Modeling Air Pollution‐Related Health Benefits of Transportation Scenarios: A Collaboration Between Academic Researchers and Environmental Justice Organizations

Abstract

Transportation is a leading contributor to greenhouse gas emissions and has become a focus for climate policies. Traffic‐related air pollution disproportionately affects environmental justice (EJ) communities—neighborhoods that have disproportionate exposure to environmental hazards, but health impact assessments rarely center EJ issues or prioritize the concerns of EJ communities. One explanation for the lack of focus on EJ communities is that both policymakers and academia have often failed to engage these communities. In this paper, academic researchers collaborate with seven EJ organizations in the northeastern US, working with collaboration advisors and facilitators, to design and evaluate potential transportation emissions reduction scenarios using air quality and health benefits modeling tools. We model and estimate the benefits of these scenarios, while working to build collaborative relationships between academic researchers and EJ organizations. The two primary outputs from this process are: quantification of health benefits attributable to emission reduction scenarios of interest to EJ organizations, and enhanced trust and community building between academic researchers and EJ organizations, with reflections on strengths, challenges, and opportunities for future work. We find the largest improvements to health result from scenarios that reduce car and truck traffic. Dialog between academic researchers and EJ organizations reinforce the disconnect between regional‐scale models and local community concerns as well as the more general gaps between statistical models and lived experience. Despite these challenges, the collaboration led to more meaningful models and valued insight for community organizations, and we recommend comparable collaborations in other settings where pollution control is being planned and evaluated in EJ communities.

Introduction

Transportation is a leading contributor to greenhouse gas (GHG) emissions and has recently become a focus for climate policies (Arter et al., 2021; U.S. Department of Transportation, 2023). Climate policies are often designed with GHG reductions as a primary goal, but the air quality benefits resulting from these policies have not always been a focus. This gap is especially problematic for environmental justice (EJ) communities, where traffic-related air pollution is one of many environmental hazards that these communities disproportionately face (Rowangould, 2013; Wier et al., 2009). Similarly, when assessing health benefits to choose the “best” policy proposal, intervention, or strategy, health impact assessments rarely prioritize the concerns of EJ communities. As a result, even successful climate policies and other mitigation strategies do not necessarily benefit EJ communities and have in the past even resulted in increased impacts around EJ communities (Thomson et al., 2008).

One explanation for the lack of focus on EJ communities is that both policymakers and academia have often failed to engage these communities (U.S. Department of Transportation, 2023). Policymakers could explicitly consider the distribution of environmental or health benefits as part of health benefits modeling, and it is also possible that meaningful engagement with EJ communities could lead to more well-informed policies and more equitable outcomes. Recent evidence has indicated that climate policies may better improve health in communities around air pollution hotspots if policies are designed to harmonize GHG emissions reductions with air quality and health outcomes near those hotspots (Cushing et al., 2018).

Another explanation for the lack of focus on EJ communities is that there is no universally agreed-upon definition for EJ communities (Rowangould et al., 2016). Criteria for EJ community definition often include geographic area demographic characteristics or community-based identification (Rowangould et al., 2016). As part of the recent Justice40 initiative, the federal government classifies census tracts as “burdened” if it meets various environmental or health thresholds, such as being above the 90th percentile in asthma or the 90th percentile for diesel particulate matter exposure and below the 65th percentile in income (National Academy of Sciences, Engineering, and Medicine, 2023).

Historically, when researchers have engaged with EJ communities, the research has mainly focused on monitoring current hazards (e.g., Kinney et al., 2000). While there is a substantial body of work documenting collaborations between these groups, the vast majority are for retrospective, or emergency response-type, efforts to remedy current harm (Williamson, 2022). Research focusing on prospective modeling of potential alternative policies, in collaboration with affected communities, is less common (Gardner-Frolick et al., 2022). As a result, community-engaged research work often fails to help allocate where forward-looking investments should be directed. We argue that prospectively designing and modeling climate and transportation policies in collaboration with EJ groups, as well as prioritizing alleviation of pollution in EJ communities, is a strategy that can achieve greater equity.

In this paper, academic researchers collaborate with seven EJ organizations to design and evaluate potential transportation emissions reduction scenarios using air quality and health benefits modeling tools. Together, we leverage our individual strengths and model the air pollution-related health benefits from the proposed scenarios. We use the Community Multiscale Air Quality (CMAQ) Modeling System and Benefits Mapping and Analysis Program in R (BenMAPR) to quantitatively model and estimate the benefits of potential scenarios, working with collaboration advisors and facilitators to create intentional collaboration and build relationships between academic researchers and EJ organizations. In addition to modeling findings, we document post-project reflections from representatives from four EJ organizations.

Methods

This project builds on earlier work from the Transportation, Equity, Climate, and Health Project that was initially designed to evaluate potential effects of the Transportation and Climate Initiative, a regional-scale policy initiative targeting GHG emissions from the transportation sector. Details are available in Arter et al. (2021), but briefly, we use modeling performed with the CMAQ model with the Decoupled Direct Method to estimate air quality impacts in the northeastern United States (Arter et al., 2021). This is a state-of-the-science model that characterizes how emissions from different sources in different locations contribute to air pollution concentrations across a region. Arter et al. modeled the effect of air emissions from five vehicle classes, 17 source areas, and five precursors at 12 km × 12 km resolution. For each source area, the receptor region for exposed populations encompassed the entire model domain. The five vehicle classes were cars, light trucks, medium trucks, heavy trucks, and buses. The 12 states were Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont, Virginia, as well as the District of Columbia, all of whom were part of the Transportation and Climate Initiative. Results were also modeled separately for four metropolitan statistical areas (MSAs) in the northeast (a MSA is a regional area defined by the U.S. Census Bureau): the Boston-Cambridge-Newton, MA and NH MSA, New York-Newark-Jersey City, NY, NJ and PA MSA, Philadelphia-Camden-Wilmington, PA, NJ, DE, and MD MSA, and the combined Baltimore-Columbia-Towson, MD and the Washington-Arlington-Alexandria, DC, VA, MD and WV MSAs. The five precursor pollutants were nitrogen oxides (NO_x), sulfur dioxide (SO₂), volatile organic compounds, primary fine particulate matter (PM_2.5), and ammonia (NH₃), and we focused on their influence on PM_2.5 and ozone (O₃) concentrations.

Note that the 12 km × 12 km (roughly 7.5 mi × 7.5 mi) resolution of our model allows us to capture differences within states and metropolitan regions, but does not capture neighborhood-scale differences that are “averaged out” within a 12 km × 12 km square. We acknowledge the tradeoff between modeling the entire northeastern US on a 12 km × 12 km grid versus using a finer-scale grid on an overall smaller geographic region, such as just one state or metropolitan area, for instance.

EJ Advisory Group Formation

We convened an EJ Advisory Group (EJAG) consisting of participants from EJ and transportation organizations in states from the Transportation and Climate Initiative. Collaboration advisors at Metropolitan Group—a social change agency with extensive expertise in intercultural engagement and strategic communication, especially in climate policy and EJ—invited organizations with which they had existing relationships to participate and also asked these organizations for any additional suggestions for invitations. We accepted all willing EJ organizations into the project.

Study participants include representatives from Alternatives for Community & Environment (Boston, MA), the Center for Latino Progress (Hartford, CT), the Connecticut Coalition for EJ (Hartford, CT), Pittsburghers for Public Transit (Pittsburgh, PA), South Ward Environmental Alliance (Newark, NJ), the Virginia EJ Collaborative (Richmond, VA), and WE ACT for EJ (New York, NY). While these organizations are primarily in urban areas, their focus is not limited to the cities. Recognizing the regional nature of transportation-related air pollution, several organizations are interested in policy concerning suburban and rural transportation. One organization is primarily focused on transportation issues, while the others are interested in transportation as well as broader environmental efforts. Additional details for each EJ organization, including their policy interests and geographic focus, are included in Table 1.

Table 1 Policy Interests for Environmental Justice Organizations

Environmental justice organization	Research interestsa	Geographic focus	Geography used for analysis	Population (millions)
Alternatives for Community & Environment	1; 2	Boston, MA MSA	Boston MSA	4.94
Center for Latino Progress	3; 4	Connecticut	Connecticut	3.61
Connecticut Coalition for Environmental Justice	1; 5	Connecticut	Connecticut	3.61
Pittsburghers for Public Transit	1; 6; 7	Pittsburgh, PA MSA	Pennsylvania	12.96
South Ward Environmental Alliance	7; 8	Newark, NJ	New Jersey	9.27
Virginia Environmental Justice Collaborative	1; 9	Virginia	Virginia	8.64
WE ACT for Environmental Justice	7; 8	New York, NY MSA	New York MSA	19.77

As a collective research team, we met with collaboration advisors and facilitators for an initial meeting in October 2021 to begin relationship building and to discuss and adopt the Jemez Principles (Solis & Southwest Public Workers Union, 1997) as a core guiding ethos for this project. EJAG members shared their policy priorities and concerns regarding transportation-related environmental issues in their geographic focus area.

Initial Scenario Development

Following the initial October 2021 meeting, EJAG members submitted detailed policy interests and problems to the academic team so that the academic team could determine how to best design scenarios to model and evaluate.

In May 2022, the academic team offered a strategy for modeling benefits of these specific scenarios using a modeling tool derived from the previously-constructed CMAQ model. We discussed our proposed approach and whether this fit the vision of individual EJAG members. The academic team also communicated details of the model and its limitations to EJAG members at this meeting. Using the 12 km × 12 km resolution CMAQ model limited the universe of potential scenarios, but had the advantage of being viable within the available time and consistent across all geographic areas of interest. This was most relevant at the geographic level, where scenarios could only be modeled at the state- or metropolitan-level and not at any more precise geographical resolution.

We designed these scenarios to reflect the aspirational goals of EJ organizations and to illustrate both the total harm and the upper bound of benefits for a particular source of pollution in a particular region. The results can be approximately scaled with the magnitude of emissions reduction.

Scenario Operationalization and Transportation Data

Following the May 2022 meeting, we finalized the specific scenarios to be modeled for each EJ organization. We grouped these scenarios into four broad categories: (a) bus electrification; (b) truck electrification; (c) increased transit use; and (d) increased active mobility—walking and cycling. Full details on the relevant transportation data for each geographic region and EJ organization is in Table 2.

Table 2 Transportation Characteristics by Environmental Justice Organization

Geographic focus	Connecticut	New York City MSA	New Jersey	Boston MSA	Pennsylvania	Virginia
Environmental Justice Organization(s)	Center for Latino Progress	WE ACT	South Ward Environmental Alliance	ACE	Pittsburghers For Public Transit	Virginia Environmental Justice Collaborative
CT Coaliation for Environmental Justice
2017: Percentage of Total Car Vehicles Miles Traveled (VMT)
VMT on School Trips	3	2	5	2	2	2
VMT on Trips <1 Mile	0.6	0.8	1	0.9	0.7	0.7
VMT on Trips <2 Miles	4	3	4	4	3	3
VMT on Trips <5 Miles	16	12	14	14	14	13
VMT on Trips <10 Miles	33	26	28	28	31	30
VMT on Trips <1 Mile, Rural Trips	0.1	0.1	0.1	0.1	0.2	0.2
VMT on Trips <2 Miles, Rural Trips	0.5	0.2	0.1	0.1	0.5	0.5
VMT on Trips <5 Miles, Rural Trips	2	1	0.8	1	3	2
VMT on Trips <10 Miles, Rural Trips	5	4	2	2	8	6
2017: Percentage of Total Truck VMT
Truck VMT <100 Miles	24	19	15	18	24	25
2021: Percentage by Type of Bus
School Buses	68	61	60	71	56	No data reported
Commercial, Transit, and Federally Owned Buses	32	39	40	29	44	No data reported

Bus Electrification

To model school and transit bus electrification, we use data on state-level bus inventories from the Federal Highway Administration—an agency of the U.S. Department of Transportation (FHWA, 2022). The number of buses is reported separately by ownership type—private and commercial or publicly owned. For the total number of buses, the number is reported separately for whether the bus is a school bus or any other bus.

For EJAG members interested in school bus electrification, we modeled the elimination of the share of bus emissions equal to the share of school buses (of the total number of buses) in that state. For those interested in transit or general bus electrification, we set baseline targets of 10% and 50% and similarly modeled eliminating that share of bus emissions. Note these are arbitrary targets used to offer initial evidence on the effect of bus electrification. We also note that bus electrification would remove tailpipe emissions, but would not eliminate particulate matter emissions from road or brake dust.

Truck Electrification

To model truck electrification, we use data on freight truck movements from Freight Analysis Framework, a project of the Federal Highway Administration. Truck movements are reported in distance bandwidths, including trips less than 100 miles.

There are currently several commercially available electric trucks with ranges between 200 and 250 miles (Freightliner, 2023; Lambert, 2020). To estimate the effect of truck electrification and in line with currently available technology, we modeled the elimination of all truck emissions on trips below 100 miles (one distance bandwidth used in the Freight Analysis Framework) because of the availability of trucks that could drive this distance daily. Note a trip with freight movement of 100 miles would necessarily require the truck to travel 200 miles back to its homebase. As with bus electrification, truck electrification would also remove only tailpipe emissions.

Increased Transit Use

We did not identify any data sets or studies that directly address the EJAG's key question concerning how expanding transit service in their particular contexts would change car commuting trips or car vehicle emissions in their regions. For example, we do not have an estimate of how car travel would change in Pittsburgh, PA if there was one additional bus route in the city. On a broader scale, one study using satellite data found a 50% reduction in population-related CO₂ emissions from subways in the 192 cities globally that have them (Dasgupta et al., 2023). Recent evidence from fare-free transit implementation finds that most new riders were previously walking or using other zero-emissions modes, so there is a negligible change in emissions (Cats et al., 2017; National Academies of Sciences, Engineering, and Medicine, 2023). It is also possible that fare-free transit increases overall emissions if ridership increases, and additional (fossil-fuel) buses are used.

Because of the lack of available data on transit improvements, we set mileage targets for vehicle trips that would be replaced by transit. We use the National Household Travel Survey from the Federal Highway Administration and set five- and ten-mile measures and modeled the elimination of the share of car emissions equal to the share of total vehicle-miles traveled below these mileage targets.

Increased Active Mobility—Walking and Cycling

We approach shifting trips to active modes—walking and cycling—in the same way we treat increased transit use and use the same data to estimate shares of travel below set benchmark distances. For walking and cycling, we use shorter benchmark distances of one and two miles.

Health Impact Assessment Modeling

The effect of exposure to air pollutants is well documented (Bell et al., 2007; Jerrett et al., 2009; Krewski et al., 2009; Laden et al., 2006). To evaluate the health impacts resulting from air quality changes, we use the Benefits Mapping and Analysis Program in R (BenMAPR), a program similar to U.S. EPA's Benefits Mapping and Analysis Program (BenMAP) (Sacks et al., 2018) that estimates the effect of air quality on health. BenMAPR is an author-created program that uses the underlying BenMAP system in an R console so that all data and analyses can be replicated and run iteratively.

BenMAPR calculates the health implications of changes in air quality associated with emissions reduction scenarios, with the potential to include a number of health outcomes for which there is baseline health data and published evidence of associations between air pollution exposure and health. In this paper, we model mortality and asthma exacerbations—a acute worsening of asthma symptoms and lung function in youth ages 5–17. For a more comprehensive discussion of previous health impact assessment methodology and description of BenMAPR, see Arter et al. (2021) and Buonocore et al. (2023).

Results Sharing and Ongoing Collaboration

Scenario- and region-specific preliminary results for each individual EJAG organization were first shared in October 2022 to hear initial reactions from EJAG members, refine scenarios, and see how these findings connected with their lived experiences in their communities. Using this feedback, the academic team prepared final results to help contextualize organization-specific results and compare the results of similar improvements across organizations. These results were shared at a November 2022 meeting and EJAG members had an additional opportunity to provide feedback at this meeting and in writing.

A final meeting was held in March 2023 with EJAG members, the collaboration advisors and facilitators, the academic team, and project funders to debrief and reflect on the collaborative efforts and to hear how EJAG members plan to use the results of this project.

Results

Baseline Health Outcomes

Characterizing traffic-related baseline mortality and asthma exacerbations by vehicle class and region can help in contextualizing the benefits from emissions reduction scenarios. Understanding that emissions from a particular mode are associated with a higher share of premature deaths in a particular region may shed some initial light on which transportation policies or investments may be most beneficial.

We first examine the number of premature deaths and asthma exacerbations associated with each transportation mode and in each geography at current levels of transportation-related air pollution. Simply, this is the number of people that prematurely die each year as a result of current levels of air pollution from the modeled vehicle classes in these regions or the number of excess exacerbations in people with existing asthma.

In every region but Pennsylvania, cars account for the largest share of premature deaths. In Pennsylvania, 32% of deaths from traffic-related air pollution are caused by light trucks, compared to 28% for cars. The highest share of deaths from cars is in Connecticut (43%) with the New York-Newark-Jersey City MSA being the lowest at 27%. One additional noteworthy finding is that 10% of fatalities from traffic-related air pollution in the New York-Newark-Jersey City MSA are caused by bus emissions—with buses accounting for just 4% in the next highest region. Complete baseline traffic-related air pollution mortality is provided in Table 3.

Table 3 Modeled Baseline Annual Transportation-Related PM_2.5 Mortality for Target Regions by Vehicle Class

	Cars	Light trucks	Medium trucks	Heavy trucks	Buses
Mortalities
Boston MSA	52 (44–61)	49 (41–57)	12 (10–14)	38 (32–45)	2 (1–2)
Connecticut	43 (36–50)	40 (33–46)	4 (3–5)	13 (10–17)	1 (1–1)
New Jersey	220 (180–250)	200 (170–230)	84 (70–97)	210 (180–240)	21 (17–24)
New York MSA	380 (320–450)	350 (300–410)	300 (250–350)	250 (210–300)	140 (120–160)
Pennsylvania	240 (200–280)	280 (230–320)	110 (91–130)	220 (180–250)	16 (13–19)
Virginia	110 (90–120)	100 (90–120)	21 (18–25)	70 (58–82)	14 (11–16)
Mortalities per 1 M people
Boston MSA	11 (9–12)	10 (8–12)	2 (2–3)	8 (6–9)	0 (0–1)
Connecticut	12 (10–14)	11 (9–13)	1 (1–1)	4 (3–5)	0 (0–1)
New Jersey	23 (20–27)	21 (18–25)	9 (8–11)	23 (19–26)	2 (2–3)
New York MSA	20 (17–23)	18 (15–21)	15 (13–18)	13 (11–15)	7 (6–8)
Pennsylvania	18 (15–21)	21 (18–25)	8 (7–10)	17 (14–19)	1 (1–1)
Virginia	12 (10–14)	12 (10–14)	2 (2–3)	8 (7–9)	2 (1–2)

The mode share of asthma exacerbations is similar to premature deaths, with cars accounting for the largest share in each region, except for Pennsylvania where cars again trail light trucks. Baseline asthma exacerbations attributable to traffic-related air pollution are in Table 4.

Table 4 Modeled Baseline Annual Transportation-Related PM_2.5 Asthma Exacerbations for Target Regions by Vehicle Class

	Cars	Light trucks	Medium trucks	Heavy trucks	Buses
Asthma exacerbations
Boston MSA	1,500	1,500	350	1,200	47
Connecticut	1,100	1,000	110	380	16
New Jersey	5,200	4,700	2,000	5,000	490
New York MSA	10,000	8,900	7,600	6,200	3,400
Pennsylvania	5,200	6,100	2,400	4,700	340
Virginia	2,700	2,700	550	1,700	340
Asthma exacerbations per 1 M people
Boston MSA	310	300	71	230	9
Connecticut	310	290	30	100	4
New Jersey	560	510	210	540	53
New York MSA	490	450	390	310	170
Pennsylvania	400	470	180	362	30
Virginia	300	310	64	200	40

Health Improvements From Scenarios Identified by EJAG Members

Detailed scenario-specific health results for each EJ organization are in Table 5. In general, we find that policies that reduce car and light truck use have the largest health benefits, consistent with their contributions to the emissions inventory and baseline transportation-related health impacts, but that reducing medium/heavy truck or bus emissions also has significant benefits. This, of course, does not consider whether reducing car and light truck trips is more or less expensive or feasible than reducing trips from other modes, like buses and medium or heavy trucks. Consistent with Coomes et al. (2022), we find that “improving transit” is among the most beneficial policies, but this should be understood as shifting some trips from cars to transit. Increasing transit ridership, on its own, does little to reduce overall emissions (and may even have deleterious consequences if additional transit service is provided by fossil-fuel-powered buses) if it is not matched with reductions in car trips.

Table 5 Modeled Benefits of Proposed Policies From Annual Transportation-Related PM_2.5

EJ organization and target policy	Mode affected	Baseline mortality	Lives saved	Baseline asthma exacerbations	Asthma exacerbations avoided
Alternatives for Community & Environment
Transit Use: 5 Miles	Cars	52 (44–61)	7 (6–9)	1,600	210
Transit Use: 10 Miles	Cars	15 (12–17)	440
Bus Electrification: 10%	Buses	2 (1–2)	0 (0–1)	47	5
Bus Electrification: 50%	Buses	1 (0.5–1)	24
Center for Latino Progress
School Bus Electrification	Buses	1 (1–1)	0 (0–1)	16	5
Zero-Emissions School Commutes	Cars	43 (36–50)	1 (1–2)	1,100	34
Connecticut Coalition for Environmental Justice
Transit Use: 5 Miles	Cars	43 (36–50)	7 (6–8)	1,100	180
Transit Use: 10 Miles	Cars	14 (12–16)	360
Transit Use: 5 Miles (Rural)	Cars	1 (1–1)	22
Transit Use: 10 Miles (Rural)	Cars	2 (2–3)	56
Pittsburghers for Public Transit
Transit Use: 5 Miles	Cars	230 (200–280)	38 (32–44)	5,200	830
Transit Use: 10 Miles	Cars	71 (59–83)	1,600
Bus Signal Priority	Buses	16 (13–19)	1 (1–1)	340	24
Bus Electrification: 10%	Buses	2 (1–2)	34
Bus Electrification: 50%	Buses	8 (7–10)	170
South Ward Environmental Alliance
Reduce Truck Emissions: MDT	Medium Trucks	84 (70–97)	8 (7–10)	2,000	200
Reduce Truck Emissions: HDT	Heavy Trucks	200 (180–240)	21 (18–24)	5,000	500
Bus Electrification: 15%	Buses	21 (17–24)	3 (3–4)	490	73
Bus Electrification: 50%	Buses	11 (9–12)	240
Virginia Environmental Justice Collaborative
Walking: 1 Mile	Cars	100 (90–120)	1 (1–1)	310	3
Walking: 2 Miles	Cars	3 (3–4)	9
Transit Use: 5 Miles	Cars	14 (11–16)	41
Transit Use: 10 Miles	Cars	30 (26–35)	91
WE ACT for Environmental Justice
Reduce Truck Emissions: MDT	Medium trucks	300 (250–350)	30 (25–35)	7,600	760
Reduce Truck Emissions: HDT	Heavy trucks	250 (210–290)	25 (21–29)	6,200	620
Bus Electrification: 7%	Buses	140 (110–160)	10 (8–11)	3,500	240
Bus Electrification: 50%	Buses	69 (58–80)	1,700

Bus electrification was a policy interest for four of seven EJ organizations. These policies yield appreciable health benefits, with larger percentage emissions reductions occurring in more densely populated areas. As an example of the influence of both size of the fleet and size of the affected population, the New York-Newark-Jersey City MSA has approximately 7 times the number of buses as Connecticut and a much larger exposed population, resulting in a baseline health burden and corresponding health benefits that are two orders of magnitude greater.

EJ Advisory Group Results Discussion, Reflection, and Recommendations

Results Discussion

In meetings between the EJAG and the academic team, EJAG advisors shared reactions to results from their respective organizational perspective.

Three main themes emerged in the discussion between the EJAG members and the academic team. The first is that EJAG members often compared results of the same scenarios across their different organizations. This was especially true for EJAG members that represented overlapping geographies, such as the Center for Latino Progress and the Connecticut Coalition for EJ. For example, noting that in Connecticut a school bus electrification scenario prevented fewer deaths than a transit scenario, these two organizations understandably wanted to better understand why this is the case. Sharon Lewis from the Connecticut Coalition for EJ asked, “why is there greater impact for (a car scenario) than the scenarios for the Center for Latino Progress?” For the academic team, this highlighted the importance of contextualizing the results with baseline data. In Connecticut, for example, emissions from buses were associated with one death per year, versus 43 for cars. Given these differences in baseline mortality attributed to traffic-related air pollution, it may not be surprising that a bus electrification scenario will avert fewer deaths.

The second theme is that there was a mismatch in geographic scale between modeled scenarios and the policies of interest to EJAG members. Using previously modeled results, emissions reduction scenarios could only be modeled at the state- or MSA-levels. These geographies were most helpful for organizations with a focus on larger geographic regions, as opposed to those with a focus on a particular city. For example, for Pittsburghers for Public Transit, modeling transit scenarios for the entire state of Pennsylvania was not as helpful as if we could have provided outputs relevant to the Pittsburgh region alone (Pittsburgh was not one of the four metropolitan regions modeled in the previous work). Mark Mitchell from the Connecticut Environmental Advisory Council said he “would have preferred to have more granular data,” including for differences within his state for different cities, like “Hartford, Bridgeport, New Haven, and Waterbury.”

The third theme, related to the second, is that hyper-local environmental exposure or health outcomes (census block, tract, or zip code) are most desired (e.g., Shukla et al., 2022). Several EJAG members expressed the desire to understand how a scenario involving truck emission reductions, for instance, affects air quality on a truck route or main artery as opposed to a neighborhood street a few blocks away. Kim Gaddy from the South Ward Environmental Alliance remarked that community members can see, smell, and taste the heavy truck emissions from trucks traveling to and from the Port of Newark, and models with 12 km × 12 km concentration data may not adequate reflect that experience. In addition, Jay Stange from the Center for Latino Progress asked, “isn't it a fact that students on buses are exposed to more air pollution than others?,” reinforcing the importance of high-resolution concentration data as well as microenvironmental exposure insights. Because transportation emissions and corresponding changes to primary pollutant concentrations are often highly localized (Chang et al., 2017; Gately et al., 2017), geographically averaged concentration estimates do not help EJ organizations understand or articulate the disparate health impacts within communities.

Reflection

Of the seven members of the EJAG who offered research questions, five organizations participated throughout the duration of the project. Four of these five were able to participate in post-project reflection interviews with the collaboration advisors and facilitators.

In post-project reflections, three of four EJAG members stated they had no concerns or reservations about participating in this project. Pittsburghers for Public Transit stated that they “came into the conversation hesitantly since the research project first evolved from the Transportation and Climate Initiative (TCI).” They had heard many concerns from frontline community organizations that TCI would create air quality harm to their communities. They further stated that some previous policy or academic collaborations with their organization felt heavy-handed and divorced from the realities of community experiences and needs. Additionally, they stated that previous community research projects often felt extractive and less collaborative.

EJAG members reported being overall satisfied with the way their scenario proposals and research questions were addressed. It was also clear, however, that the precise proposals and scenarios modeled did not exactly answer any of the questions from individual EJAG members. The dissonance between initial hopes and ultimate satisfaction reflects an understanding on the part of both the academic team and EJAG members that—on many collaborative projects—compromise is necessary and model capabilities and methodological limitations determine, in part, what questions can be answered. One EJAG member later reflected that more could have been done at the beginning to convey the capabilities of the model. Another wrote that they “eventually” understood the model's capabilities and limitations.

All four respondents reported that the results are useful and that they planned to use them for advocacy. Jay Stange stated he looked forward to sharing results with their state legislature, while Queen Shabazz from Virigina EJ Collaborative stated she was already sharing results at community advocates' meeting on transportation. The results have also been used in funding reports and applications for new funding.

Beyond the study results, the EJAG members reported to the academic team that community members were “pleased that EJ was a focus of the study.” Another commented that it was “gratifying that the researchers were open to the kinds of questions that we posed related to public transit.”

Two EJAG members also remarked on the benefits of participating in the project with other EJAG members. Laura Chu Wiens from Pittsburghers for Public Transit noted they often connect with other transit advocacy organizations. She commented it was great to connect with environment-focused organizations and to see how much their research questions overlapped with theirs. A different EJAG member noted how helpful other EJAG members were in helping them surface their own research questions and understand the results.

Recommendations From EJ Advisors for Future Work

Concerning the process for the research project, two EJAG members recommend a more discussion-oriented format for meetings instead of relying so heavily on presentations from the academic team. One commented that the presentations were long and there could be more time for breakout sessions to develop and discuss research questions, especially with other EJAG members. Combined with other reflections on the benefits from meeting and working with people from other EJAG organizations, maximizing discussion and collaboration appears desirable for future such efforts. Concerning group composition, one EJAG member noted that it would be helpful to have more youth members of the EJAG.

There were also specific recommendations concerning the substance of the project. Connecting to the geographic mismatch and regional-scale coarse resolution environmental modeling, EJAG members recommend evaluating local-scale high resolution air quality and health outcomes, or at least comparing rural regions to inner city communities. Others remarked that qualitative or ethnographic data on individuals affected by transportation emissions would be valuable to complement regional-scale results.

Queen Shabazz suggested we link the quantitative results with qualitative or ethnographic data, such as by “focusing on an individual living in a particular community who has become ill as a result of exposure to atmospheric pollution caused by diesel trucks.”

Multiple EJAG members also recommended the use of hyperlocal monitoring to measure air quality, rather than only relying on air pollution modeling. For example, because the atmospheric dispersion model does not capture the exposure to diesel emissions for children riding in school buses, students could wear lapel-style monitors to measure this exposure. While both comments above were addressing areas that were outside the scope of our study, they do provide valuable inputs to incorporate in future efforts. Others commented that low-cost and widely-distributed monitors could be used to yield more granular baseline air quality data in their communities. While this would be a substantial change in the scope of the project, including the challenge in ultimately connecting these measurements with quantified benefits of emissions reductions, these reactions reinforced the importance of multifaceted strategies to address concerns of all community organizations.

Discussion

In this paper, we use regional-scale air pollution and health benefits modeling tools to assess the health benefits of transportation emissions reduction scenarios designed with seven EJ organizations. We leveraged previously-developed air pollution modeling to examine transportation scenarios in multiple states and metropolitan areas across the northeast US.

Our approach using existing air pollution modeling offered one benefit and several costs. The benefit is simple: because there was no funding for additional air pollution modeling, the project could not have existed without using the existing modeling. The costs are twofold. First is that we were unable to model emissions reduction scenarios as precisely as any EJAG member or academic team member would have liked. We could only examine scenarios for states or metropolitan regions, which may not always be the appropriate policymaking scale since many policies of interest can be enacted at city or neighborhood scale. The second cost is that the existing modeling used a 12 km × 12 km grid for air pollution changes, which is less granular than desired for most EJAG members, who craved insights on hyperlocal impacts and benefits.

In theory, air pollution modeling could be conducted to allow for the examination of transportation policies for source regions of any size and with air quality changes at finer resolution than the 12 km × 12 km grid used in this project, though with tradeoffs between these two elements. ZIP code or census tract modeling analyses are becoming cheaper or technologically efficient so that they can be used more frequently (Shukla et al., 2022; Valencia et al., 2023). As in any project, there are tradeoffs among geographic resolution, geographic breadth, and available resources, and it is important to weigh these as early as possible in the project.

In this project, the tradeoff was clear. Because the modeling originated from a regional-scale policy context, it was more important at the project outset to conduct regional-scale modeling. For the subsequent collaborative effort, we opted to invest energy into a collaborative process rather than to conduct more refined modeling, which allowed us to build relationships and answer questions of interest to the EJAG as best as possible with the available tools. Despite limitations, the project was nevertheless perceived as beneficial by both the academic researchers and the EJ organizations and sheds some light on how state-level transportation interventions impact air quality and health outcomes. Additionally, lessons learned from this project will be beneficial for future work, including the importance of being upfront about methodological limitations.

Separate from EJAG member questions, concerns, and reflection, discussed in the previous section, there are three modeling limitations to consider in our findings. The first is that the modeling, by design, does not account for exposures in transportation microenvironments, including in vehicles and near-roadway exposures. For example, exposure to pollution from emissions near-road or in-vehicle are not captured, including for those traveling on diesel buses, near truck-heavy arteries, or otherwise exposed to idling cars, buses, or trucks. For example, the Center for Latino Progress is most interested in school bus electrification to reduce student's direct exposure to diesel exhaust during their school commute. The results presented here for school bus electrification in Connecticut do not capture these students' exposure.

The second limitation is that these results do not capture health benefits from increases in physical activity. Many of the transportation scenarios involve increasing transit or active transport (walking or biking) trips (Gerike et al., 2016; Heath et al., 2006; Raifman et al., 2021; Woodcock et al., 2013), which leads to additional physical activity, which on its own has positive health benefits including reducing the risk of all-cause mortality (Kelly et al., 2014; Samitz et al., 2011), cardiovascular disease (Nieuwenhuijsen, 2018; Wang et al., 2004), diabetes (Helmrich et al., 1991), and mental health illness (Stephens, 1988).

A third limitation is that this air pollution and health benefits modeling does not capture any other benefits to society from enhanced mobility. Improving transit service, all else equal, will help more people access employment, schools, healthcare, shopping, and other worthwhile destinations. It is important to keep in mind that health benefits resulting from transportation investments are just one piece of the total benefits. These co-benefits, which we did not quantify in this study, are likely appreciable and can further promote equity in a meaningful sense.

There are multiple additional limitations and uncertainties beyond those related to model resolution. For example, much remains unknown about transportation policies and their effects on communities. While electric vehicles are becoming more common in the US, we do not know enough about how potential air quality benefits are distributed especially given the need for electricity for charging and issues of brake and tire wear; how to shift mode choice from cars to transit or active modes, like walking and cycling; and how to make quality public transit work so that residents, especially those that are low income and historically marginalized, can access quality jobs, schools, healthcare and other destinations without relying on a car. We also do not yet know about whether potential harms from electric vehicles—such as additional fatalities because of their higher weight—outweigh environmental benefits. Future research should incorporate these various dimensions but should center dialog between analysts and representatives of EJ organizations from the outset to ensure that models yield relevant insights.

In spite of these limitations, this work begins to fill the literature gap in prospective collaboration between academic researchers and EJ organizations, which has generally been lacking. We argue that this lack of prospective collaboration results in climate policies or mitigation strategies where burdens are disproportionately borne by EJ communities (Thomson et al., 2008). However, meaningful prospective engagement will be necessary in multiple contexts. For example, the Justice40 initiative enacted under President Biden requires government programs to “conduct meaningful engagement with stakeholders to ensure community members have an opportunity to provide input on program decisions, including in the identification of the benefits of Justice40 covered programs” (U.S. Department of Transportation, 2023). Insights from our collaboration reinforce the need for clear communication related to policy scenarios of mutual interest as well as model capabilities and insights, and bidirectional conversations throughout the modeling process to ensure that outputs are interpretable and meaningful.

Acknowledgments

The authors acknowledge support from Grants from Barr Foundation (22-34969) and the New York Community Trust. Study funders had no role in study design or results.

Conflict of Interest

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

Data Availability Statement

National emissions inventories data for 2016 were downloaded from U.S. EPA Emission Modeling Platforms, publicly available at EPA (2023b). Baseline health and economic data were extracted from U.S. EPA BenMAP model which is publicly available at EPA (2023a). Mortality and population counts for individual counties for the entire U.S. for adults ≥25 years and infants <1 year old were obtained from Centers for Disease Control and Prevention Wideranging ONline Data for Epidemiologic Research—CDC WONDER, available through request at CDC (2023a); user's agreement to data use restrictions is required. Asthma exacerbations were evaluated with asthma prevalence data from Center for Disease Control and Prevention; publicly available at CDC (2018) and CDC (2023b). Emission data was processed using the Sparse Matrix Operator Kernel Emissions—SMOKE version 4.6, publicly available at Baek and Seppanen (2020). Air quality simulations were performed using the CMAQ—CMAQ model version 5.2, publicly available at EPA (2017). Source code of the BenMAPR is publicly available at jjbuonocore (2023).

Erratum

The originally published version of this article contained an error in one coauthor’s name. “Dinesch C” should be “Dinesh C.” The error has been corrected, and this may be considered the authoritative version of record.

References

Arter, C. A., Buonocore, J., Chang, C., & Arunachalam, S. (2021). Mortality‐based damages per ton due to the on‐road mobile sector in the Northeastern and Mid‐Atlantic U.S. by region, vehicle class and precursor. Environmental Research Letters, 16(6), 065008. https://doi.org/10.1088/1748‐9326/abf60b

Baek, B. H., & Seppanen, C. (2020). CEMPD/SMOKE: SMOKE v4.6 Public Release (September 2018) (SMOKEv48_Oct2020) [Software]. Zenodo. https://zenodo.org/records/1421403

Bell, M. L., Ebisu, K., & Belanger, K. (2007). Ambient air pollution and low birth weight in Connecticut and Massachusetts. Environmental Health Perspectives, 115(7), 1118–1124. https://doi.org/10.1289/ehp.9759

Buonocore, J. J., Reka, S., Yang, D., Chang, C., Roy, A., Thompson, T., et al. (2023). Air pollution and health impacts of oil & gas production in the United States. Environmental Research: Health, 1(2), 021006. https://doi.org/10.1088/2752‐5309/acc886

Cats, O., Susilo, Y. O., & Reimal, T. (2017). The prospects of fare‐free public transport: Evidence from Tallinn. Transportation, 44(5), 1083–1104. https://doi.org/10.1007/s11116‐016‐9695‐5

CDC. (2018). Asthma ‐ Table 4‐1 Current Asthma Prevalence Percents by Age, 2017 National Health Interview Survey (NHIS) Data [Dataset]. Centers for Disease Control and Prevention (CDC). Retrieved from https://www.cdc.gov/asthma/nhis/2017/table4‐1.htm

CDC. (2023a). Compressed mortality, 1999–2016 request [Dataset]. Centers for Disease Control and Prevention (CDC). Retrieved from https://wonder.cdc.gov/cmf‐icd10.html

CDC. (2023b). National Environmental Public Health Tracking Network Data Explorer [Dataset]. Centers for Disease Control and Prevention (CDC). Retrieved from https://ephtracking.cdc.gov/DataExplorer/

Chang, S. Y., Vizuete, W., Serre, M., Vennam, L. P., Omary, M., Isakov, V., et al. (2017). Finely resolved on‐road PM2.5 and estimated premature mortality in central North Carolina. Risk Analysis: An Official Publication of the Society for Risk Analysis, 37(12), 2420–2434. https://doi.org/10.1111/risa.12775

Coomes, K. E., Buonocore, J. J., Levy, J. I., Arter, C., Arunachalam, S., Buckley, L., et al. (2022). Assessment of the health benefits to children of a transportation climate policy in New York City. Environmental Research, 215, 114165. https://doi.org/10.1016/j.envres.2022.114165

Cushing, L., Blaustein‐Rejto, D., Wander, M., Pastor, M., Sadd, J., Zhu, A., & Morello‐Frosch, R. (2018). Carbon trading, co‐pollutants, and environmental equity: Evidence from California’s cap‐and‐trade program (2011–2015). PLoS Medicine, 15(7), e1002604. https://doi.org/10.1371/journal.pmed.1002604

Dasgupta, S., Lall, S., & Wheeler, D. (2023). Subways and CO2 emissions: A global analysis with satellite data. Science of the Total Environment, 883, 163691. https://doi.org/10.1016/j.scitotenv.2023.163691

EPA. (2017). Community Multiscale Air Quality (CMAQ) model version 5.2. U.S. Environment Protection Agency (EPA) Office of Research and Development [Software]. Zenodo. Retrieved from https://zenodo.org/record/1167892

EPA. (2023a). Benefits Mapping and Analysis Program (BenMAP) Downloads [Dataset]. U.S. Environment Protection Agency (EPA). Retrieved from https://www.epa.gov/benmap/benmap815

EPA. (2023b). Emissions modeling platforms [Dataset]. U.S. Environment Protection Agency (EPA). Retrieved from https://www.epa.gov/air‐emissions‐modeling/emissions‐modeling‐platforms

Freightliner. (2023). Freightliner. “EM2®|Freightliner Trucks.”. Freightliner.com. Retrieved from https://freightliner.com/trucks/em2/

Gardner‐Frolick, R., Boyd, D., & Giang, A. (2022). Selecting data analytic and modeling methods to support air pollution and environmental justice investigations: A critical review and guidance framework. Environmental Science & Technology, 56(5), 2843–2860. https://doi.org/10.1021/acs.est.1c01739

Gately, C. K., Hutyra, L. R., Peterson, S., & Sue Wing, I. (2017). Urban emissions hotspots: Quantifying vehicle congestion and air pollution using mobile phone GPS data. Environmental Pollution, 229, 496–504. https://doi.org/10.1016/j.envpol.2017.05.091

Gerike, R., Nazelle, A., Nieuwenhuijsen, M., Int Panis, L., Anaya Boig, E., Avila‐Palencia, I., et al. (2016). Physical Activity through Sustainable Transport Approaches (PASTA): A study protocol for a multicentre project. BMJ Open, 6(1), e009924. https://doi.org/10.1136/bmjopen‐2015‐009924

Heath, G. W., Brownson, R. C., Kruger, J., Miles, R., Powell, K. E., Ramsey, L. T., & Task Force on Community Preventive Services. (2006). The effectiveness of urban design and land use and transport policies and practices to increase physical activity: A systematic review. Journal of Physical Activity & Health, 3(s1), S55–S76. https://doi.org/10.1123/jpah.3.s1.s55

Helmrich, S. P., Ragland, D. R., Leung, R. W., & Paffenbarger, R. S. (1991). Physical activity and reduced occurrence of non‐insulin‐dependent diabetes mellitus. New England Journal of Medicine, 325(3), 147–152. https://doi.org/10.1056/NEJM199107183250302

Jerrett, M., Burnett, R. T., Pope, III, C. A., Ito, K., Thurston, G., Krewski, D., et al. (2009). Long‐term ozone exposure and mortality. New England Journal of Medicine, 360(11), 1085–1095. https://doi.org/10.1056/nejmoa0803894

jjbuonocore. (2023). jjbuonocore/BenMAPR: BenMAPR for O&G (Version v1) [Software]. Zenodo. https://doi.org/10.5281/zenodo.8306380

Kelly, P., Kahlmeier, S., Götschi, T., Orsini, N., Richards, J., Roberts, N., et al. (2014). Systematic review and meta‐analysis of reduction in all‐cause mortality from walking and cycling and shape of dose response relationship. International Journal of Behavioral Nutrition and Physical Activity, 11(1), 132. https://doi.org/10.1186/s12966‐014‐0132‐x

Kinney, P. L., Aggarwal, M., Northridge, M. E., Janssen, N. A., & Shepard, P. (2000). Airborne concentrations of PM(2.5) and diesel exhaust particles on Harlem sidewalks: A community‐based pilot study. Environmental Health Perspectives, 108(3), 213–218. https://doi.org/10.1289/ehp.00108213

Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., et al. (2009). Extended follow‐up and spatial analysis of the American Cancer Society study linking particulate air pollution and mortality (Vol. 140). Health Effects Institute.

Laden, F., Schwartz, J., Speizer, F. E., & Dockery, D. W. (2006). Reduction in fine particulate air pollution and mortality: Extended follow‐up of the Harvard Six Cities study. American Journal of Respiratory and Critical Care Medicine, 173(6), 667–672. https://doi.org/10.1164/rccm.200503‐443oc

Lambert, F. (2020). Kenworth launches two new electric trucks with up to 200 miles of range. Electrek. Retrieved from https://electrek.co/2020/09/13/kenworth‐electric‐trucks/

National Academies of Sciences, Engineering, and Medicine. (2023). Representing lived experience in the climate and economic justice screening tool. In Proceedings of a Workshop–in Brief. https://doi.org/10.17226/27158

Nieuwenhuijsen, M. J. (2018). Influence of urban and transport planning and the city environment on cardiovascular disease. Nature Reviews Cardiology, 15(7), 432–438. https://doi.org/10.1038/s41569‐018‐0003‐2

Raifman, M., Lambert, K. F., Levy, J. I., & Kinney, P. L. (2021). Mortality implications of increased active mobility for a proposed regional transportation emission cap‐and‐invest program. Journal of Urban Health, 98(3), 315–327. https://doi.org/10.1007/s11524‐020‐00510‐1

Rowangould, D., Karner, A., & London, J. (2016). Identifying environmental justice communities for transportation analysis. Transportation Research Part A: Policy and Practice, 88, 151–162. https://doi.org/10.1016/j.tra.2016.04.002

Rowangould, G. M. (2013). A census of the US near‐roadway population: Public health and environmental justice considerations. Transportation Research Part D: Transport and Environment, 25, 59–67. https://doi.org/10.1016/j.trd.2013.08.003

Sacks, J. D., Lloyd, J. M., Zhu, Y., Anderton, J., Jang, C. J., Hubbell, B., & Fann, N. (2018). The Environmental Benefits Mapping and Analysis Program–Community Edition (BenMAP–CE): A tool to estimate the health and economic benefits of reducing air pollution. Environmental Modelling & Software, 104, 118–129. https://doi.org/10.1016/j.envsoft.2018.02.009

Samitz, G., Egger, M., & Zwahlen, M. (2011). Domains of physical activity and all‐cause mortality: Systematic review and dose–response meta‐analysis of cohort studies. International Journal of Epidemiology, 40(5), 1382–1400. https://doi.org/10.1093/ije/dyr112

Shukla, K., Seppanen, C., Naess, B., Chang, C., Cooley, D., Maier, A., et al. (2022). ZIP code‐level estimation of air quality and health risk due to particulate matter pollution in New York City. Environmental Science & Technology, 56(11), 7119–7130. https://doi.org/10.1021/acs.est.1c07325

Solis, R., & Southwest Public Workers Union. (1997). Jemez principles for democratic organizing. SouthWest Organizing Project.

Stephens, T. (1988). Physical activity and mental health in the United States and Canada: Evidence from four population surveys. Preventive Medicine, 17(1), 35–47. https://doi.org/10.1016/0091‐7435(88)90070‐9

Thomson, H., Jepson, R., Hurley, F., & Douglas, M. (2008). Assessing the unintended health impacts of road transport policies and interventions: Translating research evidence for use in policy and practice. BMC Public Health, 8(1), 339. https://doi.org/10.1186/1471‐2458‐8‐339

U.S. Department of Transportation. (2023). Justice40 initiative. US Department of Transportation. Retrieved from https://www.transportation.gov/equity‐Justice40

U.S. Department of Transportation, Federal Highway Administration. (2022). Highway Statistics 2022—Policy. Federal Highway Administration. Retrieved from https://www.fhwa.dot.gov/policyinformation/statistics/2022/

Valencia, A., Serre, M., & Arunachalam, S. (2023). A hyperlocal hybrid data fusion near‐road PM2.5 and NO2 annual risk and environmental justice assessment across the United States. PLoS One, 18(6), e0286406. https://doi.org/10.1371/journal.pone.0286406

Wang, G., Pratt, M., Macera, C. A., Zheng, Z.‐J., & Heath, G. (2004). Physical activity, cardiovascular disease, and medical expenditures in U.S. adults. Annals of Behavioral Medicine: A Publication of the Society of Behavioral Medicine, 28(2), 88–94. https://doi.org/10.1207/s15324796abm2802_3

Wier, M., Sciammas, C., Seto, E., Bhatia, R., & Rivard, T. (2009). Health, traffic, and environmental justice: Collaborative research and community action in San Francisco, California. American Journal of Public Health, 99(S3), S499–S504. https://doi.org/10.2105/AJPH.2008.148916

Williamson, D. H. (2022). Using the community engagement framework to understand and assess EJ‐related research efforts. Sustainability, 14(5), 2809. https://doi.org/10.3390/su14052809

Woodcock, J., Givoni, M., & Morgan, A. S. (2013). Health impact modelling of active travel visions for England and Wales using an Integrated Transport and Health Impact Modelling Tool (ITHIM). PLoS One, 8(1), e51462. https://doi.org/10.1371/journal.pone.0051462