Assessing food web dynamics to improve biocontrol of pests

Biological control of pests is deeply needed to reduce chemicals inputs in agriculture. This service depends on employing natural enemies to control pest. However, the effects of management on the efficacy of biocontrol have been difficult to predict. A reason for this is the many pathways through which this service can arise; for example, natural enemy species often have complementary roles and are likely to adapt their behaviour depending on settings. This means that the regulation of biocontrol services depends on a temporal turnover of adaptations within the natural enemy community, starting with the arrival of enemies in fields. Despite this, temporal changes within natural enemy roles and their effect on biological control have only poorly been investigated. The current project addresses this by - over time - experimentally assessing how interactions within the natural enemy community build up a reliable and efficient biological control of pest. As the strength of competition between natural enemies is likely to change biocontrol levels, we will manipulate resource availability, i.e. the abundances of pest and non-pest prey. This will be done by administering manure and unfertilized treatments known to elicit such changes. As aphids are major worldwide pests that cause serious economic damage to farmers, we will use aphid infested cereal fields as a model system. Aphids are attacked by a suite of natural enemies including parasitoids, generalist and specialist predators. Thus, as the network of interactions and multitude of alternative pathways from which an efficient biological control of aphids could be built up is expected to be complex, we will interpret results from a food web perspective. In the current project we will generate high-resolution time series of pest-parasitoid-predator-non-pest prey food webs based on molecularly-derived data of trophic interactions between each of the natural enemy guilds therein. There are four aims for this research program: 1. Generating high resolution time series food webs in replicated fields during the cereal growing season. 2. Quantifying the effect of fertilizer-induced food web changes on biological control asserted by natural enemies. 3. Identifying the natural enemies contributing most strongly to biological control, and the time points and food web settings that are critical for a reliable biological control. 4. Determining sensitive periods that may immediately, or subsequently, cause the loss of biological control potential. The insights gained from this work will contribute to make biological control more predictable and show how manure treatments can be used for maximizing biological control of agricultural pests.

Assessing food web dynamics to improve biocontrol of pests To reduce pesticide inputs in agriculture, pest control tactics which utilize natural enemies to control pest are of high interest. However, the efficacy of biocontrol depends on many pathways through which this service can arise making the efficacy of this ecosystem service difficult to predict. Especially the temporal changes within natural enemy food webs and their effect on biological control have been poorly investigated so far. This project addresses this by - over time - by experimentally assessing how interactions within the natural enemy community are related to biological control of pests. We generated a high-resolution time series data set for invertebrate communities in three replicated cereal fields over two years and measured the generalist predator diet data using DNA-based techniques. Additionally, the effect of organic fertilization was tested on prey availability, predator competition and intraguild predation by adding manure in experimental plots within each field. We found that organic fertilization positively correlated with non-pest prey availability and negatively so with intraguild predation. Pest spread in the fertilized treatment decreased when intraguild predation was lower, although no effect was found for cereal aphid densities. Food web network-level specialisation was lowest during mid-season and it was uncorrelated to fertilization. Predator diversity was lower in the fertilized treatment in the first year, decreasing over the season, while rising slowly in the second, with no effect of fertilization. Prey diversity, on the contrary, had identical trends, showing a peak during mid-season. The food web parameters also differ between years, with weighted connectance declining over time only in the second one. Both trophic link rewiring and community composition dissimilarities fluctuated more in the second year, with rewiring being almost always higher. Our findings show that organic fertilization can contribute to biological control, by containing pest spread within the fields. The variations in network specialisation, species specialisation and prey diversity point to phenology, rather than fertilization, as the driver of the behavioural shifts in predators, as they adapt to prey abundance. These shifts in behaviour are also mostly responsible for the changes in the food webs, as the link rewiring was higher than community dissimilarity. From a practical perspective our findings imply that early and late cropping season are vulnerable periods for biological control, when predator behaviour is most constrained and their dietary overlap is lowest. These results show how quickly invertebrate communities and food webs change during the season, further reinforcing the need for time series data to accurately represent food webs to a gain better understanding of biological control.