This study develops and applies a multi-metric framework to quantify joint neighbourhood-scale exposure to summer surface heat and fine particulate matter (PM2.5) across large US metropolitan areas. The objective is to measure co-occurrence of environmental stressors at the census-tract level, evaluate disparities by racial composition and historical redlining, and provide a portable assessment approach for cumulative-impact analyses.
The framework combines multiple complementary metrics to describe joint exposure patterns:
Uncertainty for dependence parameters and ratios was quantified using bias-corrected and accelerated (BCa) confidence intervals derived from a 5,000-replicate metropolitan-area block bootstrap.
The analysis covered 42,304 census tracts in 48 large US metropolitan areas for summers 2015 to 2020, representing approximately 174.6 million residents. The manuscript reports that all data sources are publicly available and documented in the Methods section of the original paper.
Upper-tail dependence was estimated nonparametrically with the Capéraà–Fougères–Genest estimator and parametrically with a Gumbel copula. The study used a metropolitan-area block bootstrap with 5,000 replicates to obtain BCa confidence intervals for dependence estimates and other derived quantities. The conditional exceedance ratio was evaluated at the 80th percentile within metropolitan areas for subgroup comparisons.
Using the within-metropolitan-area 80th percentile threshold for both stressors, the authors report that 13.21% of residents in predominantly Black tracts lived in neighbourhoods that simultaneously exceeded the 80th percentile for heat and PM2.5. In comparison, 3.33% of residents in predominantly White tracts lived in such joint-exceedance neighbourhoods. The corresponding relative magnitudes for single-stressor exceedance were reported as heat-only and PM2.5-only ratios of 2.88 and 2.48 (Black versus White), respectively.
When assessing within-group conditional exceedance ratios at the 80th percentile, the study found:
Only the White group interval excluded one; the authors interpret inclusion of one in Black and Hispanic intervals as possibly due to limited power given fewer contributing Core-Based Statistical Areas (CBSAs) for those subgroup analyses.
The authors evaluated relationships with Home Owners' Loan Corporation (HOLC) grades, a historical redlining classification. After adjustment for contemporary tract racial composition, poverty, renter-occupancy, and pre-1960 housing, residents of HOLC grade D tracts had 3.97 times the odds of living in joint heat–PM2.5 hotspots compared with grade A residents (95% confidence interval 2.79 to 5.66), as reported in the manuscript.
The manuscript reports sensitivity analyses using near-surface air temperature surfaces instead of the primary surface temperature product. Under these alternative surfaces, magnitudes of some estimates attenuated but the direction and statistical significance of the primary findings were preserved.
The authors present the multi-metric framework as portable to other compound-exposure questions and suitable for cumulative-impact assessment at the neighbourhood scale. The report notes differential joint exposure burdens by race and by areas historically graded as redlined, and emphasizes that some subgroup dependence estimates may be power-limited. Specific limitations, full methods, and additional details are documented in the manuscript; where methods or data provenance are not exhaustively summarized in the source preview, the original Methods section is cited as the location of those details.
All data used in the study are publicly available with sources documented in the Methods. The authors declare no competing interests and report adherence to relevant ethical guidelines and that participant consent and necessary approvals were obtained, as stated in the manuscript.