IAP-24-039

Assessing ecosystem service trade-offs in omnivorous generalist beetles using network inference, molecular dietary analysis and nutritional networks

The interactions of generalist consumers are crucial determinants of ecosystem structure and function, but they can also be difficult to predict given the many food sources generalists access, and the way in which these interactions dynamically change over space and time. As our ecosystems are impacted by global change, it is even more important that we can understand and predict these interactions to mitigate potential impacts to ecosystem stability. This is also crucial for understanding and optimising the provision of ecosystem services and disservices, which determine the benefits we gain from these interactions, such as predation of crop pests or, conversely, herbivory of crops.
Ecological networks are a valuable means for investigating interactions across whole communities or ecosystems to understand the function and stability of ecosystems, and the properties that structure interactions across space and time. This can be particularly crucial for understanding the impact of perturbations like global change on interactions and their outcomes, such as ecosystem services. Constructing these networks empirically can, however, be time-consuming and challenging, so finding streamlined ways to do so is paramount if we are to use networks across a broad range of spatiotemporal contexts. Inferring interactions is one such method, which involves linking consumers and resources in a way that should resemble the interactions actually occurring. This is particularly important for difficult-to-observe systems such as the interactions between invertebrate consumers and their resources.

Many ground beetles (Carabidae) are highly abundant generalist omnivores, although some specialise on invertebrate prey or seeds. Given that they regularly consume weed seeds and crop pests, they are thought to be highly beneficial for agricultural productivity, but they can also consume beneficial invertebrates, such as other predators of crop pests, and they may also consume crop seeds. In order to advance our capacity for inferring interactions between invertebrate consumers and their resources, we first need to compare our inferences with known interactions. Such datasets do exist across some contexts, but we must also generate new interaction data in order to incorporate new data types within network inference models. The trophic ecology of invertebrate generalist predators, however, remains poorly understood given the difficulties associated with studying ecologically cryptic interactions of often nocturnal consumers. Molecular dietary analysis (e.g., dietary metabarcoding) circumvent this though, by allowing post-mortem reconstruction of dietary interactions long after they have occurred.

Using interaction and resource availability data, we can also begin to investigate the preferences of invertebrate consumers in the field by comparing their interactions with what we would expect them to interact with if foraging randomly. Understanding these preferences can help to refine network inference by defining likely preferential links between consumers and their resources. Such preferences are likely driven by the fundamental currency of trophic interactions: nutrients. By integrating nutrient contents into these ecological networks, we may be able to enhance our ability to infer interactions greatly.

This project will use existing ground beetle-seed interaction data to test network inference models, and then apply these to data generated from the field using dietary metabarcoding, nutritional analysis and null network modelling. This will enhance our ability to infer interactions accurately for the prediction and management of ecosystem services and disservices provided by ground beetles and other generalist consumers. This project will also generate crucial and novel insights into the foraging ecology and nutrition of generalist invertebrate omnivores beyond the reach of previous studies in this field.

The successful student will analyse an existing dataset of ground beetle-seed interactions and apply different network inference methods to determine their comparative accuracy in predicting the recorded interactions of these beetles through comparison with existing interaction data. Beetle and seed trait data will be used to refine inferences and assess the importance of traits in determining the accuracy of inference, and, simultaneously, the identity and ecosystem service implications of interactions.

To test these inference methods further and integrate new interactions and context to the data, the student will build a new empirical dataset. This will involve fieldwork at one of Newcastle University’s experimental farms (Cockle Park). During fieldwork, ground beetles and the prey and seeds available to them will be collected from different crop types and adjacent semi-natural habitats in transects across the site. Ground beetles will be identified and undergo dietary analysis in Newcastle University’s Molecular Diagnostics Facility using high-throughput robotics and cutting-edge diagnostics equipment. The gut contents of beetles will be analysed for both animal and plant DNA to determine their omnivorous diet and trophic interactions. These interactions will be compared between taxa, habitats and across time to determine how these factors influence trophic network structure and ecosystem service provision.

The seeds and prey available to them will be identified, and these data will then be used to generate null models of expected interactions. These will be compared to the actual interaction data to assess resource preferences of the beetles and how these change across space and time, and between species. Seed and prey nutrient contents (proteins, lipids, carbohydrates) will also be determined by colorimetric assays and compared between resource taxa and habitats. These will be used to construct ‘nutritional networks’ (see references) and determine how nutrients drive interactions and wider ecological network structure. The resource choice and nutrition results will be used in subsequent network inference using the interaction data generated in the project to refine predictions of interactions and build increasingly accurate models of network inference.

Year 1

• Settling in and inductions
• Reviewing the literature
• Using network inference approaches with existing beetle-seed interaction data
• Fieldwork

Year 2

• Morphological identification of invertebrates and seeds
• DNA metabarcoding of beetle gut contents
• Construction and analysis of trophic networks
• Assessment of ecosystem service provision trade-offs

Year 3

• Nutritional analysis of invertebrate prey and seeds
• Construction and analysis of nutritional networks
• Integrate nutritional data into network inference models
• Assess the nutrition as a driver of ecosystem services

Year 3.5

• Complete and refine analyses
• Explore wider implications and future steps for this work
• Complete writing of the thesis

The following skills will be developed throughout the project:
• Molecular dietary analysis
• Macronutrient analysis
• Network inference
• Null network modelling
• Ecological network analysis
• Scientific writing and publishing
• Open research
• Academic service