COS 122-1
The direct and indirect effects of removing ungulate grazing on pollination in a pastoral system

Thursday, August 14, 2014: 1:30 PM
315, Sacramento Convention Center
Daniel S. Song, Department of Biology, University of Pennsylvania, Philadelphia, PA
Pierre Liancourt, PIRE Mongolia Project ( and Academy of Sciences of the Czech Republic, Institute of Botany, Třeboň, Czech Republic
Laura A. Spence, Department of Biology, University of Pennsylvania, PIRE Mongolia Project (, Philadelphia, PA
Bazartseren Boldgiv, Ecology Group, Department of Biology, National University of Mongolia, Ulaanbaatar, Mongolia
Peter S. Petraitis, Department of Biology, University of Pennsylvania, Philadelphia, PA
Brenda B. Casper, Department of Biology, University of Pennsylvania, Philadelphia, PA

Maintaining species interactions is critical to preserving important ecological services.  Land use change can alter biodiversity directly and trophic interactions indirectly.  Commonly, the changes involve modifications to active land management, such as increases in agricultural intensity or urbanization.  However, the important consequences of ceasing, rather than increasing, management are often overlooked.  Historically grazed systems may face grazing cessation due to societal changes and could trigger important consequences for trophic interactions.  We asked how the grazing cessation affects flower and pollinator community composition and their network of interactions.

We studied the effects of excluding grazers, primarily yaks, on forb and pollinator communities in northern Mongolia.  The mountain steppe in this region has been subject to nomadic pastoralism for hundreds of years.  Our study was conducted throughout the summer flowering season.  In our split plot design we allowed one 2x2m plot in each of six pairs to be grazed during late summer through spring, while the other plot was fenced year round to exclude grazers.  We compared community composition of forbs and insect pollinators between the two treatments using constrained analysis of proximities (CAP) followed by a permutation test.  We also compared the number of insect visits and plant-pollinator network topology (ANOVA).


We observed 33 forb species and 100 pollinator morphospecies. Forb community composition differed between treatments.  The community composition of only one Order, Hymenoptera, varied between treatments, although overall insect visitor community composition did not differ.  Fewer pollinator visits were made to plots where grazing was excluded.  Visits from Hymenoptera species responded most strongly, with a decrease of 50 visits (20%).  Overall, flower abundance was greater in plots where grazing was excluded.  Thymus gobicus and Gallium verumresponded most strongly to grazing exclusion and in ungrazed plots, each gained an average of 3,483 (209%) and 1,118 (589%) flowers, respectively.  Excluding these two species, in the ungrazed plots, a total of 19 species increased in flower abundance, by an average (±SE) of 51 (±12); in ungrazed plots, 12 species increased, by 67 (±25).  Nestedness and connectance did not differ between treatments.

Cessation of ungulate grazing resulted in flower abundance and pollinator visitation changing in opposite directions.  While community composition of forbs and pollinators varied between treatments, network topology did not significantly differ between treatments.  Our results suggest that removing an antagonistic interaction, such as grazing, plant-pollinators interaction may have the potential to buffer any indirect effects to preserve an important ecological service.