Forecast of Transportation CO₂ Emissions in Shanghai under Multiple Scenarios

Abstract

A reduction in CO₂ emissions from transportation is of great significance to achieve the goal of “peak carbon and carbon neutrality” in China. For 2003–2019, this paper calculates the transportation CO₂ emissions in Shanghai and constructs an extended STIRPAT (Stochastic Impacts by Regression on Population, Affluence, and Technology) model for forecasting. The result shows that from 2003 to 2019, total and per capita CO₂ emissions from Shanghai’s transportation sector increased, but the rate of growth decreased. Oil consumption was the main source of emissions, accounting for more than 92%. The study extended the STIRPAT model to analyze the driving factors for emissions. It shows that population size, passenger turnover, per capita GDP, transportation intensity, and energy intensity are positively correlated with emissions. Energy structure (the proportion of clean energy) has a negative impact, restraining growth. Under multiple scenarios, the forecast shows that Shanghai’s transportation sector can reach a CO₂ emissions peak before 2030. However, overgrowth of the transportation sector should be avoided. Progress in green and low-carbon technology is particularly important to achieve China’s peak carbon goal. Shanghai should actively build an efficient green transportation system, continue to optimize the transportation energy structure, and promote green and low-carbon travel for residents.

1. Introduction

The issue of carbon emissions has attracted wide attention around the world. In 2019, China accounted for 28.8% of the world’s CO₂ emissions, making it the world’s largest carbon emitter [1]. At the 75th session of the United Nations General Assembly, China announced that it would strive to achieve the goal of peak carbon by 2030 and carbon neutrality by 2060. How to achieve the goal of “peak carbon and carbon neutrality” is of great concern in China.

At present, scholars are still debating whether China can achieve the goal of peak carbon. Through forecasting models or scenario analysis, some scholars believe that peak carbon can be achieved before 2030 [2,3,4,5,6,7]. Others argue that it is unlikely to achieve this goal [8,9,10,11]. Therefore, more studies are needed to discuss this issue, especially from a regional or industry perspective. Transportation is one of the top three carbon-emitting industries, and thus it is a key component of peak carbon and carbon neutrality in China. A forecast of transportation CO₂ emissions can help us understand the development trend and support the policy making of carbon emission reduction. Researchers have forecasted China’s transportation CO₂ emissions across the country [12,13,14] for provinces such as Jiangsu [15,16], Hubei [17,18,19], Hebei [20], Shandon [21], Shaanxi [22], Qinghai [23], Jilin [24], and Hainan [25] and cities such as Beijing [26,27] and Tianjin [28] using the Kaya model [13], the STIRFDT(Stochastic Impacts by Regression on Population, Affluence, and Technology) model [1,17,18,23,29,30,31], the LEAP (Long-range Energy Alternatives Planning System) method [32], the linear regression method [14], the gray model [20,22,33,34], the LMDI (Logarithmic Mean Divisia Index) method [35,36], the machine-learning method [15,27,37], the system dynamics method [28,38,39,40], and so on. These methods are often used in combination with a scenario analysis to make predictions [1,13,16,17,18,19,23,29,32,38].

Moreover, few studies have focused on the forecast of transportation CO₂ emissions in Shanghai. Shanghai’s economic development and environmental protection have been excellent in China. In 2012, Shanghai was selected as part of the second group of pilot low-carbon cities. However, as a megacity in China, the demand for transportation in Shanghai is increasing annually. It has led to increasing energy consumption and CO₂ emissions. Therefore, the development of transportation CO₂ emissions is crucial to the construction of a low-carbon city and the achievement of peak carbon in Shanghai. Yet previous studies have only focused on the calculation of and influential factors in Shanghai’s transportation CO₂ emissions between 2000–2010 [41], 1998–2012 [42], and 2001–2015 [43], without making predictions.

Due to the limitation of the study period, previous studies could not reflect on the changes in Shanghai’s transportation CO₂ emissions in recent years, nor could they be used to forecast peak carbon. Therefore, in this paper, we have three objectives: (1) to update the data to 2019, calculating Shanghai’s transportation CO₂ emissions and summarizing the changing characteristics; (2) to analyze the driving factors and establish the forecast model by using the extended STIRPAT model; (3) to set multiple scenarios and forecast Shanghai’s transportation CO₂ emissions and estimate the time to and value of peak carbon.

The contributions of this study are as follows. (1) The issue of China achieving the goal of peak carbon before 2030 is controversial. Transportation is one of the three top industries producing CO₂ emissions. Shanghai is an important megacity to build a low-carbon city and achieve the goal of peak carbon. However, the empirical research on Shanghai’s transportation CO₂ emissions is limited. This study can provide new evidence to answer this question from an industry and regional perspective. (2) The study expands the STIRPAT model from the three variables of population, affluence, and technology to six variables of population size, passenger turnover, per capita GDP, transportation intensity, energy intensity, and energy structure. In addition, in order to eliminate the influence of multicollinearity among variables, the ridge regression method was selected to establish a forecast model. (3) The driving factors in the forecast model are set at low, medium, and high change rates, regarding the current situation of Shanghai’s economic development and government planning, which provides a comprehensive scenario analysis and decision-making basis for low-carbon transportation construction in Shanghai. Shanghai can also be used as a reference for other cities in China.

2. Methods and Data Sources

2.1. Calculation of CO₂ Emissions

As mentioned above, the (IPCC) inventory method is widely used to calculate CO₂ emissions. Specifically, there are two ways to calculate emissions. The bottom-top method calculates the total fuel consumption by multiplying the mileage of various modes of transportation by the fuel consumption per kilometer; the CO₂ emissions are then obtained by multiplying this with the fuel CO₂ emissions coefficient. In contrast, the top-bottom method calculates the CO₂ emissions by multiplying the CO₂ emissions coefficient of each fuel by the national or regional transportation fuel consumption [44]. The formula for CO₂ emissions caused by transportation energy consumption is as follows:

$$\mathrm{C}{\mathrm{O}}_2={\displaystyle \sum _{i=1}^n{E}_i\times F_i}$$

(1)

where CO₂ is carbon dioxide emissions, and E_i is the energy consumption level, that is, the consumption level of fuel (i) that is directly related to CO₂ emissions, where i refers to the type of transportation energy consumption. Considering fossil fuel consumption and data availability, this paper selects eight energy types: raw coal, gasoline, kerosene, diesel, fuel oil, liquefied petroleum gas, natural gas, and electricity. Energy consumption data are mainly taken from the Regional Energy Balance Sheet (Shanghai) in the China Energy Yearbook.

F_i is the CO₂ emissions coefficient, which is the amount of carbon dioxide released per unit of fuel (i). F_i can be calculated using formula (2):

$$F_i=Q_i\times C_i\times a_i\times h$$

(2)

where Q_i is the low calorific value of fuel (i) expressed in kJ/fuel (i) units. The unit of solid and liquid fuel is kg, and the unit of gas fuel is m³. C_i is the calorific value of fuel (i), the mass of carbon in the fuel. a_i is the carbon oxidation rate of fuel (i). h is the molecular weight ratio of carbon dioxide to carbon, which is 44/12.

Calculated according to formula (2), the CO₂ emissions coefficients of various fuels in the transportation sector are shown in

Table 1. CO₂ emission coefficient of energy in the transportation sector.

Types of Energy	Standard Coal Coefficient [1] (kg.ce/kg)	Low Calorific Value [2] (kJ/kg)	Carbon Content Per Unit Calorific Value [3] (ton.c/TJ)	Carbon Oxidation Rate [4]	CO2 Emission Coefficient (kg.CO2/kg)
raw coal	0.7143	20,908	26.37	0.94	1.9003
gasoline	1.4714	43,070	18.9	0.98	2.9251
kerosene	1.4714	43,070	19.6	0.98	3.0179
diesel	1.4571	42,652	20.2	0.98	3.0959
fuel oil	1.4286	41,816	21.1	0.98	3.1705
liquefied petroleum gas	1.7143	50,179	17.2	0.98	3.1013
natural gas	1.3300 Kg.ce/m3	38,931 KJ/m3	15.3	0.99	1.9770 kg.CO2/m3
electricity (indirect emissions) [5]					0.7035 kg.CO2/kw.h

Note: Fuel with a low calorific value equal to 29,307 kJ is called 1 kg standard coal (1 kg.ce). Source: Data in [1] and [2] are from the China Energy Statistical Yearbook in 2020. Data in [3] and [4] are from the 2011 Provincial GHG Inventory Compilation Guide. Data in [5] are from the 2012 Average Carbon Dioxide Emission Factor of China’s Regional Power Grid (East China data) released by the National Development and Reform Commission.

.

2.2. Extended STIRPAT Model

The IPAT model was proposed in the 1970s to evaluate the impact of human activities on the environment. It can specifically analyze the impact of the population (P), affluence (A), and technology (T). However, the IPAT model is an identity model, assuming that there are no random factors. Further, in real scenarios, most of the influencing factors are non-monotonic and non-linear. To overcome the limitations of the IPAT model, York et al. [45] modified and extended it and proposed the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) model. The STIRPAT model is shown in formula (3):

$$I=a\times P^b\times A^c\times T^d\times e$$

(3)

Here, I is defined as environmental pressure, P is defined as the population size, A is defined as the degree of affluence, T is defined as technical level, a is the model coefficient, and b, c, and d are the indices of each influencing factor. If a = b = c = d = e = 1, the STIRPAT model is an IPAT model. Taking the logarithm of both sides of formula (3), we obtain:

$$InI=Ina+bInP+cInA+dInT+Ine$$

(4)

The STIRPAT model rejects the assumption of unit elasticity and increases randomness, which is convenient for empirical analysis. The advantage of the STIRPAT model is that it not only allows the coefficient of each variable to be estimated as a parameter, but also allows the appropriate decomposition and improvement of each variable. Therefore, the STIRPAT model is most widely used to quantify the relationship between carbon emissions and driving factors and to study the issue of peak carbon.

Based on previous studies, we extended the three factors of the STIRPAT model to six factors.

(1)

Population

Population size, age, gender, education level, and other demographic characteristics, as well as urbanization and other factors, will have different degrees of impact on traffic demand, affecting traffic carbon emissions. Two variables, population size and passenger turnover, are selected to represent the factor of Population (

Table 2. Description of STIRPAT model variables.

Variable	Symbol	Indicator Description	Unit
population size	P	the resident population at year-end	ten thousand people
passenger turnover	T	regional population/resident population at year-end	million passenger-kilometer
per capita GDP	G	number of passengers × transport distance	Yuan/person
transportation intensity	S	transportation GDP/regional GDP	%
energy intensity	E	energy consumption per unit of transportation GDP	tons/ten thousand Yuan
energy structure	N	clean energy consumption/total energy consumption	%

).

(2)

Affluence

Economic development, residential consumption, and industrial structure will cause changes in transportation demand and affect transportation carbon emissions. Two variables, per capita GDP and transportation intensity, are selected to represent the factor of Affluence (Table 2).

(3)

Technology

Energy technology will promote the improvement of energy efficiency, change the energy structure, and reduce energy consumption and carbon emissions. Two variables, energy intensity and energy structure, are selected to represent the factor of Technology.

Therefore, we can obtain an extended STIRPAT model, which can be expressed as:

$$InC=Ina+b_{1}InP+b_{2}InT+b_{3}InG+b_{4}InS+b_{5}InE+b_{6}InN+Ine$$

(5)

Here, C is for CO₂ emission, P is for population size, T is for passenger turnover, G is for per capita GDP, S is for transportation intensity, E is for energy intensity, N is for energy structure, e is for random error of the model, a is for the constant of the model, and b₁, b₂, b₃, b₄, b₅, and b₆ represent the regression coefficients of each variable (reflecting the elastic relationship between each factor and CO₂ emissions). Assuming other variables remain unchanged, changes of 1% in P, T, G, S, E, and N will cause changes of b₁%, b₂%, b₃%, b₄%, b₅%, and b₆% in C, respectively.

2.3. Scenario Analysis Method

The scenario analysis method is used to forecast the development of Shanghai’s transportation CO₂ emissions. The six driving factors in the forecast model are set at low, medium, and high change rates, regarding the current situation of Shanghai’s economic development and government planning. The latter includes the Outline of The 14th Five-Year Plan for Shanghai National Economic and Social Development and the long-term goals for 2035 (hereafter referred to as The Outline); the Shanghai 7th National Population Census Main Data Bulletin; the Shanghai Master Plan 2017–2035; the 14th Five-Year Plan for Shanghai Energy Development; the 14th Five-Year Plan for Shanghai Resource Conservation and Circular Economy Development; and the 14th Five-Year Plan for Shanghai Comprehensive Transportation Development. Due to the impact of COVID-19, the driving factors are different from historical patterns. Therefore, the transportation CO₂ emissions in 2020 are not used.

(1)

Population size (P)

The Shanghai 7th National Population Census Main Data Bulletin shows that the Shanghai permanent resident population was 24,870,895 people on November 1, 2020. Compared with the 2,3019,196 people in the sixth National Census, the total number increased by 1,851,699 people in ten years, an increase of 8.0%. During the 13th Five-Year Plan period, the average change rate of the permanent resident population in Shanghai was 0.16%. In the Shanghai Master Plan 2017–2035, it is proposed that “in order to alleviate the contradiction between rapid population growth and tight resources and environment constraints, the permanent population should be controlled within 25 million by 2020, and the target of permanent population control should be around 25 million by 2035.” As such, Shanghai’s growth will be limited in the future.

Based on this, at the medium rate, the average change rate of Shanghai population size is set at 0.15% in 2021–2025, 0% in 2026–2030, and 0% in 2031–2035. The average change rate of Shanghai population size at the low or high rate is adjusted based on the medium rate. (The average change rate of each period decreases by 0.05% at the low rate and increases by 0.05% at the high rate.) Therefore, at the low rate, the average change rate of Shanghai’s population size is set at 0.1% in 2021–2025, −0.05% in 2026–2030, and −0.05% in 2031–2035. At the high rate, the average change rate of Shanghai’s population size is set at 0.2% in 2021–2025, 0.05% in 2026–2030, and 0.05% in 2031–2035 (

Table 3. Setting of change rate.

Rate Mode	Period	Setting of Change Rate
		P	T	G	S	E	N
Low	2021–2015	0.10%	3.70%	8.00%	0.50%	−3.80%	3.70%
	2026–2030	−0.05%	9.50%	6.50%	0.00%	−3.60%	7.00%
	2031–2035	−0.05%	8.00%	5.00%	−0.50%	−3.40%	10.00%
Medium	2021–2015	0.15%	5.70%	9.00%	0.70%	−4.00%	9.00%
	2026–2030	0.00%	11.50%	7.50%	0.20%	−3.80%	12.00%
	2031–2035	0.00%	10.00%	6.00%	−0.30%	−3.60%	15.00%
High	2021–2015	0.20%	7.70%	10.00%	0.90%	−4.20%	15.00%
	2026–2030	0.05%	13.50%	8.50%	0.40%	−4.00%	18.00%
	2031–2035	0.05%	12.00%	7.00%	−0.10%	−3.80%	21.00%

).

(2)

Passenger turnover (T)

Passenger turnover is the indicator reflecting transportation workload and is equal to the number of passengers multiplied by the distance traveled, expressed in passenger–kilometer. In Shanghai’s Master Urban Plan (2017–2035), it is proposed that Shanghai should build a Comprehensive Transportation System that is “Safe, Convenient, Green, Efficient, and Economic” and improve the service of public transportation in the inner areas of the main city. By 2035, public transportation will account for over 50% of all means of transportation, green transportation will account for 85%, and 60% of rail transit stations in the inner areas of the main city will have 600 m of land coverage. It can be inferred that under normal circumstances, Shanghai passenger turnover will continue to increase.

The growth rate of Shanghai passenger turnover during the 13th Five-Year Plan period (the first four years, 2016–2019) is 54.19%, and the average annual change rate is 11.43%. However, due to COVID-19, Shanghai passenger turnover in 2020 was 134.877 billion passenger kilometers, down 47% from 2019. According to the latest data from the Shanghai Municipal Bureau of Statistics, Shanghai passenger turnover dropped 54.9% year-on-year from January to June 2022. It can be inferred that due to the impact of COVID-19, the annual growth rate of passenger turnover will decline significantly from 2021 to 2025, and then may return to the normal growth rate.

According to the above analysis, at the medium rate, the average change rate of Shanghai passenger turnover is about half that of the 13th Five-Year Plan period (the first four years) in 2021–2025, namely 5.7%. It is set at 11.5% in 2026–2030 and 10% in 2031–2035 (referring to the average change rate over the last 15 years, 10.12%). The average change rate of Shanghai passenger turnover at the low or high rate is adjusted based on the medium rate. (The average change rate of each period decreases by 2% at the low rate and increases by 2% at the high rate.) Therefore, at the low rate, the average change rate of Shanghai passenger turnover is set at 3.7% in 2021–2025, 9.5% in 2026–2030, and 8% in 2031–2035. At the high rate, the average change rate of Shanghai passenger turnover is set at 7.7% in 2021–2025, 13.5% in 2026–2030, and 8% in 2031–2035 (Table 3).

If the factors of population, affluence, and technology are combined under different growth rates, 27 scenarios for the carbon emission forecast can be set. To simplify the analysis, five typical scenarios were selected, as shown in

Table 4. Types of Scenario.

Type of Scenario		P	T	G	S	E	N
Standard scenario	Scenario 1	M	M	M	M	M	M
Technical stability—High growth	Scenario 2	H	H	H	H	M	M
Technical stability—Low growth	Scenario 3	L	L	L	L	M	M
Technological breakthrough—High growth	Scenario 4	H	H	H	H	H	H
Technological breakthrough—Low growth	Scenario 5	L	L	L	L	H	H

.

(3)

Per capita GDP (G)

In 2019, China’s per capita GDP was 70,892 yuan, and Shanghai’s was 157,279 yuan. The average change rate of Shanghai’s per capita GDP from 2003 to 2019 was 9.09%, and it was 9.08% during the first four years of the 13th Five-Year Plan. According to The Outline, during the 14th Five-Year Plan (2021–2025), the average annual growth rate of Shanghai’s GDP should reach 5%. At the medium rate, the average change rate of Shanghai’s per capita GDP is set at 9% in 2021–2025, 7.5% in 2026–2030, and 6% in 2031–2035. The average change rate of Shanghai’s per capita GDP at the low or high rate is adjusted based on the medium rate. (The average change rate of each period decreased by 1% at the low rate and increased by 1% at the high rate.) Therefore, at the low rate, the average change rate of Shanghai’s per capita GDP is set at 8% in 2021–2025, 6.5% in 2026–2030, and 5% in 2031–2035. At the high rate, the average change rate of Shanghai’s per capita GDP is set at 10% in 2021–2025, 8.5% in 2026–2030, and 7% in 2031–2035 (Table 3).

(4)

Transportation intensity (S)

From 2003 to 2019, the average annual change rate of Shanghai’s transportation intensity was −0.35%, and it was 0.7% during the 13th Five-Year Plan period (the first four years). According to the Shanghai Master Urban Plan 2017–2035, by 2035, Shanghai will grow into a standout global city—a city of innovation, humanity, and sustainability, as well as a modern socialist international metropolis with world influence. International traveling passenger volume will reach 38%, and the rate of international container transit will be no less than 13%. Each new town with over 100,000 residents will have a metro station. Efforts will be made to reduce the average commute time to less than 40 minutes in central Shanghai.

At the medium rate, the average change rate of Shanghai’s transportation intensity is set at 0.7% in 2021–2025, 0.2% in 2026–2030, and −0.3% in 2031–2035. The average change rate of Shanghai’s transportation intensity at the low or high rate is adjusted based on the medium rate. (The average change rate of each period decreases by 0.2% at the low rate and increases by 0.2% at the high rate.) Therefore, at the low rate, the average change rate of Shanghai’s transportation intensity is set at 0.5% in 2021–2025, 0% in 2026–2030, and −0.5% in 2031–2035. At the high rate, the average change rate of Shanghai’s transportation intensity is set at 0.9% in 2021–2025, 0.4% in 2026–2030, and −0.1% in 2031–2035 (Table 3).

(5)

Energy intensity (E)

According to the Comprehensive Work Plan for Energy Conservation and Emission Reduction released by The State Council, during the 13th and 14th Five-Year Plans, the national energy consumption per unit of gross domestic product will decrease by 15% and 13.5%, respectively. From 2016 to 2019, Shanghai’s transportation energy intensity decreased by 15%, with an average annual change of −4.04%. According to the 14th Five-Year Plan for Shanghai Resource Conservation and Circular Economy Development, by 2025, energy consumption is intended to be reduced (per unit GDP) by 14%. At the medium rate, the average change rate of Shanghai’s transportation energy intensity is set at −4.0% in 2021–2025, −3.8% in 2026–2030, and −3.6% in 2031–2035. The average change rate of Shanghai’s transportation energy intensity at the low or high rate is adjusted based on the medium rate. (The average change rate of each period decreases by 0.2% at the low rate and increases by 0.2% at the high rate.) Therefore, at the low rate, the average change rate of Shanghai’s transportation energy intensity is set at −3.8% in 2021–2025, −3.6% in 2026–2030, and −3.4% in 2031–2035. At the high rate, the average change rate of Shanghai’s transportation energy intensity is set at −4.2% in 2021–2025, −4.0% in 2026–2030, and −3.8% in 2031–2035 (Table 3).

(6)

Energy structure (N)

According to the 14th Five-Year Plan for Shanghai Resource Conservation and Circular Economy Development, Shanghai should vigorously develop new energy transportation tools and facilities. By 2025, new-energy vehicles will be used in public buses, cruise taxis, official vehicles of Party and government organs, cargo vehicles in central urban areas, and postal vehicles, accelerating the replacement of stock, and striving to make non-fossil energy account for about 20% of total energy consumption. In addition, the China Energy Outlook 2030 points out that the growth rate of energy demand will continue to slow in the future, the diversification of transportation energy will accelerate, and the use of clean energy will increase by 17%–26%.

From 2003 to 2019, the proportion of clean energy in Shanghai’s transportation sector increased by 1.44%, with an average change rate of 3.69%. During the 13th Five-Year Plan period (the first four years), the average change rate was 1.46%. In 2019, the proportion of clean energy in Shanghai’s transportation sector accounted for only 3.26%, which should increase in the future. Therefore, the average change rate is set at 3.7% in 2021–2025, 7% in 2026–2030, and 10% in 2031–2035. The average change rate of Shanghai’s transportation energy structure at the medium or high rate is adjusted based on the low rate. At the medium rate, the average change rate of Shanghai’s transportation energy structure is set at 9% in 2021–2025, 12% in 2026–2030, and 15% in 2031–2035. At the high rate, the average change rate of Shanghai’s transportation energy structure is set at 15% in 2021–2025, 18% in 2026–2030, and 21% in 2031–2035.

(1)

Standard scenario

It is assumed that social, economic, transportation, and other aspects of Shanghai continue to develop steadily. The CO₂ reduction in the transportation sector has been carried out in an orderly way, and the CO₂ emissions goal can be achieved. Energy intensity and energy structure improve, but green and low-carbon technology progress is limited.

(2)

Technological stability—high growth scenario

Assuming that Shanghai’s population and economy grow rapidly, the development of the transportation sector increases pressure on CO₂ emissions. The Shanghai transportation sector can implement the task of carbon emission reduction. However, green and low-carbon technology have not made significant progress. Energy efficiency has not improved significantly.

(3)

Technological stability—low growth scenario

Assuming that Shanghai’s population size, economic output, and other variables grow slowly, it is conducive to a decline in carbon emissions. The Shanghai transportation sector can implement the targets of Shanghai’s carbon reduction policies. However, green and low-carbon technology have not made significant progress. Energy efficiency has not improved significantly.

(4)

Technological progress—high growth scenario

Assuming that Shanghai’s population size, economic output, and other variables grow rapidly, the development of the transportation sector increases the pressure on carbon emissions. There has been a significant breakthrough in green and low-carbon technology. The Shanghai transportation energy intensity is greatly reduced, and the proportion of clean energy has increased. Green and low-carbon technology have been applied to all aspects of urban development, which has effectively promoted the construction of a low-carbon city.

(5)

Technological progress—low growth scenario

Assuming that Shanghai’s population size, economic output, and other variables grow slowly, it is conducive to a decline in carbon emissions. There has been a significant breakthrough in green and low-carbon technology. Shanghai’s transportation energy intensity is greatly reduced, and the proportion of clean energy has increased. Green and low-carbon technology has been applied to all aspects of urban development, which has effectively promoted the construction of a low-carbon city.

2.4. Data Sources

The research period is from 2003 to 2019. The data on transportation energy consumption in Shanghai are taken from the Regional Energy Balance Sheet (Shanghai) in the China Energy Statistical Yearbook. Other data come from the China Statistical Yearbook and Shanghai Statistical Yearbook.

3. Results

3.1. Estimation of Shanghai’s Transportation CO₂ Emissions

The total transportation CO₂ emissions in Shanghai are obtained according to the formula (1). In addition, based on the population at the end of the year, the per capita transportation CO₂ emissions can be calculated, which is shown in

Table 5. CO₂ emissions of Shanghai transportation sector (2003–2019).

Year	Total CO2 Emissions (unit:104 tons)	Per capita CO2 Emissions (unit: ton)
2003	1958.32	1.14
2004	2577.23	1.48
2005	2914.54	1.64
2006	3416.92	1.88
2007	3872.01	2.08
2008	4006.72	2.12
2009	4053.47	2.11
2010	4315.95	1.94
2011	4211.54	1.79
2012	4283.37	1.80
2013	4289.49	1.78
2014	4280.69	1.76
2015	4480.36	1.86
2016	4976.51	2.06
2017	5440.75	2.25
2018	5361.18	2.21
2019	5573.05	2.30

.

The table shows that the total and per capita of transportation CO₂ emissions in Shanghai have been growing, but the growth rate has slowed significantly. The growth rate of CO₂ emissions has dropped from more than 30% during the early years under study to less than 5% in 2019.

Further, we find that oil consumption is the main source of Shanghai’s transportation CO₂ emissions, contributing more than 92%. Specifically, fuel oil accounts for the largest proportion, followed by kerosene, diesel, and gasoline, whereas natural gas and liquefied petroleum gas account for a relatively small proportion of emissions (

Table 6. Contribution of Shanghai’s transportation energy sources to CO₂ emissions (2003–2019). (unit:%).

Year	Raw Coal	Gasoline	Kerosene	Diesel	Fuel Oil	Liquefied Petroleum Gas	Natural Gas	Electricity
2003	1.39	5.57	15.49	13.88	59.00	0.36	0.14	4.16
2004	0.78	6.02	19.33	12.99	55.79	1.16	0.14	3.78
2005	0.79	6.16	19.18	10.02	59.39	0.86	0.14	3.46
2006	0.34	5.78	22.97	10.01	56.63	0.63	0.10	3.55
2007	0.29	5.74	22.90	10.47	56.40	0.54	0.10	3.56
2008	0.28	6.56	24.05	11.22	53.08	0.51	0.12	4.16
2009	0.26	6.77	26.19	11.98	49.67	0.50	0.14	4.49
2010	0.21	6.65	27.60	12.11	47.26	0.47	0.14	5.56
2011	0.13	7.58	28.36	13.34	43.54	0.49	0.15	6.41
2012	0.08	7.72	28.14	14.30	42.75	0.48	0.17	6.36
2013	0.08	7.98	30.52	14.07	40.06	0.45	0.23	6.62
2014	0.05	8.57	31.72	13.62	38.54	0.39	0.32	6.78
2015	0.03	7.88	34.35	13.92	36.45	0.39	0.36	6.61
2016	0.03	8.15	35.45	13.28	36.00	0.35	0.27	6.48
2017	0.00	7.07	36.14	11.48	38.34	0.27	0.25	6.44
2018	0.00	3.53	39.47	11.03	38.29	0.24	0.33	7.12
2019	0.00	3.33	40.78	11.44	36.68	0.24	0.32	7.20

and Figure 1).

3.2. Calculation of the Parameters in the Forecast Model

3.2.1. Ordinary Least Square Regression Analysis

The extended STIRPAT model of Shanghai’s transportation sector was analyzed by ordinary least square regression with SPSS software. As shown in

Table 7. Ordinary least squares regression results.

Variable	Coefficient	Sig.f	VIF
Constant	−6.881	0.000 **	-
InP	0.827	0.000 **	26.606
InT	0.004	0.926	108.237
InG	1.047	0.000 **	392.948
InS	1.119	0.000 **	96.203
InE	1.102	0.000 **	161.672
InN	0.127	0.132	139.663
Sample size	17
R2	1
Adjusted R2	0.999
F.	3340.749 (p = 0.000)
D-W	2.283

Note: ** significant at the 5% level.

, the result of the regression equation is:

$$\begin{array}{ll}InC& =-6.881+0.827InP+0.004InT+1.047InG+1.119InS\\ & +1.102InE+0.127InN\end{array}$$

(6)

The result shows that the F value is 3340.7, and the corresponding significance value is 0.000 (<0.001). In addition, the adjusted R2 is 0.999, close to one, indicating that the samples are all explained by the regression equation. The D-W (Durbin–Watson) test result is 2.28, indicating that the autocorrelation in the model is weak. However, the VIF (Variance Inflation Factor) value of the model is large, indicating that the model results have a multicollinearity problem. Only by eliminating the multicollinearity of the model can we ensure the accuracy of the relationship between carbon emissions and the driving factors. Therefore, the ridge regression method is used to analyze the data further.

3.2.2. Ridge Regression Analysis

Ridge regression is an improved least-squares method, which can solve the collinearity problem by bias–variance tradeoffs, improving the accuracy and reliability of parameter estimation. The Ridge regression method can eliminate the interference of multicollinearity by adding a non-negative factor K to the main diagonal of the normalized matrix of independent variables. Although the addition of K will reduce the goodness of fit of the model, the effectiveness and stability of the regression results will be improved.

(1)

Determine the ridge parameter

Before the ridge regression analysis, the ridge parameter K should be determined by combining the ridge trace map. The selection principle of K is the minimum K value when the standardized regression coefficient of each independent variable tends to be stable. The smaller the value of K, the smaller the deviation is. When the value of K is 0, it is an ordinary linear OLS regression. Using SPSS software, the ridge track map of the extended STIRPAT model of the CO₂ emissions can be obtained (Figure 2). When the K value is 0.01, the standardized regression coefficient of the independent variable tends to be stable, so the best K value is chosen as 0.01.

(2)

Ridge regression results

Now, we calculate the ridge regression analysis results at K = 0.01. First, an ANOVA test was performed, and the significance value was <0.001. The R² and adjusted R² were 0.996 and 0.994, respectively. The F value is 458.619 and the corresponding sig. value is 0.000, which indicates that the significance of the overall regression equation is good.

All parameters are shown in

Table 8. Ridge regression results.

Variable	Coefficient	Sig.f
Constant	−7.850	0.000 **
InP	1.267	0.000 **
InT	0.256	0.000 **
InG	0.368	0.000 **
InS	0.319	0.001 **
InE	0.278	0.001 **
InN	−0.316	0.000 **
Sample size	17
R2	0.996
Adjusted R2	0.994
F	458.619 (p = 0.000)

Note: ** significant at the 5% level.

, and the regression equation is:

$$\begin{array}{ll}InC& =-7.850+1.267InP+0.256InT+0.368InG+0.319InS\\ & +0.278InE-0.316InN\end{array}$$

(7)

The results show that the factors of population size, passenger turnover, per capita GDP, transportation intensity, and energy intensity all have a positive impact on Shanghai’s transportation carbon emissions. The factor of energy structure (the proportion of clean energy) limits Shanghai’s transportation carbon emissions. The influence of various factors on the total transportation carbon emissions, from large to small, is: population size, per capita GDP, transport intensity, energy structure (proportion of clean energy), energy intensity, and passenger turnover.

3.3. Forecast Model

Based on the above results, the forecast model for Shanghai’s transportation CO₂ emissions based on the extended STIRPAT model is:

$$\begin{array}{ll}C& =\exp (-7.850+1.267InP+0.256InT+0.368InG+0.319InS\\ & +0.278InE-0.316InN)\end{array}$$

(8)

To verify the forecast effect of the model, we simulated the historical data and compared the actual value with it. As can be seen from Figure 3, the model fits well.

3.4. Forecast under Multiple Scenarios

Under the above five scenarios, the amount of Shanghai’s transportation CO₂ emissions during 2021–2035 is forecast. The results are shown in Figure 4 and

Table 9. Forecast of Shanghai’s transportation CO₂ emissions (2021–2035). (Unit: 10⁴ tons).

Year	Scenario-1	Scenario-2	Scenario-3	Scenario-4	Scenario-5
2021	5689.74	5743.62	5635.67	5643.91	5537.84
2022	5755.26	5864.77	5646.39	5662.91	5452.05
2023	5821.52	5988.47	5657.12	5681.98	5367.58
2024	5888.55	6114.78	5667.88	5701.10	5284.43
2025	5956.35	6243.76	5678.65	5720.29	5202.57
2026	6007.42	6355.67	5674.07	5724.27	5110.39
2027	6058.92	6469.58	5669.49	5728.25	5019.84
2028	6110.86	6585.53	5664.91	5732.24	4930.90
2029	6163.25	6703.57	5660.34	5736.22	4843.53
2030	6216.08	6823.72	5655.77	5740.21	4757.71
2031	6157.47	6822.81	5549.69	5644.69	4591.40
2032	6099.42	6821.91	5445.61	5550.75	4430.90
2033	6041.91	6821.00	5343.47	5458.38	4276.02
2034	5984.94	6820.10	5243.26	5367.55	4126.54
2035	5928.51	6819.19	5144.92	5278.22	3982.29

.

(1)

Under scenario 1 (standard scenario), the characteristics of all factors are mostly the same as the current ones. Shanghai’s transportation CO₂ emissions peaked at 62,160,800 tons in 2030, about 1.12 times the emissions in 2019. Therefore, Shanghai’s transportation sector can achieve peak levels by 2030.

(2)

Under scenario 2 (technological stability - high growth), the factors of population and affluence will develop rapidly, and there is no obvious breakthrough in green carbon reduction technology. The results show that transportation CO₂ emissions in Shanghai will continue to increase until 2035. After 2030, the growth rate of emissions slow. In 2035, emissions will reach 68.23 million tons, about 1.22 times that of 2019. Under this scenario, Shanghai’s transportation sector is unable to achieve its goal by 2030.

(3)

Under scenario 3 (technological stability - low growth), the factors of population and affluence will develop slowly, and there is no obvious breakthrough in green carbon reduction technology. The results show that Shanghai’s transportation CO₂ emissions will increase slowly, reaching a peak of 56.56 million tons in 2030, which is about 1.02 times the value in 2019. Emissions will decline steadily after 2030. By 2035, they will fall to 92% of the value in 2019. In this case, Shanghai’s transportation sector achieves its goal by 2030.

(4)

Under scenario 4 (technological progress - high growth), the factors of population and affluence will develop rapidly, and there is a breakthrough in green carbon reduction technology. The results show that Shanghai’s transportation CO₂ emissions do not increase much, reaching the peak of 57.40 million tons in 2030, about 1.03 times the value in 2019. CO₂ emissions will then fall steadily. By 2035, they will fall to 95% of the value in 2019. In this case, Shanghai’s transportation sector achieves its goal by 2030.

(5)

Under scenario 5 (technological progress - low growth), the factors of population and affluence will develop slowly, and there is a breakthrough in green carbon reduction technology. The results show that Shanghai’s transportation CO₂ emissions fall until 2035. In 2035, the emissions will be only 72% of their value in 2019. Shanghai’s transportation sector can reach peak carbon.

In sum, Shanghai’s transportation sector can reach peak carbon before 2030. Scenario 3 and scenario 4 are most suitable for the development of Shanghai’s transportation sector. Moreover, it can be found that extensive growth of the transportation sector is unfavorable for achieving the goal of peak carbon, and the progress of green carbon reduction technology is of great significance.

4. Discussion

This study finds that Shanghai’s total and per capita transportation CO₂ emissions fluctuated from 2003 to 2019, and the growth rate slowed significantly. Specifically, it can be divided into three stages. It increased from 2003 to 2010. From 2011 to 2014, there was a slight decline and then a slow increase. From 2015 to 2019, the growth rate continued to increase. This is consistent with the conclusion of Zhou [43]. The results are slightly different from those of Chen [42] and Wu [41], which may be due to the different study periods and the different choices of energy types for emissions. Furthermore, our study finds that oil consumption is the main source of transport carbon emissions in Shanghai, which is consistent with the findings of other scholars. Zhou [43] found that the steady growth of fuel and kerosene was the main factor contributing to the growth of transportation emissions. Wu [41] found that petroleum consumption contributed the most to Shanghai’s transportation CO₂ emissions, for example, reaching 91.80% in 2010.

This study shows that population size is the most important factor in increasing Shanghai’s transportation CO₂ emissions. However, in recent years, the growth of the permanent resident population in Shanghai has slowed, and it is close to the upper limit planned by the Shanghai government, which is conducive to the slowdown of total CO₂ emissions. In addition, an increase in the proportion of clean energy can inhibit the growth of transportation CO₂ emissions. However, due to the low proportion of clean energy in Shanghai’s transportation sector, it is difficult to reduce CO₂ emissions. In recent years, Shanghai has actively promoted a new energy transportation policy, but the study results show that the policy has not played a significant role in reducing CO₂ emissions. This reflects the immature state of the public facilities system for new energy vehicles in Shanghai and the insufficient integration of public transportation and urban planning.

The effect of various factors on Shanghai’s transportation CO₂ emissions is as follows, from large to small: population size, per capita GDP, transport intensity, energy structure (proportion of clean energy), energy intensity, and passenger turnover. In other studies, because of the partial differences in factor selection, the ranking of the degree of influence is not the same. Zhou [43] found that, according to the degree of impact, the order of factors is population scale, freight turnover, passenger turnover, the proportion of secondary industries, the proportion of tertiary industries, and per capita GDP. Chen [42] found that the factors ranked as urbanization rate, per capita GDP, and per capita consumption expenditure.

The forecasts under different scenarios show that Shanghai’s transportation sector can achieve the CO₂ emissions peak in 2030. However, there are two issues to consider. First, it is important to control the growth rate of the transportation sector, especially to get rid of the extensive sector growth model. Second, Shanghai should actively promote the progress and application of green carbon reduction technology, continue to promote new energy vehicles, and improve the green public transport system. At present, there is no sufficient environmental support for green travel in Shanghai, and the construction of Shanghai’s international shipping center aggravates the pressure of CO₂ emissions in the transportation sector. It is necessary to implement combined policies to solve the problems of peak carbon in many aspects. If the above two issues do not perform well in the following years, Shanghai’s transportation sector may fail to achieve the goal of peak carbon by 2030.

This study has two limitations. First, due to insufficient statistical methods in China’s energy statistics, the available energy consumption data of Shanghai’s transportation sector is not complete, which may lead to an underestimation of CO₂ emissions. Second, since the beginning of 2020, the COVID-19 pandemic has greatly affected people’s normal lives. Transportation CO₂ emissions seriously deviate from the normal state. In order to construct the prediction function of Shanghai transportation CO₂ emissions under normal conditions, this paper calculated CO₂ emissions until 2019. In addition, we found the latest data until 2020 and could not compare the prediction and the real number in 2021. Future research can improve the forecast model and apply it to other regions based on more sufficient data.

5. Conclusions and Policy Implications

5.1. Conclusions

This study calculates the growth in transportation CO₂ emissions in Shanghai from 2013–2019, which is from 1958.32 × 10⁴ tons to 5573.05 × 10⁴ tons. We find that CO₂ emissions continued to increase, but the growth rate slowed down significantly. Further, oil consumption was the main source of Shanghai’s transportation CO₂ emissions. The study extended the STIRPAT model and identified the driving factors in emissions. The influence of various factors on the total transportation carbon emissions, from large to small, is: population size, per capita GDP, transport intensity, energy structure (proportion of clean energy), energy intensity, and passenger turnover. Based on the extended STIRPAT model, we forecasted Shanghai’s transportation CO₂ emissions during 2021–2035 under multiple scenarios. The forecast shows that it is very possible to reach peak carbon before 2030 in Shanghai’s transportation sector. Scenario 3 and scenario 4 are most suitable for the development of Shanghai’s transportation sector. Further, it should be noted that the extensive growth of transportation is unfavorable for achieving the goal of peak carbon, and the progress of green carbon reduction technology is of great significance.

5.2. Policy Implications

(1)

From the above forecast model, we can see that the transportation intensity has a positive impact on Shanghai transportation CO₂ emissions. A 1% reduction in transportation intensity would reduce CO₂ emissions by 0.319%. To reduce transportation intensity, Shanghai should encourage the development of multimodal transport, improve transportation efficiency, and promote the coordinated and sustainable development of all modes of transportation. In addition, Shanghai should improve slow-traffic infrastructure, guarantee the right of way for slow traffic, improve the accessibility and convenience of traffic networks, and create an amenable traffic space.

(2)

From the above forecast model, we can see that energy intensity has a positive impact on Shanghai transportation CO₂ emissions. A 1% reduction in transportation intensity would reduce CO₂ emissions by 0.278%. To reduce the energy intensity of Shanghai’s transportation sector, green and low-carbon concepts should be integrated into transportation infrastructure planning, construction, operation, and maintenance. Moreover, Shanghai should shift the demand for transportation from roads with high-energy consumption and pollution to environmentally friendly transportation modes, such as railways, waterways, and urban public transport, to reduce energy consumption and transportation carbon emissions. In addition, Shanghai should accelerate the transformation and upgrading of transport vehicles to be electrified, low-carbon, and intelligent.

(3)

From the above forecast model, we can see that energy structure (the proportion of clean energy) limits Shanghai transportation CO₂ emissions. A 1% increase in transportation intensity would reduce CO₂ emissions by 0.316%. Shanghai should promote green and low-carbon technology innovation and progress, promote energy efficiency, and increase the proportion of clean energy in Shanghai’s transportation sector. Shanghai should actively expand the application of clean energy, such as electricity, natural gas, advanced bio-liquid fuels, as well as hydrogen energy, in transportation. Further, Shanghai should improve the transportation industry’s carbon emission tax and fee mechanisms, carbon emission reduction incentive mechanisms, energy trading, carbon trading systems, investment in technological innovation, and development of energy technologies.