Decoupling effect and sectoral attribution analysis of industrial energy-related carbon emissions in Xinjiang, China

Ecological Indicators 97 (2019) 1–9

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Decoupling effect and sectoral attribution analysis of industrial energyrelated carbon emissions in Xinjiang, China

T

Xinlin Zhanga, , Yuan Zhaob,c, , Changjian Wangd, Fei Wange, Fangdao Qiua ⁎

⁎

a

School of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, China School of Geographic Science, Nanjing Normal University, Nanjing 210023, China c Ginling College, Nanjing Normal University, Nanjing 210097, China d Guangzhou Institute of Geography, Guangdong Open Laboratory of Geospatial Information Technology and Application, Guangzhou 510070, China e Xinjiang Laboratory of Lake Environment and Resources in Arid Zone, School of Geography Science and Tourism, Xinjiang Normal University, Urumqi 830054, China b

ARTICLE INFO

ABSTRACT

Keywords: Decoupling relationships Sub-sector Driving factors Targeted policies

As the gateway of “the Belt and Road”, it is of great significance for Xinjiang to fulfill the carbon reduction target without compromising the steady socio-economic development. A comprehensive understanding of the decoupling relationships between industrial growth and carbon emissions as well as its driving factors is helpful to make targeted recommendations. In this paper, decoupling analysis, index decomposition analysis, and attribution analysis were applied to analyze the decoupling effect, driving factors, and contributions of sub-sectors to each driving factor, respectively. Some conclusions were drawn. (1) Xinjiang’s decoupling relationships between industrial growth and carbon emissions experienced three stages, i.e., “negative decoupling, weak decoupling, and negative decoupling”. Xinjiang’s industrial sector was in relatively unsustainable period after 2008. (2) From 2000 to 2014, energy intensity promoted the decoupling process, and industrial structure and energy structure were the main factors which inhibited the decoupling process. But after 2008, all driving factors showed negative influences. (3) After 2008, fuel processing, textile, and mining and quarrying were primarily responsible for the energy intensity’s negative influence; fuel processing was primarily responsible for energy intensity’s negative influence; production and supply, smelting and pressing of metals, and chemicals were primarily responsible for industrial structure’s negative influence. (4) This paper suggests that Xinjiang should make targeted carbon reduction policies at sub-sector level.

1. Introduction Global warming, which is primarily caused by fossil energy use, has attracted widespread attention. In order to avoid the potential damages to natural ecosystem and the survival of human beings, the international community has come to an agreement that the average increase of global temperature should be kept below 2 degrees Celsius by the end of this century (Conte Grand, 2016). China has become the world’s largest emitter of carbon emissions (Zhang and Da, 2015, Shao et al., 2016a). Due to the high-speed industrialization and urbanization, more energy will be consumed, and it is inevitable for China to emit more carbon emissions in the foresee future (Xu et al., 2016). Under such a circumstance, Chinese government has made effort to address the climate change and decrease carbon intensity. Before the Copenhagen conference in 2009, Chinese government promised to reduce its carbon intensity by 40–45% by 2020 in comparison with the 2005 level (Qiu,

2009). In 2014, a joint announcement on climate change was issued with United States, and China made a promise that the carbon emissions peak would be achieved before 2030(David, 2014). During the period of Thirteen Five-Year Plan (2016–2020), the government supported some provincial regions to be the first to achieve the carbon emissions peak (Cheng, 2016, Shao et al., 2016b). Fulfilling the carbon reduction target without compromising the steady socio-economic development highlights the urgency and importance of studies on the relationship between the economic growth and carbon emissions. In recent years, more and more researchers have applied the decoupling model to analyze the relationship between economic growth and environmental changes. And these studies can be primarily divided into two aspects. Firstly, various environmental indicators, such as carbon dioxide, soot, waste water, and solid waste, were applied to examine the decoupling level (Climent and Pardo, 2007; Yu et al., 2017; Zhao et al., 2017). Secondly, many studies focused on exploring the

Corresponding authors at: Shanghai Road 101, Tongshan District, Xuzhou City, Jiangsu Province 221116, China. E-mail addresses: [email protected] (X. Zhang), [email protected] (Y. Zhao), [email protected] (C. Wang), [email protected] (F. Wang), [email protected] (F. Qiu). ⁎

https://doi.org/10.1016/j.ecolind.2018.09.056 Received 27 February 2017; Received in revised form 10 September 2018; Accepted 27 September 2018 1470-160X/ © 2018 Elsevier Ltd. All rights reserved.

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

influencing factors leading to the decoupling changes (Dong et al., 2016; Zhao et al., 2016; Jiang et al., 2016; Zhou et al., 2017; Zhang et al., 2017a). These studies analyzed the changes in decoupling states and factors’ influences, but failed to explore the contribution of each individual sub-sector to decomposed factors. The attribution analysis model was proposed by Choi and Ang (2012) aiming to measure the contribution of each individual sub-sector to the effect of each decomposed factor. It is viewed as a perfect extension of the index decomposition analysis (IDA) model (Liu et al., 2015a). Some scholars have used the attribution analysis to do some analysis, and their studies focused on different countries or international organizations, i.e., China (Liu et al., 2015a; Su and Ang, 2014; Wang et al., 2016a), Korea (Choi and Oh, 2014), Mexico (González and Martínez, 2012), and European Union (Fernández González, 2015; Fernández González et al., 2013, 2015). Combining the decoupling analysis and IDA models is more effective (de Freitas and Kaneko, 2011), but few studies incorporated both the decoupling analysis and attribution analysis. Until now, studies combining the decoupling index and attribution analysis were much scarcer, and only Wang et al. (2016b) used decoupling index and attribution analysis to conduct a deeper understanding of the decoupling states of industrial growth from carbon emissions and the driving factors in Taiwan. Besides, above studies using attribution analysis were not based on different stages. Conducting the analysis with different stages may have different conclusions. Thus, the study period is divided into different stages based on the decoupling states in this paper. Energy consumption, carbon emissions, and the economic development are obvious different in different regions within China (Wang et al., 2014; Zhang et al., 2014). Without a sufficient understanding of the decoupling relationships at the provincial level, it is less effective to implement the national policies or strategies (Liu et al., 2012; Liu et al., 2015b). Due to its strategic location, Xinjiang becomes a crucial gateway in the second “Eurasian Land Bridge”. And with the implement of “the Belt and Road”, Xinjiang will play a more important role in social-economic cooperation with Russia, Central Asian countries, et al. Although these strategic policies promote Xinjiang’s economic development, its economic development is still characterized by high energy consumption and high carbon emissions. To be a responsible country and sponsor nation of “the Belt and Road” (Wang et al., 2017a), China should pay enough attention to its energy-related carbon emissions in order to stop the regional environmental deterioration. Xinjiang, as the important region in “the Belt and Road”, should set a good example to fulfill the carbon reduction target without compromising the steady socio-economic development. Since the implement of “Western Development” in 2000, Xinjiang’s industrial sector had been growing fast on the basis of its abundant nature resources. Due to its important status of the energy base, Xinjiang plays an important role in providing energy products for the national and regional economic development, so it is also necessary to conduct an analysis. Over 2000–2014, the total carbon emissions increased from 21.41 million tons to 112.38 million tons with an annual growth rate of 13.01%, and > 80% of the energy-relative carbon emissions was emitted by the industrial sector1. Thus, the industrial sector should be the paid enough attention. Some studies were also conducted on Xinjiang’s carbon emissions, e.g., Huo et al. (2015), Ma et al. (2013). But, the previous studies did not study the sub-sectors’ contributions to the driving factors. The differences between this paper and the existing researches mainly lies in the following aspects. Firstly, a depth study of attribution analysis was conduct at a provincial level. Secondly, according to the comparative attribution analysis of different stages, different conclusions may be drawn. Thirdly, studying the sub-sectors’ contributions to each driving factors of Xinjiang industrial carbon emissions can help make more targeted policies.

Fig. 1. Location of Xinjiang in China (This figure originates from the National Administration of Surveying, Mapping and Geoinformation).

In order to assist Xinjiang in decoupling industrial carbon emissions from industrial development, and achieving carbon reduction target and sustainable development, this paper aimed to conduct a deeper understanding of the decoupling relationship between Xinjiang’s industrial growth and carbon emissions, and then recognize the influence of each driving factor and the contribution of each industrial sub-sector. Specifically, decoupling index model was applied to examine the decoupling relationship. Sato-Vartia Logarithmic mean Divisia Index (LMDI) model was applied to analyze the decomposed factors’ influence on changes in industrial carbon intensity. Based on the decomposed results, this paper used attribution analysis to analyze the contribution of industrial sub-sectors to decomposed factors. In the end, targeted recommendations and policies can be made at sub-sector level. 2. Study area and methodology 2.1. Study area Xinjiang, located in the northwest of China (Fig. 1), has an area of 1.66 million km2 which accounts for about 1/6 of total land area of China. Based on the second national oil and gas resources evaluation, Xinjiang’s predictive reserves of coal, oil, and gas accounted for 40%, 30%, and 34% of the national onshore gross reserves, respectively (Wang and Wang, 2015). Rich resources made industry, especially the energy-intensive industrial sectors, be the primary sectors. Rich resources promoted its economic development and it also caused damage to environment (Wang et al., 2017b; Wang and Wang, 2017). 2.2. Decoupling model In order to solve the issues regarding the dependence of economic growth on material consumption, decoupling theory was proposed (Pang et al., 2014). The Word Bank and the Organization for Economic Cooperation and Development (OECD) explained decoupling concept, respectively. The Word Bank indicates that the decoupling, including both dematerialization and depollution, is the process of gradually reducing the economic activities’ effects on the environment (Bruyn and Opschoor, 1997). OECD states that a decoupling relationship existed when the growth of environmental pressures was slower than that of economic development (OECD, 2005); and there is a negative decoupling relationship, when the growth of economic development was slower than that of environmental pressures. Tapio decoupling index and OECD decoupling index are the two models to examine the decoupling states, and these two decoupling indexes had also applied to analyze the decoupling relationships between economic growth and

1 The data was calculated on the basis of Xinjiang Statistical Yearbooks (2001–2015).

2

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

carbon emissions (Chen et al., 2017; Roinioti and Koroneos, 2017). Thus, this paper examines the decoupling index between Xinjiang’s industrial growth and carbon emissions on the basis of Tapio (2005), and Wang et al. (2016b). To analyze the decoupling relationship between industrial growth and carbon emissions in Xinjiang, this paper examines the decoupling index on the following equation:

cg =

% C 1= % G

C0 G0

T / C0

1

T / G0

energy intensity effect (DEI), industrial structure effect (DIS). Carbon emission coefficient effect (DCE) reflects the industrial carbon intensity change caused by carbon emission coefficient variation. Energy structure effect (DES) reflects the industrial carbon intensity change caused by energy structure shift. Energy intensity effect (DEI) reflects the industrial carbon intensity change caused by energy intensity change. Industrial structure effect (DIS) reflects the industrial carbon intensity change caused by industrial structure shift. The equations, applying the Sato-Vartia LMDI method to estimate effect of each decomposed factor, are shown as Eqs. (B.1)–(B.6) in Appendix B.

(1)

where cg is the decoupling index between industrial growth and carbon emissions in Xinjiang; C is the industrial energy-related carbon emissions; G is the industrial added value; %△C, %△G denote the percentage change of carbon emissions and industrial added value, respectively; △C0–T, △G0–T denote the changes in carbon emissions and industrial added value from year 0 to year T. According to the decoupling index, the decoupling or negative decoupling relationship can be divided into six sub-categories basing on the change in industrial carbon emissions (ΔC), the change in industrial added value (ΔG), and the decoupling index (cg). Table A.1 in Appendix A showed the classification standards.

2.4. Attribution analysis model The contributions of each sub-sector attribute to the percentage changes in industrial carbon intensity can be assessed by means of attribution analysis model (González and Martínez, 2012). Based on the results of decomposition analysis, the attribution analysis reveals which sub-sectors were more easily regulated by the recent implemented policies and measures (Liu et al., 2015a). After deriving each factor’s influence, attribution analysis model was used to further attribute the changes in industrial carbon emissions into 15 pre-defined industrial sub-sectors in this paper. Eq. (5) expresses the single-period attribution analysis of industrial structure effect.

2.3. Decomposition model It is effective to explore the influencing factors of carbon emissions using the index decomposition. Structural decomposition analysis (SDA) and index decomposition analysis (IDA) were the two primary methods. The SDA model needs to depend on input-output model. Although the conclusions which are drawn by the means of SDA are more accurate than these applying IDA (Rose and Casler, 1996), the high requirements of data restricted SDA model’s application. LMDI model is one type of IDA model, which has the advantages of zero residuals, independent pathway, and consistent aggregation (Xu and Ang, 2013), making it widely applied. Thus, this paper used LMDI method to decompose the changes in industrial carbon intensity. Kaya identity (Liu et al., 2015a) states that the carbon emissions can be expressed as a result of four driving factors:

G E C C=P× × × P G E

I t 1, t DIS 1=

I i=1

J j=1

G I

Cij

I

J

i=1 j =1

Cij Eij

×

Eij Ei

×

CEij × ESij × EIi × ISi i=1 j=1

CIt = DCE × DES × DEI × DIS CIt 1

i=1 j =1

wijS V

N IS t 1 i L (IS t 1 Dt 1, t , IS t ) i i IS i

ISit 1 ISit 1

r IS, ij where is percentage changes of industrial structure; is the contribution of industrial sub-sector i from year t-1 to year t. Eqs. (B.7)–(B.9) in Appendix B expresses the single-period attribution analysis of other relative driving factors. In order to examine the attribution result of each decomposed factor, it can be obtained applying the following equation: I

j j=1

J

T

Dz0, T 1 =

I

J

t

1, t

T

rz0,, ijT = i=1 j =1 t=1

Dz0, t 1 rzt , ij1, t i =1 j=1 t=1

(6)

1 is the cumulative changes of a relative decomposed factor where j 0, T over the period 0–T; j = 1 r z , j is the contribution of industrial i to the cumulative changes of the decomposed factor z over the period 0–T. Dz0, T

2.5. Data sources This paper only analyzed three decomposed factors, because carbon emission coefficient remained invariable based on some previous studies (Tan et al., 2011; Wang et al., 2016b; Wang and Yang, 2015). Since 2000, the “Western Development” had begun to implemented. Then, Xinjiang’s economy gained the opportunity to grow fast. Meanwhile, its carbon emissions also grew fast (Zhang and Zhao, 2018). Thus, this paper analyzed the decoupling effect and sectoral attribution over 2000–2014. Conducting attribution analysis for a country or region required relatively detailed data. The energy consumption and added value should be decomposed into each sub-sector if it needs to conduct the analysis of each sub-sector’s contribution. The consumption of different fuel type of each industrial sub-sector was needed, including raw coal, cleaned coal, other washed coal, coke, crude oil, gasoline, kerosene, diesel oil, LPG, refinery gas, other petroleum products, and gas. Then, energy structure can be obtained. And industrial carbon emissions can be also calculated using above data. Xinjiang’s industrial carbon emissions were calculated applying the equation C t = i Eit × LCVi × CFit × Oi . Where the subscript i is the fuel type; superscript t is the year t; Ct is the industrial carbon emission; Eit is the fuel consumption; LCVi is the lower calorific value of energy fuel type i;

Ei G × i Gi G (3)

where Cij is the industrial carbon emissions of energy j consumed by sub-sector i; Eij is the consumption of energy j by industrial sub-sector i; Ei is the total energy consumption of industrial sub-sector i; Gi is the added value of industrial sub-sector i; G is the total industrial added value; CEij is the carbon emissions coefficient of fossil energy type j; ESij is the energy structure of industrial sub-sector i; EIi is the energy intensity of industrial sub-sector i; ISi is the industrial structure. Sato-Vartia LMDI (Wang et al., 2016b) method is applied to decompose the changes in industrial carbon intensity into four driving factors, as shown in Eq. (4).

DA =

t 1, t rIS , ij =

(5)

J

=

wijS V t 1 t 1, t , IS t ) ISi L (ISit 1 DIS i

J

t 1, t DIS -1

(2)

=

I

i=1 j=1

where C is carbon emissions; P is population size; G is the gross domestic product (GDP); E is the energy consumption. In order to find the driving factors for the decoupling process, the industrial carbon intensity in Eq. (3) can be expressed, on the basis of an extended Kaya identity, as the following equation:

CI =

J

(4)

where DA is the total effect caused by the four decomposed factors: carbon emission coefficient effect (DCE), energy structure effect (DES), 3

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

CFit is the carbon emissions factors of fuel type i; Oi is the oxidation rate of energy type i. Each industrial sub-sector’s added value was also collected. The annual added value of each industrial sub-sector was converted into 2000 constant price. Then the industrial structure of each sub-sector can be also obtained. All data were collected from Xinjiang Statistical Yearbooks (2001–2015). About 39 industrial sub-sectors were listed in Xinjiang Statistical Yearbooks (2001–2015), and the industrial sub-sectors of different years were not completely identical. So, based on the classification method in Lv, et al. (2012), these industrial sub-sectors in Xinjiang Statistical Yearbooks (2001–2015) were merged into 15 industrial sub-sectors (Appendix C Table C.1).

4.000

G/G

C/C

0.450

cg

3.000

0.300 0.150

cg

1.000

0.000

0.000

-1.000

G/G, C/ C

2.000

-0.150

-2.000 -0.300

-3.000 -4.000

-0.450 2013-2014

2012-2013

2011-2012

2010-2011

2009-2010

2008-2009

2007-2008

2006-2007

2004-2005

2005-2006

2002-2003

2003-2004

2001-2002

2000-2001

3. Results

Fig. 2. Changes in rate of industrial carbon emissions, industrial output, and decoupling index.

3.1. Decoupling analysis We applied Eq. (1) to calculate the decoupling index, and then obtained the decoupling states on the basis of Table A.1 in Appendix A. The results were shown in Figs. 2 and 3. Based on the decoupling index (Fig. 2) and state (Fig. 3), Xinjiang’s decoupling relationship between industrial growth and carbon emissions can be divided into three stages. For the first stage (2000–2002), this period was in negative decoupling stage. From 2000 to 2001, industrial output decreased, while industrial carbon emissions still increased, and a slight industrial recession did not lead to decrease in carbon emissions, creating a strong negative decoupling effect. From 2001 to 2002, the industrial sector recovered and output increased, while the increase in industrial carbon emissions was even larger, leading to an expansive negative decoupling effect. For the second stage (2002–2008), this period was in weak decoupling state. Both industrial output and carbon emissions continuously increased during this period. However, the annual average growth rate of industrial carbon emissions was lower than industrial output growth rate, and the decoupling index decreased first and then increased. The stress faced by carbon emissions still increased with the increases in industrial output. The decoupling index of 2003–2004 was minimum over 2000–2014, and this illustrated that this period was the relatively most sustainable period. For the third stage (2008–2014), this period was also in negative decoupling stage. In 2008–2009, both industrial output and carbon emissions decreased because of the global financial crisis in 2008 and incident of violence in 2009 (Zhang et al., 2013), resulting in weak negative decupling state. From 2009 to 2014, Xinjiang’s industrial output and carbon emissions continuously increased. However, carbon emissions grew rapidly, with an annual average growth rate of 19.15%, while the industrial output growth rate was 11.02%. This led to expensive negative decoupling state over the period 2009–2014. Overall, conducting the analysis with different stages helps understand the decoupling relationships accurately, which is the basis for making more targeted recommendations.

Fig. 3. Decoupling results between industrial carbon emissions and industrial output in Xinjiang.

Table 1 Decomposition results of changes in Xinjiang’s industrial carbon intensity.

Energy structure effect Energy intensity effect Industrial structure effect Total effect

2000–2002

2002–2008

2008–2014

2000–2014

0.977 1.241 0.903 1.095

0.973 0.536 1.107 0.577

1.084 1.130 1.209 1.482

1.031 0.751 1.208 0.936

relatively unobvious during the first two years. Production technology had not been efficiently improved, and energy efficiency of the industrial sector had not improved, yet. Initial investment had not increased the proportion of energy-intensive industries in total industrial output. Under such a background, energy intensity effect was the main driving factor for the increase, which showed a negative effect on the decoupling between industrial growth and carbon emissions. In this stage, energy efficiency decreased, thus carbon emissions stress increased despite industrial recession in 2000–2001 and even recovered in 2001–2002. Changes in industrial structure and energy structure were the inhibiting factors, and these two factors promoted the decoupling process. During the second stage (2002–2008), the value of total effect was 0.577, indicating that industrial carbon intensity decreased by 42.3%. Over 2002–2008, Xinjiang’s industrial sector increased obviously with the implementation of “Western Development”. Energy-intensive industries began to rapidly increase on the basis of Xinjiang’s rich energy.

3.2. Decomposition analysis To explain different decoupling effects between Xinjiang’s industrial growth and carbon emissions, the changes in industrial carbon intensity were decomposed into energy structure effect, energy intensity effect, and industrial structure effect. We applied the Eqs. (B.2)–(B.4) in Appendix B to calculate the energy structure effect, energy intensity effect, and industrial structure effect, respectively. The results were shown in Table 1. During the first stage (2000–2002), the value of total effect was 1.095, indicating that industrial carbon intensity increased by 9.5%. Xinjiang is a less developed area in China. Although “Western Development” had been implemented since 2000, its effect was 4

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

Besides, production technology had also been efficiently improved. Meanwhile, the proportion of oil in total energy consumption also increased. According to Table 1, the values of energy intensity effect and energy structure effect were all < 1. It illustrated that the primary factors driving the decoupling process between industrial growth and carbon emissions were the improvement in energy efficiency and optimization in energy structure. Decreases in energy intensity contributed most to the decoupling process. Adjustment of industrial structure negatively influence the decoupling process. During the third stage (2008–2014), the value of total effect was 1.482, indicating that industrial carbon intensity increased by 48.2%. The values of three decomposed factors excessed 1, and changes in these factors all exerted a negative effect on decoupling between industrial growth and carbon emissions. In this stage, more policies had been implemented. For example, Xinjiang became the new national energy strategic base, and new large-scale assistance programs were also proposed in 2010. These policies promoted the development of Xinjiang’s energy-intensive industries, leading to increase in proportion of energy-intensive industries in total industrial output. Nevertheless, less attentions were paid to improve the energy efficiency and optimize the energy structure. Thus, industrial structure effect showed an increase trend over 2008–2014, and became the most influential factors. Energy efficiency decreased, and industrial structure and energy structure were not optimized. Based on above analysis, the changes in Xinjiang’s industrial carbon intensity showed three stages of “rise, decline, and rise”, and decoupling process also experienced “negative decoupling, weak decoupling, and negative decoupling” three-stage process. From 2000 to 2014, energy intensity promoted the decoupling process, and industrial structure and energy structure were the main factors which inhibited the decoupling process. In comparison with two other decomposed factors, energy structure effect in the decoupling process were not significant. The detailed reasons for the changes in these factors’ influence would be explained from the perspective of sub-sectors using the attribution analysis.

Table 2 Attribution results of energy structure effect in Xinjiang (Unit: %). Sub-sectors

2000–2002

2002–2008

2008–2014

Mining and quarrying Foods and tobacco Textile Timber and furniture Pulp and paper Fuel processing Chemicals Non-metallic mineral products Smelting and pressing of metals Metal products General and special purpose machinery Transport equipment Electrical machinery and equipment Production and supply Other manufactures Total

−0.76 0.22 −0.19 −0.99 −0.16 0.26 0.05 0.48 0.18 1.25 −1.52 0.16 0.05 −1.24 −0.09 −2.29

−10.14 −1.35 −2.17 1.25 2.10 0.73 3.86 −1.68 7.59 4.84 −2.09 −0.05 −3.08 −2.27 −0.26 −2.74

5.89 −2.64 7.54 3.07 0.19 9.95 −1.77 −1.52 −0.86 −1.02 −4.57 −1.46 −2.36 −2.23 0.23 8.45

Table 3 Attribution results of energy intensity effect in Xinjiang (Unit: %).

3.3. Attribution analysis Applying attribution analysis, the percent changes in industrial carbon intensity can be attributed to the 15 pre-defined industrial subsectors. Then, each industrial sub-sector’s contribution can be obtained. The results were shown in Tables 2–4. The attribution analysis of energy structure effect was conducted applying Eqs. (6) and (B.8) in Appendix B, and the results were shown in Table 2. Energy structure effect was much less than the effects of other decomposed factors over 2000–2014, but it should not be ignored. As shown in Table 2, during the first stage, energy structure generally hindered the increases in industrial carbon intensity; nonmetallic mineral products contributed most to promote the increases; general and special purpose machinery contributed most to hinder the increases. During the second stage, energy structure promoted the decreases in industrial carbon intensity; the influence of mining and quarrying was most significant in promoting the decreases; smelting and pressing of metals, metal products, and chemicals were primarily responsible for hindering the decreases. During the third stage, energy structure effect began to positively influence the increases in industrial carbon intensity; fuel processing, textile, and mining and quarrying significantly contributed to the results; general and special purpose machinery was the leading factor which negatively influenced energy structure effect. Although the primary sub-sectors, contributing to the changes in the energy structure effect, had been changing in different stages, these sub-sectors belong to energy-intensive industries with the characteristic of high-carbon emissions and high consumption. Thus, readjusting the energy structure of energy-intensive sub-sectors is more effective. After 2008, fuel processing, textile, and mining and quarrying became the primary industrial sub-sectors negatively influencing the

Sub-sectors

2000–2002

2002–2008

2008–2014

Mining and quarrying Foods and tobacco Textile Timber and furniture Pulp and paper Fuel processing Chemicals Non-metallic mineral products Smelting and pressing of metals Metal products General and special purpose machinery Transport equipment Electrical machinery and equipment Production and supply Other manufactures Total

14.30 0.09 −0.02 0.04 0.02 2.10 −1.19 1.44 0.86 0.13 −0.19 0.15 0.02 6.30 0.06 24.13

−24.51 −2.03 −1.03 0.14 −0.03 −23.05 0.60 −0.88 5.66 0.07 −0.30 −0.02 −0.10 −0.88 −0.09 −46.43

−4.36 −6.05 1.08 0.04 −0.08 44.70 −3.84 −2.60 −9.55 −0.11 −0.10 0.06 0.00 −6.12 −0.05 13.00

Table 4 Attribution results of industrial structure effect in Xinjiang (Unit: %). Sub-sectors

2000–2002

2002–2008

2008–2014

Mining and quarrying Foods and tobacco Textile Timber and furniture Pulp and paper Fuel processing Chemicals Non-metallic mineral products Smelting and pressing of metals Metal products General and special purpose machinery Transport equipment Electrical machinery and equipment Production and supply Other manufactures Total

0.79 −0.53 0.23 0.00 0.00 −6.44 −0.23 −0.30 −1.39 0.01 0.00 0.00 −0.01 −1.84 −0.02 −9.73

−0.89 0.55 0.24 0.04 0.19 6.46 −0.18 1.31 1.00 0.01 0.05 0.03 0.01 1.90 −0.01 10.68

−1.32 0.95 −0.16 −0.05 −0.07 −8.40 5.82 1.71 8.45 0.04 −0.01 −0.02 0.03 13.95 −0.01 20.92

decoupling process. The energy structure of these industrial sub-sectors had been deteriorated over the period 2008–2014, and the proportion that coal accounting for had been also increasing faster. Therefore, controlling the proportion of added value of fuel processing, textile, and mining and quarrying will be of great significance for decoupling relationship between industrial growth and carbon emissions. The attribution analysis of energy intensity effect was conducted applying Eqs. (6) and (B.9) in Appendix B, and the results were shown in Table 3. Energy intensity was the primary factor leading to the 5

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

changes in industrial carbon intensity over 2000–2014. As mentioned in sub-sector 3.2, only in the second stage, energy intensity promoted the decreases in carbon intensity. According to Table 3, most industrial sub-sectors showed positive effect on the decreases during the second stage; mining and quarrying, and fuel processing were the primary contributors to the decreases; only four industrial sub-sectors negatively influenced the changes in energy intensity, and smelting and pressing of metals contributed most to the negative effect. However, energy intensity effect contributed to the increases in the industrial carbon intensity during the third stage; fuel processing was mainly responsible for the increases, and its influence was much than other sub-sectors after 2008. Based on the rich energy resources, Xinjiang’s fuel processing grew rapidly. Comparing the growth rates of its output value and energy consumption, it can observe a 5.89% annual growth rate in output value by fuel processing after 2008, and its corresponding annual growth rate in energy consumption was 17.28%. It can indicate that energy efficiency was not paid enough attention while the fuel processing grew rapidly. Improving the energy efficiency of fuel processing is very important to hinder the energy intensity’s negative influence on decoupling process between industrial growth and carbon emissions. The attribution analysis of energy intensity effect was conducted applying Eqs. (5) and (6), and the results were shown in Table 4. Industrial structure showed postive effect on decoupling process, but it changed from positive effect in the first stage to negative effect in the second and third stages. According to Table 4, during the second stage, fuel processing, production and supply, and non-metallic mineral products were the top three industrial sub-sectors contributing to the changes in industrial structure effect; only three industrial sub-sectors hindered the changes in industrial structure effect. And during the third stage, industrial structure effect had become the most influential factor. Production and supply, smelting and pressing of metals, and chemicals were the primary industrial sub-sectors contributing to the result. Since 2002, some national policies, e.g. Western Development, had brought about a significant investment in some sub-sectors (Zhang et al., 2017b) such as production and supply and chemicals, leading to significant increases in the output of these sub-sectors. Meanwhile, with the industrial restructuring, the proportion of some sub-sectors’ output decreased, so some sub-sectors’ influence also changed, such as fuel processing. Thus, controlling the growth rates and the proportion of some sub-sectors’ output will be of great significance to decoupling relationship between industrial growth and carbon emissions. Production and supply, smelting and pressing of metals, and chemicals can be the key sectors to be controlled.

converting the resource advantages into economic advantages, and technological improvements in the resource- and energy- intensive subsectors are the key to realize the sustainable and low-carbon development in Xinjiang. According to previous studies, energy intensity in many other regions in China (Li et al., 2016; Lu et al., 2015; Song et al., 2015) showed decrease trend and hindered the increases in carbon emissions. While Xinjiang’s industrial energy intensity promoted carbon emissions in recent years, which was different with many other regions in China. Weak economic foundation might be one of the determinants. In comparison with most regions in China, Xinjiang is relatively undeveloped. After 2010, the investment environment improved, and the local government increased more investment to strengthen the weak economic foundation. Although large-scale investments drove rapid growth of the industrial sectoral output, outdated industrial sub-sectors and production processes were also common. Poor technological level was obstructive to realize the sustainable and low-carbon development of Xinjiang’s industrial sector. Xinjiang’s energy intensity was relatively higher, so using new technology to improve energy efficiency was very effective way to decrease carbon emissions. Energy structure showed relatively less influence than other two decomposed factors, which is consistent with many previous studies such as (Wang and Feng, 2017; Wang et al., 2013). The proportional consumption of coal had been maintaining high level, and showed an increasing trend after 2008. Decreasing the proportion of coal in energy consumption structure was a challenge for carbon reduction. As the national energy base, Xinjiang is rich in not only coal but also petroleum and natural gas. Besides, Xinjiang’s wind and solar energy is also abundant (Fan et al., 2017; Ma et al., 2013), but the proportional consumption of clean energy is still in low level. Petroleum, natural gas, and clean energy are helpful to change the decoupling relationships between industrial growth and carbon emissions. Due to the abundance of energy types, Xinjiang has enormous potential and advantages to diversify its energy structure, and it is an effective to reduce carbon emissions. After 2008, industrial structure had been the most influential decomposed factors. Changes in industrial structure always led to the increases in carbon emissions in many less developed regions which were located in western China (Du et al., 2017). Based on the factor endowment and natural resources, less developed economy promoted Xinjiang to preferentially develop resource- and energy-intensive industrial sub-sectors such as production and supply and smelting and pressing of metals. Xinjiang’s industry highly relies on energy-intensive industrial sub-sectors which is unfavorable for decoupling process between industrial growth and carbon emissions. Over the study period, fuel processing, smelting and pressing of metal, and mining and quarrying were the main industrial sub-sectors contributing to the changes in industrial carbon intensity. Mining and quarrying showed negative influence on the changes in effects of energy structure, energy intensity, and industrial structure. Xinjiang is rich in mineral resources, which favors the development of mining and quarrying. Over the period, although the production capacity of mining and quarrying had been increasing, its added value shares decreased. In the meantime, energy efficiency had also been paid attention to improve. In the first and second stages, about 12% reduction in coal consumption. Fuel processing, and smelting and pressing of metals also belong to energyintensive industries. Over the period 2000–2014, about 49.59% of the industrial energy was consumed by these two industrial sub-sectors, but only 15.24% of the added value was created by them. The energy structure of fuel processing, and smelting and pressing of metals was not optimized. And the proportion of fuel processing’ added value decreased, and the energy efficiency was not improved. The added value share of smelting and pressing of metals increased with the improvement of energy efficiency. In the third stage, fuel processing was the only industrial subsector contributing most to the changes in the three decomposed factors. Its energy structure and energy intensity contribute to the

4. Discussion After 2008, Xinjiang’s decoupling relationship between industrial growth and carbon emissions experienced negative decoupling stage, which indicated that Xinjiang’s industries were in unsustainable development. With the support of some preferential policies, e.g. Xinjiang Aid Policy, Xinjiang’s urbanization and industrialization levels and investment are expected to grow rapidly (Wang et al., 2017b). Especially after 2010, Xinjiang was defined as the national energy strategic base, and the central government began to carry on large-scale aid to Xinjiang again. Besides, with the implement of the Belt and Road, new opportunity was also faced by Xinjiang. Due to its rich resources, resource- and energy-intensive sub-sectors had the resource and cost advantages (Huo et al., 2015). The proportion of energy-intensive industrial sub-sectors, e.g. fuel processing, possessing the characteristic of high-carbon emissions and high energy consumption, in total industrial added value became larger. Besides, industrial demand for fossil-energy increased rapidly over the period 2008–2014, that is, the energy efficiency was not improved with the development of industries. This further indicated that energy intensity negatively influenced the decoupling process, and industrial structure effect was the most influential factor to hinder the decoupling process over 2008–2014. Thus, 6

Ecological Indicators 97 (2019) 1–9

X. Zhang et al.

increases in the industrial carbon intensity. Therefore, fuel processing should be paid attention to optimize its energy structure and improve its energy efficiency.

growth and carbon emissions: Adjustment of industrial structure is urgently required. Xinjiang needs to change the industrial growth pattern and applies innovative technologies to upgrade its industrial sub-sectors. Backward enterprises with high higher energy consumption should be restricted by the means of tax policies. And subsidy policies can be also applied to support the enterprises that can effectively consume fossil energy or clean energy. High-tech industries should be given priority to develop. The enterprises, belonging to production and supply, smelting and pressing of metals, and chemicals, are the major objects to be controlled. Xinjiang can give priority to introduce enterprises involving in new energy resources, electricity generation, modern chemical manufacturing, highend equipment manufacturing, and new materials. Industrial energy structure should be optimized. Xinjiang is also very rich in oil, natural gas and other renewable energy, which is the basis for Xinjiang’s industry to optimize its energy structure. Besides increasing the proportional consumption of oil and natural gas, renewable energy should be simultaneously expanded. The development and distribution of wind and solar energy should make comprehensive plans to increase their usage. Large-scale industrial parks can be planned, and renewable energy systems can be established. The enterprises, belonging to fuel processing, textile, and mining and quarrying, should be taken as the key industrial sub-sectors to optimize their energy structure. Energy efficiency should be improved. Improving the industrial efficiency is an effective measure. Xinjiang can persistently eliminate some backward enterprises, or introduce advanced technology to update the backward production capacities. Introducing clean technology to improve the energy efficiency is also effective. Besides, some mandatory energy-saving or carbon reduction policies should also be implemented. Differentiated policies should be made for different industrial sub-sectors. For the sub-sectors that had improved their energy efficiency, e.g. mining and quarrying, the original policies can be encouraged. For the sub-sectors that energy efficiency decreased rapidly, e.g. fuel processing, more innovative policies need to be made combing their own features.

5. Main conclusions and policy recommendations 5.1. Main conclusions This study examined the decoupling relationships between industrial growth and carbon emissions in Xinjiang. Applying the multiplicative Sato–Vartia LMDI method, the changes in industrial carbon intensity was decomposed into four driving factors. And the percent changes of decomposed factors were attributed to 15 industrial subsectors in order to obtain each sub-sector’s contribution. According to the analysis of decoupling relationships, we found that Xinjiang’s decoupling relationships between industrial growth and carbon emissions experienced three stages, i.e., “negative decoupling, weak decoupling, and negative decoupling”. This finding implies that Xinjiang’s industrial sector was in relatively unsustainable period after 2008. The growth of industrial added value caused much damage to environment. The local government should take responsibility for implementing the carbon reduction policies in a more effective way. The decomposition analysis showed that energy intensity, industrial structure and energy structure inhibited the decoupling process after 2008. In the process of industrial development, Xinjiang paid not enough attention to increase the energy efficiency, optimize the industrial structure and energy structure. Attribution analysis can help us make targeted recommendations. According to attribution analysis, the primary sub-sectors, contributing to the changes in different driving factors, had been changing in different stages. After 2008, fuel processing, textile, and mining and quarrying became the primary industrial sub-sectors drove the energy intensity to negatively influence the decoupling process; fuel processing was mainly responsible for energy intensity’s negative influence on the decoupling process; production and supply, smelting and pressing of metals, and chemicals were the primary industrial sub-sectors contribute to industrial structure’s negative influence. These sub-sectors were the targeted sub-sectors to make corresponding recommendations.

Funding sources This work was supported by the National Natural Science Fund of China (Nos. 41371518, 41501144). And the Ordinary University Graduate Scientific Research Innovation Projects of Jiangsu Province (NO. KYLX16_1272).

5.2. Policy recommendations Based on the above findings, the following policy recommendations were put forward to promote the decoupling process between industrial Appendix A

Table A.1 Classification standards on the decoupling relationships. Decoupling states

ΔC

ΔG

cg

Decoupling

Strong decoupling Weak decoupling Recessionary decoupling

<0 >0 <0

>0 >0 <0

<0 <0 >0

Negative decoupling

Strong negative decoupling Weak negative decoupling Expansive negative decoupling

>0 <0 >0

<0 <0 >0

<0 <0 >0

Appendix B The equations, applying the Sato-Vartia LMDI method to estimate effect of each decomposed factor, are given as follows:

7

Ecological Indicators 97 (2019) 1–9

X. Zhang et al. I

J

WijS

V

WijS

V

WijS

V

WijS

V

DCE = EXP

ln

i=1 j=1

I

J

DES = EXP I

J

ESij, t

ln

EIi, t EIi, t 1

ln

ISi, t ISi, t 1

i=1 j=1

I

J

DIS = EXP i=1 j =1

WijS

V

(B.1)

1

ESij, t

ln

i=1 j=1

DEI = EXP

CEij, t CEij, t

(B.2)

1

(B.3)

(B.4)

L (Cij, t / Ct , Cij, t 1/ Ct 1 )

=

I i=1

J j=1

L (Cij, t /Ct , Cij, t 1/Ct 1 )

(B.5)

S-V

In Eq. (B.5), Wij is the normalized weight of energy type j in industrial sub-sector i. L(x,y)=(y-x)/ln(x/y), and x,y > 0, x ≠ y. When examining the cumulative effect of each decomposed factor, it can be obtained applying the following equations, and z denotes each decomposed factor:

(Dz )0, T = (Dz )0,1 × (Dz )1,2···(Dz )T

(B.6)

1, T

The contributions of each decomposed factor attribute to the percentage changes in industrial carbon intensity can be assessed by means of attribution analysis model (González and Martínez, 2012). The attribution analysis model is expressed as follows: I t 1, t DCE

J

I t 1, t rCE , ij

1=

=

i=1 j=1

I t 1, t DES

i=1 j=1

J

I t 1, t rES , ij

1=

I

J

=

i=1 j=1

i=1 j=1

J

t 1, t DEI 1=

J

I

J

reit , ij1, t = i=1 j=1

t−1,t

where D

i=1 j=1

wijS V t 1, t , CE t ) L (CEijt 1 DCE ij

CEijt

1

wijS V

CEijt

N CEijt 1 i L (CE t 1 D t 1, t , CE t ) ij ij CE

wijS V t 1, t , ES t ) L (ESijt 1 DES ij

ESijt

1

wijS V N ESijt 1 i L (ES t 1 D t 1, t , ES t ) ij ij ES

wijS V t 1 t 1, t , EI t ) EIi t L (EIi 1 DEI i wijS V N EI t 1 i L (EI t 1 D t 1, t , EI t ) i i EI i

CEijt

(B.7)

ESijt ESijt

1

1

1

1 (B.8)

EIit 1 EIit 1

-1 is percentage changes of each relative driving factor;

(B.9) J r t 1, t j = 1 IS, ij

is the contribution of industrial sub-sector i from year t-1 to year t.

Appendix C Table C.1 Classification of Xinjiang’s industrial sub-sectors. 15 sub-sectors Mining and quarrying Foods and tobacco Textile Timber and furniture Pulp and paper Fuel processing Chemicals Non-metallic mineral products Smelting and pressing of metals Metal products General and special purpose machinery Transport equipment Electrical machinery and equipment Production and supply Other manufactures

Mining and Washing of Coal; Extraction of Petroleum and Natural Gas; Mining and Processing of Nonferrous Metals Ores; Mining and Processing of Nonmetal Ores; Mining Activities; Mining and Processing of Ferrous Metals Ores; Manufacture of Food; Manufacture of Beverage; Manufacture of Tobacco Manufacture of Textile; Manufacture of Textile Wearing Apparel, Footwear and Caps; Leather, Fur, Feather and Related Products Manufacturing Processing of Timber, Cane, Wood, Bamboo, Grass Products; Manufacture of Furniture Printing and Copying of Medium for Record; Manufacture of Paper and Paper Products; Manufacture of Articles for Culture, Education, Sports and Entertainment Oil Processing, Coking and Nuclear Fuel Processing Manufacture of Medicine; Raw Chemical Material and Chemical Products; Manufacture of Chemical Fiber; Manufacture of Rubber Products Manufacture of Nonmetal Mineral Products Smelting and Pressing of Ferrous Metals; Smelting and Pressing of Nonferrous Metals Manufacture of Metal Products Manufacture of General Purpose Machinery; Manufacture of Special Purpose Machinery Manufacture of Automobile; Manufacture of Railroads, Ships, Aerospace and Other Transportation Manufacture of Electric Equipment and Machinery Production and Supply of Electricity and Thermal; Production and Supply of Gas; Production and Supply of Water Others

8

X. Zhang et al.

Ecological Indicators 97 (2019) 1–9

References

changes: a case study for Shanghai (China). Renew. Sustain. Energy Rev. 55, 516–536. Shao, S., Liu, J., Geng, Y., Miao, Z., Yang, Y., 2016b. Uncovering driving factors of carbon emissions from China’s mining sector. Appl. Energy 166, 220–238. Song, M., Guo, X., Wu, K., Wang, G., 2015. Driving effect analysis of energy-consumption carbon emissions in the Yangtze River Delta region. J. Cleaner Prod. 103, 620–628. Su, B., Ang, B.W., 2014. Attribution of changes in the generalized Fisher index with application to embodied emission studies. Energy 69, 778–786. Tan, Z., Li, L., Wang, J., Wang, J., 2011. Examining the driving forces for improving China’s CO2 emission intensity using the decomposing method. Appl. Energy 88, 4496–4504. Tapio, P., 2005. Towards a theory of decoupling: degrees of decoupling in the EU and the case of road traffic in Finland between 1970 and 2001. Transp. Policy 12, 137–151. Wang, C., Wang, F., 2015. Structural decomposition analysis of carbon emissions and policy recommendations for energy sustainability in Xinjiang. Sustainability 7, 7548–7567. Wang, C., Wang, F., Zhang, H., Ye, Y., Wu, Q., Su, Y., 2014. Carbon emissions decomposition and environmental mitigation policy recommendations for sustainable development in shandong province. Sustainability 6, 8164–8179. Wang, C., Wang, F., 2017. China can lead on climate change. Science 6353, 764. Wang, C., Wang, F., Zhang, X., Yang, Y., Su, Y., Ye, Y., Zhang, H., 2017a. Examining the driving factors of energy related carbon emissions using the extended STIRPAT model based on IPAT identity in Xinjiang. Renew. Sustain. Energy Rev. 67, 51–61. Wang, C., Wang, F., Zhang, X., Zhang, H., 2017b. Influencing mechanism of energy-related carbon emissions in Xinjiang based on the input-output and structural decomposition analysis. J. Geog. Sci. 27, 365–384. Wang, J., Zhao, T., Xu, X., Zhang, X., 2016a. Exploring the changes of energy-related carbon intensity in China: an extended Divisia index decomposition. Nat. Hazards 83, 501–521. Wang, M., Feng, C., 2017. Decomposition of energy-related CO2 emissions in China: an empirical analysis based on provincial panel data of three sectors. Appl. Energy 190, 772–787. Wang, Q., Hang, Y., Zhou, P., Wang, Y., 2016b. Decoupling and attribution analysis of industrial carbon emissions in Taiwan. Energy 728–738. Wang, W., Liu, R., Zhang, M., Li, H., 2013. Decomposing the decoupling of energy-related CO2 emissions and economic growth in Jiangsu Province. Energy Sustain. Dev. 17, 62–71. Wang, Z., Yang, L., 2015. Delinking indicators on regional industry development and carbon emissions: Beijing–Tianjin–Hebei economic band case. Ecol. Ind. 48, 41–48. Xu, S.C., He, Z.X., Long, R.Y., Chen, H., 2016. Factors that influence carbon emissions due to energy consumption based on different stages and sectors in China. J. Cleaner Prod. 115, 139–148. Xu, X.Y., Ang, B.W., 2013. Index decomposition analysis applied to CO2 emission studies. Ecol. Econ. 93, 313–329. Yu, Y., Zhou, L., Zhou, W., Ren, H., Kharrazi, A., Ma, T., Zhu, B., 2017. Decoupling environmental pressure from economic growth on city level: the case study of chongqing in China. Ecol. Ind. 75, 27–35. Zhang, X.L., Wang, Q., Wang, C.J., Lu, J.R., 2013. Analyses of development and industrial relevancy of energy industrial in Xinjiang. J. Univ. Chinese Acad. Sci. 4, 504–509 (In Chinese). Zhang, X., Zhao, X., Jiang, Z., Shao, S., 2017a. How to achieve the 2030 CO2 emissionreduction targets for China's industrial sector: retrospective decomposition and prospective trajectories. Global Environ. Change 44, 83–97. Zhang, X.L., Zhao, Y., Sun, Q., Wang, C., 2017b. Decomposition and attribution analysis of industrial carbon intensity changes in Xinjiang, China. Sustainability 9, 459–475. Zhang, X., Zhao, Y., 2018. Identification of the driving factors' influences on regional energy-related carbon emissions in China based on geographical detector method. Environ. Sci. Pollut. Res. 10, 9626–9635. Zhang, Y.J., Da, Y.B., 2015. The decomposition of energy-related carbon emission and its decoupling with economic growth in China. Renew. Sustain. Energy Rev. 41, 1255–1266. Zhang, Y., Wang, H., Liang, S., Xu, M., Liu, W., Li, S., Zhang, R., Nielsen, C.P., Bi, J., 2014. Temporal and spatial variations in consumption-based carbon dioxide emissions in China. Renew. Sustain. Energy Rev. 40, 60–68. Zhou, X., Zhang, M., Zhou, M., Zhou, M., 2017. A comparative study on decoupling relationship and influence factors between China's regional economic development and industrial energy–related carbon emissions. J. Cleaner Prod. 142, 783–800. Zhao, X., Zhang, X., Shao, S., 2016. Decoupling CO2 emissions and industrial growth in China over 1993–2013: The role of investment. Energy Econ. 60, 275–292. Zhao, X., Zhang, X., Li, N., Shao, S., Geng, Y., 2017. Decoupling economic growth from carbon dioxide emissions in China: a sectoral factor decomposition analysis. J. Cleaner Prod. 142, 3500–3516.

Bruyn, S.M.D., Opschoor, J.B., 1997. Developments in the throughput-income relationship: theoretical and empirical observations. Ecol. Econ. 20, 255–268. Chen, B., Yang, Q., Li, J.S., Chen, G.Q., 2017. Decoupling analysis on energy consumption, embodied GHG emissions and economic growth – the case study of Macao. Renew. Sustain. Energy Rev. 67, 662–672. Cheng, J., 2016. Decomposition of carbon cap targets at provincial level in China: a case study of Zhejiang. China Population. Resourc. Environ. 1, 23–30 (In Chinese). Choi, K.H., Ang, B.W., 2012. Attribution of changes in Divisia real energy intensity index – an extension to index decomposition analysis. Energy Econ. 34, 171–176. Choi, K.-H., Oh, W., 2014. Extended Divisia index decomposition of changes in energy intensity: a case of Korean manufacturing industry. Energy Policy 65, 275–283. Climent, F., Pardo, A., 2007. Decoupling factors on the energy–output linkage: the Spanish case. Energy Policy 35, 522–528. Conte Grand, M., 2016. Carbon emission targets and decoupling indicators. Ecol. Ind. 67, 649–656. David, M., 2014. China's peak carbon pledge raises pointed questions. Scie. Sci. 346, 903. de Freitas, L.C., Kaneko, S., 2011. Decomposing the decoupling of CO2 emissions and economic growth in Brazil. Ecol. Econ. 70, 1459–1469. Dong, B., Zhang, M., Mu, H., Su, X., 2016. Study on decoupling analysis between energy consumption and economic growth in Liaoning Province. Energy Policy 97, 414–420. Du, K., Xie, C., Ouyang, X., 2017. A comparison of carbon dioxide (CO2) emission trends among provinces in China. Renew. Sustain. Energy Rev. 73, 19–25. Fan, X.C., Wang, W.Q., Shi, R.J., Cheng, Z.J., 2017. Hybrid pluripotent coupling system with wind and photovoltaic-hydrogen energy storage and the coal chemical industry in Hami, Xinjiang. Renew. Sustain. Energy Rev. 72, 950–960. Fernández González, P., 2015. Exploring energy efficiency in several European countries. An attribution analysis of the Divisia structural change index. Appl. Energy 137, 364–374. Fernández González, P., Landajo, M., Presno, M.J., 2013. The Divisia real energy intensity indices: Evolution and attribution of percent changes in 20 European countries from 1995 to 2010. Energy 58, 340–349. Fernández González, P., Presno, M.J., Landajo, M., 2015. Regional and sectoral attribution to percentage changes in the European Divisia carbonization index. Renew. Sustain. Energy Rev. 52, 1437–1452. González, D., Martínez, M., 2012. Changes in CO2 emission intensities in the Mexican industry. Energy Policy 51, 149–163. Huo, J., Yang, D., Zhang, W., Wang, F., Wang, G., Fu, Q., 2015. Analysis of influencing factors of CO2 emissions in Xinjiang under the context of different policies. Environ. Sci. Policy 45, 20–29. Jiang, X., Dong, J., Wang, X., Li, R., 2016. The multilevel index decomposition of energyrelated carbon emission and its decoupling with economic growth in USA. Sustainability 8, 857. Li, Q., Wei, Y.-N., Dong, Y., 2016. Coupling analysis of China’s urbanization and carbon emissions: example from Hubei Province. Nat. Hazards 81, 1333–1348. Liu, N., Ma, Z., Kang, J., 2015a. Changes in carbon intensity in China's industrial sector: decomposition and attribution analysis. Energy Policy 87, 28–38. Liu, Z., Geng, Y., Lindner, S., Guan, D., 2012. Uncovering China’s greenhouse gas emission from regional and sectoral perspectives. Energy 45, 1059–1068. Liu, Z., Guan, D., Moore, S., Lee, H., Su, J., Zhang, Q., 2015b. Steps to China's Carbon Peak. Nature 279–281. Lu, Q., Yang, H., Huang, X., Chuai, X., Wu, C., 2015. Multi-sectoral decomposition in decoupling industrial growth from carbon emissions in the developed Jiangsu Province, China. Energy 82, 414–425. Lv, K.W., Miao, C.-H., Shang, W.-Y., 2012. Sectoral difference in Carbon emission of industrial energy consumption: a case study of Henan province. Econ. Geogr. 12, 15–20 (In Chinese). Ma, Z., Xue, B., Geng, Y., Ren, W., Fujita, T., Zhang, Z., Oliveira, J.A.P.D., Jacques, D.A., Xi, F., 2013. Co-benefits analysis on climate change and environmental effects of wind-power: a case study from Xinjiang, China. Renew. Energy 57, 35–42. OECD, 2005. Decoupling: a conceptual overview. Oecd Papers 5, 37–38. Pang, J., Chen, X., Wang, H., 2014. Relationship of energy consumption with economic growth in Gansu province. J. Arid Land Resourc. Environ. 28, 31–36. Qiu, J., 2009. China's climate target: is it achievable? Nature 462, 550–551. Roinioti, A., Koroneos, C., 2017. The decomposition of CO2 emissions from energy use in Greece before and during the economic crisis and their decoupling from economic growth. Renew. Sustain. Energy Rev. 76, 448–459. Rose, A., Casler, S., 1996. Input-output structural decomposition analysis: a critical appraisal. Econ. Syst. Res. 8, 33–62. Shao, S., Yang, L., Gan, C., Cao, J., Geng, Y., Guan, D., 2016a. Using an extended LMDI model to explore techno-economic drivers of energy-related industrial CO2 emission

9