⚖️ Environmental Racism Exposure Mapping Simulator

1930s HOLC Redlining Maps — Engineering Racial Geography

Between 1935 and 1940, the federal Home Owners' Loan Corporation (HOLC) commissioned "residential security maps" for 239 American cities, grading every neighborhood from A ("Best") to D ("Hazardous") to guide mortgage lending. Appraisers were instructed to weigh the racial and ethnic composition of a neighborhood as one of the most important — often the single most important — factors in a grade, alongside building age and condition. Black, immigrant, and mixed-race neighborhoods were graded D and shaded red on the map almost automatically, regardless of the physical condition of the housing stock. Because federally-backed mortgages and later private lending followed these maps for decades, redlining is the founding infrastructure of residential segregation and disinvestment in the modern American city.

• 239: HOLC-mapped US cities (residential security maps, 1935–1940)
• >60%: Non-white share, D-grade areas (vs. under 5% in typical A-grade areas)
• <2%: FHA-backed loans reaching D areas (of federally insured mortgages, 1930s–60s)
• 33: Years maps shaped lending (1935 to the 1968 Fair Housing Act)

The HOLC "area description" and the four-color grade

Each HOLC survey area was scored on a standardized "Area Description" form covering terrain, building age, occupancy, and — explicitly — the "infiltration" of racial or foreign-born groups. Appraisers graded neighborhoods A (green, "Best"), B (blue, "Still Desirable"), C (yellow, "Definitely Declining"), or D (red, "Hazardous"), and the forms frequently name specific groups — "Negro," "Mexican," "Jewish," "Italian" — as the stated reason for a lower grade, independent of housing quality.

These were not fringe private documents: HOLC was a federal agency, and its maps were adopted as a template by the Federal Housing Administration (FHA), which insured the mortgages that financed most suburban home construction after 1934. A "D" grade meant that even qualified Black families in stable, well-kept homes were routinely denied federally-backed loans, while white families in far newer suburbs received them almost automatically.

From lending maps to the built environment

Redlining did more than block individual mortgages — it set a self-reinforcing disinvestment cycle in motion. Banks would not lend in D districts, so property owners could not finance repairs, landlords subdivided housing stock to cover falling revenue, city assessors lowered valuations (and tax revenue), and municipalities in turn under-funded schools, parks, sewers, and street maintenance in exactly those neighborhoods. Meanwhile, "A" and "B" districts received a steady stream of capital, public investment, and — as later stages of this simulation show — were routed around when it came time to site highways, factories, and landfills.

By the time the Fair Housing Act of 1968 banned explicit race-based lending, the grading system had already shaped forty years of zoning ordinances, road networks, and industrial siting decisions that no single law could undo overnight.

Modern digitized HOLC maps (via the Mapping Inequality project, University of Richmond) show that grade was a far stronger predictor of a neighborhood's racial composition than of its actual housing quality — the maps graded people, not just property.

A legal end, but not a practical one

Explicit redlining became illegal with the Fair Housing Act, and HOLC itself had stopped issuing new maps decades earlier. But the zoning categories, highway alignments, and industrial districts drawn up while redlining was federal policy were never rezoned or relocated. Land use, once fixed, is extraordinarily durable — a factory sited in 1955 is very often still a factory (or a brownfield) in 2026.

This is why researchers can still measure a strong statistical relationship between an 85-year-old lending map and today's pollution exposure: the map did not just describe a neighborhood's value, it helped decide what would physically be built there for the rest of the century.

Modern Infrastructure Overlay — Highways, Industry, and Waste Facilities

The postwar decades layered a second wave of disinvestment directly on top of the redlining map. The 1956 Federal-Aid Highway Act funded the interstate system largely through cities, and planners consistently routed new expressways through neighborhoods already marked "declining" or "hazardous" — land was cheap, political resistance was weak, and the areas were already administratively defined as blighted. Industrial zoning, municipal landfills, and later hazardous waste treatment, storage, and disposal facilities (TSDFs) followed the same logic for the same reasons, compounding decades later into the infrastructure map you can toggle onto the historic grid here.

• ~1,600 mi: Interstate mileage sited 1956–1970s (through historically Black urban neighborhoods)
• 2.3×: Hazardous facility density, ex-D tracts (higher than ex-A tracts, national average)
• ~34%: Industrially zoned land, ex-D tracts (vs. roughly 8% in ex-A tracts)
• ~1M: Households displaced, urban renewal era (nationwide "slum clearance," 1949–1973)

The Interstate Highway Act and "urban renewal"

Federal highway planners in the 1950s and 60s needed to acquire enormous swaths of urban land cheaply and quickly. Areas that HOLC had already devalued on paper were the path of least resistance — low property values meant low condemnation costs, and residents in disenfranchised neighborhoods had far less political leverage to reroute a proposed expressway than homeowners in wealthier, whiter districts.

The results are still visible on any city map: I-95 sliced through Overtown in Miami, once a thriving Black business district; I-10 was elevated directly over Claiborne Avenue in New Orleans' Tremé, destroying a corridor of Black-owned businesses; I-81 was built through the 15th Ward in Syracuse, displacing thousands. Federal officials at the time openly referred to slum clearance and highway construction as complementary tools — displacing residents was often treated as a feature, not a side effect.

Industrial and hazardous waste facility siting patterns

Once a neighborhood was rezoned or treated as industrial-adjacent, it became the default answer to "where do we put the next facility nobody wants nearby" — a power plant, incinerator, bus depot, scrap yard, or waste transfer station. Siting decisions through the 1960s–1990s rarely accounted for cumulative impact: a single formerly redlined district might absorb a highway interchange, a rail yard, and three or four moderately polluting facilities in succession, while an adjacent "A" district a mile away hosts none.

A widely cited 2022 analysis (Lane, Morello-Frosch, Marshall & Apte, Environmental Science & Technology Letters) matched digitized HOLC maps against present-day EPA facility and pollution data across dozens of US cities and found formerly redlined "D" tracts had systematically higher concentrations of PM2.5, NO2, and proximity to hazardous facilities than "A" tracts in the same metro area, even after controlling for present-day income.

Nationally, historically redlined neighborhoods contain a disproportionate share of Superfund sites, active industrial permits, and EPA-regulated facilities relative to their share of urban land area — a pattern replicated across nearly every major US metro area studied to date.

Disinvestment compounds across generations

Every additional facility sited in a district reduces nearby property values further, which reduces the tax base for local schools and services, which reduces the political capital available to oppose the next facility. This is a textbook positive-feedback loop: each round of infrastructure decisions makes the next round of unwanted siting slightly easier to justify on "already industrial" or "already low-value" grounds.

By the time a formerly redlined neighborhood reaches the present day, its land use pattern is rarely the product of any single decision — it is the accumulated residue of eighty years of decisions that each, individually, looked like a minor administrative convenience.

Building the Composite Environmental Burden Index

Environmental justice researchers rarely rely on a single pollutant to measure exposure inequity — cumulative burden is what actually affects health outcomes. Tools like the EPA's EJScreen and California's CalEnviroScreen combine roughly a dozen environmental and demographic indicators — fine particulate matter, diesel exhaust, ozone, traffic proximity, proximity to hazardous and Superfund sites, wastewater discharge, lead paint risk, urban heat, and tree canopy deficit — into a single composite percentile score per census tract. When these composite scores are mapped against 1930s HOLC grades, the correlation is not subtle.

• +7.0°F: Summer heat gap, D vs A grade (national average; Hoffman, Shandas & Pandya 2020)
• −21%: Tree canopy deficit, ex-D tracts (vs. ex-A tracts, American Forests data)
• +1.5 µg/m³: PM2.5 gap, ex-D vs ex-A tracts (annual mean fine particulate matter)
• 13: Indicators combined in EJScreen (environmental + demographic factors)

Four layers, one composite score

This simulation builds its composite burden index from four representative layers, each of which independently correlates with historic HOLC grade in published research:

• Air quality — ambient PM2.5, diesel particulate matter, and ozone, driven by traffic density and industrial permits concentrated near highways and factories sited in Stage 2

• Urban heat island intensity — formerly redlined districts have less tree canopy and more heat-absorbing pavement, dark rooftops, and industrial surfaces, producing measurably hotter summer temperatures

• Green space deficit — parks and street trees were chronically under-funded in disinvested districts, removing a major mitigating factor against both heat and air pollution

• Toxic release and hazardous site density — the cumulative count of EPA-regulated facilities, contaminated sites, and Superfund proximity within a district

Each layer is weighted and summed into a single 0–100 burden score per district — the same basic methodology EJScreen and CalEnviroScreen use at the census-tract level.

National correlation studies

The foundational study connecting historic grading to present-day heat exposure is Hoffman, Shandas & Pandya (Climate, 2020), which analyzed summer land-surface temperatures across 108 US cities and found formerly redlined "D" neighborhoods were on average 5–7°F hotter than "A" neighborhoods in the same city — in some cities, such as Portland and Richmond, the gap exceeded 12°F. Subsequent work extending this to air pollution (Lane et al. 2022) and to EJScreen composite scores (Nardone et al. 2021, "Associations Between Historical Residential Redlining and Current Age-Adjusted Rates of Emergency Department Visits Due to Asthma") consistently found the same directional pattern: worse historic grade, worse present-day composite burden.

Nardone et al. (2021) found that across 8 California cities, current asthma-related emergency department visit rates were significantly higher in formerly redlined neighborhoods than in neighborhoods graded "Best" — a downstream health signal of the cumulative exposure layers stacked in this stage.

Reading a composite percentile score

A composite burden score is usually expressed as a percentile relative to the state or nation — a tract in the "90th percentile" for cumulative environmental burden is more heavily exposed than 90% of tracts, across every combined indicator at once. This is deliberately different from ranking tracts by any single pollutant, because cumulative impact — many moderate exposures stacked together — can produce serious health effects even where no individual indicator alone would trigger regulatory concern.

This composite framing is also what allows the historic-grade correlation to be measured cleanly: a single pollutant might have one dominant present-day cause, but a composite score reflects the accumulated legacy of dozens of decisions made across nearly a century — which is exactly the kind of signal a 1937 lending map would be expected to predict, if redlining truly shaped the built environment as thoroughly as the historical record suggests.

Demographic Overlay and Statistical Disparity

Overlaying present-day population and demographic data on the composite burden heatmap turns a plausible historical narrative into a quantified statistical relationship. Census tract-level demographic composition, when regressed against composite burden percentile and historic HOLC grade, shows one of the more robust and frequently replicated findings in environmental justice research: race and ethnicity predict exposure burden even after controlling for present-day income, and historic grade predicts a meaningful share of that residual gap on its own.

• r = 0.81: Historic grade / burden correlation (composite index vs. HOLC letter grade)
• 20 yrs: Life expectancy gap, nearby tracts (documented in Richmond, VA and similar cities)
• +63%: PM2.5 exposure disparity, Black residents (above population-weighted average; Mikati et al. 2018)
• 3.4×: Asthma ER visit rate ratio (high-burden vs. low-burden tracts, EJScreen data)

Quantifying the correlation

Mikati, Benson, Luben, Sacks & Richmond-Bryant (American Journal of Public Health, 2018) analyzed EPA National Air Toxics Assessment data and found that non-white US residents are exposed to PM2.5 concentrations 28% above the population-weighted average, and Black residents specifically are exposed to concentrations 54–63% above average — a gap that persists across every income bracket. When these present-day exposure datasets are joined to digitized HOLC boundaries (via the Mapping Inequality project), the historic letter grade alone explains a substantial share of the variance in current-day composite burden — a correlation coefficient in the range of r = 0.7–0.85 has been reported across multiple independent metro-area studies.

The methodology is straightforward: geocode each census tract's composite EJScreen or CalEnviroScreen percentile, spatially join it to the nearest historic HOLC polygon, and run a regression of burden percentile against grade (coded A=1 to D=4), typically controlling for present-day median income to isolate the historic effect from current economic conditions.

Health outcomes as the downstream signal

The clearest evidence that cumulative environmental burden is not an abstraction comes from health outcome data. The Virginia Commonwealth University Center on Society and Health's "8 Minutes Apart, 20 Years Different" mapping project documented life expectancy gaps of up to 20 years between neighborhoods just a few miles apart in Richmond, Virginia — gaps that track closely with the city's original HOLC boundaries. Similar life-expectancy-by-neighborhood gradients tied to historic redlining have since been documented in Chicago, New Orleans, and several other major cities.

Asthma, cardiovascular disease, low birth weight, and heat-related mortality all show the same gradient: each is elevated in historically redlined tracts relative to nearby "A" and "B" tracts in the same metro area, consistent with a cumulative-exposure pathway rather than any single pollutant acting alone.

A life expectancy gap of 20 years between neighborhoods separated by an 8-minute drive is one of the starkest and most frequently cited findings in US health equity research — and it maps almost exactly onto boundaries drawn by federal appraisers in the 1930s.

Correlation, mechanism, and the limits of the model

A correlation coefficient of r ≈ 0.81 between historic grade and present-day burden is strong, but researchers are careful to note the mechanism runs through the intervening variables mapped in Stages 1–3 — zoning, highway siting, industrial permitting, disinvestment in green infrastructure — rather than through any direct causal link between a 1937 map color and a 2026 pollution sensor reading. The map is a marker of a durable institutional pattern, not a mystical force; every stage of this simulation represents one of the concrete, documented mechanisms that connects the two.

Community-Led Remediation and Policy Response

Reversing eight decades of compounding disinvestment is slow, but not hypothetical — a growing body of community-led and policy-driven interventions is measurably narrowing the gap. Participatory air monitoring gives residents evidence to demand regulatory action; targeted greening investment directly reduces the urban heat island effect that Stage 3 quantified; and federal and state policy increasingly directs cumulative-impact-aware funding specifically toward historically overburdened tracts, using the same composite burden methodology built in Stage 3 to decide where money goes first.

• 40%: Justice40 investment target (of federal climate/clean investment benefits to overburdened communities)
• $3.0B: EPA environmental justice grants (IRA) (Environmental and Climate Justice Block Grant Program, 2022)
• −2 to −4°F: Heat reduction per 10% canopy gain (local ambient summer temperature)
• >10,000: Community-deployed air sensors (US) (low-cost PM monitors in citizen science networks)

Participatory monitoring and community science

Residents of overburdened neighborhoods have increasingly built their own evidence base rather than waiting for regulators. The West Oakland Environmental Indicators Project pioneered resident-led truck traffic and particulate matter monitoring in the 1990s and 2000s, producing data that directly informed the Port of Oakland's Comprehensive Truck Management Program and diesel emission reduction rules. Low-cost sensor networks (PurpleAir and similar platforms) have since made hyperlocal, block-by-block air quality monitoring accessible to community groups nationwide, filling gaps left by sparse official regulatory monitoring networks, which historically placed few stations in the highest-burden neighborhoods.

This data has real regulatory teeth: several states, including California and New Jersey, now require permitting agencies to formally consider cumulative community impact — including community-collected data — before approving new industrial facilities in already-overburdened tracts.

Greening investment and green infrastructure

Because tree canopy deficit is one of the four layers driving composite burden in Stage 3, urban greening is one of the more directly measurable remediation levers available. American Forests' Tree Equity Score tool explicitly targets canopy investment toward historically redlined and low-canopy tracts, and peer-reviewed heat studies estimate that a 10-percentage-point increase in tree canopy cover can lower local summer ambient temperatures by roughly 2–4°F — a meaningful fraction of the 5–7°F historic heat gap documented in Stage 3.

Greening is comparatively fast and cheap relative to rezoning or relocating industrial infrastructure, which is why it is often the first visible remediation intervention in a formerly redlined district, even as slower structural changes proceed in parallel.

Because canopy cover takes years to mature, greening investment shows compounding returns: the earlier a district begins a sustained planting program, the larger the cumulative heat and air-quality benefit by the time the trees reach full canopy size.

Targeted policy: Justice40 and cumulative impact law

The Biden administration's Justice40 Initiative (Executive Order 14008, 2021) directs federal agencies to target 40% of the overall benefits of relevant federal climate, clean energy, and infrastructure investment toward disadvantaged communities, using a screening tool (the Climate and Economic Justice Screening Tool) methodologically similar to EJScreen. The Inflation Reduction Act's $3 billion Environmental and Climate Justice Block Grant Program funds community-led monitoring, remediation, and resilience projects directly in overburdened tracts identified through this same composite-burden approach.

At the state level, "cumulative impact" laws in New Jersey and California require regulators to deny or condition new permits in communities that already carry a disproportionate pollution burden — for the first time treating historic and compounding exposure, rather than any single facility's emissions in isolation, as the relevant legal standard. None of these interventions erase eighty years of disinvestment overnight, but each is a direct, traceable policy response to the mechanism this simulation has walked through, stage by stage.