Uncovering CO₂ emission drivers under regional industrial transfer in China’s Yangtze River Economic Belt: a multi-layer LMDI decomposition analysis

Abstract

With the relocation of heavy industries moving from downstream region to upstream and midstream regions in the Yangtze River Economic Belt (YREB), it is critical to encourage coordinated low carbon development in different regions within the YREB. This paper uncovers the evolution of CO₂ emissions in different regions within the YREB for the period of 2000–2017. It decomposes regional CO₂ emission changes using the temporal and cross-regional three-layer logarithmic mean Divisia index (LMDI) method. Besides, it decomposes industrial CO₂ emission changes using the temporal two-layer LMDI method. The research results show that economic growth is the major driver for regional CO₂ emission disparities. The mitigation drivers, such as energy intensity and energy structure, lead to a more decreased CO₂ emission in the downstream region than in the upstream and midstream regions. In addition, it proposes several policy recommendations based upon the local realities, including improving energy efficiency, optimizing energy structure, promoting advanced technologies and equipment transfers, and coordinating the development in the upstream, midstream and downstream regions within the YREB.

Graphical abstract

Keywords

CO₂ emission / multi-layer LMDI decomposition / industrial transfer / governance

1 Introduction

With rapid industrialization and urbanization, China has become one of the world’s largest energy consumer and carbon emissions (CO₂) emitter [1]. To address both energy crisis and reduce the overall CO₂ emission, China has taken various measures, such as the wide application of renewable and sustainable energy, national energy efficiency efforts and industrial updates [2]. Besides, it has released national mitigation targets based upon local situations, such as the CO₂ emissions intensity and the energy intensity targets in the Five-Year Plans (FYPs) [3,4]. However, China is a very large country with diversified regional development, leading to the fact that industrial structure should be considered when setting CO₂ emission reduction targets at regional level. In particular, the Chinese government should actively encourage industrial structure updates so that those energy intensive and inefficient industries can be gradually phased out and green transformation of China’s economy can be achieved [5]. In reality, with China’s rapid development, many coastal regions have relocated their polluting businesses to inland and the western regions due to their increasingly strict environmental enforcements. Unfortunately, some inland and western regions simply accept those dirty industries in order to boom their economy, which further deteriorated local ecosystems [6]. It is therefore critical to identify differences in CO₂ emission drivers in different regions under regional industrial transfer, so that appropriate region-specific policies can be prepared to achieve the national CO₂ emission-reduction target.

Academically, studies on CO₂ emissions reduction have attracted widespread attention around the world due to the need to address global climate change. Recent studies focus on estimating global and national emissions and their evolution [7–10], identifying driving forces [11–14], and exploring mitigation pathways and policies [15–17]. In China, existing studies in this field center on driving force analysis and policy interventions [18–22]. In particular, studies on CO₂ emission drivers have been conducted at different scales, including national level [23,24], regional level [25–28], provincial level [29–31], and city level [32–34], as well as various industrial sectors [35–37]. However, these studies mainly focus on one single area or a particular industrial sector. Few studies have been conducted to combine both regional and industrial structure disparities.

Within China, the Yangtze River Economic Belt (YREB), which encompasses 11 provinces and provincial-level municipalities as an indispensable economic and ecologically fragile zone, is faced with challenges such as increasing CO₂ emissions, backward industrial structure, and ineffective regional cooperation [38]. According to the 13th FYP from 2016–2020, initiatives that promote the coordinated development of upstream, midstream, and downstream regions in the YREB should be undertaken to control the overall CO₂ emission [4]. However, only a few studies have discussed CO₂ emissions in the YREB. For example, Ren et al. [39] applied one meta-frontier dynamic slack-based measures (SBM) model and selected CO₂ emission as one evaluation indicator for measuring environmental efficiency to examine the relationship between economic development, energy input and output, and environmental effects between the YREB and the non-YREB regions. Tian et al. [40] compared the carbon emissions efficiency in the YREB with that in the non-YREB by using a modified data envelopment analysis (DEA) model and found that technology gaps and carbon emission efficiency of the provinces and cities within the YREB are significantly lower than those within the non-YREB. Tang et al. [41] uncovered the spatiotemporal disparities and driving factors of CO₂ emissions in the YREB by using a STIRPAT model and found that the per capita gross domestic product (GDP) and energy intensity have the largest positive effects on increasing CO₂ emissions, while urbanization rate has the largest negative effect on CO₂ emissions. In summary, these studies focus on the entire YREB without investigating industry-level CO₂ emissions within the YREB. Geographically, the YREB is divided into upstream, midstream, and downstream regions. In reality, the midstream region of the YREB is responsible for the connection between the upstream and downstream regions during the industrial transfer process. However, due to different resource endowments, economic development level and cultural background, different provinces are taking different measures to respond to such an industrial transfer and climate change. A rational regional cooperation mechanism is yet to be established. Consequently, it is crucial to identify the key driving forces on CO₂ emissions in different provinces and sectors within the YREB so that more appropriate mitigation policies can be proposed.

Methodologically, major research methods on CO₂ emission drivers include index decomposition analysis (IDA) [20,21,42], structural decomposition analysis (SDA) [43,44], econometric analysis [27,45], and DEA [46,47]. The factor decomposition methods are widely used in energy and environmental analysis, which help to depict the CO₂ emissions under different scenarios and to identify driving factors responsible for emissions changes. The SDA method has its advantages in analyzing detailed sectoral emissions, but requires a complete input-output table of one country or region [43]. In contrast, the IDA method has a low data requirement and is therefore more flexible. As an index decomposition method, the logarithmic mean Divisia index (LMDI) method can be applied with any available data at any level of aggregation, which has been regarded as the most appropriate method for energy and environmental decomposition analysis, due to its incomparable advantages in theoretical foundation, adaptability, ease of use, and result interpretation [42]. Especially, the multi-layer LMDI method enables the pre-defined driving forces to be first quantified at a level of sub-sectors, which are categorized according to sectors, regions or fuel types; and then at a more aggregated level, such as national scale [21]. Such a decomposition method is accurate in decomposition and consistent in aggregation. Therefore, a two-layer LMDI decomposition method and a three-layer LMDI decomposition method are employed in this paper to fit the two-dimensional and three-dimensional data sources.

Within the overall context, this paper fills the research gap by conducting a multi-layer LMDI decomposition analysis on the CO₂ emission changes of the YREB from both regional and industrial perspectives. In particular, this paper focuses on identifying key driving forces in different FYP periods, seeking to recognize regional disparity of CO₂ emissions in the upstream, midstream, and downstream regions of the YREB. The contributions of this paper are twofold. Besides, it has investigated the driving forces of CO₂ emissions changes in the YREB, including upstream, midstream, and downstream regions, and identified the regional disparities of driving effects. Hence, useful insights will be provided for the further formulation and implementation of local energy and environmental policies. Moreover it has decomposed the changes in CO₂ emissions into several driving factors by using the multi-layer LMDI model, focusing not only on region-level emissions, but also on industry-level emissions. Thus, this paper can better explain the drivers of CO₂ emissions from multiple dimensions.

2 Methods and data

2.1 Area studied

YREB is an economic entity encompassing 11 provinces and provincial-level municipalities and locates in the eastern, central, and western China. It accounts for 21.4% of the national area and almost 50% of the national economic volume. The complete data of provincial GDP and GDP per capita in the YREB are shown in the Electronic Supplementary Material (ESM, Fig. S1). The YREB is divided into three regions, that is, the upstream, midstream, and downstream regions, which significantly differ in terms of resource endowment, industrial structure, and technological level. The upstream region includes Yunnan, Sichuan, Guizhou provinces, and Chongqing municipality; the midstream region includes Jiangxi, Hubei, and Hunan provinces; and the downstream region includes Jiangsu, Zhejiang, Anhui provinces, and Shanghai municipality. In terms of economic development, the downstream region is most developed, followed by the midstream region, and the upstream region.

Green development has been proposed by to accelerate China’s industrial structure transformation [48]. The details of industrial structure changes in the upstream, midstream, and downstream regions of the YREB during 2000–2017 are shown in the ESM (Fig. S2). At present, the entire YREB is faced with several challenges, such as the destructed ecosystem functions, various environmental emissions, resource depletion, and ineffective regional cooperation. Due to its special position in China, the development of the YREB has become a national strategy since 2014 [44]. The Chinese central government issued the “Outline of the YREB Development Plan” in March 2016 [49], which clearly stipulates that the entire YREB should pursue green development. Under this strategic policy, it is crucial to reduce the total CO₂ emission by coordinating regional economic development and optimizing industrial structure.

2.2 Emission accounting method

Based on the emission accounting method of the 2006 IPCC Guidelines for National Greenhouse Gas Inventories [50], this paper calculates the total CO₂ emissions of all 45 sectors by multiplying the energy consumption data of economic sectors by their corresponding emission factors. Equation (1) demonstrates the calculation process of CO₂ emission.

\mathrm{CE}={\displaystyle {\sum }_{m=\mathrm{1}}^{\mathrm{45}}}{\displaystyle {\sum }_{l=\mathrm{1}}^{\mathrm{17}}}{\mathrm{CE}}_{ml}={\displaystyle {\sum }_{m=\mathrm{1}}^{\mathrm{45}}}{\displaystyle {\sum }_{l=\mathrm{1}}^{\mathrm{17}}}A{D}_{ml}\times {\mathrm{NCV}}_m\times {\mathrm{CC}}_m\times O_{ml},

(1)

where CE represents the overall emission; l, the energy type; m, the economic sector; CE _ml, CO₂ emission from energy type l in sector m; AD _ml, the consumption of energy type l in sector m; and NCV _m, CC _m, and O _ml, the net caloric value, carbon content, and the oxygenation efficiency, respectively.

2.3 Decomposition analysis

The LMDI decomposition method has been widely applied to identify the drivers of CO₂ emission changes [6,33,51]. With the expansion of LMDI methods, the application fields have expanded from specific industries to multi-sector and multi-space dimensions [52]. Ang and Liu [53] introduced a “two-layer” LMDI decomposition method at the sub-sector scale based upon sectors, fuel types or regions. Wu et al. [54] further proposed a new “three-layer” LMDI decomposition analysis method to fit three-dimensional data, which addressed six fuel types and six sectors in 28 provinces over the period of 1985–1999. The “three-layer” decomposition analysis method divides all the determinants into three effect groups to highlight their performances, which can provide accurate decomposition results and keep the aggregation consistency. The multi-layer LMDI decomposition method first categorizes the pre-defined driving forces according to sectors, regions, or fuel types, and then quantifies the driving forces at a more aggregated level. Compared to the single-step procedure, the multi-step procedure has the advantage of allowing a better understanding of the change mechanisms within sub-groups and comparisons be made between sub-groups. The property of consistency in aggregation remains valid if the data used in the first step of the multi-step procedure are exactly the same as those used in the single-step procedure [53]. In this paper, the driving factors of the CO₂ emissions in the YREB at the industrial level are uncovered using Ang’s “two-layer” LMDI method and CO₂ emissions at the regional level using Wu’s “three-layer” LMDI method. In addition, in order to identify the driving forces of the disparity of cross-regional CO₂ emissions, a cross-regional decomposition method is applied to investigate the causes of regional disparities in CO₂ emissions in different regions [55–57].

In total, 45 sectors are classified into the primary industry sector (agriculture: AGR), the secondary industry sector (industrial sector: IND and construction: CON), and the tertiary industry sector (service industry: SEI) based on the sector classification proposed in the Industrial Classification Standard for National Economic Activities (GB/T 4754-2017) [58]. In reality, the CO₂ emission generated by industrial sectors accounts for a large proportion of the national CO₂ emission [59]. To evaluate the CO₂ emission changes caused by different industrial sectors, the 40 sub-industrial sectors are divided into four categories according to their dependence on resources, labor, capital, and technologies, including resource-intensive industry (RII), labor-intensive industry (LII), capital-intensive industry (CII), and technology-intensive industry (TII) [60,61]. Hence, the 45 sectors are classified into seven categories. Table S1 in the ESM lists the details of industrial classifications.

2.3.1 Three-layer LMDI decomposition analysis at regional level

From regional perspective, it is necessary to consider more relevant factors. Thus, the total CO₂ emission is decomposed into seven factors, including carbon emission coefficient (CC), energy structure (ES), energy intensity (EI), industrial structure (IS), regional structure (RS), economic growth (GP), and population scale (P). The decomposition analysis is conducted by using Eq. (2).

CE={\displaystyle \sum _{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle \sum _{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle \sum _{k=\mathrm{1}}^{\mathrm{11}}}{\displaystyle \frac{{\mathrm{CE}}_{ijk}}{E_{ijk}}}\times {\displaystyle \frac{E_{ijk}}{E_{ik}}}\times {\displaystyle \frac{E_{ik}}{G_{ik}}}\times {\displaystyle \frac{G_{ik}}{G_k}}{\displaystyle \frac{G_k}{G}}\times {\displaystyle \frac{G}{P}}\times P

(2) ={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}{\mathrm{CC}}_{ijk}\times {\mathrm{ES}}_{ijk}\times {\mathrm{EI}}_{ik}\times {\mathrm{IS}}_{ik}\times {\mathrm{RS}}_k\times GP\times P,

where k denotes the Chinese province/municipality within the YREB (k = 1, 2, …, 11); i, the involved sector (i = 1, 2, …, 7) of AGR, RII, LII, CII, TII, CON, and SEI, respectively; j, the fuel type consumed by each sector (j = 1, 2, 3, representing coal, oil, and gas, respectively). CE _ijk, the CO₂ emission generated from fuel j consumed in sector i of province k; E _ijk, the fuel j consumed in sector i of province k; and G _ik, the GDP of the sector i in province/municipality k.

The indicators of driving forces are detailed as follows: ① {\mathrm{CC}}_{ijk}={\mathrm{CE}}_{ijk}/E_{ijk} represents the carbon emission intensity for fuel j consumed in sector i of province k, reflecting carbon emission coefficient of coal, oil and natural gas. ② {\mathrm{ES}}_{ijk}=E_{ijk}/E_{ik} is the proportion of the fuel j in all fuels consumption in sector i, representing energy structure. ③ {\mathrm{EI}}_{ik}=E_{ik}/G_{ik} represents the fuel consumption per unit GDP in sector i of province k, referring to energy intensity. ④ {\mathrm{IS}}_{ik}=G_{ik}/G_k reflects the proportion of each industrial output to the overall GDP of province k, representing industrial structure. ⑤ {\mathrm{RS}}_k=G_k/G is the share of the GDP of province k in the total regional GDP. This indicator reflects regional differences of economic development and is used to represent regional structure. ⑥ GP=G/P represents per capita GDP, measuring regional economic growth. ⑦ P represents population.

Based on Eq. (2), the changes of regional CO₂ emissions ( {\mathrm{\Delta CE}}^t) from year t to year t^{\prime } can be calculated by using Eq. (3).

{\mathrm{\Delta CE}}^t={\mathrm{CE}}^t-{\mathrm{CE}}^{t^{\prime }}

={\mathrm{\Delta CE}}_{\mathrm{CC}}+{\mathrm{\Delta CE}}_{\mathrm{ES}}+{\mathrm{\Delta CE}}_{\mathrm{EI}}+{\mathrm{\Delta CE}}_{\mathrm{IS}}+{\mathrm{\Delta CE}}_{\mathrm{RS}}+{\mathrm{\Delta CE}}_{GP}+{\mathrm{\Delta CE}}_P

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln ({\mathrm{CC}}^t/{\mathrm{CC}}^{t^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln ({\mathrm{ES}}^t/{\mathrm{ES}}^{t^{\prime }})

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln ({\mathrm{EI}}^t/{\mathrm{EI}}^{t^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln ({\mathrm{IS}}^t/{\mathrm{IS}}^{t^{\prime }})

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln ({\mathrm{RS}}^t/{\mathrm{RS}}^{t^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln (G{P}^t/G{P}^{t^{\prime }})

(3) +{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_{k=\mathrm{1}}^{\mathrm{11}}}L(W_{ijk}^t,W_{ijk}^{t^{\prime }})\ln (P^t/P^{t^{\prime }}),

where L(W_{ijk}^t,W_{ijk}^{t^{\prime }}) is the weight coefficient, which can be defined by Eq. (4).

(4) L(W_{ijk}^t,W_{ijk}^{t^{\prime }})=({\mathrm{CE}}_{ijk}^t-{\mathrm{CE}}_{ijk}^{t^{\prime }})/(\ln ({\mathrm{CE}}_{ijk}^t)-\ln ({\mathrm{CE}}_{ijk}^{t^{\prime }})),

where∆CE_CC denotes the temporal decomposition related carbon emission coefficient effect; ∆CE_ES, the temporal decomposition related energy structure effect; ∆CE_EI, the temporal decomposition related energy intensity effect; ∆CE_IS, the temporal decomposition related industrial structure effect; ∆CE_RS, the temporal decomposition related regional structure effect; ∆CE _GP the temporal decomposition related economic growth effect; and ∆CE _P, the temporal decomposition related population scale effect.

2.3.2 Cross-regional LMDI decomposition analysis

Referring to Ang et al. [62], a cross-regional LMDI method is applied to investigate the drivers of CO₂ emission disparities in various regions in 2017. The CO₂ emission disparities between upstream and downstream regions, midstream and downstream regions, and upstream and midstream regions are decomposed. The CO₂ emission disparity represents the difference between CO₂ emission in region r and that in region r' in the same period. For decomposing the CO₂ emission difference between upstream and downstream regions, the downstream region is chosen as region r and the upstream region as region r'; For decomposing the CO₂ emission difference between midstream and downstream regions, the downstream region is chosen as region r and the midstream region as region r'; For decomposing the CO₂ emission difference between upstream and midstream regions, the midstream region is chosen as region r and the upstream region as region r'. Based on Eq. (2), the CO₂ emission difference between two regions (∆CE ^r) can be decomposed by using Eq. (5).

{\mathrm{\Delta CE}}^r={\mathrm{CE}}^r-{\mathrm{CE}}^{r^{\prime }}

={\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{CC}}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{ES}}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{EI}}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{IS}}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{RS}}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{GP}+{\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_P

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln ({\mathrm{CC}}^r/{\mathrm{CC}}^{r^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln ({\mathrm{ES}}^r/{\mathrm{ES}}^{r^{\prime }})

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln ({\mathrm{EI}}^r/{\mathrm{EI}}^{r^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln ({\mathrm{IS}}^r/{\mathrm{IS}}^{r^{\prime }})

={\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln ({\mathrm{RS}}^r/{\mathrm{RS}}^{r^{\prime }})+{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln (G{P}^r/G{P}^{r^{\prime }})

(5) +{\displaystyle {\sum }_{i=\mathrm{1}}^{\mathrm{7}}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle {\sum }_k^{}}L(W_{ijk}^r,W_{ijk}^{r^{\prime }})\ln (P^r/P^{r^{\prime }}),

where L(W_{ijk}^r,W_{ijk}^{r^{\prime }}) is the weight coefficient, which can be defined by using Eq. (6).

(6) L(W_{ijk}^r,W_{ijk}^{r^{\prime }})={\displaystyle \frac{({\mathrm{CE}}_{ijk}^r-{\mathrm{CE}}_{ijk}^{r^{\prime }})}{(\ln ({\mathrm{CE}}_{ijk}^r)-\ln ({\mathrm{CE}}_{ijk}^{r^{\prime }}))}},

where {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{CC}} denotes the cross-regional decomposition related carbon emission coefficient effect; {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{ES}}, the cross-regional decomposition related energy structure effect; {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{EI}}, the cross-regional decomposition related energy intensity effect; {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{IS}}, the cross-regional decomposition related industrial structure effect; {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{\mathrm{RS}}, the cross-regional decomposition related regional structure effect; {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_{GP}, the cross-regional decomposition related economic growth effect; and {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_P, the cross-regional decomposition related population scale effect.

2.3.3 Two-layer LMDI decomposition analysis at industrial level

From industrial perspective, the influences of four factors are mainly focused on CC, ES, and EI and GP. To fit two-dimensional data, a two-layer LMDI decomposition method is used to uncover the driving forces of CO₂ emission changes generated by different industries. The decomposition result of the total CO₂ emission of each sector (CE ⁱ) can be calculated by using Eq. (7).

(7) {\mathrm{CE}}^i={\displaystyle {\sum }_i}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\displaystyle \frac{{\mathrm{CE}}_{ij}}{E_{ij}}}\times {\displaystyle \frac{E_{ij}}{E_i}}\times {\displaystyle \frac{E_i}{G_i}}\times G_i={\displaystyle {\sum }_i}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}{\mathrm{CC}}_{ij}\times {\mathrm{ES}}_{ij}\times {\mathrm{EI}}_i\times G_i.

The total CO₂ emission change of each sector (∆CE ^it) from year t to year t' can be decomposed by using Eq. (8).

{\mathrm{\Delta CE}}^{it}={\mathrm{CE}}^{it}-{\mathrm{CE}}^{i{t}^{\prime }}

={\mathrm{\Delta CE}}_{\mathrm{CC}}^i+{\mathrm{\Delta CE}}_{\mathrm{ES}}^i+{\mathrm{\Delta CE}}_{\mathrm{EI}}^i+{\mathrm{\Delta CE}}_G^i

={\displaystyle {\sum }_i^{}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}L(W_{ij}^{it},W_{ij}^{i{t}^{\prime }})\ln ({\mathrm{CC}}^{it}/{\mathrm{CC}}^{i{t}^{\prime }})+{\displaystyle {\sum }_i^{}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}L(W_{ij}^{it},W_{ij}^{i{t}^{\prime }})\ln ({\mathrm{ES}}^{it}/{\mathrm{ES}}^{i{t}^{\prime }})

(8) ={\displaystyle {\sum }_i^{}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}L(W_{ij}^{it},W_{ij}^{i{t}^{\prime }})\ln ({\mathrm{EI}}^{it}/{\mathrm{EI}}^{i{t}^{\prime }})+{\displaystyle {\sum }_i^{}}{\displaystyle {\sum }_{j=\mathrm{1}}^{\mathrm{3}}}L(W_{ij}^{it},W_{ij}^{i{t}^{\prime }})\ln (G^{it}/G^{i{t}^{\prime }}),

where L(W_{ijk}^{it},W_{ijk}^{i{t}^{\prime }}) is the weight coefficient, which can be defined by using Eq. (9).

(9) L(W_{ij}^{it},W_{ij}^{i{t}^{\prime }})={\displaystyle \frac{({\mathrm{CE}}_{ij}^{it}-{\mathrm{CE}}_{ij}^{i{t}^{\prime }})}{(\ln ({\mathrm{CE}}_{ij}^{it})-\ln ({\mathrm{CE}}_{ij}^{i{t}^{\prime }}))}},

where {\mathrm{\Delta CE}}_{\mathrm{CC}}^i represents carbon emission coefficient effect of different industries; {\mathrm{\Delta CE}}_{\mathrm{ES}}^i represents energy structure effect of different industries; {\mathrm{\Delta CE}}_{\mathrm{EI}}^i, energy intensity effect of different industries; and {\mathrm{\Delta C}{\mathrm{E}}^{\prime }}_G, economic growth effect of different industries.

2.4 Data sources and treatment

Data accuracy and reliability are important for this paper. To date, different CO₂ emission accounting methods have been used by different agencies [63–66]. Liu and his colleagues [67] found that China’s CO₂ emission was overestimated by 40% in some years because of the inconsistency of activity data and emission factors. There exists uncertainty in energy consumption data between national and provincial levels [63]. Moreover, most scholars use the default emission factors recommended by the Intergovernmental Panel on Climate Change (IPCC), but these emission factors cannot accurately reflect the realities in China [67]. Such factors lead to the overestimation of China’s CO₂ emission by international institutions [68,69]. To avoid such a problem in this paper, activity data and carbon emission factors are traced back to each province, each sector, and different fuel types so that local data can be applied.

The time period in this paper ranges from 2000 to 2017. Since China released different development policies in different FYP periods, and the FYP is a critical part of China’s national economic development plan, the period studied in this paper is divided into 4 sub-periods: 10th FYP, 11th FYP, 12th FYP, and part of 13th FYP. China’s 10th, 11th, 12th, and 13th FYPs refer to the periods of 2001–2005, 2006–2010, 2011–2015, and 2016–2020, respectively. Because the influences of factor changes on CO₂ emissions in different periods are explored, those periods in this paper are 2000–2005, 2005–2010, 2010–2015, and 2015–2017, respectively. The energy consumption data were obtained from the Energy Statistical Yearbooks of each province or municipality [70]. The carbon emission factors were collected from the CEADs website (www.ceads.net), which were estimated by Chinese scholars based on the measurement of 602 coal samples in China’s 100 largest coal mining areas [67]. The economic data were obtained from China Statistical Yearbooks, including GDP and industrial output values from each province and municipality [71]. All economic data were deflated into 2000 constant prices, by using GDP indexes for GDP data and producer price indexes for industrial output values.

According to Shan et al. [72], in order to simplify the accounting process, 17 kinds of fossil fuels are divided into three categories: coal, oil and natural gas. The fossil fuel classifications and emission factors are listed in Table S2 in the ESM. In this paper, the CO₂ emissions generated within all the administrative regions of the YREB are taken into account. The urban and rural consumption sectors are not considered because this paper focuses on the analysis of industrial transformation. The CO₂ emission inventory includes the CO₂ emissions generated by the consumption of 17 kinds of fossil fuels from 45 sectors within the regional boundary, but the embodied emission generated by the imported electric power and thermal energy outside the regional boundary are not considered. Besides, the energy loss, non-energy use and energy consumption generated by processing conversion are removed from the total energy consumption to avoid repeated calculation.

3 Results and discussion

3.1 Trend and disparity of CO₂ emission in the YREB

Figure 1 presents the spatial distribution and growth of CO₂ emissions in the YREB. It is clearly observed that the CO₂ emissions in the YREB significantly increased from 2000 to 2017. During the 10th FYP period, the CO₂ emission in the YREB increased from 1.08 Gt to 1.83 Gt at an average annual rate of 11.1%, representing the fastest growth stage. During the 11th FYP period, the CO₂ emission steadily increased to 2.69 Gt with an average annual growth of 8.07%. During the 12th FYP period, the CO₂ emission increased slowly to 2.98 Gt, with an annual growth rate of only 2.07%. Finally, the CO₂ emission increased to 3.19 Gt in 2017.

From a spatial distribution perspective, it is seen that the CO₂ emissions in the upstream, midstream, and downstream regions are unevenly distributed, with a clear regional disparity. In the downstream region, the CO₂ emission increased from 0.54 Gt in 2000 to 0.90 Gt in 2005, followed by the increases to 1.28 Gt in 2010 and 1.49 Gt in 2015, and finally increased to 1.59 Gt in 2017. In the midstream region, the CO₂ emission increased to 0.43 Gt and 0.66 Gt in 2005 and 2010, respectively. It then decreased from 0.75 Gt in 2011 to 0.68 Gt in 2014, and finally increased to 0.82 Gt in 2017. In the upstream region, the CO₂ emission first increased from 0.29 Gt in 2000 to 0.50 Gt in 2005, and then continued to increase to 0.84 Gt in 2012, followed by the decrease to 0.80 Gt in 2017. The CO₂ emissions in the three regions increased rapidly during the 10th FYP and 11th FYP periods and then gradually decreased during the 12th FYP. The downstream region is the main CO₂ emission region, accounting for almost 50% of the total emission within the YREB. The main reason for this is that Shanghai municipality, Jiangsu and Zhejiang provinces are more industrialized, and Anhui province is a traditional heavy industrial province. On the contrary, the upstream region has the lowest CO₂ emission due to its less developed economy.

From a sectoral perspective, it is noticed that there also clearly exist regional disparities in sectoral CO₂ emissions. Figure 2 exhibits that the highest CO₂ emissions are generated by CII in most regions, with a stable growth over time, followed by RII. In the midstream and downstream regions, the contribution of CII to the overall CO₂ emission ranks first, accounting for 48% and 64% of the total CO₂ emission within the two regions in 2017, respectively. The proportion of the CO₂ emission generated by RII in the upstream region (31%) is higher than those in the other two regions (21% and 19%, respectively), due to the relatively low industrial development level and the high dependence on resource advantages. The CO₂ emission generated by TII is mainly concentrated in the downstream region. The reason for this is that the electronic equipment and machinery sector in the downstream region is indispensable for the global industrial chain and different types of consumer electronic products are manufactured in this region with a large market share in the global market. In addition, the proportion of the CO₂ emission generated in SEI ranks third in the three regions.

3.2 Decomposition analysis at regional level

3.2.1 Regional CO₂ emission changes induced by various drivers

The three-layer LMDI method was applied to analyze the driving forces of regional CO₂ emissions in the YREB for the period of 2000–2017. The detailed decomposition results of different regions and specific provinces/municipalities are listed in Tables S3–S5 in the ESM.

Figure 3 presents the temporal decomposition results of the entire YREB region during the period studied. It is apparently seen that the YREB experienced a rapid CO₂ emission growth, which increased by nearly two times during 2000–2017. Economic growth was the major driver for the increase of CO₂ emission in the YREB, which has also been found in some other regions, or countries, such as Shanghai [33], China [73] and Turkey [74]. During the period studied, the growth rate of GDP exceeded that of CO₂ emission in the YREB. The increasing rate of CO₂ emission induced by economic growth is more than 45% in the first two stages (2000–2010) and 27.1% in the 12th FYP period, and finally dropped to 1.2% during 2015–2017. This indicates that the CO₂ emission decoupled from the income level within the YREB. Population scale, carbon emission coefficient, and industrial structure all increased CO₂ emissions of the YREB, with different contributions. Similarly, Wen and Li [75] found that population scale increased CO₂ emissions in many Chinese provinces. The contribution rate of population scale to increasing CO₂ emissions in the YREB decreased from 15.4% during the 10th FYP period, to 14.2% during the 11th FYP period, to 12.1% during the 12th FYP period, and finally to 11.7% during 2015–2017. The contribution rate of carbon emission coefficient to CO₂ emission was stable, at 0.3% throughout the entire period, with slight fluctuations. Additionally, the average annual contribution rates of industrial structure to CO₂ emission were 1.17% during the 10th FYP period, 0.61% during the 11th FYP period, and 0.11% during the 12th FYP period, while it turned to be negative at −0.30% during 2015–2017. It demonstrates that the industrial structure of the YREB has been updated to a less energy intensive one. However, there is a need to implement policies to further optimize the industrial structure in the YREB due to its contribution to CO₂ emission increase.

In contrast, the most important mitigation driver is energy intensity within the YREB. Many studies have also found that energy intensity is the main factor for CO₂ reduction, such as Zhao et al. [76], Zhu et al. [77], and Kopidou and Diakoulaki [78]. Energy intensity contributed to emission reductions in the YREB at average annual rates of - 4.22%, - 5.51%, and - 2.13% during the 11th FYP period, the 12th FYP period and 2015–2017, respectively, but increased emissions by 1.84% annually during the 10th FYP period, showing a continuous improvement. In response to the government’s call for saving energy and reducing emissions, this echoes the fact that most provinces within the YREB transferred their high pollution and high energy consuming industries to other regions, especially in the downstream region. However, energy structure has a very marginal negative effect on the overall CO₂ emission, which indicates that the energy structure within the YREB has not been significantly adjusted. Although several provinces developed more renewable energy sources, such as solar and wind power, the scale is not big enough to induce significant energy structure changes [79]. Therefore, there is a huge potential for improving energy structure in the YREB to achieve a greater CO₂ emission reduction.

Figure 4 shows regional CO₂ emission changes induced by various drivers in the upstream, midstream, and downstream regions during the four stages studied. The CO₂ emission in the downstream region experienced the fastest growth in the 10th FYP period, in which economic growth has the largest positive effect on CO₂ emission. Economic growth increased the overall CO₂ emission by 7.8%, 11.7%, 7.4%, and 5.8% annually during the four stages studied. Such a reduced average annual contribution rate by economic growth during the last three stages indicates that the economy in the downstream region has changed from a high-speed growth model to a high-quality development model. Particularly, Shanghai, as China’s largest city and one international finance center, has committed to the development of SEI by optimizing its industrial structure. The CO₂ emission in the midstream region increased in most years but decreased annually at a rate of - 2.88% from 2011 to 2014 (see Fig. 1). Especially, the CO₂ emission growth caused by economic growth is more than 50% during the 10th and 11th FYP period. The CO₂ emission in the upstream region continuously increased during 2000–2012, but began to decrease from 2012 (see Fig. 1). The average annual contribution rate of economic growth to CO₂ emission in the upstream region increases, which is mainly caused by the fact that the strategy of the western development accelerated the economic development of the upstream region. As can be seen, before 2010, the contribution of economic growth was more significant in the downstream region than those in the other regions, but such a situation began to be opposite after 2010. This phenomenon may be attributed to inter-regional industrial transfer since many high energy consumption industries gradually transferred from the downstream region to the upstream and midstream regions [80]. Energy structure only slightly decreased the CO₂ emission in the downstream region for the entire period studied. This driver even increased the CO₂ emission by 0.7% and 1.3% during the 10th and 11th FYP periods, respectively. This indicates that there is a great potential to further optimize the energy structure in the downstream region. Energy intensity offsets a large part of CO₂ emissions, and the contribution of this driver to offset the CO₂ emission in the downstream region is larger than those in the upstream and midstream regions. This indicates that the downstream region has technological advantages in improving energy efficiency. Both upstream and midstream regions should introduce more technologies to accelerate their energy efficiency improvement and prepare more mitigation policies. Although regional structure has a slightly marginal negative effect on the CO₂ emission in the three regions, the emission reduction performance of this driver in the downstream region is better than those in the other two regions. That can be attributed to the situation that more industrial transfer occurred in the different provinces/municipalities within the downstream region. Industrial structure within the three regions all increased the CO₂ emissions during the first three stages, but decreased the CO₂ emissions during 2015–2017. This reflects the contribution of innovative industrial transformation proposed in the Outline of the YREB Development Plan, even not adequately.

3.2.2 Cross-regional decomposition analysis

Based on Eqs. (2), (5), and (6), the drivers on CO₂ emissions in the three regions were decomposed for the year 2017. Table 1 lists the decomposition analysis results in the upstream, midstream, and downstream regions.

It is clearly observed that the CO₂ emission differences between the downstream region and the upstream/midstream regions are remarkable. Economic growth is the largest driver leading to the CO₂ emission difference between the two regions. The contributions of economic growth to the CO₂ emission difference between the downstream and the upstream/midstream region are 1382 Mt and 863 Mt, respectively. The reason for this is that the downstream region of the YREB is one of the most advanced economic regions in China, with a larger economic scale and more diversified industrial structure. Population scale is the second important driver inducing more CO₂ emission in the downstream region than that in the upstream/midstream regions, with contributions of 166 Mt and 202 Mt, respectively. The downstream region attracted more immigrants from both abroad and at home due to more job opportunities and higher income. Energy intensity is the largest factor on CO₂ emissions mitigation in both regions, but more CO₂ emissions were reduced by this driver in the downstream region than that in the upstream and midstream regions, with contributions of 624 Mt and 79 Mt, respectively. Energy structure is another significant driver reducing more CO₂ emission in the downstream region than that in the upstream/midstream region, with contributions of 37 Mt and 41 Mt, respectively. It indicates that the downstream region has a better energy structure and a higher energy utilization efficiency, and the upstream and midstream regions should make great efforts in energy conservation and emission reduction. Industrial structure is also one mitigation driver in 2017, but it induced more CO₂ emission in the downstream region than that in the upstream and midstream regions, respectively. Industrial structure optimization contributed immensely to emission reductions, especially in the downstream region. Industrial transfer from the downstream region to the midstream region mainly includes CII and other heavy industries. Unfortunately, technological updates did not follow such transfers, resulting in more CO₂ emissions in the midstream region. Therefore, it is critical to deploy low carbon technologies in the midstream region, particularly in those energy intensive industries.

The CO₂ emission difference between the upstream region and the midstream region is relatively small because these two regions have similar industrial structure and resource endowments. Economic growth is the only driver inducing more CO₂ emission in the midstream region than in the upstream region, with a contribution of 518 Mt. Other drivers induced less CO₂ emission in the midstream region than in the upstream region, with an amount of 14 Mt (energy structure), 341 Mt (energy intensity), 69 Mt (industrial structure) and 107 Mt (population scale). In comparison with the midstream region, the development of the upstream region is more carbon-intensive, mainly due to the implementation of the western development strategy. This national strategy has a clear focus on developing infrastructure in western China, leading to more investments in the field of energy intensive industries.

3.3 Decomposition analysis at sectoral level

In this study only the cumulative effects of industrial CO₂ emissions are investigated for the period of 2000–2017. The decomposition results of various industries in the upstream, midstream, and downstream regions are listed in Table 2. Because the effect of carbon emission coefficient is too marginal, the contribution of this driver is neglectable. Economic growth is one positive driver on CO₂ emission increases in all sectors, while energy intensity is the major mitigation driver. Energy structure is also one mitigation driver, but with much less impact.

For the AGR, economic growth increased CO₂ emissions by 93%, 154%, and 68% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 42%, 35%, and 38% in the upstream, midstream and downstream regions, respectively. Energy structure also slightly contributed to CO₂ emission reduction, with 5%, 4%, and 4%, in the upstream, midstream and downstream regions, respectively. The percentage of agricultural contribution to the overall economy is lower in the downstream region than those in the upstream and midstream regions. Such a reality leads to the fact that the percentage of increased CO₂ emission induced by agriculture in the downstream is less than that in the upstream and midstream regions. This requires both the midstream and upstream regions to further improve their agricultural energy management. In addition, it is noticeable that energy intensity in the agriculture sector has been greatly reduced due to the promotion of resource and energy efficient agriculture in the entire country. But in general, the overall emission from agricultural sector is still marginal compared with other sectors.

For resource intensive industry (RII), economic growth increased CO₂ emissions by 281%, 250%, and 255% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 92%, 122%, and 108% in the upstream, midstream and downstream regions, respectively. Energy structure also contributed to CO₂ emissions reduction, with 12%, 11%, and 15% in the upstream, midstream and downstream regions, respectively. Although the overall emission of the upstream region is much less than that in the downstream region, the overall emission from this sector is almost the same as that in the downstream region, indicating that this sector is a key emission sector in the upstream region. The midstream region is facing a similar challenge due to its rich resource endowment. Consequently, it is crucial for both regions to further improve the overall energy efficiency in this sector by applying more energy efficient technologies and improving the overall energy management performance.

For labor intensive industries (LII), economic growth increased CO₂ emissions by 117%, 183%, and 214% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 104%, 136%, and 122% in the upstream, midstream and downstream regions, respectively. Energy structure also contributed to CO₂ emissions reduction, with 7%, 11%, and 13% in the upstream, midstream and downstream regions, respectively. Unlike the resource intensive industry, this sector has a strong foundation in the downstream region. For instance, Zhejiang province is famous for its various labor-intensive products, which has attracted many immigrants from other regions. Due to the convenient harbors and mature manufacturing experiences, the downstream region will continue to maintain this sector in the near future. But the midstream region is also following such a trend. Therefore, the downstream region should promote the transfer of some labor-intensive industries to the midstream region so that some immigrants can return to their hometown to support the local economy.

For capital intensive industries (CII), economic growth increased CO₂ emissions by 205%, 243%, and 271% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 78%, 99%, and 101% in the upstream, midstream and downstream regions, respectively. Energy structure also contributed to CO₂ emissions reduction, with 8%, 17%, and 22% in the upstream, midstream and downstream regions, respectively. It is clearly noticed that the downstream region has more capital-intensive industries than the other two regions. But with the industrial transfer of such industries, the other two regions will face more challenges since capital intensive industries are generally heavy industry oriented, such as steel and iron, petrochemicals, machinery, and transportation equipment. These sectors are energy and pollution intensive and should be quickly supplied with more energy efficient technologies and equipment.

For technology intensive industries (TII), economic growth increased CO₂ emissions by 183%, 198%, and 247% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 36%, 39%, and 96% in the upstream, midstream and downstream regions, respectively. Energy structure also slightly contributed to CO₂ emission reduction, with 13%, 12%, and 20% in the upstream, midstream and downstream regions, respectively. It is apparent that the downstream region has more advantages in this sector due to its advanced technologies and leading research and development strength. This indicates that the upstream region should actively help the other two regions improve the overall energy efficiency of this sector during the industrial transfer.

For the construction sector (CON), economic growth increased CO₂ emissions by 347%, 406%, and 463% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 56%, 61%, and 65% in the upstream, midstream and downstream regions, respectively. Energy structure also slightly contributed to CO₂ emissions reduction, with 11%, 16%, and 23% in the upstream, midstream and downstream regions, respectively. Essentially, rapid urbanization is a key feature in the three regions, which lead to a large amount of CO₂ emission although it is more urbanized in the downstream region. However, with China’s west development strategy, it seems that the upstream region will continue to urbanize so that their emigrated population can return home to develop their economy. This means that the western region may face more pressures in responding to the increasing CO₂ emission. Consequently, both the middle and downstream regions should help the upstream region by transferring more advanced technologies on green buildings and green infrastructure. Moreover, it is necessary to promote more construction practitioners to seek green building certifications by showing the long-term benefits of such a green initiative.

For the service industry (SEI), economic growth increased CO₂ emissions by 463%, 461%, and 347% in the upstream, midstream and downstream regions, respectively, while energy intensity decreased CO₂ emissions by 79%, 97%, and 106% in the upstream, midstream and downstream regions, respectively. Energy structure also slightly contributed to CO₂ emissions reduction, with 5%, 4%, and 4% in the upstream, midstream and downstream regions, respectively. The downstream region has more diversified service industries and mature experiences on managing these businesses. But both the midstream and upstream regions will boom their service sector with continuing urbanization. In this regard, the downstream region should actively transfer their experiences and know-how to the other two regions, especially capacity-building activities, so that all the regions can better develop their service sector under the umbrella of low carbon development.

4 Policy implications

With the rapid development of the entire YREB, both energy consumption and CO₂ emissions have dramatically increased, leading to a great pressure on all the governments within the YREB. Policy makers have to coordinate economic development and environmental protection so that sustainable development can be achieved. In particular, due to regional disparities on industrial and energy structure, population scale, economic development, resource endowments, and technological levels, it is crucial for different provincial governments to prepare their mitigation policies by considering their own realities.

4.1 Policy recommendations for upstream region

The upstream region lacks large plains for agricultural development due to its hilly land. RII is the major CO₂ emitter in the upstream region. The proportion of the CO₂ emission generated by RII in the upstream region is higher than those in the other two regions, due to the relatively low industrial development level and the high dependence on resource advantages. CII and SEI are two other large CO₂ emitters in the upstream region. Therefore, it is necessary for the upstream region to concentrate on mitigating CO₂ emissions in these three sectors. Besides, energy structure and economic growth increased CO₂ emissions in RII. Therefore, the energy structure should be optimized in the RII sector by developing more renewable and clean energy sources. The contributions of energy structure and energy intensity to CO₂ reduction in CII were relatively small in the upstream region compared with those in the other two regions. Therefore, energy structure and energy efficiency in CII should be further improved. In total, carbon emission coefficient and energy structure contributed to the increases in CO₂ emission in the upstream region, since many provinces in the upstream region mainly rely on coal consumption. Thus, the energy structure in provinces of the upstream region should be optimized. In this regard, the city of Guiyang (capital of Guizhou province) has successfully developed its big data industry due to its rich hydropower. There is a great potential to further promote renewable and clean energy in the upstream region due to its rich endowments on hydropower, geothermal power, wind power, and biomass. In addition, the energy intensity in the upstream region should be further reduced by encouraging more energy efficient technologies and equipment. As such, local governments should consider phasing out those energy and pollution intensive industries by more strictly enforcing environmental regulations, providing financial subsidies, promoting energy efficient technologies, supporting related research activities, and organizing more capacity-building efforts. Finally, compared with both the midstream and downstream regions, research abilities in the upstream region are still weak. Consequently, it is critical for this region to transfer more advanced technologies from other regions so that they can achieve leapfrog development.

4.2 Policy recommendations for midstream region

The midstream region includes several traditional agricultural provinces due to their fertile land and rich water resources. Thus, it is necessary for this region to continue to pay more attention to agriculture, but with more applications of innovative agricultural technologies, such as water saving planting, organic fertilizer, reduced use of pesticide, and eco-farm, etc. Besides, these provinces have rich reserves on coal, rare earth, iron ore, and other resources. With the soaring demand on these resources, it is difficult for the midstream region to phase out these resource intensive industries. However, it is crucial to promote energy efficient and clean mining so that the negative impacts from these activities can be minimized. CII was the largest CO₂ emitter in the midstream region, followed by RII and SEI. Energy structure contributed to the increases of CO₂ emissions in CII and RII. Therefore, the midstream region should pay attention to optimizing the energy structure in the two sectors. Energy structure and energy intensity contributed to the increase of CO₂ emissions in SEI. Hence, energy intensity should be reduced in SEI by applying more advanced technologies and equipment. With regard to industrial structure, this region is in the transmission period. While some new high-tech industries have been operated well, such as the famous the optical valley in Wuhan, many state-owned enterprises (especially heavy industries) are still dominating their manufacturing sector. It is therefore suggested to gradually optimize its industrial structure by encouraging more service industries and high-tech industries. Fortunately, there are many universities and research institutions in this region, providing a great potential to incubate more innovations. It will be even better for these organizations to not only help the local economy but also the upstream region.

4.3 Policy recommendations for downstream region

As one of the most developed regions in China, the downstream region has the outstanding research, education, and financial ability to optimize its energy and industrial structure, as well as improving its energy efficiency. The contribution of energy intensity to CO₂ reduction is the largest in the downstream region of the three regions. However, the reduction effect of energy structure on CO₂ emissions is much smaller compared with that of energy intensity in this region. Given the fact that the growth rate of CO₂ emission in the downstream region was nearly two times of those in the upstream region and midstream region during 2000–2017, this region should play a leading role in further promoting renewable and clean energy and energy efficient technologies through investing in Research and Development (R&D) activities. Besides its own research outcomes, this region should also pay more attention to international cooperation to import the most advanced technologies from other countries and localize such technologies by considering the Chinese situations. Due to the lack of natural resource endowments, this region has to rely on importing raw materials from both upstream and midstream regions, as well as other regions. Under such a circumstance, this region cannot simply take a not-in-my-backyard approach, without considering recovering the ecological destruction in those mining sites. Rather, this region should support resource-rich regions by transferring environmentally friendly technologies and providing financial support so that resource-rich regions can operate their activities more sustainably. In addition, CII contributes greatly to the increase of CO₂ emissions in the downstream region, far surpassing the contributions from other sectors. Therefore, this region should further mitigate the emissions in CII. The reduction effect from energy structure is only around 20% of that from the energy intensity in CII. This sector is energy and pollution intensive and should be quickly equipped with clean energy and production technologies. Finally, the downstream region has more innovative low carbon development measures due to its advanced economy and more environmentally conscious public. For instance, Shanghai is the first Chinese city to promote waste separation at home so that those recyclable wastes can be reused/recycled to replace virgin materials [81]. Such a policy is remarkable and can lead to a greater CO₂ emission reduction since the mining and processing of virgin materials is usually energy and emission intensive. Therefore, such innovative practices should be transferred to the other two regions so that more cities there can initiate their own projects by considering the local situations.

5 Conclusions

The YREB is a paramount economic development belt with significant regional disparities in terms of resource endowments, economic development, industrial structure and technology level in its upstream, midstream and downstream regions. Regional CO₂ emissions in the YREB are rapidly increasing due to the rapid development of various industries and urbanization. However, previous studies have rarely paid attention to both regional and industrial CO₂ emissions in the YREB region. Therefore, it is crucial to identity the drivers of CO₂ emissions in the YREB under the background of regional industrial transfer. This study accounts the industrial CO₂ emissions in the YREB and discusses the drivers at both regional and industrial levels by employing the IPCC and multi-layer LMDI decomposition methods. Decomposition results suggest that the driving effects of economic growth, energy intensity, energy structure and industrial structure vary at regional level because of different regional development patterns within the YREB, which further lead to more CO₂ emissions after the 12th FYP period. Economic growth is the main positive driver on CO₂ emission increases, especially in the downstream region during 2000–2010. Energy intensity and energy structure are the main mitigation driver on CO₂ emissions, and the mitigation contribution of those two drivers in the downstream region are more notable. Such a reality indicates that efficiency improvement is always the major strategy for mitigating the total CO₂ emission in the YREB, and both upstream and midstream regions should pay more attention to further improving their overall energy efficiency and optimizing their energy structure, including devoting more efforts to improving energy efficiency, phasing out backward and outdated production capacity in energy-intensive industries, implementing proactive policies to promote low carbon production, etc. Industrial structures in the three regions increased the CO₂ emissions during the first three stages, but decreased the CO₂ emissions during 2015–2017. Meanwhile, the emission reduction performance of this driver in the downstream region is more excellent than that in the other two regions. Such a fact shows that industrial structure contributed to emission reduction with intensified efforts, and the innovative industrial transformation functioned. Thus, the downstream region should transfer more advanced technologies and equipment to their counterparts in both the upstream and midstream regions during the industrial transfer process in the future. Moreover, coordinating the development in the upstream, midstream, and downstream regions can further contribute to achieving the emissions reduction goals. Although this study focuses only on the YREB, policy implications from this study can be referred to by other countries facing similar challenges so that both regional and sectoral mitigation policies can be raised by considering their local realities.

Several limitations exist in this study. First, there exist some uncertainties in the decomposition results because it is difficult to guarantee the accuracy of energy consumption data and socioeconomic data released by the government. In addition, the inter-regional CO₂ emissions within the YREB have not been analyzed using the IDA method due to time and data limitations. It is necessary to further investigate such factors in the future.

References

[1] World Bank. World bank topics. 2018–09–16

[2] Wu R, Geng Y, Cui X, Gao Z, Liu Z. Reasons for recent stagnancy of carbon emissions in China’s industrial sectors. Energy, 2019, 172: 457–466

[3] National Development and Reform Commission (NDRC). Notice of the State Council on Issuing the Comprehensive Work Plan for Energy Saving and Emission Reduction in 12th Five-year Plan. Beijing: Statistics Press, 2011 (in Chinese)

[4] National Development and Reform Commission (NDRC). Notice of the State Council on Issuing the Comprehensive Work Plan for Energy Saving and Emission Reduction in 13th Five-Year Plan. Beijing: Statistics Press, 2016 (in Chinese)

[5] He J K. An analysis of China’s CO₂ emission peaking target and pathways. Advances in Climate Change Research, 2014, 5(4): 155–161

[6] Zhang X, Geng Y, Shao S, Wilson J, Song X, You W. China’s non-fossil energy development and its 2030 CO₂ reduction targets: the role of urbanization. Applied Energy, 2020, 261: 114353

[7] Ninpanit P, Malik A, Wakiyama T, Geschke A, Lenzen M. Thailand’s energy-related carbon dioxide emissions from production-based and consumption-based perspectives. Energy Policy, 2019, 133: 110877

[8] Steckel J C, Hilaire J, Jakob M, Edenhofer O. Coal and carbonization in sub-Saharan Africa. Nature Climate Change, 2020, 10(1): 83–88

[9] Oliver J, Janssens M G, Munteank M, Peters J. Trends in Global CO₂ Emissions: 2015 Report. 2015, PBL Netherlands Environmental Assessment Agency

[10] Zhou S, Duan M, Yuan Z, Ou X. Peak CO₂ emission in the region dominated by coal use and heavy chemical industries: a case study of Dezhou city in China. Frontiers in Energy, 2018, [Epub ahead of print] 10.1007/s11708-018-0558-y

[11] Chen J, Wang P, Cui L, Huang S, Song M. Decomposition and decoupling analysis of CO₂ emissions in OECD. Applied Energy, 2018, 231: 937–950

[12] Mousavi B, Lopez N S A, Biona J B M, Chiu A S F, Blesl M. Driving forces of Iran’s CO₂ emissions from energy consumption: an LMDI decomposition approach. Applied Energy, 2017, 206: 804–814

[13] Wang H, He J. China’s pre-2020 CO₂ emission reduction potential and its influence. Frontiers in Energy, 2019, 13(3): 571–578

[14] Lopez N S, Chiu A S, Biona J B. Decomposing drivers of transportation energy consumption and carbon dioxide emissions for the Philippines: the case of developing countries. Frontiers in Energy, 2018, 12(3): 389–399

[15] Fujii M, Fujita T, Dong L, Lu C, Geng Y, Behera S K, Park H S, Chiu A S F. Possibility of developing low-carbon industries through urban symbiosis in Asian cities. Journal of Cleaner Production, 2016, 114: 376–386

[16] Kuramochi T. Assessment of midterm CO₂ emissions reduction potential in the iron and steel industry: a case of Japan. Journal of Cleaner Production, 2016, 132: 81–97

[17] Song X, Geng Y, Li K, Zhang X, Wu F, Pan H, Zhang Y. Does environmental infrastructure investment contribute to emissions reduction? A case of China. Frontiers in Energy, 2020, 14(1): 57–70

[18] Song M, Guo X, Wu K, Wang G. Driving effect analysis of energy-consumption carbon emissions in the Yangtze River Delta region. Journal of Cleaner Production, 2015, 103: 620–628

[19] Hao H, Geng Y, Hang W. GHG emissions from primary aluminum production in China: regional disparity and policy implications. Applied Energy, 2016, 166: 264–272

[20] Li A, Zhang A, Zhou Y, Yao X. Decomposition analysis of factors affecting carbon dioxide emissions across provinces in China. Journal of Cleaner Production, 2017, 141: 1428–1444

[21] Gu S, Fu B, Thriveni T, Fujita T, Ahn J W. Coupled LMDI and system dynamics model for estimating urban CO₂ emission mitigation potential in Shanghai, China. Journal of Cleaner Production, 2019, 240: 118034

[22] Zheng J, Mi Z, Coffman D M, Milcheva S, Shan Y, Guan D, Wang S. Regional development and carbon emissions in China. Energy Economics, 2019, 81: 25–36

[23] Liang W, Gan T, Zhang W. Dynamic evolution of characteristics and decomposition of factors influencing industrial carbon dioxide emissions in China: 1991–2015. Structural Change and Economic Dynamics, 2019, 49: 93–106

[24] Xu S C, He Z X, Long R Y. Factors that influence carbon emissions due to energy consumption in China: decomposition analysis using LMDI. Applied Energy, 2014, 127: 182–193

[25] Tian X, Bai F, Jia J, Liu Y, Shi F. Realizing low-carbon development in a developing and industrializing region: impacts of industrial structure change on CO₂ emissions in southwest China. Journal of Environmental Management, 2019, 233: 728–738

[26] Chen J, Yuan H, Tian X, Zhang Y, Shi F. What determines the diversity of CO₂ emission patterns in the Beijing-Tianjin-Hebei region of China? An analysis focusing on industrial structure change. Journal of Cleaner Production, 2019, 228: 1088–1098

[27] Xu X, Yang G, Tan Y, Zhuang Q, Tang X, Zhao K, Wang S. Factors influencing industrial carbon emissions and strategies for carbon mitigation in the Yangtze River Delta of China. Journal of Cleaner Production, 2017, 142: 3607–3616

[28] Chen L, Xu L, Yang Z. Accounting carbon emission changes under regional industrial transfer in an urban agglomeration in China’s Pearl River Delta. Journal of Cleaner Production, 2017, 167: 110–119

[29] Liao C, Wang S, Zhang Y, Song D, Zhang C. Driving forces and clustering analysis of provincial-level CO₂ emissions from the power sector in China from 2005 to 2015. Journal of Cleaner Production, 2019, 240: 118026

[30] Geng Y, Zhao H, Liu Z, Xue B, Fujita T, Xi F. Exploring driving factors of energy-related CO₂ emissions in Chinese provinces: a case of Liaoning. Energy Policy, 2013, 60: 820–826

[31] Shan Y, Liu J, Liu Z, Xu X, Shao S, Wang P, Guan D. New provincial CO₂ emission inventories in China based on apparent energy consumption data and updated emission factors. Applied Energy, 2016, 184: 742–750

[32] Zheng X, Wang R, He Q. A city-scale decomposition and decoupling analysis of carbon dioxide emissions: a case study of China. Journal of Cleaner Production, 2019, 238: 117824

[33] Shao S, Yang L, Gan C, Cao J, Geng Y, Guan D. Using an extended LMDI model to explore techno-economic drivers of energy-related industrial CO₂ emission changes: a case study for Shanghai (China). Renewable & Sustainable Energy Reviews, 2016, 55: 516–536

[34] Liu Z, Liang S, Geng Y, Xue B, Xi F, Pan Y, Zhang T, Fujita T. Features, trajectories and driving forces for energy-related GHG emissions from Chinese mega cites: the case of Beijing, Tianjin, Shanghai and Chongqing. Energy, 2012, 37(1): 245–254

[35] Gao Z, Geng Y, Wu R, Chen W, Wu F, Tian X. Analysis of energy-related CO₂ emissions in China’s pharmaceutical industry and its driving forces. Journal of Cleaner Production, 2019, 223: 94–108

[36] Wang Y, Zhu Q, Geng Y. Trajectory and driving factors for GHG emissions in the Chinese cement industry. Journal of Cleaner Production, 2013, 53: 252–260

[37] Shao S, Liu J, Geng Y, Miao Z, Yang Y. Uncovering driving factors of carbon emissions from China’s mining sector. Applied Energy, 2016, 166: 220–238

[38] National Development and Reform Commission (NDRC). Guiding Opinions of the State Council on Relying on Golden Waterways to Promote the Development of the Yangtze River Economic Belt. 2014

[39] Ren F R, Tian Z, Shen Y T, Chiu Y H, Lin T Y. Energy, CO₂, and AQI efficiency and improvement of the Yangtze River Economic Belt. Energies, 2019, 12(4): 647

[40] Tian Z, Ren F R, Xiao Q W, Chiu Y H, Lin T Y. Cross-regional comparative study on carbon emission efficiency of China’s Yangtze River Economic Belt based on the meta-frontier. International Journal of Environmental Research and Public Health, 2019, 16(4): 619

[41] Tang D, Zhang Y, Bethel B J. An analysis of disparities and driving factors of carbon emissions in the Yangtze River Economic Belt. Sustainability, 2019, 11(8): 2362

[42] Zhang X, Zhao X, Jiang Z, Shao S. How to achieve the 2030 CO₂ emission-reduction targets for China’s industrial sector: retrospective decomposition and prospective trajectories. Global Environmental Change, 2017, 44: 83–97

[43] Mi Z, Meng J, Green F, Coffman D M, Guan D. China’s “exported carbon” peak: patterns, drivers, and implications. Geophysical Research Letters, 2018, 45(9): 4309–4318

[44] Zhang S, Li H, Zhang Q, Tian X, Shi F. Uncovering the impacts of industrial transformation on low-carbon development in the Yangtze River Delta. Resources, Conservation and Recycling, 2019, 150: 104442

[45] Zhao X, Ma Q, Yang R. Factors influencing CO₂ emissions in China’s power industry: co-integration analysis. Energy Policy, 2013, 57: 89–98

[46] Lin B, Fei R. Regional differences of CO₂ emissions performance in China’s agricultural sector: a Malmquist index approach. European Journal of Agronomy, 2015, 70: 33–40

[47] Liu Z, Geng Y, Dong H, Wilson J, Micic T, Wu R, Cui X, Qian Y, You W, Sun H. Efficient distribution of carbon emissions reduction targets at the city level: a case of Yangtze River Delta region. Journal of Cleaner Production, 2018, 172: 1711–1721

[48] Li K, Lin B. Economic growth model, structural transformation, and green productivity in China. Applied Energy, 2017, 187: 489–500

[49] National Development and Reform Commission (NDRC). Outline of the Yangtze River Economic Belt Development Plan. 2016, General office of the State Council of China

[50] Intergovernmental Panel on Climate Change (IPCC). 2006 Intergovernmental Panel on Climate Change Guidelines for National Greenhouse Gas Inventories. 2006

[51] Jeong K, Kim S. LMDI decomposition analysis of greenhouse gas emissions in the Korean manufacturing sector. Energy Policy, 2013, 62: 1245–1253

[52] Ang B W, Wang H. Index decomposition analysis with multidimensional and multilevel energy data. Energy Economics, 2015, 51: 67–76

[53] Ang B W, Liu F L. A new energy decomposition method: perfect in decomposition and consistent in aggregation. Energy, 2001, 26(6): 537–548

[54] Wu L, Kaneko S, Matsuoka S. Driving forces behind the stagnancy of China’s energy-related CO₂ emissions from 1996 to 1999: the relative importance of structural change, intensity change and scale change. Energy Policy, 2005, 33(3): 319–335

[55] Ang B W, Zhang F Q. Inter-regional comparisons of energy-related CO₂ emissions using the decomposition technique. Energy, 1999, 24(4): 297–305

[56] Li A, Hu M, Wang M, Cao Y. Energy consumption and CO₂ emissions in eastern and central China: a temporal and a cross-regional decomposition analysis. Technological Forecasting and Social Change, 2016, 103: 284–297

[57] Zhang F Q, Ang B W. Methodological issues in cross-country/region decomposition of energy and environment indicators. Energy Economics, 2001, 23(2): 179–190

[58] National Bureau of Statistics of China (NBSC). Industrial Classification for National Economic Activities (GB/T 4754–2017). Beijing: China Statistics Press, 2017 (in Chinese)

[59] Tian Y, Xiong S, Ma X, Ji J. Structural path decomposition of carbon emission: a study of China’s manufacturing industry. Journal of Cleaner Production, 2018, 193: 563–574

[60] Luo L, Zhao F. Optimization of industrial layout and high-quality development of the Yangtze River Economic Belt: based on the perspective of industrial transfer between regions. Reform, 2019, 2: 27–36 (in Chinese)

[61] Li L, Ma Y. Spatial-temporal pattern evolution of manufacturing geographical agglomeration and influencing factors of old industrial base: a case of Jilin province, China. Chinese Geographical Science, 2015, 25(4): 486–497

[62] Ang B W, Xu X Y, Su B. Multi-country comparisons of energy performance: the index decomposition analysis approach. Energy Economics, 2015, 47: 68–76

[63] Guan D, Liu Z, Geng Y, Lindner S, Hubacek K. The gigatonne gap in China’s carbon dioxide inventories. Nature Climate Change, 2012, 2(9): 672–675

[64] Gregg J S, Andres R J. A method for estimating the temporal and spatial patterns of carbon dioxide emissions from national fossil-fuel consumption. Tellus. Series B, Chemical and Physical Meteorology, 2008, 60(1): 1–10

[65] Liu Z, Geng Y, Lindner S, Guan D. Uncovering China’s greenhouse gas emission from regional and sectoral perspectives. Energy, 2012, 45(1): 1059–1068

[66] Wang Z, Yin F, Zhang Y, Zhang X. An empirical research on the influencing factors of regional CO₂ emissions: evidence from Beijing city, China. Applied Energy, 2012, 100: 277–284

[67] Liu Z, Guan D, Wei W, Davis S J, Ciais P, Bai J, Peng S, Zhang Q, Hubacek K, Marland G, Andres R J, Crawford-Brown D, Lin J, Zhao H, Hong C, Boden T A, Feng K, Peters G P, Xi F, Liu J, Li Y, Zhao Y, Zeng N, He K. Reduced carbon emission estimates from fossil fuel combustion and cement production in China. Nature, 2015, 524(7565): 335–338

[68] Shan Y, Guan D, Zheng H, Ou J, Li Y, Meng J, Mi Z, Liu Z, Zhang Q. China CO₂ emission accounts 1997–2015. Scientific Data, 2018, 5(1): 170201

[69] Shan Y, Huang Q, Guan D, Hubacek K. China CO₂ emission accounts 2016–2017. Scientific Data, 2020, 7(1): 54

[70] National Bureau of Statistics of China (NBSC). China Energy Statistical Yearbooks (2000–2017). Beijing: China Statistics Press, 2018 (in Chinese)

[71] National Bureau of Statistics of China (NBSC). China Statistical Yearbooks (2000–2017). Beijing: China Statistics Press, 2018 (in Chinese)

[72] Shan Y, Guan D, Liu J, Mi Z, Liu Z, Liu J, Schroeder H, Cai B, Chen Y, Shao S, Zhang Q. Methodology and applications of city level CO₂ emission accounts in China. Journal of Cleaner Production, 2017, 161: 1215–1225

[73] Wang M, Feng C. Using an extended logarithmic mean Divisia index approach to assess the roles of economic factors on industrial CO₂ emissions of China. Energy Economics, 2018, 76: 101–114

[74] Akbostancı E, Tunç G I, Türüt-Aşık S. Drivers of fuel based carbon dioxide emissions: the case of Turkey. Renewable & Sustainable Energy Reviews, 2018, 81: 2599–2608

[75] Wen L, Li Z. Provincial-level industrial CO₂ emission drivers and emission reduction strategies in China: combining two-layer LMDI method with spectral clustering. Science of the Total Environment, 2020, 700: 134374

[76] Zhao X, Zhang X, Shao S. Decoupling CO₂ emissions and industrial growth in China over 1993–2013: the role of investment. Energy Economics, 2016, 60: 275–292

[77] Zhu X, Zou J, Feng C. Analysis of industrial energy-related CO₂ emissions and the reduction potential of cities in the Yangtze River Delta region. Journal of Cleaner Production, 2017, 168: 791–802

[78] Kopidou D, Diakoulaki D. Decomposing industrial CO₂ emissions of Southern European countries into production- and consumption-based driving factors. Journal of Cleaner Production, 2017, 167: 1325–1334

[79] Wang M, Feng C. Understanding China’s industrial CO₂ emissions: a comprehensive decomposition framework. Journal of Cleaner Production, 2017, 166: 1335–1346

[80] Xiao S, Dong H, Geng Y, Tian X, Liu C, Li H. Policy impacts on Municipal Solid Waste management in Shanghai: a system dynamics model analysis. Journal of Cleaner Production, 2020, 262: 121366

[81] Dong H, Geng Y, Yu X, Li J. Uncovering energy saving and carbon reduction potential from recycling wastes: a case of Shanghai in China. Journal of Cleaner Production, 2018, 205: 27–35

RIGHTS & PERMISSIONS

Higher Education Press