A Study on the Impact of Urbanization Level on Carbon Emissions Based on the STIRPAT Model: A Case Study of Five Major Urban Agglomerations in Northern China

Abstract

With the acceleration of urbanization, city clusters of different levels have emerged. They play a significant role in the process of carbon emissions. Based on the improved STIRPAT model and using the five most important city clusters in northern China as the research object, this study investigates carbon emissions (carbon emissions refer to CO₂ emissions) from 2005 to 2021 and the urbanization factors that affect carbon emissions. The study found that energy urbanization has the most significant positive impact on carbon emissions from the overall perspective, while population urbanization has the most substantial negative effect. The total emissions of the Beijing–Tianjin–Hebei urban cluster far exceeding those of the other four clusters. The factor with the least impact on carbon emissions in the Beijing–Tianjin–Hebei city cluster, the Central Plains city cluster, and the Guanzhong Plain city cluster is energy urbanization. The impact of technological urbanization factors can be positive or negative, the role of spatial urbanization factors is not significant. The northern city clusters should start with transforming energy structure, promote industrial upgrading through innovation, encourage green production and consumption, increase the use of efficient scientific and technological products and services, and promote the development of new urbanization.

1. Introduction

Global warming has become the focus of global attention, and controlling greenhouse gas emissions has become an important task. With the development of global economy and the growth of population, the problem of carbon emission is becoming increasingly prominent. According to the IPCC Fifth Assessment Report, global average surface warming depends on cumulative carbon dioxide emissions [1]. Human activities lead to an increase in carbon dioxide emissions and the occurrence of more extreme climate events, such as heat wave, heavy precipitation, drought, and sea level rise [2]. In the report on the work of the Chinese Government in 2024, the important task of actively and steadily promoting the goal of carbon peak and carbon neutrality is put forward—under the goal of “dual carbon”, China has actively promoted the low-carbon energy transformation to reduce carbon emissions and meet the challenges of climate change [3].

Urban energy consumption is the main source of greenhouse gas emissions, and more than 70% of carbon dioxide emissions come from cities [4], and the development of urbanization is an important driving force for this phenomenon. As highly populated and economically dense regions, cities are not only the main achievers of high-quality economic and social development but also the main places coping with external change risks [5]. Therefore, in the process of urbanization, it is crucial to control the carbon emissions of cities.

Urbanization is usually accompanied by population growth, and population agglomeration to cities also promotes the urbanization process [6]. Increasing urbanization and population growth interact and together have important impacts on the environment [7], which will reduce the quality and benefits of urbanization and hinder the sustainable development of the city when the ecological environment carrying capacity is exceeded [8]. Urban agglomeration reflects the new characteristics of interdependence and win–win cooperation between cities. Close economic ties and resource dependence make the carbon emission problem more prominent. Yao Shimou defines urban agglomeration as: There are a considerable number of cities with different natures, types, and levels of scale within a specific geographical scope, relying on certain natural environmental conditions, with one or two super-large or megacities as the core of regional economy, and with the help of modern means of transportation and the accessibility of comprehensive transportation networks, as well as highly developed information networks, the internal relations between urban individuals occur and develop. Together, it forms a relatively complete urban “complex” [9]. Therefore, to promote the urbanization process on the basis of protecting the ecological environment, we must pay more attention to the coordinated development of urban agglomerations.

Urban agglomeration is considered to be the core space for future regional development. At present, the Yangtze River Delta, Pearl River Delta, and Beijing–Tianjin–Hebei city clusters have become the core engines of China’s rapid economic development, and their total GDP has accounted for more than 40% [10]. In the context of accelerating economic globalization and the rapid development of urbanization process, only the industrial agglomeration ability and scale advantages of urban agglomerations can participate in the competition and cooperation of cities on a global scale. By forming a strong economic community and a community of common destiny, city clusters can better cope with the challenges brought by globalization [11]. Due to the difficulty in the transformation of traditional energy structure and industrial structure, the carbon emissions of China’s urban agglomerations account for more than 75% of the country’s total emissions, making them the main source of carbon emissions and an important region to undertake the responsibility of carbon neutrality [12]. Therefore, China needs to deeply understand the specific sources and influencing factors of carbon emissions in the process of urbanization, so as to formulate targeted emission reduction measures and sustainable development strategies to cope with the environmental challenges brought by urbanization.

Currently, society pays particular attention to the development of urban agglomerations, but few people pay attention to the differences between urban agglomerations. In China, 19 city clusters are widely distributed, but the Beijing–Tianjin–Hebei city cluster, Shandong peninsula city cluster, Central and southern Liaoning city cluster, Guanzhong plain city cluster, and the Central Plains city cluster are relatively concentrated in their locations and have similar industrial structures. This provides a perspective for this research: Where are the similarities and differences in carbon emissions and influencing factors among these five urban clusters located in northern China? Obviously, studying the relationship between urban agglomerations and carbon emissions can not only guide and manage urbanization processes, achieve sustainable development but can also provide important references for global response to climate change.

2. Literature Review

With the rapid development of global urbanization and the increasingly serious problem of global warming caused by carbon emission, it is more urgent to study the relationship between urbanization and carbon emission. Empirical studies on the relationship between urbanization and carbon emissions can be traced back to Parikh et al. [13]. At present, researchers have widely used various models to explore the relationship between urbanization and carbon emissions. These models cover urban population growth, urbanization degree, city size, energy structure, transportation mode, industrial structure, and many other aspects. Therefore, in-depth research on the relationship between urbanization level and carbon emissions of urban agglomerations is helpful to promote the sustainable development of urban agglomerations, enhance the overall competitiveness of the region, and promote the sustainable development of cities.

Domestic and foreign scholars have conducted in-depth research on the relationship between carbon emission and urbanization, adopted a variety of models, and have fully studied the theory of urbanization. Lin Boqiang et al. [14] used the co-integration method to study the long-term equilibrium relationship between CO₂ emissions and major variables. Cheng Yeqing et al. [15] used spatial autocorrelation analysis method and spatial panel measurement model to explore the spatio-temporal pattern and main influencing factors of carbon emission intensity at the provincial scale in China. Jahanger [16] used STIRPAT model framework and quantile regression method to analyze the results to study the emissions of BRICS countries. Smarnika Ghosh [17] used the stochastic effects STIRPAT model framework of population, affluence and technology regression to analyze the promotion of environmental sustainability in developing countries. Qu Tingting et al. [18] used the bidirectional fixed effect model, instrumental variable model, spatial Durbin model, and intermediary effect model to analyze the empirical results and concluded that there is an “inverted U-shaped” relationship between urbanization and carbon emissions.

In the existing literature, scholars have widely used a variety of models to study the relationship between urbanization and carbon emissions, among which STIRPAT model is generally regarded as one of the most widely used and the best effect models. The STIRPAT model plays a significant role in revealing the influencing factors and changing rules of carbon emissions and is of great value for further understanding of carbon emissions. First, the STIRPAT model is clearly structured to quantify the contribution of population, affluence and technology to environmental impact, helping to uncover the links between these factors. Secondly, based on statistical regression analysis, the STIRPAT model can quantify the influencing factors through data validation and parameter estimation, which improves the reliability of the research results. In addition, the model has a wide range of applications, can be applied to the study of different regions and environmental problems, and has strong universality and practicability.

At the national level, urbanization and carbon emissions require all countries to work together to cope with climate change and environmental pollution. For example, Zhang Zhonghua et al. [19] used STIRPAT assessment model to analyze the data of economy, population, electricity and energy consumption and carbon emissions of the five BRICS countries and studied and demonstrated the impact of factors affecting carbon emissions of each member country on carbon emissions. The influence factors are analyzed and compared. Zhou Wenwen [20] used the STIRPAT model as an analytical framework to study the direct and indirect effects of attention and content of carbon-related news on carbon emission and carbon transfer. Ghazala Aziz [21] used the STIRPAT model to explore in depth the interactions between environmental technologies, environmental taxes, and renewable energy consumption, highlighting their combined impact on the ecological footprint in the East Asian context. Xu Fangjin [22] used STIRPAT model to investigate whether there is an environmental Kuznets curve (EKC) relationship between urbanization and pollutant discharge (including industrial wastewater and industrial SO₂).

At the provincial level, the research fully understood the urbanization level and development trend of each province. Based on the extended STIRPAT model, He Letian et al. [23] identified six independent variables, including square per capita GDP, population size, and total industrial output value, and studied the effects of the six variables on the carbon emissions of the industrial sector in Heilongjiang Province. Xu Penghui et al. [24] analyzed the influencing factors of carbon emissions from industrial land in Liaoning Province based on the extended STIRPAT model. Wang Qi [25] used STIRPAT model to predict the peak carbon emission of Hebei Province, which has important practical significance and practical value.

At the city level, cities are the most concentrated areas of carbon emission. Domestic scholars have specifically studied the spatio-temporal evolution of carbon emission intensity and the role of influencing factors at the city level. For example, Zhao Fan et al. [26] used the STIRPAT model to analyze the energy consumption of the Yangtze River Economic Belt and study the impact of strong carbon emissions. Guo Yungong et al. [27] used Ling regression function to fit the STIRPAT model, studied the relationship between economic development and carbon emission footprint in depth, and provided scientific basis for the development of adaptive management strategies in the future. Sun Jie et al. [28] selected variables from seven aspects: population size, wealth level, urbanization level, technology level, industrial structure, international trade level and scientific and technological innovation ability, and established STIRPAT model to study the main factors affecting carbon emissions in Nanjing.

From the existing literature, most of the studies on regional differences of carbon emissions in China focus on a specific region, province, and city, while the research content from the perspective of urban agglomerations is less. Moreover, China’s urbanization level research involves geography, economics, demography, ecology and other fields, because of the diversity of research perspectives, the measurement standards of urbanization level are different, including population urbanization, economic urbanization and other different indicators. Urbanization is of great significance to urban construction. Through the change in urbanization level and the rationalization of urban scale and spatial structure, low-carbon cities can be built, which can not only have a positive impact on the sustainable development of cities but also reduce urban carbon emissions.

The improvement of urbanization level will not only affect individual cities, but also profoundly affect the sustainable development and future trend of the entire urban agglomeration. Urban agglomerations can be divided according to different types. Taking the northern urban agglomerations as an example, the optimized and upgraded urban agglomerations include the Beijing–Tianjin–Hebei urban agglomerations, etc. The development focuses on large-scale urban agglomerations—such as Shandong Peninsula urban agglomerations, Central Plains urban agglomerations, and Guanzhong Plain urban agglomerations—and the development of urban agglomeration—such as the central and southern Liaoning urban agglomeration. Each type of urban agglomeration has its specific geographical, economic, and social conditions, and corresponding policies and measures need to be taken according to local conditions to achieve sound development and prosperity of urban agglomeration.

At present, the world is experiencing the third energy transition for sustainable development [29]. In the context of global energy transition and rapid urbanization, northern urban agglomerations are dominated by heavy industry with large energy consumption. Moreover, their production structure is relatively undiversified, and they mainly rely on traditional energy resources (such as coal), thus playing an important role in political and economic development. Urban agglomerations in northern China generally face severe environmental challenges, including air pollution, water use and land use. Therefore, the study of northern urban agglomerations can provide a key reference for the national carbon emission reduction. Most of the cities in northern urban agglomerations are resource-exhausted cities, which are faced with prominent environmental problems and serious pollution problems. By studying the northern urban agglomeration, we can promote the construction of ecological civilization and promote the formulation of scientific carbon emission reduction policies.

The research focus on carbon emissions in these five urban clusters is different. When most scholars study several important urban agglomerations in China, the Beijing–Tianjin–Hebei city cluster is usually included [30,31,32,33]. Due to the fact that the Beijing–Tianjin–Hebei city cluster includes China’s megacities, such as Beijing and Tianjin, there is considerable research on carbon emissions in this city cluster [34,35,36,37]. In contrast, there are many fewer carbon emissions studies on the Shandong Peninsula city cluster [38], Guanzhong Plain city cluster [39], Central Plains city cluster [40,41] (some studies also include these three city clusters in the Yellow River Basin [42,43]), and central and southern Liaoning city cluster [44,45].

In terms of selecting urbanization factors, urbanization has triggered a trend of population concentration in cities, which is crucial to driving the development and growth of urban economies. At the same time, in order to build a livable ecological environment, we need to strengthen ecological environmental protection, promote the construction of low-carbon cities, and reduce urban pollution emissions. In order to achieve these goals, science and technology play an important role. In addition, energy urbanization is one of the key areas. Although the energy structure is developing in the direction of cleanliness, the technical support for new energy consumption and clean and efficient use of coal is not sufficient, which restricts the promotion of energy strategy [46]. Therefore, we need in-depth analysis of energy urbanization in order to promote energy transition and sustainable energy development. In summary, population urbanization, economic urbanization, spatial urbanization, ecological urbanization, science and technology urbanization, and energy urbanization are important aspects of China’s urbanization process. By strengthening scientific and technological innovation, strengthening policies, optimizing land use planning and promoting low-carbon development, we can achieve sustainable urbanization and provide people with a better living environment and development opportunities.

Based on the above viewpoints, in order to further study and understand the development mode and characteristics of the relationship between urban agglomerations and carbon emissions, this study selects the Beijing–Tianjin–Hebei urban agglomerations, the Shandong Peninsula urban agglomerations, the Central and southern Liaoning urban agglomerations, the Guanzhong Plain urban agglomerations, and the Central Plains urban agglomerations among the northern urban agglomerations with potential and importance as research objects and establishes the STIRPAT model. Compared to the existing literature, the model used in this research is an extended STIRPAT model. In terms of model setting, this study is similar to existing literature in the selection of population and economic factors but has significantly expanded on technical factors. This study comprehensively referred to the development positioning of urban agglomerations by the Chinese government, the actual situation of each urban agglomeration, the selection of urbanization factors, and the availability of data sources. Finally, from the perspectives of population urbanization, economic urbanization, technological urbanization, energy urbanization, ecological urbanization, and spatial urbanization, this study analyzed the influencing factors of carbon emissions in the five major urban agglomerations in northern China and explored the role of urbanization level on carbon emissions. At the same time, a horizontal comparison was made on their differences and characteristics in carbon emissions. This is an improvement on the methodology of this study compared to the existing literature.

Through this study, we can better understand the differences in carbon emissions between different urban agglomerations, reveal the spatial and temporal characteristics of carbon emissions in urban agglomerations, provide a scientific basis for urban planning and policy formulation, and provide a scientific basis for the formulation of targeted emission reduction measures to promote the green and low-carbon development of urban agglomerations. These urban agglomerations have significant influence on regional development and cooperation. By analyzing and comparing them, we can better grasp the law and trend of the development of urban agglomerations and provide scientific reference and guidance for the sustainable development of urban agglomerations.

Compared with the existing literature, this paper makes two contributions:

• To analyze the relationship between carbon emission and urbanization level from the perspective of urban agglomeration so as to enrich the research perspective of the relationship between urbanization and carbon emission and provide a new perspective and method for further discussion in related fields.
• The variables selected by the existing STIRPAT model are supplemented, and analysis and comparison are conducted from the perspective of six urbanization types, aiming to further explore the potential factors affecting carbon emissions in the process of urbanization so as to provide a more comprehensive and in-depth understanding of the complex relationship between urbanization level and carbon emissions of urban agglomerations.

3. Model Analysis

The STIRPAT model is an extended form of the IPAT model. The IPAT model was originally proposed by demographers Ehrlich and Holder in 1971 as a model of the impact of human drivers on the environment from the perspective of technology improvement and environmental deterioration, namely IPAT model. In 1998, Dietz and Rosa revised it into a regression model of random effects on the environmental impacts of population, affluence and technology, namely STIRPAT model [47]. As a model to analyze the causal relationship between human social activities and environment, STIRPAT model has been widely used to study environmental pollution and social resource use.

3.1. STIRPAT Model Improvement

Given the flexibility of the STIRPAT model, it allows the introduction of additional control variables for in-depth analysis of their impact on the environment. However, to ensure that these new variables must be conceptually consistent with the multiplicative form specified by the model. Combined with the research theme of this paper, this study introduces several variables into the STIRPAT model.

The basic model expression is:

$$I=a{P}^b{A}^c{T}^{d}e$$

(1)

In the formula, I represents the state of environmental impact, P represents the number of people, A represents wealth, and T represents the technology level. a is the total coefficient, b, c, and d represent the exponents of the above three, respectively, and e is the error of the entire model.

In order to introduce more independent variables into the real situation and eliminate the multicollinearity among explanatory variables, logarithmic analysis can be carried out on Formula (1) to analyze the impact of each variable on environmental conditions in a more comprehensive way; that is, Formula (2) is as follows:

$$\ln I=\ln a+b\ln P+c\ln A+d\ln T+\ln e$$

(2)

With lnI as the dependent variable; lnP, lnA, and lnT as the independent variables, lna as the constant term; and lne as the error term, this model provides a concise causal analysis framework to decompose the impact of human and social factors on the environment and is widely used in environmental pollution, social resources and other aspects [48].

This study divides urbanization into population urbanization, economic urbanization, science and technology urbanization, space urbanization, ecological urbanization, and energy urbanization and expresses it with different variables; detailed instructions can be found in

Table 1. Description of variables.

Urbanization Type	Definition	Unit
Population urbanization	Total population at the end of the year	10,000 people
Economic urbanization	per capita gross regional product (GDP)	CNY
Sci-tech urbanization	The number of green patents for science and technology urbanization (OGP)
Energy urbanization	Dust emissions (DE)	Ton
	Total society electricity consumption (TEC)	billion kWh
Spatial urbanization	The number of road lights for spatial urbanization (NEL) per 10,000 people
Ecological urbanization	Green area	Hectares

. This study aims to promote the sustainable development of the city; improve the quality of life of urban residents; and realize the coordinated development of the economy, society, and environment of the city through the research and analysis of different types of urbanization.

In order to improve the intuitiveness of the model, this study adjusts some variable symbols in the original model. Specifically, this paper treats lna as a constant term. Population urbanization refers to the process of population migration from rural areas to urban areas and gathering in the city, which also promotes the expansion and development of the city. Population urbanization will affect carbon emissions in various ways, including population size, urbanization process, consumption patterns, population migration, and so on. The increase in population directly leads to an increase in energy consumption and carbon emissions in production and daily life. The consumption effect caused by urbanization dominates, and urbanization brings about the construction and use of high energy consuming infrastructure, increasing carbon emissions [49,50]. With the development of the economy entering a moderately prosperous society, the consumption patterns of residents have undergone earth shaking changes, and population migration will also affect carbon emissions. For example, migration from rural to urban areas may lead to an increase in energy consumption, as urban life typically requires more energy support. These are all caused by population changes and increased urbanization rates. Therefore, in this study, the total population of each province and city at the end of the year is selected as the proxy variable of population urbanization and is represented by P in the formula.

Economic urbanization means that with the advancement of the urbanization process, the urban economic structure has also undergone obvious changes when the population gathers in the city [51]. This study chooses GDP per capita as the proxy variable of affluence level A, expressed by GDP.

Sci-tech urbanization refers to a model that takes sci-tech innovation and development as the guide, pays attention to the integration and promotion of sci-tech factors in the process of urbanization, and promotes the sustainable development of the city by building an intelligent, green, and informationized urban ecosystem. In practical operation, the government and enterprises improve carbon reduction performance by implementing green innovation strategies, cultivating green patents, and so on. In addition, after green patents are transformed into technological innovation, governments and enterprises can reduce carbon emissions and improve carbon reduction performance through technologies such as end of pipe treatment; clean production; and carbon capture, utilization, and storage (CCUS) [52]. Therefore, in this study, the number of obtained green patents is selected as the proxy variable of technology factor T, and the rest are the extended variables of IPAT model.

Energy urbanization refers to a mode that focuses on the development, utilization and management of energy resources in the process of urbanization in order to achieve sustainable development of urban energy supply and optimize energy structure. In most cities in China, power generation is still dominated by thermal power generation, and dust emissions and CO₂ emissions share the same origin, both of which are generated by the combustion of fossil fuels. Dust and electricity, as indicators that can directly measure energy use, are convenient for statistics and analysis. Therefore, from the perspective of clean energy and industrial pollution, this study selected dust emissions and total electricity consumption as proxy variables of energy urbanization, expressed by DE and TEC, respectively [53].

Spatial urbanization refers to a mode that pays attention to spatial planning and utilization efficiency in urban development to realize the optimization of urban spatial structure, the diversification of functions and the sustainable development of urban form. In urban spatial research, nighttime lighting data is often used to reflect the carbon emissions level of a region. From a practical perspective, the indicator of the number of urban road lighting lights per 10,000 people reflects the spatial popularity of government social security and living facilities and can also indirectly reflect nighttime lighting data. Therefore, in this study, the number of road lighting lights per 10,000 people was selected and represented by NRL in the formula [54].

Ecological urbanization refers to the development mode that pays attention to ecological environment protection and sustainable development in urban planning and construction, so as to realize the harmonious coexistence between the city and the natural environment. Among the elements of ecological urbanization, increasing urban carbon sinks and reducing urban carbon emissions are mainly achieved through increasing urban green spaces. The most direct benefit of urban green space is carbon sequestration and oxygen release, effectively increasing the total amount of urban carbon sink. Vegetation, wetlands, and microorganisms in green spaces absorb carbon dioxide through photosynthesis, convert it into organic carbon, and store it in plant bodies and soil. The green space system is the only natural carbon sink in urban areas; secondly, green space systems indirectly reduce carbon emissions by alleviating urban heat island effects and reducing urban energy consumption [55,56]. Therefore, in this study, green area is selected and represented by GA in the formula.

Based on the above explanatory variables, Formula (2) is extended, and Formula (3) can be obtained:

$$\ln E_{C{O}_2}=\ln a+b\mathrm{lnP}+\mathrm{clnGDP}+\mathrm{dlnOGP}+\mathrm{elnDE}+\mathrm{flnTEC}+\mathrm{hlnNRL}+\mathrm{glnGA}$$

(3)

It should be noted that the STIRPAT model used in this research has certain limitations. Because this study used time series data from 2005 to 2021, stationarity testing and descriptive statistical analysis are required before regression. In addition, the seven indicators selected in this research are all from statistical yearbooks and actual data published by the government, and the magnitude of each indicator varies greatly. After taking the logarithm, attention should be paid to the issue of multicollinearity between variables. Therefore, after obtaining the regression parameters using the ordinary least squares method, it is necessary to conduct a multicollinearity test, and then perform a fitting effect test. Based on this approach, this research conducts a comprehensive and individual study of five city clusters in the Empirical Analysis and Discussion Section.

3.2. Data Source

The variables selected in this study—such as the total population at the end of the year, the per capita regional product, the number of green patents obtained, dust emissions, the electricity consumption of the whole society, the number of road lights per 10,000 people, and the area of green space—are derived from the China Urban Statistical Yearbook and the China Urban Construction Statistical Yearbook. In this paper, data from 2005 to 2021 are selected as statistical samples. The total carbon emission data came from the relevant studies of Cong et al. [57], which were calculated and sorted out by the authors. Carbon emissions are the total amount of carbon dioxide emitted directly and indirectly by a city.

The geographical location of urban agglomeration selected in this paper is shown in Figure 1.

This paper aims to study the impact of the overall urbanization level of the five major urban agglomerations in northern China on carbon emissions and make a brief horizontal comparison among urban agglomerations to understand the differences and characteristics of each urban agglomerations in terms of carbon emissions.

4. Empirical Analysis and Discussion

4.1. Overall Analysis of Carbon Emissions in Urban Agglomeration

With the acceleration of the global urbanization process, the development and construction of urban agglomerations have attracted extensive societal attention. In the process of urban agglomeration urbanization, the linear growth trend of carbon emissions has aroused people’s attention to the goal of carbon peak. As an important goal in mitigating climate change and promote sustainable urban development, carbon peaking is of great significance to the five major urban agglomerations in northern China. Therefore, it is particularly urgent to deeply study the factors affecting carbon emissions, especially in the context of rapid urbanization, and put forward effective carbon emission reduction measures and policies. According to the data in Figure 2, from 2005 to 2021, carbon emissions in 2005 were 238.57 million tons, and in 2021, carbon emissions increased to 387.66 million tons, an increase of about 0.6 times. Compared with Wang’s research [58], the carbon emissions of these five urban agglomerations can account for 45% to 55% of the entire northern region of China, and their proportion has increased in recent years. The total carbon emissions of the five major urban agglomerations in northern China showed a linear growth trend, which highlights the challenge and importance of achieving the target of carbon peak.

This growth trend shows that urbanization, while bringing economic prosperity and social development, is also accompanied by an intensification of environmental problems, especially the continuous increase in carbon emissions. Therefore, in order to achieve the target of carbon peak, it is necessary to further explore the causes and mechanisms leading to an increase in carbon emissions.

Therefore, through the detailed division of different urbanization levels of urban agglomerations, this study mainly studies five major urban agglomerations in northern China, and establishes a regression model, focusing on the impact of different urbanization types on carbon emissions, so as to better understand the environmental impact factors behind urban development, provide a scientific basis for formulating corresponding policies and measures, and promote low-carbon-transformation and sustainable development of urban agglomerations.

Before the regression analysis, the stationarity test and descriptive statistical analysis of the data should be carried out. After the least square estimation results are obtained, the multicollinearity test and fitting effect test are carried out to ensure that the obtained results are reliable and effective. Then, the regression equation is established and the regression equation of the northern urban agglomeration is obtained. After the overall analysis of the relationship between urbanization and carbon emissions, the regression equation of each urban agglomeration is further established and a more detailed analysis is carried out.

4.2. Model Checking

4.2.1. Stationary Analysis

Due to the use of 2005–2021 data for fitting, it is necessary to verify the stationarity of time series to avoid the phenomenon of pseudo-regression. Therefore, the ADF unit root test is carried out for each panel data series. The ADF test results of the data after first-order difference show that p = 0.000 < 0.01, the null hypothesis is rejected with more than 99% certainty, and the sequence is stable.

4.2.2. Descriptive Statistics

In view of the data sources and statistical caliber, this paper selects 72 prefecture-level cities in five major urban agglomerations in the north from 2005 to 2021. The descriptive statistics of each variable are shown in

Table 2. Descriptive statistics.

	N	Minimum	Maximum	Mean Value	Standard Deviation
lnP	1190	0.41	8.18	5.68	1.14
lnGDP	1190	1.10	12.01	9.44	1.62
lnOGP	1190	0.60	10.29	4.43	1.81
lnDE	1190	2.64	15.46	8.90	1.85
lnTEC	1190	0.26	7.12	4.11	1.31
lnNRL	1190	0.69	13.03	9.18	2.47
lnGA	1190	0.00	12.05	7.52	1.56

:

4.3. Regression Model

The least square method is a commonly used statistical method that can perform regression analysis on data to obtain estimates of relevant parameters. Through the application of the least square method, we can better understand the relationship between the quality and benefit of urbanization and provide an important reference for promoting the sustainable development of cities. The specific results are shown in

Table 3. The least square method estimates the result.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	471.224	131.079		3.595	0.001
lnP	−61.399	17.244	−0.248	−3.561	0.001	0.374	2.677
lnGDP	−0.161	0.031	−0.360	−5.143	0.000	0.370	2.702
lnOGP	0.202	0.041	0.368	4.873	0.000	0.320	3.128
lnDE	0.319	0.028	0.600	11.582	0.000	0.678	1.474
lnTEC	0.731	0.080	0.919	9.096	0.000	0.178	5.617
lnNRL	−0.019	0.017	−0.058	−1.109	0.271	0.659	1.518
lnGA	−0.217	0.044	−0.368	−4.993	0.000	0.335	2.986

. In addition, in order to ensure the correctness of the formula, the multicollinearity test and fitting effect test should be carried out first.

(1)

Multicollinearity test

As can be seen from the results of Table 3, the variance inflation factor (VIF) of each explanatory variable is less than 10, indicating that there is no strong multicollinearity among explanatory variables, so the results are relatively reliable. The regression coefficient estimates for each explanatory variable in the table can be used to determine the degree of primary and secondary influence of each influencing factor.

(2)

Test of fitting effect

The data show that R² is the decision coefficient of fitting linear regression, which is 0.860, indicating that the change of lnI can explain about 86.0% of the variables. The adjusted R² is 0.847. In this model, the difference between R² and adjusted R² is small, indicating that the standard estimation error is not overfitted. The error value between the adjusted R² and the standard estimate. The larger the R² value, the higher the goodness of fit between the model and the data.

Through the double verification of the multicollinearity test and the fitting effect test, it is concluded that the estimation results of the least square method adopted in this study are reliable and effective, which provides strong support for the in-depth study of urbanization development and urban agglomeration construction.

According to the results, the regression linear equation is established, and the results are as follows:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=471.224-61.399\mathrm{lnP}-0.106\mathrm{lnGDP}+0.202\mathrm{lnOGP}+0.319\mathrm{lnDE}+0.731\mathrm{lnTEC}-0.019\mathrm{lnNRL}-0.217\mathrm{lnGA}$$

(4)

4.4. Regression Result Discussion

In this part, the improved STIRPAT model is used to conduct a regression analysis of the relationship between urbanization level and carbon emissions in five major urban agglomerations in northern China. The results show that the urbanization level affecting carbon emissions is ranked from large to small (according to the absolute value of the coefficient) as follows: energy urbanization (electricity consumption of the whole society, dust emissions), ecological urbanization (green space), scientific and technological urbanization (number of green patents), economic urbanization (gross regional product per capita), and population urbanization (total population at the end of the year). According to the analysis of the results, some conclusions can be obtained.

Energy urbanization has the largest positive impact on carbon emissions, and energy urbanization is subdivided into clean energy and industrial pollution. Clean energy is directly related to the optimization of energy structure and environmentally friendly development. The use of clean energy can reduce the dependence on fossil energy and thus reduce carbon emissions. Industrial pollution is related to the treatment of waste gas, wastewater, and solid waste generated in the production process, which directly involves the pollution of the atmosphere and water and has a negative impact on the environment. The subdivision of energy urbanization into clean energy and industrial pollution helps to more accurately analyze its impact on carbon emissions and promote the development of energy urbanization in a low-carbon and clean direction to achieve the sustainable development goals. Therefore, the variables of clean energy and industrial pollution are expressed in terms of total social electricity consumption and dust emissions, respectively.

Electricity consumption in the whole society is the most significant factor, and every 1% increase in electricity consumption in the whole society will lead to a 0.731% increase in carbon emissions. The power system is an energy-intensive industry, relying mainly on fossil fuels such as coal and natural gas, the combustion of which releases large amounts of carbon dioxide. Dust emissions have a positive impact on carbon emissions, because industrial production and other activities usually release a large amount of carbon emissions, but also accompanied by dust and particulate matter emissions. Therefore, an increase in dust emissions will lead to an increase in carbon emissions, thus showing a positive impact. Therefore, energy urbanization reduces carbon emissions through clean energy use and energy efficiency improvement, but requires large-scale development of renewable energy sources and optimization of energy system management.

The impact of technological urbanization (the number of green patents) on carbon emissions shows a positive trend, that is, the higher the level of science and technology, the more carbon emissions will increase. In the regression model, the coefficient of the variable green patent number is 0.202, indicating that in this regression model, the increase in the number of green patents will affect the increase in carbon emissions. Technological urbanization represents the progress of environmental technology and innovation, but its practical application is affected by many factors, including technology cost and other aspects.

The coefficient of eco-urbanization (green area) is −0.217—that is, the increase in eco-urbanization level contributes to a reduction in carbon emissions. At the same time, the variable of green area also has a negative impact on carbon emissions, indicating that large green area can improve urban air quality and mitigate the urban heat island effect, thus playing an important role in reducing carbon emissions.

Population urbanization (total population at the end of the year) has a negative impact on carbon emission in the model. With the intensification of population urbanization, urban development becomes more concentrated, resource utilization becomes more efficient, and more environmental protection measures and technologies may be adopted, thus reducing carbon emission. In addition, population urbanization may also mean that more people are gathered together, and that it is easier to implement policy measures such as public transportation, energy conservation, and emission reduction, which is conducive to reducing carbon emissions.

The coefficient of economic urbanization (gross regional product per capita) is −0.161, indicating that the improvement of economic urbanization level can have a certain inhibition effect on carbon emissions. Since an improvement in economic development level is often accompanied by people’s attention to environmental protection and the application of environmental protection technology, a higher per capita GDP can promote the protection of the environment and reduce carbon emissions.

The coefficient of spatial urbanization (the number of road lights) is 0.019, and the impact on carbon emissions is often negative. Spatial urbanization is one of the core contents of urban planning. Through rational utilization of urban land resources, urban sustainable development can be realized. Effective urban planning can promote the efficient use of land resources, reduce the damage to the natural environment, and thus reduce carbon emissions. Scientific and rational spatial urbanization planning is the key to the construction of low-carbon and sustainable cities and provides important support for the realization of urban sustainable development goals.

4.5. Regional Differences of Carbon Emission and Carbon Emission Factors in Urban Agglomerations

In order to further analyze the carbon emissions of the urban agglomerations involved in this study and better understand the evolution trend of carbon emissions in the urbanization process of the five major northern urban agglomerations, this study will conduct a detailed analysis of the influencing factors of carbon emissions during the study period and clearly show the change trend of carbon emissions in the form of visual charts so as to provide intuitive and powerful support for research. As shown in Figure 3, the total carbon emissions of urban agglomerations are calculated according to the sum, and it is obvious that the carbon emissions of the five urban agglomerations continued to rise from 2005 to 2021.

By comparing the total carbon emissions of the five urban agglomerations from 2005 to 2021, it is found that the total carbon emissions of the five urban agglomerations from 2005 to 2021 are large and show an upward trend. As can be seen from the figure, in recent years, the total carbon emission of the Beijing–Tianjin–Hebei urban agglomeration has far exceeded that of the other four northern urban agglomerations, always ranking first and remaining high, or will show an upward trend. The carbon emission trends of the Shandong Peninsula city cluster, Guanzhong plain city cluster, central and southern Liaoning city cluster, and Central Plains city cluster are very similar. The carbon emission of Shandong Peninsula urban agglomeration is always relatively low, and the growth is slow and stable. The Central Plains city group is located between the Beijing–Tianjin–Hebei City group and the Guanzhong City group, with a steady upward trend as a whole. In recent years, the carbon emissions of Shandong Peninsula city cluster, south central Liaoning city cluster and Guanzhong Plain city cluster are close to each other in terms of total amount and growth trend. In summary, the total amount of carbon emissions may continue to increase in the future, and it is urgent to find a positive solution to the problem of carbon emissions.

After an in-depth study of the carbon emissions of urban agglomerations, this study will analyze the characteristics and trends of carbon emissions of the Beijing–Tianjin–Hebei city agglomeration, Shandong Peninsula city agglomeration, Central Plains city agglomeration, central and southern Liaoning City agglomeration, and Guanzhong Plain city agglomeration one by one so as to understand the differences and influencing factors among different urban agglomerations more comprehensively.

4.5.1. Beijing–Tianjin–Hebei City Cluster

The cities involved in this study include Beijing, Tianjin, Anyang, Baoding, Tangshan, Langfang, Shijiazhuang, Qinhuangdao, Zhangjiakou, Chengde, Cangzhou, Hengshui, Xingtai, and Handan. Before discussing the relationship between carbon emissions and urbanization in the Beijing–Tianjin–Hebei city cluster, this study first conducted an overall analysis of carbon emissions in the region and found that it showed a trend of continuous growth. As shown in Figure 4, the carbon emissions of the Beijing–Tianjin–Hebei city cluster increased from 1113.31 million tons in 2005 to 129.954 million tons in 2010, to 1532.979 million tons in 2015, and to 1829.584 million tons in 2021, showing a significant growth trend. The calculation results of this article are relatively close to the research results of Fu [59]. By 2021, the carbon emissions of Beijing, Tianjin, and Shijiazhuang will show a triangular concentration distribution, forming a three-pillar situation, and the carbon emissions of the cities in the Beijing–Tianjin–Hebei city cluster will also increase to varying degrees.

Due to the multiple validations of the stationarity test, multicollinearity test, and fitting effect test for the Beijing–Tianjin–Hebei city cluster, we can confirm the reliability and validity of the least square estimation results adopted in this study. The fitting results of the Beijing–Tianjin–Hebei city cluster are as shown in

Table 4. Estimation results of least square method for Beijing–Tianjin–Hebei city cluster.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	2.533	1.316		1.926	0.056
lnP	0.057	0.097	0.029	0.589	0.556	0.396	2.526
lnGDP	0.148	0.116	0.101	1.274	0.204	0.157	6.377
lnOGP	0.141	0.046	0.282	3.043	0.003	0.115	8.715
lnDE	−0.130	0.032	−0.143	−4.046	0.000	0.788	1.270
lnTEC	−0.139	0.056	−0.178	−2.454	0.015	0.188	5.311
lnNRL	−0.030	0.019	−0.066	−1.615	0.108	0.596	1.677
lnGA	0.656	0.061	0.721	10.744	0.000	0.219	4.556

.

The regression equation of the Beijing–Tianjin–Hebei city cluster is as follows:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=2.533+0.057\mathrm{lnP}+0.148\mathrm{lnGDP}+0.141\mathrm{lnOGP}-0.130\mathrm{lnDE}-0.139\mathrm{lnTEC}-0.030\mathrm{lnNRL}+0.656\mathrm{lnGA}$$

(5)

Through the significance test, it is found that the three variables, P, GDP, and NRL, have no significant impact on the total carbon emissions of the Beijing–Tianjin–Hebei urban agglomeration. According to the standardized regression coefficient, the factors affecting the total carbon emissions are listed in order from largest to smallest: ecological urbanization (green space), scientific and technological urbanization (green patent number), energy urbanization (dust emission, electricity consumption).

According to the regression equation of the Beijing–Tianjin–Hebei city cluster, the green space area and the number of green patents have a positive impact on the carbon emissions of the Beijing–Tianjin–Hebei city cluster, and the corresponding regression coefficients are 0.656 and 0.141, respectively. In other words, with each 1 unit increase in the above explanatory variables, the corresponding carbon emissions will increase by the above numerical units, respectively. The whole society’s electricity consumption and dust emissions are negative, and the impact on carbon emissions is negatively correlated—that is, carbon emissions will decrease with the increase in the whole society’s electricity consumption. According to the regression results, carbon emissions will decrease by 0.139% for every 1% increase in energy intensity. This conclusion is slightly different from previous studies [60,61,62]—that is, the increase in energy consumption does not affect the increase in CO₂ emissions. This is because the Beijing–Tianjin–Hebei region is located in the core area of China. Due to the extensive use of clean energy, the increase in clean energy even exceeds the consumption of fossil fuels, and the deterioration of air quality and heavy pollution weather have significantly decreased. Therefore, the energy factor shows a slightly negative effect.

From the perspective of ecological urbanization, in this regression model, green space area has a positive impact on carbon emissions, and the relationship between environmental factors may be affected by a variety of factors, leading to different results from conventional expectations. The Beijing–Tianjin–Hebei city cluster has taken a number of measures to address environmental challenges and improve the ecological environment, including increasing green space, promoting sponge city construction and urban parks. However, the expansion of urban construction land in the early stage and the increase in green area may require land leveling and building demolition, which may consume certain energy and produce certain carbon emissions. In addition, the greening project itself also needs to invest certain resources and energy, such as water resources, fertilization, maintenance during plant growth, etc., which will produce a certain amount of energy consumption and carbon emissions.

For the Beijing–Tianjin–Hebei city cluster, due to the spatial pattern of multiple sub-nuclei, carbon emission management cooperation among sub-nuclei should be strengthened to develop unified or coordinated carbon reduction strategies. According to the spatial pattern of the “three cores”, the development of green technology and industry can be strengthened, and the coordination between policy and economy can be encouraged. In addition, although Beijing is more of a political center, it can also use its resource advantages to promote economic and environmental policy coordination with cities such as Tianjin and Tangshan.

4.5.2. Shandong Peninsula City Cluster

In this study, the cities involved in Shandong Peninsula city cluster include Jinan, Qingdao, Yantai, Zibo, Weihai, Weifang, Dongying, and Rizhao. After the previous analysis of the relationship between carbon emissions and urbanization in the Beijing–Tianjin–Hebei city cluster, this study will turn to the discussion of relevant issues in the Shandong Peninsula city cluster. First of all, as shown in Figure 5, the carbon emissions of Shandong urban agglomeration in different time periods were relatively low in 2005. After rapid development and improvement in urbanization level, the carbon emissions of Shandong Peninsula urban agglomeration will increase by 125.2862 million tons by 2021. At first, the polarization of carbon emissions in the Shandong Peninsula urban agglomeration was not obvious, but by 2021, a dual-core distribution with Jinan and Weihai as the center was formed, and the carbon emissions of all cities in the Shandong urban agglomeration showed a significant upward trend, which indicated the demand for sustainable development and highlighted the urgency and importance of reducing carbon emissions. In terms of carbon emission trends and urban spatial distribution, this study aligns with the research findings of References [38,63].

After multiple validations, including a stationarity test, multicollinearity test, and fitting effect test, the reliability and validity of the least square estimation results adopted in this study were obtained. The fitting results of the Shandong Peninsula urban agglomeration are shown in

Table 5. Estimation results of least square method for urban agglomeration of Shandong Peninsula.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	7.343	0.495		14.824	0.000
lnP	−0.015	0.034	−0.050	−0.439	0.661	0.230	4.349
lnGDP	0.105	0.036	0.346	2.950	0.004	0.213	4.701
lnOGP	0.050	0.018	0.422	2.859	0.005	0.134	7.460
lnDE	−0.012	0.011	−0.064	−1.021	0.309	0.735	1.360
lnTEC	0.072	0.019	0.327	3.863	0.000	0.409	2.447
lnNRL	0.009	0.006	0.110	1.667	0.098	0.673	1.486
lnGA	−0.098	0.023	−0.379	−4.334	0.000	0.383	2.612

.

Regression equation of Shandong Peninsula urban agglomeration:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=7.343-0.15\mathrm{lnP}+0.105\mathrm{lnGDP}+0.05\mathrm{lnOGP}-0.012\mathrm{lnDE}+0.072\mathrm{lnTEC}+0.009\mathrm{lnNRL}-0.098\mathrm{lnGA}$$

(6)

It can be seen from the table that P, DE, and NRL are not significant. According to the standardized regression coefficient, the factors affecting the total carbon emissions of the Shandong Peninsula urban agglomeration are: economic urbanization (gross regional product per capita), ecological urbanization (green space area), energy urbanization (electricity consumption of the whole society), and science and technology urbanization (number of green patents) in the order from largest to smallest. Among them, the negative impact is ecological urbanization (green space), that is, carbon emissions will decrease with the increase in green space. Science and technology urbanization (number of green patents), economic urbanization (gross regional product per capita) and energy urbanization (electricity consumption in the whole society) had a positive impact on the carbon emissions of the Shandong Peninsula urban agglomeration, and the corresponding regression coefficients were 0.05, 0.105, and 0.072, respectively. The results of this research on the degree and direction of influence of factors such as energy consumption, economic scale, and ecological green space are similar to the conclusions of reference [64,65,66].

However, this research considered technological factors. From the perspective of science and technology urbanization, the coefficient of green patent number is 0.05, which means that there is a positive relationship between the number of green patents and carbon emissions. With the application and promotion of green technology, a large amount of energy and resources may be required to support its production and use process, resulting in a certain degree of carbon emissions. For example, the development and application of clean energy may require the construction of more energy facilities, including the transition phase of using traditional energy sources, which may lead to higher carbon emissions.

4.5.3. Central and Southern Liaoning City Cluster

The cities involved in this study include Shenyang City, Dalian City, Anshan City, Fushun City, Benxi City, Yingkou City, Liaoyang City, Tieling City, and Panjin City. As shown in Figure 6, in 2005, the city of Shenyang in the central and southern Liaoning city cluster exceeded 30 million tons, showing a unique situation. By 2021, the nine cities in the central and southern Liaoning city cluster will all exceed 30 million tons. Moreover, Shenyang is still far ahead in carbon emission. Compared with 2005, the total carbon emissions of the central and southern Liaoning urban agglomeration increased by 17.0108 million tons, and the carbon emissions showed a significant growth trend.

This study focuses on the impact of urbanization on carbon emissions and their correlation in order to understand the relationship between carbon emissions and urbanization in the Central and Southern Liaoning city clusters. Through the multiple validations of the stationarity test, multicollinearity test, and fitting effect test for the Central and Southern Liaoning city cluster, the reliability and validity of the estimated results of the least square method adopted in this study were confirmed. The fitting results of the Central and Southern Liaoning city cluster are shown in

Table 6. Estimation results of least square method for urban agglomerations in central and southern Liaoning.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	7.599	0.457		16.632	0.000
lnP	−0.110	0.037	−0.353	−2.947	0.004	0.261	3.832
lnGDP	0.039	0.039	0.113	1.001	0.318	0.291	3.428
lnOGP	0.081	0.018	0.689	4.382	0.000	0.144	6.901
lnDE	0.019	0.016	0.087	1.202	0.231	0.712	1.404
lnTEC	−0.027	0.009	−0.220	−2.889	0.004	0.329	3.035
lnNRL	0.014	0.006	0.167	2.182	0.031	0.664	1.507
lnGA	0.022	0.021	0.107	1.031	0.304	0.308	3.243

.

Regression equation of urban agglomeration in central and southern Liaoning:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=7.599-0.110\mathrm{lnP}+0.039\mathrm{lnGDP}+0.081\mathrm{lnOGP}-0.027\mathrm{lnTEC}+0.014\mathrm{lnNRL}+0.022\mathrm{lnGA}$$

(7)

Through the significance test, it is found that GDP, DE, and GA have no significant impact on the total carbon emissions of the central and southern Liaoning urban agglomeration. According to the standardized regression coefficient, the factors affecting the total carbon emissions are listed in order from largest to smallest: science and technology urbanization (number of green patents), population urbanization (total population at the end of the year), energy urbanization (total electricity consumption of the whole society), and spatial urbanization (number of road lights per 10,000 people).

Among them, the total population at the end of the year and the electricity consumption of the whole society have a negative impact. The number of green patents and the number of road lights per 10,000 people had a positive impact on the carbon emissions of the central and southern Liaoning urban agglomeration, and the corresponding regression coefficients were 0.081 and 0.014, respectively. In other words, with each 1-unit increase in the above explanatory variables, the corresponding carbon emissions will increase by the above numerical units.

In general, a decrease in population may lead to a decrease in carbon emissions, as a decrease in population reduces energy consumption and the demand for living activities. However, there may be some special circumstances in the case of urban agglomerations in central and southern Liaoning that lead to the correlation between population reduction and increased carbon emissions. For example, compared to other regions, the central and southern Liaoning city cluster may face the problem of economic restructuring and energy structure, especially in the process of production reduction and transformation of heavy industry and resource industries. The energy supply of the central and southern Liaoning city cluster mainly depends on fossil fuels such as coal. Population decline may lead to a decline in energy demand, but if the energy mix is not effectively changed and remains over-reliant on coal, carbon emissions may be difficult to reduce. Therefore, the process of energy mix transformation and renewable energy development is crucial for reducing carbon emissions.

There is actually very little research on carbon emissions in the Central and southern Liaoning city cluster. Although the indicators selected in this research in terms of population level, economic level, and technological level are different from Zhou’s research [44], the degree and direction of the factors are relatively close to the results. This study is the first to study the impact of factors such as energy consumption level, spatial urbanization level, and ecological greening level on carbon emissions in the Central and southern Liaoning city cluster.

4.5.4. Guanzhong Plain City Cluster

The cities of Guanzhong Plain urban agglomeration involved in this study include Xi’an, Baoji, Tongchuan, Weinan, Yangling, Shangluo, Tianshui, Pingliang, Qingyang, Yuncheng, and Linfen. The factors affecting the carbon emissions of Guanzhong Plain urban agglomeration will be further analyzed below. The carbon emissions of Guanzhong Plain urban agglomeration are shown in Figure 7. In 2005, carbon emissions totaled 286.18 million tons; in 2010, the minimum carbon emission was 335.508 million tons; in 2015, the minimum carbon emission was 390.8773 million tons; and in 2021, the minimum carbon emission was 468.22 million tons, showing a gentle upward trend. By 2021, most of the areas with the most serious carbon emissions in the Guanzhong Plain urban agglomerations will be concentrated in the middle of the urban agglomerations, such as Xi’an, Weinan and Shangluo, and Linfen, and Yuncheng in the eastern region will have the least carbon emissions, showing that the central part of the urban agglomerations has the most carbon emissions and the eastern part has the least carbon emissions.

The reliability and validity of the least square method were confirmed through multiple validations using the stationarity test, multicollinearity test, and fitting effect test for the Guanzhong Plain urban agglomeration. The fitting results of the Guanzhong Plain urban agglomeration are shown in

Table 7. Estimation results of least square method for urban agglomeration in Guanzhong Plain.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	7.492	0.337		22.204	0.000
lnP	−0.082	0.031	−0.209	−2.694	0.008	0.387	2.583
lnGDP	0.148	0.027	0.453	5.485	0.000	0.340	2.941
lnOGP	0.014	0.012	0.117	1.117	0.266	0.212	4.715
lnDE	−0.084	0.009	−0.499	−9.640	0.000	0.865	1.155
lnTEC	−0.046	0.011	−0.290	−4.011	0.000	0.443	2.259
lnNRL	−0.004	0.007	−0.034	−0.495	0.622	0.505	1.979
lnGA	0.071	0.022	0.300	3.176	0.002	0.261	3.838

.

Regression equation of Guanzhong Plain urban agglomeration:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=7.492-0.082\mathrm{lnP}+0.148\mathrm{lnGDP}+0.014\mathrm{lnOGP}-0.084\mathrm{lnDE}-0.046\mathrm{lnTEC}-0.004\mathrm{lnNRL}+0.071\mathrm{lnGA}$$

(8)

According to the standardized regression coefficient, the factors affecting the total carbon emissions are listed in order from largest to smallest: economic urbanization (gross regional product per capita), ecological urbanization (green space), population urbanization (total population at the end of the year), and energy urbanization (total electricity consumption and dust emissions).

According to the regression equation of Guanzhong Plain urban agglomeration, GDP and green space area have a positive impact on carbon emissions of Guanzhong Plain urban agglomeration, and their corresponding regression coefficients are 0.148 and 0.071, respectively. At the end of the year, the total population, electricity consumption of the whole society, and dust emissions are negative, and the impact on carbon emissions is negatively correlated.

From the perspective of energy urbanization, the coefficient of electricity consumption in the whole society is −0.046 and that of dust emission is −0.084. With social and economic development and technological progress, traditional high-carbon energy (such as coal) is gradually replaced by clean and low-carbon energy (such as wind, solar, nuclear, etc.). This upgrading and transformation of the energy structure makes the increase in electricity consumption not necessarily accompanied by an increase in carbon emissions but may lead to a reduction in carbon emissions. With the improvement of environmental protection awareness and the implementation of environmental management policies, many industries and enterprises have adopted effective dust control measures to reduce dust emissions in the atmosphere.

Chen et al. [67] pointed out that for the Guanzhong Plain city cluster, technological scale promotes increases in carbon emissions, output scale, and energy scale, which continues to drive an increase in carbon emissions. However, this research is different. Chen’s research is based on the Kaya identity and is an expansion of GDIM decomposition method, which is limited by many factors and cannot examine indicators such as population and ecology. After incorporating indicators such as population urbanization and ecological urbanization into the STIRPAT model in this research, the role of technological factors is reduced. However, it cannot be denied that both studies [39,68,69] point out that economic level is an absolute positive factor, so in the future development process of the Guanzhong Plain city cluster, it will still be necessary to always pay attention to the growth relationship between economic development and energy consumption.

4.5.5. Central Plains City Cluster

This study covers the cities of the Central Plains city cluster, including Zhengzhou, Kaifeng, Luoyang, Nanyang, Anyang, Shangqiu, Xinxiang, Pingdingshan, Xuchang, Jiaozuo, Zhoukou, Xinyang, Zhumadian, Hebi, Puyang, Luohe, Sanmenxia, Jiyuan, Changzhi, Jincheng, Yuncheng, Xingtai, Handan, Liaocheng, Heze, Huaibei, Bengbu, Suzhou, Fuyang, Bozhou. In this study, four time points (2005, 2010, 2015, and 2021) were selected for comparison. As shown in Figure 8, the carbon emissions of the Zhongyuan Urban agglomeration also showed a trend of continuous growth during 2005–2021. From 2005 to 2021, the carbon emissions of each city in the Zhongyuan Urban Agglomeration increased significantly. The total carbon emissions of the Central Plains urban agglomerations also increased significantly, increasing to 915.5628 million tons, of which the Central Plains urban agglomerations, especially Xingtai and Handan in Hebei Province and Yuncheng in Shanxi Province, emitted the most carbon.

The reliability and validity of the least square method were verified through the multiple validations of the stationarity test, multicollinearity test, and fitting effect test for the Central Plains urban agglomeration. The fitting results of the Central Plains urban agglomeration are shown in

Table 8. Estimation results of least square method for Central Plains urban agglomeration.

	Unnormalized Coefficient		Standardization Coefficient	t	Significance	Collinearity Statistics
	Beta	Standard Error	Beta			Allowance	VIF
(constant)	287.199	41.559		6.911	0.000
lnP	0.755	0.054	0.320	13.959	0.000	0.612	1.633
lnGDP	0.030	0.002	0.437	16.896	0.000	0.259	3.859
lnOGP	0.210	0.209	0.027	1.006	0.315	0.228	4.377
lnDE	0.006	0.001	0.161	7.601	0.000	0.959	1.042
lnTEC	0.466	0.428	0.030	1.091	0.276	0.354	2.827
lnNRL	0.014	0.002	0.260	7.941	0.000	0.698	1.433
lnGA	0.001	0.003	0.006	0.313	0.754	0.458	2.184

.

Regression equation of Central Plains urban agglomeration:

$${\mathrm{lnE}}_{{\mathrm{CO}}_2}=287.199+0.755\mathrm{lnP}+0.03\mathrm{lnGDP}+0.21\mathrm{lnOGP}+0.006\mathrm{lnDE}+0.466\mathrm{lnTEC}+0.014\mathrm{lnNRL}+0.001\mathrm{lnGA}$$

(9)

According to the regression equation of the Central Plains urban agglomeration, economic urbanization (GDP), population urbanization (population), spatial urbanization (number of road lights per 10,000 people), and energy urbanization (dust emissions) have a positive impact on the carbon emissions of the Central Plains urban agglomeration.

The economic urbanization level of Guanzhong Plain urban agglomeration is relatively high, and the increase of GDP may lead to the increase of carbon emissions. This is because as one of the important industrial bases in China, the economic development of Guanzhong Plain urban agglomeration is mainly dominated by manufacturing, and the increase of energy demand and energy consumption leads to the increase of carbon emissions. And higher GDP boosts infrastructure and urban expansion. Large-scale infrastructure construction usually requires a lot of energy and resources, such as the production of building materials, transportation and construction of energy consumption will bring about an increase in carbon emissions. At the same time, urban expansion will also bring land use changes and increased transportation demand, further aggravating carbon emissions.

At the same time, the population urbanization of the Guanzhong Plain urban agglomeration has accelerated the process of energy consumption, increased transportation demand, industrial development and building energy consumption, resulting in an increase in carbon emissions. The increase in carbon emissions caused by population urbanization is mainly due to the increase in energy consumption, transportation demand, industrial development and building energy consumption. It is necessary to reduce carbon emissions through effective planning and management to promote sustainable urban development.

The positive influence of energy urbanization on carbon emissions of Central Plains urban agglomeration is caused by the increase of industrial production and energy consumption. With the acceleration of urbanization, industrial activities and energy demand have increased, and a large number of factories and enterprises have concentrated in the region, releasing a large amount of greenhouse gases. At the same time, building energy consumption and transportation demand are rising, exacerbating the increase in carbon emissions. Energy urbanization has a positive impact on the carbon emissions of the Central Plains urban agglomeration. Effective measures should be taken to reduce carbon emissions and promote sustainable development.

These results show that urbanization factors have a positive promoting effect on CO₂ emissions. This is quite similar to the results of previous studies [40,41,67,70,71,72], but the factors selected in this study are more comprehensive. This indicates that the Central Plains city cluster is still in the stage of development and growth, and urbanization is still on the rise. Compared to the other four city clusters in north China, there is still considerable space for improvement in the carbon emissions growth of the Central Plains city cluster.

5. Conclusions and Suggestion

By studying the impact of urbanization factors on carbon emission in five of the most important northern urban agglomerations in China from 2005 to 2021, the following results are obtained:

(1)

From the analysis of the five urban agglomerations as a whole, carbon emissions are generally on the rise, the main driving factor affecting carbon emissions is energy urbanization, and the strongest inhibiting effect is population urbanization. The main reason is that China’s northern city clusters still rely on traditional energy sources such as coal, oil, and natural gas. Although some regions have begun to promote clean energy use, the vast majority of regions still rely mainly on traditional energy sources, which greatly stimulates the growth of carbon emissions. Population urbanization refers to the process of gradual migration and concentration of rural population to cities, and population urbanization has an inhibitory effect on carbon emissions, indicating that the population urbanization level of the five northern urban agglomerations is relatively low on the whole. For these five urban agglomerations, population concentration and transfer in the process of urbanization have little impact on carbon emissions. The electricity consumption of the whole society is usually inversely proportional to carbon emissions, that is, with the increase of electricity consumption, carbon emissions are reduced, this phenomenon shows that the transformation of energy structure has achieved certain results in many cities, transforming the energy structure and increasing the proportion of clean energy is the only way to reduce carbon emissions.

(2)

From the perspective of urban agglomerations, it is generally believed that energy consumption has a significant impact on carbon emissions, but an interesting phenomenon has emerged in this research: the role of energy urbanization factors varies among different urban agglomerations. In the Beijing–Tianjin–Hebei city cluster, the Central and southern Liaoning South city cluster, and the Guanzhong plain city cluster, energy urbanization factors play a negative role; in the Shandong Peninsula city cluster and the Central Plains city cluster, energy urbanization factors play a negative role. This indicates that different urban agglomerations have different policies in dealing with energy urbanization factors.

(3)

In terms of urban agglomerations, the urbanization factors that have the greatest positive impact on carbon emissions of the Beijing–Tianjin–Hebei city cluster, Shandong Peninsula city cluster, central and southern Liaoning city cluster, Guanzhong Plain city cluster and Central Plains city cluster are ecological urbanization, economic urbanization, scientific and technological urbanization, economic urbanization, and population urbanization, respectively. The Shandong Peninsula city cluster, Guangzhou Plain city cluster, and Central Plains city cluster in the Yellow River Basin have some commonalities in terms of carbon-emission-influencing factors.

Therefore, the emission reduction policy recommendations for different urban agglomerations are as follows:

(1)

In the process of urbanization, urban agglomerations in the north should reduce the consumption of polluting energy, change the energy structure, increase investment in clean energy, reduce dependence on traditional energy, give play to the role of science and technology, and promote digital and intelligent operation.

(2)

In the process of ecological urbanization, the Beijing–Tianjin–Hebei city cluster needs to control the degree of energy utilization, reduce unnecessary energy consumption, promote ecological urbanization by scientific and technological means as much as possible, and improve energy efficiency. It is also necessary to give play to the positive significance of ecological urbanization, improve the ecological environment, increase the carbon absorption capacity of the ecosystem, and reduce the impact of carbon emissions.

(3)

In the process of scientific and technological urbanization, the central and southern Liaoning city clusters should adopt green technologies in a moderate and standardized manner, reduce the negative impact caused by the redundant use of green technologies, constantly optimize the energy structure, and give play to the role of science and technology in emission reduction. It is also necessary to promote the effective implementation and promotion of green patents, provide greater possibilities for the transformation of the energy structure, achieve digital management and intelligent operation with the help of advanced technologies, improve the efficiency of urban resource utilization, and reduce carbon emissions.

(4)

The Central Plains urban agglomeration needs to make reasonable planning for the urban population size, control the population size, optimize the urban layout, promote the optimization and upgrading of the transportation system, and reduce the carbon emission caused by population urbanization.

(5)

For Guanzhong Plain urban agglomeration and Shandong Peninsula urban agglomeration, compared to other regions, they may face the problems of economic restructuring and energy structure, so they need to promote green production and consumption, encourage low-carbon technological innovation, promote emission reduction through economy, and reduce energy consumption and carbon emissions.

The relevant conclusions and policy recommendations for these five city clusters are shown in

Table 9. A summary table of conclusions and improvement for city cluster.

City Cluster	Beijing–Tianjin–Hebei	Shandong Peninsula	Central and Southern Liaoning	Guanzhong Plain	Central Plains
Population urbanization	*	*	−	−	+
Economic urbanization	*	+	*	+	+
Sci-tech urbanization	+	+	+	*	*
Energy urbanization	−	+	−	−	+
Spatial urbanization	*	*	+	*	+
Ecological urbanization	+	−	*	+	*
Improvement direction	1. Reduce unnecessary energy consumption and improve energy efficiency 2. Vigorously develop green and low-carbon technologies 3. Reasonably utilize land for greening purposes	1. Continue to expand urban green space area 2. Reduce fossil energy consumption 3. Vigorously develop green and low-carbon technologies	1. Properly control population size 2. Reasonably arrange urban architectural space 3. Reduce unnecessary energy consumption and improve energy efficiency 4. Vigorously develop green and low-carbon technologies	1. Properly control population size 2. Reduce fossil energy consumption 3. Reduce unnecessary energy consumption and improve energy efficiency 4. Reasonably utilize land for greening purposes	1. Properly control population size 2. Significantly reduce fossil energy consumption and energy intensity while promoting economic growth 3. Reasonably arrange urban architectural space

+/− represents the direction of influence of the factor and * represents the negligible or insignificant impact of the factor.

.

In the process of realizing urbanization, the impact of carbon emissions needs to be considered in all aspects. Population urbanization requires rational planning of urban population density, optimization of urban layout, and reduction of transportation demand and carbon emissions. Economic urbanization requires promoting green production and consumption, encouraging low-carbon technological innovation, and reducing energy consumption and carbon emissions. Technological urbanization can realize digital management and intelligent operation with the help of advanced technologies, improve the efficiency of urban resource utilization and reduce carbon emissions. Ecological urbanization emphasizes ecological protection and restoration, building eco-friendly cities, reducing the exploitation and destruction of natural resources, and reducing carbon emissions. Energy urbanization needs to promote the use of clean energy, reduce the consumption of fossil fuels, and reduce carbon emissions. Spatial urbanization should pay attention to urban spatial planning, advocate low-carbon travel and urban-rural integration development, and reduce carbon emissions.

At present, this research has only conducted preliminary analysis on these five urban agglomerations in northern China, and the selected indicators are common—that is, the influencing factors of all urban agglomerations are the same. However, from this study, it can be found that the carbon emissions factors (especially the ones with the greatest impact) for these five urban agglomerations are slightly different—in particular, the previously believed impact of energy urbanization factors is not significant in some urban agglomerations. Therefore, future research will gradually delve into targeted studies of various urban agglomerations To decompose and localize the specific influencing factors of each urban agglomeration. For the Beijing–Tianjin–Hebei city cluster, the influencing factors will be refined from the aspects of ecological urbanization, technological urbanization, etc.; for the Shandong Peninsula city cluster, the influencing factors will be refined from the aspects of economic urbanization, ecological urbanization, etc.; for the Central and southern Liaoning city cluster, the influencing factors will be refined from aspects such as technological urbanization and population urbanization; for the Guanzhong Plain city cluster, the influencing factors will be refined from the aspects of economic urbanization, ecological urbanization, etc.; for the Central Plains city cluster, the influencing factors will be refined from aspects such as economic urbanization and population urbanization.

In addition, the research on the carbon emission impression factors of urban agglomeration is still in infancy, and most of the relevant research focuses on the urbanization of the Beijing–Tianjin–Hebei city cluster, the Yangtze River Delta city cluster, the Pearl River Delta city cluster, and other city cluster. Due to the difficulties in transforming traditional energy and industrial structures, carbon emissions in urban agglomerations in China are the main source of carbon emissions and an important region responsible for carbon neutrality. Therefore, future research will have a broader understanding of the specific sources and influencing factors of carbon emissions in the urbanization process of other urban agglomerations.