Scale-dependent effects of landscape context on urban bee diversity

Abstract

As urbanization continues throughout much of the world, there is great interest in better understanding the value of urban and residential environments to pollinators. We explored how landscape context affects the abundance and diversity of bees on 50 residential properties in northern Georgia, USA, primarily in and around Athens, GA. Over 2 years of pan trap sampling we collected 4938 bees representing 111 species, from 28 genera in five families, constituting 20% of the species reported for the state. Development correlated positively with bee diversity at small (< 2.5 square km) scales, and positively with six of eight individual bee species’ abundances. Agriculture often correlated positively with bee diversity at larger spatial scales (> 2.5 square km), and negatively at smaller spatial scales. Forest cover correlated negatively with bee diversity at small spatial scales, but positively at larger scales. This trend was also largely true for individual bee species abundances. Bee communities differed between sites by predominant land cover types (agriculture, forest and development). Simper and indicator species analysis revealed which species contributed heavily to the observed patterns and helped to determine group distinctions. Implications for insect conservation Our results show that residential landscapes can support high bee diversity and that this diversity is sensitive to landscape context at different scales. Although development appears to have a negative effect on bee diversity overall, some bee species are favored by the open conditions characteristic of developed areas. Moreover, forest remnants appear to be valuable habitats for many species and are thus important to regional bee diversity. Urban planning that prioritizes and incorporates forest remnant conservation will promote bee abundance and diversity.

Forests are critically important to global pollinator diversity and enhance pollination in adjacent crops

“…In some cases, opposite conclusions can be reached about the relationship between forest cover and pollinator diversity depending on the scale investigated. In the southeastern USA, for example, Janvier et al (2022) found urban bee diversity to be negatively related to the amount of forest within the surrounding 250 m but positively related to forest cover at 2 km. Similarly, in Brazil, Montagnana et al (2021) reported a positive effect of forest cover on bee diversity at large scales (5 km) but found landscape heterogeneity to be more important at the local scale (1 km).…”

“…In the southeastern USA, for example, Janvier et al . (2022) found urban bee diversity to be negatively related to the amount of forest within the surrounding 250 m but positively related to forest cover at ~2 km. Similarly, in Brazil, Montagnana et al .…”

“…Most studies have found that bee communities generally respond most strongly to forest cover at scales between 300 and 2000 m (Taki, Kevan & Ascher, 2007; Winfree et al ., 2007 a ; Watson, Wolf & Ascher, 2011; Gutiérrez‐Chacón et al ., 2018; Landaverde‐González et al ., 2017; Janvier et al ., 2022; Park et al ., 2015; de Sousa et al ., 2022; Campbell et al ., 2022). Smaller species are generally thought to respond to smaller scales than larger species due to differences in foraging range and dispersal ability (Kendall et al ., 2022).…”

Pollinator cultivar choice: An assessment of season-long pollinator visitation among coreopsis, aster, and salvia cultivars

“…Clearly, there is increasing interest in planting for pollinators (Braman and Griffin, 2022). A recent study (Janvier et al, 2022) reported results from pan trap sampling 50 residential sites in and around Athens, Clarke Co., GA and documented 110 species of bees occurring in these urban and peri urban habitats. Twentytwo of the bee species collected directly from flowers reported in the present study were also represented in Janvier et al, thus further demonstrating the similarity and reliability of our collection methods.…”

Forest bees need compositionally diverse canopies

“…Using the gBuffer function of the 'rgeos' package (59), we calculated the proportion of forest in the landscape surrounding each site at 10 spatial scales (from 200 to 2000 m, using 200 m increments), hereafter referred to as landscape forest cover (LFC). These spatial scales are appropriate for bee community responses to multi-scale landscape factors (60) and are within accepted distances of bee dispersal and foraging (55,61). We used Pearson correlation between species richness and LFC at all scales to determine the spatial scale at which most species respond (62; Supplementary Table S5).…”