Street-tree ecological inequality under climate stress: spatial diagnostics and candidate planning contexts in New York City

Le résumé fourni par la source

Urban street-tree planning often relies on canopy cover or tree counts, obscuring whether neighborhoods differ in inventory abundance, recorded condition, and species composition and how these differences coincide with climate and social burdens. We analyzed 2,327 Census tracts in New York City by integrating the 2015–2016 street-tree census with land cover, built form, social vulnerability, and heat exposure. We quantified area-standardized alive-tree density, mean ordinal condition, and Shannon diversity, and evaluated their spatial structure and ecological-context associations using complementary spatial and machine-learning diagnostics, including validation across boroughs. The three outcomes were weakly correlated (ρ = 0.05–0.18) yet significantly clustered (Moran’s I = 0.306–0.355), demonstrating that street-tree disadvantage is multidimensional and spatially concentrated. Density showed the strongest and most transferable ecological-context signal (borough-holdout R 2 = 0.616), whereas diversity transferred weakly (R 2 = 0.088) and condition showed no borough-holdout predictive skill (R 2 = -0.010), suggesting that relevant microsite and management information is missing at tract scale. An equal-weight co-burden index combining low street-tree quality, social vulnerability, heat exposure, built-environment intensity, and canopy fragmentation identified 396 top-20% candidate tracts among 1,976 eligible tracts, including 198 in the top 10%, with the greatest concentrations in the Bronx, Upper Manhattan, Central Brooklyn, and parts of Queens. By distinguishing observed inventory patterns from outcomes requiring field verification, the framework provides a transparent first-stage screen for climate-equity planning rather than evidence of ecological function or intervention effectiveness.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.