Skip to main content
NCBI home page
GeoHealth logoLink to GeoHealth
. 2025 Sep 16;9(9):e2025GH001431. doi: 10.1029/2025GH001431

Source‐Specific Air Pollution Emissions Inequalities From 2011 to 2020 in Virginia

Lucas R F Henneman 1,✉, Ryah Nadjafi 1, Xiaorong Shan 1, Jenna R Krall 2
PMCID: PMC12439277  PMID: 40964502

Abstract

Air quality has improved in recent decades across most of the United States. However, decreases in pollution have not been uniform, potentially exacerbating inequalities in air pollution exposure by race and ethnicity. These inequalities exist, in part, because of spatial differences in source(s), for example, power plants or roadways. Determining which sources are driving inequality across racial and ethnic groups is critical to determining which policies (e.g., targeting power plant vs. vehicle emissions) would reduce inequalities. Our study determines which pollutant sources should be decreased to address inequalities in four pollutants (NOx, SO2, VOCs, and PM2.5) in the Commonwealth of Virginia. We derived emissions from eight source categories for 134 Virginia counties from the National Emissions Inventory and the MOtor Vehicle Emissions Simulator mobile source emissions model. We used race and ethnicity data from the American Community Survey from 2011 to 2020. We applied the Atkinson Index to obtain a single summary of inequality for each source‐pollutant pair (e.g., NOx from electricity generation) across all race and ethnic groups. Most source category emissions were unequally distributed for at least once pollutant. Compared to other sources, electricity generation resulted in the largest inequalities across pollutants. Mobile sources increased in inequality from 2011 to 2020 even as emissions decreased. These results show the importance of identifying sources that contribute most to inequalities when developing policies to promote environmental justice.

Plain Language Summary

Our study unfolds in Virginia, where lawmakers passed the Environmental Justice Act in April 2020 to promote environmental justice. The Act's implementation is challenging because it is unclear how specifically to modify pollution levels to reduce inequality. To inform the Act's implementation, we connected information about where air pollution sources are located with population groups that lived nearby. We found that certain sources are more likely to emit air pollution near populations of certain racial/ethnic groups. Most source categories—including dust and fires, power plants, industrial activities, cars, and residential sources—emit at least one pollutant closer to either Hispanic or Black populations than other groups. Air pollution emissions changes since 2011 improved equality for some sources but not others. For example, emissions from power plants decreased but power plant emissions are still more likely to be near Hispanic and Black people. Similarly, automobile emissions decreased but became more unequally distributed, meaning the gap between the most and least exposed group increased over time. By identifying sources whose emissions are more likely to be breathed in by certain populations, we have provided evidence to promote environmental justice in Virginia. Our approach establishes a framework for similar assessments in other locations.

Key Points

  • Inequality was examined for source categories including dust and fires, electricity generation, industry, mobile, and residential

  • Compared to other source categories, electricity generation resulted in the largest inequalities across pollutants

  • Mobile sources increased in inequality from 2011 to 2020 even as emissions decreased

1. Introduction

Exposure to air pollution is associated with adverse health outcomes including mortality (Samet et al., 2000; US EPA, 2016, 2017, 2022), and air pollution continues to pose a risk to human health despite decreasing concentrations in the US (Dominici et al., 2022; Wu et al., 2020). Air pollutants known to impact health include fine particulate matter (PM2.5) and gaseous nitrogen oxides (NOx) and sulfur dioxide (SO2). NOx and volatile organic compounds (VOCs) react in the atmosphere to form gaseous ozone, another pollutant known to impact health (Seinfeld & Pandis, 2006). Harmful pollutants are emitted from a wide range of sources, including both anthropogenic sources (e.g., cars, power plants, industrial facilities) and natural sources (e.g., dust and wildfires, which are both influenced by human activity) (Seinfeld & Pandis, 2006; Thakrar et al., 2020). Types of sources (“source categories”) have unique spatial patterns (McDuffie et al., 2020, 2021), for example, power plants emissions are from discrete points, and automobile emissions are primarily from roadways.

Communities of color bear a disproportionate share of these pollutant exposures that negatively affect health and life span. Exposure inequality stems from a combination of historical industrial, governmental, and commercial operations and/or policies (Gohlke et al., 2023; Levy et al., 2006; Tessum et al., 2019). For example, communities of color and Hispanic communities have been specifically targeted with new polluting sources, contributing to disparities in exposure to environmental pollutants by race and ethnicity (Cushing et al., 2021; L. R. F. Henneman et al., 2023; Morello‐Frosch & Obasogie, 2023; Perera, 2018; Tessum et al., 2019). Redlining, the 20th century government‐endorsed practice of offering loans depending on prospective homeowners' skin color, locked people of color into “undesirable” neighborhoods, which were often near air pollution sources (Gonzalez et al., 2023; Lane et al., 2022). Systemic inequalities in pollution exposure between communities of color and white communities exist across all income levels, with racial disparities in PM2.5 exposure being similar for low‐income and high‐income groups (Jbaily et al., 2022; Tessum et al., 2021; Wang et al., 2022). Mikati et al. (2018) found that the difference in proportional burden of point source particulate matter emissions between Black residents and White residents exceeded the difference in proportional burden between those living below and above poverty, suggesting that the disparity between racial groups is not entirely explained by differences in poverty levels (Mikati et al., 2018). These disparities result in different health impacts attributable to pollution exposure by race and ethnicity (Bowe et al., 2019).

Previous studies have assessed exposure inequalities over time using either individual pollutants from all sources (e.g., total PM2.5 mass or NO2) (Kerr et al., 2023; Lane et al., 2022; Tessum et al., 2021) or individual pollutants from specific source categories (e.g., PM2.5 from coal‐fired power plants) (L. R. F. Henneman et al., 2023; Kerr et al., 2024). Yet, comparing between source categories is important for developing an understanding of which source categories are driving inequality and comparing source‐specific policies aimed at reducing disparities. Additionally, because each source category emits a unique combination of air pollutants, source‐specific health impacts vary. Thus, comparing exposures to pollutant‐source categories (e.g., NOx from fires with NOx from industrial sources) facilitates an understanding of which sources are driving inequality, while accounting for sources as complex mixtures of environmental pollutants.

Since the 1990 Clean Air Amendments, which accounted for disparities in air pollution‐related health effects in Section 108(a) (2) by highlighting “…the kind and extent of all identifiable effects on public health…,” presidents have used Executive Orders (EOs; e.g., EO 12898 in 1994 and EO 14096 in 2023) to make environmental justice (EJ) policy more explicit (Executive Office of the President, 1994, 2023). While federal support for addressing race‐based disparities has varied over time, states such as Virginia have also used legislation to create policies to address disparities. The Commonwealth of Virginia enacted the Environmental Justice Act in April 2020 to promote environmental justice, yet its implementation is challenging because it is unclear how to modify pollution levels to reduce inequality. Virginia defines EJ as “the fair treatment and meaningful involvement of every person, regardless of race, color, national origin, income, faith, or disability, regarding the development, implementation, or enforcement of any environmental law, regulation, or policy.” The definition of an “Environmental Justice Community” as defined in the Act applies to 53% of Virginia's geographic area and 59% of its population. Identifying specific regulatory targets, for example, air pollution sources that impact some communities more than others, would provide an avenue to promote environmental justice in Virginia.

This study aims to quantify inequalities in emissions by pollutant‐source categories, that is, combinations of pollutants (NOx, PM2.5, SO2, VOCs) and eight source categories (e.g., power plants, wildfires, industry) from 2011 to 2020 in Virginia. Our approach leveraging the Atkinson Index quantifies holistic disparities considering multiple racial and ethnic groups to enable direct comparisons among source‐pollutant pairs. To our knowledge, this is one of the first studies to identify pollutant‐source pairs most responsible for disparities, which can inform source‐based strategies to reduce existing environmental injustice.

2. Methods

Annual county‐level air pollution emissions data were obtained from the United States Environmental Protection Agency's (EPA) National Emissions Inventory (NEI) for the four most recent years available: 2011, 2014, 2017, and 2020 (U.S. EPA, 2021). We extracted emissions of NOx, PM2.5, SO2, and VOCs. These pollutants were chosen because they are EPA‐regulated criteria pollutants (NOx, PM2.5, and SO2), and/or they contribute to secondary ozone or PM2.5 formation (NOx, SO2, and VOCs). We summed NEI source sector emissions to eight broad categories intended to align with existing regulatory programs or related natural sources. These include [names used throughout are in brackets]: agriculture and biogenics [Agr. and Biogenics], dust and fires [Dust and Fires], electricity generation [Electricity], industry [Industry], non‐road mobile [NR Mobile], on‐road mobile—diesel [OR Mobile (Diesel)], on‐road mobile—non‐diesel [OR Mobile (non‐diesel)], and residential [Residential]. The list of original NEI sectors assigned to our eight broad source categories can be found in Table S1 in Supporting Information S1.

In each NEI year, EPA used a different version of its MOtor Vehicle Emissions Simulator (MOVES) model to calculate mobile source emissions estimates for mobile sources which would create inconsistencies when using NEI for mobile sources (U.S. EPA, 2021). To improve consistency of on‐road mobile source emissions estimates (which constitute a large fraction of emissions for multiple pollutants, especially NOx and VOCs), we replaced the original NEI emissions with temporally‐consistent emissions estimated using MOVES3 (U.S. EPA, 2020) (the most recent version of MOVES at the start of this work), which incorporates Virginia‐specific vehicle miles traveled data (method described in SI). Our application of MOVES3 produces larger emissions than NEI in 2011, with similar magnitudes in the other years, with larger differences observed in counties with larger vehicle miles traveled (Figure S1 in Supporting Information S1). These comparative results are consistent with EPA's comparisons between MOVES3 and previous versions (Foley et al., 2023).

To assess disparities in pollution by race and ethnicity, we obtained county and census tract racial (Census table B02001) and ethnic (table B03003) populations from the American Community Survey (ACS) using the tidycensus package in R for years 2011, 2014, 2017, and 2020. From these tables, we obtained eight racial groups [American Indian and Alaska Native Alone (referred to here as “AIAN”), Asian Alone (“Asian”), Black or African American Alone (“Black”), Native Hawaiian and Other Pacific Islander Alone (“NHPI”), Some Other Race Alone (“Some other race”), Two or more races, and White Alone (“White”)] as well as ethnic groups including Hispanic or Latino (“Hispanic”) versus non‐Hispanic.

2.1. Atkinson Index and Inequality

We used the Atkinson Index (AIx) to quantify differences in exposure across racial and ethnic groups. We assigned pollutant‐source exposure as the emissions in each county. This approach assumes that exposure is greatest to those living closest to the emissions source, that is, people living within the county. This and other distance‐based approaches have been applied throughout the literature (Cushing et al., 2021; Mikati et al., 2018; Nunez et al., 2024). Such approaches have the benefit of relating polluting sources to nearby populations, and limitations of this method are discussed below.

The AIx measures relative inequality on a scale from zero to one, where zero represents complete equality amongst all members of a population, and one represents complete inequality, and thus provides a concise summary of inequality for each pollutant‐source category. Unlike other approaches (e.g., average emission burden or exposures for each population group), the AIx provides a single summary to quantify inequality across multiple racial or ethnic groups. Therefore, it can be used as a screening tool to identify those pollutants and sources that are driving inequality. Although initially used to calculate income inequality, the AIx has been used to quantify inequality in pollution including for exposure to ambient PM2.5 and NO2 and overall PM2.5 health burdens (Fann et al., 2011, 2018; Levy et al., 2006; Rosofsky et al., 2018). To adapt AIx to the study of pollution as in previous work, we applied AIx to the inverse exposure (emissions or pollution), so that greater values are more desirable. AIx is subgroup decomposable (i.e., it enables quantifying each group's inequality relative to other groups), which facilitates comparing inequalities between race and ethnic groups (Levy et al., 2006). This between‐group AIx is defined as:

AIx=1−∑j=1nfjyj′‾y′‾1−ε11−ε (1)

where n is the number of population subgroups (e.g., number of racial or ethnic groups), fj is the fraction of the subgroup out of the population, yj′‾ is the average inverse exposure for each subgroup in the population, and y′‾ is the average inverse exposure of the population. The inequality aversion parameter ε measures the degree of societal concern about inequality and has a range of zero to infinity.

Although AIx ranges from 0 to 1, the specific magnitude of the between‐group AIx can be difficult to interpret because it is dependent on the number of subgroups, each subgroup's value (yj′‾), and the inequality aversion parameter (ε). The AIx is effectively a weighted power mean, where the power is related to ε. The weights in the power mean are the fractions in each population group (e.g., large or small proportions of the population) and the quantity being averaged is the relative inverse exposure value (e.g., amount of “clean air”) for each population group. To add intuition for interpreting AIx, we applied the AIx to simulated data for two groups, with various differences in air quality between the groups and varying inequality aversion ε (Figure 1). As the difference in air quality between the groups grows (more inequality), so does the AIx, and there is a leveling off at higher differences between groups that is dependent on the inequality aversion, ε. The darkest line at the bottom corresponds with a low societal preference for reducing inequality, and the AIx in that case is always 0. The lightest line at the top denotes a high societal preference for reducing inequality, and the AIx is 1 for any differences between groups (the line jumps from 0 to 1 as soon as a difference between groups exists). In general, relative to lower values of ε, higher ε (more aversion to inequality) yields AIx closer to 1, indicating greater inequality.

Figure 1.

Figure 1

Open in a new tab

Atkinson Index differences by tuning parameter ε.

To be consistent with the air quality environmental justice literature, we calculated the AIx for two aversion parameters: ε = 0.75 and ε = 2 (Fann et al., 2018; Rosofsky et al., 2018). Because our primary goal was to consider overall between‐group inequality, we calculated the between‐group AIx across all eight racial groups, leveraging AIx's ability to provide one single measure of inequality across all groups. We also separately calculated AIx between (a) Hispanic and non‐Hispanic and (b) Black versus all other to target inequality related to historical policies targeting these groups.

In sensitivity analyses, we tested the influence of the county‐level spatial resolution in the main analysis, and we used population‐weighted exposure (PWE) as an estimate of burden to complement the AIx calculations. We assessed PWE of (a) census tract‐level inequalities in total emissions for NOx, VOC, PM2.5, and SO2 using a fine resolution emissions data set and (b) county and census tract‐level inequalities in PM2.5 and NO2 ambient concentrations using fine‐scale concentration estimates.

In the first sensitivity analysis, we used fine‐scale (1 km × 1 km) total emissions from the Neighborhood Emissions Modeling Operation (NEMO) (Ma & Tong, 2022), which is derived from the 2017 NEI, but does not provide source‐specific emissions. We used area‐weighting to assign NOx, VOC, PM2.5, and SO2 emissions to each Virginia Census tract and county. While EPA maintains measured and estimated facility‐specific emissions information, we use spatially aggregated emissions to protect against potential exposure error that would accompany ignoring atmospheric transport in assigning facilities to populations living, for example, in the same administrative boundary.

Air pollutant emissions are highly relevant for regulatory policies seeking to reduce emissions from specific sectors, whereas air pollution concentrations reflect ambient exposures relevant for human health. In the second sensitivity analysis, to compare results for our main investigation of emissions versus air pollution concentrations, we incorporated publicly available PM2.5 and NO2 concentrations derived from a combination of ground‐based and satellite observations and chemical transport modeling. PM2.5 concentrations (Shen et al., 2024) were available in 2020 on a 0.01° × 0.01° resolution grid, and NO2 concentrations (Cooper et al., 2022) were available in 2019 on a 1 km × 1 km resolution grid. We assigned average concentrations to census tracts and counties using area weighting.

PWE for a given demographic subgroup j is defined as:

PWEj=∑i=1Iyi∗Pj,iPj,total (2)

where i=1,2,…,I denotes locations (counties or census tracts), Pj,i is the population (P) of a given demographic j in each location, and Pj,total is the total population of the demographic group (Harper et al., 2013; Kerr et al., 2021; Levy et al., 2006; Tessum et al., 2021). The exposure, yi, is the county (or census tract) value. We defined PWEall as the overall population weighted exposure in each location, using the proportion of the total population regardless of race and ethnicity. We assessed the absolute difference between a group's PWE and the total population as:

PWEjabsolute=PWEj−PWEall (3)

We also defined the relative difference between a group's PWE and the total population as:

PWEjrelative=PWEjPWEall (4)

PWEjabsolute is in emissions units, with a value of 0 representing no difference between subgroup j’s exposure and the reference population. PWEjrelative is in fractional units, with a value of one denoting no difference from the reference population.

3. Results

Virginia is diverse, with substantial differences between counties in racial and ethnic distributions; for example, in 2017, Virginia counties ranged from (0.39%, 77%) Black, (0.72%, 39%) Hispanic, and (0.17%, 20%) Asian (Figure S2 in Supporting Information S1). Asian and Hispanic populations were more concentrated within the Northern Virginia region near Washington, DC. In contrast, Black residents of Virginia were more concentrated in the central and southern counties.

There were marked differences in emissions by pollutant species, consistent with pollutants known to be emitted by each source (Figure 2). NOx was predominantly emitted by mobile sources (36% in 2020, the most recent NEI year available); PM2.5 by dust and fires, electricity generation, and industry (together contributing 69% in 2020); SO2 by electricity generation and industry (89%); and VOC from agriculture and biogenics (81%). Virginia achieved substantial NOx and SO2 emissions reductions between 2011 and 2020, with total NOx emissions decreasing by 56% driven by decreases in mobile pollution primarily attributable to adoption of NOx control technologies (Frey, 2018) and SO2 decreasing by 84% driven by decreases in coal use in electricity generation (L. R. F. Henneman et al., 2023). Comparatively, PM2.5 and VOC emissions varied from year‐to‐year without a consistent trend across the study period. Notably, decreases in dust and fire PM2.5 emissions were balanced by increases in electricity generation PM2.5.

Figure 2.

Figure 2

Open in a new tab

Air pollution emissions by source in Virginia in 2011, 2014, 2017, and 2020. On‐road mobile emissions were calculated using MOVES3 and the remaining sources were from EPA's National Emissions Inventory.

Total emissions of all pollutants were along transportation corridors and in metropolitan areas, and NOx, VOC, and PM2.5 emissions were elevated in population‐dense Washington, DC suburbs and in the eastern part of the state near Richmond and Hampton Roads (Figure S3 in Supporting Information S1). For most source categories in 2020 (the most recent year), spatial distributions of county emissions are similar across pollutants, aligning with expectations that the spatial emissions variability is driven by the source locations (Figure 3). Industrial and residential sources stand out as having unique spatial distributions across pollutants because of the heterogeneity present in these sources. Unlike emissions from most other sources, agriculture and biogenics emissions were highest in some rural areas, reflecting the expected distribution of those sources.

Figure 3.

Figure 3

Open in a new tab

County‐level source‐specific NOx, PM2.5, SO2, and VOC emissions normalized by the maximum county's emissions by source in 2020. Counties with normalized emissions greater than 0.5 are colored the same. SO2 emissions for sources the Agr. and Biogenics category were zero in 2020.

3.1. Exposure Inequality

Using the Atkinson Index, we quantified the equitable distribution of pollutant and source‐specific county emissions (year 2020 shown in Figure 4; other years and results for ε = 2 in Figure S4 through Figure S10 in Supporting Information S1). We present AIx normalized by the largest AIx for each pollutant and each group (e.g., inequality across all groups, Hispanic vs. other, and Black vs. other), which enables the identification of sources contributing to inequitable exposure of each pollutant. Emissions magnitudes vary across source categories for each pollutant, and changes from 2011 to 2020 did not occur consistently across all sources and locations, leading to substantial differences in the effects of the emissions reductions on calculated AIx across all groups (Figure 5). Using these figures, we explore the extent that each source's emissions are inequitably distributed and how they have changed over time.

Figure 4.

Figure 4

Open in a new tab

Atkinson Index normalized by the largest AIx for each group‐pollutant pair assessed at the county level for each pollutant and source in 2020 with ε = 0.75. AIx was calculated for three comparisons: across all groups, comparing the Black population with all other individuals, and comparing the Hispanic population with all other individuals.

Figure 5.

Figure 5

Open in a new tab

AIx (horizontal axis) and emissions magnitudes (point size) in 2011 and 2020. Relative emissions are emissions of each pollutant normalized by the largest pollutant‐source pair in 2011 (i.e., a value of 0.50 means 50% of the mass emitted by the most emitting source in 2020).

Relative to other source categories, agriculture and biogenics have large VOC emissions, and dust and fires have large PM2.5 emissions relative to other source categories (Figure 2; we discuss agriculture, biogenics, dust, and fires together since they are primarily natural sources). Inequality is elevated for agriculture and biogenics NOx and VOCs as well as dust and fires SO2 for the Black population, driven by the presence of these sources in areas with large Black populations. Similarly, inequality across all groups is elevated for dust and fires NOx and VOC (Figure 4). Most VOCs from agriculture and biogenics come from vegetation and soil, biogenic sources that are difficult source to control. Prescribed fires, an important tool in wildfire management, contribute a large fraction of the dust and fires category, and these are not equally distributed because these practices are only undertaken in some areas in Virginia. Care should be taken in interpreting these source categories (agriculture and biogenics and dust and fires) because these emissions estimates are uncertain and methods to quantify them have changed over time.

Electricity generation emissions, emanating from discrete point locations, are inequitably distributed for all pollutants—especially NOx, SO2, and VOCs—and groups (Figure 4). Between 2011 and 2020, power plant closures, air quality controls, and increased use of natural gas led to NOx, SO2, and VOC emissions decreases. The changes drove large decreases in emissions inequity (Figure 5) in each pollutant except VOCs. We found an increase in AIx for VOCs, which is likely tied to proliferation of natural gas power plants throughout Virginia, which have led to more VOC emissions source locations. Out of all the sources, we have most confidence in changes in electricity generation emissions over time because emissions from most power plants are measured and reported directly to EPA.

The industrial category is characterized by relatively large inequality in PM2.5 and SO2 emissions. Industrial PM2.5 was more unequally distributed among Hispanic communities, and industrial SO2 was more unequal for the Black population. We observe small changes in PM2.5, SO2 and VOC emissions and small corresponding increases in inequity for SO2 and VOCs. Industrial NOx decreased 40% from 2011 to 2020, with AIx decreasing to near 0 by 2020.

Of the three mobile source categories, on‐road non‐diesel and non‐road (e.g., marine emissions and farming/construction equipment) tend to have higher inequity than on‐road diesel across pollutants, especially PM2.5 (Figure 4). This suggests that diesel traffic is not as large a driver of inequality as other mobile sources in Virginia. Across all mobile source categories and pollutants, emissions magnitudes decreased since 2011 and inequality increased slightly, suggesting relatively larger emissions decreases near populations with lower emissions already.

Residential sources are associated with inequality in NOx emissions across all groups and in the Hispanic population (Figure 4). Most NOx in this category is from residential natural gas combustion. Inequity in residential NOx increased slightly from 2011 to 2020 while emissions of all other pollutants decreased in magnitude and AIx (Figure 5), although changes are small and should be interpreted with caution given changing emissions estimation methods.

Inequality in years 2011, 2014, and 2017 show some differences from 2020 (Figure S4 through Figure S6 in Supporting Information S1), but the most unequally distributed sources (electricity generation, non‐road mobile, industry, and agriculture and dust, depending on pollutant and population group) are broadly consistent across years. Normalized AIx values calculated with ε = 2 are similar to those calculated with ε = 0.75 (Figure S7 through Figure S10 in Supporting Information S1).

3.2. Sensitivity Analyses

Using the 1 km2 resolution NEMO emissions inventory, we find that all racial‐ethnic subgroups except for the White population live in areas with greater total NOx, PM2.5, SO2, and VOC emissions than the population average (Figure S11 in Supporting Information S1). Asian, Hispanic, and those identifying as some other race groups live in areas with the greatest relative emissions of most pollutants. Higher PWEjrelative for SO2 relative to other pollutants across most groups indicates inequitable siting of point emissions sources such as power plants and industrial sources, which emit most of the SO2 in Virginia.

The NEMO inventory enables an investigation of emissions inequity across all sources at the census tract level, which is not feasible with the source‐specific county‐level NEI. For most groups, emissions at the tract level are slightly more inequitably distributed than on the county level as measured by PWEjrelative. While the difference between PWEjrelative at county and tract levels varies by demographic group and pollutant, the direction is consistent. This suggests that the county‐level analysis we perform below using NEI identifies inequalities present at finer spatial scales even if it may underestimate exposure differences among groups. Similar conclusions come from PWEjabsolute (Figure S12 in Supporting Information S1), which shows larger differences between county and tract‐level exposures than PWEjrelative.

To interpret emissions inqualities related to concentrations, we computed population‐weighted exposure to PM2.5 and NO2 concentrations. PM2.5 concentrations exhibited less spatial variability than emissions (Figure S13 in Supporting Information S1) and a different spatial pattern than PM2.5 emissions, with greater relative concentrations than emissions in western Virginia. NO2 concentration patterns are similar to NOx emissions patterns. Few exposure inequalities exist for PM2.5 on the county or census tract level (Tables S2, S3, and Figure S14 in Supporting Information S1). In contrast, PWEjrelative for NO2 concentrations are very similar to PWEjrelative for NOx emissions (Tables S4, S5, and Figure S14 in Supporting Information S1). These results are consistent with NO2's relatively shorter atmospheric lifetime (exposure is highest nearby sources) and regional contributions to secondary PM2.5 formed in the atmosphere. These factors lead to NO2 concentration gradients (and inequities) that are consistent with emissions and smooth PM2.5 fields that are not highly correlated with primary emissions and associated inequities.

3.3. Limitations

We assessed the equity of emissions sources based on their co‐location with population groups. Such approaches have precedent; for example, in port vehicle retrofits in California (Densberger & Bachkar, 2022), but there are several limitations to such proximity‐based approaches (Shan et al., 2024). First, the approach ignores directional atmospheric transport and nonlinear chemistry; for example, fresh NOx emissions may inhibit O3 and secondary PM2.5 formation in NOx‐rich environments (most cities) and have the opposite effect in rural areas. Additionally, some pollutants transform in the atmosphere to form new pollutants faster than others, so emissions of one pollutant may relate to exposure differently than other pollutants. The sensitivity analyses using 1 km2 emissions data assigned to census tracts do not necessarily reduce exposure error because these data are still emissions, and not concentration data. Future work could reduce error by applying atmospheric models that directly assess each source category's contributions to air pollution concentrations. Our work examines differences in emissions as a proxy for exposures, but it does not account for other factors that contribute to health disparities for certain racial and ethnic groups.

EPA's approaches for estimating species‐source emissions within NEI have changed over time. We addressed this limitation by modeling mobile source emissions using the same version of MOVES, but we did not address inconsistencies over time in other sources, including industrial and natural sources.

As we discussed in the Methods, potential exposure measurement error may come from using emissions as the exposure of interest. To mitigate this, we focus more on between‐source comparisons within a year compared to between‐year comparisons. The AIx is a useful summary to identify pollution sources that might be driving inequality in pollution exposures, and, subsequently, future policies to reduce inequality can focus on those sources.

4. Discussion

Using a consistent methodology and applying multiple approaches to assess population exposure, we identified inequalities in most source categories, though the pollutant driving the inequality and the groups affected varied by source. To our knowledge, this study represents the first application of Atkinson Index—which enables direct inequality comparisons across many groups—to emissions source‐pollutant pairs. This screening has identified targets for reducing air pollution inequalities in Virginia. Across all pollutants, electricity generation was the source driving most inequality. Even though mobile sources decreased during our study period, inequality has increased since 2011.

We found some variation in inequality by pollutants, with most sources showing elevated inequality for NOx; non‐road mobile and electricity generation showing elevated inequality for PM2.5; industry and electricity generation emissions showing elevated inequality for SO2; and dust and fires and electricity generation showing elevated inequality for VOCs. Agricultural NOx and VOC emissions are unequally distributed near the Black population. These results show the specific source categories that should be targeted to reduced inequalities related to each pollutant.

To our knowledge, this is the only study comparing the overall inequality of specific sources and their changes over time in Virginia. In a nationwide analysis, Nunez et al. (2024) used hierarchical models to find that total emissions and emissions changes from 1970 to 2010 differed across race and income groups (Nunez et al., 2024). Their nationwide findings covering a different period generally reflect agreement with our identification of electricity generating and industrial source categories as contributors to racial‐ethnic inequality with additional evidence of most sources linked with inequality in at least one population. Similarly, Tessum et al. (2021) identified nearly all sources as contributing to exposure inequality in Virginia using population‐weighted exposure in 2014 (Tessum et al., 2021). Our work extends these previous analyses by applying the AIx for a comprehensive inequality metric and quantifying changes over time.

Emissions changes between 2011 and 2020 improved equality for some sources but not others. Electricity generation was associated with the greatest inequality for NOx, PM2.5, and SO2 in 2011. Electricity generation decreased its NOx and SO2 emissions substantially and subsequently improved the equitable distribution of NOx and SO2, although it remains one of the most inequitably distributed source categories. Mobile sources saw increased inequality even as overall emissions decreased substantially owing to national rules such as the Heavy Duty Vehicle Rule and Tier 2 Emissions standards, which required reductions in sulfur present in diesel and gasoline fuels (U.S. EPA, 1999, 2000). Industrial SO2 and VOC emissions changed very little in magnitude, but changing spatial patterns in these sources led to increased inequality in 2020.

Historically, Virginia's SO2 emissions were dominated by fuel combustion by coal power plants (electricity generation sector). Emissions from Virginia's 12 coal facilities declined across the period with scrubber installations and increased reliance on natural gas as the primary fossil fuel. Surprisingly, PM2.5 emissions from electricity generation increased over the period—EPA estimates that increased emissions from natural gas electricity generation have outpaced decreasing emissions from coal electricity generation.

Not all air pollution sources have the same impact on health, and recent evidence suggests that some carry a greater relative risk (e.g., biomass burning, traffic, and coal burning), although the most harmful sources are not consistent across studies (Hopke et al., 2020; Krall et al., 2016; Thurston et al., 2016; Weichenthal et al., 2021). Furthermore, air pollution emissions and downstream atmospheric processes vary across space (e.g., urban vs. rural), meaning that some populations are more exposed to air pollution than others (Manisalidis et al., 2020). Thus, both because of different chemical profiles from sources and because air pollution sources are not distributed equally among people, the health burden of air pollution is greater for some sources compared to others. Further work could incorporate health risks of individual pollutants to determine how our observed inequalities impact health.

Future work could apply a similar approach to estimate of exposure inequality over a larger geographical area (e.g., the entire United States) or a smaller one (e.g., the Richmond metro area). The methodology can be extended to other air pollutants of concern to human health including air toxics using the NEI. We utilized broad source categories in this analysis, but future work can incorporate facility‐level information from NEI, along with appropriate air quality modeling, to better assign inequality to individual facilities. Our work demonstrates that focusing on sources driving inequality in pollutant emissions may better decrease inequality compared to reducing total emissions from all sources.

Conflict of Interest

The authors declare no conflicts of interest relevant to this study.

Supporting information

Supporting Information S1

GH2-9-e2025GH001431-s001.pdf (836.6KB, pdf)

Acknowledgments

This work was funded by The Thomas F. and Kate Miller Jeffress Memorial Trust, Bank of America, Trustee. We acknowledge the effort of the undergraduate and Masters students at George Mason who worked on this project: Sara Alhassani, Gabi Armeda, Amira Benkahla, Emily Csizmadia, Amber Elston, Lia Knowlton, Jonathan Ogebe, Zainab Syed, Trajan Smeeth, and Keerthana Vallamkona. In addition, this project was supported by funding from the National Institute of Health Grant R01AG074359. Research described in this article was conducted under contract to the Health Effects Institute (HEI), an organization jointly funded by the United States Environmental Protection Agency (EPA) (Assistance Awards No. R‐82811201) and certain motor vehicle and engine manufacturers. The contents of this article do not necessarily reflect the views of HEI, or its sponsors, nor do they necessarily reflect the views and policies of the EPA.

Henneman, L. R. F. , Nadjafi, R. , Shan, X. , & Krall, J. R. (2025). Source‐specific air pollution emissions inequalities from 2011 to 2020 in Virginia. GeoHealth, 9, e2025GH001431. 10.1029/2025GH001431

Data Availability Statement

National Emissions Inventory data was downloaded from the EPA's website (US EPA, 2024); each year's emissions were downloaded separately. NEMO emissions data were requested from the original authors (Ma & Tong, 2022). PM2.5 and NO2 concentrations were procured from public repositories associated with the published manuscripts (Atmospheric Composition Analysis Group, 2025a, 2025b; Cooper et al., 2022; Shen et al., 2024). Code to reproduce the primary analyses are available on Github: https://github.com/DEVA‐GRP/VA_emissions_AI. Data to support the code is available at Zenodo (L. Henneman, 2025).

References

References

  1. Atmospheric Composition Analysis Group . (2025a). SatPM2.5 (satellite‐derived PM2.5) [Dataset]. Retrieved from https://sites.wustl.edu/acag/datasets/surface‐pm2‐5/
  2. Atmospheric Composition Analysis Group . (2025b). Surface NO2 [Dataset]. Retrieved from https://sites.wustl.edu/acag/datasets/surface‐no2/
  3. Bowe, B. , Xie, Y. , Yan, Y. , & Al‐Aly, Z. (2019). Burden of cause‐specific mortality associated with PM2.5 air pollution in the United States. JAMA Network Open, 2(11), e1915834. 10.1001/jamanetworkopen.2019.15834 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cooper, M. J. , Martin, R. V. , Hammer, M. S. , Levelt, P. F. , Veefkind, P. , Lamsal, L. N. , et al. (2022). Global fine‐scale changes in ambient NO2 during COVID‐19 lockdowns. Nature, 601(7893), 380–387. 10.1038/s41586-021-04229-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cushing, L. J. , Chau, K. , Franklin, M. , & Johnston, J. E. (2021). Up in smoke: Characterizing the population exposed to flaring from unconventional oil and gas development in the contiguous US. Environmental Research Letters, 16(3), 034032. 10.1088/1748-9326/abd3d4 [DOI] [Google Scholar]
  6. Densberger, N. L. , & Bachkar, K. (2022). Towards accelerating the adoption of zero emissions cargo handling technologies in California ports: Lessons learned from the case of the Ports of Los Angeles and Long Beach. Journal of Cleaner Production, 347, 131255. 10.1016/j.jclepro.2022.131255 [DOI] [Google Scholar]
  7. Dominici, F. , Zanobetti, A. , Schwartz, J. , Braun, D. , Sabath, B. , & Wu, X. (2022). Assessing adverse health effects of long‐term exposure to low levels of ambient air pollution: Implementation of causal inference methods. Research Report (Health Effects Institute), 211, 1–56. [PMC free article] [PubMed] [Google Scholar]
  8. Executive Office of the President . (1994). Executive order 12898: Federal actions to address environmental justice in minority populations and low‐income populations. Retrieved from https://www.govinfo.gov/content/pkg/FR‐1994‐02‐16/html/94‐3685.htm
  9. Executive Office of the President . (2023). Executive order 14096: Revitalizing our nation’s commitment to environmental justice for all, 14096 §. Retrieved from https://www.federalregister.gov/documents/2023/04/26/2023‐08955/revitalizing‐our‐nations‐commitment‐to‐environmental‐justice‐for‐all
  10. Fann, N. , Coffman, E. , Timin, B. , & Kelly, J. T. (2018). The estimated change in the level and distribution of PM2.5‐attributable health impacts in the United States: 2005–2014. Environmental Research, 167(April), 506–514. 10.1016/j.envres.2018.08.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Fann, N. , Roman, H. A. , Fulcher, C. M. , Gentile, M. A. , Hubbell, B. J. , Wesson, K. , & Levy, J. I. (2011). Maximizing health benefits and minimizing inequality: Incorporating local‐scale data in the design and evaluation of air quality policies. Risk Analysis, 31(6), 908–922. 10.1111/j.1539-6924.2011.01629.x [DOI] [PubMed] [Google Scholar]
  12. Foley, K. M. , Pouliot, G. A. , Eyth, A. , Aldridge, M. F. , Allen, C. , Appel, K. W. , et al. (2023). 2002–2017 anthropogenic emissions data for air quality modeling over the United States. Data in Brief, 47, 109022. 10.1016/j.dib.2023.109022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Frey, H. C. (2018). Trends in onroad transportation energy and emissions. Journal of the Air and Waste Management Association, 68(6), 514–563. 10.1080/10962247.2018.1454357 [DOI] [PubMed] [Google Scholar]
  14. Gohlke, J. M. , Harris, M. H. , Roy, A. , Thompson, T. M. , DePaola, M. , Alvarez, R. A. , et al. (2023). State‐of‐the‐science data and methods need to guide place‐based efforts to reduce air pollution inequity. Environmental Health Perspectives, 131(12), 125003. 10.1289/EHP13063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gonzalez, D. J. X. , Nardone, A. , Nguyen, A. V. , Morello‐Frosch, R. , & Casey, J. A. (2023). Historic redlining and the siting of oil and gas wells in the United States. Journal of Exposure Science and Environmental Epidemiology, 33(1), 76–83. 10.1038/s41370-022-00434-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Harper, S. , Ruder, E. , Roman, H. A. , Geggel, A. , Nweke, O. , Payne‐Sturges, D. , & Levy, J. I. (2013). Using inequality measures to incorporate environmental justice into regulatory analyses. International Journal of Environmental Research and Public Health, 10(9), 4039–4059. 10.3390/ijerph10094039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Henneman, L. (2025). Data to support “source‐specific air pollution emissions inequalities from 2011‐2020 in Virginia” by L. Henneman et al [Dataset]. 10.5281/zenodo.15046380 [DOI]
  18. Henneman, L. R. F. , Rasel, M. M. , Choirat, C. , Anenberg, S. C. , & Zigler, C. (2023). Inequitable exposures to U.S. coal power plant–related PM2.5: 22 years and counting. Environmental Health Perspectives, 131(3), 037005. 10.1289/EHP11605 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hopke, P. K. , Croft, D. P. , Zhang, W. , Lin, S. , Masiol, M. , Squizzato, S. , et al. (2020). Changes in the hospitalization and ED visit rates for respiratory diseases associated with source‐specific PM2.5 in New York State from 2005 to 2016. Environmental Research, 181, 108912. 10.1016/j.envres.2019.108912 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Jbaily, A. , Zhou, X. , Liu, J. , Lee, T.‐H. , Kamareddine, L. , Verguet, S. , & Dominici, F. (2022). Air pollution exposure disparities across US population and income groups. Nature, 601(7892), 228–233. 10.1038/s41586-021-04190-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kerr, G. H. , Goldberg, D. L. , & Anenberg, S. C. (2021). COVID‐19 pandemic reveals persistent disparities in nitrogen dioxide pollution. Proceedings of the National Academy of Sciences of the United States of America, 118(30), e2022409118. 10.1073/pnas.2022409118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kerr, G. H. , Goldberg, D. L. , Harris, M. H. , Henderson, B. H. , Hystad, P. , Roy, A. , & Anenberg, S. C. (2023). Ethnoracial disparities in nitrogen dioxide pollution in the United States: Comparing data sets from satellites, models, and monitors. Environmental Science & Technology, 57(48), 19532–19544. 10.1021/acs.est.3c03999 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kerr, G. H. , Meyer, M. , Goldberg, D. L. , Miller, J. , & Anenberg, S. C. (2024). Air pollution impacts from warehousing in the United States uncovered with satellite data. Nature Communications, 15(1), 6006. 10.1038/s41467-024-50000-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Krall, J. R. , Mulholland, J. A. , Russell, A. G. , Balachandran, S. , Winquist, A. , Tolbert, P. E. , et al. (2016). Associations between source‐specific fine particulate matter and emergency department visits for respiratory disease in four U.S. cities. Environmental Health Perspectives, 125(1), 97–103. 10.1289/EHP271 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lane, H. M. , Morello‐Frosch, R. , Marshall, J. D. , & Apte, J. S. (2022). Historical redlining is associated with present‐day air pollution disparities in U.S. cities. Environmental Science and Technology Letters, 9(4), 345–350. 10.1021/acs.estlett.1c01012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Levy, J. I. , Chemerynski, S. M. , & Tuchmann, J. L. (2006). Incorporating concepts of inequality and inequity into health benefits analysis. International Journal for Equity in Health, 5(1), 2. 10.1186/1475-9276-5-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Ma, S. , & Tong, D. Q. (2022). Neighborhood emission mapping operation (NEMO): A 1‐km anthropogenic emission dataset in the United States. Scientific Data, 9(1), 680. 10.1038/s41597-022-01790-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Manisalidis, I. , Stavropoulou, E. , Stavropoulos, A. , & Bezirtzoglou, E. (2020). Environmental and health impacts of air pollution: A review. Frontiers in Public Health, 8, 14. 10.3389/fpubh.2020.00014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. McDuffie, E. E. , Martin, R. V. , Spadaro, J. V. , Burnett, R. , Smith, S. J. , O’Rourke, P. , et al. (2021). Source sector and fuel contributions to ambient PM2.5 and attributable mortality across multiple spatial scales. Nature Communications, 12(1), 3594. 10.1038/s41467-021-23853-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. McDuffie, E. E. , Smith, S. J. , O’Rourke, P. , Tibrewal, K. , Venkataraman, C. , Marais, E. A. , et al. (2020). A global anthropogenic emission inventory of atmospheric pollutants from sector‐ and fuel‐specific sources (1970–2017): An application of the community emissions data system (CEDS). Earth System Science Data, 12(4), 3413–3442. 10.5194/essd-12-3413-2020 [DOI] [Google Scholar]
  31. Mikati, I. , Benson, A. F. , Luben, T. J. , Sacks, J. D. , & Richmond‐Bryant, J. (2018). Disparities in distribution of particulate matter emission sources by race and poverty status. American Journal of Public Health, 108(4), 480–485. 10.2105/AJPH.2017.304297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Morello‐Frosch, R. , & Obasogie, O. K. (2023). The climate gap and the color line—Racial health inequities and climate change. New England Journal of Medicine, 388(10), 943–949. 10.1056/NEJMsb2213250 [DOI] [PubMed] [Google Scholar]
  33. Nunez, Y. , Benavides, J. , Shearston, J. A. , Krieger, E. M. , Daouda, M. , Henneman, L. R. F. , et al. (2024). An environmental justice analysis of air pollution emissions in the United States from 1970 to 2010. Nature Communications, 15(1), 268. 10.1038/s41467-023-43492-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Perera, F. (2018). Pollution from fossil‐fuel combustion is the leading environmental threat to global pediatric health and equity: Solutions exist. International Journal of Environmental Research and Public Health, 15(1), 16. 10.3390/ijerph15010016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rosofsky, A. , Levy, J. I. , Zanobetti, A. , Janulewicz, P. , & Fabian, M. P. (2018). Temporal trends in air pollution exposure inequality in Massachusetts. Environmental Research, 161, 76–86. 10.1016/j.envres.2017.10.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Samet, J. M. , Dominici, F. , Zeger, S. L. , Schwartz, J. , & Dockery, D. W. (2000). The national morbidity, mortality, and air pollution study. [PubMed]
  37. Seinfeld, J. H. , & Pandis, S. N. (2006). Atmospheric chemistry and physics: From air pollution to climate change. Atmospheric Chemistry and Physics, 2nd. Retrieved from http://www.knovel.com/knovel2/Toc.jsp?BookID=2126 [Google Scholar]
  38. Shan, X. , Casey, J. A. , Shearston, J. A. , & Henneman, L. R. F. (2024). Methods for quantifying source‐specific air pollution exposure to serve epidemiology, risk assessment, and environmental justice. GeoHealth, 8(11), e2024GH001188. 10.1029/2024GH001188 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Shen, S. , Li, C. , van Donkelaar, A. , Jacobs, N. , Wang, C. , & Martin, R. V. (2024). Enhancing global estimation of fine particulate matter concentrations by including geophysical a priori information in deep learning. ACS ES&T Air, 1(5), 332–345. 10.1021/acsestair.3c00054 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Tessum, C. W. , Apte, J. S. , Goodkind, A. L. , Muller, N. Z. , Mullins, K. A. , Paolella, D. A. , et al. (2019). Inequity in consumption of goods and services adds to racial–ethnic disparities in air pollution exposure. Proceedings of the National Academy of Sciences of the United States of America, 116(13), 6001–6006. 10.1073/pnas.1818859116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tessum, C. W. , Paolella, D. A. , Chambliss, S. E. , Apte, J. S. , Hill, J. D. , & Marshall, J. D. (2021). PM2.5 polluters disproportionately and systemically affect people of color in the United States. Science Advances, 7(18), eabf4491. 10.1126/sciadv.abf4491 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Thakrar, S. K. , Balasubramanian, S. , Adams, P. J. , Azevedo, I. M. L. , Muller, N. Z. , Pandis, S. N. , et al. (2020). Reducing mortality from air pollution in the United States by targeting specific emission sources. Environmental Science and Technology Letters, 7(9), 639–645. 10.1021/acs.estlett.0c00424 [DOI] [Google Scholar]
  43. Thurston, G. D. , Burnett, R. T. , Turner, M. C. , Shi, Y. , Krewski, D. , Lall, R. , et al. (2016). Ischemic heart disease mortality and long‐term exposure to source‐related components of U.S. fine particle air pollution. Environmental Health Perspectives, 124(6), 785–794. 10.1289/ehp.1509777 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. U.S. EPA . (1999). Technical support document for the tier 2/gasoline sulfur ozone modeling analyses.
  45. U.S. EPA . (2000). Technical support document for the heavy‐duty engine and vehicle standards and highway diesel fuel sulfur control requirements: Air quality modeling analyses. United States EPA. [Google Scholar]
  46. US EPA . (2016). Integrated science assessment for oxides of nitrogen–health criteria (No. EPA/600/R‐15/068).
  47. US EPA . (2017). Integrated science assessment for sulfur oxides–health criteria (No. EPA/600/R‐17/451). Retrieved from https://assessments.epa.gov/isa/document/&deid=338596
  48. U.S. EPA . (2020). Policy guidance on the use of MOVES3 for state implementation plan development, transportation conformity, general conformity, and other purposes (No. EPA‐420‐B‐20‐044). OTAQ. Retrieved from https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1010LXH.pdf [Google Scholar]
  49. U.S. EPA . (2021). 2017 national emissions inventory: January 2021 updated release, technical support document (No. EPA‐454/R‐21‐001). Research Triangle Park, NC. Retrieved from https://www.epa.gov/sites/default/files/2021‐02/documents/nei2017_tsd_full_jan2021.pdf [Google Scholar]
  50. US EPA . (2022). Supplement to the 2019 integrated science assessment for particulate matter (final report, 2022) (reports and assessments no. EPA/600/R‐22/028). Retrieved from https://assessments.epa.gov/risk/document/&deid%3D354490 [PubMed]
  51. US EPA, O. (2024). Get air emissions data [Data and Tools]. Retrieved from https://www.epa.gov/air‐emissions‐inventories/get‐air‐emissions‐data‐0
  52. Wang, Y. , Apte, J. S. , Hill, J. D. , Ivey, C. E. , Patterson, R. F. , Robinson, A. L. , et al. (2022). Location‐specific strategies for eliminating US national racial‐ethnic PM2.5 exposure inequality. Proceedings of the National Academy of Sciences of the United States of America, 119(44), e2205548119. 10.1073/pnas.2205548119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Weichenthal, S. , Lavigne, E. , Traub, A. , Umbrio, D. , You, H. , Pollitt, K. , et al. (2021). Association of sulfur, transition metals, and the oxidative potential of outdoor PM2.5 with acute cardiovascular events: A case‐crossover study of Canadian adults. Environmental Health Perspectives, 129(10), 107005. 10.1289/EHP9449 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wu, X. , Braun, D. , Schwartz, J. , Kioumourtzoglou, M. A. , & Dominici, F. (2020). Evaluating the impact of long‐term exposure to fine particulate matter on mortality among the elderly. Science Advances, 6(29), eaba5692. 10.1126/sciadv.aba5692 [DOI] [PMC free article] [PubMed] [Google Scholar]

References From the Supporting Information

  1. US EPA . (2015). 2011 national emissions inventory, version 2 technical support document. Retrieved from https://www.epa.gov/sites/default/files/2015‐10/documents/nei2011v2_tsd_14aug2015.pdf
  2. US EPA . (2018). 2014 national emissions inventory, version 2 technical support document. Retrieved from https://www.epa.gov/sites/default/files/2018‐07/documents/nei2014v2_tsd_05jul2018.pdf
  3. Wickham, H. (2016). ggplot2: Elegant graphics for data analysis (2nd ed.). Springer. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

  1. Atmospheric Composition Analysis Group . (2025a). SatPM2.5 (satellite‐derived PM2.5) [Dataset]. Retrieved from https://sites.wustl.edu/acag/datasets/surface‐pm2‐5/
  2. Atmospheric Composition Analysis Group . (2025b). Surface NO2 [Dataset]. Retrieved from https://sites.wustl.edu/acag/datasets/surface‐no2/
  3. Henneman, L. (2025). Data to support “source‐specific air pollution emissions inequalities from 2011‐2020 in Virginia” by L. Henneman et al [Dataset]. 10.5281/zenodo.15046380 [DOI]

Supplementary Materials

Supporting Information S1

GH2-9-e2025GH001431-s001.pdf (836.6KB, pdf)

Data Availability Statement

National Emissions Inventory data was downloaded from the EPA's website (US EPA, 2024); each year's emissions were downloaded separately. NEMO emissions data were requested from the original authors (Ma & Tong, 2022). PM2.5 and NO2 concentrations were procured from public repositories associated with the published manuscripts (Atmospheric Composition Analysis Group, 2025a, 2025b; Cooper et al., 2022; Shen et al., 2024). Code to reproduce the primary analyses are available on Github: https://github.com/DEVA‐GRP/VA_emissions_AI. Data to support the code is available at Zenodo (L. Henneman, 2025).

Articles from GeoHealth are provided here courtesy of Wiley

RESOURCES