Main

In light of increasing concerns about combating climate change, China’s top government official pledged to peak national carbon dioxide emissions before 2030 and to be carbon neutral by 2060. Such an ambitious quest for carbon neutrality requires a profound domestic energy structure transformation. For instance, the share of non-fossil energy in China’s primary energy mix may need to increase from less than 16% in 2020 to over 80% by 20601,2, with solar and wind energy estimated to make up almost 37% and 36.5% of the electricity generation mix in 20603, up from merely 3% and 6% in 2020, respectively4. In contrast, coal share is expected to drop from 57% to around 10% in the coming four decades5,6.

Progressive energy (for example, renewables and/or natural gas) substitution for coal towards carbon mitigation can simultaneously reduce considerable air pollutant emissions, hence bringing remarkable air quality and associated human health co-benefits7,8,9,10,11. Although the air quality co-benefits of low-carbon energy policies have been extensively evaluated12,13,14,15, academic efforts have been started only recently to explore the potential air quality (and human health to a much lesser degree) synergies of China’s carbon neutral target5,16,17,18. In addition to emissions intensity, meteorology also plays a vital role in near-surface air pollution via atmospheric dispersion and diffusion, whereas atmospheric aerosols would substantially perturb the meteorological conditions directly through scattering and absorbing solar radiation (aerosol–radiation interactions (ARI)) and indirectly by modifying microphysical properties of clouds19. Specifically, those radiatively active aerosols, particularly light-absorbing black carbon, have been well proven to greatly deteriorate atmospheric dispersion conditions by warming the upper air and blocking sunlight that would otherwise warm the surface20,21,22. The atmospheric heating and surface dimming induced by aerosols make the air increasingly stagnant, thereby greatly weakening air pollutants dilution23. However, because most existing studies use either meteorology–chemistry decoupled (for example, WRF-CAMx) or reduced-form (for example, CHEER-AIR) regional air quality models, the underlying self-reinforcing feedback mechanisms and the amplifying air quality and human health benefits due to weakened aerosol–meteorology interactions under China’s carbon neutral target have not been fully analysed5,16,17,18,24.

Unlike the more frequently recognized air quality and human health co-benefits along with renewables penetration-induced carbon mitigation5,16,17,25, improved wind–solar power performance under China’s carbon neutral target has been much less analysed. As explained above, existing meteorology–chemistry decoupled integrated assessment frameworks are incapable of capturing the feedbacks of aerosols mitigation on meteorology and hence on solar and wind resource potential. Therefore, current related studies primarily rely on historical observational data to build statistic correlations between solar radiation and air pollution across China26,27,28,29,30. Sweerts et al. (2019)29 estimated solar energy losses owing to China’s air pollution based on historical surface radiation data for the periods of 2003–2014 and 1960–2016, respectively. Similarly, using top-of-atmosphere insolation and aerosol optical depth reanalysis data, Chen et al. (2022)30 identified the major drivers of interannual variations in solar potential for the period of 1995–2019 and discussed the possible implications for realizing China’s 2060 carbon neutral target. Despite the valuable insights from existing literature, current history-oriented statistical studies cannot fully characterize the self-reinforcing feedbacks or unravel the underlying atmospheric physics and chemistry mechanisms between low-carbon energy transition and improved renewable energy performance in the context of China’s carbon neutral pursuit.

Therefore, a systematic investigation of aerosols mitigation-induced positive feedback amplification between air pollution reduction, human health and renewable energy performance enhancement due to China’s quest for carbon neutrality is largely missing, hindering a more comprehensive understanding of the multi-aspect impacts, interactions and hence potential local motivations to pursue carbon neutrality in countries facing dual challenges of domestic air pollution and global climate change.

Here we integrate China’s carbon neutrality pathways developed from the GCAM-China (China-focused Global Change Assessment Model) and DPEC (Dynamic Projection for Emission in China) model, a meteorology–chemistry online coupled air quality model (WRF-Chem) and the Global Exposure Mortality Model (GEMM) to systematically evaluate the air quality, human health and renewable energy performance synergies from China’s ambitious target of achieving carbon neutrality by 2060. WRF-Chem model is used in this study because, unlike most other widely used air quality models (for example, GEOS-Chem), it is capable of capturing the feedbacks of changing aerosols on meteorology31,32, hence allowing us to assess future wind and solar performance responses to achieving carbon neutrality, together with the air quality and associated human health impacts, and the underlying physical mechanisms (Extended Data Table 1). More analytical details are described in Methods.

Aerosols’ impacts on air quality and renewable energy

By resolving high spatio-temporal resolution details of primary air pollutant emissions, local meteorology and chemical mechanisms, our simulations reveal considerable spatial heterogeneities in changing solar radiation, wind speed and near-surface PM2.5 (particles at or less than 2.5 μm in diameter) concentrations between with and without (w/o) considering aerosols’ radiative effects (Fig. 1). The baseline simulation is validated using nationwide air quality monitoring data (Supplementary Fig. 1), which show a reasonably good model performance in capturing the spatial pattern and temporal variations in PM2.5. Under the 2015 baseline, China has been suffering from high aerosol concentrations in both human emissions-intensive city clusters in eastern China and the dust source region in northwestern China.

Fig. 1: Aerosol distribution and its radiative effects on solar radiation, wind speed and air quality changes in 2015.
figure 1

a–d, Simulated annual mean PM2.5 surface concentrations in China (a) and changes in solar radiation (b), wind speed (c) and PM2.5 surface concentrations (d) in comparison to the hypothetical case of no aerosols’ radiative effects. Basemap data from the Resource and Environment Science and Data Center, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences (https://doi.org/10.12078/2023010103)37.

Full size image

High-loading aerosols covering China attenuate the solar energy that is supposed to heat the surface via absorbing and scattering solar radiation33. The radiative effects of atmospheric aerosols are estimated to reduce annual mean solar radiation reaching the ground surface by approximately 11 W m−2. In highly polluted regions such as the North China Plain and the Yangtze River Delta, grid-level reduction can reach over 35 W m−2, resulting in up to 15% of local solar energy density losses. Reduced solar radiation reaching the ground surface due to aerosol extinction could further lead to a substantial drop in the convective motion and thus planetary boundary layer height, which is vital for the vertical dispersion and dilution of air pollutants (Supplementary Fig. 2). The atmospheric heating by light-absorbing aerosols, such as black carbon, further compound the stagnant air conditions and cause a lower planetary boundary layer height by forming a temperature inversion. Worse still, the weakened vertical exchange consequently leads to a lower near-surface wind speed with less downward momentum transfer (Fig. 1). Under a more unfavourable atmospheric dispersion and ventilation condition, air pollutants emitted near the surface tend to be accumulated in a much shallower boundary layer, rising national area-weighted (A-W) PM2.5 surface concentrations by 0.85 μg m−3, even as high as 2.1 μg m−3 for population-weighted (P-W) concentrations. In highly polluted and densely populated regions (for example, Beijing–Tianjin–Hebei and Sichuan–Chongqing), increment in provincial-level P-W PM2.5 surface concentrations are generally around 4 μg m−3 (Extended Data Fig. 1), which may consequently result in considerable health damages. Notably, aerosol-induced solar radiation declines demonstrate a similar spatial pattern as that of PM2.5 concentrations, with provincial solar radiation reductions generally exceeding 15 W m−2 (>6%) across China’s energy demand centres (Fig. 1 and Extended Data Fig. 2). In comparison, wind speed changes demonstrate a clearly different spatial pattern and are much smaller than solar radiation changes. Grid-level wind speed and hence resource performance can either increase or decrease due to aerosols’ radiative effects, though it is dominated by decreasing trends in polluted regions due to weakened vertical flux of momentum.

We thereby highlight that the underlying aerosols’ radiative effects would not only result in notable air quality degradation but also aggravate the geospatial mismatch between national energy supply and demand due to predominant aerosol-driven reductions in solar and wind resources across China’s energy demand centres. Such aerosols’ radiative effects would therein provide critical targeted opportunities to simultaneously improve local air quality and enhance renewable energy performance along with China’s pursuit of carbon neutrality.

Amplified air quality and health under carbon neutrality

Increasing renewable substitution for fossil fuels under carbon neutrality brings progressive air pollutant emissions mitigation (Supplementary Fig. 3), causing notable PM2.5 surface concentration reductions and avoided premature deaths in comparison to the 2060 baseline: ~14.6 μg m−3 and strikingly over ~38.8 μg m−3 of A-W and P-W PM2.5 surface concentration reductions, respectively (Fig. 2, Extended Data Fig. 3, Supplementary Fig. 4 and Supplementary Table 1).

Fig. 2: Air quality and human health co-benefits under China’s carbon neutral target.
figure 2

a, Provincial P-W (coloured bars) and A-W (solid lines) PM2.5 surface concentrations in comparison to national (NAAQs, magenta dashed lines) and WHO air quality guidelines (WHO-10, green dashed lines; WHO-5, red dashed lines) under the 2060 baseline and 2060 carbon neutral scenarios across mainland China. Filled dots and open dots represent net changes in A-W and P-W PM2.5 concentrations between the 2060 carbon neutral and 2060 baseline scenarios, respectively. Magenta, green and red numbers represent the national proportion of population exceeding NAAQs, WHO-10 and WHO-5, respectively. Provinces in different power grids are illustrated in different colours. b, Provincial avoided premature deaths, with the white dots representing the mean values and the grey shadow (error bands) and error bars representing the 2.5th and 97.5th percentiles (n = 10,000).

Full size image

Such a prominent drop of PM2.5 concentration not only results from direct emissions reduction but also a more favourable dispersion condition. Quantitatively, planetary boundary layer height and ventilation coefficient in eastern China could be increased by up to ~7% and 6%, respectively, due to substantially weakened aerosol–meteorology interactions. In particular, aerosols’ radiative effects alone can amplify PM2.5 concentration reductions in Beijing, Tianjin and Hebei regions by 4.2 (5.4%), 4.7 (6.2%) and 2.8 (5.9%) μg m−3 of A-W and 7.1 (6.5%), 5.6 (6.9%), 4.3 (6.6%) μg m−3 of P-W, respectively, and 1.0 (6.5%) and 2.5 (6.3%) μg m−3 of A-W together with 2.9 (7.5%) and 3.7 (7.4%) μg m−3 of P-W PM2.5 concentration reductions in Sichuan–Chongqing regions (Supplementary Fig. 5). Across the nation, ARI amplifies on average ~0.5 and 1.5 μg m−3 of national A-W (3.8%) and P-W (4.0%) PM2.5 concentration reductions, respectively. Greater reductions in national and regional P-W than A-W PM2.5 concentrations highlight that air quality improvement resulting from China’s quest for carbon neutrality predominantly occurs in areas with large population density, thereby bringing ~1.13 (95% confidence interval: 0.97–1.29) million (~57%) reductions in PM2.5 associated premature deaths in comparison to the 2060 baseline, among which ARI alone contributes ~23,000 (21,000–26,000) reductions (Fig. 2, Extended Data Fig. 4, Supplementary Fig. 6 and Supplementary Table 2).

Without ambitious climate target and air pollution control, 780 million national population (~65%) in 2060 cannot meet China’s current National Ambient Air Quality Standards (NAAQs) for annual average PM2.5 concentrations (35 μg m−3), with 8.9 million resulting from aerosols’ radiative effect (Fig. 2 and Supplementary Fig. 7). However, when coupling carbon neutral climate policies with ambitious air pollution control (2060 Carbon neutral), except for 1.4% (17.8 million) of national population, almost the whole country can attain NAAQs. Thereby, carbon neutrality allows ~760 million of the population to reach NAAQs, with ~8.24 million contributed by reduced ARI effects. Likewise, China’s carbon neutrality pledge decreases the proportion of the population exposed to air pollution above the World Health Organization (WHO)-10 and WHO-5 air quality guidelines by 42 and 0.9 percentage points, respectively, particularly across northern and central power grids.

Self-reinforcing renewable energy synergies

Carbon neutrality-oriented aerosols mitigation can not only improve local air quality but also enhance future solar and wind resources via perturbed earth–atmosphere energy balance, particularly in China’s primary energy demand centres. Renewable energy enhancement will be increasingly beneficial in consort with rapidly growing solar and wind power penetration in China’s energy mix as the country moves towards carbon neutrality.

As indicated in Fig. 3, we observe an almost consistent renewable power increase with decreasing hourly variability across the nation supposing achieving carbon neutrality in 2060 compared with lax climate and air pollution control due to enhanced solar radiation and wind speed (that is, 2060 Baseline; Supplementary Table 3). In particular, solar radiation increases generally exceed 6% across eastern provinces, while wind speed increases are much smaller (for example, up to ~2%). As a result, provincial solar power capacity factor (CF) increase is generally above 5% (up to 10%) with improving stability (ΔCV < −0.04, coefficient of variability) in energy-demanding central, eastern and northern grids. In comparison, wind power increases upon achieving carbon neutrality are much less prominent, with the maximum provincial enhancement being around 6% and wind power variability can either increase or decrease. Across the country, central, southern and eastern grids demonstrate evident solar and wind performance enhancement with notable simultaneous CF increases and variability mitigation. Enhancing solar and wind resources in eastern provinces could thus potentially alleviate China’s geospatial mismatch between energy demand and energy supply centres (Supplementary Fig. 8). Primarily due to strong interconnections between aerosols and solar radiation and associated vertical momentum, aerosols’ feedbacks can notably enhance national solar and wind power CF by ~2.6% (0.4–9.7% across provinces) and ~0.6% (−0.1–5.7%), while increasing renewable resource stability, therein decreasing solar and wind CV by 1.2% (0.2–5.7%) and 0.4% (−0.9% to 1.7%), respectively.

Fig. 3: Renewable energy-related meteorology and performance changes under China’s carbon neutral target.
figure 3

a, Provincial solar radiation (in red) and 10 m wind speed (in blue) absolute (bars) and percentage (lines) changes upon achieving carbon neutrality in comparison to the 2060 baseline across mainland China. Please note for wind speed changes, the axis is inverted (bottom positive, top negative). b,c, Corresponding changes in solar (b) and wind power (c) performance represented by percentage changes in CF (x axis) and ΔCV (y axis). We generally find increasing resource performance and decreasing variability at the provincial level. Basemap data from the Resource and Environment Science and Data Center, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences (https://doi.org/10.12078/2023010103)37.

Full size image

Renewables to tackle the remaining environmental concerns

Due to aerosols’ feedbacks, regions (for example, central and eastern grids) receiving substantial air quality benefits are often simultaneously experiencing notable renewable energy (solar, in particular) resource enhancement (Figs. 4 and 5), which are not only densely populated and heavily polluted but are also deeply energy demanding. That is, when achieving carbon neutrality, both air quality improvement and renewable power enhancement occur in places where it is usually most needed.

Fig. 4: Multi-aspect synergies and self-reinforcing renewables under China’s quest for carbon neutrality.
figure 4

Changes in provincial PM2.5 concentrations, associated premature deaths and renewable performance of China’s carbon neutral scenario in comparison to the 2060 baseline, together with the remaining multi-aspect air pollution and health impacts upon achieving the 2060 carbon neutral target. Carbon neutrality achievement cannot fully address various environmental consequences; multi-aspect synergies and self-reinforcing renewables may provide additional local incentives for further air quality and human health improvement. (6) Remaining P-W PM2.5 and (7) Remaining premature mortality refer to a P-W PM2.5 surface concentration (6) and PM2.5 associated premature mortality upon achieving carbon neutrality (7), respectively. tce, tonne of standard coal equivalent.

Full size image
Fig. 5: Conceptual model of multi-aspect synergies and renewable self-reinforcing effects.
figure 5

Grey arrows indicate the linkages between renewables penetration, air quality, greenhouse gas emissions and renewable energy performance, primarily in eastern China. Red and green arrows show respective increase and decrease of each parameter. Dashed red arrows indicate generally weaker wind speed increases in comparison to solar radiation. Grey rings indicate the self-reinforcing renewables may provide additional incentives for post-carbon neutral renewable energy promotion to address the remaining air pollution, together with other air pollution mitigation practices (for example, efficiency improvement and emissions mitigation from industrial processes). Chinese map from the Resource and Environment Science and Data Center, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences (https://doi.org/10.12078/2023010103)37. World map underneath sourced from Natural Earth (https://www.naturalearthdata.com/).

Full size image

That said, even achieving carbon neutrality in 2060 with progressive aerosol reduction and improved regional dispersion condition, 57% (724 million) and 99% (1.25 billion) of the national population are still exposed to air pollution above WHO-10 and WHO-5, respectively, among which 6.9 million and 1.2 million are ARI induced. Therefore, air pollution exposure under carbon neutrality can still lead to 0.84 (95% confidence interval: 0.70–0.98) million PM2.5 associated premature deaths, highlighting the value of further air quality improvement (Extended Data Fig. 4). As the low-carbon energy transition and associated aerosol reduction primarily take place across eastern provinces, China’s carbon neutrality does not bring much air quality improvement within the northwestern grids (for example, Xinjiang and Gansu). These provinces thereby may end up turning to places with the worst air quality suffering from notable dust aerosols (Extended Data Fig. 3 and Supplementary Fig. 4). Meanwhile, the remaining PM2.5 associated premature deaths, upon realizing carbon neutrality, are still relatively high in eastern provinces, highlighting the value of air pollution control post-2060 (Extended Data Fig. 4). As air pollution mitigation potential from end-of-pipe controls are shrinking34, progressive low-carbon energy transitions beyond carbon neutrality may provide additional air pollution mitigation opportunities, while simultaneously improving local renewable energy performance. This, however, is likely to be possible in consort with a breakthrough in energy storage technology in the next four decades. If so, enhanced solar and wind power performance from renewable-driven aerosol reduction can provide additional local incentives for a more progressive low-carbon energy transition so as to further improve local air quality and reduce associated health damages post-carbon neutral, together with other non-fossil air pollution mitigation practices (for example, mitigation from industrial processes; Supplementary Fig. 9).

Discussion

Transforming China’s fossil-dominated energy structure towards a low-carbon energy mix is widely considered as the cornerstone to achieving carbon neutrality, which could bring considerable air quality and human health co-benefits due to increasing renewables penetration in China’s energy sector, while simultaneously enhancing renewable generation via aerosols reduction and hence weakened aerosol–meteorology interactions.

Integrating a renewable-driven carbon neutral roadmap with a regional meteorology–chemistry coupled air quality model, we find that due to both emissions reduction and aerosols reduction-induced better atmospheric dispersion, China’s renewable energy penetration under carbon neutrality will reduce national A-W PM2.5 surface concentrations by ~14.6 μg m−3, thereby reducing population exposure by as high as 38.8 μg m−3 and avoiding 57% (~1.1 million, 95% confidence interval: 0.97–1.29 million) PM2.5 associated premature deaths in comparison to lax climate and air pollution controls in 2060 (that is, 2060 Baseline), or reduce A-W and P-W PM2.5 surface concentrations by ~16.9 μg m−3 and ~47.1 μg m−3, respectively, while avoiding 66% (~1.6 million, 95% confidence interval: 1.38–1.83 million) premature deaths in comparison to 2015 baseline. Ageing populations will often completely offset the health benefits from China’s carbon neutrality target, though it could be fully compensated by reduced baseline mortality rate (Supplementary Fig. 10). Meanwhile, decreasing aerosols will perturb the earth–atmosphere balance and lead to increasing solar radiation (for example, over 6% across eastern provinces and a national average increase of ~3%) and strengthened vertical flux of momentum (for example, up to 2% increases in wind speed with a national average increase of 0.3%), resulting in increasing solar and wind power CF with decreasing variability.

Particularly, regions receiving the largest air quality, human health and renewable resource synergies are concentrated in central, northern and eastern power grids, where provincial air pollution and population exposure reductions are generally around 80% (>40 μg m−3), the corresponding provincial avoided premature deaths are mostly above 60% (~50,000) and the solar and wind power CF increases can reach up to 10% and 6%, respectively, with decreasing hourly variability. Substantially larger reductions in P-W than A-W PM2.5 surface concentrations upon achieving carbon neutrality and the predominant renewable resource increases in heavily energy-demanding eastern provinces highlight that China’s carbon neutrality-induced air quality, human health and renewable energy synergies primarily occur in places where such improvements are most needed, therefore helping to alleviate the geospatial mismatch between energy demand and energy supply.

China’s quest for carbon neutrality brings considerable multi-aspect synergies supported by progressive low-carbon energy transitions and amplified by weakened aerosol–meteorology interactions and improved atmospheric dispersion condition across eastern provinces. Yet, post-carbon neutral air quality (~12.1 μg m−3) is still beyond the most recent WHO air quality guidelines, with 57% and 99% of the population exposed to PM2.5 concentrations exceeding 10 and 5 µg m−3, respectively. Thus, carbon neutrality does not completely address all environmental concerns we evaluate here, and additional air pollution mitigation strategies may still be needed from the human health perspective. Post-carbon neutral renewables facilitation may provide additional air pollution mitigation opportunities, together with strategies such as continuous energy efficiency improvement and emissions mitigation from non-fossil sources. That said, we highlight substantial uncertainties surrounding political and public willingness worldwide to meet the stringent WHO air quality guidelines upon that time. Current PM2.5 concentrations are mostly above 5 µg m−3 across global major cities, even in those most developed regions35. By then, factors such as an increasing value of a statistical life36, the cost of available air pollution control technologies and the maturity of energy storage technology that affects the true costs of additional low-carbon energy transformation, will jointly determine the cost-effectiveness of post-carbon neutral air pollution abatement. Along that line, increasing renewables potential (especially solar) with decreasing resource variability in consort with aerosols reduction may provide extra local incentives for a more progressive low-carbon energy transition.

Our study reveals that China’s carbon neutrality pursuit can produce an important self-reinforcing feedback loop between increasing renewables substitution for fossil fuels, aerosols reduction and renewable energy performance enhancement, which is amplified by perturbed earth–atmosphere energy balance and improved atmospheric dispersion condition. Therefore, the multi-aspect air quality and human health synergies and self-reinforcing low-carbon energy development, can potentially provide important local incentives for global developing and polluting countries to progressively transform towards, and even beyond, carbon neutrality.

Methods

Scenario design

Integrating China’s carbon neutrality pathways, an online meteorology–chemistry coupling air quality model and the GEMM, our study quantifies and highlights considerable national air quality, human health and renewable energy performance synergies in China’s quest for carbon neutrality.

We obtain Chinese baseline and carbon neutral emissions scenarios from the GCAM-China (China-focused Global Change Assessment Model) and DPEC (Dynamic Projection for Emission in China) model, developed in Chen et al. (2021)5 and Tong et al. (2020)38. As described in earlier studies, GCAM-China is expanded on GCAM (that is, an integrated assessment model representing the dynamics of both human and earth system) with greater spatial resolution for China and is used to simulate province-specific energy consumption and energy mix under China’s carbon neutrality goal. DPEC model is designed in Tong et al. (2020)38 to integrate the energy system, socio-economic projection and emissions inventory, thus provides scenario-specific air pollutants and CO2 emissions. Baseline scenarios are designed with lax climate and air pollution policy following a highly unequal and isolated socio-economic pathway (SSP4–RCP6.0 with no new clean air actions beyond 2015; SSP, Shared Socioeconomic Pathway; RCP, Representative Concentration Pathway), which provides the 2015 reference point for other scenarios (that is, 2015 Baseline). Aerosol can perturb the energy budget by directly scattering or absorbing solar radiation39,40, which can further result in a notable drop in convection, lower near-surface wind speed and unfavourable air pollution dispersion condition and hence further rise near-surface PM2.5 concentrations. Multiple studies have demonstrated that higher ambient PM2.5 concentrations directly lead to lower incident solar radiation27,30,39,40, and Supplementary Fig. 15 illustrates a close link between PM2.5 concentrations and near-surface wind speed. We thereby also develop a counterfactual scenario that excludes aerosols’ radiative effects (that is, 2015 Baseline-noARI) so as to explore the underlying mechanisms of how aerosols’ radiative effects affect PM2.5 concentration reductions and renewable energy performance (that is, solar and wind power). The 2060 Baseline scenario is also designed to represent emissions changes under the business-as-usual pathway with (that is, 2060 Baseline) and without (that is, 2060 Baseline-noARI) including aerosols’ radiative effects. Meanwhile, we have the scenario designed to pursue China’s carbon neutral commitment with the best available end-of-pipe control technologies with (that is, 2060 Carbon neutral) and without (that is, 2060 Carbon neutral-noARI) including ARI (SSP1-China’s net-zero CO2 emissions in 2060) (Extended Data Table 1). Differences between with and without aerosols’ radiative effects indicate ARI’s amplifying effects, whereas differences between 2060 Baseline and 2060 Carbon neutral represent total effects of aerosols mitigation together with aerosols’ feedbacks.

Online coupled regional meteorology–chemistry simulation

We further integrate emissions scenarios with the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) to simulate PM2.5 surface concentrations under each scenario. Unlike most other chemistry–meteorology decoupled atmospheric chemistry models used in integrated assessment frameworks, WRF-Chem model is capable of capturing the feedbacks of changing aerosols on meteorology (for example, solar radiation and wind speed) and hence on solar and wind power performance. The model has been widely used to reproduce air pollution and evaluate the radiative impacts of aerosols41. Some key physical parameterization options for the model are listed in Supplementary Table 4. For the chemical scheme, the Carbon-Bond Mechanism version Z photochemical mechanism and the four-bin version of the MOSAIC (Model for Simulating Aerosol Interactions and Chemistry) aerosol model are applied. Aerosols were assumed to be spherical particles. The sizes of aerosols were divided into four discrete size bins defined by their lower and upper dry particle diameters. Aerosols in each size bin were assumed to be internally mixed. Aerosol optical properties were computed as a function of wavelength based on Mie theory. Aerosol’s radiative effects were simulated using parameters including aerosol optical depth, single scattering albedo and asymmetry factor that were online estimated in the model42.

The WRF-Chem model domain, centred at (100.0° E, 37.0° N) with a 27 km × 27 km grid resolution, covers China and its surrounding areas (Supplementary Fig. 12). There were 37 vertical layers from the ground level to the top pressure of 50 hPa in which more than 16 layers were located below 2 km to well describe the vertical structure of the boundary layer. The simulation was conducted for January, April, July and October, with averaged results from these four seasons representing annual mean conditions. For simulations driven by the 2015 meteorology, the initial and boundary conditions of meteorological fields were updated from the 1-hour ERA5 (the fifth-generation atmospheric reanalysis of the European Centre for Medium-Range Weather Forecasts) data in 2015, with a 0.25° × 0.25° spatial resolution. For simulations driven by the 2060 meteorology, the initial and boundary meteorological conditions were derived from the downscaled and bias-corrected climate model simulations in 2060, with a 6-hour temporal resolution and a 1.25° × 1.25° spatial resolution43 (Supplementary Fig. 13). Meteorological conditions are reinitiated every 48 hours to prevent drifting effects of the simulated meteorology43. The chemical outputs from previous runs were used as the initial conditions for the following run. A similar modelling configuration and settings have been successfully adopted in our previous works32. Both natural and anthropogenic emissions were included for the regional WRF-Chem modelling in the present work. We applied the aforementioned DPEC dataset for characterizing anthropogenic emissions, in which emission sources were classified into five sectors: power plants, residential combustion, industrial process, on-road mobile sources and agricultural activities. The natural dust emissions were calculated online using Goddard Chemistry Aerosol Radiation and Transport (GOCART) module based on surface wind speed44.

As illustrated in Supplementary Fig. 1, the baseline simulation is validated using nationwide air quality monitoring data, which show a reasonably good model performance in capturing the spatial pattern and temporal variations in PM2.5. Spatial and temporal correlation of wind speed and solar radiation are validated against the ERA5 reanalysis data (Supplementary Figs. 14 and 15).

On the basis of simulated grid-level PM2.5 surface concentrations, we estimate provincial and national area-weighted\(\left({C}_{\mathrm{AW}}\left(R\right)=\frac{{\sum }_{i=1}^{n}{\mathrm{Area}}_{i}\,\left(R\right)\times{C}_{i}\left(R\right)}{{\sum }_{i=1}^{n}{\mathrm{Area}}_{i}\left(R\right)}\,\right)\) and population-weighted \(\left({C}_{\mathrm{PW}}(R)=\frac{{\sum }_{i=1}^{n}{\mathrm{Pop}}_{i}\,\left(R\right)\times{C}_{i}\left(R\right)}{{\sum }_{i=1}^{n}{\mathrm{Pop}}_{i}\left(R\right)}\,\right)\) PM2.5 surface concentrations in comparison to China’s national air quality standards (NAAQs) (35 μg m−3), the previous (10 μg m−3) and recently updated (5 μg m−3) WHO air quality guidelines for annual average PM2.5 surface concentrations. \({C}_{\mathrm{AW}}(R)\) and \({C}_{\mathrm{PW}}(R)\) refer to national or provincial area-weighted and population-weighted PM2.5 surface concentrations in region R (province or the whole country), Areai(R) and Popi(R) refer to grid area and population in each grid i within region R and Ci(R) refers to model-simulated grid-level PM2.5 surface concentrations in grid i in region R. Population-weighted PM2.5 surface concentration (CPW\()\) is often referred to as population exposure. Supplementary Table 5 shows China’s inter-provincial regional power grids45.

Air pollution associated human health impacts evaluation

We then further quantify grid-level PM2.5-associated premature mortality under each scenario using equations (1) and (2). We extract grid-level age-specific population for China in 2015 under different SSP scenarios from Wang and Sun (2022)46 and retrieve national baseline mortality rates from the Global Burden of Diseases study 2017 (https://gbd2017.healthdata.org/gbd-search/). We generate 10,000 shapes of relative risks (RR) and calculate the 95% confidence intervals of premature mortality under each scenario by perturbing the value of θ based on its mean and standard error described in Burnett et al. (2018)47.

$${\mathrm{Mort}}_{j,i}={\mathrm{POP}}_{j,i,k}\times{\mathrm{Mortbase}}_{j,k}* ({1-1/{\mathrm{RR}}}_{j,i})$$
(1)

where Mortj,i is the number of premature mortality in grid i from disease j, POPj,i,k is the number of exposed targeted population in grid i from disease j for adults’ age group k (for example, 25–29), Mortbasej,k is the baseline mortality rate for respective adults group k for disease j. RRj,i is the relative risk for disease j in grid i. Here we use more recent GEMM developed in Burnett et al. (2018)47, which covers the global range of exposures and is widely considered to be suitable for evaluating health impacts for China’s high PM2.5 concentrations. GEMM quantifies premature mortality for noncommunicable diseases and lower respiratory infections for adults (≥25 years old), with the threshold of no health effects at 2.4 μg m−3. Equation (2) below estimates the corresponding RR under GEMM. Supplementary Table 6 summarizes the parameters used in the GEMM model.

$${\rm{RR}}({{z}})={{\rm{e}}}^{\frac{{\rm{\theta }}\times \log \left(\frac{{{z}}}{\rm{\alpha }}+1\right)}{1+{{\rm{e}}}^{-\frac{{{z}}-{\rm{\mu }}}{{{v}}}}}},{{z}}=\max (0,{{c}}-2.4)$$
(2)

In equation (2), c represents the long-term ambient PM2.5 concentration; θ, μ and v represents parameters that determine the shape of RR in GEMM and are specified for each age group; e is natural logarithm; 2.4 μg m−3 is the threshold for PM2.5, below which there will be no effect; and z is defined as the max value between 0 and c − 2.4.

In addition to GEMM, we use two other different health impacts models (GEMM-5 diseases and integrated exposure response (IER) functions) to evaluate human health impacts as sensitivity analysis. GEMM-5 diseases model is constructed in Burnett et al. (2018)47 that includes the same five diseases as IER: ischaemic heart disease, stroke, chronic obstructive pulmonary disease, lung cancer and lower respiratory infections. Detailed parameters for GEMM-5 diseases are listed in Table S2 in Burnett et al. (2018)47, and the same RR equation (equation (2)) is used for GEMM-5 diseases models. Disease-specific IER RRs are also obtained from the global burden of disease studies for age-specific population groups; then equation (1) is used to quantify disease-specific premature deaths. We further obtain the age distribution and baseline mortality rate changes in 2060 from earlier studies48, with the respective age distribution and mortality rate under 2060 SSP4 used for the 2060 baseline scenario and under the 2060 SSP1 used for the 2060 carbon neutral scenario according to how these two scenarios are designed (Extended Data Table 1).

Solar and wind power performance evaluation

We calculate the capacity factor (CF) for solar power using scenario-specific outputs from the WRF-Chem model, applying the same method used in Jerez49 and Feron50 (equations (3)–(5)). All variables used and their corresponding units are listed in Supplementary Table 7. CF illustrates the performance of the photovoltaic cell under actual environmental conditions in comparison to rated power capacity. I and ISTC (1,000 W m−2) indicate downward shortwave radiation (W m−2) under actual environmental conditions and standard test conditions, respectively; whereas PR indicates the effects of actual temperature on the performance of photovoltaic panels. Tcell (°C) is the cell temperature, TSTC is the cell temperature under standard test conditions (25 °C) and γ equals 0.005 °C−1. Tcell relates to ambient temperature at 2 m (T2m, °C), downward shortwave radiation (I, W m−2) and wind speed at 10 m (u10m, m s−1), where c1 = 4.3 °C, c2 = 0.943, c3 = 0.028 °C (W m−2)−1, c4 = 1.528 °C (m s−1)−1 (ref. 49). In equations (3)–(5), the subscripts i, j, k denote the variable value at (i, j) position and at the hour k.

$$\mathrm{CF}_{i,j,k}={P}_{\mathrm{R}_{i,j,k}}\left(\frac{{I}_{i,j,k}}{{I}_{\mathrm{STC}}}\right)$$
(3)
$${P}_{\mathrm{R}_{i,j,k}}=1-\gamma ({T}_{\mathrm{cell}_{i,j,k}}-{T}_{\mathrm{STC}})$$
(4)
$${T}_{\mathrm{cell}_{i,j,k}}={c}_{1}+{c}_{2}{T}_{2\mathrm{m}_{i,j,k}}+{c}_{3}{I}_{i,j,k}-{c}_{4}{u}_{10\mathrm{m}_{i,j,k}}$$
(5)

Wind power CF is evaluated based on the atmospheric kinetic energy equation of the atmosphere and an idealized power curve for wind turbine using equations (6) and (7)51. Assuming the wind turbine hub height is of 100 m, the wind speed at 100 m (u100m) is derived from wind speed at 10 m (u10m) using the wind profile formula:

$${u}_\mathrm{100m}={u}_\mathrm{10m}{\left(\frac{100}{10}\right)}^{0.14}$$
(6)

Following earlier studies52, we assume the cut-in wind speed (uin) is 3 m s−1 (below the cut-in speed the wind power turbine will not rotate), the cut-out wind speed (uout) is 25 m s−1 (above the cut-out speed the wind power turbine will shut down for safety reasons), and the rated wind speed (urated) is 12 m s−1; hourly capacity factor is calculated by equation (7) (the meaning of i, j, k are the same as above):

$${\mathrm{CF}}_{i,j,k}=\left\{\begin{array}{c}0({u}_{\mathrm{100m}_{i,j,k}} < {u}_{\mathrm{in}}\,{\mathrm{or}}\,{u}_{{\mathrm{100m}}_{i,j,k}} > {u}_{\mathrm{out}})\\ \displaystyle\frac{{u}_{{\mathrm{100m}}_{i,j,k}}^{3}-{u}_{\mathrm{in}}^{3}}{{u}_{\mathrm{rated}}^{3}-{u}_{\mathrm{in}}^{3}}({u}_{\mathrm{in}}\le {u}_{{\mathrm{100m}}_{i,j,k}}\le {u}_{\mathrm{rated}})\\ 1({u}_{\mathrm{rated}} < {u}_{{\mathrm{100m}}_{i,j,k}}\le {u}_{\mathrm{out}})\end{array}\right.$$
(7)

We further use the coefficient of variation (CV) to quantify the hourly variability of solar and wind energy capacity factors. The hourly coefficient of variation is defined as the ratio of the hourly CF standard deviation to the hourly average CF in each month (equation (8))30.

$$\mathrm{CV}=\frac{\mathrm{std}(\mathrm{CF})}{\mathrm{avg}(\mathrm{CF})}$$
(8)