Pollination services from native bees: What scale matters?

Background/Question/Methods

The loss of natural habitat is associated with reduced pollination services from native bees, but the relative importance of habitat availability at the local versus the landscape scale is not well understood. Further, bee species of different body sizes may respond at different scales. The goals of our study are to measure the relative importance of habitat loss at these two scales for native, wild bee pollinators of highbush blueberry (Vaccinium corymbosum), and to determine whether body size predicts the scale of response. We worked at 16 blueberry farms in southern New Jersey that varied in the proportion of non-crop vegetation within a 300m radius (“local scale”) and within a 1500m radius (“landscape scale”). The study was designed so that landscape and local variables were uncorrelated. Each farm was visited on three different days during bloom. On each visit, we observed pollinators visiting blueberry flowers and net-collected pollinator specimens during standardized transect surveys. As a measure of pollinator abundance, we pooled observations of bees visiting flowers (n=824) and specimens netted from flowers (n=421). As a measure of diversity, we used the number of bee species net-collected at a given farm. The data were analyzed in a repeated measures design.

Results/Conclusions

When bees were divided by body size into large (n=724) and small (n=521) body size categories, abundance of large bees showed a significant negative response to land use at 300 m (p=0.02) and a much stronger negative response to land use at 1500 m (p=0.0001), while abundance of small bees had a significant negative response to land use only at the local scale (p=0.004, vs. p=0.10 for landscape).  Species richness was consistent with these patterns. Large bees significantly increased species richness at the landscape scale (p=0.02) but did not show as strong an effect of local scale (p=0.09). Small bee species richness was strongly affected by agriculture at the local scale (p=0.0009) but had little response to the landscape scale (p=0.20). These results suggest that scale matters: the relationship between crop-visiting pollinators and human land use varied with both the scale at which land use is examined, and the body size of the bee. All trends were negative, however, suggesting that conservation of ecosystem services will require land use planning at multiple scales.