Environment, Ecology, and Urban Plant Diversity: A Dose-Response Analysis of Residential Green Space on Self-Reported Mental Health Using BRFSS and Land Cover Data
Residential green space is an established determinant of population mental health, yet most studies employ binary or categorical exposure measures, obscuring whether the relationship is linear or exhibits threshold effects. To estimate the dose-response relationship between residential plant diversity and self-reported mentally unhealthy days in a nationally representative US adult sample. We linked 297,195 adults from the 2019 BRFSS to state-level Plant Diversity Index (PDI) values derived from NLCD 2019, comprising tree canopy, shrub/scrub, and herbaceous vegetation. Negative binomial regression with restricted cubic splines (5 knots) estimated non-linear dose-response curves, adjusted for age, sex, race/ethnicity, education, employment, income, physical health, and census region. VIF assessed multicollinearity. Component-wise analysis tested each vegetation class independently. Results: Mean mentally unhealthy days was 3.39 (SD = 7.41). The composite PDI was significantly associated with mentally unhealthy days (IRR = 1.111, 95% CI: 1.031–1.198, p = 0.006). Component analysis revealed opposing effects: shrub/scrub was associated with more unhealthy days (IRR = 2.263, p < 0.001) while herbaceous cover was protective (IRR = 0.578, p = 0.036). The dose-response curve was non-linear with a threshold near 30% PDI and contextual reversal at high rural vegetation density. The PDI association was significant in urban residents (IRR = 1.125, p = 0.004) and confirmed by hurdle modelling (OR = 1.077, p = 0.023). Residential plant diversity is associated with mental health burden. Ecological composition — not merely quantity — determines the mental health relevance of plant cover. Findings support targeted canopy investment in low-income urban areas.