Temporal–Spatial Characteristics of Carbon Emissions and Low-Carbon Efficiency in Sichuan Province, China

Abstract

Clarifying the temporal and spatial characteristics of regional carbon emissions and low-carbon efficiency is of great significance for the realization of carbon peaking and carbon neutrality. This study calculated the carbon emissions in Sichuan Province from 2015 to 2022 based on four major units: energy activity, industrial production, forestry activity, and waste disposal, and its time evolution characteristics and key sources were investigated. Meanwhile, based on the Super-SBM-Undesirable model, the low-carbon efficiency of Sichuan Province and its 21 cities (states) was evaluated, and its spatial heterogeneity characteristics were investigated. The empirical results reveal the following: (1) energy activity was the main contributor to regional carbon emissions, with thermal power generation and industrial energy terminal consumption as the key sectors. Inter-regional power allocation could indirectly reduce the regional emission intensity. The carbon emissions of industrial production showed significant aggregation in cement and steel production. The forest carbon sink had a significant effect on alleviating the regional greenhouse effect. The carbon emissions of waste disposal were small. (2) From 2015 to 2022, the low-carbon efficiency of Sichuan Province showed an overall upward trend. Chengdu had a high level of economic development, a reasonable industrial organization, and a continuous increase in its urban greening rate. Heavy industrial cities such as Panzhihua and Deyang made great efforts to eliminate backward production capacity and low-carbon transformation of key industries. Therefore, they were the first mover advantage regions of low-carbon transformation. Zigong, Mianyang, Suining, and Leshan enjoyed favorable preferential policies and energy-saving space, and were developmental regions of low-carbon transformation. But they need to actively deal with the problem of industrial solidification. The low-carbon efficiency of plateau areas in western Sichuan was relatively low, but they have unique resource endowment advantages in clean energy such as hydropower, so the development potential is strong. Cities such as Ya’an and Bazhong faced a series of challenges such as weak geographical advantages and the risk of pollution haven. They were potential regions of low-carbon transformation.

1. Introduction

Climate change has become a major issue concerning human survival and sustainable development [1]. The AR6 Synthesis Report: Climate Change 2023 issued by the United Nations Intergovernmental Panel on Climate Change (IPCC) further clarified that greenhouse gas emissions generated by human activities are the main cause of global warming [2]. The influence of human activities has caused extensive and rapid changes in the atmosphere, ocean, cryosphere, and biosphere. Specifically, in recent years, the problem of carbon emissions caused by energy systems using fossil energy as their main fuel has become increasingly prominent [3]. At the 75th session of the United Nations General Assembly, the Government of China further strengthened national actions and targets to deal with climate change. It is proposed that China’s carbon emissions will reach their peak by 2030, and the goal is to achieve carbon neutrality by 2060. The proposal of carbon peaking and carbon neutrality goals is a crucial step for China to help the world achieve net zero emissions, which conforms to the development trend of green recovery and low-carbon transformation after the global epidemic [4]. Considering that China is still in the period of economic rise and carbon emission growth, it is necessary to consider the contradiction between carbon emission constraints and economic and people’s livelihood development as a whole.

In recent years, the construction of western China has gradually formed a new development pattern. It is worth noting that the “Land and Space Plan of Sichuan Province (2021–2035)” issued by the Ministry of Land and Resources of China clearly states that Sichuan Province is the strategic hinterland of China’s development, and an important region to support the implementation of national strategies such as the Western Development in the new era and the development of the Yangtze River Economic Belt. In this regard, the important position of Sichuan Province in the process of Chinese national rejuvenation is increasingly prominent. In addition, as a clean energy base and energy transmission hub, Sichuan Province should actively respond to China’s emission reduction commitment to the international community [5]. However, at this stage, it is not obvious that the advantage of clean energy endowment in Sichuan Province can be actively transformed into the advantage of low-carbon transformation [6]. In 2022, the proportion of coal fuel consumption in Sichuan Province was 32.96%, and the proportion of oil fuel consumption was 22.76%. Heavy industries represented by steel, building materials, and chemicals are still the important foundation for regional economic and social development. Meanwhile, Sichuan Province is a populous province in China, with a permanent population of 83.74 million in 2022, and the carbon emission contribution of residential life cannot be ignored. In this regard, the regional development path is still characterized by high energy consumption and high carbon emission. The main reasons for the above phenomenon are that, on the one hand, the economic and social development level of the central region where Sichuan Province is located is lower than that of the eastern coastal regions as a whole. On the other hand, the development heterogeneity of different cities (states) in Sichuan Province is significant. In 2022, the GDP of Chengdu, ranked first, reached CNY 2081.75 billion, 5.74 times that of Mianyang, ranked second, and 45.01 times that of Aba Tibetan and Qiang Autonomous Prefecture, ranked last. In this regard, regional industrialization and urbanization are still advancing, which depends on the stable and high-level power supply of fossil energy [7]. Clean and low-carbon production methods and lifestyles need to be popularized. Meanwhile, Sichuan Province, as the main starting point of national energy allocation projects such as “Power Transmission from West to East”, exports a large amount of hydropower every year to ensure energy supply security in the eastern coastal areas [8]. In 2022, the provincial power export volume was 155.81 billion kWh, accounting for 32.15% of the total power production. Therefore, the lack of local consumption of clean energy will also affect the regional low-carbon development process.

In this case, investigating the characteristics of regional carbon emissions and low-carbon efficiency is of great significance for the realization of carbon peaking and carbon neutrality. Specifically, carbon emission calculation is an important prerequisite for accurately determining the key sectors of carbon emission, and regional low-carbon efficiency evaluation is an important basis for clarifying the current situation of regional sustainable development. Meanwhile, the economic development and resource endowments of different cities (states) in Sichuan Province are unbalanced, so it is difficult for them to go down the same path to achieve low-carbon transformation [9]. Therefore, this study comprehensively considers the analysis of temporal and spatial heterogeneity, which can provide an empirical basis for the formulation of differentiated low-carbon development plans [10]. From this perspective, the major contributions of this study are as follows: (1) the analysis perspective of existing research mainly focuses on the emission characteristics of a single sector such as energy activity [11,12,13], so the results cannot reflect the differences in carbon emission characteristics caused by the heterogeneity of production factors among multiple sectors. In this study, regional carbon emission calculation is carried out based on four major units: energy activity, industrial production, forestry activity, and waste disposal, which is conducive to scientifically investigating the key emission sources from a holistic perspective. Meanwhile, this study considers the indirect emissions of regional electricity trading and forestry carbon sinks, which is helpful to further investigate the attribution of carbon emission responsibility and the emission reduction effect of forest activity. (2) When carbon emission is regarded as an undesirable output, the existing low-carbon efficiency research mainly focuses on the direct emission contribution of carbon sources [14,15,16,17,18,19]. In order to comprehensively evaluate the impact of carbon emission on low-carbon efficiency, the indirect carbon sequestration effect of forest carbon sinks should also be considered. In addition to this, as the development of new quality productivity becomes a new national strategy in China, the important role of technological innovation in regional low-carbon development is becoming increasingly significant. Therefore, it is necessary to consider the energy-saving and emission reduction effects of scientific and technological progress when evaluating low-carbon efficiency. In this study, energy efficiency, low-carbon scientific and technological innovation, and land greening level are comprehensively used as the evaluation basis of low-carbon efficiency of Sichuan Province and its 21 cities (states). This will help to limit the scale and speed of economic development to the allowable range of resource-carrying capacity and environmental capacity. Meanwhile, the calculation results of carbon emissions based on multiple emission units can provide basic data for the undesirable output of DEA model, thus improving the accuracy of efficiency evaluation.

With the strategy of “Chengdu-Chongqing Dual City Economic Circle” put forward, it will further promote the coordinated optimization and upgrading of regional industrial structure and realize the effective improvement in the economic development level in Chengdu–Chongqing region. However, the increase in energy consumption and carbon emission sources caused by economic growth is a double-edged sword, which makes the impact of regional development strategy on the overall low-carbon efficiency uncertain [20]. In order to effectively promote the coordinated and integrated development of “Chengdu-Chongqing Dual City Economic Circle”, in the process of regional industrial adjustment, it should fully consider the optimization, transformation, acceptance, and distribution of industries in vulnerable areas [21], rather than just transferring high-energy consumption and high-pollution industries to neighboring cities [22]. In addition to sharing industrial synergy benefits, all regions should also implement a sharing model in environmental responsibility, so as to realize equal treatment and shared responsibility for economic and social development and energy conservation and emission reduction tasks [23]. In this regard, clarifying the temporal and spatial characteristics of regional carbon emissions and low-carbon efficiency will provide an important reference for the design of differentiated transformation paths.

The rest of this article is structured as follows: The relevant literature review is in Section 2. Model and data are in Section 3. The empirical analysis results are in Section 4. The main conclusions and policy implications are in Section 5.

2. Literature Review

2.1. Calculation of Regional Carbon Emissions

At present, the calculation of regional carbon emissions is mainly based on the emission factor method, mass balance method, and actual measurement method.

(1)

The emission factor method mainly investigates the emission level by multiplying the basic activity data of different emission sources by the corresponding carbon emission factors [24]. It is the most widely used calculation method, and is mainly applicable to the macro-level analysis at the national and provincial levels. Its advantages are that it is simple, clear and easy to understand, with mature calculation formulas and sufficient basic activity data from different statistical sources. Meanwhile, the existing emission factor database can provide the default parameters needed for calculation, and there is a large number of application examples for reference. Its disadvantage is that the calculation is mainly carried out at the macro level, and its response ability to the changes in the emission system is poor. Zhou [25] investigated the carbon emission characteristics of 31 provinces, municipalities, and autonomous prefectures in China, except Tibet, based on the emission factor method. However, this study mainly focused on carbon emissions from energy activities, so it was impossible to identify the key emission sources and the evaluation results were not comprehensive. Based on the emission factor method, Darwish [26] investigated the carbon emission characteristics of four cold metropolitan areas, such as Chicago and Boston, and made a comparative analysis with other areas in the United States. However, the research perspective was limited to the carbon emission contribution of land use practice and transportation behavior.

(2)

The mass balance method mainly calculates the share of new chemicals consumed to meet the capacity of new equipment or replace the removed gas in national production and residential life every year [27]. This method is mainly suitable for situations of rapid social and economic development, frequent replacement of emission equipment, and complex natural emission sources. Its advantage is that it can calculate the carbon emissions of facilities and equipment in different sectors at the micro level, thus improving the comprehensiveness of the analysis results. Its disadvantage is that there are many intermediate emission processes that need to be taken into consideration, which make it easy to increase system errors and difficult to obtain detailed basic data. Ryoo [28] and Fiehn [29] used the mass balance method based on flight observation data to investigate the carbon emission characteristics of Sacramento, California, and the Silesia coal basin, respectively. The errors mainly came from the uncertainty of atmospheric background mole fraction and the change in planetary boundary layer height during morning flight, and the biosphere flux also increased the difficulty of quantitative analysis. Based on the traditional mass balance method, Pitt [30] assumed that the scale emission is not limited to a clearly defined area, and comprehensively considered the carbon emission effect around the research area, thus proposing a new modeling method. Then, the carbon emission characteristics of London were investigated, and compared with the emission inventory calculated based on traditional methods. The new method did not need to separate the city from the surrounding emission sources, so it had wider applicability.

(3)

Based on the measured basic data of emission sources, the actual measurement method summarizes the relevant carbon emissions, which specifically includes an on-site measurement method and an off-site measurement method [31]. The method is mainly suitable for emission sources in small areas and is capable of obtaining first-hand monitoring data. Its advantage is that there are few intermediate links, so the results are accurate. Its disadvantage is that data acquisition is relatively difficult, the investment is large, and the accuracy is influenced by sample representativeness. Chen [32] investigated the annual carbon emission characteristics of 247 sewage treatment plants in 7 regions of China based on the actual measurement method and determined the emission reduction potential of regional sewage treatment. Weltman [33] investigated the carbon emissions of household solid fuel consumption in Haryana, India based on the actual measurement method. The results showed that there was a great difference between the field carbon emissions in daily cooking activities and the measurement values obtained in the laboratory.

2.2. Evaluation of Regional Low-Carbon Efficiency

Low-carbon efficiency is an important index to investigate the level of regional sustainable development. The comprehensive indicator evaluation method, as well as the data envelopment analysis (DEA) method and its extended model (SBM model, super efficiency model, etc.), are the main methods used to evaluate efficiency.

(1)

The comprehensive indicator evaluation method mainly constructs a comprehensive index system from the aspects of economy, technology, environment, and policy. Furthermore, combined with objective weighting methods such as the entropy weight model and subjective weighting methods such as the Delphi model, the regional low-carbon efficiency is quantitatively evaluated [34]. Liu [35] established a low-carbon efficiency evaluation system for regional tourism development based on three dimensions: economic support level, low-carbon development level, and policy support level, and further used the Delphi and Analytic Hierarchy Process (AHP) methods to carry out an empirical evaluation of the Daxinganling area. However, because this study mainly determined the index weight based on expert experience, it inevitably had subjective defects. Ye [36] established a low-carbon efficiency evaluation system of regional power system based on three dimensions: power generation side, power grid side, and load side, and determined the weights through a multi-scenario dispatching simulation and an index sensitivity analysis. On this basis, an empirical evaluation for the power system in a certain area of Zhejiang Province, China was carried out.

(2)

The DEA model mainly evaluates efficiency by constructing an input–output index system. Specifically, the input mainly includes energy, population, assets, and other elements, and the output mainly includes desirable outputs such as GDP and undesirable outputs such as pollutants. DEA model analysis has the characteristics of dimensionless data processing, which can reduce model deviation and avoid subjective interference. Therefore, it has been widely used in the research field of efficiency evaluation. Gouveia [37] used the DEA model to evaluate the low-carbon economic efficiency of 23 beneficiary EU countries. However, the input and output indicators considered in this study only include four elements: EU co-financing, eligible total expenditure, decided eligible expenses, and greenhouse gas emission reduction, so the comprehensiveness of the evaluation results needs further improvement. However, this study mainly used emission reduction to evaluate the adverse output of economic development; that is, it was treated as a positive indicator in an actual evaluation. After data transformation, the model could only be solved under the condition of variable scale return, so the solution analysis had limitations. Keivani [38] further introduced the undesirable output into the DEA model and carried out an empirical evaluation for the low-carbon efficiency of regional petrochemical plants from 2011 to 2017. However, the traditional DEA model often assumes that the input, desirable output, and undesirable output need to be adjusted in equal proportion, which is contrary to the actual situation. Based on slack-based measurement (SBM), Zha [39] used the SBM-Undesirable model to evaluate the development efficiency of urban low-carbon tourism economy in Hubei Province from 2007 to 2013. However, in this study, the efficiency values of decision-making units (DMUs) on the frontier of effective production were all equal to 1, so it was impossible to carry out a differentiated ranking for these effective DMUs. Tao [40] and Zhang [41] further introduced the super-efficiency setting into DEA model; that is, the evaluation values of effective DMUs were allowed to be greater than one. On this basis, the green and low-carbon efficiencies of 30 selected provinces in China were quantitatively analyzed, respectively. However, the input factors considered in the above research are limited to labor, capital investment, and energy, and the comprehensiveness of the indicators need to be further improved.

So far, the research on the calculation of regional carbon emissions and evaluation of regional low-carbon efficiency has made some progress in theory and practice, but the existing research has limitations in the following aspects: (1) The emission factor method is widely used in regional carbon emission calculations due to its advantages of intuitive understanding, mature theory, and sufficient basic activities. However, existing research mainly focuses on energy activity, and there is a lack of comprehensive inventory analysis for economic and social units such as industrial production and waste disposal [42,43,44,45]. Meanwhile, the changes in the carbon pools caused by forestry activity and the indirect carbon emissions driven by regional energy allocation also need to be further analyzed. In this study, a regional carbon emission calculation system consisting of energy activity, industrial production, forestry activity, and waste disposal is established. Meanwhile, the “up-bottom” research framework is adopted, and official statistics are used as the basic activity data for calculation. This is helpful to eliminate the estimation difference caused by multivariate statistical sources and to avoid the uncertainty of analysis results caused by the complexity of actual measurement under the “bottom-up” framework. (2) The DEA method mainly determines the optimal objective weight based on the actual data of DMUs and does not need to set complex functional relations between variables, so it has certain advantages compared with comprehensive index evaluation method [46]. This study expands the evaluation system of regional low-carbon efficiency. Specifically, for the input factors, energy, labor, technology, and capital investment are comprehensively considered. For the output factors, it includes the desirable output: GDP, and the undesirable output: carbon emissions, which is composed of emission sources and absorption sinks. In addition, a slack-based measure and super-efficiency setting are introduced into the traditional DEA model to establish a Super-SBM-Undesirable model. In this regard, it is helpful to solve the problem of non-zero relaxation of input and output, helping to rank the efficiency level of the effective DMUs in sequence [47].

3. Model and Data

3.1. Calculation Method of Regional Carbon Emissions

With the low-carbon development becoming a universal consensus in the world, regional carbon emission calculation has gradually attracted the attention of scholars in related research fields [48]. At the level of national assessment, the IPCC issued the “2006 IPCC Guidelines for National Greenhouse Gas Inventories” and revised the guidelines in 2019. It divides carbon emission sources into four major units: energy, industrial process and product use, agriculture, forestry and other land use, and waste disposal [49]. The IPCC guidelines provide the latest methods and rules for establishing the national carbon emission inventory, and their methodology system has a profound and significant impact on all countries in the world. The International Organization for Standardization formulated “ISO14064-1 Specification and Guide for Quantification and Reporting of Carbon Emissions and Removal” [50]. It divides the emission sources into direct emission and removal of carbon sources, indirect emission of transportation, indirect emission of products used by organizations, etc. In each category, the guide distinguishes biomass emissions from non-biomass emissions. On the basis of international guidelines, the Office of the National Coordination Group for Climate Change Countermeasures of China published the “Study on Greenhouse Gas Inventory in China”. This study identifies the key emission sources, analyzes the availability of activity data and emission factors, and demonstrates the applicability of the IPCC inventory accounting method to China. In addition, at the level of regional assessment, the Energy Research Institute of the National Development and Reform Commission of China issued “Guidelines for the Compilation of Provincial Greenhouse Gas Inventory (Trial)”. The guidelines provide the division method of provincial emission sources and the reference carbon emission accounting parameters [51]. Based on the above-mentioned reference guidelines, this study establishes a regional carbon emission calculation system consisting of energy activity, industrial production, forestry activity and waste disposal, which mainly focuses on the macro level of the region. Meanwhile, the basic activity data come from official statistics at the national level, provincial level, departmental level, and urban level, so as to ensure the availability and authenticity of multiple pieces of data. In this regard, the emission factor method is used as the specific calculation method.

3.1.1. Energy Activity

(1)

Fuel Combustion

The carbon emissions from fuel combustion mainly come from the intentional oxidation of fuel inside or outside the equipment when providing heat or mechanical work for a certain process [52]. The energy-using units mainly include energy conversion processes such as thermal power generation and heating, as well as energy terminal consumption processes in sectors such as agriculture, industry, construction, transportation, commerce, other service industries, and residential life. The specific calculation model is shown in Formula (1).

$$C_{fc}={\displaystyle \sum _m{\displaystyle \sum _n{E}_{m,n}\times V_{fc,n}\times F_{fc,n}\times K_{fc,n}\times \frac{44}{12}}}$$

(1)

In this formula, C_fc is the carbon emissions during the process of fuel combustion; E_m,n is the consumption level of the n-th fossil fuel by the m-th energy-using unit; V_fc,n is the combustion calorific value of the n-th fossil fuel; F_fc,n is the content of carbon atom per unit combustion calorific value of the n-th fossil fuel; K_fc,n is the carbon oxidation rate of the n-th fossil fuel during combustion; 44/12 is the mass conversion coefficient between the carbon atom and the CO₂ molecule.

(2)

Inter-regional power allocation

The calculation process of fuel combustion unit covers the carbon emission caused by thermal power generation. In the process of dividing the carbon emission responsibility, the actual energy user is usually regarded as the carbon emission responsibility body. Therefore, it is necessary to calculate the indirect carbon emissions in the process of inter-regional power allocation based on the difference between externally imported electricity and locally exported electricity [53]. The specific calculation model is shown in Formula (2).

$$C_{ipa}=\left(Q_{ipa,out}-Q_{ipa,in}\right)\times F_{ipa}$$

(2)

In this formula, C_ipa is the carbon emissions during the process of inter-regional power allocation; Q_ipa,out is the externally imported electricity; Q_ipa,in is the locally exported electricity; F_ipa is the carbon emission factor of inter-regional power allocation, which is expressed by the average carbon emission factor of regional power grid supply.

3.1.2. Industrial Production

After removing the carbon emissions from industrial energy terminal consumption, raw materials, catalysts, additives, or other substances will also produce certain carbon emissions in the process of participating in chemical reactions or physical changes in industrial production [54].

(1)

Cement production

The carbon emissions of cement production mainly come from the chemical reaction of spherical intermediate products such as clinker; that is, the CO₂ emitted when calcium carbonate is heated or calcined into lime [55]. The specific calculation model is shown in Formula (3).

$$C_{cp}=E_{cp}\times R_c\times F_{cp}$$

(3)

In this formula, C_cp is the carbon emissions during the process of cement production; E_cp is the production volume of cement products; R_c is the proportion of clinker used in cement production; F_cp is the carbon emission factor of cement production.

(2)

Steel production

There are two main sources of carbon emissions in steel production. One is the production process of pig iron that has not been converted into steel, and the other is the steel-processing processes such as basic oxygen furnace (BOF), electric arc furnace (EAF), and open-hearth furnace (OHF) [56].

The specific calculation model of carbon emissions from pig iron production is shown in Formula (4).

$$C_{pip}=E_{pip}\times F_{pip}$$

(4)

In this formula, C_pip is the carbon emissions during the production process of pig iron; E_pip is the production volume of pig iron that has not been converted into steel; F_pip is the carbon emission factor of pig iron production.

The specific calculation model of carbon emissions from steel processing is shown in Formula (5).

$$C_{sp}=E_{BOF}\times F_{BOF}+E_{EAF}\times F_{EAF}+E_{OHF}\times F_{OHF}$$

(5)

In this formula, C_sp is the carbon emissions during the processes of steel processing; E_BOF, E_EAF, and E_OHF are the production volumes of BOF, EAF, and OHF, respectively; F_BOF, F_EAF, and F_OHF are the carbon emission factors of BOF, EAF, and OHF, respectively.

(3)

Glass production

The carbon emissions of glass production mainly come from the melting process of raw materials such as limestone. Meanwhile, considering that recycled waste glass and broken glass will also be used as raw materials in production, the proportion of this part should be removed during the calculation [57]. The specific calculation model is shown in Formula (6).

$$C_{gp}=E_{gp}\times F_{gp}\times (1-R_{rwb})$$

(6)

In this formula, C_gp is the carbon emissions during the process of glass production; E_gp is the production volume of glass products; F_gp is the carbon emission factor of glass production; R_rwb is the proportion of recycled waste glass and broken glass used in glass production.

(4)

Calcium carbide production

The carbon emissions of calcium carbide production mainly come from the chemical reaction of carbon-containing raw materials [58]. The specific calculation model is shown in Formula (7).

$$C_{ccp}=E_{ccp}\times F_{ccp}$$

(7)

In this formula, C_ccp is the carbon emissions during the process of calcium carbide production; E_ccp is the production volume of calcium carbide; F_ccp is the carbon emission factor of calcium carbide production.

3.1.3. Forestry Activity

Plant biomass constitutes an important carbon pool in the earth’s ecosystem. The carbon emissions of forestry activity mainly come from the change in the biological carbon pool [59].

(1)

Biomass growth

The annual increase in the carbon pool caused by biomass growth mainly comes from the carbon fixation effect of forest plants [60]. The specific calculation model is shown in Formula (8).

$$C_{bg}=A_{fl}\times Q_{abp}\times R_{dmc}$$

(8)

In this formula, C_bg is the annual increase in the carbon pool caused by biomass growth; A_fl is the regional forestry land area; Q_abp is the average biological growth per unit of forestry land area in the region; R_dmc is the carbon content ratio of dry matter.

(2)

Wood harvesting

Wood harvesting will cause significant carbon loss and further lead to a reduction in the annual carbon pool [61]. The specific calculation model is shown in Formula (9).

$$C_{wh}=Q_{wh}\times B_{wh}\times \left(1+P_{aur}\right)\times R_{dmc}$$

(9)

In this formula, C_wh is the annual reduction in the carbon pool caused by wood harvesting; Q_wh is the annual wood harvesting volume; B_wh is the coefficient of converting wood removal into biological removal; P_aur is the ratio of underground biomass to aboveground biomass.

(3)

Natural disturbance

The annual reduction in the carbon pool caused by natural disturbance mainly focuses on the reduction in the carbon pool caused by forest fires and insect disasters [62]. The specific calculation model is shown in Formula (10).

$$C_{nd}=A_{nd}\times B_{agb}\times \left(1+P_{aur}\right)\times R_{dmc}\times F_{dbl}$$

(10)

In this formula, C_nd is the annual reduction in the carbon pool caused by natural disturbance; A_nd is the influence area of natural disturbance; B_agb is the average aboveground biomass of the area affected by disturbance; F_dbl is the proportion of biomass loss in the process of fire or insect disaster disturbance.

3.1.4. Waste Disposal

Residents will produce a lot of solid waste while consuming energy resources to meet their daily needs. The carbon emission effect in the process of waste disposal poses a serious threat to the quality of the ecological environment. The carbon emission of waste disposal mainly comes from the oxidation reaction in the process of solid waste incineration [63]. The specific calculation model is shown in Formula (11).

$$C_{wd}=E_{wd}\times P_{tct}\times P_{mc}\times C_{ce}\times {\displaystyle \frac{44}{12}}$$

(11)

In this formula, C_wd is the carbon emissions during the process of waste disposal; E_wd is the quantity of solid waste incineration; P_tct is the proportion of total carbon content in solid waste; P_mc is the proportion of mineral carbon in the total carbon content; C_ce is the combustion efficiency of solid waste incinerator.

The basic activity data of the regional carbon emission calculation come from China Statistical Yearbook, China Energy Statistical Yearbook, China Forestry and Grassland Statistical Yearbook, China Environmental Statistical Yearbook, World Steel Statistical Data, National Forest Resource Inventory Data, Sichuan Statistical Yearbook, Sichuan Ecological Environment Bulletin, etc. The data on emission factors come from the “2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories” issued by IPCC, the “Guidelines for the Compilation of Provincial Greenhouse Gas Inventory (Trial)” issued by the National Development and Reform Commission of China, and the baseline emission factors of regional power grids issued by the Ministry of Ecology and Environment of the People’s Republic of China.

3.2. Evaluation Method of Regional Low-Carbon Efficiency

3.2.1. Super-SBM-Undesirable Model

The DEA model has been developed for more than 40 years since it was put forward by the American logistics scientist Charnes in 1978 [64]. The DEA model is derived from the concept of relative efficiency evaluation, and it is a nonlinear parametric programming method to evaluate the relative efficiency of a series of DMUs with the same multi-input and multi-output [65]. The efficiency value of each DMU is defined as the ratio of the output weighted sum to the input weighted sum, and the model system structure is shown in Figure 1. The DEA model regards the interior of the system as a black box. Therefore, it does not need to consider the intermediate links and intermediate data from input to output, and only needs to use the initial input data and the final output data to evaluate the relative effectiveness of DMUs. In addition, when carrying out efficiency evaluation based on the DEA model, there is no need to define the real production operation function of DMUs or estimate the parameters in advance, so it has certain advantages over other efficiency evaluation methods. In this regard, the DEA model has become an important tool and a means for management evaluation and decision analysis.

In 1978, Charnes established the first CCR model based on constant return to scale (CRS). Subsequently, in 1984, Banker put forward the BCC model based on variable returns to scale (VRS), which set off an upsurge in research of the DEA model [66]. When efficiency evaluation is carried out based on the traditional DEA model, it often happens that multiple DMUs are all in a 100% effective state, especially when the number of input and output indicators is large, the number of effective DMUs is large and the efficiency values are all one. In this case, it is difficult to further distinguish and compare the differences in efficiency levels. In order to solve this problem, Andersen put forward the super efficiency DEA model in 1993 [67]. The core idea of the super efficiency model is to replace the reference set of the DMU to be evaluated with the production frontier composed of other DMUs, and the DMU to be evaluated is not included in the reference set. At this time, the efficiency values calculated for effective DMUs are generally greater than one, while the production frontier and efficiency values of invalid DMUs will not change. This is helpful to rank the efficiency values of all DMUs in sequence [68].

The traditional DEA model is based on radial and angular measurement methods. The radial spatial distribution of the DEA model will cause the relaxation or crowding of input factors. When the input or output factors show non-zero relaxation characteristics, the radial DEA model will overestimate the evaluation efficiency. Meanwhile, the calculation based on input perspective and output perspective will obtain different results, so it is not accurate. In this regard, Tone proposed an SBM-Undesirable model [69]. It is different from the traditional model in that the slack-based variables are included in the objective function, which is helpful to solve the problem of non-zero relaxation of input and output. In addition, it comprehensively considers the undesirable output problems that may occur in the production process.

It is assumed that a specific production system contains n DMUs, which is expressed as DMU_j (j = 1, 2, …, n) [70]. Each DMU contains m units of inputs x_i(i = 1, 2, …, m), s₁ units of desirable outputs y_r^g(r = 1, 2, …, s₁), and s₂ units of undesirable outputs y_r^b(r = 1, 2, …, s₂) [71]. On this basis, the objective function is set to be non-oriented, and the Super-SBM-Undesirable model based on CRS is shown in Formula (12) [72].

$$\begin{array}{l}{\rho }^{\ast }=\min {\displaystyle \frac{\frac{1}{m}{\displaystyle \sum _{i=1}^m\frac{{\overline{x}}_i}{x_{i0}}}}{\frac{1}{s_1+s_2}\left({\displaystyle \sum _{r=1}^{s_1}\frac{\overline{y^g}}{y_{r0}^g}+{\displaystyle \sum _{r=1}^{s_2}\frac{\overline{y^b}}{y_{r0}^b}}}\right)}}\\ s.t.\left\{\begin{array}{c}\overline{x}\ge {\displaystyle \sum _{j=1,\ne 0}^n{\theta }_j{x}_j}\\ \overline{y^g}\le {\displaystyle \sum _{j=1,\ne 0}^n{\theta }_j{y}_j^g}\\ \overline{y^b}\ge {\displaystyle \sum _{j=1,\ne 0}^n{\theta }_j{y}_j^b}\\ \overline{x}\ge x_0,\overline{y^g}\le y_0^g,\overline{y^b}\ge y_0^b,\theta \ge 0,{\displaystyle \sum _{j=1,\ne k}^n{\theta }_j}=1,\overline{y^g}\ge 0\end{array}\right.\end{array}$$

(12)

In this formula, ρ* is the efficiency value of DMUs, which can be greater than one and is a dimensionless quantity; θ is the weight variable of DMUs. The subscript “0” in the model refers to the evaluated DMU. Only when ρ* ≥ 1, DMU is effective; when ρ* < 1, it means that the evaluated DMU has efficiency loss compared with the DMUs on the production frontier, and it needs to adjust the input or output to enter the effective state.

3.2.2. Indicator Selection and Data Source

In this study, energy utilization efficiency, low-carbon scientific and technological innovation, and land greening level are comprehensively used as the evaluation basis of low-carbon efficiency [73]. The production frontier of the DEA model is constructed with the goal of minimizing net carbon emissions and maximizing economic development level while investing the same amount of production factors [74]. The higher the low-carbon efficiency, the more conducive it is to achieve carbon peaking as soon as possible and even further achieve carbon neutrality. In this regard, this study uses the panel data of Sichuan Province and its 21 cities (states) from 2015 to 2022 to evaluate the low-carbon efficiency from the provincial and urban levels, respectively. The evaluation system framework is shown in Figure 2.

(1)

Energy input. The total annual energy consumption of Sichuan Province and its 21 cities (states) is used to represent the energy input level [75]. In order to compare the energy consumption of different years more intuitively and eliminate the impact of differences in energy units, the consumption of primary energy such as coal, oil, and natural gas in different years is uniformly converted into standard coal for evaluation during data analysis.

(2)

Labor input. Due to the unavailability of data on indicators such as education level and labor efficiency of the labor force, the measurement of labor input is based on the annual number of employed personnel in Sichuan Province and its 21 cities (states) [76].

(3)

Technology input. The development speed of advanced technology depends on the investment level of scientific research funds [77]. Therefore, the technology input is expressed as the proportion of the annual scientific and technological expenditure of Sichuan Province and its 21 cities (states) to the regional public budget expenditure.

(4)

Capital input. Capital stock can effectively represent capital investment. However, because the relevant data of capital stock cannot be directly found, it is necessary to estimate the capital stock in statistical analysis. At present, academic circles usually use the perpetual inventory method (PIM) pioneered by Goldsmith [78] to measure the capital stock, and its basic formula is shown in Formula (13).

$$K_t=K_{t-1}\left(1-\delta \right)+I_t$$

(13)

In this formula, K_t and K_t−1 are the capital stock in year t and year t − 1, respectively; δ is the depreciation rate of fixed assets; I_t is the actual investment in fixed assets in year t.

In this study, the total amount of regional fixed capital formation is selected as the index to measure the annual capital stock, which is calculated based on the PIM. In order to eliminate the influence of price fluctuation in different years, the capital stock is uniformly converted to the price level in 2015 based on the fixed asset investment price index.

(5)

Desirable output. The regional GDP of Sichuan Province and its 21 cities (states) from 2015 to 2022 is selected as the desirable output variable of low-carbon efficiency calculation [79]. In order to avoid the impact of price changes, based on the GDP deflator, the original data are uniformly converted into a comparable GDP based on the price level in 2015.

(6)

Undesirable output. The undesirable output index is selected as the regional net carbon emission, and its level evaluation is based on the calculation method established in Section 3.1. Energy activity and industrial production are selected as the main carbon emission sources [80]. Meanwhile, in order to consider the driving effect of the forest carbon sink on the improvement in low-carbon efficiency, the carbon absorption in the process of forest growth is also included in the empirical evaluation.

The basic activity data of the above indicators come from the China Statistical Yearbook, China Energy Statistical Yearbook, China City Yearbook, Sichuan Statistical Yearbook, and the statistical data of 21 cities (states) in Sichuan Province.

4. Empirical Analysis

4.1. Regional Carbon Emission Characteristics

According to the calculation method of regional carbon emissions established in Section 3.1, the carbon emission inventory of Sichuan Province from 2015 to 2022 is shown in

Table 1. Carbon emission inventory of Sichuan Province from 2015 to 2022 (Unit: 10,000 tons).

Emission Unit		2015	2016	2017	2018	2019	2020	2021	2022
Energy activity	Thermal power generation	3750	2709	2697	2630	3457	3656	4572	5946
	Supply heat	434	414	559	591	646	923	725	380
	Agricultural industry	575	526	502	413	432	483	486	527
	Industry	19,710	16,991	16,875	14,405	14,461	13,508	13,492	11,996
	Construction industry	222	346	366	427	441	420	498	563
	Communications and transportation industry	2004	2869	3021	2970	3111	2978	3089	3080
	Commercial industry	911	938	1046	874	864	766	676	710
	Other service industries	748	678	721	800	714	620	688	730
	Residential life	2541	2563	2526	2639	2758	2832	2914	3071
	External power import	434	408	560	897	996	1056	1479	1365
	Local power export	−10,877	−11,303	−12,272	−12,072	−11,877	−12,237	−12,162	−13,379
Industrial production	Cement production	5665	5885	5572	5548	5718	5849	5708.35	5274
	Steel production	5102	4950	5199	5792	5359	5420	5351	5280
	Glass production	20	27	28	27	46	47	48	49
	Calcium carbide production	79	83	71	113	125	57	59	54
Forestry activity	Biomass growth	−9185	−9185	−9185	−9683	−7600	−7676	−7714	−7720
	Wood harvesting	184	218	242	249	264	241	327	312
	Forest fires	3	2	11	17	7	16	3	3
	Insect disasters	13	13	20	16	3	4	216	167
Waste disposal	Solid waste incineration	8	10	13	15	19	20	29	30
Net carbon emission		22,341	19,142	18,572	16,668	19,944	18,983	20,484	18,438

.

During the research period, the net carbon emissions in Sichuan Province showed a fluctuating downward trend on the whole, decreasing from 223.41 million tons in 2015 to 184.38 million tons in 2022, which is consistent with the findings of Sun et al. [81]. In 2019, the net carbon emission of Sichuan Province reached the first maximum value, which could be explained as the accelerated implementation of regional economic and social development goals at the end of the 13th Five-Year Plan. In 2020, under the influence of the COVID-19 epidemic, regional carbon emissions showed a significant downward trend, which is similar to the findings of Sikarwar et al. [82] based on the global perspective. In 2021, with the precise implementation of policies on the impact of the COVID-19 epidemic, the recovery of social operation led to a rebound in regional carbon emissions, which agrees with the findings of Ray et al. [83]. In 2022, with the COVID-19 Omicron mutant gradually becoming the main epidemic strain in the global epidemic, its strong immune escape characteristics induced significant pathogenicity and promoted a new round of epidemic outbreak. It led to a social shutdown and had a negative inhibitory effect on regional carbon emissions, which is different from the results of Liu et al. [84] on global emission monitoring. This was mainly due to the fact that China’s epidemic blockade control policy lasted for a longer time. In 2022, the net carbon emission of Sichuan Province was 184.38 million tons, of which the direct carbon emission was 395.37 million tons, the indirect emission reduction in inter-regional power allocation was 133.79 million tons, and the absorption of forestry carbon sink was 77.2 million tons.

The direct carbon emissions from energy activity and industrial production were 283.68 million tons and 106.57 million tons, respectively, accounting for 71.75% and 26.95% of the total direct carbon emissions of 395.37 million tons in 2022. In this regard, they were the main contributors to regional carbon emissions, which is consistent with the findings of Wang et al. [85]. The main reasons for the above phenomenon were the following: on the one hand, energy activity ran through all aspects of national economic and social development. Under the influence of scale effect, fuel combustion had made a significant contribution to regional greenhouse effect. On the other hand, industries such as equipment manufacturing, building materials, energy, and the chemical industry, with high energy consumption and carbon emissions, had always been the pillar industries and modern economic base of Sichuan Province. In addition to this, the direct emissions of carbon sources in forestry activity and waste disposal were relatively small, only 4.82 million tons and 300,000 tons, respectively, which were almost negligible in the total.

(1)

Energy activity

For energy activity, the carbon emissions of energy processing and transformation and energy terminal consumption were 63.26 million tons and 206.77 million tons in 2022, respectively. Because energy consumption was closely related to the production activities of the three major industries and the diverse fields of residential life, its contribution to the greenhouse effect was particularly remarkable under the drive of high utilization level, which agrees with the findings of Afroz et al. [86]. Specifically, 93.99% of carbon emissions from energy processing and transformation were concentrated in thermal power generation. Meanwhile, the proportion of energy lost to the environment in conventional power plants is high [87]. Therefore, the carbon emission effect caused by energy consumption during the transformation process cannot be ignored. As for the carbon emissions of energy terminal consumption, the sectoral composition is shown in Figure 3. From 2015 to 2022, the average proportion of carbon emissions of energy terminal consumption from the primary industry represented by agriculture, forestry, animal husbandry, and fishery accounted for 2.13% of the total emission. The main reason for the above phenomenon was that the daily business activities of the primary industry were less dependent on energy utilization, which is consistent with the findings of Wang et al. [88]. During the research period, the average proportion of carbon emissions of energy terminal consumption from the industrial sector in the secondary industry accounted for 64.88% and became the primary emission source of terminal consumption. The main reason for the above phenomenon was that the carbon emissions in the combustion process mainly depended on the carbon content of fuel. Industrial production depended on high-carbon fossil energy such as coal and heavy oil, and at the same time, it often had high-level energy demand. The larger consumption scale and carbon oxidation ratio would inevitably lead to the simultaneous increase in carbon emissions. During the research period, the transportation industry and residential life in the tertiary industry also contributed a certain amount of CO₂, accounting for an average of 12.59% and 11.88%, respectively. The main reasons for the above phenomenon were the continuous progress of urbanization in Sichuan Province and the lifestyle changes caused by the improvement in residents’ wealth, which is different from the results of Xiao et al. [89] on the Yangtze River Delta Integration Demonstration Area. This was mainly because the living standard and average education level of residents in the western region were lower than those in the eastern coastal areas, and the popularization of the low-carbon life concept needed to be strengthened. Specifically, transportation facilities such as cars, trucks and even high-speed rail and airplanes were still not decoupled from the consumption of fossil fuels such as gasoline and diesel. Meanwhile, the natural gas consumption level of residential life in fields such as heating and cooking was second only to the industrial sector.

During the research period, the indirect carbon emission caused by external power import increased from 4.34 million tons in 2015 to 13.65 million tons in 2022, and the indirect carbon emission reduction caused by internal power export increased from 108.77 million tons in 2015 to 133.79 million tons in 2022. From a system perspective, the power transaction process in Sichuan Province effectively and indirectly reduced the regional emission intensity, which is different from the results of Luo et al. [90] on non-energy producing areas. This was mainly due to the fact that hydropower generation in Sichuan Province ranked first in China. In 2022, the total regional power generation reached 484.62 billion kWh, of which 155.81 billion kWh was transported to Central China and East China through the “West-to-East Power Transmission Project” to support regional energy demand. It was helpful to promote the whole life-cycle emission reduction in energy utilization in China through clean power production and form a new pattern of sustainable development while ensuring the safety of energy supply. From 2015 to 2022, Sichuan Province delivered a total of 1120.06 billion kWh of electricity through the “point-to-network” mode, replacing and reducing carbon emissions by about 961.79 million tons, making a positive contribution to the realization of carbon peaking and carbon neutrality. However, at the same time, the continuous increase in regional electricity imports in recent years highlighted the problem of insufficient local energy consumption. On the one hand, it was caused by seasonal energy overproduction and energy shortage due to intermittent output of water energy and insufficient power storage capacity. On the other hand, the connecting line between Sichuan and other provinces was mainly a one-way outgoing transmission line, and the lack of energy channels into Sichuan led to the difficulty of cross-regional power mutual assistance and emergency power supply coordination.

(2)

Industrial production

During the research period, the carbon emissions of industrial production in Sichuan Province reached the maximum in 2018, and then showed a downward trend. The main reason for the above phenomenon was that the supply-side capacity restriction policy in Sichuan’s industrial sector had been tightening. In this regard, it was conducive to improving the overall concentration level of the industry, which is similar to the findings of Song et al. [91]. The sectoral composition of carbon emissions of industrial production is shown in Figure 4. In 2022, the carbon emissions in the production process of cement, steel, glass and calcium carbide were 52.74 million tons, 52.80 million tons, 490,000 tons, and 540,000 tons, respectively, which indicated that the carbon emissions from industrial production showed significant aggregation in the two major sectors of cement and steel production. Specifically, the carbon emissions from cement production were the largest. This was mainly because Sichuan Province was a major cement-producing province. Referring to the Statistical Yearbook of Sichuan Province, regional cement output reached 130.70 million tons in 2022, ranking first in the western region. In the process of cement production, when limestone is heated and calcined, it will lead to significant direct emission of CO₂ [92]. Although the carbon emissions of steel production also exceeded 50 million tons, the emission level was relatively low compared with other provinces with the same production scale. The main reason for the above phenomenon was that Sichuan Province was a concentrated area of electric furnace steel industry in China. According to the official data released by the Sichuan Provincial Economic and Information Department, the proportion of regional short-process steelmaking reached about 40% in 2022, which was much higher than the national average. Its production principle of “steelmaking with steel” could eliminate the sintering and reduction process of iron ore, thus effectively controlling energy consumption and carbon emissions, which is consistent with the findings of Hasanbeigi et al. [93]. From 2015 to 2022, the carbon emission level of glass and calcium carbide production was generally low. On the one hand, this was because glass manufacturers partially used recycled waste glass or broken glass as raw materials, and its proportion in mass production was as high as 40~60% [94]. In this regard, it could indirectly reduce about half of the carbon emissions in the smelting process through resource saving. On the other hand, the production of calcium carbide in Sichuan Province was low, and in 2022 it was only 468,000 tons. Meanwhile, calcium oxide is the main raw material for calcium carbide production in China. During the production process, most of the carbon will be contained in calcium carbide products by the electrothermal method, and only the remaining part will react with excess O₂ and be further converted into CO₂. Due to the characteristics of the economic structure, the carbon emission effect of industrial production in Sichuan Province always existed, and it had not been significantly curbed during the research period.

(3)

Forestry activity

During the research period, the overall change trend of the carbon emissions of forestry activity in Sichuan Province was relatively stable. The change in forest area in Sichuan province was not significant, so the variation of the forest carbon sink caused by continuous growth of forest plants was small. Forestry activity had significant ecological barrier functions, mainly due to the fact that forest vegetation was the mainstay of natural ecosystems and played a leading role in absorbing and converting CO₂. Biomass related to woody plants constituted an important carbon pool of the regional ecosystem, and the forest carbon sink in Sichuan Province had a significant effect on alleviating the regional greenhouse effect, which is similar to the findings of Koondhar et al. [95] based on the national perspective. In 2022, the regional forest carbon sink was 77.20 million tons, which could hedge the direct carbon source emissions of 395.37 million tons by 19.53%. This was mainly due to the active promotion of large-scale “greening the whole Sichuan” actions during the 13th Five Year Plan period. According to China Environmental Statistics Yearbook, the provincial forest resource coverage rate reached 40.26% in 2022, far higher than the national average of 24.02%. In the consumption of forest resources, the carbon emission caused by biomass loss was relatively small. In 2022, the annual carbon pool reductions caused by wood harvesting, forest fires, and insect disasters were 3.12 million tons, 30,000 tons, and 1.67 million tons, respectively. The impact of the above discrete events mainly depended on the actual situation in specific areas in different years. By transferring carbon from the carbon sink of living biomass to the pool of dead organic matter, CO₂ emission was eventually formed and the carbon composition of the ecosystem was redistributed. Specifically, in 2022, the timber output in Sichuan Province was 2.89 million cubic meters. Due to the decentralized management mode and high logging cost, the vacancy rate of regional logging quota remained high, so the annual carbon pool reduction caused by it was not significant. In addition, under the warning that the area affected by forest fires in 2020 was as high as 1487.20 hm², Sichuan Province had made great efforts to carry out normalized forest and grassland fire prevention and control work, scientifically and orderly implemented combustible removal and planned burning, and actively promoted supervision and inspection, fire prevention drills, and information construction. Benefitting from the above measures, the forest area damaged by fire in 2022 was controlled to 261.90 hm², and the damage rate was only 0.01%. As a whole, the reduction in the carbon pool caused by forest fires was almost negligible. However, at the same time, 2022 was a year of high incidence of pests and diseases in Sichuan Province, and quarantine pests such as pine wood nematodes and red imported fire ants showed a continuous diffusion trend. The disaster area of forestry pests was as high as 49,840 hm², which is 40.85 times that of the disaster area of 1220 hm² in 2020. The main causes were the rapid spread of alien invasive species and the inherent defects of local ecological fragile areas, which agrees with the findings of Li et al. [96].

(4)

Waste disposal

For waste disposal, carbon emissions were mainly concentrated in the incineration process of solid waste. From 2015 to 2022, the carbon emissions of waste disposal in Sichuan Province showed an overall upward trend. On the one hand, it was mainly because with the victory of China in the fight against poverty, the urbanization process in Sichuan Province continued to advance, and the population scale was also expanding year by year. Economic diversification led to a large number of surplus rural labor entering the city for work. Meanwhile, regional per capita disposable income and the proportion of enjoyment consumption continued to increase, which inevitably led to the continuous expansion of waste treatment scale, which is similar to the findings of de Araújo et al. [97] based on the global perspective. Moreover, there were still some problems in the terminal treatment of domestic waste in Sichuan province, such as low-level and single technology, which promoted the carbon emission rate of waste disposal to increase. In this regard, the innovation of carbon emission control technology for regional waste disposal is urgent. The carbon emission from waste incineration increased from 80,000 tons in 2015 to 300,000 tons in 2022, with an average annual increase of 20.78%. Referring to the bulletin of ecological environment in Sichuan Province, the harmless treatment rate of municipal solid waste in Sichuan Province reached 100% in 2022, of which the incineration capacity accounted for 88.40%. Harmless treatment can eliminate the hazards of carcinogens and water pollution, but carbon emission is inevitable in the process of high-temperature incineration and oxidation. In this case, carbon capture and storage technology are important means to control the concentration of greenhouse gases. On the other hand, with the continuous improvement in residents’ living standards, combustible components (such as paper, plastics, etc.) in regional domestic garbage gradually increased, which is consistent with the findings of Kang et al. [98] on 294 cities in China.

4.2. Regional Low-Carbon Efficiency

According to the evaluation method of regional low-carbon efficiency established in Section 3.2, the efficiency values of Sichuan Province and its 21 cities (states) from 2015 to 2022 are shown in

Table 2. Low-carbon efficiency of Sichuan Province and its 21 cities (states) from 2015 to 2022.

Region	2015	2016	2017	2018	2019	2020	2021	2022
Chengdu City	0.9730	0.9621	0.9749	1.0078	1.0885	1.0996	1.0195	1.0956
Zigong City	1.0590	1.0205	1.0096	1.0229	1.0279	1.0291	1.0111	1.0272
Panzhihua City	1.0088	0.9983	0.9996	1.0161	1.0828	1.0942	1.0140	1.0820
Luzhou City	0.6940	0.6570	0.6491	0.6557	0.6630	0.6648	0.6441	0.6552
Deyang City	0.9762	0.9658	0.9827	1.0039	1.0237	1.0353	1.0028	1.0229
Mianyang City	0.8655	0.8985	0.9159	0.9471	0.9556	0.9574	0.9273	0.9461
Guangyuan City	0.6579	0.6926	0.7188	0.7582	0.8037	0.8154	0.8052	0.8132
Suining City	0.6733	0.7015	0.7614	0.8387	0.8411	0.8430	0.8222	0.8291
Neijiang City	0.8787	0.8472	0.8423	0.8566	0.8554	0.8565	0.8361	0.8472
Leshan City	0.8511	0.8302	0.8188	0.8168	0.8162	0.8170	0.8034	0.8105
Nanchong City	0.6958	0.6779	0.6924	0.7414	0.7766	0.7888	0.7697	0.7776
Meishan City	0.7061	0.7009	0.7039	0.7279	0.7474	0.7590	0.7389	0.7512
Yibin City	0.7496	0.7351	0.7346	0.7330	0.7495	0.7509	0.7319	0.7435
Guang’an City	0.7383	0.7036	0.7045	0.7354	0.7404	0.7454	0.7334	0.7422
Dazhou City	0.6121	0.6155	0.6049	0.6050	0.6150	0.6196	0.6056	0.6137
Ya’an City	0.5705	0.5732	0.6122	0.6381	0.6566	0.6668	0.6388	0.6541
Bazhong City	0.7112	0.6970	0.7337	0.7586	0.7871	0.7979	0.7679	0.7859
Ziyang City	0.8254	0.8629	0.9183	0.9650	1.0780	1.0910	1.0408	1.0809
Aba Tibetan and Qiang Autonomous Prefecture	0.3216	0.3245	0.3322	0.3438	0.3537	0.3559	0.3450	0.3521
Ganzi Tibetan Autonomous Prefecture	0.5948	0.5909	0.6230	0.6517	0.6543	0.6575	0.6473	0.6562
Liangshan Yi Autonomous Prefecture	0.6739	0.6515	0.6613	0.6808	0.6813	0.6819	0.6639	0.6726
Sichuan Province	0.7263	0.7264	0.7470	0.7677	0.7900	0.8110	0.7750	0.7950

. The distribution of low-carbon efficiency in typical years (2015, 2018, 2020 and 2022) is shown in Figure 5.

(1)

Analysis based on provincial perspective

From the provincial perspective, except for the short-term fluctuation caused by black swan events such as the COVID-19 epidemic, the low-carbon efficiency of Sichuan Province maintained a continuous upward trend from 2015 to 2022, which is similar to the findings of Wang et al. [99] based on the national perspective. This was mainly due to the strategic goal of promoting high-quality development in Sichuan Province. Since the 13th Five-Year Plan, regional development had adhered to the goal of building a national clean energy demonstration province, coordinated development and ecological security, and continuously promoted the green and low-carbon transformation of economic system, industrial system and energy system. In 2022, the evaluation value of low-carbon efficiency in Sichuan Province reached 0.7950, but it still needed to be further improved compared with the effective production frontier. The main reasons for the above phenomenon were as follows: firstly, the economic and social development of Sichuan Province had not been decoupled from fossil energy consumption. Due to their molecular characteristics, coal and oil had higher carbon emission coefficients than other types of fossil energy. It inevitably led to a large amount of CO₂ emission through oxidation reaction during the combustion process, which agrees with the findings of Ansari et al. [100]. Under the constraints of technical level and infrastructure, energy transformation would take a long time span, and the dependence of regional development on fossil energy could not be fundamentally changed for at least a few years. Therefore, the advantages and benefits of energy structure optimization and upgrading on regional carbon emission reduction had not been effectively highlighted in the short term. Secondly, the contribution of population expansion to carbon emissions could not be ignored, which is different from the results of Bianco et al. [101] on the European Union. This was mainly due to the fact that China had the second largest population in the world, while developed countries such as the UK, France, and the Netherlands had small population bases, so the scale effect of carbon emissions was not significant. Since the 21st century, the slowdown of population growth caused by the family planning policy had effectively controlled the contribution of population factor to carbon emissions in China. However, in recent years, with the increasing aging problem in China, the new birth policy of “two children” and even “three children” had been fully implemented. On the one hand, the expansion of population size would inevitably require more resources to meet the needs of human production, circulation and consumption. Therefore, it would inevitably lead to the increase in direct and indirect energy utilization caused by the consumption of goods and services, and eventually led to the increase in carbon emissions. On the other hand, the expansion of population scale would further increase the impact of human beings on the ecological environment, and more forest and land resources would be put into industrial production and residential life, thus weakening the carbon sequestration of forest ecosystems [102].

Since 2016, the improvement in low-carbon efficiency in Sichuan Province had gradually accelerated, mainly due to significant changes in the domestic and international economic environment. In order to adapt to the new pattern of low-carbon development and accelerate the transformation to a comprehensive, coordinated, and sustainable economic mode, Sichuan Province gradually entered a new stage of industrialization. Meanwhile, the government launched a large-scale and in-depth energy-saving and carbon reduction campaign. It had led to a steady decline in energy consumption per unit of GDP in the province, and remarkable achievements in carbon emission reduction. In 2020, the promotion of regional low-carbon efficiency slowed down under the impact of COVID-19 epidemic. In 2021, the epidemic turned into sporadic occurrence, and the ecological environment protection work had to give way to the recovery of production and living activities, so the regional low-carbon efficiency showed a short-term downward trend. In 2022, with the maturity of epidemic control policies, social development gradually entered the post-epidemic era, and regional low-carbon efficiency resumed its upward trend, which is consistent with the findings of Tang et al. [103]. With the proposal of the “Chengdu-Chongqing Dual City Economic Circle” strategy, it will form an important growth pole of high-quality development in western China. Under the path of sustainable development, Sichuan Province will continue to change its economic development model and improve its growth quality in the future. Therefore, the low-carbon efficiency of Sichuan Province will be further improved during the 14th Five-Year Plan period.

(2)

Analysis based on urban perspective

From a regional perspective, during the period of 2015 to 2022, the low-carbon efficiency of 21 cities (states) in Sichuan Province was significantly different, showing the characteristics of high in the southeast and low in the northwest. It is similar to the distribution law of urban–rural integration development efficiency in the Chengdu–Chongqing area obtained by the research of Jiang et al. [104]. This was mainly because the integration of urban and rural development was a necessary step to achieve the “double carbon” objective and promote the thorough transition to a green and low-carbon economy and society. In 2022, the low-carbon efficiency values of Chengdu, Panzhihua, Deyang, and Ziyang were on the effective production frontier. Among them, Chengdu, Deyang and Zigong had a high level of economic development, a long history of building cities and reasonable industrial institutions. Meanwhile, the urbanization development was moving towards the middle and late stages, the urban infrastructure was perfect, and the residents were highly educated and had strong environmental awareness. Therefore, these cities were the first mover advantage regions of low-carbon transformation in Sichuan Province. Specifically, as the capital of Sichuan Province, Chengdu had convenient transportation and a per capita GDP close to that of a moderately developed country. Its location advantage and decision-making autonomy further expanded its advantages in national support, opening up to the outside world, and attracting foreign investment. Meanwhile, Chengdu was the only mega city in Sichuan Province, which was more conducive to the development of productive service industries with high added value. In 2022, the output value of its tertiary industry was CNY 2962.84 billion, accounting for 52.20% of the regional GDP of CNY 5674.98 billion, ranking first among 21 cities (prefectures). Meanwhile, the expenditure on science and technology in Chengdu was CNY 15,185.37 million, accounting for 6.24% of the total public budget expenditure of CNY 24,350.11 million, ranking among the top in Sichuan Province, which provided effective green technology support for the development of emerging industries. In addition, in 2017, Chengdu became one of the third batch of low-carbon pilot cities in China, innovatively implemented the “Carbon Benefits Tianfu” plan and demonstrated the construction of the urban carbon sink capacity system. The green space rate in the central city of Chengdu continued to increase, and the regional forest coverage rate reached 40.50% in 2022. Under the combined influence of the above factors, Chengdu’s per capita carbon emissions in 2022 ranked the lowest among the top 10 cities in China, including Beijing, Shanghai, Guangzhou, and Shenzhen, which is similar to the findings of Li et al. [105]. Its low-carbon efficiency continued to rise from 0.9730 in 2015 to 1.0956 in 2022, ranking first in the province and showing an initial trend of carbon peaking. Deyang was the second largest industrial city in Sichuan Province. However, since 2015, Deyang had gradually developed the construction of a national comprehensive demonstration city for energy-saving and emission reduction financial policies. Specifically, it carried out a series of explorations around low-carbon urban industries, clean transportation, green buildings, intensive services, reduction in major pollutants, and large-scale utilization of renewable and new energy. Meanwhile, it deeply explored the potential for emission reduction in five major areas: thermal power, cement, flat glass, coal-fired boilers, and the elimination of backward production capacity, and achieved remarkable results. Therefore, its low-carbon efficiency was among the top in the province.

As an important producer of coal and steel resources in China, Panzhihua’s special geographical location, resource endowment conditions, and economic development layout had caused the iron and steel industry and energy industry to occupy a large proportion in regional industrial development. In the early stage, long-term resource overload development led to a sustained increase in regional carbon emissions and a significant deterioration of the ecological environment, which is consistent with the findings of Wang et al. [106]. However, during the 13th Five-Year Plan period, Panzhihua focused on promoting the development of circular economy, and carried out mandatory cleaner production audits for enterprises in four key industries including metallurgy, iron ore mining, steel, and nonferrous metal smelting. By strengthening source control and comprehensive utilization, it effectively promoted the green and low-carbon development of industrial industry. Meanwhile, relying on the advantages of wind energy and hydropower resources, Panzhihua made great efforts to optimize the energy structure and vigorously develop clean energy. It gradually formed a new energy development pattern with hydropower as the pillar, thermal power as the support, wind power, photovoltaic power generation, and garbage power generation as the supplement, and the application of biogas energy, natural gas, and solar photovoltaic achieving comprehensive breakthroughs. In addition, in the field of public publicity, Panzhihua took energy-saving publicity week and low-carbon day as carriers to strengthen the publicity and implementation of laws and regulations on energy conservation and emission reduction. It focused on creating a strong high-quality development atmosphere in the whole society, helping the public to establish a green development concept. Under the combined influence of the above factors, its low-carbon efficiency had consistently ranked second in the province since 2019. During the research period, the economic development level of Ziyang was relatively low and its energy consumption was small, resulting in a relatively insignificant contribution to carbon emissions. Moreover, in recent years, regional governments have taken a series of environmental protection measures. Specifically, at the end of 2017, the “Comprehensive Work Plan for Energy Conservation and Emission Reduction in Ziyang City (2017–2020)” was issued, and in 2018, the “Deadline Compliance Plan for Environmental Air Quality in Ziyang City” was formulated. As a result, the regional low-carbon efficiency was high during the research period. However, considering that Ziyang is currently in an important period of accelerated industrialization and urbanization, its future low-carbon transformation work still faces considerable pressure and difficulty, which is similar to the findings of Lin et al. [107] on less-developed areas.

Zigong, Mianyang, Suining, and Leshan had dense populations and were in the middle stage of urbanization, with relatively complete urban infrastructure. Due to the advantages in national policy support, energy conservation and emission reduction space, they were the developmental regions of low-carbon transformation in Sichuan Province. However, in terms of industrial structure, the proportion of the secondary industry in these cities reached 37.92%, 41.75%, 47.87%, and 42.97%, respectively in 2022, showing the characteristics of prominent heavy industry development. The industrial sector was the main body of regional carbon emissions, especially the development of six high energy-consuming industries: mining, automobile, steel, transportation, chemical industry, and building materials. Meanwhile, some enterprises believed that low-carbon development would inevitably bring high costs and low profits, and had little knowledge of carbon trading, leading to serious resistance. Therefore, although the low-carbon efficiency of these cities was relatively high during the research period, it will be difficult to change the slow trend of carbon reduction if the concept of high-quality development is not creatively and quickly rooted in people’s minds in the future, which is consistent with the findings of Chen et al. [108]. During the 14th Five-Year Plan period, they can adopt the method of “breaking the whole into parts” when facing the problem of industrial development solidification. Specifically, for regions with less emphasis on low-carbon transformation, the focus should be on building pilot parks with low-carbon, net-zero carbon, and even zero carbon characteristics, implementing innovative solutions such as industry low-carbon access standards, and gradually forming a demonstration effect of low-carbon development. For regions making significant efforts in the low-carbon transformation, the focus should be on eliminating backward production capacity and resolving excess capacity, implementing environmental restoration and governance, or actively building ecological tourism towns with low-carbon education and energy culture as the core.

The low-carbon efficiency of plateau areas in western Sichuan, such as Aba Tibetan and Qiang Autonomous Prefecture and Ganzi Tibetan Autonomous Prefecture, was relatively low, and the evaluation values in 2022 are 0.3521 and 0.6562, respectively. This was due to the fragile ecological environment, complex geological structure, low level of urbanization, extremely low proportion of technology-intensive industries and service industries, and backwards technology level in these prefectures, which is similar to the findings of Wang et al. [109] on the Qinghai–Tibet Plateau. Meanwhile, residents’ education level was generally low and their awareness of environmental protection was low. The pressure of resources, survival, and environment significantly hindered the process of low-carbon transformation. As a result, it was difficult to promote clean production methods, and the implementation effect of emission reduction measures was not satisfactory. However, from the overall perspective of social development, in recent years, with the implementation of projects such as “Shared bicycles turning into sports fields”, “Reforestation of Degraded Land in Northwest Sichuan” and “Photovoltaic Poverty Alleviation” in the western region, the concept of green and low-carbon life had been actively advocated. It promoted precise poverty alleviation by green poverty reduction, made up for the shortcomings of the ecological compensation policy that relied solely on the low compensation standard of the government, and enabled residents to enjoy the benefits brought by science and technology and environmental protection, thus promoting the sustainable development of ethnic areas. Meanwhile, the wind, solar, and water resources in the western Sichuan Plateau are unique, among which hydropower is the most reliable clean energy in Sichuan, and it is also an important guarantee to achieve regional energy structure upgrading [110]. Therefore, the low-carbon development in the above prefectures has a good prospect in the future. Dazhou and Ya’an were also potential areas of low-carbon transformation in Sichuan Province, with evaluation values of 0.6137 and 0.6541 in 2022, respectively. They faced a series of challenges, such as weak geographical advantages, a strong dependence on resources, insufficient endogenous motivation for transformation, and insufficient attraction to low-carbon innovative talents. Some grassroots parties and government agencies were still in the early stages of understanding carbon peaking and carbon neutrality, equating low-carbon development with stagnant development. In this regard, cognitive misconceptions inevitably constrained the construction of low-carbon cities, which is consistent with the findings of Zhao et al. [111]. However, at the same time, these regions were rich in forest resources. Especially, the forest coverage rate of Ya’an exceeded 69%, ranking first in the province, and was awarded the title of “National Forest Tourism Demonstration City” in 2018. It could provide strong support for the increase in regional carbon sink reserves. In the future, these cities should fully rely on the advantages of natural resources to promote the deep integration of primary and tertiary industries. In addition, the potential areas of low-carbon transformation also included underdeveloped cities such as Guangyuan, Bazhong, and Neijiang. In these cities, the phenomenon of pollution haven still existed, and loose environmental regulation situation attracted high-carbon enterprises transfer in pursuit of profits, which put great pressure on the improvement in regional low-carbon efficiency. It is similar to the findings of Guan et al. [112] on the least developed region. In the future, they should pay attention to avoiding the blind path of economic development before environmental governance.

5. Conclusions and Policy Implications

This study calculated the carbon emissions in Sichuan Province from 2015 to 2022 based on four major units: energy activity, industrial production, forestry activity, and waste disposal, and the time evolution characteristics and key sources of carbon emissions were investigated from a systematic perspective. Meanwhile, based on the Super-SBM-Undesirable model, the low-carbon efficiency of Sichuan Province and its 21 cities (states) was evaluated, and its spatial heterogeneity characteristics were further investigated. The main research conclusions are as follows: (1) energy activity was the main contributor to regional carbon emissions, with thermal power generation and industrial energy terminal consumption as the key sectors. Inter-regional power allocation indirectly reduced the regional emission intensity. The carbon emissions of industrial production showed significant aggregation in cement and steel production. The forest carbon sink had a significant effect on alleviating the regional greenhouse effect. The carbon emissions of waste disposal were small. (2) From 2015 to 2022, the low-carbon efficiency of Sichuan Province showed an overall upward trend. Chengdu had a high level of economic development, a reasonable industrial organization, and a continuous increase in greening rate. Heavy industrial cities such as Panzhihua and Deyang made great efforts to eliminate backward production capacity and low-carbon transformation of key industries. Therefore, they were the first mover advantage regions of low-carbon transformation. Zigong, Mianyang, Suining, and Leshan enjoyed favorable preferential policies and energy-saving and carbon-reducing space, and were developmental regions. The low-carbon efficiency of plateau areas in western Sichuan was relatively low, which was caused by the fragile natural environment, insufficient market potential, and backward technical level. Cities such as Ya’an and Bazhong faced a series of challenges such as weak geographical advantages, insufficient endogenous motivation for transformation and the risk of pollution haven, so they were potential areas.

Based on the research findings, the following policy implications are proposed: (1) Sichuan Province should accelerate the electrification transformation and clean utilization of coal, and promote the complementary and comprehensive utilization of solar photovoltaic, wind power, nuclear power and hydropower. Energy infrastructure such as energy storage and peak shaving devices and high-voltage transmission and transformation power grids should be improved to solve the consumption problem caused by seasonal and random output of clean energy. For key industries such as steel and cement production, low-carbon transformation guidelines should be issued. Specifically, non-carbonate raw materials such as industrial solid waste can be used to replace raw materials for cement production. New steel materials or renewable materials can be used to replace traditional steel, and the proportion of electric furnace steel production should be further increased. The construction of a forest carbon sink should be actively supported by projects such as land greening and ecological restoration. The normalization of forest and grassland fire prevention work, as well as daily monitoring of quarantine pests such as pine wood nematodes and red imported fire ants should be continuously carried out. The monitoring and management of waste incineration should be strengthened, and the level of reduction, harmlessness, and resource treatment of urban domestic waste should be improved. (2) The first mover advantage regions should fully leverage their radiating and driving effects, and target the increase in energy consumption towards natural gas and new energy. Their scientific and technological research should focus on key areas such as CCUS, efficient utilization of renewable energy, energy storage and distributed utilization. The developmental regions should seek new growth points on the basis of traditional pillar industries, strengthen regional cooperation in low-carbon technological innovation, and design policies that are beneficial to improving the competitiveness of clean energy enterprises. The potential areas should actively learn from advanced experience and introduce low-carbon advantageous industries and innovative talents from the first mover advantage regions. Meanwhile, efforts should be made to create new development models such as recreational tourism, green agriculture, and cultural and creative industries, and to transform ecological advantages into development advantages. In addition, the western plateau region should actively explore new paths for achieving a win-win situation between green, low-carbon, and comprehensive prosperity. The construction of green high-energy industries such as big data centers and ecological industries such as agriculture and animal husbandry with plateau characteristics should be promoted.

This study offers some innovative insights, but it also has significant limitations that may open up new areas for further study. First, this study mainly uses the reference emission factors provided in the official guidelines. In the future, carbon emission calculation can be carried out based on the measured field data to further improve the accuracy. Moreover, limited by the availability of data, the indicator system we built in this study, although objectively reflecting the low-carbon efficiency, it can further be optimized.