Delimiting plant diversity that is functionally related via interactions with diurnal pollinators: An expanded use of rarefaction curves

- How much plant diversity is functionally related by interactions with pollinators?
- Three plant communities located in geographically distant hotspot regions were studied.
- Co-flowering communities are not necessarily as diverse as the region where they occur.
- Pollinators made an incomplete selection of available floral resources.

The way in which taxonomic diversity relates to functional diversity is important for understanding the mechanisms that sustain ecosystem function and services. We investigated how an explicit consideration of plant-pollinator interactions influences our view of plant diversity. We studied three plant species-rich communities located in different biodiversity hotspot regions: two soil-specific plant communities with Mediterranean-climate, sandy dolomite outcrops in Andalusia (Spain) and serpentine seeps in California (USA), and a third community in the sub-tropical dry scrublands in Yucatan (Mexico). Sampling at three spatial scales (region, site, plot) and rarefaction analyses were used to characterize and compare spatial and temporal variation of entomophilous plant diversity based on species presence (“static plant diversity”), flower display sizes along the season (“dynamic flower diversity”), and pollinator visitation (“interaction-effective diversity”). The studied communities differed in the static diversity of plants, with sub-tropical dry scrublands being less diverse than the two Mediterranean communities. Reduction of static diversity at local scale was stronger in the richest Mediterranean communities and, thus, static diversity was similar among regions when considering finer (site-level) spatial scales. In addition, the two Mediterranean communities displayed more seasonal variation, thus reducing differences in dynamic diversity among regions, i.e. when considering finer temporal scales. These results suggested that, at finer spatio-temporal scales expected to be relevant for interactions with pollinators, plant communities are not necessarily as diverse as the region where they occur. Accordingly, interaction-effective diversity based on pollinator visitation was in all cases lower than expected relative to the diversity of flowers. Thus, diversity of visited flowers does not perfectly track diversity of flowering species but instead reaches asymptotes at much lower values than expected in the richest communities. Regional species diversity may support the functionality of interactions at broader spatial and temporal scales than they actually occur.