全球变化知识资源中心

Carbon dioxide emissions standards for US power plants will influence the fuels and technologies used to generate electricity, alter emissions of pollutants such as sulphur dioxide and nitrogen oxide, and influence ambient air quality and public health. We present an analysis of how three alternative scenarios for US power plant carbon standards could change fine particulate matter and ozone concentrations in ambient air, and the resulting public health co-benefits. The results underscore that carbon standards to curb global climate change can also provide immediate local and regional health co-benefits, but the magnitude depends on the design of the standards. A stringent but flexible policy that counts demand-side energy efficiency towards compliance yields the greatest health benefits of the three scenarios analysed.

On 2 June 2014, the US Environmental Protection Agency (EPA) proposed CO₂ emissions standards for existing power plants in the Clean Power Plan¹. When finalized in summer 2015, affected states will use the federal standards to develop state implementation plans for decreasing CO₂ emissions from the power sector. As an abundant greenhouse gas, CO₂ is a major contributor to climate change. Power plants in the USA fired by fossil fuels emitted 2 billion tonnes of CO₂ in 2012², representing 39% of total national emissions — more than any other single source. Standards to reduce CO₂ emissions for existing US power plants can result in near-term public health benefits locally and regionally by decreasing emissions of co-pollutants, including sulphur dioxide (SO₂), nitrogen oxides (NO_x), mercury (Hg) and fine particulate matter (PM_2.5).

We linked power sector model results with air quality and epidemiological models to quantify the air quality and public health benefits of changes in emissions of co-pollutants under different scenarios for power plant carbon standards. The analysis is based on emissions estimates for each of the 2,417 fossil-fuel-fired power plants in the USA, from the Integrated Planning Model (IPM), for a reference case and three policy scenarios (http://www.icfi.com/insights/products-and-tools/ipm; Supplementary Information: Emissions modelling). These emissions estimates were used as inputs for the spatially explicit Community Multiscale Air Quality Model (CMAQ v. 4.7.1) to project resulting changes in air quality at a 12 × 12 km resolution for the continental USA (http://www.epa.gov/AMD/Research/RIA/cmaq.html; Supplementary Information: Air quality modelling). The CMAQ results for ozone (O₃) and PM_2.5 were used as inputs for the Environmental Benefits Mapping and Analysis Program-Community Edition (BenMAP-CE v. 1.08) to estimate public health co-benefits for each scenario compared to the 2020 reference case (http://www.epa.gov/airquality/benmap/ce.html; Supplementary Information: Health co-benefits modelling). We isolate the co-benefits attributable to the carbon standards by comparing changes in air quality and health co-benefits in the year 2020 for each scenario with a reference case that includes all existing and planned air quality policies for the power sector. The results show that, for two of the three policy scenarios, carbon standards for existing power plants can substantially decrease emissions of harmful co-pollutants, and improve air quality and public health beyond what would occur under existing air quality policies.

To facilitate comparison with the goals of the Clean Power Plan, we report estimated changes in CO₂ emissions to 2005 levels, the baseline year used in the plan. The Bipartisan Policy Center (BPC) and the Natural Resources Defense Council (NRDC) developed the reference case that was used for our analysis. We selected two policy scenarios that were generated by BPC (scenarios 1 and 3) and one that was developed by NRDC (scenario 2). As we were interested in a wide range of policy approaches, researchable scenarios were selected that incorporate contrasting policy assumptions. The policy differences in the scenarios include different approaches to CO₂ emissions reductions, investments in end-user energy efficiency, and inclusion of options for compliance such as co-firing, fuel-switching and cross-state trading.

The reference case uses the energy demand projections in the Annual Energy Outlook for 2013³ as the benchmark. Current EPA clean air policies are fully implemented under this scenario, including the Mercury and Air Toxics Standard (MATS) and the Clean Air Interstate Rule. Moreover, existing state-level requirements for power sector emissions reductions and renewable energy portfolio standards are implemented under this scenario. By 2020, minor changes in energy generation sources under the reference case result in an estimated decrease in annual CO₂ emissions of 15.2% compared with 2005 levels (Table 1).

Full table

Detailed boiler unit-level IPM emissions were used for the reference case and the three scenarios as input to CMAQ to estimate anticipated changes in air quality associated with changing power plant emissions. We used CMAQ output to determine spatial patterns of expected changes in ground-level O₃ and PM_2.5 for 2020. These pollutants have well-understood health and environmental consequences that are documented extensively in the peer-reviewed literature^{5, 6}.

Scenario 1 results in a modest increase in average annual PM_2.5 (Fig. 2a) and peak ground-level O₃ concentrations (Fig. 3a) compared with the reference scenario. This pattern of 'emissions rebound' at several coal-fired power plants occurs when facilities that exhibit high emissions are made more efficient and therefore run more frequently and for longer periods than in the reference case⁷.

a, Scenario 1; and b, scenario 2.

We used the PM_2.5 and O₃ concentrations from the CMAQ air quality simulations for the continental USA and compared them with the 2020 reference case to estimate and map the health co-benefits for each of the policy scenarios. These estimates do not include the direct health benefits resulting from mitigating climate change (for example, reduced heat-related illness). Concentration–response functions were derived for six health co-benefit outcomes, on the basis of extensive published literature on the health effects of air pollution. The six outcomes are: PM_2.5-related changes in premature deaths; myocardial infarctions (heart attacks); cardiovascular hospital admissions (excluding myocardial infarctions); respiratory hospital admissions; O₃-related changes in premature deaths; and hospital admissions associated with respiratory illness. We selected this subset of health outcomes from the numerous effects associated with PM_2.5 and O₃ because they are supported by concentration–response functions derived from investigations that examined populations from multiple cities simultaneously under different conditions across the USA, large cohort studies of residents from different locations, or meta-analyses of studies that have taken place in many different locations. These health outcomes contribute to most of the monetized benefits accompanying air quality management^{4, 8, 9, 10, 11}.

In BenMAP-CE, we linked data on population, age structure, baseline prevalence and incidence rates of the health co-benefit outcomes of interest to estimate changes in outcomes at the county and state levels for the continental USA for each of the three carbon standard scenarios, compared with the 2020 reference case. We report the central estimate and 95% confidence intervals for each health outcome, based on only concentration–response function uncertainties, given a lack of quantitative information on other model uncertainties. Population data are from Woods & Poole¹²; baseline hospitalization and myocardial infarction data are from the Healthcare Utilization and Cost Program¹³; and mortality rate projections for 2020 are from the US Centers for Disease Control and Prevention WONDER database (http://wonder.cdc.gov/natality-current.html).

The concentration–response functions we derived relate changes in air quality to changes in the rate of an adverse health outcome (Supplementary Information: Concentration–response functions). The functions are based on published epidemiological literature (Supplementary Table 1) and are expressed as a change in the risk of each outcome per unit concentration change of a given pollutant over a given time period. Unless indicated otherwise, we based all values shown here on central estimates.

Our results show that scenario 1 has the lowest health co-benefits in the continental USA of the three scenarios considered (Table 2). Under this scenario, estimated decreases in hospitalizations were modest and there was a slight increase in premature deaths and heart attacks from the 2020 reference case. This represents a negative co-benefit of 10 additional premature deaths per year (Table 2), which corresponds to −0.2 premature deaths avoided per million tonne decrease in CO₂ (Fig. 1). This pattern is likely to be due to the increase in SO₂ emissions and resulting PM_2.5 concentrations that are projected for this scenario.

Full table

The estimated health co-benefits vary widely across the USA and under the three scenarios, with all states experiencing some benefit under scenario 2. For all three scenarios, areas with the highest health benefits have the greatest air quality improvements and large exposed populations.

Scenario 1 results in small changes in the number of premature deaths relative to the 2020 reference case for most counties (Supplementary Fig. 2a). At the state level, based on central estimates, the health co-benefits include 21 to −33 premature deaths eliminated annually (Fig. 4a), 5 to −10 hospitalizations averted per year and 2 to −2 heart attacks avoided each year.

a, Scenario 1; and b, scenario 2.

Different policy approaches to US carbon standards for power plants produce markedly different changes in PM_2.5 and O₃, and associated health co-benefits. The magnitude and direction of the changes in health co-benefits parallel the changes in annual emissions of SO₂ and NO_x for each scenario (Fig. 1). In each scenario, the geographic distribution of state-level health co-benefits is consistent with air quality changes coupled with population distribution (Figs 2–4; Supplementary Fig. 2). Our analysis shows that the design of carbon standards for US power plants can have a marked impact on air quality and associated health outcomes for local communities and states. Scenario 2 — which is the most similar of our three scenarios to the Clean Power Plan proposal of the EPA in terms of stringency, policy structure and anticipated changes in power generation — results in the greatest estimated emissions reductions, air quality improvements and health co-benefits (Fig. 1). Its top performance is due to lower total fossil fuel generation, greater substitution of natural gas for coal and more new demand-side energy efficiency. In contrast, carbon standards that largely rely on retrofitting existing power plants, as illustrated in scenario 1, could increase SO₂ emissions from the power sector, resulting in potential increases in air pollution beyond what is expected to occur in the reference case. As illustrated by scenario 3, a lower ratio of health co-benefits per tonne of CO₂ emissions controlled can occur when the standards result in carbon pollution controls that continue or increase reliance on coal generation by means of CCS, and provide no new programmatic investment in demand-side energy efficiency.

Carbon standards implemented for existing US power plants that result in improvements in air quality can lead to immediate local and regional health co-benefits. For the USA and other countries with sizeable greenhouse-gas emissions along with air pollution challenges, the link between climate policy, air quality and public health could provide a key catalyst to act on climate change.

1. Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units 79 FR 34829 (US EPA, 2014); https://federalregister.gov/a/2014-13726
2. National Greenhouse Gas Emissions and Sinks: 1990–2013 Publication No. 430-R-14–003 (US EPA, 2014); http://go.nature.com/GSeOar
3. Annual Energy Outlook 2014 with Projections to 2040 (DOE/EIA, 2014); http://go.nature.com/BPZObe
4. Regulatory Impact Analysis for the Proposed Carbon Pollution Guidelines for Existing Power Plants and Emission Standards for Modified and Reconstructed Power Plants Publication No. EPA-452/R-14-002 (US EPA, 2014); http://go.nature.com/M5JBzy
5. Jerrett, M. et al. Long-term ozone exposure and mortality. New Engl. J. Med. 360, 1085–1095 (2009).
   • CAS
   • ISI
   • PubMed
   • Article
6. Fann, N. et al. Estimating the national public health burden associated with exposure to ambient PM_2.5 and ozone. Risk Anal. 32, 81–95 (2012).
   • ISI
   • PubMed
   • Article
7. Burtraw, D., Linn, J., Palmer, K. & Paul, A. The costs and consequences of greenhouse gas regulation under the Clean Air Act. Am. Econ. Rev. 104, 557–562 (2014).
   • ISI
   • Article
8. Muller, N. Z. & Mendelsohn, R. Efficient pollution regulation: Getting the prices right. Am. Econ. Rev. 99, 1714–1739 (2009).
   • ISI
   • Article
9. Fraas, A. & Lutter, R. Efficient pollution regulation: Getting the prices right: Comment. Am. Econ. Rev. 102, 602–607 (2012).
   • ISI
   • Article
10. The Benefits and Costs of the Clean Air Act from 1990 to 2020 (US EPA, 2011); http://www.epa.gov/cleanairactbenefits/feb11/summaryreport.pdf
11. Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards Publication No. EPA-452/R-11-011 (US EPA, 2011); http://go.nature.com/6cBliV
12. Population by Single Year of Age (Woods & Poole Economics, 2008).
13. HCUPnet, Healthcare Cost and Utilization Project (Agency for Healthcare Research and Quality, 2007).

Download references

Financial support for this work was provided by the William and Flora Hewlett Foundation, the Grantham Foundation and Mistra's Indigo Program. The authors thank colleagues M. Weiss, K. Driscoll, M. Hale, J. Macedonia, S. Pan, S. Sekar and S. Yeh.

Affiliations

1. Department of Civil and Environmental Engineering, Syra