Exploring the Nature of Associations Between SES and Greenness in the Green Heart Louisville Study

Abstract

BACKGROUND AND AIM: Residential area greenness is often associated with socioeconomic status (SES) of nearby residents. However, the nature of this association has not been thoroughly explored at the individual level, or by greenness types at various scales. We conducted an in-depth assessment of associations between SES and greenness in Louisville, Kentucky, in a study area similar to many urban residential neighborhoods in the eastern United States. METHODS: As part of the Green Heart Louisville Study, we collected data on SES: income, education, employment, property ownership status, for 730 participants residing in a neighborhood study area of 3.5 sq miles with high variability of SES and greenness. We collected aerial-based high-resolution indices of greenness - NDVI, biomass, leaf surface area, leaf area index, and canopy cover. We used multiple linear regression and random forest models to examine associations between SES and surrounding greenness estimated at parcels of participant homes, spatial radii around homes, and census units, and we compared the relationship between SES and greenness at the parcel, area, and neighborhood scales. RESULTS:We found that surrounding greenness was significantly associated with income, population density, and distance to major roads and that these factors explain a significant percentage of greenness variability in both parcel and area-level models. Much stronger associations were observed among non-renters. The associations varied with the spatial unit examined, but less so with different measures of greenness. Adjusted multiple linear regression models showed high consistency across local-level greenness, while parcel-level models evince greater variability. CONCLUSIONS:We observed significant associations between SES and greenness in our study area, with results varying based on the spatial context and participant characteristics. While not representative of all urban areas, these results could help inform statistical adjustments of future work when examining associations between greenness and health outcomes and better enable sustainable and culturally competent greenness intervention strategies. KEYWORDS: Greenness, SES, remote sensing, trees, built environment