Analysis of the carbon emission driving factors and prediction of a carbon peak scenario——A case study of Xi'an city

Abstract

The continuous growth of global carbon emissions has become the focus of attention in political and academic circles in various countries. Understanding the driving factors of change in urban carbon emissions and predicting the peak of carbon emissions is of great significance for guiding the formulation of urban as well as national carbon emission reduction policies. Using Xi'an as an example, this study analyses the changing trend of its carbon emissions over the past 20 years. Based on carbon emissions and total economic volume, a Tapio decoupling elasticity analysis model was constructed, the decoupling coefficient of Xi'an from 2000 to 2020 was calculated, and the decoupling status of economic growth and carbon emissions were analysed. Using the Kaya identity and logarithmic mean divisia index (LMDI) decomposition to analyse the driving factors of the city's carbon emissions, combined with a multi-scenario forecasting method, three different scenarios were subdivided, and the approximate time of Xi'an's carbon peak was estimated. The results show that from 2000 to 2020, the overall carbon emissions in Xi'an showed an upwards trend. In recent years, the decoupling status of economic growth and carbon emissions in Xi'an has been ideal, and the effect of carbon emission reduction is obvious. Population and per capita gross domestic product (GDP) have a positive driving effect on carbon emissions, and energy intensity has a negative driving force on carbon emissions. During early years, the carbon intensity of energy consumption showed a positive effect on carbon emissions. With the improvement of the energy structure, the intensity of energy consumption inhibits the growth of carbon emissions. Under the three scenarios of low carbon, baseline and high carbon, the carbon peak years will be achieved approximately in 2016, 2025 and 2035, and the corresponding carbon peaks are approximately 29.5 million tons, 29.66 million tons and 31 million tons, respectively.

Carbon emission; Carbon peak; Decoupling state; LMDI; Multi-scenario forecast.

1. Introduction

Global climate change caused by excessive carbon dioxide emissions has become a real problem that countries around the world are facing [1, 2]. As the country with the largest carbon emission volume, China's carbon emissions at the urban level cannot be ignored. According to statistics, the global urban carbon emissions account for approximately 75% of the total emissions, and the urban area in China is slightly higher than this proportion [3, 4]. Technological improvement and energy transformation can effectively reduce regional carbon emissions, for example, in the automotive field, optimizing vehicle performance through energy-saving technologies can effectively reduce fuel consumption [5, 6] or using hydropower and solar energy to promote energy transformation, can protect and utilize more renewable energy, and effectively reduce the environmental pollution caused by energy consumption [7, 8, 9, 10]. However, in terms of carbon emissions at the city level, it is necessary to clarify the internal relationship between carbon emissions and urban development to make better use of renewable energy and related emission reduction technologies. With the continuous urbanization of China, especially after the reform and opening up, the development speed of Chinese cities has reached an unprecedented height. A large amount of urbanization construction and rapid urban expansion are accompanied by a large amount of fossil energy consumption, and hence, the problem of urban environmental pollution is becoming increasingly serious [11, 12]. Therefore, understanding the driving factors of urban carbon emissions and roughly predicting the peak time of carbon emissions in cities can provide a theoretical basis for cities to rationally formulate carbon emission reduction policies and boost confidence in carbon emission reductions.

At present, carbon emission-related research has attracted extensive attention from relevant scholars. Scholars pay attention to the long-term temporal and spatial changes in carbon emissions through energy statistical data, night light data and land use data [13, 14, 15, 16], and through refining simulations and performing spatial heterogeneity analysis. With the deepening of carbon emission research, scholars have begun to study typical cities and developed economies to analyse the driving factors of carbon emissions in urban areas and to predict their carbon peak time. Fan et al. (2020) [17] used the LMDI decomposition method to decompose the carbon emissions of energy consumption in the Beijing-Tianjin-Hebei region into five factors: population, economy, industrial structure, energy structure, and energy consumption structure. The results show that population and per capita GDP promote the growth of carbon emissions, while the overall industrial structure effect, energy intensity effect and energy consumption structure factors inhibit the growth of carbon emissions. Liu et al. (2021) [18] studied the carbon peaking and classified emission reduction paths of cities in the Yangtze River Economic Belt. The results show that energy-related and industry-related carbon emissions in the Yangtze River Economic Belt are the main sources of carbon emissions. It is expected that the Yangtze River Economic Belt will reach its carbon peak before 2030, whereby, cities are classified, and carbon emission reduction paths are proposed. Differences in the selection of study areas and the methods used by different scholars have led to slightly different analytical results. The specific literature review is shown in Table 1.

Table 1.

Overview of research on the decomposition of carbon emission factors and prediction of carbon peaking.

Research topics	Year	Author	Research method	Main conclusion
Carbon emission decomposition	2020	Li B et al.	GDIM	GDP has the highest cumulative contribution rate to China's construction industry carbon emissions [19].
	2019	Li Z G et al.	GDIM	Output scale, energy consumption scale, and investment scale have a strong role in promoting carbon emissions [20].
	2019	Ma X J et al.	GDIM	Output scale effect, technological progress effect, energy consumption scale effect and per capita carbon emission effect are the main factors leading to the increase of industrial carbon emissions [21].
	2018	Wang H et al.	IDA	The narrowing disparity among countries in per capita consumption level drove the inequality reduction, while the expanding disparity in consumption-based emission intensity largely hindered it [22].
	2022	Wang Q et al.	IDA	Advances in energy-saving technologies can effectively curb the growth of carbon emissions [23].
	2022	Wu X T et al.	LMDI	Economic output effect is the main factor affecting carbon emissions in the EGT area, followed by the population effect and energy structure effect, while energy intensity effect slows carbon emissions [24].
	2022	Chen Y	LMDI	The carbon emission intensity of the counties in China continues to grow, and population and economic growth lead to certain levels of carbon emissions [25].
Carbon emission forecast	2018	C.B. Wu et al.	STIRPAT model	It is verified that the STIRPAT model can be used for the prediction of carbon dioxide, and the fitting results are good [28].
	2022	Zhang S et al.	STIRPAT model	the forecast results show that the carbon emissions of the construction industry in Jiangsu Province will generally show a downwards trend in the future [29].
	2021	Zhang F et al.	Multi-scenario forecast	Under all the SSP scenarios, the western region was always the first to reach its peak value, followed by the central region and then the eastern coastal zone [30].
	2021	Li J H et al.	Multi-scenario forecast	The forecast results show that carbon emissions will increase significantly from 2000 to 2030, and carbon emission reduction policies should rely more on the domestic product sector [31].
	2022	Zou X et al.	LEAP model	The results show that in the absence of other mitigation policy interventions, Shanxi's CO2 emissions will increase year by year, reaching 1,646.2 million tons by 2035 [32].
	2022	Zhang Z X et al.	LEAP model	The CO2 emissions of the five northwestern provinces under the baseline scenario, policy scenario and green scenario will peak in 2035, 2031 and 2027, respectively [33].

To decompose the driving factors of carbon emissions and quantitatively analyse the contribution of factors affecting urban carbon emissions, scholars mainly use the generalized dividing index decomposition model (GDIM), index decomposition analysis model (IDA) and LMDI model to decompose the driving factors [19, 20, 21, 22, 23, 24, 25]. As shown in Table 1, scholars such as Li B et al, Wang H et al, and Wu X T et al applied the GDIM, IDA, and LMDI model to decompose the influencing factors of carbon emissions, respectively. Among them, the LMDI model has been widely used because of its many advantages. It eliminates unexplained residuals and quantifies the impact of factors such as population size, energy intensity, energy structure, and industrial structure on regional carbon emissions through decomposition, and is considered one of the main methods of research in the field of carbon emissions [26, 27]. At the carbon emission forecast level, the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) model, multi-scenario analysis and Long-range Energy Alternatives Planning System (LEAP) model are considered the main adopted methods [28, 29, 30, 31, 32, 33]. For instance, C.B. Wu et al., Zhang F et al., and Zou X et al. applied STIRPAT, the multi-scenario forecasting method and the LEAP model for carbon emission prediction, respectively. In recent years, Xi'an city has promulgated the 14th Five-Year Plan and the outline of the 2035 Visionary Goals, and energy conservation and emission reduction efforts have continued to increase. Considering the characteristics of urban development, this paper selects the LMDI decomposition method to decompose the driving factors for the research cities in this study and applies it to the field of urban spatial decomposition analysis. Based on this, the multi-scenario prediction method is selected to predict the urban carbon peak, which visually reflects the differences in factors affecting urban carbon emissions and the development trend of carbon emissions under different scenarios.

The significant differences among Chinese cities in terms of industrial base, energy structure and economic development level lead to spatial differences in carbon emissions and their influencing factors among regions [34, 35]. Consequently, it is necessary to study the driving factors of carbon emissions and predict carbon peaks in combination with regional characteristics and urban development trends, which will help decision-makers intuitively understand the current status of urban carbon emissions and formulate appropriate energy-saving policies according to local conditions. As the only national central city in northwest China, Xi'an is located in an important energy-rich area in western China [36]. While urban development is entering a new stage, understanding the relationship between carbon emissions and economic development and analysing the driving factors are of great significance for balancing the relationship between energy and the low carbon economy and promoting the carbon peak in northwest China. This paper analyses the evolution of the long-term trend of carbon emissions in the city over the past 20 years, calculates the decoupling state of carbon emissions and economic growth, decomposes the driving factors of carbon emissions, and predicts their peak value. In contrast to previous studies, this paper mainly involves the following aspects. First, compared with the developed cities and regions in the central and eastern regions, there are relatively few studies for the northwest region, and because of the huge urban development differences between the eastern and western regions, it is of representative significance to select important megacities in the western region as the study area. Second, introducing the carbon intensity of energy consumption as a new decomposition factor can intuitively reveal the impact of the carbon intensity of energy consumption on carbon emissions. Third, combining the decoupling state of carbon emissions with the decomposition results of driving factors comprehensively expounds the current urban development model and carbon emissions.

2. Methodology

To achieve the research purpose of this study, this paper adopts the following four methods: calculate the carbon emissions of energy consumption using the formula provided by the Intergovernmental Panel on Climate Change (IPCC); calculate the decoupling coefficient between economic growth and carbon emissions through the Tapio decoupling model; use the LMDI decomposition model to analyse the driving factors of carbon emissions; and conduct a multi-scenario analysis to predict the carbon peak. The research framework is shown in Figure 1, and the specific details of the research methods are shown below.

Figure 1.

Research flow chart.

2.1. Estimation of carbon emissions

City-level carbon emissions are mainly derived from energy consumption carbon emissions. According to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Volume II (Energy) Carbon Emissions Calculation Method, an estimate of carbon emissions is established. Based on this estimate, to calculate the carbon emissions of energy consumption from 2000 to 2020, the calculation formula is as follows.

$$C=\frac{44}{12}{\sum }_{i=1}^n{K}_i{E}_i$$ (1)

where $C$ is the carbon emission, $E_i$ is the energy consumption of the $i$ th fuel, and $K_i$ is the carbon emission factor of the $i$ th fuel.

According to the data records of the Xi'an Statistical Yearbook, this paper selects 11 types of energy sources, such as raw coal, washed coal, coke, natural gas, gasoline, kerosene, diesel oil, fuel oil, liquefied petroleum gas, heat, and electricity, to calculate carbon emissions.

2.2. Tapio model

The practical significance of carbon emission decoupling is to realize the continuous reduction of energy consumption on the basis of economic growth. It is an idealized process, and its essence is to measure whether economic growth is at the expense of the environment. This study uses the Tapio decoupling index to measure the decoupling status of carbon emissions and economic growth and analyses the decoupling relationship [37]. The calculation method of the decoupling coefficient is shown in Eq. (2).

$$\epsilon =\left(\frac{\Delta V}{V}/\frac{\Delta GDP}{GDP}\right).\left(\frac{{\Delta CO}_2}{{CO}_2}/\frac{\Delta V}{V}\right)=\frac{\raisebox{1ex}{${\Delta Co}_2$}\!\left/ \!\raisebox{-1ex}{${CO}_2$}\right.}{\raisebox{1ex}{$\Delta GDP$}\!\left/ \!\raisebox{-1ex}{$GDP$}\right.}$$ (2)

where $\epsilon$ is the Tapio decoupling elasticity index, $V$ is the intermediate variable for calculating the decoupling index, and ${CO}_2$ refers to carbon emissions.

According to the classification of the Tapio decoupling elasticity index model, the decoupling state can be subdivided into 8 types of states, as shown in Figure 2. As seen from the schematic diagram, the second and fourth quadrants are in two very extreme states, which are strong decoupling and strong negative decoupling. The first and third quadrants are shown as intermediate states, which are six decoupling states: expansion negative decoupling, expansion connection, weak decoupling, weak negative decoupling, recessionary connection, and recessionary decoupling. When $\Delta C$ >0, the smaller the $\epsilon$ index is, the lower the dependence of carbon emissions on economic growth; conversely, when $\Delta C$ <0, the lower the $\epsilon$ index is, the more this indicates that carbon emissions are more dependent on economic growth.

Figure 2.

The diagram of the Tapio decoupling index.

2.3. LMDI carbon emission factor decomposition model based on the kaya identity

The Kaya identity is the mainstream carbon emission factor decomposition model, which combines economic development, population, policies and other factors with carbon emissions to analyse regional carbon emissions [38, 39, 40]. Using the Kaya identity combined with the change characteristics of carbon emissions in Xi'an city, the LMDI decomposition model of carbon emissions in Xi'an city is constructed. The specific decomposition model is shown in Eq. (3)

$${CE}_t=\sum _{i=1}^n{P}_t\times \frac{G_t}{P_t}\times \frac{E_t}{G_t}\times \frac{{CE}_t}{E_t}=\sum _{i=1}^{n}P\times Q\times M\times N$$ (3)

where ${CE}_t$ is the carbon emissions of Xi'an in year $t$, $G_t$ is the regional GDP, $E_t$ is the total energy consumption, $P$ is the population size, $Q$ is the per capita GDP, $M$ is the energy consumption intensity, and $N$ is the carbon intensity of energy consumption.

According to the results of the LMDI decomposition model, the comprehensive effect $\Delta C$ of carbon emission growth is divided into 4 parts: population size effect $\Delta C_P$; per capita GDP effect $\Delta C_Q$; energy intensity effect $\Delta C_M$ and carbon intensity effect of energy consumption $\Delta C_N$. The decomposition formulae for the different drivers are shown in Eqs. (4), (5), (6), and (7).

$$\Delta C_P=\sum _{i=1}^n\frac{C^t-C^{t-1}}{\ln C^t-\ln C^{t-1}}\ln \left[\frac{P\left(t\right)}{P\left(t-1\right)}\right]$$ (4)

$$\Delta C_Q=\sum _{i=1}^n\frac{C^t-C^{t-1}}{\ln C^t-\ln C^{t-1}}\ln \left[\frac{Q\left(t\right)}{Q\left(t-1\right)}\right]$$ (5)

$$\Delta C_M=\sum _{i=1}^n\frac{C^t-C^{t-1}}{\ln C^t-\ln C^{t-1}}\ln \left[\frac{M\left(t\right)}{M\left(t-1\right)}\right]$$ (6)

$$\Delta C_N=\sum _{i=1}^n\frac{C^t-C^{t-1}}{\ln C^t-\ln C^{t-1}}\ln \left[\frac{N\left(t\right)}{N\left(t-1\right)}\right]$$ (7)

The total carbon emission effect of Xi'an can be expressed as Eq. (8).

$$\Delta C=\Delta C_P+\Delta C_Q+\Delta C_M+\Delta C_N$$ (8)

2.4. Carbon peak scenario prediction model

2.4.1. Scenario forecasting method

The coupled model comparison program, which was designed to better analyse past, present, and future climate change, has become a key element of national and international climate change assessments. The scientific combination scenario of shared socioeconomic pathways (SSPs) and representative concentration pathways (RCPs) proposed has been widely used in studies on climate and carbon emissions [41]. Specific scenarios are defined in Table 2. In recent years, China has issued the 14th Five-Year Plan and Vision 2035 at the national level and has actively implemented policies in the areas of energy conservation and environmental protection, people's livelihoods and well-being, and planning the future development of the economy and society. Taking into account the general trend of social development, the multi-scenario setting corresponds to the composite scenario framework in the RCPs-SSPs programme, simulating three possible future scenarios and the corresponding policies, which helps to highlight the theoretical basis for the scenario setting and serves as a better reference for the scenarios, making the three scenarios more scientific. To scientifically simulate the future carbon emission trend of Xi'an, combined with the RCP-SSP scenario, three carbon emission scenarios were set based on the city's social development trend, emission reduction policies and specific actual conditions [42], as shown in Table 3 below.

Table 2.

RCP-SSP scenario definition [42].

Model	Scenario	Definition
Shared Socioeconomic Pathways	SSP1	Significantly improved energy efficiency for sustainable development
	SSP2	Moderate growth in population, efforts to achieve Sustainable Development Goals
	SSP3∼SSP4	Slow economic development and high population growth rate
	SSP5	Maintain high economic growth rate and not commit to energy conservation and emission reduction
Typical Concentration Pathway	RCP8.5	The government does not get involved in formulating carbon emission reduction policies at all
	RCP6	The government actively intervenes, but the emission reduction effect is not significant
	RCP4.5	The government actively intervenes, and achieve a certain emission reduction effect
	RCP2.6	Humans are actively responding to climate change, and greenhouse gases are declining year by year

Table 3.

Scenario setting for carbon emissions in Xi'an.

scenarios	Main basis	Scenario Description	Scenario support
Baseline	RCP6-SSP2	Active government intervention and certain low carbon policies to accelerate the optimization of the energy mix.	Issuing the 14th Five-Year Plan and other policies, proposing industrial upgrading, energy transformation and other related measures [43, 44, 45].
High carbon	RCP6-SSP5	Urban population and GDP per capita are increasing, carbon emission reduction is not obvious and carbon emissions continue to grow.	China's urbanisation process, with rising urbanisation rates and continued growth in urban population [46, 47].
Low carbon	RCP4.5-SSP2	The government attaches great importance to environmental protection, vigorously deploying energy-saving and emission reduction policies and accelerating the research and development of low-carbon technologies.	The country is currently stepping up its efforts to invest in renewable and clean energy to ensure technical support [48, 49, 50].

Combined with the basic method of the above factor decomposition model, the above three possible carbon emission scenarios are set, and the changes in future urban carbon emissions are analysed with 2020 as the base year. The changes in the driving factors of the city's carbon emissions in year $t+1$ can be expressed as follows Eqs. (9), (10), (11), and (12):

$$P_{t+1}=P_t.\left(1+\alpha \right)$$ (9)

$$Q_{t+1}=Q_t.\left(1+\beta \right)$$ (10)

$$M_{t+1}=M_t.\left(1+\gamma \right)$$ (11)

$$N_{t+1}=N_t.\left(1+\delta \right)$$ (12)

Based on this sequence of formulas, the city's carbon emissions in year $t+1$ can be expressed as Eq. (13):

$$C_{t+1}=P_{t+1}.Q_{t+1}.M_{t+1}{N}_{t+1}=P_t\left(1+\alpha \right).Q_t\left(1+\beta \right).M_t\left(1+\gamma \right).N_t\left(1+\delta \right)$$ (13)

where $\alpha 、\beta 、\gamma 、\delta$ represent the change rate of population size, the change rate of per capita output, the change rate of energy consumption intensity and the change rate of energy consumption carbon intensity, respectively. Therefore, the change rate of the city's carbon emissions can be expressed as Eq. (14):

$$\phi =\left(1+\alpha \right).\left(1+\beta \right).\left(1+\gamma \right).\left(1+\delta \right)-1$$ (14)

Future carbon emissions can be derived based on the predicted values of these four factors.

2.4.2. Scenario parameter setting

Population size. The population growth rate of Xi'an from 2010 to 2019 is shown in Figure 3. Over the years, the city's population growth rate has been maintained at approximately 2%. Due to the implementation of the two-child policy in 2016, the city's population growth rate has increased significantly and then sharply declined. The rate is expected to return to 2020 levels, which was a growth rate of approximately 2.5%. With the introduction of the three-child policy and the improvement of related supporting measures, as well as the impact of the new coronavirus epidemic on society, it is expected that the city's population growth rate will increase slightly during the next few years and then gradually decrease. Referring to the previous literature [51, 52] and the current policy orientation, the detailed prediction parameters are shown in Table 4 below.

Figure 3.

Change rate of population, per capita GDP, energy intensity, and carbon intensity of energy consumption.

Table 4.

Carbon emission scenario setting.

Scenario	Time	P (%)	Q (%)	M(%)	N (%)
Baseline	2020–2025	2.7	6.0	-3.5	-4.5
	2025–2030	2.1	5.5	-3.2	-4.2
	2030–2035	1.5	5.0	-2.9	-3.9
High carbon	2020–2025	3.5	7.0	-3.1	-4.2
	2025–2030	2.7	6.2	-2.9	-4.1
	2030–2035	1.9	5.4	-2.7	-4.0
Lower carbon	2020–2025	2.3	5.5	-3.9	-4.7
	2025–2030	1.8	5.1	-3.6	-4.5
	2030–2035	1.3	4.7	-3.3	-4.3

GDP per capita. China's economic development has gone through a period of rapid development [53]. The city's per capita GDP growth rate is shown in Figure 3, and it demonstrates the growth in recent years. In 2020, the city's GDP exceeded one trillion yuan, an increase of 5.2% over the previous year. According to the "14th Five-Year" Development Outline issued by the Xi'an Municipal People's Government, it is estimated that by 2025, the city's economic aggregate will reach more than 1.4 trillion yuan, and the per capita GDP will also show a slow decline [54]. Referring to the prediction parameters and policy trends of Xi'an city by predecessors [52], the specific prediction parameters are shown in Table 4 below.

Energy intensity. Energy intensity is an important indicator of energy efficiency, representing energy consumption per unit of GDP. The energy intensity of Xi'an city was -8.94% and -7.85% during the "Twelfth Five-Year Plan" and "Thirteenth Five-Year Plan" periods, respectively. The rate of change of energy intensity showed a continuous downwards trend, and this trend weakened yearly [55]. According to Xi'an's 14th Five-Year Plan and Outline of Vision 2035, the carbon emission intensity will be continuously reduced, and the industrial structure will be further optimized [56]. With the adjustment of the energy structure and the improvement of energy utilization efficiency, it can be predicted that the city's energy intensity will decline steadily. The specific forecast parameters are shown in Table 4.

Carbon intensity of energy consumption. The carbon intensity of energy consumption mainly depends on the structure of energy consumption. To achieve the carbon peaking target as scheduled, the State Council issued the "Carbon Peaking Action Plan before 2030", pointing out that by 2025, the proportion of nonfossil energy consumption will reach approximately 20%, and by 2030, the proportion of nonfossil energy consumption will account for approximately 25% [57]. At present, Xi'an city has clarified the transformation tasks of manufacturing and high-energy-consuming industries, accelerated the proportion of investment in new energy industries, and improved the energy consumption structure. The specific parameters are shown in Table 4 below.

3. Study areas and data

3.1. Study areas

The Guanzhong Plain urban agglomeration is the eighth national-level urban agglomeration approved in China following the Beijing-Tianjin-Hebei, middle reaches of the Yangtze River, Chengdu-Chongqing, Harbin-Chongqing, Yangtze River Delta, Central Plains, and Pearl River Delta urban agglomerations. The National Development and Reform Commission proposed building Xi'an, the core city of the Guanzhong Plain urban agglomeration, as a national central city during the same year. Therefore, Xi'an became the ninth national central city and the only approved mega city in the northwest region. At the same time, as the hub of the northwest region, it plays an important role in leading the economic development process of the entire northwest region. Under national strategies such as "One Belt One Road", the development of Xi'an has very broad prospects. The northern part of Xi'an is an alluvial plain, and the southern part is a denuded mountain. It has a continental monsoon climate with distinct seasons. In 2020, Xi'an had a total area of 10,108 square kilometres, 11 districts and 2 counties under its jurisdiction, with a permanent population of 12.9529 million and a total GDP of 1,002.039 billion yuan. Figure 4 shows the extent of the study area.

Figure 4.

Scope of the study area.

3.2. Data sources

This paper selects a total of 11 energy sources to calculate the annual carbon emissions of Xi'an from 2000 to 2020. The energy consumption data come from the 2001 to 2021 Xi'an Statistical Yearbook. Population data and GDP data also come from the Xi'an Statistical Yearbook. The auxiliary data are the administrative boundary data of prefecture-level cities of 1:4 million in China, which can be queried through the website of the National Basic Geographic Information System.

4. Results and analysis

4.1. The spatiotemporal characteristics of carbon emissions

Based on Eq. (1), we calculate the city's carbon emissions for 20 years. As shown in Figure 5, from 2002 to 2016, the city's carbon emissions showed an upwards trend, reaching a peak in approximately 2016, which is approximately three times its level in 2000, and carbon emissions increased rapidly. Since 2016, the city's carbon emissions have begun to slightly decline yearly, and carbon emissions have gradually decreased.

Figure 5.

Carbon emissions and their growth rates in Xi'an from 2000 to 2020.

According to the change in the regional carbon emission growth rate, the period from 2000 to 2020 can be divided into three stages: the high-speed growth stage (2000–2006), medium-growth stage (2008–2014), and negative growth stage (2016–2020). From 2000 to 2006, carbon emissions in Xi'an increased by an average of 21%. At this stage, the city overpursued urban development and economic benefits, and it was a period of trading economic growth at the expense of the environment, which brought significant damage to the natural environment, and defined as a high growth stage. From 2008 to 2014, the city’s economic development entered a new normal, the city’s development goals were no longer at the expense of the environment, and the energy structure adjustment and industrial structure upgrade were strengthened. Then, the growth rate of carbon emissions slowed down significantly. During this period, the average growth rate dropped to 7.4%, which is defined as the medium growth stage. From 2016 to 2020, the state successively introduced new environmental protection policies and advocated new development models, and public awareness of environmental protection was greatly enhanced. Adhering to the new concept of green and low-carbon, vigorously promoting the construction of green and low-carbon industries and energy-saving cities, environmental problems have been greatly improved. During this period, the city's carbon emissions showed a downwards trend, which is an important step towards the goal of building a low-carbon city, and is one step defined as the negative growth stage. The spatial and temporal evolution of regional carbon emissions reflects the phased changes in urban development and indicates the different tasks of urban development at different stages.

4.2. Analysis of the decoupling status of carbon emissions and economic growth

Calculation of decoupling factor according to Eq. (2). Table 5 reveals the decoupling elasticity results of economic growth and carbon emissions in Xi'an from 2000 to 2020. Overall, the change trend of the decoupling index $\epsilon$ in the city in the past 20 years has been unstable, with rising and falling trends alternating. The decoupling states are mainly expanded negative decoupling, weak decoupling, expanding connection, and strong decoupling. To further understand the difference in the decoupling relationship between the city’s economic growth and carbon emissions, it can be roughly divided into two main development stages to observe changes in decoupling.

Table 5.

Decoupling status of carbon emissions in Xi'an.

Year	% $\Delta c$	% $\Delta GDP$	$\epsilon$	Decoupling state
2001	0.153	0.120	1.270	expansive negative decoupling
2002	0.006	0.111	0.054	weak decoupling
2003	0.124	0.108	1.150	expansive connection
2004	0.266	0.152	1.750	expansive negative decoupling
2005	0.059	0.156	0.378	weak decoupling
2006	0.108	0.144	0.750	weak decoupling
2007	0.020	0.186	0.108	expansive connection
2008	-0.036	0.197	-0.183	strong decoupling
2009	0.199	0.139	1.432	expansive negative decoupling
2010	-0.056	0.158	-0.354	strong decoupling
2011	-0.040	0.157	-0.225	strong decoupling
2012	0.055	0.132	0.417	strong decoupling
2013	0.187	0.119	1.571	expansive negative decoupling
2014	-0.146	0.110	-1.327	strong decoupling
2015	-0.314	0.060	-5.230	strong decoupling
2016	-0.006	0.073	-0.073	strong decoupling
2017	0.283	0.138	2.051	expansive negative decoupling
2018	-0.024	0.127	-0.189	strong decoupling
2019	-0.069	0.096	-0.719	strong decoupling
2020	-0.054	0.062	-0.871	strong decoupling

From 2000 to 2010, the decoupling status was basically three decoupling statuses: expansionary negative decoupling, expansionary connection and weak decoupling, and the two decoupling states were mostly expansionary negative decoupling and weak decoupling. This set of statuses indicates that the level of economic development in Xi'an city and carbon emission levels are both in a positive stage and have a positive promoting effect. The general direction of development has changed from the expansionary negative decoupling to the weak decoupling and expansionary connection state, from the early carbon dioxide growth rate greater than the economic growth rate to the year-on-year positive growth of both and the economic growth rate greater than the carbon dioxide growth rate, to achieve economic growth and carbon emissions. The volume growth is stable and positive.

From 2010 to 2020, the decoupling state was basically strong decoupling, and there were expansionary negative decoupling and weak decoupling for a few years. Economic growth and carbon dioxide growth showed the reverse trend. This growth shows that economic growth is no longer accompanied by an increase in carbon dioxide emissions, continuous economic growth has effectively curbed the growth of carbon emissions, and the effect of emission reduction is obvious, which is an ideal decoupling state. This has a lot to do with the government's proposal for high-quality economic development, vigorously developing the tertiary industry, and actively promoting the transformation of the energy structure.

4.3. Decomposition analysis of the driving factors of carbon emissions

By decomposing the driving factors of carbon emissions by Formula 3, specific decomposition formula according to 4–7, the yearly decomposition results of the carbon emission factors in Xi'an are obtained and shown in Figure 6. Through the comprehensive analysis of the decomposition results of the city's carbon emission factors over the past 20 years, the population scale effect and the per capita GDP effect have made greater contributions to the growth of urban carbon emissions. At the same time, the energy intensity effect can restrain the growth of carbon emissions to a large extent. The carbon intensity effect of energy consumption shows a positive promotion of carbon emissions for a few years and shows a restraining effect on carbon emissions in most years. Improvement of the energy structure and optimization of the industrial structure. According to the data characteristics and changes in the city's carbon emissions, the research can be divided into three stages, namely, 2000–2005, 2005–2010, and 2010–2020.

Figure 6.

Yearly effect of carbon emission factor decomposition.

The first stage (2000–2005). Overall, both the population scale effect and the per capita GDP effect promote the increase in regional carbon emissions, and the per capita GDP effect is the main driving force of carbon emissions. The energy intensity effect can effectively suppress the increase in regional carbon emissions, but at this stage, the contributions of the energy intensity effect are all below -700,000 tons, indicating that the comprehensive utilization efficiency of energy is low, and the economic benefits of energy utilization are still greatly improved. At this stage, the carbon intensity effect of energy consumption is mostly manifested in the promotion of regional carbon emissions, and the energy structure urgently needs to be improved. The second stage (2005–2010). At this stage, the overall population size effect and per capita GDP effect still show a promoting effect on carbon emissions, and the energy intensity and energy consumption carbon intensity effects show a restraining effect on carbon emissions. Compared with the first stage, the contribution of the per capita GDP effect and energy intensity in this stage is much greater than that in the first stage, and the carbon intensity effect of energy consumption has also changed from promotion to inhibition in the previous stage, indicating that the economy has developed rapidly in this stage. The per capita GDP has increased significantly, and at the same time, the energy and industrial structures have improved to a certain extent. The third stage (2010–2020). The effect of population size and per capita GDP is still the main driving force of regional carbon emissions, and the effect of energy intensity and carbon intensity of energy consumption inhibits the growth of regional carbon emissions. At this stage, the city shifted from a rapid economic development model to a high-quality economic development model and no longer pursued a large economic volume. In addition, due to the prominent environmental problems brought about by rapid development, government departments have listed environmental governance as an important task for urban development, continuously optimized the industrial structure, and accelerated the transformation of the energy structure.

4.4. Prediction and analysis of the carbon peak scenario

The peak carbon emissions under different scenarios were calculated based on the forecast parameters and Eqs. (9), (10), (11), (12), (13), and (14). Figure 7 shows the forecast results of carbon emissions in Xi'an from 2020 to 2035. It can be seen from the figure that there are significant differences in the peak carbon emissions and carbon peak times under different scenarios. The carbon peak time of the low-carbon scenario is approximately in 2016, and the carbon emission peak is 29.5 million tons. Under the baseline scenario, the carbon peak time is approximately in 2025, and the carbon emission peak is 29.66 million tons. Under the high-carbon scenario, the carbon peak time is approximately in 2035, and the carbon emissions peak is 31 million tons. Generally, different carbon emission scenarios represent different development patterns, and peak carbon emissions are also the result of the combined effect of different urban development patterns and related policies.

Figure 7.

Carbon peak time under different scenarios.

Specifically, we discuss the time of the peak carbon emissions in Xi'an. The carbon peak time in the low-carbon scenario was achieved in approximately 2016. Afterwards, according to the set scenario parameters, carbon emissions began to show a decline. The goals of industrial transformation, energy structure optimization, and energy utilization efficiency improvement have been basically completed, and carbon emission reduction effects have been achieved. Obviously, it is an ideal carbon emission scenario. The baseline scenario achieves a carbon peak in approximately 2025. The parameters of the baseline scenario are based on the current urban development trend, the country's general policy and the government's implementation. It is the most realistic scenario. In other words, according to the current development trend, the city's carbon peak year will most likely be achieved in approximately 2025, which will be earlier than the country's 2030 carbon peak target. During this period, regarding the 14th Five-Year Plan of Xi'an, a series of feasible carbon emission reduction policies were formulated and implemented, including general economic and social development goals, industrial upgrading, industrial structure improvement, and vigorous development of new energy and other policy measures. The high-carbon scenario predicts that the carbon peak year will be achieved in approximately 2035. It implies that the government does not interfere with social development at all, does not issue relevant emission reduction policies, consumes many natural resources, pursues economic and social development and pollutes the environment recklessly. Carbon reduction is a significant pressure and is one of the worst development scenarios.

Using the baseline scenario as an example, Xi'an should continue its efforts to ensure that the carbon peak can be achieved by 2025, and the following tasks are expected to be completed by 2025. (1) Control the population size and ensure that the total population size of the city is maintained at approximately 10 million. (2) The economy will continue to achieve high-quality development. While maintaining stable economic growth, the per capita GDP will be approximately 108,000 yuan. (3) The energy structure will be optimized, speeding up industrial transformation, and effectively developing low-carbon technologies to reduce energy consumption per unit of GDP by approximately 12% and nonfossil energy to account for approximately 20%.

5. Discussion

5.1. Main findings

Based on the analysis and forecast of long-term carbon emissions in Xi'an, it is found that during the study period, the total carbon emissions in the city tended to be stable and showed a downwards trend. After 2010, a strong decoupling of economic growth and carbon emissions was basically achieved, and according to the decomposition results of the driving factors, energy intensity and energy consumption carbon intensity had obvious inhibitory effects on carbon emissions. However, compared with the developed eastern regions, such as Beijing-Tianjin-Hebei, Shanghai, and the Yangtze River Economic Belt, and based on the results, the city still presents a new model of green development, and the effect of carbon emission reduction is obvious, although the western cities are relatively backwards in urban development. This result is basically the same as that of some developed regions [58, 59, 60]. At present, Chinese society has entered a new stage of economic development, the growth rate of carbon emissions has slowed down significantly [61], and climate change in urban areas has also been improved [62]. To achieve the national carbon peak target as scheduled, all cities need to work together.

5.2. Policy implications

Based on the above research and taking into account Xi'an's own urban characteristics and development prospects, this paper puts forward some low carbon recommendations to provide theoretical support for the city to achieve peak carbon as scheduled. As reducing the energy intensity and carbon intensity of energy consumption can effectively curb the growth of carbon emissions in the city, urban emission reduction policies should be formulated with the objective of reducing the energy intensity and carbon intensity of energy consumption. For example, within the city region, the integration and innovation of green industries such as new energy vehicles, new energy materials and big data will be accelerated, and the proportion of clean energy consumption in urban areas will be increased. These policies will not only promote energy and industrial transformation at the city level but will also enable cities to develop in a smarter direction to achieve the goal of sustainable urban development.

5.3. Limitations

This paper has certain limitations, which need to be further explored to improve it. The LMDI factor decomposition model is a relatively intuitive decomposition method that is widely used in the analysis of regional carbon emissions. This paper takes cities as an example to decompose carbon emission factors and can also be extended to larger regions, such as applying this model to analyse the driving factors of carbon emissions in China, the European Union, and the United States. However, in the process of decomposing the driving force of carbon emissions, due to the differences in regional development characteristics and regional statistical data, different scholars choose slightly different types of model decomposition, such as industrial structure, energy structure, and energy intensity, resulting in a slight difference in the decomposition results. In addition, this paper makes a scenario prediction on the carbon emissions of Xi'an. The setting of the scenario parameters refers to the parameter change rate of Xi'an in the past 10 years and takes into account the previous research results, current social development trends and national policies, which is persuasive. Subsequent research can be combined with simulation prediction software, such as the LEAP model, to analyse the difference in the results obtained by the two methods.

6. Conclusion

This study calculates the carbon emissions of Xi'an over the past 20 years, uses the Tapio decoupling method to analyse the decoupling status of the city's carbon emissions and economic growth, and uses the LMDI decomposition model to decompose the driving factors of the city's carbon emissions. Finally, combined with multiple scenarios, analytical methods predict peak carbon emissions. The current status and driving factors of carbon emissions in Xi'an are discussed, and the carbon emissions trends of the city under different scenarios are expounded. The relevant conclusions of this study are as follows.

• (1)According to the calculation results of the carbon emissions, the city's carbon emissions have generally been on the rise during the past 20 years, and carbon emissions have grown significantly. In recent years, due to the impact of relevant energy conservation and emission reduction policies, the growth rate of carbon emissions has shown negative growth, and the city's carbon emissions have shown a slight downwards trend.
• (2)By analysing the decoupling relationship between the city's carbon emissions and economic growth, we found that the city's decoupling states are mainly four decoupling states: expansive negative decoupling, weak decoupling, expansive connection, and strong decoupling. After 2010, the decoupling state is basically strong decoupling, which is an ideal decoupling state, indicating that the realization of economic growth is no longer at the expense of the environment, and the city's low-carbon development has achieved initial results.
• (3)Through the LMDI decomposition model, the influencing factors of carbon emissions are decomposed. Two factors, population size and per capita GDP, have a positive effect on carbon emissions and promote the increase in carbon emissions. Energy intensity can effectively restrain the increase in regional carbon emissions. The carbon intensity of energy consumption was shown to promote the growth of carbon emissions in the early years. With the improvement of the energy structure and industrial structure, the carbon intensity of energy consumption gradually shows a reverse effect on carbon emissions.
• (4)According to the carbon peak forecast under different scenarios, the city is expected to achieve a carbon emission peak approximately in 2025 in the baseline scenario, which is earlier than the national 2030 carbon peak target plan. Under the low-carbon scenario, the carbon peak would have been achieved in approximately 2016, while under the high-carbon scenario, the city's carbon peak year is estimated to be approximately in 2035.

Declarations

Author contribution statement

Zhang Yao: Conceived and designed the experiments.

Zhang Yuxin: Performed the experiments; Wrote the paper.

Zhang Yongjian: Analyzed and interpreted the data.

Gong Chao; Kong Yaqian: Contributed reagents, materials, analysis tools or data.

Funding statement

This work was supported by Youth Science Foundation Project of the National Natural Science Foundation of China [51806133].

Data availability statement

Data included in article/supp. material/referenced in article.

Declaration of interest’s statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.