Historical Smelters: Tracing Environmental Inequities

With mid-20th century industrialization and urbanization in the United States, secondary lead smelters grew rapidly to meet the demand for lead and support the recycling economy. Often located in or near urban centers, their emissions disproportionately affected nearby residents. Lead, a potent neurotoxin, is especially harmful to young children, causing developmental delays and neurological damage (Levin et al., 2021). Despite gradual recognition of these hazards, many smelter sites were never formally documented by federal or local environmental agencies, leaving significant gaps in remediation efforts.

Environmental justice has since emerged as a critical framework for examining the unequal distribution of environmental burdens. Historically marginalized and socioeconomically vulnerable communities across the country remain at higher risk for exposure to legacy pollutants such as lead. Recognizing these persistent inequities, this study focuses on the Northeastern U.S., a region with a dense population and a long history of industrial activities. This area played a significant role in lead smelting operations during their peak period, making examining lingering environmental health risks significant. This study aims to identify and assess potential disparities in lead exposure and related environmental burdens affecting socioeconomically vulnerable communities near these legacy sites. By analyzing the spatial relationship between historical lead smelter locations, demographic characteristics, and existing environmental justice indices, the research seeks to illuminate how these past industrial activities continue to shape public health concerns today.

Secondary lead smelters, which extracted lead from recycled sources like used batteries, operated across the United States throughout the 20th century. Their emissions were rarely tracked in real-time, and many of these sites were not included in federal databases until researchers—such as Eckel et al. (2001)—identified over 400 previously undocumented smelter locations. These facilities tended to cluster in industrial zones—often in neighborhoods disproportionately affected by discriminatory housing policies like Redlining. By systematically segregating racial and ethnic minorities, Redlining funneled these populations into areas closer to industrial activities, compounding their exposure to environmental hazards.

Given lead’s toxicity and its disproportionate impacts on marginalized communities, accurately mapping and documenting these historical smelter sites is essential. This study employs GIS and spatial statistics to integrate several key datasets, including EJScreen (the U.S. Environmental Pollution Agency’s environmental justice index), historical Redlining maps, and the latest American Community Survey (5-year estimates). Concentrating on the Northeastern U.S.—a historically significant hub for lead smelting—enables a targeted examination of both legacy contamination and current community vulnerabilities. By creating 1-mile and 3-mile buffers around smelter locations, the research assesses how far-reaching the effects of past pollution may still be, highlighting areas where additional public health policies, environmental remediation, or focused interventions may be most urgently needed.

Past industrial operations, particularly secondary lead smelting, have contributed to lasting contamination in urban environments. Eckel et al. (2001) revealed that over 400 former lead-smelting sites were unknown to federal authorities, underscoring the hidden extent of lead pollution throughout the United States. Their sampling of soil at these sites found elevated lead levels, drawing attention to a regulatory gap that has allowed legacy contamination to persist. Subsequent studies, like those conducted by Wu and Johnston (2019), confirmed that smelter-related lead can linger in residential soils for decades and often clusters in “hot spots,” thus posing significant neurodevelopmental risks to children. These findings highlight the enduring public health implications of historical industrial activities and reinforce the necessity of systematic site identification and assessment.

A growing body of work underscores the disproportionate burden of such environmental hazards on historically marginalized or low-income populations. Greenberg and Schneider (2024) specifically examined the intersection of public housing and environmental hazards, noting that certain communities—often already disadvantaged by socioeconomic factors—face higher exposure risks from facilities like waste storage sites or high-traffic roadways. Aligning with these insights, Chan et al. (2020) found that oil and gas facilities in Los Angeles County tend to be situated in neighborhoods with higher cumulative environmental burdens, particularly those with predominantly Black populations. This body of research collectively affirms the environmental justice principle that industrial pollution is not uniformly distributed but is often concentrated in socioeconomically vulnerable communities.

Alongside quantitative assessments, community engagement has emerged as a key strategy for addressing legacy contamination. Johnston et al. (2019) illustrated how collaborations between researchers and local advocacy groups can help residents living near toxic facilities—such as the Exide lead-acid battery smelter—understand the specific health risks they face. Their “Truth Fairy” project not only enabled the collection of biological samples (i.e., baby teeth) to assess lead exposure but also empowered community members through evidence-based advocacy. By involving residents directly, such initiatives facilitate more transparent policy-making and remediation efforts, strengthening the accountability of regulatory agencies.

The use of spatial data and analytics is central to these investigations. Wu and Johnston (2019) demonstrated how hot spot analysis can effectively pinpoint areas of elevated pollution, providing a framework for interventions. Similarly, Greenberg and Schneider (2024) highlighted the potential of newer data-rich tools—such as EJScreen—to capture the interplay between environmental stressors and demographic factors. Drawing on the CalEnviroScreen model, Chan et al. (2020) used cumulative measures to identify neighborhoods with overlapping social and environmental stressors.

The current EPA EJScreen provides data for individual environmental and demographic indicators but lacks a comprehensive cumulative score to evaluate overall environmental burdens and vulnerabilities, which motivated the development of USEnviroscreen as an integrative measure. The EJScreen tool provides valuable individual environmental and demographic indicators but lacks some of the comprehensive components found in the CalEnviroScreen (CES) framework. Notably, CES "Sensitive Population Indicators" are absent from EJScreen, limiting its ability to fully capture population vulnerabilities. To address this, the USEnviroscreen was developed by adapting overlapping indicators from CES while acknowledging the need for additional data to enhance its accuracy and applicability at a national scale.

Building on this literature, the present research integrates EJScreen data, American Community Survey demographics, and historical Redlining data to produce a comprehensive environmental justice score akin to CalEnviroScreen 4.0. Like Chan et al. (2020), this score considers multiple stressors (e.g., pollution levels, socioeconomic vulnerability) to quantify cumulative burdens. By focusing on the Northeastern U.S.—a region with a dense population and extensive industrial history—this study will investigate whether communities within a 1-mile and 3-mile radius of former smelter sites experience higher environmental burdens. Recent studies highlight that lead contamination tends to be most concentrated within relatively small perimeters surrounding smelting facilities, which choose 1-mile and 3-mile buffers methodologically sound. For the selection of the extent of the buffer zone, EPA (1995) has established criteria for evaluating sensitive environments and resources mostly within a 1-mile radius of the study and analysis, so the peaks in lead concentrations in soil and air of interest have been observed primarily at the historic smelter site, also within a 1-mile radius, which suggests that there is an area of high exposure in the vicinity.

In contrast, In selecting an area with a 3-mile radius as the study area, the EPA concluded in the 2019 Era Report that cleanup activities at hazardous waste sites within this area produce measurable benefits. This scope effectively captures localized impacts, including residential health risks and community development opportunities, making it a key focus for analyzing the impacts of lead contamination and remediation efforts. Using both buffers therefore captures not only the core area of most acute pollution but also the peripheral neighborhoods that may still experience elevated lead exposure relative to baseline levels. The approach echoes Eckel et al. (2001) in seeking to fill information gaps about unrecognized or insufficiently studied smelter locations, while also incorporating the spatial hot spot methods noted by Wu and Johnston (2019). Ultimately, this research contributes both an empirical analysis of continued lead exposure risks and a practical framework for targeting environmental policies and remediation efforts where they are most urgently needed.

This research employs a spatially integrated approach that combines multiple datasets to assess cumulative environmental burdens and socioeconomic vulnerabilities around legacy lead smelter sites in the Northeastern United States. The methodology involves three core phases: (1) developing a nationwide environmental burden score, termed USEnviroscreen, by adapting the CalEnviroScreen 4.0 (CES) model; (2) conducting spatial analysis to identify Census block groups near historical smelter sites and assess the potential impacts on surrounding communities; and (3) creating interactive maps and performing correlation analysis to visualize relationships between environmental risks, socioeconomic indicators, and legacy pollution.

This study incorporates multiple datasets at the block group level to examine the spatial relationships between historical lead smelter locations, demographic characteristics, and environmental burdens. First, environmental justice indicators are drawn from the U.S. Environmental Protection Agency’s EJScreen dataset, which offers nationwide metrics on pollution burdens and socioeconomic factors. Second, demographic information—specifically population by racial and ethnic groups—is sourced from the 2022 American Community Survey (ACS) 5-Year Estimates (Table B03002), providing robust, small-area estimates of community composition. Finally, historical Redlining (Home Owners’ Loan Corporation) data, which assigns residential security grades to neighborhoods, is used to assess past housing discrimination’s influence on present-day environmental disparities. By integrating these datasets, the analysis can evaluate how socioeconomic and racial inequalities intersect with legacy pollution to shape environmental health outcomes.

Table 1: Dataset and Source Information

Adapting the CalEnviroScreen (CES) 4.0 methodology (Zeise et al., 2021) for a nationwide application, USEnviroscreen calculates a cumulative environmental burden score for each Census block group by combining two main components: Pollution Burden and Population Characteristics. In CES, Exposure Indicators and Environmental Effects Indicators make up Pollution Burden, while Sensitive Population Indicators and Socioeconomic Factors comprise Population Characteristics.

Following the CES weighting scheme, the calculation of Pollution Burden and Population Characteristics involves aggregating individual indicators within each category using weighted averages to derive cumulative scores, which are then scaled for comparative analysis.

• Pollution Burden =[1×(Avg. Exposure Indicators)+0.5×(Avg. Environmental Effects Indicators)]÷(1+0.5)

• Population Characteristics = Avg. Socioeconomic Factors

Because Sensitive Population Indicators were unavailable, only Socioeconomic Factors were used. These scores remain on a 0–100 scale before further scaling.

Both Pollution Burden and Population Characteristics are scaled to a 0–10 range by dividing each block group’s value by the national maximum (for that component) and multiplying by 10. The final USEnviroScreen Score is then:

USEnviroScreen Score=(Scaled Pollution Burden)×(Scaled Population Characteristics),

resulting in a 0- 100 range. Higher scores indicate greater overall environmental burden and vulnerability.

One example block group in Maryland was selected to illustrate how an overall USEnviroScreen score is calculated. Its block group number is 245102604043 (BG4043).

Shown below are:

• An area map for the block group and surrounding block groups.

• Tables for the indicators of Pollution Burden and Population Characteristics with percentile scores for each of the indicators.

• A table showing how a USEnviroScreen score was calculated for the example area.

The approach used to calculate the USEnviroScreen Score for BG4043 is shown below in tabular form.

The study region is defined as the Northeastern United States, comprising seven states: Connecticut, Maryland, Massachusetts, New Jersey, New York, Pennsylvania, and Rhode Island (Figure 1). For the analysis, all Census Block Groups located within the counties containing smelter sites and those within the 1-mile and 3-mile buffer zones around each smelter were clipped to establish the research area.

A list of documented and previously undocumented lead smelting locations is gathered from academic literature (e.g., Eckel et al. 2001), and U.S. Environmental Protection Agency Office of Land and Emergency Management(2017). Each site is geocoded to generate point features.

Using ArcGIS Pro or equivalent GIS software, 1-mile and 3-mile buffer zones are established around each smelter point. These distances capture both proximate and slightly more distant community exposures. Census block groups are then spatially joined to these buffers. Block groups whose centroids (or significant population areas) fall within a buffer are categorized as “Near” (1 mile or 3 miles) versus “Not Near” (outside the buffers).

• American Community Survey 5-Year Data (2022, Table B03002) supply detailed racial/ethnic demographics. By joining these block group attributes, it becomes possible to analyze whether communities of color experience higher exposures to lead.
• HOLC Redlining Grades (A–D) are georeferenced and overlaid on current Census block group boundaries to examine the legacy effect of historical housing discrimination on present-day environmental risks.

Mitchell, L.H. Co., located at 216 Klagg Avenue in Trenton, NJ, is documented as a historical lead smelter from as early as 1927. Utilizing ArcGIS Pro, two circular buffers were generated at 1 mile (purple) and 3 miles (red) around the site to approximate varying degrees of potential community exposure. These boundaries encompass several Census block groups in northern Trenton, each potentially subject to residual lead contamination.

An initial review of ACS5Y2022 B03002 data indicates that several block groups within the 1-mile radius and 3 miles have higher proportions of Black residents compared to citywide averages. Block groups with darker shading often overlap with pie charts indicating a higher proportion of minority populations, suggesting potential environmental justice concerns. This visual overlay helps identify neighborhoods that may merit deeper investigation or targeted interventions—especially if high USEnviroscreen scores coincide with elevated lead risks. Although we do not perform a correlation analysis, the mapping hints at a possible overlap between minority populations and high pollution burdens.

Overlaying HOLC Redlining maps reveals that portions of the 1-mile and 3-mile zone coincide with historically “hazardous” or “definitely declining” areas. This finding aligns with broader scholarship indicating that formerly redlined neighborhoods often remain near industrial zones, contributing to ongoing environmental health disparities.

Interactive maps hosted on ArcGIS StoryMaps allow the public to visualize the data layers simultaneously, featuring pop-ups that display each block group’s demographic and EJScreen metrics. However, the limited availability of sensitive health indicators in EJScreen restricts the depth of the analysis. Future research could incorporate local health outcome data or soil sampling results to bolster our understanding of legacy lead exposure.

While the ultimate objective of this research is to examine all historical lead smelter sites across the Eastern United States, time constraints necessitated a focused case study on the Mitchell, L.H. Co. location in Trenton, NJ. By overlaying 1-mile and 3-mile buffers around this site with Census demographic data, USEnviroscreen scores, and historical Redlining maps, preliminary patterns of environmental risk and socioeconomic vulnerability were revealed. In particular, block groups with higher proportions of Black and Latinx residents often coincided with elevated USEnviroscreen scores, highlighting potential environmental justice concerns rooted in both industrial legacies and housing discrimination.

In recent years, the EPA Office of Inspector General (OIG) evaluated the agency’s response to contamination stemming from historical lead smelter sites, referred to as the “Eckel sites.” In its 2014 report, the OIG emphasized that while progress has been made—such as completing site assessments and introducing a Lead Smelter Strategy—major gaps remain in linking urban contamination directly to legacy smelting, conducting community outreach, and establishing transparent tracking systems (U.S. Environmental Protection Agency, 2014). These findings underscore the need for more robust procedures and clearer guidance in managing legacy pollution threats.

Despite this case study's indicative findings, the analysis remains limited by data constraints, particularly the lack of sensitive population indicators in EJScreen and minimal soil or air sampling data. As a result, the observed spatial patterns—while suggestive of environmental and socioeconomic disparities—can not definitively confirm lead exposure risks without further investigation. Future studies can enhance the USEnviroscreen scoring methodology by incorporating more granular health metrics, employing formal correlation or regression analyses, and performing community-based environmental sampling to validate the potential contamination and its impacts.

Ultimately, the insights gained from this Trenton case study serve as a proof of concept for a more expansive evaluation of lead smelter sites throughout the Eastern U.S. By refining and broadening these methods and data sources—and by partnering with local stakeholders—researchers and policymakers can better address the lingering effects of historical lead smelting. Such efforts are crucial to ensuring equitable public health outcomes for all communities still bearing the burden of legacy industrial pollution.