COS 83-8
Navigating the urban matrix: Pest population dynamics in an urban agroecosystem

Thursday, August 8, 2013: 10:10 AM
101I, Minneapolis Convention Center
Theresa Wei Ying Ong, Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI
John H. Vandermeer, Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI
Background/Question/Methods

Urban agriculture represents a unique ecosystem for which the study of spatially explicit questions is especially convenient. From a metapopulation perspective, urban gardens resemble habitat patches for garden pests dispersed within inhospitable urban space. In this study, we explored how pest population dynamics and connectivity between patches is influenced by both local management choices and landscape level factors such as the intensity of the built environment. Populations of pea aphids were surveyed in community gardens located throughout Ann Arbor, MI in both rural and urban settings with high and low natural enemy diversity. Coefficient of variation was calculated for gardens using time series data in order to compare stabilities of individual populations. Finally, we developed a simple metapopulation model to explore how patch population stability at the local level could influence patch connectivity at the landscape level.

Results/Conclusions

Pest populations in urban gardens with high natural enemy diversity are less variable than sites with less diversity. Gardens in rural areas are overall less variable than those in urban areas, but there is no clear difference for sites with high or low natural enemy diversity. Our results indicate that diversity of natural enemies has a strong negative influence on pest population variability, thus increasing stability of those sites. Increased stability in individual gardens may reduce the probability of spillover to neighboring gardens. However, these results only apply to urban sites where habitat may be of lower quality and aphid populations at higher risk to attack by natural enemies. Model results indicate that high levels of variability in pest populations increase risk of dispersal to neighboring patches as well as extinction in individual patches.