Air pollution disparities and equality assessments of US national decarbonization strategies

Abstract

Energy transitions and decarbonization require rapid changes to a nation's electricity generation mix. There are many feasible decarbonization pathways for the electricity sector, yet there is vast uncertainty about how these pathways will advance or derail the nation's energy equality goals. We present a framework for investigating how decarbonization pathways, driven by a least-cost paradigm, will impact air pollution inequality across vulnerable groups (e.g., low-income, minorities) in the US. We find that if no decarbonization policies are implemented, Black and high-poverty communities may be burdened with 0.19-0.22 μg/m³ higher PM_2.5 concentrations than the national average during the energy transition. National mandates requiring more than 80% deployment of renewable or low-carbon technologies achieve equality of air pollution concentrations across all demographic groups. Thus, if least-cost optimization capacity expansion models remain the dominant decision-making paradigm, strict low-carbon or renewable energy technology mandates will have the greatest likelihood of achieving national distributional energy equality. Decarbonization is essential to achieving climate goals, but myopic decarbonization policies that ignore co-pollutants may leave Black and high-poverty communities up to 26-34% higher PM_2.5 exposure than national averages over the energy transition.

Fig. 1. Annual generation mix (PWh) 2010–2050 by technology for each decarbonization scenario resulting from the ReEDS model.

We highlight that the renewable and low carbon technology mandates accommodate additional energy needs primarily through expanded wind and solar generation investments. We see that the base case, US NDC, and 80% renewable energy decarbonization pathways retain coal generation through 2050.

Fig. 2. National operating emissions across scenarios 2010–2050 (megatonnes, Mt).

The emissions shown here are: a CO₂ operating emissions, b NO_x operating emissions, c SO₂ operating emissions, and d PM_2.5 operating emissions. Note that the y-axes are not consistent. We see the base case (black line) as an upper bound on all emissions types and Scenarios C (solid yellow line) and E (dotted blue line) as a lower bound across all emissions types. Scenario E emissions reach close to zero by 2035, its mandate year, and remains close to zero by 2035–2050 for CO₂, NO_x, and SO₂ emissions. However, PM_2.5 emissions in Scenario E rise because of investments in biopower, which help maintain the 100% renewable grid but still have co-pollutants that will be emitted.

Fig. 3. Regional distribution of PM_2.5 across decarbonization scenarios for 2020, 2035, and 2050.

Total PM2.5 air pollution from power plants across scenarios: a 2020, b 2035, and c 2050. By 2035, the only scenarios with 100% of regions below the threshold of 0.25 µg/m³ include the aggressive carbon cap (Scenario C) and the ones with a technology mandate with a 2035 goal of 100% low carbon or renewable technologies (Scenarios E and G). The distribution of NO_x and SO₂ for 2020, 2035, and 2050 can be found in SI Figs. S-8 and S-9.

Fig. 4. Population weighted average annual PM_2.5 concentrations (in µg/m³) across different race and ethnicity groups for each scenario from 2020 to 2050.

We see that Black communities in the US are exposed to higher concentrations of PM_2.5 in 2020, which is consistent with historical impacts. Over the energy transition, Black communities are exposed to higher concentrations of PM_2.5 until a technology mandate is >80% renewable energy (Scenario D in 2050, Scenario E in 2035, or Scenario F in 2050), 100% low carbon energy (Scenario G in 2035 and Scenario H in 2050), or carbon cap that requires emissions of CO₂ <400 Mt (Scenario C in 2030).

Fig. 5. Population weighted average annual PM_2.5 concentration across poverty rates within census tracts for each scenario 2020–2050.

We see that census tracts with poverty rates >70% are burdened with the highest concentrations of PM_2.5 across all scenarios and timelines. Figures S-18 and S-19 show the NO_x and SO₂ concentrations across poverty rate groups, respectively.

Fig. 6. Population weighted annual average PM_2.5 across each scenario 2020 to 2050 for the highest (>$150k), mid ($100k-$125k), and lowest (<$25k) income groups in µg/m³.

The other income groups fall in between the highest and lowest bounds. Results from all scenarios can be found in SI Fig. S-12, and NO_x and SO₂ across income groups in Figs. S-13 and S-14, respectively.

Fig. 7. Overview of equality analysis methods using a capacity expansion model (ReEDS) and a reduced complexity air pollution transport model (InMAP) to investigate where emissions are settling after being emitted from power plants.

Regions are downscaled from ReEDS regions to census tract level, and annual average air pollution concentrations are estimated from emissions at the ReEDS region level. Figure S-2 shows InMAP emission inputs. The shapefiles for the maps produced in this image are sourced from NREL ReEDS, InMAP output file, and ArcGIS.