Analysis on the constraint mechanism of transportation carbon emissions in the Pearl River Delta based on ‘Dual carbon’ goals

ABSTRACT Transportation not only provides support for economic and social development, but also has an important impact on emissions. Based on the DPSIR model, the constraint index system of transportation carbon emissions in the Pearl River Delta under ‘Dual carbon’ is constructed, and the hierarchical structure model is constructed by using the DEMATEL-ISM integration method. The constraint mechanism of transportation carbon emission in the Pearl River Delta is analyzed from the three dimensions of the correlation and importance, system hierarchical structure and constraint path of the constraints. The study found that there are six levels of constraints. The economic development level, carbon emission scale and the ‘Dual carbon’ goals are the core factors of the entire system. Developing public transport and intelligent transportation and increasing investment in new energy infrastructure and technology are conducive to the development of transportation in the Pearl River Delta and the realization of ‘double carbon’ goals.


Introduction
The enormous challenges and expected impacts of climate change have emerged. In response to the global climate change, from the United Nations Framework Convention on Climate Change in 1992 to the Kyoto Protocol in 1997, and then to the famous Paris Agreement signed by countries all over the world in 2016, a world pattern of Climate Governance has been formed after 2020 (Gao et al., 2017;Gao & Song, 2021;Li, 2019). Currently, more than 130 countries and regions in the world have successively put forward the relevant goals of achieving emission peak and carbon neutrality (Palmer, 2019;Shu & Xin, 2021). At the 75th United Nations General Assembly on Sep. 22, 2020, China made a solemn commitment to achieve carbon peak by 2030 and carbon neutrality by 2060. The establishment of the 'Dual carbon' goals has injected new impetus into China's low-carbon development in the new stage and pointed out a new direction. Focusing on high-quality development, fully considering regional differences and synergies, promoting the deep integration of low-carbon development and supply-side structural reforms is an important way to achieve the 'Dual carbon' goal (Zhang, 2021).
In 2020, the annual GDP of Guangdong Province exceeded 11 trillion yuan. Since it rose to the first place in 1989, the total GDP of Guangdong Province has ranked CONTACT Yunlong Liu longyun768@126.com first for 32 consecutive years in China, and it is the only province that has exceeded 10 trillion yuan (General Office of the Standing Committee of Guangdong Provincial People's Congress, 2021). From the perspective of GDP, industrial output value and the patents, the Pearl River Delta urban agglomeration has consistently ranked the top three among the nineteen urban agglomerations from 2006 to 2019 (Economic Geography, 2021). At the same time, as an important part of the 'four-beam and eight-column', Guangdong Province has reached a new level of 11 trillion, and the transportation also plays an important role in supporting Guangdong's 'economic building' (Yang, 2021). Due to the large economic scale, it is a very arduous task for the Pearl River Delta to ensure high development quality while maintaining large-scale economic growth under the 'Dual carbon' goals. Therefore, we should jump out of the limitations of traditional thinking, dig deep into the high-quality supply sources through system and mechanism innovation, and regard low-carbon development as a more efficient and desirable new development model (Zhaoxiang, 2021). However, with the development of regional economy and the advancement of urbanization, the city scale is constantly expanding; the travel distance of people is increasing, and the demand for transportation is also increasing. These changes make the contradiction between economic development, transportation demand and environment increasingly prominent (Liu et al., 2020). Therefore, systematically analyzing the 'Dual carbon' goals on the carbon emission constraints of transportation in the Pearl River Delta region, and deeply studying the carbon emission constraint mechanism of transportation in the region under the 'Dual carbon' goals are of practical significance to promote the development of transportation industry while realizing the 'double carbon' goal, so as to promote the high-quality economic development of the Pearl River Delta. Currently, the international community generally takes the vigorous development of public transport as an important means to deal with climate change. For example, the proportion of public transport travel in foreign big cities such as New York, London, Tokyo and Singapore has reached more than 80%, and this proportion is still relatively low in Chinese cities (Zhang et al., 2018). Liu et al. believe that the implementation of policies such as rail transit, new energy vehicles and intensive land use can promote the further optimization of urban transportation structure and the realization of carbon neutrality goals (Liu, Zhao et al., 2021). Liu et al. believed that the lowcarbon multimodal transport mode can rely on the technical and economic advantages of integrating multiple transportation modes, and can significantly improve the enterprise's low-carbon environmental protection, transportation organization level, cost reduction and efficiency (Liu, Shangguan et al., 2021). David et al. took the transportation carbon emissions of the European Union, the United States, China, and India as the research objects, and analyzed and measured the difficulty of each country in achieving the transportation emission reduction targets (Stern et al., 2012). Ren et al. extended STIRPAT model to analyse pathways for carbon emissions reduction in Beijing for the 'Dual Carbon' targets (Ren et al., 2022), and He et al. Use the IPAC-AIM model to analyse the carbon emissions reduction in Beijing and the results show that Beijing can achieve the goals of carbon neutrality (He et al., 2019). Papagiannaki et al. studied the influencing factors of vehicle carbon emissions in Greece and Denmark through the LMD method (Papagiannaki & Diakoulaki, 2009). Zhang et al. used the ESDA method to study the temporal and spatial distribution pattern of transportation carbon emissions, and constructed a GWR model . It was found that the transportation carbon emissions space in China's provinces from 2000 to 2013 showed certain clustering characteristics, but little change; urbanization rate and transportation structure are the main driving factors. Similarly, some scholars have analyzed the impact of the driving factors of carbon emissions in the transportation of China (Xu & Lin, 2015). The results showed that energy efficiency had a dominant effect on reducing carbon emissions, and urbanization was a factor that cannot be ignored in affecting carbon dioxide emissions. Based on the STIRPAT model and the ridge regression method, Wu et al. found that the urbanization development level was a major factor leading to the growth of transportation carbon emissions, and the per capita GDP had little impact (Wu et al., 2015). Through the dynamic VAR (Vector AutoRegression) method, Xu studied the driving factors of carbon dioxide emissions in the transportation of China, and believed that energy efficiency and urbanization had a significant impact on carbon dioxide emissions (Xu & Lin, 2015). After a regression study on the passenger and cargo turnover and transportation carbon emissions of different transportation modes in China, the United States, the European Union and Japan, Chai concluded that in the EU and Japan, the impact of transportation structure on traffic carbon emission was the most obvious, while China has a lot of space to optimize transportation structure for carbon emission reduction (Chai et al., 2017). Preeti and Suresh constructed a fuel composition framework of active structural energy intensity to study the energy consumption and transportation carbon emissions of road transport in Delhi, and obtained the optimal scenario based on carbon emission reduction targets through scenario analysis (Aggarwal & Jain, 2016). Zhao used ASIF methodology and the principle of index factor analysis to quantitatively analyze the impact of changes in freight structure on transportation carbon emissions (Zhao et al., 2012). Ma used the principal component analysis method to analyze the influencing factors of transportation carbon emissions in Guangdong Province, and believed that the economic development level, urban transportation scale and transportation capacity played a major role in the carbon emissions of urban passenger transportation in Guangdong Province (Ma et al., 2018). Taking the EU as an example, Zhang et al. analyzed the differences in the realization path of the 'Dual carbon' goals of the international community. Then they comprehensively applied a variety of quantitative analysis methods to systematically study the carbon emission situation at the provincial level in China, and finally put forward the idea of regional differentiation path to achieve the 'double carbon' goal (Zhang & Bai, 2021). Combined with the relevant theories of sustainable development transportation, low-carbon transportation and green circular transportation, Hu et al. constructed an evaluation index system of green transportation development based on PSR model, and conducted a case study on the evaluation of green transportation development in Jiangxi Province by using AHP method and comprehensive evaluation method (Hu et al., 2017).
In this paper, the DPSIR model is used to select the 'Dual carbon' goal constraints on transportation carbon emissions in the Pearl River Delta from five subsystems: driving force, pressure, state, influence and response. The DEMATEL method is used to calculate the centrality and causality of the relevant constraints, and the comprehensive influence degree and logical relationship between each constraint factor are analyzed. Combined with the ISM method, a multi-level hierarchical structure model of related factors is constructed to analyze the influence level, influence path and action mechanism between the constraints, so as to deeply explore the development of transportation in the Pearl River Delta under the 'twocarbon' goal.

Relationship between economic development and transportation carbon emission in the Pearl River Delta
As one of the most open and economically dynamic regions in China, the Pearl River Delta urban agglomeration has an obvious trend of mutual promotion, integration and interaction between transportation and economic development (Fan, 2018). Taking Guangzhou, the core city of the Pearl River Delta as an example, according to the 2019 'Guangzhou Yearbook', the main indicators of transportation in the whole society are 1.362 billion tons of the completed freight volume and 2.182914 trillion tons of the cargo turnover, with an increase of 6.6% and 1.6% over 2018, respectively. The annual port cargo throughput was 626.873 million tons, with an increase of 12.6% over 2018. The passenger throughput of Guangzhou Baiyun International Airport was 73.3861 million person-times, and the airport cargo and mail throughput was 2.5485 million tons, with an increase of 5.2% and 2.2% respectively over 2018. In the whole year, the income of postal business was 68.384 billion yuan, with an increase of 32.0% over 2018, and the total postal business was 135.145 billion yuan, with an increase of 30.0% over 2018. In the same year, the GDP of Guangzhou was 2362.86 billion yuan, with an increase of 6.8% over 2018. The main indicators of transportation have increased synchronously with the total economy. The development of transportation in the Pearl River Delta has become an important support to drive the regional economic and social development. At the same time, with the economic and social development of the Pearl River Delta, people's demand for transportation is still growing. Based on 2013, it is estimated that by 2030, the demand for intercity passenger transport, urban passenger transport and freight will increase by 2.3, 1.6 and 1.8 times, respectively (Energy Research Institute of National Development and Reform Commission, 2017). However, as an important source of urban carbon emissions, transportation carbon emissions will be more prominent in 2020. Affected by the outbreak of New Coronavirus (COVID-19), countries in the world have adopted restrictive mobility policies to varying degrees, which greatly reduced the activity of the global economy. The carbon dioxide emissions were reduced by 4%-7% in 2020 compared with 2019 (Le Quéré et al., 2020;O'Neill, 2020;Ula, 2020). Moreover, with the economic development, according to the prediction of the World Energy Council, the proportion of carbon emissions from world transportation will reach 50% and 80% in 2030 and 2050, respectively (International Energy Agency, 2009). Generally, due to the close relationship between the economic development of the Pearl River Delta and the transportation industry, if effective emission reduction measures are not taken for the constraints of transportation carbon emissions, the transportation carbon emissions of the Pearl River Delta are likely to increase in the future, which will also put pressure on the realization of the 'Dual carbon' goals of transportation.

Identification and selection of constraints
Since the 'Dual carbon' goals has a complex system with multiple levels, wide goals and dynamics for the constraints of transportation carbon emissions, according to the principles of the DPSIR model, the Diving force, Pressure, State, Impact and Response are used to identify and select the constraints. The interaction between transportation activities and resources and environment is comprehensively analyzed, and the causal relationship and action mechanism between the causes and results of environmental problems is dynamically reflected to reveal the 'response' of transportation carbon emissions to the 'Dural carbon' goal. The connotation analysis of the five subsystems is: Driving force refers to the pressure caused by the daily needs of human society, economy or culture on the ecosystem, mainly manifested in social progress, economic development, urbanization development, motorization level improvement and pressure from population growth. Pressure refers to the human activities that actually affect the low-carbon transportation system under the driving force, such as the scale and demand of transportation, car ownership, energy consumption, exhaust emission, noise pollution and the degree of road facilities. State refers to the situation of low-carbon transportation system after being affected by pressure, which is manifested as low-carbon transportation. Impact refers to the resource and environmental pollution or social and economic losses caused by the system state in turn, specifically referring to the impact of low-carbon transportation systems and transportation concepts on social and economic systems, energy and environmental systems, residents' daily life, and public physical and mental health. Response refers to the measures taken by human beings to promote the development of low-carbon transportation, mainly manifested as the improvement measures at the institutional, technical, cognitive and structural levels, which can affect the driving force, pressure, state and impact. The process is shown in Figure 1.
Based on the five subsystems of the driving force, pressure, state, impact and response, through summarizing the relevant literature, combined with the '14th Fiveyear Plan for Guangdong Comprehensive Transportation System' and the characteristics of the Pearl River Delta Tian et al., 2020;Zhang et al., 2019), a total of 25 constraints are identified and selected according to the frequency statistical method, as shown in Table 1.

Construction of hierarchical structure model of transportation carbon emission constraints based on DEMATEL-ISM integration method
DEMATEL method evaluates the relationship and influence of constraints by analyzing the constraints of uncertain relationship in the system. The ISM method decomposes the complex system into different levels and simplifies the system. By layering and ordering the intricate constraints, the influencing levels and paths between the constraints are analyzed (Kim & Nguyen, 2021;Venkatesanarayanan et al., 2021). Combining the advantages of the two methods, this paper comprehensively analyzes the constraint mechanism of transportation carbon emissions in the Pearl River Delta under the 'double carbon' goal from the influence level, path and degree of each constraint factor.

Calculate the combined impact matrix and the overall impact matrix
Through the column sum maximum method, formula (1) is used to normalize the direct impact matrix D to obtain the normalized impact matrix C. C = D max 1≤i≤n n j=1 d ij i = 1, 2, · · ·, n; j = 1, 2, · · ·, n (1) Based on the normalized impact matrix C, considering the direct and indirect impact degree between various constraints, the comprehensive impact matrix S is obtained.
where, I is the corresponding order identity matrix, and the comprehensive impact matrix S is obtained as S = (s ij ) n×n , as shown in Table 3. Through formula (3), the overall impact matrix G is constructed, which is used to make up for the comprehensive impact matrix S that only reflects the relationship between different constraints and fails to consider the influence of constraints on itself.

Calculate the influence degree, influenced degree, centrality and cause degree of each constraint
The influence degree y i is the column sum in the comprehensive impact matrix S, which represents the value of the comprehensive impact of a certain constraint on other constraints. y i = n j=1 s ij i = 1, 2, · · ·, n; j = 1, 2, · · ·, n The influenced degree b j is the column sum in the comprehensive impact matrix S, which represents the value of the comprehensive impact of a constraint by other constraints. b j = n i=1 s ij i = 1, 2, · · ·, n; j = 1, 2, · · ·, n The centrality z i reflects the total degree that the constraint affects other constraints and is affected by other constraints, and represents its importance in the system.
The degree of cause q i is used to judge the causal relationship between the constraint factors. The positive constraint factor is the cause, and otherwise it is the result.

Build the reachability matrix
The reachability matrix K expresses the degree of reachability from one constraint to another constraint through direct or indirect influence relationship in matrix form.
The ISM based on the DEMATEL method can use formula (8) to calculate the reachability matrix K based on the overall influence matrix G. The threshold is calculated mathematically, that is, the threshold is equal to the sum of the mean and standard deviation of the elements in the matrix S (Zhang & Wang, 2020). After calculation, the threshold is 0.075, and the calculation results are shown in Table 5.

Build a system hierarchy model
The constraints are layered according to the reachability matrix K to determine the system hierarchy model. The reachability matrix K is decomposed by the method of extracting top-level variables. Set that P(k ij ) is the reachable set, that is, the set of all constraints that can be reached from the constraint k ij ; F(k ij ) is the antecedent set, that is, the set of all constraints that can reach the constraint k ij . When formula (8) is satisfied, the constraints corresponding to all elements in P(k ij ) are the top layer, then delete the rows and columns where the corresponding constraint factors are located, and repeat this step until the bottom layer.
According to the above calculation and analysis, the hierarchical structure model of transportation carbon emission constraints in the Pearl River Delta region under the 'Dural carbon' goal is obtained as shown in Figure 2.

Analysis on the constraint mechanism of transportation carbon emissions in the Pearl River Delta region under the 'Dural Carbon' goal
The hierarchical structure model of transportation carbon emission constraints in the Pearl River Delta under the 'Dual carbon' constructed by the DEMATEL-ISM integrated method divides the 25 constraints into 6 levels. According to the hierarchical structure model, the constraint mechanism of transportation carbon emissions in the Pearl River Delta is analyzed from three aspects: the correlation and importance of constraints, the hierarchical structure of constraint system and constraint path.

Correlation and importance analysis of constraints
According to the calculation results in Table 4, from the perspective of the correlation with other constraints, the order from the largest to the smallest is: economic development level, carbon emission scale, 'Dual carbon' goals, carbon emission structure, transportation scale, vehicle ownership, renewable energy transportation infrastructure construction, vehicle growth, proportion of new energy vehicles, green travel service system, air quality, emission standards, investment in new energy technologies, total economic output, bus travel priority concept, transportation demand, transportation accessibility, intelligent transportation, public transport as a share of motorized travel, urbanization level, fuel standards, transportation environmental risks, population density, per capita GDP, and travel mode and rightof-way allocation. Among them, the economic development level, carbon emission scale and the centrality of 'Dual carbon' goals rank among the top three in Table 4, indicating that the above three constraints are closely related and highly correlated with other constraints. They are the core of the whole system and the key concern of the carbon emission constraint mechanism of transportation in the Pearl River Delta.
The influence degree and influenced degree of the urbanization level, travel mode and right-of-way allocation, transportation environmental risks, bus travel priority concept, economic development level and motor vehicle ownership are significantly different. The first four have a significantly greater influence degree than the influenced degree, and they mainly affect other constraints, while the latter two are more likely to be influenced.
'Dual carbon' goals, transportation environmental risks, bus travel priority concept, urbanization level, air quality, motor vehicle growth, travel mode and right-of-way allocation, emission standards, economic aggregate, transportation accessibility, transportation demand, public transport as a share of motorized travel, carbon emission structure and green travel service system are greater than 0. It shows that the above constraints play a causal role in the system and are the cause factors, and the others are the result factors.

Analysis of the system hierarchy structure of the constraints
The first and second layers are the direct constraints of the system, including emission standards, economic aggregate, transportation demand, per capita GDP, proportion of new energy vehicles, investment in new energy technologies, carbon emission scale, vehicle ownership, urbanization Level, renewable energy transportation infrastructure construction, population density, transportation scale and economic development level. It can be seen that the construction of renewable energy transportation infrastructure such as charging piles and the increase in investment of new energy technologies have improved the safety, reliability and convenience of new energy vehicles. This will help enhance people's recognition of new energy vehicles, increase the proportion of new energy vehicles, and directly affect the scale of transportation carbon emissions.

Constraint path analysis
There are multiple constraint paths in the hierarchical structure model, among which the dominant is the constraint path starting from the lowest constraints: Bus travel priority concept → public transport as a share of motorized travel → travel mode and right-of-way allocation → green travel service system → air quality→ carbon emissions structure (in combination with the 'Dual carbon' goals) →fuel standards→ emission standards and proportion of new energy vehicles. This is the main path for the development of the transportation industry under the 'Dual carbon' goals. By developing public transportation, strengthening environmental monitoring, improving protection, raising emission standards, controlling the traditionally powered motor vehicles, and promoting the new energy vehicles, they can provide a reference path for the development of the transportation industry in the Pearl River Delta under the new situation.
There are five constraint paths starting from indirect constraints, such as traffic accessibility→vehicle growth volume→transportation scale→motor vehicle ownership and transportation demand. The convenience of traffic accessibility will affect the motor vehicles and the transportation demand affects transportation carbon emissions. Smart transportation → economic development level → economic aggregate and motor vehicle ownership. The improvement of circulation efficiency brought about by the development of smart transportation will promote economic development and increase transportation demand.

Conclusion
In this paper, the DPSIR model is used to identify and select the 'Dual carbon' goals for the constraints of transportation carbon emissions in the Pearl River Delta. According to the experts' scores, the influence intensity of 25 constraints is assigned, and the DEMATEL-ISM method is used to analyze and evaluate the relationship and influence of the constraints. By layering and organizing the complex constraints, the influence levels and paths among the constraints are obtained, and the constraint mechanism of transportation carbon emissions in the Pearl River Delta under the 'Dual carbon' goals is analyzed. The specific conclusions are: First, the constraint hierarchical model divides 25 constraint factors into 6 levels. The first and second layers are the direct constraints of the system, which can significantly affect the 'Dual carbon' goals in a short period of time; the third and fourth layers are indirect constraints of the system, which need to be controlled; the fifth and sixth layers are the deep constraints of the system, which are root constraints and require long-term monitoring. Second, the top three constraints of centrality are the economic development level, carbon emission scale, and the 'Dual carbon' goal, which are the core factors of the entire system and should be concentrated. Third, through constraint path analysis, developing public transportation and smart transportation and increasing investment in new energy infrastructure and technology will help the development of transportation in the Pearl River Delta and the realization of the 'Dual carbon' goals.

Disclosure statement
No potential conflict of interest was reported by the author(s).