51661125010), by the Program for New Century Excellent Talents in University (no. NCET-12-0059), by the National Natural Science Foundation of China (no. 41571521), and by the Fundamental Research Funds for the Central Universities (no. 2015KJJCA09). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.resconrec.2017.05.010. References Baynes, T.M., Bai, X., 2012. Reconstructing the energy history of a city. J. Ind. Ecol. 16 (6), 862–874. Chen, Z.M., Chen, G.Q., 2013. Demand-driven energy requirement of world economy 2007: a multi-region input-output network simulation. Commun. Nonlin. Sci. Numer.

13