1. Temperate deciduous forests in the United States are located in the most densely populated states across the northern and mid-Atlantic east coast. Land development and associated human activities result in small forests that are susceptible to anthropogenic influences, such as urbanization and non-native plant invasion. 2. The overall objective of this study was to assess spatial and temporal drivers of forest vegetation structure and diversity in small forests embedded across developed landscapes. We assessed woody plant composition across spatial gradients (i.e., urbanization and non-native plant invasion) and a temporal gradient (i.e., time since canopy closure) across 38 forests along the east coast of the United States in northern Delaware and southeastern Pennsylvania. 3. Surprisingly, we found the invasion gradient was not related to the urban gradient across our forests. Across all forests, the canopy consisted of native species (98% of all trees), whereas the forest understory was most vulnerable to nonnative plant invasion (65% of all woody plant stems). Greater native species richness in forest canopies and understories with increasing urbanization supports the conclusion that urban forests maintain native species and are not inherently degraded ecosystems. Non-native plant invasion has a strong influence on understory plant communities, and the duration of intact forest canopy had a strong negative correlation with non-native plant invasion suggesting intact forests can resist invasion. 4. Synthesis. This is the first study to compare simultaneously the importance of invasion and urbanization in determining plant community composition in forests embedded across developed landscapes, and to discover that younger forests harbor more invasive plants and urban forests maintain native plants. K E Y W O R D S nonmetric multidimensional scaling, non-native plant invasion, temperate deciduous forest, urban forests, urbanization, woody plant community | 2367

Despite their diversity and abundance, the importance of native, forest bee communities to pollination services and inherent biological diversity conservation is often overlooked. We studied forest bee communities in Delaware and Pennsylvania, USA, to better understand how forest bee community structure varies with changing land use and microhabitat quality among small, urban and suburban forest fragments in the mid‐Atlantic United States. Our hypotheses were that (1) microhabitat quality would affect relative abundance of bee taxa, (2) surrounding landscape composition would drive local patch colonization–extinction dynamics, and (3) forest patch size would not affect community structure and composition. We found a lack of spatial autocorrelation among forest patches, indicating the importance of individual fragments in the autonomous generation and/or maintenance of bee communities. Community analyses revealed the importance of both landscape context and microhabitat quality in defining forest bee communities. By partitioning beta diversity into its constituent components, we also found that landscape composition drove changes in the relative abundance of taxa, while both landscape and microhabitat characteristics significantly influenced species turnover. Neither landscape composition nor quality of the microhabitat influenced initial site occupancy or colonization–extinction dynamics. Bisected by highways, suburban neighborhoods, agricultural lands, and urban development, many urban and suburban forest patches in the eastern United States are similar in composition to our study sites. As many native forest bees have limited capacity for long‐range movement between forest patches, these remaining forest fragments are critical to the conservation of unique and speciose forest bee communities.

No abstract

No abstract