One of the features that distinguishes urban agroecosystems from rural agroecosystems is the remarkable heterogeneity in habitat structure and resource availability at the local scale. In community gardens, several gardeners grow crops in a communal space, but each gardener typically manages the composition and arrangement of crops in their own plot in ways that differ from management of neighboring plots. In this field experiment, I investigated the role of garden plot-level structural complexity. Specifically, I consider the importance of vegetation density and connectivity in determining the abundance and richness of herbivorous insects and their natural enemies. I worked in 8 community gardens in Santa Cruz and Monterey counties. I selected 4 plots per garden in which to conduct vegetation surveys, insect surveys, and an experimental manipulation artificially enhancing connectivity between crops. I assigned each plot to one of four treatments: control-low vegetation density (CH), control-high vegetation density (CL), +strings-low vegetation density (SL), and +strings-high vegetation density (SH). The +strings (S) manipulation consisted of adding 1.5 x 1.5 m grids of bamboo stakes strung together with butcher’s twine to increase the connectivity between crops in S treatment plots. Garden vegetation and insect surveys were conducted before and after the manipulation.
Results/Conclusions
I recorded nearly 2000 individual insects representing 5 orders of herbivores and 5 orders of enemies. The most abundant orders of enemies were Hymenoptera (n = 340), Araneae (n = 58), and Diptera (n = 47), while herbivores were most represented in the order Hemiptera (n = 469). On average, herbivore abundance (chi-squared = 13.608, df = 1, p-value = 0.0002252) and herbivore richness (chi-squared = 18.479, df = 1, p-value = 1.718e-05) were higher in treatment plots after applying the strings manipulation. Average enemy richness was lowest in SL treatment plots and higher in CL, CH, and SH plots (chi-squared = 9.5927, df = 3, p-value = 0.02237). These results support the hypothesis that greater vegetation density supports natural enemy richness. Further, herbivores may be more sensitive to local-level changes in connectivity within urban garden plots, while natural enemies may be less sensitive to local-level changes. Implications for biological pest control are not yet clear as natural enemy richness can facilitate and interfere with pest control depending on intraguild interactions. Moreover, dispersal via spatial complexity can lead to stability in predator-prey population dynamics (Huffaker 1958), so greater herbivore abundance and richness may not mean less effective biological pest control.