Can Both Ecosystem Service Providers and Rare Species Be Restored With the Same Plants? An Experimental Approach Using Plant-Pollinator Networks

EPA Grant Number: F13A10020
Title: Can Both Ecosystem Service Providers and Rare Species Be Restored With the Same Plants? An Experimental Approach Using Plant-Pollinator Networks
Investigators: MacLeod, Molly Katherine
Institution: Rutgers, The State University of New Jersey
EPA Project Officer: Lee, Sonja
Project Period: September 2, 2014 through September 2, 2016
Project Amount: $84,000
RFA: STAR Graduate Fellowships (2013) RFA Text |  Recipients Lists
Research Category: Academic Fellowships , Fellowship - Ecology


Plant-pollinator networks can facilitate the convergence of biodiversity and ecosystem services-based approaches to conservation, but apparent network patterns may be driven by skewed species abundance distributions. This research will use an experimental approach to networks to meet the following objectives: (1) Experimentally test the attractiveness of 17 native plant species, in single-species plots, to rare and key ecosystem service-providing (crop-pollinating) bees, determining whether both groups prefer the same plants; (2) identify annual variation in the network structure and in the identity of core generalist plant or bee species; (3) determine whether variation in partner identity is driven by bee foraging preferences or by species abundance distributions; and (4) explore the role of species abundance distributions on the outcome of simulated species extinctions from pollination networks.


In the field experiment, pollinators have been collected from 102 monospecific plots of 17 plant species over three field seasons. The plant species are established in equal relative abundance, thus removing a sampling effect driven by a skewed plant species abundance distribution and allowing a direct measurement of pollinator preference. To meet the first research objective, rare and ecosystem service-providing (ES) bee species were defined using independent databases on the bees of the region. Mixed models were used to determine whether plant species varied in the abundance of rare or ES bees, and a rank correlation tested whether plant species that attract a high (or low) abundance of rare bees also attract high (or low) abundance of ES bees. Statistical models will determine the role of species abundance distributions in apparent annual network variation and on the simulation of species extinctions.

Expected Results:

This research will expand on preliminary results that suggest that rare and ES bee species are separate sets of species that have similar preferences that are not due to plant species abundance alone. The research will expand upon this analysis and will explore changes in the plant-pollinator network over time. It is expected that (1) network structure, including the identity of ‘core’ generalist plant or bee species, will be similar across years, indicating that the same plant species can be used to restore rare and ES bee species, and that these plant species will be important across years despite any annual changes in the bee community. (2) Apparent annual variation in pairwise interactions (“rewiring”) will be driven more by species abundance distributions than by actual changes in preferences. (3) The relative importance of generalist and specialist species to network robustness to species loss will be driven more by relative abundance than by diet breadth.

Potential to Further Environmental/Human Health Protection

This work will contribute to EPA’s Pollinator Protection Plan, and will identify methods for conservation of both rare species and key ES species with the same management actions and will develop educational tools to broadly communicate the results. The results will inform the design of existing and future pollinator restorations mandated by the Pollinator Protection Research Act of 2007 and funded by the U.S. Farm Bill. Although Farm Bill programs are not explicitly intended as ES or rare species conservation programs, they can serve both purposes if suitably designed. The multiyear experimental approach also will facilitate the development of novel educational tools for landowners to design pollinator restoration that accounts for and minimizes temporal fluctuations in the populations of the target species.

Supplemental Keywords:

biodiversity, ecosystem services, pollinators

Progress and Final Reports:

  • 2015
  • Final