Urban fragmentation leads to lower floral diversity, with knock-on impacts on bee biodiversity

Urbanization, a major driver of anthropogenic land-use change, negatively impacts biodiversity through habitat fragmentation, degradation, pollution, and increased impervious surfaces. Bees and flowering plants, closely interacting organisms, are particularly vulnerable. While a correlation exists between flowering plant richness and pollinator richness, the causal relationship and the effects of urban fragmentation remain unclear. This study addresses this gap by examining how bee and flowering plant species richness, phylogenetic diversity, and functional trait diversity interact in response to urban fragmentation. The researchers hypothesized that urban fragmentation would negatively affect both bees and plants, with a bottom-up effect (plants influencing bees) dominating their interaction. Understanding these relationships is crucial for effective urban biodiversity conservation strategies.

Previous research extensively documents the strong correlation between flowering plant species richness and pollinator species richness, with parallel declines observed in insect pollinators and insect-pollinated plants in some regions. Studies show that increased flowering plant richness enhances pollinator community stability, while pollinator biodiversity impacts plant diversity, density, reproduction, and seedling abundance. Plant-pollinator interactions are complex networks influenced by species' phenotypic traits, functional traits regulating interactions, and biotic filtering. Urbanization's effects on bee species richness vary (negative, neutral, or positive), influencing bee functional diversity. Similarly, urbanization’s impact on native plant species richness is mixed, although some cities show high native plant richness. Despite this research, the co-variation of bee and floral diversity within urban landscapes remains poorly understood.

The study sampled bee and flowering plant communities monthly from June to August 2017 in eight urban semi-natural sites in Halle (Saale), Germany. Transect sampling was used to collect wild bee specimens (845 individuals, 63 species) and flowering plant specimens (58 species). Bee traits (body size, sociality, tongue length, nesting behavior, voltinism, and lecty) and plant traits (nectar holder depth, flower color, shape, breeding system, flower sex timing, and longevity) were measured. Four community diversity metrics (abundance, species richness, functional diversity, and phylogenetic diversity) were calculated. Local (patch) environmental variables (bare soil cover, patch size) and landscape-scale variables (habitat heterogeneity metrics at a 1000m radius, including proportion of different land cover types, edge density, fragmentation) were quantified. Linear mixed effects models (LMMs) and generalized linear mixed models (GLMMs) were used to analyze relationships between bee and plant diversity metrics and environmental variables. Fourth-corner analysis was used to explore relationships between bee functional traits and environmental variables. Piecewise structural equation modeling (SEM) was employed to infer causality in plant-bee relationships within the urban ecosystem. All quantitative predictors were standardized, and model assumptions were checked.

The study found a positive relationship between bee species richness and flowering plant richness, and between bee abundance, flowering plant richness, and bare soil cover. Bee functional diversity decreased with increasing community weighted mean (CWM) nectar holder depth. CWM bee body size was positively related to CWM plant nectar holder depth. Urban fragmentation negatively impacted flowering plant species richness, functional diversity, and phylogenetic diversity. Piecewise SEM confirmed the strong relationships between bee biodiversity, local ground nesting resources, floral richness, and urban fragmentation, showing a bottom-up effect (plants influencing bees). Urban fragmentation indirectly negatively affected bee diversity through its effect on flowering plant richness. Fourth-corner analysis revealed various associations between bee functional groups and environmental variables. For example, oligolectic bees were positively associated with residential cover, while polylectic bees were positively associated with flowering plant richness and the proportion of allotment gardens and parks. Ground nesting bee abundance was positively associated with local flowering plant richness and bare soil cover.

The findings demonstrate a strong interdependence between bee and flowering plant diversity in urban areas. The negative impact of urban fragmentation on flowering plants, in turn, negatively affects bee diversity, highlighting the importance of considering indirect effects mediated by species interactions. The bottom-up effect suggests that plants may have alternative reproductive strategies reducing their dependence on pollinators compared to pollinators' dependence on plants. These results support the need for local management practices to increase native plant diversity in urban areas to benefit wild bees. The study also showed that diverse land-use surrounding urban green spaces can benefit various bee functional groups, emphasizing the importance of both local and landscape-scale resource availability for bee conservation in cities.

This study reveals strong relationships between bee and flowering plant diversity in urban semi-natural habitats and the negative impact of urban fragmentation on both. The bottom-up effect and the indirect negative impact of fragmentation on bees through its effects on plant richness are significant findings. Increasing native plant diversity in urban areas through simple management practices is crucial for bee conservation. Future research could focus on exploring the effects of fragmentation at different spatial scales for various bee species and the interactions of multiple resources (food and nesting).

The study was conducted in a single city, limiting the generalizability of the findings. The sampling period (June-August) might not fully capture seasonal variations in bee and plant communities. The use of transect walks could potentially introduce bias; while the variable transect method was used to mitigate spatial variation, the potential for uneven sampling remains. The analysis focused on specific functional traits, possibly overlooking other factors influencing plant-bee interactions.