Consumption-Based Carbon Emissions of Tianjin Based on Multi-Scale Input–Output Analysis

Cities are a major source of carbon emissions and should play an important role in reducing carbon emissions. This study applies the method of multi-scale input–output analysis (MSIO) to analyze the consumption-based carbon emissions of Tianjin in 2012. This method can estimate the carbon emissions embodied in imported products. The results reveal that the production-based carbon emissions of Tianjin were 1.52 × 108 tonnes CO2 in 2012, which had increased over 50% since 2007. Meanwhile, the consumption-based carbon emissions of Tianjin city were 2.55 × 108 tonnes CO2, 1.71 times higher than those in 2007 and 1.67 times the amount of production-based carbon emissions in 2012. Regarding the total embodied carbon emissions involved in the Tianjin economy in 2012, about 6% were from foreign countries, over 60% were from other regions of China, and only one-third were territorial-based or production-based carbon emissions. Correspondingly, Tianjin respectively exported 11% and 34% of the total embodied carbon emissions to foreign countries and other regions in China, while over half were embodied in the local final demand. Tianjin was a carbon budget importer in domestic trade and an exporter in international trade in both 2007 and 2012. However, when both domestic and international trades are considered, Tianjin had shifted from a carbon budget exporter in 2007 to an importer in 2012. Since 2007, the carbon nexus between Tianjin and other regions in China had become much closer (carbon emissions embodied in domestic trade increased 103.47%), but the connection with foreign countries became looser (carbon emissions embodied in international trade decreased 21.96%). Compared to Beijing in 2012, it is evident that there were less carbon emission transfer issues for Tianjin city.


Introduction
Urbanization has developed rapidly in China since 1978. China's urbanization rate increased from 17.92% in 1978 to 57.35% in 2016 [1]. The rapidly growing urban population brought in a lot of carbon emissions, which had increased 190% from 1996 to 2012 [2]. China's target for carbon reduction is to decrease the carbon emissions per unit GDP by 60 to 65% of 2005 levels by 2030 [3]. In addition, the central government declared to reach peak carbon emissions by 2030. As one of China's four directly controlled municipalities, Tianjin is a traditional industrial city and a major seaport in the According to basic environmental input-output theory [19,20], the consumption-based carbon emissions (E) of Tianjin's local final demand can be calculated as follows: where X is the diagonalized matrix of the output column vector in Tianjin's input-output table.
(I − A) −1 is the Leontief inverse matrix, where I is an identity matrix and A is the technical coefficient matrix. Y represents final demand matrix. The vector ε L means the direct carbon emission coefficients of all sectors in traditional input-output analysis, which equals FX −1 . F represents the carbon emission vector emitted directly by local sectors, and X −1 is the inverse matrix of the diagonalized output matrix.
In the MSIO method developed by Chen and his team [16,18,21,22], ε L was expanded to include coefficients that originate from external carbon emission imports as well (the detailed algorithm can be found in [16]). Therefore, Equation (1) can be rewritten into L, D, and M are used to denote the local system of Tianjin, the domestic system, and the foreign system, respectively. ε D = [ε D 1 , ε D 2 . . . ε D n ] is a vector of embodied carbon emission intensities for domestically imported products from economic sectors of other Chinese regions. ε M = [ε M 1 , ε M 2 . . . ε M n ] is a vector of embodied carbon emission intensities for foreign-imported products from economic sectors of foreign countries. Z D and Z M are economic flow matrices of domestic imports and foreign imports that are used as the intermediate inputs of local economic sectors. It is indicated that the carbon emission intensity vector for the local product (ε L ) includes not only the local direct carbon emission intensity (FX −1 ), but also carbon emission intensities caused by domestic and international material imports (ε D Z D X −1 and ε M Z M X −1 ). As long as the carbon emissions intensities of local products in Tianjin are multiplied by the corresponding economic flow, the carbon emissions transfer along with the flow can be obtained.

Data Sources
As illustrated in the last section, the data needed by the MSIO analysis of Tianjin are: (a) the local carbon emissions of Tianjin's economy, which have been derived from China Emission Accounts and Datasets [23]; (b) the economic flow data of Tianjin, which have been obtained from the official input-output table of Tianjin [24] (the unit is Chinese Yuan, which is referred to as Yuan hereafter; the sectors are shown in Appendix A Table A1); (c) the embodied carbon emission intensity databases of the world (excluding China) and China, which have been cited from Shao et al. [16] (The corresponding sectors of Tianjin's economic input-output sectors in the global and Chinese carbon emission intensity database are also shown in Appendix A Table A1). The official input-output table of Tianjin does not distinguish between foreign imports and domestic imports. The method in Shao et al. has been applied to obtain the domestic and foreign import matrices of Tianjin [16].

Embodied Carbon Emissions Intensities of Tianjin in 2012
Embodied carbon emissions intensity is defined as the total carbon emissions driven per unit of product output. The embodied carbon emissions intensities of the world (except China), China, and Tianjin for 42 sectors in 2012 are illustrated in the Figure 1. The averaged embodied carbon emissions intensities of the world (except China), China, and Tianjin were 0.72, 2.05, and 1.96 tonnes CO 2 /10 4 Yuan, respectively. Tianjin's average embodied carbon emissions intensity was a little lower than that of China, but nearly treble that of the world economy (except China), which suggests that Tianjin triggered much more carbon emissions than the world's average level when outputting the same amount of value. Technological progress and economic growth usually lead to a decrease in the embodied carbon emissions intensity. While pursuing GDP growth, the government should also promote the development of low-carbon technologies.  Table A1).
The output-weighted embodied carbon emissions intensities of Tianjin's 42 sectors in 2012 are shown in Figure 1, too. The sectors belonging to the second industry (from Sector 2 to Sector 30) have higher embodied carbon emission intensities than those of the other two industries. As a modern industrial city, Tianjin relies heavily on secondary industry. It has been reported that secondary industry had contributed over 50% of Tianjin's total output in 2012 [25]. Therefore, it is important to develop low-carbon technologies for secondary industry to reduce the carbon emissions of Tianjin. Sector 25 (Electricity, Heat Production, and Supply) had the highest local carbon emissions intensity of 15.83 tonnes CO2/10 4 Yuan. Sector 27 (Water Production and Supply) ranked second with an embodied intensity of 6.99 tonnes CO2/10 4 Yuan. Chen et al. also found that energy-or resourceintensive industries had the highest carbon emission intensities in their study on Beijing 2007 [21]. The embodied carbon intensities of Sector 12 (Recycling), Sector 20 (Manufacture of Communication Equipment, Computer and Other Electronic Equipment), and Sector 30 (Transportation, Storage, Posts, and Telecommunications) are significantly magnified, indicating that these sectors are key industries to reduce Tianjin's carbon emissions.

Carbon Emissions Embodied in the Local Final Demands of Tianjin in 2012
The production-based carbon emissions of Tianjin in 2012 were 1.52×10 8 tonnes CO2. Among the 42 sectors, Sector 25 (Electricity, Heat Production, and Supply) ranked first with 7.58×10 7 tonnes CO2 carbon emissions (49.90%), which was followed by Sector 14 (Metal Smelting and Rolling Processing) with 4.14×10 7 tonnes CO2 carbon emissions (27.25%). The consumption-based carbon emissions embodied in the local final demands of Tianjin in 2012 were 2.55×10 8 tonnes CO2, which was 1.67 times the production-based carbon emissions. Therefore, if the study only takes the production-based carbon emissions into consideration, it would lead to an underestimation of Tianjin's carbon emission footprint.
According to the input-output table, local final demands of Tianjin are divided into rural household consumption, urban household consumption, government consumption, fixed capital formation, and inventory increase, respectively. Their embodied carbon emissions are from three  Table A1).
The output-weighted embodied carbon emissions intensities of Tianjin's 42 sectors in 2012 are shown in Figure 1, too. The sectors belonging to the second industry (from Sector 2 to Sector 30) have higher embodied carbon emission intensities than those of the other two industries. As a modern industrial city, Tianjin relies heavily on secondary industry. It has been reported that secondary industry had contributed over 50% of Tianjin's total output in 2012 [25]. Therefore, it is important to develop low-carbon technologies for secondary industry to reduce the carbon emissions of Tianjin. Sector 25 (Electricity, Heat Production, and Supply) had the highest local carbon emissions intensity of 15.83 tonnes CO 2 /10 4 Yuan. Sector 27 (Water Production and Supply) ranked second with an embodied intensity of 6.99 tonnes CO 2 /10 4 Yuan. Chen et al. also found that energy-or resource-intensive industries had the highest carbon emission intensities in their study on Beijing 2007 [21]. The embodied carbon intensities of Sector 12 (Recycling), Sector 20 (Manufacture of Communication Equipment, Computer and Other Electronic Equipment), and Sector 30 (Transportation, Storage, Posts, and Telecommunications) are significantly magnified, indicating that these sectors are key industries to reduce Tianjin's carbon emissions.

Carbon Emissions Embodied in the Local Final Demands of Tianjin in 2012
The production-based carbon emissions of Tianjin in 2012 were 1.52 × 10 8 tonnes CO 2 . Among the 42 sectors, Sector 25 (Electricity, Heat Production, and Supply) ranked first with 7.58 × 10 7 tonnes CO 2 carbon emissions (49.90%), which was followed by Sector 14 (Metal Smelting and Rolling Processing) with 4.14 × 10 7 tonnes CO 2 carbon emissions (27.25%). The consumption-based carbon emissions embodied in the local final demands of Tianjin in 2012 were 2.55 × 10 8 tonnes CO 2, which was 1.67 times the production-based carbon emissions. Therefore, if the study only takes the production-based carbon emissions into consideration, it would lead to an underestimation of Tianjin's carbon emission footprint.
According to the input-output table, local final demands of Tianjin are divided into rural household consumption, urban household consumption, government consumption, fixed capital formation, and inventory increase, respectively. Their embodied carbon emissions are from three sources: direct carbon emission, domestic import, and foreign import. As is shown in Figure 2, the carbon emissions driven by gross fixed capital formation held the top proportion (65.56%), and urban household consumption ranked second (19.85%), followed by government consumption (8.77%), inventory increase (4.25%), and rural household consumption (1.58%). This is partially due to the rapid development of Tianjin's city construction in 2012. Tianjin had put great efforts in developing Binhai New Area, where 135 major urban infrastructure projects were completed throughout the whole year [26]. inventory increase (4.25%), and rural household consumption (1.58%). This is partially due to the rapid development of Tianjin's city construction in 2012. Tianjin had put great efforts in developing Binhai New Area, where 135 major urban infrastructure projects were completed throughout the whole year [26].
In 2012, the urban population was 2.75 times larger than the rural population of Tianjin, and the per capita disposable income of urban residents was 2.19 times that of rural residents [1]. However, the ratio of carbon emissions embodied in urban household consumption to rural household consumption was as high as 12.56. It can be known that the residents in the urban area of Tianjin caused much more carbon emissions compared to people living in the rural area. Scholar Dhakal got a similar result in his research, showing that highly urbanized and economically important cities had merely 18% the population of China, but produced 41% of the GDP and contributed 40% of CO2 emissions in 2006 [27].    In 2012, the urban population was 2.75 times larger than the rural population of Tianjin, and the per capita disposable income of urban residents was 2.19 times that of rural residents [1]. However, the ratio of carbon emissions embodied in urban household consumption to rural household consumption was as high as 12.56. It can be known that the residents in the urban area of Tianjin caused much more carbon emissions compared to people living in the rural area. Scholar Dhakal got a similar result in his research, showing that highly urbanized and economically important cities had merely 18% the population of China, but produced 41% of the GDP and contributed 40% of CO 2 emissions in 2006 [27]. had transferred over 280 million passengers and 477 million tonnes of cargo [28], which were accompanied by massive virtual carbon emission flows. As Tianjin becomes an important node in the Belt and Road Initiative, the carbon emissions trade of the transportation-related sector would be expected to increase. Sector 28 (Construction Industry) has domestically imported the most carbon emissions (7.39×10 7 tonnes CO2) from other regions in China. Sector 25 (Electricity, Heat Production, and Supply) ranks second (5.56×10 7 tonnes CO2). Tianjin has imported a large amount of electricity from Inner Mongolia, Shaanxi, and other western provinces through the national West-East Power Transmission Project, which triggered a large amount of carbon emissions transfer.

Comparison of Embodied Carbon Emissions Flows of Tianjin in 2007 and 2012
As shown in Figure 4, the production-based carbon emissions of Tianjin in 2012 had increased 51.37% since 2007 while the consumption-based carbon emissions had increased 170.78%. It is revealed that the consumption-based carbon emissions grew much faster than the production-based carbon emissions in Tianjin. The production-based carbon emissions of Tianjin (1.00×10 8 tonnes CO2) were a little higher than the consumption-based carbon emissions (9.41×10 7 tonnes CO2) in 2007. However, the consumption-based carbon emissions (2.55×10 8 tonnes CO2) of Tianjin were 1.67 times the production-based carbon emissions (1.52×10 8 tonnes CO2) in 2012. According to Mi et al.'s theory [6], Tianjin had become a consumption-based city.  Among 42 sectors, Sector 18 (Manufacture of Transport Equipment) had the largest foreign-exported and foreign-imported carbon emissions. At the same time, its domestic exports embodied the largest carbon emissions. Sector 30 (Transportation, Storage, Posts, and Telecommunications) also had a close carbon emission relationship with other regions, which ranked second in both domestic-exported and foreign-imported carbon emissions. Both Sectors 18 and 30 are transportation-related industries. Tianjin Port is China's third largest port and one of the most important transportation hubs in northern China. Tianjin city undertakes many transportation tasks every year. In 2012, Tianjin had transferred over 280 million passengers and 477 million tonnes of cargo [28], which were accompanied by massive virtual carbon emission flows. As Tianjin becomes an important node in the Belt and Road Initiative, the carbon emissions trade of the transportation-related sector would be expected to increase. Sector 28 (Construction Industry) has domestically imported the most carbon emissions (7.39 × 10 7 tonnes CO 2 ) from other regions in China. Sector 25 (Electricity, Heat Production, and Supply) ranks second (5.56 × 10 7 tonnes CO 2 ). Tianjin has imported a large amount of electricity from Inner Mongolia, Shaanxi, and other western provinces through the national West-East Power Transmission Project, which triggered a large amount of carbon emissions transfer.

Comparison of Embodied Carbon Emissions Flows of Tianjin in 2007 and 2012
As shown in Figure 4, the production-based carbon emissions of Tianjin in 2012 had increased 51.37% since 2007 while the consumption-based carbon emissions had increased 170.78%. It is revealed that the consumption-based carbon emissions grew much faster than the production-based carbon emissions in Tianjin. The production-based carbon emissions of Tianjin (1.00 × 10 8 tonnes CO 2 ) were a little higher than the consumption-based carbon emissions (9.41 × 10 7 tonnes CO 2 ) in 2007. However, the consumption-based carbon emissions (2.55 × 10 8 tonnes CO 2 ) of Tianjin were 1.67 times the production-based carbon emissions (1.52 × 10 8 tonnes CO 2 ) in 2012. According to Mi et al.'s theory [6], Tianjin had become a consumption-based city.  As for carbon emissions embodied in local final demands of Tianjin (see Figure 2), the carbon emissions embodied in gross fixed capital formation in 2012 were 3.21 times as high as those in 2007. In 2012, Tianjin's construction industry output and annual building construction area had increased by 275.10% and 48.07% since 2007, respectively. Meanwhile, Tianjin's total fixed asset investment in 2012 was 3.71 times that in 2007. All these have contributed to the rapid increase in the embodied carbon emissions of gross fixed capital formation.
The carbon emissions embodied in urban household consumption and rural household consumption had increased by 99.54% and 60.76% during 2007-2012, respectively. Despite the growths in both Tianjin's urban and rural households' carbon footprints , the gap between per capita carbon emissions of urban households and rural households in Tianjin became even larger (from 3.71 to 7.14 tonnes CO2). China is promoting urbanization nationwide, and the urban population will be increasing fast in the near future. In addition, the Chinese government is trying their best to reduce the urban-rural gap through a series of favorable agricultural measures, and the living standards of rural residents are gradually improving. These would further enlarge the carbon footprint for Tianjin. Therefore, green lifestyle and energy-saving technology should be promoted in Tianjin to achieve The carbon emissions embodied in urban household consumption and rural household consumption had increased by 99.54% and 60.76% during 2007-2012, respectively. Despite the growths in both Tianjin's urban and rural households' carbon footprints, the gap between per capita carbon emissions of urban households and rural households in Tianjin became even larger (from 3.71 to 7.14 tonnes CO 2 ). China is promoting urbanization nationwide, and the urban population will be increasing fast in the near future. In addition, the Chinese government is trying their best to reduce the urban-rural gap through a series of favorable agricultural measures, and the living standards of rural residents are gradually improving. These would further enlarge the carbon footprint for Tianjin. Therefore, green lifestyle and energy-saving technology should be promoted in Tianjin to achieve low-carbon development. The

Comparison of Embodied Carbon Emissions Flows of Beijing and Tianjin in 2012
As Beijing is one of the most developed cities in China, a comparison is made between Beijing and Tianjin to see if Tianjin could learn something from Beijing (see Figure 4). The carbon emissions data of Beijing are from one of our previous studies [16]. The consumption-based carbon emissions of Beijing in 2012 were 2.35 × 10 8 tonnes CO 2 , roughly the same as those of Tianjin (2.55 × 10 8 tonnes CO 2 ). However, Beijing's consumption-based carbon emissions were 2.50 times the production-based carbon emissions, while the ratio in Tianjin was only 1.67. This shows that Beijing had more carbon emission transfer issues compared to Tianjin. This paper has also compared the carbon emission inflows and outflows of Beijing and Tianjin in 2012. According to the results, the total carbon emissions related to Tianjin in 2012 were 4.67 × 10 8 tonnes CO 2 . As for the carbon emissions inflows, the embodied carbon emissions from local, domestic-imported, and foreign-imported products were respectively sharing 32.49%, 61.37%, and 6.15% of total carbon emissions. For carbon emissions outflows, 54.51% were embodied in the local final demand, 34.22% were exported to other regions in China, and 11.27% were exported to other countries in the world. Direct carbon emissions from local sources in the inflows of Tianjin (1.52 × 10 8 tonnes CO 2 ) were 1.66 times those of Beijing (9.13 × 10 7 tonnes CO 2 ), which means that Tianjin had emitted more carbon emissions than Beijing during the production process. Beijing obviously had much larger embodied carbon emissions in domestic trade than Tianjin. The carbon emissions embodied in the domestically imported and exported trade of Beijing were 4.38 and 7.63 times those of Tianjin, respectively. This indicates that Beijing had a closer carbon relationship with other Chinese regions than Tianjin. 8.20 tonnes CO 2 ) and 61.59% (from 0.66 to 1.07 tonnes CO 2 ) during 2007-2012, respectively. It can be seen that the carbon emissions gap between urban and rural residents in Tianjin was very huge, which became even larger during 2007-2012.

Conclusions
In the total amount of carbon emissions related to Tianjin in 2012 (4.67 × 10 8 tonnes CO 2 ), 32.49% were from local emissions, and 61.37% and 6.15% carbon emissions were respectively from domestic imports and foreign imports. Approximately half of the total amount of CO 2 had been embodied in the local final demand, one-third of carbon emissions were exported to other regions in China, and about 11% carbon emissions had been exported to other countries. By comparing to the results of Tianjin in 2007, the carbon nexus between Tianjin and other regions in China had become much closer, and the connection of Tianjin with other foreign countries became looser.
Although Tianjin was a carbon budget importer in domestic trade and a carbon budget exporter in international trade in both 2007 and 2012, carbon trades of Tianjin had changed significantly. The carbon emissions embodied in domestic imports and exports had respectively increased by 152.69% and 50.71%. As for foreign trades, internationally imported and exported carbon emissions had respectively decreased by 36.64% and 10.68%. Transportation and resources-related sectors such as Sector 30 (Transportation, Storage, Posts, and Telecommunications) and Sector 25 (Electricity, Heat Production, and Supply) had embodied large amounts of carbon emissions trade.
Tianjin plays an important role in integrating the Bohai Economic Belt and the Beijing-Tianjin-Hebei region. It is also projected as a key area in the Belt and Road Initiative. When compared to Beijing's carbon flows in 2012, it can be seen that Beijing contacts more with other regions in China than Tianjin. Tianjin should seize the historical opportunity to expand communication with other Chinese regions and foreign countries. However, at the same time, Tianjin should pay attention to carbon emissions embodied in transportation-related industries and the virtual carbon emission transfer embodied in trade. To promote energy conservation and emission reduction technology or green transports in the transportation sector, increasing imports from economies with a low carbon emission intensity would be helpful to reduce Tianjin's carbon footprint.