Exploring the Spatiotemporal Heterogeneity of Carbon Emission from Energy Consumption and Its Influencing Factors in the Yellow River Basin

Abstract

Scientific estimation and dynamic monitoring on the heterogeneity of carbon emission from energy consumption (CEEC) is the basis for formulating and implementing regional carbon reduction strategies to realize the goal of carbon neutrality and high-quality development. This study analyzes the temporal and spatial differences of CEEC and its driving factors in the Yellow River Basin (YRB) from 2000 to 2018 based on the Log-Mean Divisia Index (LMDI) time decomposition method and the multi-regional (M-R) space decomposition method. The results indicate the following: The amount of CEEC of the YRB increased greatly from 2000 to 2012, and then expressed a convergence trend after 2012, with obvious spatial differences. The economic development is the leading factor that promotes the increase in CEEC in the YRB, energy intensity is the main force for the reduction in CEEC, and their influencing effectiveness varies significantly in different periods and provinces. Spatially, the larger economic development in Shandong, Henan, and Sichuan causes the higher level of CEEC, and the regulation of energy intensity in Shanxi, Ningxia, and Inner Mongolia is important for the reduction in their CEEC. The impact effectiveness of economic structure and energy structure on CEEC in the YRB is relatively weak, and they are potential factors for the reduction in CEEC. Therefore, the corresponding emission reduction measures in nine provinces of the YRB should focus on reducing energy intensity, building a green energy system, and strengthening “green” economic development to achieve high-quality development in the YRB. This study is designed to explore the spatiotemporal variations and influencing factors of carbon emissions in the nine provinces of the YRB, which is of great significance for achieving low-carbon development in the region.

1. Introduction

The concern about climate change and the reduction in carbon emissions has been important topic for academic research and the international community [1,2]. Carbon emissions from large-scale energy consumption are the main source [3] and have profound impacts on the surface natural environment process and human society [4]. How to reduce and effectively control carbon emissions has become an important factor for countries around the world when formulating environmental policies. China ranks first for energy consumption worldwide [5] and plans to achieve carbon neutrality by 2060. Based on this goal, the scientific understanding of the spatiotemporal heterogeneity characteristics of China’s regional carbon emission from energy consumption (CEEC) and the identification of the main affecting factors can provide a scientific reference for different regions to formulate reasonable and different carbon emission reduction measures.

In recent years, scholars have carried out extensive work and made great achievements in exploring the temporal and spatial heterogeneity of carbon emission and its influencing factors, using the method of regression analysis [6], correlation analysis [7], factor decomposition analysis [8], causal analysis [9], spatial econometric analysis [10], social network analysis [11], and other methods. Among them, decomposition technologies of the structural decomposition analysis (SDA) and index decomposition analysis (IDA) have been the two most commonly used methods [12,13]. The SDA method decomposes carbon emission of specific years based on an input–output table [14,15,16], and it is extremely inconvenient for application in China because the input–output table is generated every 5 years [17]. The IDA method only requires sector aggregation data [18,19], which is particularly suitable for models containing time series data, so it is widely used in China [20]. The Log-Mean Divisia Index (LMDI) method in the IDA decomposes the carbon emission into several factors without residual decomposition, and it is also very effective in decomposing some incomplete datasets [21,22]. Therefore, the LMDI method is popular in the study of time differences in carbon emission [23,24].

Compared with the research on temporal difference decomposition, there has been less research on spatial difference decomposition of carbon emission [25]. At present, spatial decomposition and comparation methods include the bilateral–regional (B-R), radial–regional (R-R), and multi-regional (M-R) space decomposition methods. The B-R model conducts decomposition analysis by generating a comparison group and addressing the two-region pairs in the group, and the calculation difficulty increases greatly with the increase in the number of research units [26,27,28]. The R-R model conducts decomposition analysis by using benchmark reference that may be a specific region in the comparing group or some other benchmark outside the group [29,30], and the B-R model is not convincing because of the subjectivity of the selection of the benchmark area [31]. For this reason, Ang et al. [32] put forward the M-R model, which selects the average level of the study regions as the benchmark to avoid the subjectivity of identifying the benchmark in R-R model.

The distinctions of total amount and structure of CEEC are obvious, and different influencing factors play different roles in different regions [33]. For example, the effectiveness of a region’s economic development and urbanization on carbon emission includes two diametrically opposed views that economic development and urbanization processes promote carbon emission [34,35] and that economic development and urbanization processes inhibit carbon emissions through improving energy utilization efficiency and intensity [36]. Some scholars pointed out that the above differences are common due to the differences of the inverted “U” relationships between urbanization level and carbon emission in different natural stages [37,38]. Additionally, the study regions in the previous study mainly relate to the national level, provincial level, six provinces in the central region, and urban agglomerations [39]. However, there is less research on the basin level and the Yellow River basin (YRB) [40]. The YRB is the most important energy production and supply base in China, and its coal output and the national primary energy output accounted for 70% and 40% of the nation’s total [41]. The comprehensive management and industrial transformation to achieve the goal of carbon emission reduction has become an important development path since the ecological protection and high-quality development of the YRB was established as national strategy in 2019. For the sake of carbon neutrality, ecological protection, and high-quality development of the YRB, local governments are facing great pressure in terms of emission reduction. Therefore, exploring the temporal and spatial differentiation of CEEC and its influencing factors in the YRB is not only conducive to energy conservation and emission reduction at the national level but also of great significance for high-quality development of the YRB.

The research objectives include: (1) to explore the spatiotemporal distribution pattern of carbon emissions in nine provinces of the YRB; (2) to analyze the influencing factors of carbon emissions that are of great significance for achieving low-carbon development; (3) to put forward countermeasures and suggestions for reducing CEEC in the YRB. This research reexamines the issue of regional disparities in carbon emissions from the perspective of carbon sources not only helps to supplement and expand existing research but also deepens the understanding of regional disparities and their sources of carbon emissions.

2. Materials and Methods

2.1. Study Area

The YRB is located in the middle of China along the Yellow River spanning the three ladders of China (Figure 1). The YRB includes 9 provinces, 115 prefecture-level cities, and 963 counties. There are Sichuan, Qinghai, Gansu, and Ningxia in the upper reaches; Inner Mongolia, Shaanxi, and Shanxi in the middle reaches; and Henan and Shandong in the lower reaches. The population and regional gross domestic product (GDP) of the YRB were 421 million people and RMB 254 million in 2020, accounting for 25.07% and 29.88% of the national total, respectively. As a major energy production and supply base in China, its coal output accounted for 70% of the national annual output, and there were 9 of the country’s 14 large coal energy production areas with bases of over 100 million tons.

2.2. Methods

2.2.1. Carbon Emission Calculating Method

In this research, the IPCC method was selected to calculate the CEEC by considering the categories of coke, gasoline, crude oil, fuel oil, diesel oil, crude coal, kerosene, and natural gas. The calculation equation is shown below [42,43].

$$C{O}_2={\displaystyle \sum }_{i=1}^9{E}_i\cdot F_i\cdot K_i\cdot 44/12$$

(1)

where $C{O}_2$ is the amount of CEEC (tons); $E_i$ is the energy consumption amount of energy type i (calculated as standard coal, 10⁴ t); $F_i$ is standard coal coefficient of energy fuel i; and $K_i$ is the carbon emission coefficient of energy type i. The standard coal conversion coefficients and energy carbon emission coefficients are shown in

Table 1. The standard coal conversion coefficients and energy carbon emission coefficients of various energy sources.

Energy Type	Coke	Gasoline	Crude Oil	Fuel Oil	Diesel Oil	Crude Coal	Kerosene	Natural Gas
Standard coal conversion coefficients	0.9714	1.4714	1.4286	1.4286	1.4571	0.7143	1.4714	1.33
Carbon emission coefficient	0.855	0.5538	0.5857	0.6185	0.5921	0.7559	0.5714	0.4483

.

2.2.2. LMDI Factor Decomposition Model

LMDI decomposition method is a branch of IDA method [22,25], which is more suitable for driving-factor analysis with time series property indicators, and has the advantage that there is no residual term after decomposition. The carbon emission (C) can be decomposed into a combination of k driving effects (${\mathrm{X}}_1$, ${\mathrm{X}}_2$, ${\mathrm{X}}_3$, …, ${\mathrm{X}}_{\mathrm{k}}$), and the decomposition form is as follows [18]:

$$C={\mathrm{X}}_1\times {\mathrm{X}}_2\times {\mathrm{X}}_3\times \dots \times {\mathrm{X}}_{\mathrm{k}}$$

(2)

In this paper, the total CEEC is decomposed into:

$$C={\displaystyle \sum }_{i=1}^3{\displaystyle \sum }_{j=1}^{n}G\times \frac{G_i}{G}\times \frac{E_i}{G_i}\times \frac{E_{ij}}{E_i}\times \frac{C_{ij}}{E_{ij}}$$

(3)

where G is GDP, $G_i$ is ith industry’s GDP, $E_i$ is the amount of ith industry’s energy consumption, $E_{ij}$ is the amount of the jth energy consumption in the ith industry, and $C_{ij}$ is the amount of the jth energy type’s carbon emission in the ith industry. The factors of G, $\frac{G_i}{G},\frac{E_i}{G_i},\frac{E_{ij}}{E_i}, and \frac{C_{ij}}{E_{ij}}$ represent economic scale, energy intensity, industrial structure, and carbon emission coefficient, respectively. The total changes for carbon emission ($\mathsf{\Delta }{C}_{tot}^t$) are decomposed into economic growth ($\mathsf{\Delta }{C}_g^t$), industrial structure ($\mathsf{\Delta }{C}_s^t$), energy intensity ($\mathsf{\Delta }{C}_{ei}^t$), and energy structure ($\mathsf{\Delta }{C}_{es}^t$).

In time [0, t], the change amount of carbon emission is decomposed into the following driving effects:

$$\mathsf{\Delta }{C}_{tot}^t=C^t-C^0=\mathsf{\Delta }{C}_{x_1}^{t-0}+\mathsf{\Delta }{C}_{x_2}^{t-0}+\mathsf{\Delta }{C}_{x_3}^{t-0}\dots \dots +\mathsf{\Delta }{C}_{x_k}^{t-0}$$

(4)

where $\mathsf{\Delta }{C}_{x_1}^{t-0}, \mathsf{\Delta }{C}_{x_2}^{t-0}, \mathsf{\Delta }{C}_{x_3}^{t-0}, \mathrm{and} \mathsf{\Delta }{C}_{x_k}^{t-0}$ are the decomposition share of each driving effect indicating the contribution of each driving factor ($X_i$), which can be calculated by the following equation.

$$\mathsf{\Delta }{C}_{x_i}^{t-0}=\frac{C_i^t-C_i^0}{ln{C}_i^t-ln{C}_i^0}ln\frac{X_i{}^t}{X_i{}^o}$$

(5)

2.2.3. The M-R Spatial Decomposition Model

In this study, the time difference of carbon emission in the LMDI ($\mathsf{\Delta }{C}_{tot}^t=C^t-C^0$) is replaced by the spatial difference between the CEEC of j province ($\mathsf{\Delta }{C}^{R_j}$) and the average level of the YRB (${\mathsf{\Delta }\mathrm{C}}^{R_{\mu }})$; Equation (3) is also correspondingly changed into the following form [25,32,44]:

$$\mathsf{\Delta }{C}_{tot}^{R_j-R_{\mu }}=\mathsf{\Delta }{C}^{R_j}-{\mathsf{\Delta }\mathrm{C}}^{R_{\mu }}=\mathsf{\Delta }{C}_{es}^{R_j-R_{\mu }}+\mathsf{\Delta }{C}_{ei}^{R_j-R_{\mu }}+\mathsf{\Delta }{C}_s^{R_j-R_{\mu }}+\mathsf{\Delta }{C}_g^{R_j-R_{\mu }}$$

(6)

where $\mathsf{\Delta }{C}_{es}^{R_j-R_{\mu }},\mathsf{\Delta }{C}_{ei}^{R_j-R_{\mu }}, \mathsf{\Delta }{C}_s^{R_j-R_{\mu }},\mathsf{\Delta }{C}_g^{R_j-R_{\mu }}$ are the spatial difference of energy structure, energy intensity, industrial structure, and economic growth, respectively. Accordingly, the spatial contribution of each driving factor ($X_i$) can be calculated by the following equation.

$$\mathsf{\Delta }{C}_{x_i}^{R_j-R_{\mu }}=\frac{C_i^t-C_i^0}{ln{C}_i^t-ln{C}_i^0}ln\frac{X_i{}^{R_j}}{X_i{}^{R_u}}$$

(7)

2.3. Data Sources and Processing

The energy data of each province represent the end consumption of each industry and are from the China Energy Statistical Yearbook from 2000 to 2019. The mean interpolation method was used to fill and replace the outlier and missing data in Ningxia province. The average low calorific value and standard coal coefficient of each fossil energy source were taken from the 2019 China Energy Statistical Yearbook. The GDP of each province and the output value of each department (including agriculture, forestry, animal husbandry, fishery, industry, construction, wholesale and retail, transportation, warehousing and postal services, accommodation and catering, and other industries) were sourced from the provincial statistical yearbooks (2000–2019) and corrected based on constant prices in 2000.

3. Results

3.1. Spatiotemporal Characteristics of CEEC in the YRB

The amount of CEEC in the YRB from 2000 to 2018 was calculated and is shown in Figure 2. On the whole, the amount of CEEC in the YRB rose rapidly from 2000 to 2012, and then showed a fluctuating trend. The CEEC in the YRB was 460.708 million tons in 2000 and reached a peak of 1524.58 million tons in 2012, and then fluctuated on the line of 1400 million tons from 2013 to 2018. Within the basin, the distribution pattern of “downstream > upstream > midstream” was visible from 2000 to 2012, which then changed into the distribution pattern of “upstream > downstream > midstream” from 2013 to 2018. On the whole, the spatial distribution of CEEC in the YRB is in an unbalanced state, and carbon emission of the provinces located in middle and lower reaches is generally higher than that in the upper reaches. The CEEC of the provinces located in upper reaches of the YRB has continued to rise since 2000, and the growth trend of CEEC in the middle and lower reaches is consistent with the overall trend, reaching the maximum in 2012. The provinces with resource-based economic development, such as Gansu, Inner Mongolia, and Ningxia in the upstream, are dominated by energy-consuming industries, and the emission reduction intensity has been higher than those provinces in the middle and downstream. Among them, the growth rate of CEEC in Inner Mongolia was 400% from 2000 to 2018, with the largest growth rate. Since the 12th Five Year Plan in 2015, the middle and lower reaches of China have focused on energy conservation, emission reduction, and carbon reduction and achieved remarkable results; Among them, Shandong’s carbon emissions decreased from 437.6326 million tons in 2012 to 291.25 million tons in 2018.

This research selected four time sections at four-year intervals, and a CEEC distribution map and per capita CEEC distribution map were drawn (Figure 3 and Figure 4). There are obvious differences in CEEC among the provinces in the YRB. In 2000, the provinces with higher CEEC values were Shandong, Henan, and Shanxi. In 2006, Shandong and Henan had higher carbon emissions. In 2012, Shandong was far ahead of Shanxi, Henan, and other provinces. In 2018, Inner Mongolia and Shandong were high-carbon provinces, while Qinghai Ningxia, Gansu, and other provinces were light carbon emission areas. There was still a large gap between the economic development level of inland areas and that of these areas. It could also be seen from Figure 4 that the provinces of Ningxia, Inner Mongolia, and Shanxi had the highest carbon emissions per capita, while the provinces of Henan, Sichuan, Gansu, and Shaanxi had lower carbon emissions per capita. Therefore, both the total carbon emissions and per capita carbon emissions in the YRB show a large spatial difference. For example, in 2018, the Inner Mongolia Autonomous Region had high carbon emissions, while Shandong Province had high carbon emissions, but its per capita carbon emissions were moderate.

The trend surface maps [44] were drawn to reveal the spatial differentiation pattern of CEEC in the YRB (Figure 5). In the surface maps, the X and Y positive directions represent the east and north, respectively, and Z represents the amount of CEEC. The trend surface maps of CEEC indicate that the spatial distribution pattern of CEEC in the YRB is obviously different, showing an increasing trend from west to east at different time sections, but different from north to south. In 2000, there was an increasing trend in the direction from the north to the south in the YRB. In 2006 and 2012, the upward trend from the north to the south slowed down. In 2018, the distribution changed into the “U” shape. Shandong, Henan, Shanxi, and other places had higher CEEC, while the inland areas had lower emissions. In recent years, the increase in CEEC in Inner Mongolia and the higher carbon emissions in Sichuan Province were the main reasons for the north–south “U” distribution.

3.2. Temporal Influencing Factors of CEEC in the YRB

The influencing factors of CEEC in the YRB in the period from 2000 to 2018 are disintegrated using the Formulas (4)–(7) of LMDI model, and the cumulative contribution rate of economic growth ($\mathsf{\Delta }{C}_g$), energy intensity ($\mathsf{\Delta }{C}_{ei}$), industrial structure ($\mathsf{\Delta }{C}_s$), and energy structure ($\mathsf{\Delta }{C}_{es}$) were investigated (Figure 6). The results are as follows.

Economic growth is the leading influencing factor promoting the growth of CEEC in the YRB. During 2000–2018, the GDP of the YRB increased by more than 5 times (constant price calculation), and the cumulative impact on CEEC reached 117%, which is the largest of all influencing factors, indicating that economic development is the main factor leading to the growth of CEEC in the YRB. Additionally, the lifting effect of economic growth on CEEC in the YRB has continued to increase from 2000 to 2018.

Energy intensity is the main driving force to reduce CEEC. The energy intensity showed a positive effect in the period from 2000 to 2006, indicating that the energy utilization efficiency was not high in this period. However, the energy intensity showed a negative effect in the period from 2007 to 2018. After offsetting some positive effects, the cumulative contribution rate of energy intensity during the whole investigation period reached −21%, and the absolute value of energy intensity showed an increasing trend, indicating that the inhibition of energy intensity on CEEC is increasingly strengthened.

The impact effectiveness of economic structure and energy structure on CEEC in the YRB is relatively weak, and the change trend of the action direction is consistent. Among them, the role of industrial structure in promoting CEEC continued to increase from 2000 to 2008, reached the maximum in 2008, and then continued to decrease. In 2015, industrial structure achieved a positive to negative transformation, with a cumulative impact of 4% in the study period. In 2016, the industrial structure of the YRB was transformed from “two three one” to “three two one” for the first time. The proportion of the three industries was 0.10:0.43:0.47, indicating that the industrial structure adjustment in the YRB has a positive impact on CEEC reduction. Compared with the economic structure, the energy structure has the smallest impact on the CEEC in the YRB, with the cumulative impact rate of only 0.3%. Except 2013, the energy structure had a promotion effect on CEEC from 2000 to 2014, which then turned into an inhibition effect. The absolute value of the inhibition effect is expressed as a continuous increasing trend, and its effect was greater than that reflected in the promotion effect. In general, the influencing effect of industrial structure and energy structure is relatively weak, but with the increase in time, both of them showed inhibition, indicating that their regulation has great potential for energy conservation and emission reduction.

The influencing factors of CEEC from different periods are decomposed and shown in Figure 7. There are obvious differences in the degree and direction of effectiveness of each influencing factor on the change of CEEC in different periods. Economic growth has been the leading force in promoting CEEC in all periods; energy intensity, energy structure, and economic structure all show inhibition as time goes on, in which the inhibition effectiveness of energy intensity is the most significant.

During 2000–2006, the amount of CEEC in the YRB experienced a rapid growth, with the growth rate reaching 126.07%. All factors contributed to the increase in CEEC. Among them, economic growth led to a 102.73% increase in CEEC, which was the largest factor. During this period, the provinces in the YRB were still in the stage of extensive growth, with low energy utilization efficiency and unreasonable economic structure, relying on an industrial development model of “high energy consumption“.

During 2007–2012, the amount of CEEC in the YRB increased by 46.38%, and economic growth led to an increase of 81.78%. Energy intensity, energy structure and economic structure were transformed during this period, which led to a reduction of about 35.4%. Among them, energy intensity played the most important role, leading to a reduction of 31.04%, which indicates that the improvement of technological progress leads to the improvement of energy utilization efficiency and then leads to a decline in energy intensity.

During 2013–2018, the amount of CEEC in the YRB achieved a negative growth, with a reduction of 9.91%. During this period, the economic growth slowed down, and the economic development led to a 42.97% increase in CEEC, which was the lowest in all periods; Under the joint action of other influencing factors, CEEC decreased by 52.88%. Among them, the role of economic structure is prominent, resulting in lower CEEC of 16.48%. The impact of energy intensity is still the largest, with a contribution rate of −34.02%. During this period, the YRB paid more attention to green development, and all provinces sped up the adjustment of backward production capacity and actively used clean energy to improve energy utilization efficiency.

3.3. Spatial Influencing Factors of CEEC in the YRB

The CEEC of nine provinces in the YRB in 2000 and 2018 were decomposed into four parts using the M-R model, as shown in Figure 8. In 2000, the CEEC of Shandong, Henan, Shanxi, and Sichuan had higher values. In 2018, except for the four provinces mentioned above, the CEEC of Inner Mongolia exceeded the average level. Energy intensity and economic growth are the two main factors leading to the spatial differences of CEEC in the YRB.

The value of economic growth ($\mathsf{\Delta }{C}_g$) of Shandong, Henan, and Sichuan is larger than those of the other six provinces. The GDPs of these three regions are far higher than the average level of the YRB, so the energy consumption and CEEC are higher. Specifically, the amount of the energy consumption of Shandong Province in 2000 was 74.98 million tons, which was the largest in all regions, and rose to 219.987 million tons in 2018. However, the GDPs of underdeveloped regions such as Ningxia, Gansu, and Shanxi were lower than the average level of the YRB, which led to the negative effect of economic development on CEEC. Economic growth has the most obvious negative effects in all regions, reaching −69.574 million tons and −24.282 million tons in 2000 and 2018, respectively.

The value of energy intensity ($\mathsf{\Delta }{C}_{ei}$) of Shanxi, Ningxia, and Inner Mongolia is obviously higher than the average level of the YRB. These regions are important coal-production bases in China with large energy production and consumption amounts that caused the increase in CEEC. Among them, Shanxi’s energy intensity effect reached 58.128 million tons in 2000, ranking first in the YRB, followed by Shanxi province. In 2018, Inner Mongolia leaped to the top, with the energy intensity effect reaching 132.752 million tons. The reason was that Inner Mongolia had a single industrial structure and relies too heavily on coal and other resource consumption industries. It can be concluded that the CEEC in these areas will be significantly reduced if the energy efficiency is raised to the average level of the YRB.

The industrial structure ($\mathsf{\Delta }{C}_s$) indicates the impact of regional economic structure differences on CEEC. If the regional industry is relatively developed, the industrial structure value is significantly positive. The industrial structure values of Shandong, Henan, and Shanxi were significantly positive in these two years, and the industrial level of these provinces was higher in the YRB. The industrial structure values of other provinces exhibited negative effects, mainly because their industries were relatively backward. The difference between the regional CEEC caused by the energy structure ($\mathsf{\Delta }{C}_{es}$) values with the average level of the YRB was small.

4. Discussion

4.1. Spatiotemporal Differentiation of CEEC in the YRB

The research result that the CEEC in the YRB has not reached the peak but has shown a trend of convergence is same with the previous research [45], and the research results that the CEEC in the YRB rose rapidly from 2000 to 2012 are proved by the research conclusions of many scholars [41,44,46]. The main reason is the application of the national strategy of Western Development. In 2006, the State Council deliberated the “Eleventh Five-Year Plan” for Western Development, and heavy industries with coal as the main fuel consumed a large amount in the early stage of Western Development, so that the CEEC in the upper and middle reaches of the YRB continued to rise from 2000 to 2012. In 2012, the State Council issued the 12th Five-Year Plan for Energy Conservation and Emission Reduction, and the nine provinces in the YRB have formulated their own emission reduction plans [43]. Since 2012, the middle and lower reaches of China have focused on energy conservation, emission reduction, and carbon reduction and achieved remarkable results. Among them, Shandong’s carbon emission decreased from 437.6326 million tons in 2012 to 291.2492 million tons in 2018. The CEEC in the YRB is in an unbalanced state, and the carbon emission of Shandong is at a high level, which is consistent with the previous research [47].

The distribution pattern of CEEC in the YRB presented the characteristic of “downstream > upstream > midstream” from 2000 to 2012, which has been confirmed by a relevant study [44], and the distribution pattern changed to “upstream > downstream > midstream” from 2013 to 2018. The CEEC of Gansu, Qinghai, and Ningxia in the west is smaller, which is consistent with the research of Huang and Liu [48]. The CEEC of Shanxi, Shaanxi, Henan, and Inner Mongolia increased from 2000 to 2018 because these provinces are resource-rich, which is proved by the research of Wu et al. [49] and Liu et al. [50]. In recent years, the CEEC of Qinghai, Shaanxi, Gansu, Ningxia, and Mongolia has increased significantly, which is consistent with the previous research result showing the deterioration of the ecological environment in the YRB due to the development of resource-intensive industries [51].

4.2. Influencing Factors of CEEC in the YRB

At present, the positive driving effect of economic development on CEEC has become a common consensus in the academic community [40,44], and this study also supported the conclusion. The effect of energy intensity on carbon emission has a significant inhibitory effect and has become a key factor restricting the increase in regional carbon emission. The inhibition effect of energy intensity on CEEC increased significantly in the period from 2007–2012, which is consistent with the previous research [44]. The energy intensity has obvious spatial differences and is the main inhibiting factor of CEEC in Shanxi and Shaanxi, which has been proved by previous studies [52,53]. This study also indicates that the reducing effectiveness of economic structure and energy structure on the reduction in CEEC in the YRB is relatively weak. The reason is that the proportion of the traditional industries with high energy consumption in the YRB is still large, which has formed a great constraint on the effective adjustment of industrial structure [54,55]. After 2010, the role of industrial structure and energy has been highlighted. Energy-driven economy and urbanization driven by the secondary industry have occupied a leading position and become the main driving force affecting CEEC in the YRB. At the same time, along with the industrial transfer from the eastern to the central and western, cities within the YRB mainly develop energy-intensive industries based on their energy endowment advantages [43]. The energy advantages promote the agglomeration of energy-intensive industries, and the industrial isomorphism is significant [56,57]. However, the economic structure transformation, the industrial structure conversion, and the replacement of the old drivers are difficult to enact in the short period. Additionally, the coal-based energy and industrial structure dominated by secondary industry have become the bottleneck factors restricting the “low-carbon” and “green” model [40].

4.3. Policy Recommendations

Based on the analysis results of spatiotemporal differentiation and influencing factors of CEEC in the YRB, this study proposes the following policy recommendations:

(1) Change the economic development mode. All regions in the YRB should actively optimize the economic structure, vigorously develop green industries, take high-quality development as guidance, and handle the relationship with the ecological environment while developing the economy. The development of the overall service industry in the YRB is relatively backward. The upper reaches of the YRB are economically less developed regions in China, which can selectively undertake industries from the eastern region. The provinces of Shandong, Henan, and Shanxi should cultivate strategic emerging industries, actively develop new formats and models, promote the development of new industrialization with the goal of a green cycle, and achieve low-carbon economic development.

(2) Improve energy utilization efficiency and effectively reduce energy intensity. Reducing energy intensity is the main way to reduce the amount of CEEC. All provinces in the YRB should actively encourage technological innovation, and local enterprises should strengthen the new energy-saving technologies and actively eliminate equipment with high energy consumption. Most provinces in the upper reaches of the YRB are located in the Loess Plateau with good lighting conditions and resources, so it is suggested to make full use of the advantages of solar energy resources, develop clean energy, and accelerate the construction of the green energy system. It is suggested to deal with the excessive dependence on coal; focus on eliminating a number of “high energy consumption, low efficiency” enterprises; improve the access threshold for investment in plant construction; and continue to strengthen the energy conservation management of key energy-consuming units to reduce energy intensity in In Shanxi, Ningxia, Inner Mongolia and other resource-based regions. Especially, Shandong Province should give full play to the demonstration role of the national comprehensive experimental zone for the replacement of the old drivers.

(3) Fast industrial transformation and upgrading, promoting low-carbon development of industries. Upstream regions should rely on their resource endowment advantages; vigorously develop clean energy such as wind energy, electricity, and natural gas; and gradually achieve clean and low-carbon development. The middle reaches of the Yellow River Basin are rich in coal resources, further developing low-carbon environmental protection technologies; promoting the application of carbon capture, utilization, and storage technologies; and reducing energy carbon emissions. The lower reaches of the Yellow River Basin should make good use of the existing industrial foundation and further develop and explore new materials and technologies.

(4) Coordinated development and win–win cooperation. The CEEC in the YRB has obvious spatial differences and strong spillover. Only by reducing the differences in carbon emissions in various regions can the nine provinces achieve the green coordinated development. All provinces and regions in the YRB should strengthen industrial cooperation, work together to optimize and upgrade industries, and strengthen industrial complementarity among provinces. All provinces should work together to control air and environmental pollution, strengthen the exchange and sharing of pollution treatment technologies, and realize the effective flow of green energy technology and green industrial technology in the basin.

4.4. Limitations and Future Research

This research studied the CEEC and its influencing factors in the YRB from 2000 to 2018 in provincial level using the LMDI and M-R method, and the important role and mechanism of the regional background on the spatial effect of CEEC have been well revealed and expanded from the provincial scale [53]. The spatial heterogeneity of CEEC among provinces is mainly determined by the regional resource endowment, economic development level, energy intensity, and structure mode. In general, the research methods and contents of this paper can help to formulate precise policies of energy conservation and emission reduction in the YRB. However, there are also some limitations in this study. Firstly, the study level is only at the provincial level, and there are huge differences in the economic development modes and energy and resource endowments within provinces in the YRB, which brings great difficulties to the determination of precise emission reduction areas and the formulation and implementation of policies. Therefore, it is suggested to analyze the temporal and spatial heterogeneity of CEEC and its influencing factors from a multi-scale perspective. Secondly, the data used in this study are social statistical data from the yearbooks in the period from 2000 to 2018. An increasing number of scholars have used night light data for multi-scale estimation of carbon emission [58] and have achieved good estimation results. Future research can optimize the analysis of factors and mechanisms affecting carbon emissions in the YRB in terms of data accuracy and comprehensive indicators by introducing updated multi-source data such as remote sensing data and geographic big data. Thirdly, this study relies on provincial administrative boundaries to divide regions, which is not fully in line with the spatial boundaries of industrial clusters, factories, and enterprises and other carbon emission entities. In the future, it is possible to more accurately measure the space-time evolution and spatial effects of CEEC by dividing the physical boundaries [59]. Fourthly, the positive driving effect of economic development on CEEC in the YRB has formed a common consensus, but the impact effectiveness of other factors and their influence levels are still controversial. Therefore, further research is needed on the social and economic development gaps that may lead to these differences within the region.

5. Conclusions

This study analyzes the temporal and spatial heterogeneity of CEEC and its influencing factors in the YRB from 2000 to 2018 the LMDI and M-R model and draws the following conclusions:

(1) In general, the CEEC in the YRB increased greatly from 2000 to 2012 and expressed a convergent trend after 2012. The distribution pattern of “downstream > upstream > midstream” was presented in the basin from 2000 to 2012, and then “upstream > downstream > midstream” was presented from 2012 to 2018. In terms of space, there are obvious differences in CEEC among provinces. Shandong has always ranked first in terms of CEEC. The CEEC of Qinghai, Gansu, Ningxia, and other inland provinces is lower than those of other provinces, and they are increasing from west to east in different time sections.

(2) Economic development is the leading factor for the increase in CEEC, and energy intensity is the main force to reduce CEEC in the YRB; energy structure and economic structure are important potential factors for reducing CEEC in the YRB. From different periods, the driving effect of economic development on CEEC was the largest in 2001–2005, and the energy structure had the most obvious inhibition on CEEC in 2011–2015. The economic structure showed inhibition for the first time in 2011–2015, while the energy structure began to play an inhibition role in 2006–2010. Since 2010, the industrial structure and energy structure has had an inhibitory effect on CEEC, but the role of energy structure has not been obvious, indicating that structural factors have great potential in future energy conservation and emission reduction.

(3) The spatial heterogeneity of CEEC among provinces is obvious in the YRB. The CEEC of Shandong, Shanxi, Henan, and Inner Mongolia is significantly higher than the average level of the YRB, which is mainly caused by economic development and energy intensity. Shanxi, Ningxia, and Inner Mongolia urgently need to improve energy utilization efficiency.

(4) The transformation of economic growth and consumption patterns requires more attention to promote the implementation of new energy strategies.