Air quality and heath co-benefits of low carbon transition policies in electricity system: the case of Beijing–Tianjin–Hebei region

OSeMOSYS model is designed to calculate a portfolio of energy technologies that meet the lowest net present value cost of an energy system for a given demand and constraint (Howells et al 2011). The BTH electric power system constructed simplifies the real electric power system including generation, transmission, distribution and use. We assume that the electricity in BTH is produced by 12 power generation technologies and each power station uses the single type of fuel. Different from existing studies that merely considered local power supply, we take EELE as an important source of power supply to cover the gap in power demand according to the actual situation. Considering the difference of power load between seasons, working days or non-working days, day and night, the model has 16 time slices to describe time, and the time accuracy of the model takes 2 years as a step.

As shown in figure 1, four kinds of energy, coal, natural gas, biomass and garbage, are supplied to the power station, and the energy conversion from chemical energy to electricity is completed through 12 kinds of power generation technologies. Both the external power and the power generated by the power station need to be transported to different substations through the power grid, and then distributed to various power terminals.

Zoom In Zoom Out Reset image size

Technology and economic parameters, technology exploitable capacity, installed capacity, CO₂ emission factor, capital cost of power generation technologies, fuel costs, electricity demand reserve rate, and discount rate are introduced in section S1 and table S1 of supplementary information (SI).

Based on energy use estimated by the OSeMOSYS model under different scenarios, fuel related air pollutant emissions, such as SO₂, NO_X, VOC, and PM_2.5, are calculated by emission coefficients which refer to Zhang et al (2021). Air pollutant emissions in BTH electric system under the BAU scenario are based on the Air Benefit and Cost and Attainment Assessment System-Emission Inventory version 2.0 (ABaCAS-EI v2.0) (Li et al 2023). The historical changes in air pollutants in ABaCAS-EI v2.0 have been deeply explored from the aspects of 70 control policies, source contributions, and spatial-temporal characteristics. More details of air pollutant emissions could be found in the section S2 of SI. We use the ERSM prediction system, built by Tsinghua University (see Zhao et al 2018, Li et al 2023), to estimate PM_2.5 concentrations in BTH area through 2060.

An ERSM model, effectively quantifies the response of PM_2.5 concentrations to various emission changes through building a 'real-time' relationship between concentrations and precursor emissions on the basis of abundant simulations on chemical transports. A pf-ERSM, a multiregional prediction system, addresses the indirect effects of interaction among regions in accord with regional emission changes through incorporating polynomial functions. Specifically, the response of PM_2.5 concentrations includes local chemical generation and regional transportation as well as interaction among regions. The pf-ERSM prediction system is based on the chemical transports simulated by community multi-scale air quality model with a grid spacing of 27 km * 27 km in East Asia for January, April, July, and October, four typical months. The response variables, provincial annual PM_2.5 concentrations, are calculated by the average of four typical months, associated with control factors including provincial emission rates of VOC, SO₂, NOx, ammonia (NH₃), and PM_2.5. Thus, we use pf-ERSM prediction system to estimate BTH PM_2.5 concentrations.

Scenario setting is a key part of low-carbon pathway analysis. We identify five key factors that affect the decarbonization of power systems and set up three main categories of scenarios, totaling 12 scenarios (see table 1).

The BAU scenario examines the model's choice of power generation technology in the context of a rapid decrease in the cost of renewables in the future, and assumes that renewables is only used 80% of the technology exploitable capacity. Technology is in continuous development and progress, and phased assumptions are made on energy efficiency, transmission efficiency and capacity factor on the basis of the techno-economic assumptions in the previous section, and the remaining parameters remain unchanged. These parameters are listed in table S2 of SI.

There are 6 single scenarios: (1) technological progress scenario (Tech). Technical progress includes the improvement of capacity factor, the improvement of energy efficiency and the development of energy storage (Liu et al 2023). The parameters involved in energy efficiency include power generation efficiency and transmission efficiency. Referring to the 'Power Generation Costs 2020' released by IRENA (IRENA 2020), set future capacity factor based on historical trends in past capacity factor. The parameter settings are summarized in table S2 of SI. (2) Carbon pricing scenario (Carb). This scenario sets two different levels of carbon prices which change as shown in figure S3 of SI. The national carbon market trading data is obtained from Tianjin Climate Exchange, and the national carbon price is predicted based on the historical trading price trend of mature pilot areas, referred to as the low-carbon price scenario (L-Carb). The EU carbon market is the most mature market in the world. Currently, the EU carbon price has broken through the 100-euro mark. Drawing on the EU carbon price level, we assume that China's carbon price can reach $120/t by 2060, referred to as high-carbon price scenario (H-Carb). (3) Gas-fired generation scenario (Gas). With the continued retirement of coal-fired power units, gas-fired power generation will gradually replace coal-fired power generation, the flexible power supply will be completely powered by gas. Therefore, the maximum capacity of coal-fired power will be gradually set to 0 at the later stage of the modeling period, and a completely coal-free renewable energy power system will be obtained. (4) Renewable Energy Development scenario (RE). The installed capacity of renewable energy reaches the maximum limit that can be utilized, that is technology exploitable capacity. The scenario highlights the increasing proportion of electricity generated from renewable sources. At the same time assuming that the degree of external transfer of electricity to further improve the cleanliness, in 2050, the proportion of green power transfer 100%. (5) CCUS scenario (CCUS). The installation of CCUS in advanced coal-fired power units can capture 90% of CO₂ emissions. In this scenario, all generating units that emit CO₂ in 2060 will be equipped with CCUS. Here, referring to the hypothesis of Chen et al (2021) on CCUS cluster deployment of China's coal-fired power plants under the carbon neutrality goal, CCUS technology would be widely applied in 2030. The cost of installing CCUS in thermal power units will be 712 $/KW, while the cost of biomass power generation units will increase by 850 $/KW, with an annual decrease rate of 3%. The above costs refer to the cost assumption of installing CCUS units by Du et al (2023).

The implementation effect of a single policy has certain limitations, which would be not enough to achieve the goal of carbon neutrality. The single scenarios thus are combined to enhance carbon reduction effectiveness. Carbon pricing and CCUS both use price transmission mechanisms to force the transformation of the energy and power system. We assume that carbon pricing and CCUS would be only chosen between them. Thus, five combined policy scenarios are designed as follows. (1) S1 represents comprehensive technological progress and renewable energy. Explore the impact of two positive promoting effects on the transformation of energy and power systems; (2) S2 represents a combination of technological advances, carbon pricing, and renewable energy. (3) S3 considers gas-electric substitution on the basis of S2, that is, a high-proportion renewable energy power system with gas-electricity as the main flexible power supply under the comprehensive influence of multiple factors. (4) S4 considers renewable energy, technological advancements, and CCUS technologies in a comprehensive manner. And (5) S5 considers gas-electric substitution on the basis of S4.

The change in installed capacity of gas and electricity presents an 'inverted U-shaped' pattern from 2020 to 2060. The installed capacity under the H-Carb and Gas scenarios during 2020–2060 exhibits an increasing trend, and changes in installed capacity under RE scenario is relatively stable. Table S4 of SI shows key indicators including total installed capacity, percentage of reproducible installation, coal-fired power generation, and gas-fired power generation, etc., under BAU and single scenarios in 2060. The total installed capacity under the single scenarios (except RE and Tech) generally fluctuates around the BAU scenario of 310.5 GW in 2060. The installed capacity of RE reaches up to 350.8 GW in 2060, and the installed capacity of Tech scenario are only 282 GW. The differences in installed capacity under the single scenarios are related to the mechanisms of the five factors as follows. The Tech could reduce CO₂ emissions by reducing energy use during generation, transmission and distribution, increasing the available capacity of the power plant and meeting more electricity demand with the same capacity. Increasing the installed capacity of renewable energy can reduce the installed capacity of thermal power, thereby reducing the CO₂ emissions of the energy power system. Replacing coal-fired power generation with gas and electricity aims to reduce CO₂ emissions by changing the proportion of clean energy installed. CCUS technology has a relatively small impact on the power structure. RE and Tech scenarios have a relatively high proportion of renewable energy installed capacity, which can reach around 83%, fully demonstrating the importance of these two factors in improving energy structure.

Trends in the installed structure associated with the combined scenarios are consistent with the corresponding single scenarios. The proportion of installed capacity of renewable energy in S1 and S2 would be 90.9% in 2060. S3 achieves the highest proportion of renewable energy power systems with 91.8% in 2060, followed by S4 with 91.3%. Based on the installation of coal, gas and biomass power under combined scenarios (see table S6), compared to S1–S3 scenarios, S4 and S5 scenarios adding CCUS units gradually decrease the installed capacity of traditional coal-fired power, while the installed capacity of traditional biomass energy would be only reduced after the expansion of the installed capacity. This indicates that biomass power generation would shift towards low-carbon intensive BECCS power plants after experiencing a mature period in its initial scale, that is, BECCS power plants are with cost advantages in 2042. S4 and S5 are generally considered to be the most ideal energy and power optimization paths, which are specifically introduced in figure 2. However, S4 results in 26.53 GW of coal-fired power installed capacity in 2060, which is higher than the coal-fired power installed capacity associated with S1 and S2, but lower than the BAU level (40.5GW). This is because that CCUS technology delays the retirement of coal-fired power generation. Under S5, importantly, all coal-fired power generation would be out of the power system in 2060.

Zoom In Zoom Out Reset image size

Figure S6 of SI depicts CO₂ emissions of the electric power generation under the BAU and single scenarios in BTH area from 2020 to 2060. Under the BAU, CO₂ emissions of the electric power generation in BTH region generally exhibit an 'inverted U-shaped' trend during the period of 2020–2060, reaching a peak (428.3 Mt) in 2038 and gradually decreasing associated with an increase in renewable energy generation capacity. However, the amount of CO₂ emissions in 2060 is over 341.5 Mt, which is still far from achieving carbon neutrality target. In the single scenario, except for the L-Carb price scenario, which will not achieve the goal of peaking CO₂ emissions before 2030, all other scenarios will peak in 2028. CCUS scenario results in the largest reductions in CO₂ emissions through 2060, while other single scenarios have limited reductions, especially L-Carb, almost coinciding with the BAU curve.

All of the composite scenarios can peak before 2030, and S4 and S5 scenarios have great potential to achieve carbon neutrality in 2060. Compared to CO₂ emissions reductions associated with the single scenarios, combined scenarios result in larger CO₂ emission mitigation (see figure 3). For example, S1 reduces 87.6 Mt more emissions than the RE scenario, 122.5 Mt more emissions than the Tech, and 210.1 Mt more compared emissions than the BAU. The CO₂ emissions from 2020 to 2060 under S2 and S3 scenarios are at similar level, although the flexible power supply under these scenarios is different (one is ultra-supercritical coal-fired power generation, the other is natural gas combined cycle power generation). This indicates that the emission reduction effect of gas-fired power replacing coal-fired power is limited. Similarly, in the comparison of S4 and S5, the magnitude of CO₂ reductions associated with gas electricity replacing coal-fired power is still small. Thus, changes in CO₂ emissions under S4 and S5 are similar through 2060. Under combined scenarios, CCUS, the main technology to keep the flexibility of power system, is more effective than carbon pricing. In 2060, CO₂ emissions associated with S4 and S5 would be reduce to about 12 Mt. Carbon neutrality in 2060 would be fully achieved if direct air capture, biomass carbon cycle, and carbon sink, etc. are taken into account.

Zoom In Zoom Out Reset image size

Similar to CO₂ emissions, all air pollutant (SO₂, NO_X , VOC, and PM_2.5) emissions of power generation sector in BTH region through 2060, under all policy scenarios are reduced relative to the BAU. This finding is supported by previously studies that suggested that the air quality co-benefits of some climate policy options exist (West et al 2013, Thompson et al 2016, Li et al 2018, Zhang et al 2024). Explanation of air pollutant reductions is associated with a decrease in fuel use in power generation sector, which is similar to CO₂ emission reductions. Thus, PM_2.5 concentrations in BTH area under all policy scenarios decrease at different level. The combined scenarios decrease more PM_2.5 concentrations in BTH area compared to the single scenarios.

Figure 4 shows regional PM_2.5 concentrations in absolute and relative terms in 2060 under combined scenarios. S4 and S5 which potentially achieve carbon neutrality in 2060 improve air quality in BTH region more effectively than S1–S3. While the regional average PM_2.5 concentrations are 43.94 μg m⁻³ in 2060 BTH area under BAU scenario, S4 and S5 drop PM_2.5 conventions by 13.95% and 14.56%, respectively. Specifically, S4 and S5 results in Beijing and Tianjin exposing to PM_2.5 concentrations below 35 μg m⁻³ in 2060 (meeting the national Class II ⁹ standard for PM_2.5 conventions), while Hebei is likely to expose to PM_2.5 concentrations over 48 μg m⁻³. This is because that air pollutant emissions of thermal power generation in Hebei are not the main cause of air pollution. This indicates that energy-intensive sectors in Hebei, such as nonmetallic mineral product manufacturing, smelting and pressing of metals and chemical industry, calls for attention. Furthermore, the differences of regional PM_2.5 concentrations in BTH area under S1–S5 exceed these under the BAU scenarios. For instance, S4 and S5 enlarge regional differences in PM_2.5 concentrations to more than 18 μg m⁻³ in 2060, while the BAU level is merely 12.53 μg m⁻³ in 2060. This implies that regionally differentiated policies for air pollution control would be better than the uniform policies to promote environmental equity (Zhang et al 2021, 2024). Thus, the BTH area would make coordinated efforts to the improvement in air quality based on their own characteristics of industry and resources. Figure S5 shows changes in air quality in BTH area associated with the single low carbon transition policies in electricity system from the BAU level in 2060. The single scenarios result in BTH area exposing to average PM_2.5 concentrations at the range of 38.5–43.91 μg m⁻³ in 2060. This indicates that single scenarios are far inferior to the combined scenario, especially S4 and S5. Thus, S4 and S5 policies are highly recommend for regional air quality improvement.

Zoom In Zoom Out Reset image size

Improvement in air quality benefits to human health (Thompson et al 2014, Zhang et al 2023). Following the estimation of Cao et al (2011) that an increase of 10 μg m⁻³ PM_2.5 concentrations leads to a 0.9% increase in cardiovascular, total respiratory, and lung cancer mortality, we calculate avoided mortalities under different low carbon transition policies in electricity system in 2060 BTH area from the BAU level as shown in supplementary table S7. Combined policies (S1–S5) contribute to about 5192, 5192, 5422, 8189 and 8528 avoided premature mortalities in BTH area in 2060, respectively. The level of regional health benefits associated with combined policies is much higher than that of regional health benefits associated with the single policies in 2060, which is caused by more air quality improvement under combined scenarios as above. Human health benefits in BTH area associated with S4 and S5 policies in 2060 are much larger than the benefits associated with S1–S3 policies, especially S5. The level of human health benefits in BTH area in 2060 under S5 is higher than that of Zhang et al (2023). This is because S5 which is highly recommended is consist with more fuel related effective measures and policies. Additionally, human health benefits in Beijing under each policy scenario are larger than these in Tianjin and Hebei, which indicates that further end-of-pipe control measures for Tianjin and Hebei to reduce non-combustion air pollutant emissions are needed.