Carbon dioxide mitigation co-effect analysis of clean air policies: lessons and perspectives in China’s Beijing–Tianjin–Hebei region

In 2018, the Beijing–Tianjin–Hebei (BTH) area launched the Blue Sky Protection Campaign (BSPC) to control atmospheric pollution. CO2 emissions could be significantly reduced due to the co-effects of implementing the BSPC. This paper employs the Greenhouse Gas and Air Pollution Interactions and Synergies Asia model to quantitatively evaluate the CO2 reductions when implementing the BSPC in the BTH region. The results indicate that CO2 emissions can be reduced by 20.7 Mt (equivalently, a 19.7% reduction in the corresponding baseline scenario), 6.8 Mt (3.8%), and 80.2 Mt (9.2%) by 2020 for Beijing, Tianjin, and Hebei, respectively, as a co-benefit of the BSPC. By 2030, it is estimated that the CO2 emission reductions will be 37.8 Mt (26.6%), 4.85 Mt (2.5%), and 69.9 Mt (8.6%) for Beijing, Tianjin, and Hebei, respectively. NOx presents the highest co-effects with CO2 in each region. From the key sector perspective, sectors of power and heating in Beijing, residential combustion in Tianjin, and industrial combustion in Hebei are the most important sector that presents the highest co-effects on CO2 emission reductions due to the application of BSPC. We suggest that the implementation of BSPC, specifically the energy control measures in the power and heating, residential combustion, and industrial combustion sectors for Beijing, Tianjin, and Hebei, respectively, have high synergies and can simultaneously reduce CO2 and other atmospheric emissions. The results contribute to city-level policymaking on facilitating air pollution control and climate change mitigation among different governmental departments.


Introduction
In recent decades, greenhouse gas and air pollutant emissions have been increasing rapidly in China as a result of the dramatic economic growth that relied on fossil fuels (Lei et al 2011, Zhao et al 2012. China surpassed the United States and became the world's largest carbon emitter in 2007 (Wen and Shao 2019). Substantial efforts have been made to reduce carbon emissions. Based on the 13th Five-Year Plan (FYP), China is making efforts to improve the proportion of non-fossils fuels in the total energy consumption to 15% by 2020, and decrease the share of coal consumption to below 58% by 2020 (Xu et al 2017). In addition, a national carbon trading scheme was launched in order to effectively cope with climate change through market-based mechanisms in 2017 (Lo 2013). The Chinese government has made ambitious commitments in terms of reducing CO 2 emissions. Specifically, the carbon intensity (measured by tons of CO 2 emissions per unit of GDP) of 2020 and 2030 is declared to decrease by 40%-45% and 60%-65%, respectively, compared with 2005 level (Xu et al 2017, Dong et al 2018. To achieve these goals, a series of measures have been taken to decrease the consumption of fossil fuels and develop cleaner energy such as hydropower, wind energy, and nuclear power (Liu et al 2015).
The Chinese government has also implemented some air pollutant control plans. In 2012, China's Ministry of Environmental Protection adopted the Air Quality Standards (DEP 2012), and PM 2.5 pollution was monitored in different regions. The Chinese government released the Air Pollution Prevention and Control Action Plan (APPCAP) in 2013, so as to reduce urban PM 10 levels by 10% and decrease the PM 2.5 concentrations of the Beijing-Tianjin-Hebei (BTH), the Pearl River Delta (PRD), and the Yangtze River Delta (YRD) by 15%-25% from 2012 to 2017 (Ma et al 2020). The APPCAP was the first plan that specified the air quality goals in China (APPCAP 2013) and it may have been one of the most influential environmental policies in the past 5 years (Cai et al 2017). Specifically, it was calculated that with the implementation of the APPCAP, the emissions of SO 2 , NO X , PM 2.5 , and non-methane volatile organic compound (NMVOC) in 2017 had decreased by 36%, 31%, 30%, and 12%, with NH 3 emissions increased by 10% from the 2012 levels in BTH, respectively, based on the Weather Research and Forecasting model and Community Multiscale Air Quality model (Cai et al 2017). Also, results from Zhang et al (2018) indicated that the annual mean PM 2.5 concentration had decreased by more than 30% throughout the country since the implementation of the APPCAP using arithmetic mean and percentile methods. It has been concluded that under the APPCAP, industrial combustion and industry process play important roles in SO 2 reduction while it is the transportation sector dominates the NO x reduction, and the industrial sources and domestic combustion sources contribute the most to the PM 2.5 reduction. It is worth noting that the APPCAP has reduced 352.7 Mt (39.9 ± 5.3%) of CO 2 emissions when compared to the emissions in 2012 (Lu et al 2019). In 2018, the Chinese government launched a 3 year plan named the Blue Sky Protection Campaign (BSPC) to reduce air pollution. The specific goals are to decrease SO 2 emissions, NO x emissions, and PM 2.5 concentration by 15%, 15%, and 18%, respectively, compared with 2015 levels. In addition, the goal of helping reduce greenhouse gas emissions are also mentioned under the BSPC but with no controls specified. The BSPC mainly focuses on the three key regions of BTH, YRD and Fenwei Plain. Specific measures were designed to reduce emissions of atmospheric pollutants, with core contents in energy related measures, including eliminating highly polluting facilities, optimizing the energy structure, reducing the use of traditional fossil fuel in power, industrial, and residential sectors, and increasing the proportion of clean energy in total energy use. It is worth noting that these control policies would also have impacts on other air pollutants besides SO 2 , NO x , and PM 2.5 . For example, Cai et al (2017) presented that implementing the APPCAP had decreased the emissions of NMVOC by 12% and increased NH 3 emissions by 10% by 2017, and would decrease NMVOC emissions by 22% and increase NH 3 emissions by 3% by 2020 from the 2012 levels in the BTH region. Maji et al (2020) concluded that the APPCAP had led to a reduction in CO concentration by 80 µg m 3 yr −1 while the O 3 concentration was increased by 1.3 µg m −3 yr −1 from 2014 to 2018 in Beijing. In addition, many studies have shown the opposite trends of PM 2.5 and O 3 precursor emissions under the APPCAP (Fenech et al 2019, Zhao et al 2020. This is mainly because O 3 is a typical secondary pollutant and the increase in O 3 could be due to non-linear chemistry effects (e.g. NO x Titration), as well as because of changes in PM 2.5 (Dong et al 2019). As the BTH region is one of China's most developed regions owing to its status in terms of economic development as well as one of the regions suffering from the most severe air pollution (Qi et al 2017, Hao et al 2018, both the APPCAP and the BSPC have identified the BTH region as a priority area for air pollution control and prioritized an investigation of the policy impacts. As policies that intended to address either air pollution or CO 2 mitigation may well impact on the other, generating co-effects or trade-offs (Rao et al 2016), it is important to investigate the existing indirect impacts of air pollution regulations on CO 2 mitigation and utilize the lessons learned from previous policies.
The synergistic effects between air pollution mitigation and CO 2 emission reduction has been confirmed by many studies (Mittal et al 2015, Peng et al 2017. Wagner and Amann (2009) applied the Greenhouse Gas and Air Pollution Interactions and Synergies (GAINS) model to assess the effectiveness of CO 2 reduction measures in the Kyoto Protocol and concluded that CO 2 reduction goals could be reached while reducing emissions of SO 2 , NO x , and PM 2.5 by an additional 5%. Shrestha and Pradhan (2010) explored the synergies in Thailand by using a cost optimization method, which showed that by reaching the target of decreasing CO 2 emissions by 30%, SO 2 would decrease by 43%. Many studies have suggested that policies targeting sources of air pollution can also lead to CO 2 reductions benefits. For China, Lu et al (2019) identified that the energy-related measures under the APPCAP had reduced 47.3 ± 0.8% of SO 2 , 32.5 ± 2.6% of NO x , 15.2 ± 0.2% of PM 2.5 , and 39.9 ± 5.3% of CO 2 emissions, compared to the emissions in 2012 using the GAINS model, and suggested that PM 2.5 and NO x had a high co-effect on reducing CO 2 emissions in the BTH region. Alimujiang and Jiang (2020) analyzed the co-effects of the electric vehicles promoting policy in Shanghai and discovered that substituting electric vehicles for conventional fuel ones could aid in both atmospheric pollutant reduction and CO 2 reduction. Yan et al (2013) analyzed the synergy between the reductions of atmospheric pollutant and CO 2 emissions under the Clean Development Mechanisms (CDMs), which aims to offer the opportunity for developed countries to make lowcost greenhouse gas reducing investments in developing countries. It was concluded that the coal to gas policy has a high CO 2 emission reduction co-effects. In addition, Yang et al (2018) concluded the measures for mitigating PM 2.5 pollution can have CO 2 reduction potentials based on a multi-objective analysis in China's iron and steel industry. Similar synergistic effects between air pollutants and CO 2 emissions have also identified by other studies under various air control policy (Dong et al 2019. The air pollution control measures under the BSPC aims to improve the air quality. To realize this goal, specific measures including clean energy transformation in the energy system is implemented in different sectors. It is clear that emissions of greenhouse gases, particularly CO 2 , could benefit from this policy. As most studies focus on the air quality improvement after the implementation of the APPCAP, the impacts and co-effects of implementing the BSPC have not been evaluated. To the best of our knowledge, we provided the first study on the synergistic effects of the BSPC on CO 2 mitigation in the BTH region. In this research, the GAINS Asia model was applied to examine the CO 2 emission reductions, which is a co-effect of implementing the BSPC, in the BTH region of China. The results of the GAINS Asia model can provide a comprehensive CO 2 emission inventory for the year of 2015 in the BTH region, which is the most up to date one as far as we know. Future CO 2 emissions trends from different sectors after implementing the BSPC are also projected up to 2030 based on the model. We aim to answer these questions: (a) How much have CO 2 emissions reduced since the implementation of the BSPC in the BTH region? (b) Which factors contribute to the changes in CO 2 emissions? (c) What are the differences in synergies between different pollutants and the major sectors? This work will advance the scientific understanding of co-effects on CO 2 emissions when implementing air quality policy and will provide evidence of the synergies for future decisionmaking in air pollution control.

Model description
The GAINS Asia model, which was developed by the International Institute of Applied Systems Analysis, is a comprehensive assessment system that evaluates the interactions and the effectiveness of different atmospheric management policies. This model takes into account activities (such as power, industry, transportation, etc) and air pollution controls for different pollutants from different source sectors at 5 year intervals (Amann et al 2011. We employed this model to assess the CO 2 emission reductions due to the co-effects of implementing the BSPC. Details regarding the emission calculation principle of the model are provided in section 1 of the supporting information.

Scenario settings
In this study, the base year is 2015 and the evaluated period covers 2015-2030 in 5 year intervals. Two scenarios are proposed to study the policy effectiveness: the baseline scenario and the policy scenario. Based on the World Energy Outlook 2018 Current Policy Scenario (WEO-2018-CPS) projected by the GAINS Asia model, the baseline scenario in our study assumed that the BSPC is not implemented in the BTH region. We recalibrated the activities of energy, agricultural, and industrial processes for 2020 based on the trend from the 'WEO-2018-CPS' and the 13th FYP targets published by the BTH regional government and China's renewable report (NDRC 2016, NEA 2017a). The APPCAP is not part of our baseline scenario due to the 5 year interval evaluation and that the 2020 targets set for The policy scenario assumes that the BSPC is implemented in the BTH region in order to predict the effects of this control policy from 2015 to 2020. The policy packages of the policy scenario are shown in table S1 (available online at stacks.iop.org/ERL/16/015006/mmedia) in section 2 of the supporting information. Specifically, the policy packages are divided into four categories: the power and industry sector (with core content such as improving industrial structure and the associated distribution for steel, cement, coke, glass, and coal power plants); the transport sector (with core content such as increasing the number of new energy vehicles and implementing national vi (B) standard gasoline and diesel for traffic vehicles); the building sector (with core content such as replacing untreated coal of heating by households); and the cross sector (such as the pollution controls for stationary sources). The policy packages of the BSPC and its corresponding emission control measures in the GAINS Asia model are summarized in table S2 in section 3 of the supporting information. Within the policy scenario setting step, various air pollution sources (e.g. power, industry, transportation, residential building, and agriculture) and associated activities will be considered. More specifically, parameters (e.g. energy consumption by different fuels and sectors, industrial process activities, and utilization rate of atmospheric pollution control technology) are forecasted for 2020 using the consumption trends from the yearbooks (NBS 2016(NBS , 2017(NBS , 2018(NBS , 2019. For example, a reduction in coal consumption by urban residents and commercial combustion sectors is attributed to the following factors: (a) in the urban residential heating sector, we assumed that 80% of the heat demand was derived from the district heating system (Xiong et al 2015); (b) hard coal grade 2 is applied instead of hard coal grade 3, and the usage of natural gas and electricity increased. These assumptions are consistent with the BSPC regulations. From 2020 to 2030, it is assumed that the same levels of policy strictness will be maintained in the BTH region and the governments will continue to implement specific policy measures at the same rate. The detailed ways to determine the key parameters and default parameters of the GAINS Asia model used in our study are present in section 4 of the supporting information.

Quantification of CO 2 co-effects
Two indicators are introduced to evaluate the synergistic effects of CO 2 and air pollutants under the BSPC. The indicator of S 1 , which is named the 'degree of co-effect' , represents the ratio of CO 2 reduction to one pollutant reduction (equation (1)).
S 1 and S 2 has been suggested in some studies to assess the synergistic level (Nathan and Kristin 2010, Liu et al 2013, Lu et al 2019. The higher S 1 is, the stronger the co-effect is. However, S 1 could be high when both ∆E CO2 and ∆E P are low in certain cases, indicating low reduction potentials. And also S 1 could be low when there are large changes in both air pollutant and CO 2 emissions. Therefore, the 'relative degree of co-effect' indicator S 2 has been used as the indicator for assessing the synergies in our study. E CO2 is the CO 2 emissions (Mt). E p represents the pollutant p emissions (Kt). When we implement the BSPC, assuming the same effect of pollutant reduction level, the greater the indirect emission reduction ratio of CO 2 , the higher the co-effects. This means the higher is S 2 , the greater is the CO 2 mitigation potential of decreasing one-unit pollutant.

Uncertainty
All analyses conduced for this study are based on the meteorology of 2015. It is known that the interannual variability in meteorological conditions causes variations for individual years. In addition, in absence of local information from each province, the baseline projection of economic activities for each province up to 2030 has been developed from international data sources. Activity projections, especially the future development of the most polluting activities in Hebei (production levels of the various sectors in heavy industry, agricultural activities, etc) have important impacts on future emission projections and the mitigation potential. Further work that incorporates local information will be required to enhance the robustness of the findings.  emissions of these air pollutants in different sectors and scenarios of each region are shown in figures S1-S3 of the supporting information. Compared to the emissions in the 2020 baseline scenario, the BSPC would reduce emissions of SO 2 by 62.7 kt (equivalent to a 49% reduction of the emissions in the corresponding baseline scenario), 24 kt (13%), and 246.3 kt (23.9%), respectively, in Beijing, Tianjin, and Hebei by 2020. NO x emissions reduce by 22 kt (10%) in Beijing, 11 kt (4.1%) in Tianjin, and 213 kt (12%) in Hebei by 2020. PM 2.5 emissions reduce by 50.4 kt (52%), 22 kt (23%), and 349 kt (40%) for Beijing, Tianjin, and Hebei by 2020, respectively. In addition, the emissions of SO 2 , NO x , and PM 2.5 are also lower in the policy scenarios than in the baseline scenarios for 2025 and 2030. The BSCP can significantly reduce air pollutant emissions in the BTH region.

Projection for CO 2 emission by key sectors
The predicted CO 2 emissions of the BTH region in the baseline and policy scenarios are presented in figure 1. In 2015, Hebei produced substantially more CO 2 than Beijing and Tianjin (about 807 Mt, 92 Mt, and 169 Mt, respectively). Figure 1 shows the primary emission tendencies in the baseline and policy scenarios. CO 2 emissions of Beijing are expected to grow significantly if the BSPC is not implemented, and it is projected to increase by 14%, 39%, and 55% relative . If the BTH area achieves the policy targets by 2020, it is predicted that Beijing, Tianjin, and Hebei will reduce CO 2 emissions by 20.7 Mt (equivalent to a 19.7% reduction in the corresponding baseline scenario), 6.8 Mt (3.8%), and 80.2 Mt (9.2%), respectively. CO 2 emissions from the BTH area would be reduced by 107.7 Mt (9.3%), of which Hebei accounts for 74.5%. By 2030, it is expected that Beijing, Tianjin, and Hebei will reduce their emissions by 37.8 Mt (26.6%), 4.85 Mt (2.5%), and 69.9 Mt (8.6%) compared with the 2030 baseline scenario, respectively (figure 2). It is worth noting that Tianjin and Hebei presents smaller reductions by 2030 compared with 2020 and 2025. On the one hand, it is due to that the baseline scenario slightly decreases by 2030 in Hebei, generating smaller policy benefits. On the other hand, for both Tianjin and Hebei, it is mainly because by 2030, increased levels of activity will require more power plants to be built for the increasing power generation requirements, leading to a smaller reduction potential. CO 2 emissions from the BTH area in 2030 would be reduced by 112.6 Mt (9.8%), of which Hebei accounts for 62.0%. From 2020 to 2030, the absolute emission reductions and the relative emission reduction ratio of CO 2 will remain relatively stable in each of the BTH region. Specifically, Beijing: 21-35 Mt (20%-25%); Tianjin: 5-6.7 Mt (2.5% to 3.8%); Hebei: 70-92 Mt (9%-11%) (figure 2). This indicates that the indirect effects of BSPC on CO 2 emission reductions is stable and continuous. In each scenario, the emissions of Hebei would be several times higher than those of Beijing and Tianjin. The high emissions are mainly due to the fact that most of the industrial activity occurs in Hebei, and therefore, a greater amount of fossil fuels are consumed.
In addition, it is worth noting that although significant reductions in CO 2 emissions have been projected in Beijing when compared with the corresponding baseline scenario of each year, the CO 2 emissions in the 2030 policy scenario only show a slight increase compared to those in 2015. This shows that after the implementation of the BSPC, Beijing's CO 2 emissions have remained at a relatively stable level from 2015 to 2030. Further, in the BTH region, the CO 2 emissions have also remained at a stable level since 2015, suggesting that the CO 2 emission in BTH area peaked in 2015. This is consistent with the conclusions drawn from Zhao et al (2018), indicating that the increasing CO 2 emission trends in the BTH region from 2000 to 2014 and Yuan (2019), which show that the CO 2 emissions presents a downward trend from 2014 to 2016 based on a 2 year interval study in the BTH region. And then the CO 2 emissions will keep relatively stable (decrease within 1% compared with 2015) in 2020 and 2025, and decrease slightly by 3% in 2030 under the BSPC compared with 2015 in the BTH region. The CO 2 emissions under the policy scenarios in each region will not change significantly from 2015 to 2030 under the BSPC, however, the APPCAP would significantly reduce 352.7 Mt (39.9 ± 5.3%) of CO 2 emissions from 2012 to 2017 (Lu et al 2019). This also indicates that the peak for CO 2 emissions would have come in the BTH region. Compared with the study of Tao et al (2019) which suggests that China's total CO 2 emissions will peak by 2030, it could be concluded that the relatively stable CO 2 emissions in the BTH region contribute to the peak of CO 2 emissions being achieved before 2030 in China and the CO 2 emissions from the BTH region have a positive effect on controlling China's total CO 2 emissions.
The CO 2 emissions of the power and heating sector, which is a significant source of CO 2 emissions in the BTH region, accounts for 22.4%, 44.7%, and 32.5% of the total emissions in 2015 for Beijing, Tianjin, and Hebei, respectively. CO 2 emissions in the power and heating sector in Beijing would decline by 49% and 54% compared with the baseline scenario for 2020 and 2030, respectively, while the emissions in Tianjin and Hebei would undergo minimal change. In Beijing, the main reason for the reductions was the closure of the thermal power plants and coal-fired units in the power sector under the BSPC, which caused a dramatic decline in coal consumption. It is estimated that by 2025, Beijing's power demand will be 35 million kW, in which nearly 75% will depend on external sources (Liu 2020). The main source of external transmission is from Shanxi and Inner Mongolia, which are still largely powered by coal (Jiang et al 2019a). Therefore, it may have negative effects of indirect emissions transmission on Shanxi and Inner Mongolia based on the generation structures used in the two provinces. In Hebei and Tianjin, although some thermal power units have been renovated or shut down, the thermal power output has increased since 2017 (TSB 2019), and coal consumption in the power sector is also increasing (GCMR 2019). The increasing demand due to its own industrial structures and economic development status is consistent with the analysis report on the mid-term evaluation and future prospect of power coal control during the 13th FYP period (NCEPU 2019), which predicts that the coal consumption in China's power sector will peak in 2020 and then plateau. Current ultra-low emissions from power plants only target general air pollutants, such as SO 2 , but do not target CO 2 . Therefore, CO 2 emissions in the power and heating sector in Tianjin and Hebei would not decrease significantly by 2020 and 2030 compared with the baseline scenarios. In addition, the residential combustion sector in Beijing and the sector of industrial combustion in Tianjin and Hebei also play important roles in CO 2 emissions. Their emissions account for 27.5%, 28.7%, and 28.9%, respectively, of the total emissions in 2015.
As is shown in figure 1, significant CO 2 emission reductions was projected in the residential combustion sector, indicating that implementing energy replacement from coal to gas and electricity can achieve significant reductions to both CO 2 and air pollutants emissions (figures S1-S3). Specifically, the reductions in the residential combustion sector will account for 39%, 73%, and 57% of the total CO 2 emissions reductions by 2020 for Beijing, Tianjin, and Hebei, respectively. Further, by 2030, it is projected to accounts for 44%, 88%, and 43% for each region, respectively. In addition, the reductions in the power and heating plants of Beijing will account for 57% of the total reductions by 2020. And the reductions in the Industrial combustion sector for Tianjin and Hebei will account for 20% and 19% of the total reductions by 2020, respectively. Overall, the degrees of emission reduction potential vary in different sectors and regions.
Our results demonstrate that BSPC can reduce CO 2 emissions while reducing air pollutants. For controlling air pollutants (SO 2 , NO x , and PM 2.5 ), as shown in figures S1-S3 of the supporting information, reductions mainly occur due to the energy structure optimization and transformation (e.g. control coal consumption, coal to clean policy in residential combustion and thermal power plant sector), the transportation structure optimization and transformation (e.g. increasing new energy vehicles), and the industrial structure optimization and transformation (e.g. control of production capacity in highpollution and high-emission industries). Therefore, the optimization and transformation of these major sectors has great impacts on CO 2 emission reductions.

Clean air co-effects on CO 2 mitigation 3.2.1. Evaluation by pollutants
The metric S 2 is applied to evaluate the co-effects of CO 2 emission reductions and other atmospheric pollutants reductions (listed in table 1) under policy scenarios of different regions (figure 3). The S 2 values range from 0.38 to 2.94 in Beijing, from 0.07 to 0.99 in Tianjin, and from 0.23 to 1.20 in Hebei and BTH. The values of S 2 for BTH are close to those of Hebei because Hebei dominates the total emissions of the BTH area.
High values of S 2 indicate high synergies due to the co-effects of implementing the BSPC. Our results show that NO x has the highest co-effects with CO 2 in Beijing, Tianjin, Hebei, and BTH in each of the policy scenarios because of the higher reduction potential of CO 2 per unit of NO x emission reductions. This finding is consistent with the results from Feng et al (2018), which indicated that a co-effect for CO 2 control exists when NO x emission control measures are implemented and identified that the energy efficiency technologies have significant co-effects on reducing NO x and CO 2 emission levels simultaneously in China's cement industry with the Integrated MARKAL-EFOM System (TIMES) model. In our study, the values of S 2 for NO x in Beijing are significantly higher than those of Tianjin and Hebei, which indicates a stronger co-effect on the CO 2 emission reductions due to application of NO x control measures in Beijing. Higher S 2 value of NO x in Beijing means higher reduction potential of CO 2 emissions while reducing per unit of NO x emission.

Evaluation by key sectors
Higher S 2 values of NO x in Beijing compared with Tianjin and Hebei are mainly attributed to Beijing's implementation of policies that substantially decrease the coal consumption in power and heating and residential sectors. By 2020, the two sectors will account for 38.6% and 54% of Beijing's total reduction in NO x , and 57% and 39% of Beijing's total CO 2 reductions, respectively. In addition, for NO x reductions, the two sectors will totally account for 39% and 29% of total NO x reductions by 2020 for Tianjin and Hebei, respectively. For SO 2 reductions, the two sectors will totally account for 70% and 94% of total SO 2 reductions by 2020 for Tianjin and Hebei, respectively. Table 2 presents the CO 2 co-effects evaluated by S 2 in major sectors. The values in bold represent the maximum S 2 of each key sector in the corresponding pollutant, which indicates stronger synergetic effects. The results of S 2 show that the power and heating sector has a major synergistic role in reducing CO 2 in Beijing, which mainly due to the closure of the thermal power plants and coal-fired units in the power sector. And whether the additional external sources of electricity that Beijing traded with Inner Mongolia and Shanxi would cause their additional increase in CO 2 emissions depends on the generation structures of the electricity in the two provinces. However, the sector of residential combustion presents the highest co-effect on reducing CO 2 in Tianjin. For Hebei, the industrial combustion sector has the major synergies. The discussion in section 3.2.1 indicates that NO x and CO 2 have high synergetic effects and therefore as a major emitter of NO x , the transport sector may play a role in the co-effects in the BTH region. However, the data suggest otherwise. For example, in Beijing, although the NO x emissions from the transportation sector is substantial (figure S2), the CO 2 emission reductions from the transportation sector is minimal, which yields small values of S 2 .
It is worth noting that the co-benefits to both SO 2 and CO 2 from changes in power sector in Beijing is potential large, as indicated by the relatively large S 2 value in the power and heating sector. It indicates a higher reduction potential of CO 2 emissions while reducing per unit of SO 2 emission in the power and heating sector of Beijing, representing high synergies. However, for Tianjin and Hebei, it is the sector of residential combustion and industrial combustion that presents high co-effects, respectively. This is mainly due to the different intensity of policies implemented by each region under the BSPC, resulting in different reduction potential of air pollutants in different sectors and regions. The results show that under the BSPC, different sectors would have different reduction potentials (He et al 2018). Our results show agreement with the conclusions reached by Akimoto et al (2010), who showed that power sectors have high emission reduction potentials of CO 2 , and the conclusions reached by Lu et al (2019), who showed that the industry and residential sectors will significantly produce large amounts of CO 2 reductions due to the implementation of the APPCAP. In addition, these studies are in agreement with the research results of China's CDM projects, which indicates that the coalto-gas conversion policy in the industry enables high CO 2 reduction synergies (Nathan and Kristin 2010, Qie et al 2011).

Conclusion and policy implications
Our study demonstrates that BSPC can reduce CO 2 emissions while reducing air pollutants. These synergic results can further increase the overall benefits of the BSPC and reduce the costs for mitigating CO 2 in future. The BSPC can reduce CO 2 emissions by 20.7 Mt, 6.8 Mt, and 80.2 Mt by 2020 compared with the 2020 baseline scenario and reduce CO 2 emissions by 37.8 Mt, 4.85 Mt, and 69.9 Mt by 2030 compared to the 2030 baseline scenario in Beijing, Tianjin, and Hebei, respectively. In the BTH region, the relatively stable CO 2 emissions remained since 2015 after implementing the BSPC, suggesting that the CO 2 emission in the BTH region peaked in 2015. Therefore, it is reasonable to conclude that the relatively stable CO 2 emissions in the BTH region contribute to the peak of CO 2 emissions being achieved before 2030 in China and the CO 2 emissions from the BTH region have positive effects on controlling China's total CO 2 emissions.
Major sectors have great impacts on decreasing both air pollutants (SO 2 , NO x , and PM 2.5 ) and CO 2 emissions, which includes the structure optimization of energy and industry. Significant CO 2 emission reductions were observed in the residential combustion sector in each of the BTH region. Fully implementing these control policies for residential combustion were projected to reduce CO 2 emissions by 8.1 Mt, 6.4 Mt, and 46 Mt by 2020 and reduce CO 2 emissions by 15.7 Mt, 7.15 Mt, and 30.3 Mt by 2030 in Beijing, Tianjin, and Hebei, respectively. Significant CO 2 emission reductions were also presented in the power and heating sector in Beijing which is due to the closure of the thermal power plants and coalfired units in the power sector under the BSPC.
The results of S 2 show that NO x reduction in the BSPC has the highest co-effect with CO 2 in Beijing, Tianjin, Hebei, and BTH. The power and heating sector has a major synergistic role in reducing CO 2 in Beijing. The residential combustion presents the highest co-effects on CO 2 in Tianjin. For Hebei, the industrial combustion sector has the major synergies.
Our conclusions indicate the implementation of the BSPC, specifically the control measures in the power and heating, residential combustion, and industrial combustion sectors for Beijing, Tianjin, and Hebei, respectively, have high synergies and can simultaneously reduce CO 2 and other atmospheric emissions.

Data availability statements
All data that support the findings of this study are included within the article (and any supplementary files).