Study on the impact of comprehensive urbanization on urban civil building CO2 emissions in China

With the rapid development of China, urbanization has become an important research topic of China’s CO2 emissions. To fill the gap in considering the spatial correlation of the comprehensive urbanization that includes multi-dimensional factors on CO2 emissions from urban civil buildings (UBEC), this study constructs a comprehensive evaluation indicator of urbanization from four aspects including population, economy, society, and land urbanization by using the entropy method. The spatial spillover effect of UBEC and the impact of comprehensive urbanization on UBEC are also studied by using the spatial panel model in this paper. This study finds out that UBEC has obvious spatial spillover effects. During the early years of the study period, the eastern coastal areas had greater carbon emissions, while in recent years, they have gradually transitioned to the northwestern regions. Comprehensive urbanization has a significant promotion effect on it. And foreign direct investment and per capita energy consumption also have positive impact on UBEC. This study provides a reference for measuring the effects of urbanization on sector-specific CO2 emissions and may be useful for energy efficiency and emission abatement efforts in China.


Introduction
With the rapid pace of economic development, terrible side effects such as severe polluting problems are caused worldwide by abusing and overusing coal and fossil fuel Song et al. 2019a, b;Zhu et al. 2020). People are increasingly concerned about global warming caused by pollutant emissions . CO 2 emissions in greenhouse gases are the main factor of climate change (Yang et al. 2021). Therefore, decreasing CO 2 emissions has been one of the key tasks for countries around the world (Song et al. 2019a, b). And China is one of the world's largest carbon producers (Chen et al. 2020a, b, c;Song et al. 2020). From an end-use of energies perspective, industry, transport, and construction are listed as the three industries with the largest energy consumption. The Intergovernmental Panel on Climate Change (IPCC) reveals that buildings are estimated to account for 31% of the world's total energy-related CO 2 emissions by 2020, rising to 52% by 2050 (Hou et al. 2021). The CO 2 emissions from China's construction have grown from 668 million tons to 2.04 billion tons between 2000 and 2017, which is approximately 17-21% of the China's CO 2 emissions ). In the lifetime of a building, CO 2 emissions from the operation phase of buildings contribute to 2/3 of the total CO 2 emissions from building, most of them from urban buildings (Mengjie 2019). As the economy grows and urbanization accelerates, these numbers are on the rise. To reach China's commitment to reduce carbon emissions intensity by 60-65% by 2030 compared to 2005 (Chen et al. 2020a, b, c;Ding et al. 2019;Wu et al. 2020), the problem of reducing carbon emissions from urban civil buildings needs to be resolved urgently.
China is in a phase of rapid urbanization. The direct effects of urbanization are the concentration of urban population, the aggregation of land use, and the clustering of economic activities (Huo et al. 2021). Based on the urbanization rate represented by the proportion of urban population, China's urbanization rate increases from 17.92 in 1978 to 60.6% in 2019 (Lin and Zhu 2021). Urbanization is a critical influence on CO 2 emissions from buildings. Regarding urbanization research, most of the urbanization is characterized by the ratio of the urban population. But from the current point of view, urbanization is multifaceted, and if it is only studied from the perspective of the population, it may be relatively single. Most of the existing studies on the effects of urbanization on the CO 2 emissions of buildings have been conducted on the entire life cycle of buildings. Due to the entire life cycle of the building, the operating phase accounts for a relatively large proportion. It is important to study the changes of CO 2 emission from the operation phase of the building separately. There are relatively few studies in this field that consider spatial correlation, and most studies use non-spatial econometric models for calculation and analysis. In addition, urbanization should be considered from multiple angles to analyze its impact on building CO 2 emissions comprehensively.
To fill these gaps, this paper makes a focus on the operation phase of CO 2 emissions from urban buildings. The contributions are listed below: first, this paper analyzes the effect of comprehensive urbanization on CO 2 emissions from the operation phase of urban civil buildings, while most of the precious articles study the entire life cycle of buildings (Hou et al. 2021;Su et al. 2021;Zhang et al. 2019;Zhang et al. 2020). Second, a comprehensive urbanization indicator is constructed based on the urbanization indicators of the four aspects including population, land, economy, and society, which are used to characterize the comprehensive development level of urbanization. Third, the influence of comprehensive urbanization on building CO 2 emissions is explored by taking spatial correlation into account based on the spatial panel model. Rarely articles have studied the association between CO 2 emissions and urban civil buildings using the spatial panel model (He et al. 2020;Huo et al. 2020;Huo et al. 2021;Li et al. 2021;Wang and Feng 2018). This paper is structured as follows: the "Literature review" section gives a review of the literature review on urbanization and building carbon emissions. The "Method and data" section introduces the calculation methods of CO 2 emissions from the operation phase of urban buildings, as well as the theoretical models and data sources of various variables used in this research. The "Results and discussion" section presents the empirical analysis. The "Conclusion" section draws to conclude.

Literature review
In the field of CO 2 emissions, there are two main types of research on urbanization. One is to research the effect of urbanization on carbon emissions by taking the ratio of urban population to the total population of the region as an urbanization indicator. Ding and Li (2017) use the LMDI model for calculation and analysis. The study finds that the mechanism of urbanization impact on carbon emissions has significantly regionally heterogeneous. Bai et al. (2019) use the fixed-effect two-stage least squares model based on data from four dimensions of urbanization. And the study concludes that urbanization has increased CO 2 emissions. Han et al. (2019) analyze the factors that influenced the intensity of CO 2 emission based on the extended kaya model. The study finds that urbanization and employment rate of urban population are the major influencing factors and have the effect of decreasing the intensity of CO 2 emission. Chen et al. (2019); Liu and Liu (2019a); Muhammad et al. (2020); Wu et al. (2017); and others study the association between urbanization and CO 2 emissions regionally and in groups, while considering regional heterogeneity.
The other emphasizes the harmonization of population, economy, society, and ecological environment based on the new urbanization strategy first proposed in Zhejiang Province in 2006 (Lin and Zhu 2021). The studies characterize urbanization from multiple perspectives, and then study the relationship between it and CO 2 emissions. Liu et al. (2018) use the Tobit model to analyze the impact on 10 typical urban agglomerations from the four aspects of urbanization including population, space, industry, and economy. The study finds that the impact of urbanization in different dimensions on different urban agglomerations varies greatly. Wang et al. (2019a, b) also establish a comprehensive urbanization indicator system in four aspects to constitute the quality urbanization. And based on the Geographically Weighted Regression Model (GWR), the study finds that the urbanization quality has significant temporal and spatial distribution differences among provinces. Zhou et al. (2019) use a spatial agglomeration function, gray correlation model, and Kuznets curve model to analyze the economy, population, and spatial urbanization. The study finds that economy urbanization has the most significant impact on CO 2 emissions. Wang and Zhao (2018) establish a comprehensive urbanization indicator system from three aspects including population, technology, and industrial institutions. The country is also divided into three zones according to the level of urbanization transition and studied separately by using a modified STIRPAT model. The study finds that China's urbanization is uneven, and that the dominant industrial structure has different effects on CO 2 emissions.
As a major sector of consumption of energy, the construction industry has attracted many scholars' attention. The whole life cycle of a building includes the stages of material production, building construction, building operation, and building abandonment.  Zhao et al. (2019) argue that CO 2 emissions from the constructions' operation phase are the largest proportion of the building's entire life cycle. Few scholars study the association of urbanization and CO 2 emissions from buildings, especially during the operation phase of buildings. Zhang et al. (2021) establish urbanization indicators from the three aspects including population, economy, and technology. And consider the region-specific heterogeneity to study the impact of urbanization on CO 2 emissions from buildings. Huo et al. (2020) establish a comprehensive indicator system of urbanization based on the quantitative and structural dimensions in three dimensions: population, economy, and space. The study uses the STIRPAT model to study the multiple effects of urbanization on CO 2 emissions from the urban construction sector. Furthermore, Huo et al. (2021) also use a panel threshold regression model to examine the dynamic mechanisms underlying the effects of urbanization on carbon emissions from urban civil buildings. Wang et al. (2019a, b) use a geographically weighted regression model to examine the association of urbanization with carbon emissions in six sectors, including the construction sector. Liu et al. (2020) use the system dynamics model to predict the long-run carbon emissions from buildings, and simulate the impact of different policies on the CO 2 emissions from buildings in Beijing. The study explains in more detail the urban building stock evolution and reveals the effects of the policies.
In summary, the construction operation stage is considered the largest carbon emissions in the buildings industry by many studies. However, few studies conduct the effect of urbanization on building CO 2 emissions at this stage. To fill the gaps in this research field, this paper conducts further research on it. It mainly reflects in two aspects: (1) To measure urbanization comprehensively, a comprehensive urbanization index is constructed based on the urbanization evaluation system from the four aspects including economy, population, land, and society. (2) On the basis of considering the spatial correlation and spatial spillover effects, a spatial panel model is used to explore the impact of urbanization on the CO 2 emissions from urban civil buildings.

Model specifications
According to the "First Law of Geography," the variables are not independent of each other in the study of the dimensionality of China's provinces, but have extensive connections. Considering this connection, the analysis of the problem can be more accurate. Therefore, this paper intends to analyze the effect of urbanization on CO 2 emissions from the operation of urban civil buildings (UBEC) by using a spatial panel model.

Spatial correlation
Having spatial correlation is the prerequisite for applying spatial econometrics. The measure of spatial selfcorrelation is divided into global and local statistics. Global autocorrelation statistics are quantitative indicators that reflect whether there is spatial autocorrelation for all provinces, while local autocorrelation statistics are quantitative indicators that reflect whether there is spatial autocorrelation among other surrounding provinces for one province (Lu et al. 2021;Tiyan and Hanchen 2010). This paper uses the global and local Moran index to test the spatial autocorrelation of UBEC. The calculation method is as follows: Global Moran's I: where n represents 30 provinces, w ij is the spatial weight matrix (for measuring the spatial distance between two regions), x i and x j are the independent variables, respectively, and x is the average values of the independent variables.
The range of Moran's I is (−1,1). (0,1) means positive autocorrelation, while (−1,0) is the opposite. The Moran's I is close to 0, indicating that the distribution in space is random and there is no spatial autocorrelation.
Local Moran's I: The meaning of Local Moran's I is similar to Global Moran's I. Positive I i means that the high (low) value of area i is surrounded by surrounding high (low) areas; negative I i means that the high (low) value of area i is surrounded by surrounding low (high) areas (Qiang 2014).

Spatial panel model
The general form of the spatial panel model is as follows: Among them,y i, t − 1 is the first-order lag of the explained variable y i, t ; w 0 i x t θ is the spatial lag of the explanatory variable; λ t is the time effect; m i is the i-th row of the disturbance term spatial weight matrix M; and w ij represents an n×n spatial weight matrix, which measures the spatial distance between regions.
Spatial econometric models allow for three kinds of interactions. One is the interaction of the endogenous dependent variables with each other. Another is the interaction that occurs with exogenous independent variables. And the third one is the interaction between disturbance terms  . Therefore, different interaction effects refer to three spatial econometric models.
In this study, after a series of spatial model tests, the spatial Dubin model is finally selected to calculate and explain the impact of urbanization on UBEC.
Liu and Liu (2019a) illustrate the spatial spillover effect mechanism. The spatial Dubin model is calculated as follows: Because the effect of explanatory variable varies across region, the total effect means the row (column) mean of the coefficient matrix of the explanatory variable. The direct effect refers to the diagonal mean of the coefficient matrix of the explanatory variable. The indirect effect is the difference between the total and direct effect (LeSage and Pace 2014).
The spatial weight matrix selected is the spatial inverse distance matrix (take the reciprocal of the distance between regions) in this paper, and records the distance between region i and region j as d ij , which can be expressed as: The establishment of the spatial Dubin model (SDM) is as follows: where UBEC means the CO 2 emissions from the operation phase of urban civil buildings (ten thousand tons); urb means the comprehensive urbanization; FDI means the foreign direct investment (100 million yuan); PUEI means the energy consumption per capita (ten thousand tons of standard coal per person); and ε it means the disturbance term.

Calculation of UBEC
According to the "Professional Knowledge of Housing Construction Engineering" and the "Regulations on Energy Conservation of Civil Buildings," according to the nature of using, buildings are classified into civil buildings, industrial buildings, and agricultural buildings. Civil buildings are divided into residential and public buildings (Mengjie 2019). The energy consumption of civil buildings is classified into urban, rural, and northern heating energy consumption. Generalized building energy consumption is usually measured as the whole life cycle of the building, while it refers to the energy consumption from the operation phase of the building in the narrow sense. In 2000, the Ministry of Construction formally unified the definition of building energy consumption, and defined building energy consumption as energy consumption from the operation phase of building (Jia 2016). The scope of this paper is the operational phase of URBan civil buildings. Energy consumption and CO 2 emissions are calculated at this stage, and heating energy consumption in the north is not calculated separately. The emission factor method is used in this paper to calculate UBEC. The calculation is shown in Equation (7): where UBEC is the total CO 2 emissions from the operation phase of urban civil buildings. e j is the energy consumption of urban civil buildings in the energy balance sheet (physical quantity). δ j is the heat conversion factor. ε j is the j-th energy CO 2 emission coefficient of urban civil buildings.
In energy statistics, there are no statistics on energy consumption from the operation phase of buildings. In terms of the energy consumption characteristics of each industry, the energy consumption of wholesale, retail and accommodation, catering, other tertiary industries, and living consumption after deducting transportation energy consumption mainly occur in the operation phase of buildings. It can be used as the energy consumption of the building (Min et al. 2012).
The average low calorific value uses the data from the appendix of the "China Energy Statistical Yearbook." The CO 2 emission coefficient is the data given in the 2006 edition of the "IPCC National Greenhouse Gas Inventory Compilation Guidelines," as shown in Table 1:

Establishment of a comprehensive index system for urbanization
With the rapid development of China, urbanization is no longer limited to population urbanization but is developing in a more diverse direction. Economic urbanization, social urbanization, and land urbanization have increasingly become the focus of scholars' research. This paper uses the method mentioned in Wang et al. (2019a, b) to combine the above four aspects of urbanization to construct a new comprehensive urbanization indicator system, which includes almost all aspects of urbanization, and analyzes the process of China's provincial urbanization comprehensively. The indicators involved are shown in Figure 1.
The entropy method determines the index weight according to the degree of change of each index value. It can avoid errors caused by the subjective preference of experts in the subjective weighting method. Compared with other subjective weighting methods, the accuracy rate is higher, the objectivity is stronger, and the results can be better explained. The entropy method avoids errors caused by the subjective preference of experts in the subjective weighting method. This method was introduced by Shannon in 1948 to the discipline of information management to express information or uncertainty (Shen et al. 2015). In this paper, this method is used to objectively weigh the indicators mentioned in the previous chapter to obtain a new comprehensive urbanization index to measure the development of China's urbanization process, and then to study its impact on UBEC. The main entropy method's calculation steps are as follows: (1) Index normalization Suppose there are m samples and n indexes, then x ij is the j-th index of the i-th sample (i=1, ……, m, j=1, ……, n). Because the measurement units of each indicator are not standardized, they need to be standardized before calculating the comprehensive urbanization indicator. Processing methods for positive indicators and negative indicators are different, and the formula is as follows: Positive indicators: Negative indicators: (2) Calculate the weight of the i-th sample value under the j-th index in the index: (3) Calculate the entropy value of the j-th index: where =1/ ln(n) > 0, e j ≥ 0. (4) Calculate information entropy redundancy (difference): (5) Calculate the weight of each indicator: (6) Calculate the comprehensive score of each sample: For weights and a comprehensive score of the provincial indicators in the comprehensive urbanization indicator system, this paper uses Beijing in 2017 as an example. A variable with a positive sign means that it has a promoting effect on comprehensive urbanization, that is, comprehensive urbanization will become larger as the variable becomes larger. Variables with negative signs are the opposite. The result is shown in Table 2.

Results and discussion
Analysis of calculation results of UBEC According to the calculation method, the UBEC emissions can be obtained. The overall change trend of UBEC is shown in Figure 2. In the past 10 years, UBEC shows an overall increasing trend, and the growth rate has begun to decline after 2011.
Since 2014, the growth rate has been increasing again. In 2017, UBEC was 104,822×104 t, an increase of 7.3% over 2016.
To analyze the spatial movement of China's building carbon emissions in recent years, this study uses ArcGIS 10.2 to draw the center of gravity of UBEC (Yan et al. 2021). When the carbon emission center shifts to a certain direction, it means that the spatial distribution of regional carbon emission is uneven during this period (Yuewu et al. 2016). It can be seen from Figure 3 that from 2008 to 2017, the center of gravity of UBEC across the country has not moved much, but it mainly concentrates in Henan Province, and the overall trend is moving from the northeast to the northwest. This shows that the UBEC has obvious spatial and regional uneven characteristics, and further confirms the necessity of studying carbon emissions from a spatial perspective.

Spatial distribution of comprehensive urbanization
The comprehensive urbanization is an indicator for comprehensively measuring the process of China's urbanization. Figure 4 shows the development process of comprehensive urbanization in 30 provinces of China. It can be concluded that provinces with high urbanization in the early stage were mainly in the eastern coastal region, and they gradually transferred to the western region in the later stage. High urbanized provinces increase obviously. And it reflects the rapid development of China's economy, society, land, and many other aspects.

Spatial correlation analysis
Global Moran's I are listed in Table 3. The Moran's I are all positive and almost greater than 0.2, which is significant at 1% level. It suggests that UBEC has positive spatial effect.
As shown in Figure 5, most provinces in China are distributed in one and three quadrants, which further proves the positive spatial effect of UBEC among provinces in China.
Since the Moran scatter plots cannot judge the statistical significance of clustering, additional analysis using LISA is necessary . Figure 6 shows the LISA and salience graphs used to examine the local autocorrelation of UBEC in 2008. In H-H type provinces, Beijing, Shanghai, and Jilin are significant at 1% level. Jiangsu, Inner Mongolia, and Hebei are significant at 5% level and Liaoning is significant at 0.1% level. In H-L type provinces, Guangdong is significant at 5% level. In L-H provinces, Tianjin is significant at 5% level. In L-L type provinces, Year urban building CO2 emissions growth rate Figure 2 The trend of UBEC Sichuan and Guizhou are significant at 1% level, while Xinjiang, Gansu, Qinghai, Chongqing, Guangxi, and Hainan are significant at 5% level. Figure 7 shows the LISA and salience graphs used to examine the local autocorrelation of UBEC in 2017. In the H-H type provinces, Jilin and Hebei are significant at the level of 5%, and Liaoning is significant at the level of 0.1%. In the H-L type provinces, Xinjiang is significant at 0.1% level, and Guizhou is significant at 5% level. Compared with 2008, Beijing and Shanghai are moved to the L-H type provinces. In L-L type provinces, Sichuan is significant at 1% level, and Hainan, Guangxi, and Gansu are significant at 5% level.
Through comparative analysis, we can see that high emission concentration areas mainly concentrated in highly developed areas and the northern provinces in 2008. The northern provinces, including the old industrial bases in the northeast, are mostly heavy industries and lack innovation in the treatment of carbon emission pollution. The population density in highly developed areas was relatively high. China was in the stage of high-speed economic development at that time, and people's environmental protection consciousness was not strong. With the strengthening of China's strength in various fields, the gradually increase in people's awareness of environmental protection, and the implementation of policies such as the great development of the western part of the country, the growth of carbon emissions has slowed down significantly by 2018. The high emission concentration area gradually shifts to the western region.

Spatial model diagnostics
The previous article verifies the positive spatial correlation of UBEC among Chinese provinces. It shows that a spatial panel model needs to use, but some tests are required to complete the selection of which spatial panel model to use.
Hausman test First, the Hausman test is used to choose whether to use a fixed or random effect. The test results are listed in Table 4. The statistical value is 22.95 and the p value is 0.00001. The original hypothesis is rejected at 1% significance level, so the former should be chosen.
To test which of the three effects of regional fixed effect, time, and double fixed effect is the most appropriate for the study in this paper, the effect test is carried out. From the results, regional and double fixed effects' p value is 0.0508, and the null hypothesis is not rejected at the 5% significance level. For the test of time double fixed effects, the p value is 0.00001, rejecting the null hypothesis at the 1% significance level. It is more appropriate to choose a regional fixed-effect model. Table 4, for the LM test, the p values of LM-error and LM-lag are significant, rejecting the non-spatial hypothesis. For the Wald test, the p values of Wald-SAR and Wald-SEM are both Figure 3 Variations in center of gravity for CO 2 concentration significant at the 1% level, rejecting the hypothesis that SDM can degenerate into SAR model or SEM model. The LR test results are consistent with the Wald test results. Table 4 shows the test results. To sum up, this study uses the spatial Dubin model (SDM).

Spatial Dubin model estimation
The spatial Dubin model reflects the impact of comprehensive urbanization (URB), foreign direct investment (FDI), and per capita energy consumption (UEI). And it also reflects the spatial spillover effect of the above indicators on UBEC.
After the previous analysis, this study finally selected the regional fixed-effect spatial Dubin model, and used stata15.0 to complete the calculation and analysis. The estimated results are shown in Table 5.  The spatial autoregressive coefficient is positive and significant at the 1% level. It means that the UBEC has positive spatial spillover effect, which has been confirmed in numerous literatures (Liu and Liu 2019b;Huang et al. 2020;Pang et al. 2021). That is, every increase of one unit of UBEC in neighboring areas will increase the area's carbon emissions by 0.364%.
Comprehensive urbanization (URB) is significant at the 1% level and effectively suppresses the increase in UBEC. Based on the data, 1% increase of local comprehensive urbanization will reduce 0.89% of the UBEC. However, the increase in comprehensive urbanization in neighboring areas has contributed significantly to local carbon emissions. One percent increase in comprehensive urbanization in neighboring areas will increase 4.091% of the UBEC. This result shows that while developing comprehensive urbanization, more attention should be paid to avoid carbon transfer between provinces.
Foreign direct investment (FDI) has a promoting effect on the UBEC, and it is significant at the 10% level. One percent increase in local foreign direct investment will increase 0.027% of the carbon emissions of local urban civil buildings. FDI in neighboring regions also promotes local carbon emissions. The results show that f FDI drives China's economic development, and it also drives the increase in UBEC.
Per capita energy consumption (PUEI) also promotes the UBEC, and it is significant at the 1% level. One percent increase in local energy consumption per capita will increase 1.344% of the CO 2 emissions of local urban civil building. The PUEI of neighboring areas has no significant effect on local CO 2 emissions, but its coefficient is positive, which also has a promoting effect to some extent. This result shows that PUEI plays a leading role in increasing UBEC. It can be said that reducing PUEI is an effective way to reduce emissions.

Estimation and analysis of spatial spillover effects
Considering that there may exist deviations to test spillover effect by using point estimation. The direct, indirect, and total effects of UBEC are estimated by using the partial differential method in this section (Li et al. 2019). The estimated results are shown in Table 6.
It can be seen from Table 6 that the estimated results of the direct, indirect, and total effects of comprehensive urbanization are −0.7, 5.73 and 5.03, respectively. For the direct effect, 1% increase in comprehensive urbanization will reduce the carbon emissions from urban civil buildings by 0.7%. It means that the level of local comprehensive urbanization will increase, because of the changes in urban population, land use structure, tertiary industry, the development of economy, the improvement of education level, and the living standard of residents, which will restrain the increase of UBEC. However, the indirect effects show positive impacts. One percent increase in the comprehensive urbanization of neighboring areas will increase 5.73% of the UBEC in local cities and towns. The main reason is that with the vigorous implementation of environmental protection policies, the growth of comprehensive urbanization in neighboring areas has caused the shift of some industries with high carbon emissions and increased the UBEC in local cities and towns. The total effect obtained by adding the direct and indirect effects is positive. This result is compatible with several researches (Kasman and Duman 2015;Li et al. 2015). So, comprehensive urbanization has a significant role in promoting UBEC.
The estimated results of the direct effect, indirect effect, and total effect of FDI are 0.03, 0.11 and 0.14, respectively. In this article, FDI is used to express the degree of trade openness. The impact of trade openness on our country mainly has two aspects. On the one hand, foreign investment can bring advanced technology to developing countries and make them develop rapidly in line with the trend. On the other hand, the effect is the opposite. To attract foreign investment, developing countries lower the standards of environmental regulations, which may make them become "pollution haven" for developed countries (Chen et al. 2020a, b, c). Regardless of direct effect, indirect effect, or total effect, foreign direct investment can promote carbon emissions from urban civil buildings. Specifically speaking from the perspective of the overall effect, 1% increase in foreign direct investment will increase the carbon emissions from urban civil buildings by 0.14%. In other words, the influence of foreign capital on my country is more inclined to the second aspect. This is consistent with recent studies (Zhang and Zhang 2018;Sun et al. 2017).
The direct effect, indirect effect, and total effect of per capita energy consumption are estimated to be 1.36, 0.39, and 1.75, respectively. From the perspective of the overall effect, every 1% increase in per capita energy consumption will increase the carbon emissions from urban civil buildings by 1.75%. There are two main reasons. On the one hand, our country uses a single energy structure. For example, heating and industrial production in northern are dominated by coal. The combustion of coal will produce a large amount of CO 2 , which will have extremely adverse effects on the environment. On the other hand, China's energy utilization rate is low, and the rapid development of the economy is at the expense of energy use.

Conclusion
This study uses the entropy method to construct an indicator that aims to comprehensively describe China's urbanization process-comprehensive urbanization in four dimensions including population, economy, land, and society urbanization. By combining the characteristics of geographical space distribution, and considering the spatial correlation and spatial spillover effects, this paper comprehensively analyzes the effect of comprehensive urbanization on the UBEC. First, the UBEC has obviously positive spillover effects in space. The increase of CO 2 emissions from neighboring areas will drive the increase of local CO 2 emissions. Therefore, it is essential to take spatial effects into account.
Second, the areas with high comprehensive urbanization are mainly located in the eastern coastal areas. The reason is that the eastern coastal areas are rich in resources and technology advanced. However, with the accelerated development of China and the implementation of national policies, highly comprehensive urbanization is gradually transitioning to the western region. China is showing a good trend of all-round development.
Third, comprehensive urbanization promotes the UBEC. This indicates that the development direction of China's promotion of comprehensive urbanization and reduction of the UBEC is inconsistent. Therefore, we cannot blindly pursue urbanization development. We must make practical and reasonable decisions to achieve a win-win goal, for example, planning urban construction area and population rationally, increasing urban green area, and increasing support for clean energy companies.
Fourth, foreign direct investment has increased the UBEC. This shows that the government must do a good job of controlling relevant policies to avoid becoming a "pollution paradise" for developed countries, comprehensively and reasonably allocating foreign capital, to support more investment to be used in clean energy industries but less investment for high-carbon emission industries.
Finally, energy consumption per capita has significantly increased the UBEC, which shows that China needs to vigorously develop high-tech industries and accelerate the development and utilization of clean energy. Adhere to "green development" as the core, and not forget that low carbon is the prerequisite for all development while developing the economy and other aspects.
This study takes the UBEC in 30 provinces in China except for Tibet, Hong Kong, Macao, and Taiwan in the past 10 years as the research object and explores the effects of comprehensive urbanization on the UBEC from the spatial effect's perspective. It enriches the current research in this field. However, the research scope of this paper is only in the operation phase of urban civil buildings, and carbon emissions in other phases are not considered, and further research is needed. ***, **, and * represent significant at 1%, 5%, and 10%, respectively