Divergent social patterning of directly measured environmental exposures across Rhode Island communities
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Environmental noise and air pollution are both shaped by road traffic and the built environment, and exposure assessment increasingly folds them into composite indices or proxies both by traffic exposure. Whether the two share a social distribution has rarely been tested against direct measurement of several exposures in the same communities, and community noise is almost always characterized by A-weighted levels alone, which discount low-frequency energy.
Methods
At 176 sites across Rhode Island, spanning the contiguous urban area of Providence, Central Falls, and Pawtucket together with four rural municipalities, we measured the acoustic environment under A- and C-weighting (LAeq, LCeq), fine particulate matter (PM2.5), night- time illuminance, and relative humidity across four session types over roughly one year (704 site-sessions). Exposures were linked to census-tract composition (American Community Survey), and mixed-effects models were fitted for each of eight area-level markers of disadvantage, adjusting for campaign and session. Relative humidity was carried through the identical model as a negative control.
Results
A-weighted noise was consistently higher in more disadvantaged tracts, rising with non-White, poverty, renter, and no-vehicle shares and falling with income and older-resident share (six of eight markers significant; 1.3 to 1.8 dBA per standard deviation; 6.6 dBA between the least and most racially diverse neighborhoods). C-weighted levels followed the same gradient on every marker and exceeded their A-weighted counterparts at block-group scale for renter occupancy and vehicle absence. Night-time illuminance was also socially patterned, whereas short-term PM2.5 was roughly an order of magnitude weaker and relative humidity showed no gradient. The acoustic gradient persisted within the urban core alone.
Conclusions
Measured burden was carried by the acoustic environment, including its low- frequency component, and by night-time light, not by short-term particulates. The exposure metric and the averaging time determine which disparities are visible at all.
Impact statement
Composite environmental indices and traffic surrogates assume that a neighborhood burdened on one exposure is burdened on all. Measuring noise, particulate matter, and night-time light directly at the same 176 sites shows otherwise: the social gradient was carried by the acoustic environment and by night-time light, not by short-term particulates. Reporting C-weighted alongside A-weighted levels, which to our knowledge has not previously been examined as a distributional measure, shows that the low-frequency burden falls on the same disadvantaged communities and is understated by the conventional metric. Exposure assessment choices therefore determine which inequalities become visible.
