Response diversity as a mechanism for stabilizing pollinator communities and pollination function against land-use change

Background/Question/Methods

Biodiversity may stabilize ecosystem services against environmental change if multiple species providing the same service respond differently to a given change.  This response diversity hypothesis has been investigated primarily by using species abundances as a proxy for the function they deliver, thus leaving the response of the function itself unexplored. Here, we report both abundance-based and function-based measures for a diverse community of pollinators. We used native bees pollinating cultivated cranberry (Vaccinium macrocarpon) as a model system to investigate species-specific responses to land-use change. We measured land use as the proportion of agricultural land cover surrounding each cranberry bog. To measure the abundance of each bee species, we net-collected bees foraging on cranberry flowers at 16 farms in 2009 and 2010. To measure the pollination function provided by each bee species, we presented virgin cranberry flowers to individual bees and counted pollen grains deposited in a single flower visit. We then estimated the pollination function provided by each species at each farm as the product of the mean number of pollen grains deposited per species per visit, and its abundance at the farm. We tested for response diversity in each outcome variable (bee abundance and pollination function provided) using generalized linear models with the predictor variables proportion of surrounding land cover in agriculture, year, and bee species. A significant interaction term between bee species and land cover indicates response diversity.

Results/Conclusions

We collected 2995 individual bees of 50 species; of these, 9 species, represented by 2824 specimens, had sufficient data for use in statistical analyses. We found significant response diversity for bee abundance (bee species x land cover interaction, F_{8, 160} = 8.95;  P = < 0.001). Seven of the nine dominant pollinator species decreased in abundance with increasing agricultural land cover, with the slopes of these relationships differing among species. The remaining two species increased in abundance with increasing agricultural land cover.  The response diversity was even stronger when abundance was translated into function (bee species x land cover interaction, F_{8, 262} = 14.04;P = < 0.001). This is because the species with highest pollen deposition value, Bombus impatiens, increased with agriculture, while the species with the lowest mean pollen deposition value also had the strongest negative response to agriculture. Our results suggest that bee diversity may mitigate the potentially negative effects of land-use change on pollination services.