Air pollution disparities and equality assessments of US national decarbonization strategies

The study addresses how national electricity decarbonization strategies, typically designed under least-cost optimization, affect the distribution of air pollution across demographic and socioeconomic groups within the United States. While decarbonization will reduce aggregate emissions, it may exacerbate or alleviate existing inequalities in exposure to co-pollutants (NOx, SO2, PM2.5). The authors frame the work within distributional energy justice, focusing on equitable distribution of risks and benefits from energy transitions. They identify a gap in electricity planning models, which often consider environmental metrics at system or national levels and largely omit social dimensions such as equality and justice. The purpose is to develop a forward-looking framework to quantify how the benefits of decarbonization (reductions in co-pollutants) and the remaining burdens are distributed across vulnerable groups (by race/ethnicity, income, poverty) under alternative national policy pathways, thereby informing policies that can achieve both climate and equality goals.

Prior work shows that national-level GHG mitigation often yields co-benefits in reduced PM2.5, but the spatial and demographic distribution of these co-benefits remains uncertain. Historical analyses document higher PM2.5 exposure among low-income, Black, Asian, and Hispanic communities due to legacy policy and siting inequities. Mainstream power sector planning tools (e.g., least-cost capacity expansion models) typically exclude equality considerations or treat them post hoc, potentially missing local impacts. Some studies integrate equality or inter-country equity into planning objectives, or internalize health damages, but often do not examine distributional air pollution outcomes for distinct demographics at high spatial resolution or across diverse decarbonization scenarios. The review underscores a need for forward-looking, subnational equality assessments linked to decarbonization planning that capture distributional outcomes across vulnerable groups.

The analysis couples a national capacity expansion model with environmental sustainability and equality assessments over 2010–2050. Electric power system model: The National Renewable Energy Laboratory's Regional Energy Deployment System (ReEDS) is used to simulate least-cost electricity generation mixes subject to load, operational, and transmission constraints. ReEDS solves sequentially by year (limited foresight) and implements exogenous policy constraints. Decarbonization scenarios: Eight scenarios are evaluated: A) Base case with existing policies only; B) US NDC carbon cap; C) 1.5 °C pathway carbon cap; D) 80% renewable by 2050; E) 100% renewable by 2035; F) 100% renewable by 2050; G) 100% low-carbon (RE + NG-CCS + nuclear) by 2035; H) 100% low-carbon by 2050. Technology costs follow NREL 2019 ATB mid-case. Renewable and low-carbon mandates ramp linearly from 20% in 2020 to the target year. Environmental sustainability (operating emissions): Annual national and regional operating emissions are computed as E = Σ_n g_{nt} e_n, where g_{nt} is generation by technology n and year t, and e_n is the operating emission rate (g/kWh). Emissions include CO2, NOx, SO2, and PM2.5. Renewable and nuclear operating emissions are assumed zero; technology-specific rates (e.g., coal, gas CT/CC, CCS, biopower) are sourced from literature (Table 2). Air pollution modeling and equality assessment: Power sector emissions by ReEDS region are downscaled to the Intervention Model for Air Pollution (InMAP) spatial grid via area-weighting, then transported and transformed using InMAP to estimate annual average ambient concentrations of NOx, SO2, and total PM2.5 (primary + secondary) across a variable 1–48 km grid. Concentrations are averaged to census tracts and combined with demographic data from the American Community Survey (ACS) to compute population-weighted concentrations for groups by race/ethnicity, poverty rate bands, and income categories. Groupings include income intervals (e.g., <$25k, $100k–$125k, >$150k) and poverty-rate bins (0–10% up to 70–100%). The analysis examines temporal trends (2020–2050) and scenario differences in exposure disparities.

- Under no new decarbonization policies (Scenario A), Black and high-poverty communities face 0.19–0.22 µg/m³ higher PM2.5 concentrations than the national average during the transition and up to 26–34% higher exposure relative to national averages over 2020–2050. - National mandates exceeding 80% renewable or 100% low-carbon deployment achieve near-equality in ambient concentrations across demographic groups when mandate years are met. Scenarios C (1.5 °C cap), E (100% RE by 2035), and G (100% low carbon by 2035) achieve PM2.5 <0.25 µg/m³ across all regions by 2035; Scenarios F and H achieve this by 2050. - In 2020, PM2.5 exposures >1.0 µg/m³ are concentrated in the Midwest and Eastern US. By 2035, only scenarios with aggressive caps/mandates (C, E, G) reduce all regions below 0.25 µg/m³; by 2050, F and H also reach this threshold. - Aggregate emissions: Scenario A provides an upper bound across pollutants. By 2035, Scenarios C, E, and G have CO2 operating emissions <100 Mt; coal is entirely retired by 2035 and gas or gas-CCS contributes <10% of generation, with NOx ≤0.02 Mt, SO2 <0.003 Mt, and PM <0.002 Mt. In Scenario E, PM2.5 rises from 2035 to 2050 due to biopower deployment to sustain 100% RE, so Scenarios C and F have lower PM2.5 by 2050. - Generation patterns: In Scenario A by 2050, coal = 7.5% (0.41 PWh), natural gas = 20.0% (1.08 PWh), onshore wind = 33.8% (1.83 PWh), solar PV = 20.9% (1.14 PWh). Scenario C retires coal by 2035 and reduces gas to ~0.2% by 2050; carbon cap in Scenario B allows coal to rise from 1.67% (2040) to 4.67% (2050). - Racial disparities: Black communities have the highest PM2.5 exposure until technology mandates exceed ~80% or national CO2 falls below ~400 Mt. Scenarios C and E reduce racial inequality fastest (by ~2030). By 2050, even the least-improving policy case (Scenario B) lowers Black community PM2.5 to 0.25 µg/m³ versus 0.38 µg/m³ in the base case, implying at least a 33% improvement versus Scenario A in 2050. - Poverty disparities: Census tracts with >70% poverty have the highest PM2.5 concentrations until >80% RE or 100% low-carbon deployment is achieved; without additional policies (Scenario A), high-poverty tracts experience up to 26% higher PM2.5 over the transition. - Income disparities: Differences in PM2.5 and NOx exposures between highest and lowest income groups are <0.2 µg/m³; SO2 disparities are larger (up to 0.3 µg/m³ before 2040 in Scenario A), likely due to coal plant siting. - Efficiency of co-pollutant reduction: Scenario D (80% RE by 2050) has the highest ratio of PM2.5 concentration reduction per unit CO2 reduction among racial/ethnic groups, indicating relatively faster local pollutant improvements per CO2 decrease, despite not reaching zero emissions. - All scenarios with carbon-oriented policies outperform the base case, yielding at least a 20% reduction in co-pollutant concentrations for the lowest income group in 2050 versus Scenario A.

The findings demonstrate that least-cost decarbonization pathways can reduce national emissions yet still perpetuate local exposure disparities until aggressive caps or mandates are fully realized. Equality in ambient air pollution exposure is not achieved before mandate years (2035 or 2050) in the evaluated scenarios. Trade-offs emerge: sustaining a 100% renewable system via biopower can reintroduce local SO2 and PM2.5 emissions that affect nearby communities and partially undermine equality. The 1.5 °C carbon cap scenario yields fewer national emission reductions than some 2035 technology mandates by mid-century but still achieves equalized ambient concentrations (<0.25 µg/m³) by 2050, similar to 100% RE (2050) and low-carbon mandates. Historical exposure patterns strongly influence distributional outcomes under least-cost planning, disproportionately burdening Black and high-poverty communities during the transition. The results highlight the need to integrate equality objectives or strict technology mandates into planning to ensure equitable distribution of air quality benefits, and to anticipate trade-offs between aggregate decarbonization targets and local co-pollutant exposures.

This work introduces a forward-looking framework that couples least-cost capacity expansion with reduced-complexity air quality modeling to evaluate distributional equality of co-pollutant exposures across US demographic groups under eight national decarbonization strategies. It shows that strict technology mandates (≥80% renewable or 100% low-carbon) or aggressive carbon caps are most likely to deliver equality in ambient concentrations, but only once mandate years are reached. Without explicit equality-focused objectives, least-cost planning can leave Black and high-poverty communities with higher exposures during the transition. Future research should explore alternative optimization paradigms that embed equality in objective functions, examine trade-offs among equality (e.g., exposure, costs), environmental (e.g., land, water), and economic objectives, quantify health damages and mortality outcomes, improve spatial allocation of emissions to plant level, include dynamics in demographics and migration, and extend analyses across sectors beyond electricity.

- Demographics static over time: Population (2010) and equality metrics (2018) are held fixed; no modeling of migration or demographic shifts, potentially affecting future exposure distributions. - ReEDS modeling uncertainty: Sequential, least-cost solves with limited foresight introduce uncertainty in plant siting, generation, technology costs, demand evolution, and policy changes over a 40-year horizon. - Spatial allocation of emissions: Area-weighting from ReEDS regions to InMAP grids spreads emissions uniformly within regions and does not represent plant-level locations or stack parameters. - Reduced-complexity air quality model: InMAP entails approximation error (e.g., mean fractional bias ≈ -17%), though validated within acceptable bounds; secondary formation and deposition are simplified relative to full CTMs. - Sectoral scope: Only electricity sector emissions are considered; disparities from transportation, industry, and residential sources are excluded, which can be significant for some groups (e.g., Hispanic, Asian, Indigenous communities). - Context: Results pertain to a developed-country electricity system; implications may differ in emerging economies. - Assumptions on operating emissions: Renewable and nuclear operating emissions assumed zero; biopower and CCS rates are literature-based with inherent variability.