Energy-related CO2 emissions in Fujian Province.
China has implemented several FYPs to set goals and directions for long-term economic development. As official data from 2018 to 2020 are not yet available, our study period covers 2000–2017. We chose 2000, 2005, 2010, 2015, and 2017 as five specific years to compare the spatial differences among nine prefecture-level cities in Fujian Province.
Total CO2 emissions in Fujian during 2000–2017 are shown in Fig. 1. Total CO2 emissions increased rapidly from 78.4 Mt to 432.11 Mt during the study period, with an average annual growth rate of 6.70%. During 2000–2005, CO2 emissions increased from 143.2 Mt to 194.69 Mt, with an average annual growth rate of 6.34% (moderate rate). Between 2005 and 2010, CO2 emissions increased more rapidly, with an annual growth rate of 11.41%. Slower growth in CO2 emissions was observed from 2010 to 2015, with an average annual growth rate of 4.89%. Finally, during 2015–2017, CO2 emissions increased only slightly and with fluctuations, with an average annual growth rate of 0.93%. The observed change in CO2 emissions is related to the transformation of economic development patterns, with recent energy saving and emission reduction policies gradually achieving some success. CO2 emissions from the industry sector accounted for 86.6% of total regional CO2 emissions during this period. Therefore, this sector was the largest CO2 emitter in Fujian Province, followed by the transportation sector, whose share in regional CO2 emissions fluctuated slightly between 2.28% and 5.71% from 2000 to 2017. The third-largest CO2 emitter was the household sector, which accounted for 4.22% of total CO2 emissions during 2000–2017. The influence of other sectors on total CO2 emissions in Fujian Province was negligible so will not be discussed further in this paper.
Figure 2 presents the energy-related CO2 emissions of nine cities in 2000, 2005, 2010, 2015, and 2017. Most cities showed growing trends in CO2 emissions in all years, except for Longyan City from 2010 to 2017 (12th and 13th FYPs) and Xiamen City from 2015 to 2017 (13th FYP), where CO2 emissions decreased. Moreover, the average annual growth rates of CO2 emissions in the nine cities were 15.06% (Ningde), 12.60% (Putian), 10.39% (Zhangzhou), 7.69% (Fuzhou), 6.70% (Quanzhou), 6.55% (Sanming), 3.77% (Xiamen), 3.23% (Longyan), and 2.11% (Nangping). In 2017, the two cities with the highest CO2 emissions, i.e., Quanzhou and Fuzhou, were responsible for 39.22% and 16.61% of total CO2 emissions, respectively. In comparison, all other cities accounted for less than 10% of total CO2 emissions, i.e., 9.8% (Sanming), 9.16% (Zhangzhou), 7.14% (Xiamen), 5.92% (Ningde), 5.46% (Longyan), 4.32% (Putian), and 2.37% (Nanping). Therefore, clear spatial differences in CO2 emissions were observed in Fujian Province, with changes in CO2 emissions in Quanzhou and Fuzhou playing a crucial role in provincial-level CO2 emissions.
Spatial-temporal decomposition analysis of energy-related CO2 emissions
Figures 3–8 show the results of the temporal and spatial decomposition analysis. The temporal decomposition of ICEs, ECEs, and HCEs for the nine cities was calculated for 2000–2017 using Eqs. 3–11 (Figs. 3, 5, and 7). The relative contribution rates of different drivers are shown in Tables A1, A3, and A5 in supplementary material. The spatial decomposition of ICEs, ECEs, and HCEs for the nine cities in 2000, 2005, 2010, 2015, and 2017 was further calculated using Eqs. 12–20 (detailed calculation results are shown in Tables A2, A4, and A6 in supplementary material). We then investigated the spatial relationship among three main driving factors (∆CEt, ∆CG, and ∆CUr) in different cities, as shown in Figs. 4, 6, and 8.
Spatial-temporal decomposition of energy-related CO2 emissions from the industry sector
Figure 3 shows that ∆CG was the dominant factor leading to higher CO2 emissions in each city during 2000–2017. ∆CG was largest in Quanzhou (113.04 Mt), followed by Fuzhou (42.00 Mt), and the relative contribution rate ranged from 50.30% (Fuzhou) to 25.89% (Sanming). This result agrees with those of previous studies (Chang et al., 2019, Chen et al., 2019, Li et al., 2017, Quan et al., 2020, Shi et al., 2019, Xue et al., 2019). ∆CEt was the dominant factor leading to lower CO2 emissions in all cities except for Ningde and Putian. Quanzhou showed the largest ∆CEt during 2000–2017 (−94.56 Mt), followed by Sanming (−36.96 Mt). The relative contribution rate ranged from −47.07% (Nanping) to −22.2% (Fuzhou). Improvements in productivity and development, and the deployment of clean energy can reduce energy intensity, which is crucial for reducing industrial carbon emissions. ∆CUr was the second largest positive effect on the increase in ICEs, which was especially more prominent in coastal cities (e.g., Quanzhou, Fuzhou, and Putian), for example, ∆CUr in Quanzhou reached 46.25 Mt during 2000–2017. Urbanization can increase ICEs via infrastructure construction, industrial development, and urban population expansion. ∆CP was positive in most cities (most significantly in Xiamen), indicating that the increase in population size from 2000 to 2017 in Xiamen contributed to the observed changes in ICEs, and that the population load in Xiamen is relatively heavy. ∆CIs was also positive in each city, however, there was a discrepancy in the contribution rate among cities, with Sanming, Zhangzhou, Ningde, and Nanping exhibiting higher contribution rates. ∆CEs and ∆CQi had little impact on the changes in ICEs in all cities in Fujian Province from 2000 to 2017.
The spatial decomposition results of the three main driving factors are denoted by dots in Fig. 4, and joined by arrows. The three factors correspond to the X-axis (∆CEt), Y-axis (∆CG), and Z-axis (∆CUr), and the plotting area is divided into four quadrants and two axes according to the projection of each dimension. According to Fig. 4(b), cities located in quadrant Ⅰ exhibited lower ∆CG and ∆CUr than the regional average, cities located in quadrant Ⅱ had higher ∆CG and lower ∆CUr than the regional average, cities located in quadrant Ⅲ had higher ∆CG and ∆CUr than the regional average, and cities located in quadrant Ⅳ had lower ∆CG and higher ∆CUr than the regional average. The data displayed in Fig. 4(c) and Fig. 4(d) can be interpreted in a similar way. The spatial decomposition results provide a comprehensive description of the variations in ICEs in the nine cities. Movement over time toward a lower value along an axis indicates a positive development, i.e., ∆CEt in Xiamen (Fig. 4[c]), whereas movement over time toward a higher value along an axis indicates a negative development, i.e., ∆CEt in Quanzhou (Fig. 4[c]).
The difference between ∆CEt and the corresponding average value was much greater than the difference between ∆CUr or ∆CG and their corresponding average values. Moreover, the evolutionary trends of the scattered dots were diverse, in that some cities were always located in one quadrant and some cities covered multiple quadrants. Therefore, we classified the nine cities into four groups according to the spatial decomposition results. The first group, which includes Putian, Nanping, and Ningde, were cities predominantly located in one quadrant (i.e., quadrant Ⅲ). These cities moved toward a lower value over time (Fig. 4[b–d]), indicating higher ∆CEt, ∆CUr, and ∆CG values than the regional average, this inhibitory effect increased over time. The second group includes cities located in both quadrant Ⅰ and quadrant Ⅱ, i.e., Fuzhou and Xiamen. The ∆CEt value of Fuzhou and Xiamen was negative but increased over time, indicating higher energy use efficiency than the regional average, which curbed the excessive growth of carbon emissions. The ∆CG and ∆CUr values of Fuzhou and Xiamen were also higher than the regional average, indicating that economic development and urbanization promoted carbon emissions in these cities. The third group includes cities located in both quadrant Ⅲ and quadrant Ⅳ, i.e., Longyan City. Here, the ∆CEt was positive, whereas ∆CG and ∆CUr were negative, indicating that improving the energy use efficiency is imperative for reducing industrial carbon emissions in this city. Quanzhou, Zhangzhou, and Sanming cities belong to the fourth group as they were located across three or more quadrants. Quanzhou City exhibited the most distinctive characteristics of change, whereby ∆CG and ∆CEt were both positive and ∆CUr changed from negative to positive in 2010, indicating that all three factors had low influence. Therefore, policy priorities to reduce ICEs in each city should be developed according to the spatial decomposition analysis of the driving forces in each city.
The increase in ICEs in Fujian Province was generally affected by the increase in ∆CG, where the ∆CG value in industrially developed areas (e.g., Quanzhou, Zhangzhou, Fuzhou, Sanming, and Xiamen) was higher than the regional average. Since the reform and opening-up, to further meet the demand for economic growth, the government of Fujian Province has been increasing construction and investment in the industry sector, which in turn has increased energy consumption and related CO2 emissions. ∆CEt was the main factor responsible for the mitigation of ICEs in the study area, energy intensity decreased in each city, except for Ningde City, indicating that carbon emission reduction targets played an important role in enabling regions to improve energy use efficiency by adjusting their energy consumption structure and promoting the use of clean energy.
Spatial-temporal decomposition of energy-related CO2 emissions from the transportation sector
Figure 5 shows that ∆CG was also the dominant factor influencing the increase in TCEs in Fujian Province. Quanzhou showed the largest ∆CG of 3.35 Mt during 2000–2017, whereas Putian showed the lowest ∆CG of 0.53 Mt, and the relative contribution rate ranged from 62.27% (Nanping) to 31.19% (Xiamen). ∆CUr had the second highest positive effect on the growth of TCEs, especially in Quanzhou (1.37 Mt), Fuzhou (0.62 Mt), and Zhangzhou (0.43 Mt). The relative contribution of ∆CEs to the transportation sector was higher than that to the industry sector, with the relative contribution rate ranging from 15.98% (Zhangzhou) to 6.36% (Putian), therefore, the energy structure of the transportation sector requires adjustment. Unlike for ICEs, ∆CEt showed a positive effect on TCEs in most cities except Putian City (−0.09 Mt), the relative contribution rate ranged from 22.45% (Xiamen) to 0.67% (Fuzhou), indicating that the energy use efficiency of the transportation sector requires further improvement. Moreover, ∆CP was positive in coastal cities (e.g., Quanzhou, Fuzhou, Zhangzhou, and Xiamen), indicating that an increase in population size promoted the increase in TCEs. ∆CIs was negative in most cities except Zhangzhou (0.17 Mt), the relative contribution rate ranged from −17.95% (Fuzhou) to −2.88% (Longyan) indicating that development of the transportation sector has a suppressive effect on TCEs.
Figure 6 presents the spatial decomposition results of TCEs, according to which the nine cities were divided into three groups. The first group includes the cities predominantly located within one quadrant (quadrant III), which includes Putian, Nanping, Ninde, and Xiamen. Putian and Nanping consistently appeared in quadrant Ⅲ (Fig. 6[b–d]). For Xiamen, which was consistently located in quadrant Ⅰ, the positive effect of ∆CEt and ∆CUr was larger and increased over time, indicating that the transportation sector was driven by population migration during urbanization. However, the energy efficiency of the transportation sector did not improve, which led to an increase of CO2 emissions. The second group includes cities mostly located in quadrant II, i.e., Fuzhou, where ∆CUr and ∆CG were higher than the regional average. The third group includes the cities mostly located in quadrant Ⅳ, i.e., Longyan. Here, ∆CUr and ∆CG were consistently negative and increased over time, whereas ∆CEt was positive and increased over time. The fourth group includes cities located across three or more quadrants, i.e., Quanzhou, Zhangzhou, and Sanming, which all exhibit similar characteristics. Specifically, ∆CG had a positive effect in these cities, whereas ∆CEt changed from positive to negative in different years. Income improvement generally drives population mobility and the circulation of goods, and the population tends to flow from developing areas to developed areas. This contributes to development of the transportation sector, and the increasing energy consumption required for transportation leads to an increase in carbon emissions. Therefore, ∆CG was higher than the regional average in industrially developed regions (e.g., Quanzhou, Zhangzhou, Xiamen, Fuzhou, and Sanming). ∆CEt increased TCEs, and the ∆CEt value in Xiamen was much higher than the regional average. These results suggest that it is necessary to improve energy use efficiency in the transportation sector.
Spatial-temporal decomposition of energy-related CO2 emissions from the household sector
Figure 7 shows that ∆CG was the dominant factor leading to an increase in HCEs, and the relative contribution rate ranged from 147.66% (Fuzhou) to 74.18% (Xiamen). ∆CEt was the dominant factor responsible for the mitigation of HCEs in each city, and the relative contribution rate ranged from −117.01% (Fuzhou) to −34.87% (Longyan). ∆CUr had the second highest positive influence on the increase in HCEs in most cities, and the relative contribution rate ranged from 49.64% (Zhangzhou) to 15.12% (Xiamen), indicating that urbanization led to a concentration of population in urban areas and an improvement of people’s living standards, and thus to an increase in HCEs. ∆CEs were positive in each city, indicating that the energy structure led to an increase in HCEs. ∆CP showed a significant positive effect in developed cities, for example, the relative contribution rate in cities with recent rapid population growth (Xiamen, Fuzhou, and Quanzhou) was 44.14%, 19.71%, and 10.69%, respectively, indicating that population growth is a significant factor driving increased HCEs.
The nine cities are divided into three groups in Fig. 8. The first group includes Xiamen and Fuzhou, which were characterized by negative ∆CEt that progressively decreased compared to the regional average, positive ∆CUr that was higher in Xiamen than in Fuzhou, and positive ∆CG, the gap between ∆CG and the regional average decreased over time. The second group includes cities that covered two quadrants, including Nanping, Putian, Ningde, and Longyan. In these cities, ∆CUr and ∆CG were consistently negative and ∆CEt was consistently positive, indicating that the energy efficiency of these cities was lower than the regional average. The third group includes cities that moved across three or more quadrants, including Quanzhou, Zhangzhou, and Sanming. In Zhangzhou, ∆CEt and ∆CUr were negative, and the ∆CEt gap between the cities and the regional average was relatively small. Conversely, ∆CG was positive, and the gap continued to increase. With an increase in household income and living standards, people prefer to purchase more carbon-emitting products, which leads to an increase in HCEs. ∆CG was more evident in high-income areas, such as Quanzhou, Zhangzhou, Xiamen, and Fuzhou. The observed decrease in energy intensity was likely related to people’s awareness of environmental protection and adjustment of their consumption structure. ∆CEt was below the provincial average in developed regions (e.g., Xiamen, Fuzhou, Quanzhou, and Zhangzhou), indicating that improved living standards can promote a reduction of energy intensity in the household sector.