UC Davis

Soil transitions underly many terrestrial landscapes and are well-established drivers of plant community structure. Yet, edaphic characteristics are frequently excluded from hypotheses of how environmental features structure ecological communities for higher trophic levels. In this dissertation, I test whether a particular type of soil, serpentine, structures bee communities and the flowers they depend on. Building on a foundation of classic ecological studies, I establish where and why flowers and bees distribute themselves across the mosaic landscapes of California serpentine soils by using both community and functional ecology tools. Functional trait ecology plays a key role in revealing what mechanisms filter local bee communities. And finally, I categorize how these bee and flower communities produce emergent patterns of interaction using network approaches. This dissertation demonstrates that soil, and perhaps many other underlying environmental conditions, not only filter and structure plant communities, but also the wild bee communities that visit them.

Chapter One documents and categorizes wild bee and flower communities across serpentine and non-serpentine meadows in California’s coastal mountain ranges. Soil characteristics are well-established environmental drivers of local plant community diversity and composition, but their effects on mobile organisms in higher trophic levels are less well understood. We analyzed whether pollinator (mostly wild bee) and floral communities—distributed across a discrete serpentine soil boundary in California’s coast ranges—differed according to the underlying soil type (serpentine or non-serpentine). We found that wild bee richness and abundance were significantly lower in non-serpentine meadows, but the taxonomic composition of bee communities did not differ between soil types. Spring floral richness, abundance, and composition were also lower in non-serpentine meadows, but floral communities had different compositions on the two soil types. In both bee and floral communities, there were strong phenological effects on abundance, richness, and composition over the course of the spring season that were parallel. Both bees and flowers were markedly more diverse and abundant in serpentine meadows, particularly in the late spring when non-serpentine meadows stopped flowering.

Chapter Two utilizes a functional trait framework to test whether soil characteristics affect the functional diversity of wild bee communities, either directly or indirectly via changes to vegetation-related nesting and foraging habitat quality. We demonstrate that soil type affects the functional diversity of bee communities, with those on serpentine soils being more functionally rich. Rather than direct soil-bee impacts, soil type appears to indirectly filter bee communities via interactions between vegetation-based habitat quality and bee nesting and foraging traits. We use a fourth-corner analysis to show that nesting and foraging habitat quality correlates with particular bee functional traits. Specifically, above-ground nesting bees are filtered out of serpentine meadows, and late-flying bees are filtered out of non-serpentine meadows. Despite a growing literature on landscape drivers affecting bee community functional richness, the indirect pathways filtering bees are rarely quantified. In contrast to the strong direct filters that infertile soils exert on plant communities via functional response traits, the indirect effects on pollinators are more complex. Soil fertility’s indirect effects of “cascading up” to structure the functional diversity of other higher trophic communities may be a broader pattern, but evidence is scant.

Chapter Three takes a network perspective to understand how interactions (and not only species) turn over across serpentine and non-serpentine grasslands. Ecologists have long sought to understand and dissect interaction networks among co-occurring species and to understand if these interactions assemble into networks with common emergent properties like nestedness, modularity, and specialization. In addition, how interactions turn over—and consequently affect network structure and resilience to perturbations—across other environmental gradients is a field of important, and current, focus. We analyzed how soil type affected the network structure and interaction turnover of plant-pollinator communities. We demonstrate that plant-pollinator network structure differed significantly between serpentine and non-serpentine meadows. We also utilized a parallel analysis of network microstructure, where we focused on the turnover of interactions between networks. The difference in overall network structure appears to be driven by interaction turnover; serpentine networks exhibited lower interaction turnover between themselves than non-serpentine networks, and interaction rewiring—when shared species switch interaction partners—contributed to turnover between serpentine networks more frequently. Interaction rewiring contributed very little to interaction turnover in general, and only ~20% of interaction turnover between sites was among shared species. And when compared to a regional pooled meta-web, serpentine networks were more unique than non-serpentine networks. Although serpentine meadows are more unique in a regional context, the more frequent rewiring and lower interaction turnover between serpentine meadows drives less specialized, more resilient plant-pollinator networks in these low-resource environments.