Golubski, A. J.,Westlund, E. E.,Vandermeer, J.,Pascual, M.

Analyses of ecological network structure have yielded important insights into the functioning of complex ecological systems. However, such analyses almost universally omit non-pairwise interactions, many classes of which are crucial for system structure, function, and resilience. Hypergraphs are mathematical constructs capable of considering such interactions: we discuss their utility for studying ecological networks containing diverse interaction types, and associated challenges and strategies. We demonstrate the approach using a real world coffee agroecosystem in which resistance to agricultural pests depends upon a large number of TMIIs. A hypergraph representation successfully reflects both the importance of species imposing such effects and the context -dependency of that importance in terms of how it is affected by removal of other species from the system.