Linking environmental injustices in Detroit, MI to institutional racial segregation through historical federal redlining

Abstract

Objectives: To identify the most pervasive environmental exposures driving environmental disparities today associated with historical redlining in Detroit.

Methods: We overlaid Detroit's 1939 Home Owners' Loan Corporation (HOLC) shapefile from the Mapping Inequality project onto the EPA EJScreen and the DOT National Transportation Noise maps to analyze differences in current demographic and environmental indicators between historically redlined (D-grade) and non-redlined neighborhoods using simple linear regression and a boosted classification tree algorithm.

Results: Historically redlined neighborhoods in Detroit experienced significantly higher environmental hazards than non-redlined neighborhoods in the form of 12.1% (95% CI: 7.2-17.1%) higher levels of diesel particulate matter (PM), 32.2% (95% CI: 3.3-69.3%) larger traffic volumes, and 65.7% (95% CI: 8.6-152.8%) higher exposure to hazardous road noise (L_EQ(24h) >70 dBA). Historically redlined neighborhoods were situated near 1.7-times (95% CI: 1.4-2.1) more hazardous waste sites and twice as many (95% CI: 1.5-2.7) risk management plan (RMP) sites than non-redlined neighborhoods. The lifetime cancer risk from inhalation of air toxics was 4.4% (95% CI: 2.9-6.6%) higher in historically redlined communities, and the risk of adverse respiratory health outcomes from air toxics was 3.9% (95% CI: 2.1-5.6%) higher. All factors considered together, among the environmental hazards considered, the most pervasive hazards in historically redlined communities are proximity to RMP sites, hazardous road noise, diesel PM, and cancer risk from air pollution.

Conclusions: Historically redlined neighborhoods may have a disproportionately higher risk of developing cancer and adverse respiratory health outcomes from air toxics. Policies targeting air and noise pollution from transportation sources, particularly from sources of diesel exhaust, in historically redlined neighborhoods may ameliorate some of the impacts of structural environmental racism from historical redlining in Detroit.

Keywords: Air pollution; EJSCREEN; Environmental justice; Noise; Redlining; Structural racism.

Fig. 1. Scan of the 1939 Home Owners’ Loan Corporation (HOLC) map of the city of Detroit.

Source: https://dsl.richmond.edu/panorama/redlining/#loc=5/39.1/-94.58&text=downloads.

Fig. 2. Boxplots of environmental indicators from the EPA EJScreen by 1930s HOLC grade.

From left to right, top to bottom: proximity to Risk Management Plan (RMP) sites (count), proximity to hazardous waste sites (count), hazardous noise indicator (% road noise exposure>70 dBA), traffic volume (count), diesel particulate matter (PM, µg/m³), air cancer toxics risk (lifetime risk in 1 million), respiratory hazard index (ratio—unitless), lead paint indicator (% housing stock pre-1960s), PM_2.5 (µg/m³), ozone (ppb), proximity to Superfund sites (count), and proximity to wastewater sites (count). Grade A signifies the “best” neighborhoods; grade B signifies “still desirable” neighborhoods; grade C signifies “definitely declining” neighborhoods; and grade D signifies “hazardous” neighborhoods. Redlined neighborhoods (grade D) are shaded gray; non-redlined neighborhoods (grades A, B, and C) are not shaded. Note that the y-axes are not consistent due to the different units of the environmental indicators. Diamond represents mean.

Fig. 3. Results of the full boosted classification tree analysis.

The left panel displays the relative variable importance of environmental and demographic indicators strongly associated with a redlined neighborhood compared to a non-redlined neighborhood. Predictors that exceed the dashed line (i.e., randomness threshold of 100/14, or 7.14%) are considered important factors that are most indicative of an area having been redlined. Gray bars are environmental indicators; white bars are demographic indicators. The right panel displays the partial dependency plots for the four environmental indicators above the randomness threshold in order of decreasing relative importance (from left to right, top to bottom): proximity to risk management plan sites (count); hazardous noise indicator (% road noise exposure >70 dBA); diesel particulate matter (PM, µg/m³); and air cancer toxics risk (risk in 1 million). The black line signifies the smoothed partial dependency plot.

Fig. 4. Map of major streets (i.e., interstate and state highways) built post-1939 in Detroit.

Green and red shaded areas represent Grade A and D 1939 Home Owners’ Loan Corporation (HOLC) neighborhoods, respectively. Yellow streets represent interstate highways. Gray streets represent state highways. Uncolored streets represent streets constructed pre-1939. Bridges and tunnels (from left to right): Gordie Howe International Bridge (currently under construction); Ambassador Bridge; and Detroit-Windsor Tunnel.