No evidence for landscape composition effects on biological control in Wisconsin soy agroecosystems despite evidence of effects on natural enemy abundance
Arthropods in agricultural fields can be affected by the composition and configuration of the surrounding landscape. The diversity and abundance of insect predators and parasitoids is widely accepted to be higher in areas with a greater proportion of non-crop, “semi-natural” area in the vicinity. As a consequence, in landscapes with more non-crop areas, previous research also shown the ability of entomophagous arthropods of crop pests to consume pest insects to be higher. However, these two relationships are not always tested in the same studies - leaving the correlation to be assumed to be causation. Thus we investigated how landscape composition simultaneously affected insect crop pest abundance, natural enemy abundance and community composition, and biological control. We also assessed other potential determinants of crop arthropod abundance and community composition, including prey density, growing degree days, and seasonal cumulative precipitation. Our study was conducted in 60 commercial soybean fields across Wisconsin in 2013 and 2014, focused on soybean aphid (Aphis glycines) and its generalist predators. We used direct observation and field counts of soybean aphid and sweep sampling and unbaited sticky card traps to collect predators. Biological control assays were performed using clip cages and egg cards.
Using mixed effects models and multivariate analysis, we found significant negative effects of growing degree days on soybean aphid abundance and no effects of any landscape variables. Similarly, we found overall negative effects of growing degree days on natural enemy abundance. However, we also found that if natural enemies are present, landscape variables have significant effects on them depending on the arthropod functional group of interest. Yet we found no significant effects of landscape variable included in our analysis with biological control potential as measured by either assay. Our results suggest that other factors are contributing to the potential of generalist predators to be effective biological control agents in Wisconsin soybean agroecosystems. Future work by our group will test whether on-farm management appears to be driving biological control instead.