Host–parasitoid networks in urban gardens across gradients of garden management and land-use in the California central coast

Background/Question/Methods

Ecological networks depict interactions (links) between species (nodes), illuminating community interaction structure. Network quantification reveals that landscape changes (e.g. agricultural intensification, urbanization) can alter network structure with consequences for ecosystem service delivery. Parasitoid wasps are one of several natural enemies that attack herbivorous insects in urban community gardens, and parasitoid-herbivore interactions may be shaped by both local garden management characteristics and landscape surroundings. This study focuses on interactions between herbivores (hosts) and their parasitoids and asks: 1) How do local management and landscape context influence host abundance, parasitism, and host–parasitoid network structure? and 2) Does host–parasitoid network structure predict host parasitism? Over two summers (2017 and 2018), we sampled herbivores from Brassica oleracea at 22 urban gardens and reared parasitoids from herbivores under laboratory conditions. We identified herbivores and parasitoids to morphospecies and used observed interactions to construct bipartite networks. We surveyed local management characteristics (e.g. canopy cover, vegetation, ground cover) and quantified land-cover types (e.g., urban cover, remnant natural habitat, agriculture) surrounding gardens. We used generalized linear and generalized linear mixed effects models to explore relationships between host abundance, parasitism (at the level of site, plant, and host individual), network metrics, and local and landscape characteristics.

Results/Conclusions

We sampled 60,245 herbivores, representing 9 species, and reared 612 parasitoids from 1,995 hosts, including 23 morphospecies. The most abundant herbivores were cabbage aphids (n = 57,346), whiteflies (n = 1,312), and cabbage worms (n = 812). Cabbage worm abundance increased with flower species richness in gardens, but no other factors influenced herbivore abundance. Site-level whitefly parasitism was higher in sites with greater whitefly abundance. Similarly, plant-level prevalence of cabbage aphid parasitism increased with host abundance as well as urban cover in the landscape and decreased with grass cover in gardens. Network vulnerability (average number of parasitoids per host) was higher in gardens with more grass cover and lower in gardens with more brassica plants. Host-level cabbage aphid parasitism increased in gardens with higher network interaction richness, but no other network metrics were predictive of parasitism. Overall, we found that local management more strongly predicted host abundance, and host abundance was a strong predictor of parasitism. Local (grass) and landscape (urban cover) characteristics influenced parasitism and network metrics but results varied depending on host species and the scale at which parasitism was measured.