Abstract
Examination of the greenspace-human health relationship operates in at least four dimensions: what is considered greenspace? which moderators and mediators are included? what outcomes are measured? and which units of analysis (e.g., individuals, cities) are studied? We examined three of these four dimensions in a cross-sectional study of 496 of the 500 most populated US cities (total population size = 97,574,613, average population per city = 197,920). Spatial average models tested the effect of two greenspace measures (Normalized Difference Vegetation Index greenness and tree cover) on two outcomes (obesity and mental health), while adjusting for income, race and ethnicity, sprawl, age, sex, physical inactivity, median age of housing, and total population. We conducted analyses at the city scale, which is an understudied unit of analysis, and compared findings to individual- and neighborhood-level studies. In two of four models, greenspace was associated with better health. We found race and ethnicity moderated this relationship with varying results. In full sample analyses, cities with greater percentages of non-Hispanic Whites showed links between higher tree cover and lower obesity but marginal relationships between higher greenness and lower obesity. In subsample analyses with majority-non-Hispanic Black cities, higher tree cover was associated with lower obesity and better mental health. These findings advance previous research by showing that race and ethnicity moderate the greenspace-health link at the city level.