Carbon Emissions in the Yellow River Basin: Analysis of Spatiotemporal Evolution Characteristics and Influencing Factors Based on a Logarithmic Mean Divisia Index (LMDI) Decomposition Method

Abstract

The “14th Five-Year Plan” period is a critical period and a window to obtain emission peak and carbon neutrality in China. The Yellow River Basin, a vital location for population activities and economic growth, is significant to China’s emission peak by 2030. Analyzing carbon emissions patterns and decomposing the influencing factors can provide theoretical support for reducing carbon emissions. Based on the energy consumption data from 2000–2019, the method recommended by Intergovernmental Panel on Climate Change (IPCC) is used to calculate the carbon emissions in the Yellow River Basin. The Logarithmic Mean Divisia Index (LMDI) decomposition method decomposes the influence degree of each influencing factor. The conclusions are as follows: First, The Yellow River Basin has not yet reached the peak of carbon emissions. Regional carbon emissions trends are different. Second, Shandong, Shanxi, Henan and Inner Mongolia consistently ranked in the top four in total carbon emissions, with low carbon emission efficiency. Third, Economic development has the most significant contribution to carbon emissions; other factors have various effects on nine provinces.

1. Introduction

The International Energy Agency reported that China’s carbon emissions exceeded 11.9 billion tons in 2021, accounting for 33% of the global share. The 14th Five-Year Plan has set the goals of lowering carbon emissions, formulating and improving the dual control degree of total energy consumption and intensity, and accelerating the construction of the “1+N” policy system. These provide strategic support for the implementation of achieving peak carbon and carbon neutrality. As China’s ecological security barrier, due to its various topographical features, the Yellow River Basin has a remarkable ecological status. Qinghai, Sichuan, Gansu, Ningxia, Inner Mongolia, Shanxi, Shaanxi, Henan and Shandong are crucial ecological function areas with abundant energy resources, and they have a large share of secondary industries. The phenomenon of industrial structure homogeneity is prominent, with the dominant characteristics of energy chemicals, raw materials, agriculture and animal husbandry, etc. The problem of imbalance in economic development, scientific and technological innovation efforts, clean energy use, etc., still exists. The economic expansion mode, which involves significant energy consumption, emissions and low efficiency, has led to the increasing concentration of carbon emissions and exacerbating the conflict between population, resources and the environment. Due to the conflicting restrictions of decreasing carbon intensity and achieving emission peak and carbon neutrality, it is critical to comprehend the current state of carbon emissions and the influencing factors, to promote the high-quality development of the Yellow River Basin and provide a scientific basis for formulating, implementing and evaluating carbon reduction strategies in the basin. Therefore, the scientific measurement and analysis of carbon emission trends and spatiotemporal evolution patterns can help determine the time of emission peak in each province of the Yellow River Basin. Using the LMDI method to analyze the contribution of energy structure, energy intensity, industrial structure, population size and economic growth, this can provide more comprehensive theoretical support for studying regional energy use and industrial structure transformation. Finally, promoting green production and lifestyle can help China achieve emission peak and carbon-neutral goals.

Based on the fossil energy consumption data of nine provinces in the Yellow River Basin from 2000 to 2019, this paper measures the total carbon emissions of each province based on the carbon emissions calculation method recommended by IPCC. It decomposes the factors influencing carbon emissions by applying the LMDI decomposition method, intending to promote carbon emission reduction and accelerate high-quality development of the Yellow River Basin by combining each province’s and region’s actual situation.

The remaining structure is as follows: Section 2 offers a literature review. Section 3 introduces the methodology. Section 4 analyzes carbon emission trends and carbon emission spatial and temporal evolution patterns and introduces relevant factors, such as energy intensity and carbon emission per capita GDP. Section 5 analyses the decomposition of carbon emission influencing factors, Section 6 contains discussions, and Section 7 contains conclusions.

2. Literature Review

Carbon emissions contributes a lot to global climate change. About 45% of atmospheric carbon emissions come from fossil fuel combustion and land use. At present, the research on carbon emissions mainly has the following aspects:

2.1. Accounting for Carbon Emissions

The principles for estimating carbon emissions are divided into production-based and consumption-based principles. Under the production-based principle, carbon emissions are calculated for domestic production, including exports [1]. It ignores where the end-use or end-consumer of the product is located [2]. Under this principle, there are two accounting methods, bottom-up and top-down. The bottom-up approach, in which energy emissions datasets are created from energy data, includes industrial energy consumption, household energy consumption, industrial processes [3], electricity and maritime transport [4]. This approach calculates cities’ carbon emissions and demonstrates the overall potential for mitigating climate change in cities. In the top-down approach, in response to the lack of data and poor quality, Shan [5] developed an energy balance sheet-based carbon emission inventory approach for Chinese cities. It determines better the significant elements that influence carbon emissions, through data on 47 socioeconomic sectors, 17 fossil fuels and 9 primary industrial products. However, inter-regional trade makes it necessary for the region’s carbon emissions to meet its own consumption needs and those of other regions. Calculating carbon emissions based on the consumption-based principle is more accurate, i.e., carbon emissions from production are assigned to the product’s final customers. Carbon emissions under this principle include imports but not exports, and rich countries may reduce their carbon emissions through international trade. Peter developed a consumption-based carbon emissions database that accounts for implied emissions consisting of domestic fossil fuel use and imports minus exports [6], thereby quantifying the role of international trade in carbon emissions and mitigating global air pollution.

2.2. Spatial Relationship of Carbon Emissions

Due to the differences in economic, social and energy, geographic proximity, the cross-regional flow of production factors, technological spillover effects and mutual imitation of production innovations, surrounding regions are more closely connected [7]. Both the Theil index [8] and spatial data analysis [9] indicate that carbon emissions are spatially correlated and heterogeneous [10,11] and show the Matthew effect at the city level [12]. The distribution pattern of carbon emissions in different industries shows different characteristics. The carbon emission efficiency distribution in China’s building industry shows a tendency of high in the east and low in the west; the carbon emissions of traditional agricultural provinces are relatively high [8]. The carbon emissions of household energy consumption exhibit evident spatial dependence and spatial heterogeneity in terms of the spatial distribution characteristics of high-value agglomerations and the gap between settlement types [13]. The Markov chain results show an obvious spatial spillover effect of carbon emission efficiency in China, with provinces with high carbon emission efficiency positively impacting neighboring provinces, and provinces with low carbon emission efficiency harming neighboring provinces [9]. Using spatial econometrics to study carbon emissions exhibited a significant spatiotemporal correlation. Increases in carbon emissions in nearby regions cause local carbon emissions to rise, but economic growth in neighboring regions suppresses increases in local carbon emissions through indirect effects [14].

2.3. Influencing Factors of Carbon Emissions

Understanding the factors affecting carbon emissions can better make recommendations for carbon reduction. Scholars used the spatial panel model [15] to investigate China’s carbon emission intensity convergence. They verified the positive impact of an industrial structure upgrade [16], energy consumption structure [17] and technological progress [18] on carbon emissions reduction and the negative influence of provincial GDP, and population density on carbon emission reduction [17]. Similarly, AMDI and LMDI are being used to investigate the factors that influence carbon emissions in Greece, and it is found that energy intensity contributes to the reduction of carbon emissions. At the same time, the population effect increases carbon emissions [19]. However, the AMDI model contains residuals. When the value in the data is zero, the model is invalid [20]. The LMDI method solves this problem and is applied to various sectors. In the industrial sector, the constraints of energy consumption structure, energy intensity and industrial structure to emission are increasing [21]. In the transportation sector, the crucial factor in limiting carbon emissions is the state of technology effect; the energy intensity effect promotes the process of decoupling carbon emissions, while the energy consumption structure hinders the process of decoupling carbon emissions [22]. The per capita arable land area and rural population in the agricultural sector are main factors impacting CO₂ emissions [23]. The impact of several absolute and relative variables on carbon emissions is considered by GDIM, which is a new index decomposition model [24]. In the factor analysis of the construction industry, energy use contributes to carbon emissions in a positive way that grows yearly. Energy intensity has a huge potential for future emission reductions. In the power sector, GDP significantly contributes to carbon emission reductions. Factors such as energy consumption structure, industrial structure, energy intensity, GDP and population size in various sectors will have a specific impact on carbon emissions.

2.4. Prediction of Carbon Emissions Trends

Predicting carbon emissions can also help us better test whether the results of implementing carbon emission reduction policies are ideal and analyze which circumstances the carbon emission reduction effect is better. In recent years, system dynamics models have frequently been used in urban carbon emissions prediction research [25,26] to dynamically simulate the carbon emission trends of various provinces and cities in China under different conditions. It found the most favorable situation for emission reduction and provided new ideas for emission reduction for planning decision-making departments. In the long run, China’s carbon emissions will peak in 2036. The date of emission peak in the agricultural sector is 2026, according to the carbon Kuznets curve (CKC), a theoretical model for predicting peaks. The reason why emission peak is delayed can be attributed to the delay in the peaks of the three pillar sectors of industry, construction and transportation [27]. The spatial Markov probability transition matrix results demonstrate that Chinese cities have a “club convergence” phenomenon in carbon emissions, which will gradually improve over time [28]. Although carbon emission is a long-term process, forecasting carbon emission trends and setting short-term emission reduction goals can help slow the rise of carbon emissions. Through the STIRPAT model and the ridge regression model, carbon emission prediction models under seven development scenarios are established. In the next five years, Beijing will have the highest carbon emissions in the next five years under rapid urbanization. The direct impact of carbon trading on energy conservation in the secondary and tertiary industries has promoted a substantial drop in carbon emissions [29]. For the first time, Sun [30] constructed a “decomposition prediction” model in the field of Ensemble Empirical Mode Decomposition (EEMD) carbon emission prediction, combined with a back-propagation neural network based on particle swarm optimization (PSOBP). It improved the accuracy of short-term carbon emission prediction. Due to the emergence of new information in the system development process, it is more accurate to predict and analyze carbon emission data through the grey prediction model of the rolling mechanism. It found that China’s carbon emissions are expected to remain steady in the next few years [31].

In summary, the existing literature conducts carbon emission accounting based on production principles and consumption principles, considering the implied carbon emissions caused by international trade. Analyzing the spatial link of carbon emissions and the spatial spillover effects that follow gives a reference for the collaborative promotion of carbon emission reduction among diverse regions, determining the impact of energy intensity, economic development, and other factors on carbon emissions in various sectors. According to a novel grey rolling mechanism based on the new information priority principle, China’s carbon emissions will remain relatively stable over the next few years. Existing literature rarely studies the carbon peaking situation of the Yellow River Basin as a whole. This paper compares the time when the Yellow River Basin reaches its peak carbon emissions and when each province reaches its peak. The influencing factors need to be decomposed to find the reasons for the gap in carbon peaking time and provide more accurate measurements for the final realization of the carbon peaking in the Yellow River Basin.

3. Methodology

3.1. Research Methods

3.1.1. Carbon Dioxide Emission Measurement

According to a 2021 report from International Energy Agency (IEA), the great bulk of carbon emissions are caused by the usage of fossil fuels. This paper selected nine categories of coal, coke, crude oil, gasoline, kerosene, diesel, fuel oil, natural gas and electricity for analysis, considering the integrity of local data. IPCC guidelines assist countries worldwide in compiling complete national greenhouse gas emission inventories. Meanwhile, it helps maintain compatibility, comparability and consistency between carbon measurement methods across countries. In this paper, carbon dioxide emissions from energy consumption in the Yellow River Basin are calculated through the method recommended by IPCC. The specific calculation formula is as follows:

$${\mathrm{CO}}_2={\displaystyle \sum }_{\mathrm{i}=1}^7{\mathrm{E}}_{\mathrm{i}}·{\mathrm{NCV}}_{\mathrm{i}·}{\mathrm{CC}}_{\mathrm{i}}·{\mathrm{CoF}}_{\mathrm{i}}·\frac{44}{12}$$

(1)

where CO₂ is carbon dioxide emission; i represents coal, coke, crude oil, gasoline, kerosene, diesel, fuel oil, natural gas and electricity; E_i is the energy consumption; NCV_i is the average low calorific value; CC_i is the carbon content per unit calorific value; CoF_i is the oxidation factor and refers to the carbon oxidation rate during energy combustion; and 44 and 12, respectively, refer to the molecular weight of carbon dioxide and carbon.

3.1.2. Logarithmic Mean Divisia Index (LMDI) Decomposition Method

The Logarithmic Mean Divisia Index (LMDI) decomposition method is a type of Index Decomposition Analysis (IDA). Because of its robustness, ease of use and adaptability, it can overcome the shortcomings of the residual term or improper decomposition of the residual term after being decomposed by other methods. It is considered the most preferred decomposition analysis to quantify the factors driving fossil energy consumption and compare the influence degree. Therefore, this paper analyzes the causes of carbon emissions through LMDI. The formula is as follows:

$$\mathrm{I}=\frac{{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{w}}_{\mathrm{ij}}\left({\mathrm{x}}_{\mathrm{i}}- \overline{\mathrm{x}}\right) \left({\mathrm{x}}_{\mathrm{j}}- \overline{\mathrm{x}}\right)}{{\mathrm{S}}^2{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=\mathrm{i}}^{\mathrm{n}}{\mathrm{w}}_{\mathrm{ij}}}.$$

(2)

where ${\mathrm{CE}}^{\mathrm{n}}$ is the total carbon emission in the n-th year, I represents the industrial sector, i = 1,2,3; ${\mathrm{CE}}_{\mathrm{ij}}^{\mathrm{n}}$ represents the carbon emission of j energy from industry i in year n, G represents the output value, P represents the total population of the region, ${\mathrm{CI}}_{\mathrm{ij}}^{\mathrm{n}}=\frac{{\mathrm{CE}}_{\mathrm{ij}}^{\mathrm{n}}}{{\mathrm{E}}_{\mathrm{ij}}^{\mathrm{n}}}$ represents the carbon emission coefficient of energy j of industry i in the n-th year, ${\mathrm{ES}}_{\mathrm{ij}}^{\mathrm{n}}=\frac{{\mathrm{E}}_{\mathrm{ij}}^{\mathrm{n}}}{{\mathrm{E}}_{\mathrm{i}}^{\mathrm{n}}}$ represents the proportion of energy consumption of j energy in the industry i in year n, ${\mathrm{EI}}_{\mathrm{i}=}^{\mathrm{n}}\frac{{\mathrm{E}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{G}}_{\mathrm{i}}^{\mathrm{n}}}$ denotes the energy consumption intensity of industry i in year n, ${\mathrm{IS}}_{\mathrm{i}}^{\mathrm{n}}=\frac{{\mathrm{G}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{G}}^{\mathrm{n}}}$ represents the proportion of industrial output value of I and ${\mathrm{EG}}^{\mathrm{n}}=\frac{{\mathrm{G}}^{\mathrm{n}}}{{\mathrm{P}}^{\mathrm{n}}}$ represents GDP per capita.

Through the LMDI method, the change in carbon emissions can be decomposed into six categories: carbon emission coefficient ($\mathsf{\Delta }$CI), energy structure ($\mathsf{\Delta }$ES), energy intensity ($\mathsf{\Delta }$EI), industrial structure ($\mathsf{\Delta }$IS), economic growth ($\mathsf{\Delta }$EG) and population size ($\mathsf{\Delta }$PP). Since the change in carbon emission coefficient is not evident during the research period, the carbon emission coefficient is not included in the research. The formula is as follows:

$${\mathsf{\Delta }\mathrm{CE}}^{\mathrm{t}}={\mathrm{CE}}^{\mathrm{t}}-{\mathrm{CE}}^0=\mathsf{\Delta }\mathrm{CI}+\mathsf{\Delta }\mathrm{ES}+\mathsf{\Delta }\mathrm{EI}+\mathsf{\Delta }\mathrm{IS}+\mathsf{\Delta }\mathrm{EG}+\mathsf{\Delta }\mathrm{PP}$$

(3)

$$\mathsf{\Delta }\mathrm{CI}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{CI}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{CI}}_{\mathrm{i}}^0}$$

(4)

$$\mathsf{\Delta }\mathrm{ES}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{ES}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{ES}}_{\mathrm{i}}^0}$$

(5)

$$\mathsf{\Delta }\mathrm{EI}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{EI}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{EI}}_{\mathrm{i}}^0}$$

(6)

$$\mathsf{\Delta }\mathrm{IS}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{IS}}_{\mathrm{i}}^{\mathrm{n}}}{{\mathrm{IS}}_{\mathrm{i}}^0}$$

(7)

$$\mathsf{\Delta }\mathrm{EG}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{EG}}^{\mathrm{n}}}{{\mathrm{EG}}^0}$$

(8)

$$\mathsf{\Delta }\mathrm{PP}={\displaystyle \sum }_{\mathrm{i}}{\mathrm{w}}_{\mathrm{i}}\ln \frac{{\mathrm{PP}}^{\mathrm{n}}}{{\mathrm{PP}}^0}$$

(9)

$${\mathrm{w}}_{\mathrm{i}}=\frac{{\mathrm{C}}_{\mathrm{i}}^{\mathrm{n}}-{\mathrm{C}}_{\mathrm{i}}^0}{{\mathrm{lnC}}_{\mathrm{i}}^{\mathrm{n}}-{\mathrm{lnC}}_{\mathrm{i}}^0}$$

(10)

3.2. Data Source

This paper’s original energy consumption data is from the regional energy balance sheet (physical table) of the China Energy Statistical Yearbook and China Statistical Yearbook from 2001 to 2020. Guidelines for Compiling Provincial Greenhouse Gas Inventories (Trial) and IPCC Guidelines for National Greenhouse Gas Inventories 2006 provide carbon content and carbon oxidation rate per unit calorific value. The complete quantity of power used must be translated to standard coal. The carbon emission coefficient is calculated by multiplying the total amount of standard coal by the coal emission coefficient. The Statistical Yearbook of China provides population, GDP, and added value for the three industries, with GDP adopting a constant price model in 1999. Energy consumption is computed using actual consumption data and a standard coal conversion coefficient, while energy consumption intensity and carbon emission intensity are calculated using the total energy consumption to carbon emission ratio.

4. Differences in Carbon Emission Trends among Provinces

4.1. Regional Carbon Emission Trends

4.1.1. Peak Period of Carbon Emissions

Due to the influence of development basis, geographical location, resource conditions, and other factors, the Yellow River Basin has not reached its emission peak. The nodes of carbon emissions reaching the peak during the study period show noticeable differences among provinces. Among them, Qinghai and Sichuan provinces in the upper reaches, Shaanxi province in the middle reaches and Henan province in the lower reaches was the first to peak from 2011 to 2014. These provinces have inland energy and grain production bases and areas with higher economic development and less developed economies, meaning that the factors influencing their carbon peak are likely to differ. Shandong’s total carbon emissions are slowing among provinces that have not yet reached their peak, while Ningxia, Inner Mongolia, Shanxi, and Gansu continue to increase.

4.1.2. Peak Path of Carbon Emissions

Each province’s peak process of carbon emission is diverse during the study period. The increase in total carbon emissions in Qinghai province presents a trend of “slow–fast–slow–fast,” and the growth rate of carbon emission before peaking is significantly faster than in the previous stage. After a relatively short growth period, the growth rate of carbon emission in Henan province tended to be stable in 2007. This may be because China listed energy conservation and emission reduction as a binding index in the economic growth plan, and Henan intensified energy conservation and emission reduction work. The momentum of the excessive development of heavy industry products was curbed. It then lingered in the higher emission zone for a long time before peaking. From 2006 to 2010, Sichuan strictly controlled the blind development of high-polluting industries and accelerated the elimination of backward production capacity. Shaanxi is developing low-carbon technologies and carrying out pilot projects in the province. Before reaching the peak, the carbon emission rate in the two provinces is stable without significant fluctuations.

4.1.3. Post-Peak Carbon Emission Situation

As seen from Figure 1, from 2000 to 2019, the changing trend of carbon emissions of nine provinces and regions after reaching the peak is also diverse. After reaching the peak, Qinghai showed a “double peak” phenomenon, falling and rising, indicating a rebound trend. Qinghai’s overwhelmingly dominant livestock sector, which generates many carbon emissions, and the small size of farming, which has relatively weak carbon sequestration capacity, are likely to rise again after emissions fall. Although Henan province declined slowly, it showed a steady downward trend. This may be because Henan has achieved results in phasing out outdated production capacity, reducing emissions in the four major sectors of industry, construction, transportation and public institutions, and enhanced forest carbon sink function. Although Sichuan experienced a rapid decline in the past, there has been an upward trend in recent years. While the transformation of social and economic greening is accelerating rapidly, Sichuan, a province with a large population, resources, and economy, has a weak ideological understanding and capacity. It will lead to a significant gap in the development basis and policy actions of emission peak and carbon neutrality, and the phenomenon of carbon emissions rebound. After carbon peaked in Shaanxi province, carbon emissions first showed a downward trend, and then a stable trend. From 2011 to 2019, Shaanxi formed the CCUS technology development mode with multi-party participation. It has promoted the pilot construction of a national climate-resilient city and established the carbon emission environmental impact assessment system, which to some extent promoted the work of emission peak and carbon neutrality.

4.2. Evolution of Spatial and Temporal Patterns of Regional Carbon Emissions

According to the classification of natural discontinuities, the total carbon emission is divided into five levels. As seen from the temporal and spatial pattern evolution chart of carbon emissions (Figure 2), the regional distribution of carbon emissions at the provincial level is highly unbalanced. Carbon emissions are high in the upper reaches and low in the lower reaches. The number of high and low carbon emissions areas remains stable. The high carbon emissions area is distributed in Qinghai province, and the low carbon emissions area is distributed in Shandong Province. Overall, the basin is China’s primary energy (coal) and secondary energy (electricity) production and supply base. The energy structure is always short of oil, has less gas and is relatively rich in coal. The provinces and regions are mainly resource-based industries and traditional manufacturing industries. The main reason the upper reaches have the lowest carbon emissions is that the upper reaches, especially the Sanjiangyuan area, as a source of the Yellow River, are constrained by the strict environmental protection policy. They restricted by terrain conditions, it is difficult to carry out large-scale industrial construction in the upper reaches like the middle and lower plains, and carbon emissions are relatively low. The industrial vitality of the downstream region is significantly higher than that of midstream and upstream. In the process of inland transfer, most of the undertakings are mid-to-low-end industries in the eastern coastal areas, which has become one of the constraints of emission reduction.

Specifically, from 2005 to 2010, the number of sub-high value regions of total carbon emissions increased, and the growth rate remained relatively short. The period from 2010 to 2015 coincided with the formation of the Paris Agreement negotiations. During this period, sub-low-value areas increased, mainly distributed in the middle and upper reaches. This indicates that China actively formulated relevant policies and assumed the responsibility for emission reduction. From 2015 to 2019, the distribution pattern of carbon emissions remained unchanged. However, in terms of the total amount, the growth rate slowed down compared with the previous two periods, which may be related to the proposal of new development concepts and high-quality development. Meanwhile, the National Climate Change Plan (2014–2020) released in 2014 also provides a goal-oriented approach to low-carbon development.

4.3. Regional Carbon Emission Characteristics

For the convenience of analysis, the whole research period is roughly divided into four stages according to China’s five-year plan: 2001–2005, 2006–2010, 2011–2015, 2016–2019. Because of the lack of data in 2020, for data comparability, the data for every period is average. The proportion of carbon emission, energy consumption, energy consumption intensity and per capita GDP of each province in the Yellow River Basin were compared (

Table 1. Comparison of relevant indicators among provinces and regions in the Yellow River Basin from 2001 to 2005.

Period	Figures	Qinghai	Sichuan	Gansu	Ningxia	Inner Mongolia	Shanxi	Shaanxi	Henan	Shandong
2001–2005	proportion of carbon emissions (%)	1.00	8.85	6.53	2.68	10.83	24.09	7.05	15.37	23.59
	proportion of energy consumption (%)	1.00	8.73	6.82	2.66	10.71	23.36	7.35	15.21	24.16
	energy consumption intensity (ton/thousand CNY)	204.8	121.0	358.6	504.2	336.6	685.2	225.9	171.2	160.9
	GDP per capita (CNY/person)	6860.83	6653.6	5629	6843.5	9999.0	7725.2	6656.7	6959.8	12,344.31

,

Table 2. Comparison of relevant indicators among provinces and regions in the Yellow River Basin from 2006 to 2010.

Period	Figures	Qinghai	Sichuan	Gansu	Ningxia	Inner Mongolia	Shanxi	Shaanxi	Henan	Shandong
2006–2010	proportion of carbon emissions (%)	1.03	8	6.77	2.87	14.41	18	6.95	15.4	26.57
	proportion of energy consumption (%)	1.01	7.96	6.90	2.82	14.20	7.27	17.54	15.14	27.16
	energy consumption intensity (ton/thousand CNY)	217.7	110.0	388.9	576.8	342.4	574.7	222.3	168.7	192.7
	GDP per capita (CNY/person)	11,528.2	12,253	9576	10,900	23,380.2	12,113	12,239	13,129.6	20,591.7

,

Table 3. Comparison of relevant indicators among provinces and regions in the Yellow River Basin from 2011 to 2015.

Period	Figures	Qinghai	Sichuan	Gansu	Ningxia	Inner Mongolia	Shanxi	Shaanxi	Henan	Shandong
2011–2015	proportion of carbon emissions (%)	1.14	7.22	8.42	4.1	16.38	17.01	6.31	13.3	26.13
	proportion of energy consumption (%)	1.14	7.49	8.75	0.90	16.58	6.76	17.11	13.49	27.80
	energy consumption intensity (ton/thousand CNY)	183.2	73.8	370.4	140.5	279.7	477.2	149.5	115.2	158.2
	GDP per capita (CNY/person)	18,969.4	22,103	16,162	17,224	41,949.9	21,231	17,463	21,958.9	31,892.58

and

Table 4. Comparison of relevant indicators among provinces and regions in the Yellow River Basin from 2016 to 2019.

Period	Figures	Qinghai	Sichuan	Gansu	Ningxia	Inner Mongolia	Shanxi	Shaanxi	Henan	Shandong
2016–2019	proportion of carbon emissions (%)	1.07	5.57	8.72	4.99	17.2	18.85	4.85	10.68	28.07
	proportion of energy consumption (%)	1.04	5.67	8.71	4.89	16.69	4.96	18.31	10.50	29.24
	energy consumption intensity (ton/thousand CNY)	136.3	45.4	311.3	636.4	243.7	465.7	89.2	72.8	139
	GDP per capita (CNY/person)	25,903.4	30,766	21,773	22,950	55,413.2	29,597	21,713	30,781.6	43,004.9

). In terms of the proportion of carbon emissions, Shandong, Shanxi, Henan, and Inner Mongolia have always been in the top four. In 2019, the total carbon emissions of these four provinces accounted for 74.74% of the total amount of the Yellow River Basin, and their GDP accounted for 45.09%. The carbon emission efficiency of these four provinces is relatively low. These four provinces are distributed along the Yellow River Basin and have different natural resource endowments and industrial structures, so the factors leading to their large emissions are also different.

From the perspective of energy consumption intensity in each province, Ningxia and Shanxi’s energy consumption intensity is much higher than other provinces. On the one hand, due to relatively poor economic growth, Ningxia and Shanxi’s energy-consuming industrial structure is not rational. On the other hand, the energy-consuming industrial structure is not rational due to the rich coal resources in Shanxi and a high degree of dependence on coal. In addition, Ningxia’s economic growth level is the worst among the provinces. Still, its energy consumption is large, so the energy consumption of Shanxi and Ningxia is much higher than that of other provinces. The provinces with higher energy consumption are mainly in the central region with relatively high economic growth, while the provinces with lower energy consumption are in the western region with relatively backward economic growth. As can be seen from the linear change trend results, energy carbon emissions are rising and this trend will continue in the future. This is because the energy carbon emissions have great economic development inertia. In the short term, it is difficult to achieve both economic growth and energy reduction but the constraining of the increase of energy consumption can be attempted, ultimately achieving emission reduction targets.

5. LMDI Model Decomposition

The driving factors of carbon emission have various roles in different provinces. Using the LMDI method to decompose the carbon emissions of each province from 2001 to 2019, obtains the influence of five factors of energy consumption structure, industrial structure, energy intensity, population size and economic development on the change of carbon emissions and further obtains the annual average impact of each factor at each stage (the yearly average of the change in carbon emissions caused by each factor and the percentage of carbon emissions at the beginning of the stage). The evolution characteristics of each element and its influence on the four stages are different in different provinces.

5.1. Energy Structure

Throughout the historical research period, the energy structure of various provinces and regions has generally changed in the direction that will lead to a reduction in carbon emissions (Figure 3). Specifically, the energy consumption structure of Ningxia has the highest contribution to carbon emissions, but it shows a downward trend. From 2016 to 2019, the contribution rate was 7.4%, a decrease of 34% compared to 2011 to 2015. The second is Qinghai, which contributed 2.45% to carbon emissions from 2016 to 2019, a decrease of 44.44% compared with 2011 to 2015. From 2016 to 2019, the negative to the positive impact of energy consumption structure in Henan and Sichuan on carbon emission reduction was determined. Only Shaanxi’s energy consumption structure is suited to lowering carbon emissions in all four stages. Ningxia has the most significant energy consumption intensity. Coal use has increased due to the “transmitting electricity from west to east” policy in Ningxia. China’s most prominent modern coal industry demonstration area has begun transforming a hydrogen energy leader to reduce environmental pressure. New materials, new energy and ten other industries are focused to build substantial new industrial bases in the country. Since 2015, the output value of high-tech industries has grown by over 20%. Although Shaanxi is a significant energy province, its scientific and technological advantages provide strong support for breaking through the “resource curse” and transforming into low-carbon and high-efficiency. The proportion of energy and chemical industries continues to decline, showing great potential for carbon emission reduction. Following the financial crisis of 2008, China made significant investments in the infrastructure sector and high energy-consumption sectors, such as metallurgy and construction materials, use much energy in Henan. After 2015, renewable energy development in Henan intensified and the energy consumption structure began to develop toward a green direction; the impact on carbon emission reduction is gradually turning positive. In 2016–2019, the country’s most extensive clean energy base will be built in Sichuan Province, and the energy consumption structure will be transformed into low-carbon.

5.2. Industrial Structure

Each province’s industrial structure has generally shifted toward lowering carbon emissions (Figure 4). Specifically, the industrial structure changes in Shanxi, Shandong and Gansu are conducive to carbon emission reduction in 2016–2019. The industrial structure of Shaanxi, Qinghai and Ningxia contribute the most to carbon emissions. Among them, the equipment manufacturing industry is the leading industry in Shaanxi. The high-tech industry has a low concentration, limited by the “black” industrial structure, which easily enhances carbon emissions. From 2006 to 2010, the contribution rate of Qinghai’s industrial structure reached 15.36%, while the contribution rate from 2016 to 2019 was only 1.92%, a drop of 86.8%. Qinghai’s pillar industries are concentrated on energy-dependent industries, such as high oil, power and other industries. The secondary sector contributes a larger share of carbon emissions. The “Belt and Road” promotes exchanges with countries and regions along the route, realizes the industrial transfer and adjustment of industrial structure, simultaneously promotes the undertaking and emergence of high-end manufacturing technologies, optimizes industrial layout, and lowers industrial structure’s contribution to carbon emissions. With the support of special funds in the province, Gansu Province began to focus on new materials and new energy in 2014 and made joint efforts to develop strategic emerging industries as a whole. Shandong has changed the development route of traditional industries since 2013, cultivated and developed emerging industries and conducted comprehensive experiments on converting new and old kinetic energy, which have produced good results. Shanxi, in 2016–2019, reformed the supply-side structure continuously and traditional industries have continued to improve, which has an enhanced role in carbon emission reduction.

5.3. Energy Intensity

The contribution of energy intensity to carbon emissions in each province has been decreasing throughout the period (Figure 5). From 2016 to 2019, Shanxi, Inner Mongolia, Henan, Sichuan, Shaanxi, Gansu and Qinghai had a favorable impact on carbon emissions. It still has a positive impact in Shandong. However, compared with 2011–2015, the degree of influence has dropped by 70%, and only Ningxia’s energy intensity’s contribution to carbon emissions has maintained a high level, reaching 37%. The primary energy sources for Ningxia’s industrial development are coal and electricity. The energy structure is unreasonable, and the use efficiency is low, which makes the changes in the secondary industry’s share have a more outstanding influence on carbon emissions. From 2001 to 2009, Shandong Province was dominated by heavy chemical industries, and the extensive growth of high-emission enterprises resulted in a greater contribution of energy intensity to carbon emissions. From 2016 to 2019, Shandong Province’s new energy and the renewable energy industry have outgrown. The utilization of high-quality resources outside the province has reached a new level, the situation of heavy energy consumption has eased, and its contribution to carbon emissions has also weakened. Among the remaining provinces, Shanxi is promoting energy transformation from black to green and Inner Mongolia is actively promoting clean energy transformation. Guided by the low carbonization of the energy system, the energy-saving industries and green lifestyles, Sichuan has comprehensively improved energy utilization efficiency and enhanced its role in reducing carbon emissions. In general, the “dual control” target of energy consumption in 2016–2019 has strengthened the inhibitory effect of energy intensity in each province.

5.4. Population Size

The favorable impact of each province’s population size on carbon emissions showed a rising trend in general, but the impact of individual provinces showed a specific difference (Figure 6). Only Sichuan Province is favorable for carbon emission reduction at all stages, and the impact is gradually weakening, which matches the changing trend of the total population. Henan Province’s contribution to carbon emissions during 2006–2010 and 2011–2015 was negative and finally turned into a positive impact. The population size contribution in Qinghai to carbon emissions demonstrated an upward trend, and the contribution rate reached 43.04% from 2016 to 2019. Although the contribution rate of Ningxia ranks second, the contribution rate has gradually reduced by 18.84% since the “Eleventh Five-Year Plan.” In particular, the lower Yellow River area carries 40.36% of the population with 10.17% of the area. The dense spatial structure restricts the efficiency of industrial production and determines that cities along the Yellow River are areas with high carbon emission density. On the one hand, population increase can increase the density of buildings, which is conducive to the centralized supply of energy, thereby improving energy efficiency and reducing reduction. On the other hand, the excessive expansion of the population has increased the demand for electricity consumption and transportation, highlighting negative externalities, such as increased energy consumption and severe environmental pollution. At the same time, changes in residents’ production and lifestyle have brought more carbon emissions.

5.5. Economic Growth

Each province’s economic expansion has contributed to an increase in carbon emissions, and the trend will continue (Figure 7). Specifically, economic growth in Qinghai, Sichuan, and Shaanxi was the most responsible for carbon emissions, exceeding 100% in the last stage. Shandong’s economic growth contributes the least to carbon emissions, and the growth rate shows a trend of slowing down year by year, while Ningxia is shifting in a way that will help reduce carbon emissions. After 2000, the implementation of strategies, such as the development of the western region, has prompted the Yellow River Basin to enter a stage of rapid development, leading to a significant rise in energy consumption and carbon emissions. The Copenhagen Conference in 2009 included carbon emission reduction into the economic growth indicators. Under the background of “low carbon,” the growth rate of carbon emission began to slow down. Even so, economic expansion contributed significantly to carbon emissions in the Yellow River Basin. The contribution rate of Qinghai’s economic growth to carbon emissions has risen rapidly from 198.61% in 2000–2005 to 469.49% in 2016–2019, an increase of 139.69%, which is consistent with the rising trend of Qinghai’s GDP. Shaanxi’s economic growth is dominated by high-carbon industries, especially Yulin City in northern Shaanxi plays a vital role in “West–East Coal Transportation” and “West–East Electricity Transmission”. The contribution rate of Sichuan’s economic growth increased from 39.21% in 2000–2005 to 122.92% in 2016–2019, an increase of 213.49%. Shandong began to build a blue economic zone in 2011, with rapid economic growth, but carbon emissions also showed a fast upward trend. The rate at which economic growth contributed to carbon emissions was 14.1%. From 2013 to 2016, the blue economic zone was completed. Affected by the overheated development during the construction period, high energy consumption and pollution levels began to occur, and the contribution rate rose to 15.53%.

6. Discussion

Accelerate industrial upgrading in the central cities, radiate and drive industrial transformation in small and medium-sized cities. The midstream and upstream of the Yellow River Basin is China’s coal industry concentration area, with a significant amount of coal and coal power, and is planning to build modern coal chemical projects. Strengthening environmental protection management of the coal industry in the midstream and upstream, promoting the coordinated development of the coal industry and environmental protection are preconditions for carbon emission reduction in the Yellow River Basin. Restricted by geographical conditions, the economic connection degree of the provinces is not high, the regional division of labor and cooperation consciousness are not strong and the efficient, coordinated development mechanism is not perfect. The central city should be taken as the development engine, combined with the differences in the natural environment and social and economic conditions along the Yellow River. A circular and efficient industrial system with reasonable structure should be built and the current situation dominated by resource-based and heavy chemical industry structure should be changed, realizing industrial division and cooperation among the upper, middle and lower reaches of urban agglomeration. Full play to the supporting and driving role of leading enterprises should be given, and overall plans for the supply and demand side of the energy chain, the upstream and downstream of the industrial chain, and the front and rear ends of the supply chain should be made. Build an industrial ecosystem with a reasonable division of labor, complete supporting facilities and firm support, and accelerate the process of advantageous green industries with centralized distribution, clustered chains, and intensive and efficient development. Improve the system of regional industrial policies and formulate new standards for assessing local governments’ ecological protection and economic development. Differentiated assessment systems based on location and resource endowments should be established, the construction of environmental supervision systems, introduce relevant strategies and policies should be promoted, and the government’s environmental control and supervision mechanism to prevent and control pollution should be strengthened. The digital empowerment of industries and the encouragement the deep integration of new technologies, such as big data, AI and 5G with green and low-carbon industries can be accelerated. The penetration and coverage of digital technology to industrial development should be enhanced, a local clean energy industry’s Internet platform system should be built and digital systems on the energy supply and demand side should be supported. The downstream cities should take advantage of technological innovation to improve the basin’s overall industrial innovation input level through an intergovernmental cooperation mechanism and establish a cross-regional coordination mechanism and improve the carbon trading market’s supervision system in order to realize coordinated carbon emissions management on a whole basin scale.

Innovative breakthroughs should be sought and should guide the low-carbon energy transition. Energy transformation should be “open source” and should “reduce expenditure.” On the one hand, renewable energy should be vigorously developed, turn traditional elemental energy into standby regulatory energy, and gradually form an energy system with new energy. On the other hand, we need to limit the total consumption of high-carbon fossil energy and reduce carbon emissions through energy conservation and efficiency improvement. Systems for budgeting energy consumption, monitoring energy consumption, trading energy and evaluating energy conservation should be improved. Henan and other central regions have fared better in transformation and upgrading than western regions. Although the energy intensity of Henan province has been dramatically reduced, it is still higher than rich countries’ average in the past few years, so it still has great potential. Henan and other central provinces should expand the reduction effect of energy intensity through emission reduction technology progress. The diversified allocation of low-carbon energy to achieve carbon emission reduction should be increased, energy demand growth should be limited, and demand-side management should be improved. In addition, coal is still the most common energy source in China’s provinces, especially in western regions like Inner Mongolia; economic growth is dominated by energy-intensive businesses, which uses much energy and emits much carbon. Through energy structure optimization, there is still a lot of room for emission reduction, making it difficult to achieve the emission reduction target. The dependence of industry on coal and the use of fossil fuels as a source of energy should be decreased. The production of renewable energy should be further promoted. The introduction and innovation of advanced technologies should be encouraged. A new technology is the friction-welded joint made of 316L stainless steel. It takes the heat of fusion from friction and does not cause any gas emissions to the environment [32]. To some extent, it can reduce the amount of energy used and carbon emitted in the western region.

The financing channels for a low-carbon economy should be expanded, and the decoupling of economic growth from carbon emissions should be accelerated. Economic growth is a necessary material basis for improving carbon emission efficiency, which can enhance financial support for implementing regional carbon emission reduction policies and provide personnel, technology, and management guarantee for improving carbon emission efficiency. In midstream and downstream, where the economic growth level is higher, in addition to paying attention to the decoupling of its economic development and carbon emissions, it is necessary to support a creative carbon financial market and improve the carbon financial market in the midstream and upstream. In terms of decoupling from economic growth, we should actively guide key industries to further implement cleaner production transformation and accelerate the promotion and application of carbon capture, utilization, and storage technology (CCUS). Additionally, a circular economy should be developed and should strive to cultivate new growth points in ecological and environmental protection industries. Steady progress will be made in the pilot development of national low-carbon and climate-resilient cities, and efforts will be made to establish mechanisms for innovative research and development of low-carbon technologies by governments, research institutions and enterprises. In terms of expanding financing channels for a low-carbon economy, we will vigorously develop green finance. Support eligible enterprises in green and low-carbon industries to go public for financing and issue bonds. Guide banks and other financial institutions to actively use the support tools to provide funds for low-carbon industrial projects in the upstream and midstream by promoting the exchange of resources in Lanxi, Hu, Bao, Eyu and Guanzhong, encouraging urban agglomeration, integrating financial resources, optimizing financing structure, strengthening financial interaction, pushing green low carbon financial product, and creating support conditional region green financial innovation pilot test area, carrying out the national climate investment and financing.

National environmental protection awareness should be cultivated in order to build a green, low-carbon, sustainable modern cities. For provinces like Henan and Shandong, high population density and rapid urbanization resulted in an increase in carbon emissions, which should be divided into important carbon emission reduction areas, key reduction areas and reduction concern areas according to the differences between provinces. For important carbon emission reduction areas, improving the quality of urbanization should be prioritized, reducing per capita carbon emission efficiency, appropriately guiding population evacuation and transfer, and reducing carbon emission rebound. For the crucial areas, we should encourage the usage of renewable energy by residents, strengthen the publicity of carbon peak and carbon neutrality policies, demonstrate green and low-carbon social acts, and lead the new trend of green and low-carbon life. For carbon emission reduction concern areas, the problem should be addressed, such as increasing residents’ carbon emissions in urban fringe areas and urban agglomerations and cultivating residents’ awareness of low carbon. For sites with a relatively sparse population, such as Qinghai, Inner Mongolia and Ningxia, there is ample opportunity to reduce carbon emissions in the early stage of population agglomeration and economic growth. Boost energy-saving and emission-reduction technological innovation through rapid upgrading of public transportation and other infrastructure and release the scale effect of population and economy to promote carbon reduction as soon as possible. Individuals with a higher degree are more motivated to live a low-carbon lifestyle, so carbon emission efficiency in areas with a higher proportion of high education population is more elevated. The government should actively advocate and lead low-carbon consumption to produce a continuous demonstration effect in abandoning luxury consumption. At the same time, enterprises, public institutions and the private sector should be encouraged to adapt to the new consumption concept, increase the publicity of low-carbon consumption and accelerate the transformation of consumers from traditional consumption concept to a wide low-carbon consumption concept. Although the population size of Sichuan province has an inhibitory effect on carbon emissions, along with the aging problem gradually intensified, the new “three-child” policy has been fully implemented. The future population size effect on energy consumption and carbon emissions are driving impact of Sichuan province remains to be further strengthened. Still, from the current situation of population growth, the implementation of the birth encouragement policy has not produced the expected population float in Sichuan Province. With the improvement of residents’ quality of life, the green low-carbon lifestyle and consumption concept will gradually take root in people’s consciousnesses. Therefore, it is less likely that the population scale effect will cause a significant carbon emission increase in the future.

7. Conclusions

Using the method recommended by IPCC, the carbon emissions of nine provinces in the Yellow River Basin were measured, and the trends and spatial distribution of carbon emissions were observed. It was found that: (1) The Yellow River Basin has not yet reached its emission peak. The carbon emissions of Qinghai, Sichuan, Henan and Shaanxi took the lead in reaching the peak in 2011–2015, and the carbon emission trends during the peaking process and after the peaking are different. Only Shandong’s carbon emission rate has slowed down among the provinces that have not yet reached the peak, and the rest have maintained the initial growth rate. (2) High-value areas of carbon emissions are mainly located in the upper reaches of the Yellow River Basin, and the low-value areas are mainly located in the lower reaches. From 2011 to 2015, the number of sub-high-value areas decreased. Although the spatial pattern did not change from 2016 to 2019, the increase in total carbon emissions has slowed down.

Shandong, Shanxi, Henan and Inner Mongolia’s carbon emissions are always in the top four. In 2019, these four provinces took up 74.74% of the total carbon emissions, while their GDP accounted for 45.09% of the entire Yellow River Basin. Ningxia and Shanxi have substantially greater energy intensity but much lower per capita GDP than other provinces. The energy carbon emissions of all provinces in the Yellow River Basin are rising and this trend will continue.

The LMDI method to can be used to decompose five factors influencing carbon emission. It was found that: (1) Economic growth has the most significant impact on carbon emissions, especially in Qinghai, where economic growth’s contribution rate to carbon emissions in 2019 reached 469.49%, followed by Sichuan and Shaanxi. Shandong’s economic growth contributed the least to carbon emissions. (2) Although energy consumption structure has a detrimental impact on carbon emission reduction, the overall adjustment of energy consumption structure is developing in a direction that promotes the decrease of carbon emissions. (3) Except for Ningxia and Henan, other provinces and regions have contributed to reducing carbon emissions. In 2019, energy intensity’s contribution rate to carbon emission in Henan was only 0.75%. Although Ningxia contributed 37%, it also showed a downward trend. (4) The rate at which the industrial structure contributes to carbon emissions is about 1% in most provinces. The industrial structure of Shanxi, Shandong and Gansu restrained carbon emissions in 2019. Although the contribution rate of other provinces is still positive, the contribution rate is gradually weakening. (5) Energy efficiency can be improved when the population size reaches a certain level. However, from the current results, only Sichuan’s population size has a weak limiting influence on carbon emissions, with a contribution rate of −0.08%. Carbon emissions are still influenced by population size in other provinces, with Qinghai contributing the most to carbon emissions.