Heterogeneous energetic, chemical, and biological human inputs concomitant with urbanization and intensive agriculture impact the function and composition of terrestrial ecosystems worldwide. Population and community level responses to anthropogenically-driven landscape change are taxon and context specific so that predicting biotic responses in coupled human-natural systems presents a substantial challenge. Our aim is to characterize effects of human landscape changes on spatial distributions, abundances, and ultimately patterns of gene flow and dispersal in specialist insects across trophic levels. Specialists constitute the majority of described insects and are expected to be more sensitive to habitat modification than generalists. We are focusing on two systems in which fruiting tree species (Prunus serotina and Juglans nigra) are host to specialist fruit flies (Rhagoletis cingulata and R. suavis), which are in turn host to specialized parasitoid wasp taxa. We assessed 250 random 150 x 150 meter sampling plots in a 360 km2 urban/ agricultural landscape for host tree presence, followed by two seasons of insect collections. We have compared mean densities of the three trophic levels across broad- and fine-scale landcover categories. We have also utilized geographical tools to assess patterns of spatial autocorrelation and the explanatory power of land cover with differing spatial scale.
Results/Conclusions
Preliminary results show that both plant-herbivore-parasitoid systems under study are sensitive to spatial patterns of human development. Analyses of organismal densities across broad- (natural, agricultural, and city) and fine-scale (plot-level landcover) indicate that, while host tree densities are significantly higher in city than agricultural areas, mean fly density over two seasons was lower in commercial/ residential settings for both systems across scales. Both R. cingulata and R. suavis densities exhibit significant clusters of negative spatial autocorrelation predominately in areas of urban and high-density residential development, which is consistent with a hypothesis of dispersal limitation in these areas. Spatial regression analyses indicate that landcover proportions including urban development and cropland at 2-3 km around plots best predict fly densities for agricultural areas alone, while the best predictive landcover models for fly densities in the city are at a finer scale (500m). This latter finding is also consistent with limited dispersal in cities. Results for parasitoids are mixed; wasps infested R. suavis across all landcover categories, but did not infest R. cingulata in the city. These results provide promising evidence that the most intense anthropogenic development serves to decouple associations across trophic levels.