Urbanization has a homogenizing effect on biodiversity and leads to communities with fewer native species and lower conservation value. However, few studies have explored whether or how land management by urban residents can ameliorate the deleterious effects of this homogenization on species composition. We tested the effects of local (land management) and neighborhood‐scale (impervious surface and tree canopy cover) features on breeding bird diversity in six US metropolitan areas that differ in regional species pools and climate. We used a Bayesian multiregion community model to assess differences in species richness, functional guild richness, community turnover, population vulnerability, and public interest in each bird community in six land management types: two natural area park types (separate and adjacent to residential areas), two yard types with conservation features (wildlife‐certified and water conservation) and two lawn‐dominated yard types (high‐ and low‐fertilizer application), and surrounding neighborhood‐scale features. Species richness was higher in yards compared with parks; however, parks supported communities with high conservation scores while yards supported species of high public interest. Bird communities in all land management types were composed of primarily native species. Within yard types, species richness was strongly and positively associated with neighborhood‐scale tree canopy cover and negatively associated with impervious surface. At a continental scale, community turnover between cities was lowest in yards and highest in parks. Within cities, however, turnover was lowest in high‐fertilizer yards and highest in wildlife‐certified yards and parks. Our results demonstrate that, across regions, preserving natural areas, minimizing impervious surfaces and increasing tree canopy are essential strategies to conserve regionally important species. However, yards, especially those managed for wildlife support diverse, heterogeneous bird communities with high public interest and potential to support species of conservation concern. Management approaches that include the preservation of protected parks, encourage wildlife‐friendly yards and acknowledge how public interest in local birds can advance successful conservation in American residential landscapes.

Societal Impact Statement: People plant, remove, and manage urban vegetation in cities for varying purposes and to varying extents. The direct manipulation of plants affects the benefits people receive from plants. In synthesizing several studies of urban biodiversity in Los Angeles, we find that cultivated plants differ from those in remnant natural areas. This highlights the importance of studying cultivated plants in cities, which is crucial for the design and planning of sustainable cities. Residents have created a new urban biome in Los Angeles, and this has consequences for associated organisms, ultimately resulting in a responsibility for society to determine what type of biome we wish to create.Summary: Urbanization is a large driver of biodiversity globally. Within cities, urban trees, gardens, and residential yards contribute extensively to plant biodiversity, although the consequences and mechanisms of plant cultivation for biodiversity are uncertain. We used Los Angeles, California, USA as a case study for investigating plant diversity in cultivated areas. We synthesized datasets quantifying the diversity of urban trees, residential yards, and community gardens in Los Angeles, the availability of plants from nurseries, and residents’ attitudes about plant attributes. Cultivated plant diversity was drastically different from remnant natural areas; compared to remnant natural areas, cultivated areas contained more exotic species, more than double the number of plant species, and turnover in plant functional trait distributions. In cultivated areas, most plants were intentionally planted and dominated by exotic species planted for ornamental purposes. Most tree species sampled in Los Angeles were available for sale in local nurseries. Residents’ preferences for specific plant traits were correlated with the trait composition of the plant community, suggesting cultivated plant communities at least partially reflect resident preferences. Our findings demonstrate the importance of cultivated species in a diverse megacity that are driven in part through commercial distribution. The cultivation of plants in Los Angeles greatly increases regional plant biodiversity through changes in species composition and functional trait distributions. The pervasive presence of cultivated species likely has many consequences for residents and the ecosystem services they receive compared with unmanaged or remnant urban areas.

In urban areas, anthropogenic drivers of ecosystem structure and function are thought to predominate over larger‐scale biophysical drivers. Residential yards are influenced by individual homeowner preferences and actions, and these factors are hypothesized to converge yard structure across broad scales. We examined soil total C and total δ¹³C, organic C and organic δ¹³C, total N, and δ¹⁵N in residential yards and corresponding reference ecosystems in six cities across the United States that span major climates and ecological biomes (Baltimore, Maryland; Boston, Massachusetts; Los Angeles, California; Miami, Florida; Minneapolis‐St. Paul, Minnesota; and Phoenix, Arizona). Across the cities, we found soil C and N concentrations and soil δ¹⁵N were less variable in residential yards compared to reference sites supporting the hypothesis that soil C, N, and δ¹⁵N converge across these cities. Increases in organic soil C, soil N, and soil δ¹⁵N across urban, suburban, and rural residential yards in several cities supported the hypothesis that soils responded similarly to altered resource inputs across cities, contributing to convergence of soil C and N in yards compared to natural systems. Soil C and N dynamics in residential yards showed evidence of increasing C and N inputs to urban soils or dampened decomposition rates over time that are influenced by climate and/or housing age across the cities. In the warmest cities (Los Angeles, Miami, Phoenix), greater organic soil C and higher soil δ¹³C in yards compared to reference sites reflected the greater proportion of C₄plants in these yards. In the two warm arid cities (Los Angeles, Phoenix), total soil δ¹³C increased and organic soil δ¹³C decreased with increasing home age indicating greater inorganic C in the yards around newer homes. In general, soil organic C and δ¹³C, soil N, and soil δ¹⁵N increased with increasing home age suggesting increased soil C and N cycling rates and associated¹²C and¹⁴N losses over time control yard soil C and N dynamics. This study provides evidence that conversion of native reference ecosystems to residential areas results in convergence of soil C and N at a continental scale. The mechanisms underlying these effects are complex and vary spatially and temporally.

Managed turfgrass is a common component of urban landscapes that is expanding under current land use trends. Previous studies have reported high rates of soil carbon sequestration in turfgrass, but no systematic review has summarized these rates nor evaluated how they change as turfgrass ages. Here we conducted a meta-analysis of soil carbon sequestration rates from 63 studies globally, comprised mostly of C3 grass species in the U.S., including 24 chronosequence studies that evaluated carbon changes over 75 years or longer. We showed that turfgrass established within the last ten years had a positive mean soil C sequestration rate of 5.3 Mg CO₂ ha^-1 yr^-1 (95% CI = 3.7-6.2), which is higher than rates reported for several soil conservation practices. Areas converted to turfgrass from forests were an exception, sometimes lost soil carbon, and had a cross-study mean sequestration rate that did not differ from 0. In some locations, soil C accumulated linearly with turfgrass age over several decades, but the major trend was for soil C accumulation rates to decline through time, reaching a cross-study mean sequestration rate that was not different from 0 at 50 years. We show that fitting soil C timeseries with a mechanistically derived function rather than purely empirical functions did not alter these conclusions, nor did employing equivalent soil mass versus fixed-depth carbon stock accounting. We conducted a partial greenhouse gas budget that estimated emissions from mowing, N-fertilizer production, and soil N₂O emissions. When N fertilizer was applied, average maintenance emissions offset 32% of C sequestration in recently established turfgrass. Potential emission removals by turfgrass can be maximized with reduced-input management. Management decisions that avoid losing accrued soil C-both when turfgrass is first established and when it is eventually replaced with other land-uses-will also help maximize turfgrass C sequestration potential.