Recreational urban parks support diverse assemblages of plants that contribute ecosystem services to billions of individuals in cities throughout the world. Drivers of ecosystem services in parks are complex, as climate and human preferences interact with multiple species of vegetation types. Yet, informal observations suggest that recreational parks are built consistently to a specific typology. Here we ask: what are the patterns of ecosystem services and vegetation biodiversity in cities of varying climate in the United States, and how do these patterns result in a consistent typology of recreational park? We hypothesized that increased climatic aridity would exclude species not adapted to warm, dry climates, thereby reducing taxonomic alpha diversity and shifting community composition, while a similar municipal preference within the United States for specific suites of service-based traits will lead to the convergence of ecosystem services in recreational parks among cities, regardless of differences in climate. We tested this hypothesis by surveying lawn species, comprised of herbaceous turf and spontaneous plants, and woody species in fifteen recreational parks across Baltimore MD, Riverside CA, and Palm Springs CA, three cities that contain multiple recreational parks but differ in regional aridity. With increasing aridity, taxonomic alpha diversity decreased and plant physiology shifted, yet no differences were observed among most ecosystem service-based traits. Among the cities surveyed, no significant differences were observed in functional dispersion of woody and spontaneous species or most ecosystem service traits.  Taxonomic composition differed in each city for all vegetation types, while ecosystem service composition differed between Baltimore and the two more arid cities of Riverside and Palm Springs. Our results suggest that across the U.S., ecosystem service traits are consistent, even when arising from unique compositions and abundances of species in recreational parks.  We interpret these results as an interaction between aridity and human preferences for services, creating a pattern of biodiversity where taxonomic alpha and beta diversity vary among regions while specific suites of ecosystem services remain available.

All woody species were counted in sampled recreational parks.All herabcous species (turf and spontaneous) were sampled recording braun-blanquet cover classes through 16 random placements of a 500 cm² quadrat in each recreational park. Data is presented in two site by species matricies, one for woody species and one for herbaceous.