A process-based framework for quantifying sampling completeness of species interaction networks

Michael D. Catchen , Timothée Poisot , Laura J. Pollock, Andrew Gonzalez 

Species interaction networks underpin ecosystem function and persistence, but their study is hindered by lack of empirical knowledge about interactions. Many interactions occur in nature that have not been observed or recorded. We develop a process-based framework for simulating species interaction accumulation curves that accounts for spatiotemporal variation species interaction networks. We do this by explicitly linking interaction realization and detection rates to species abundance. We implement this framework in a software package, SpeciesInteractionSamplers.jl, which enables researchers to evaluate monitoring strategies and assess the completeness of empirical interaction datasets under biologically realistic assumptions. Together, this framework and software provide a foundation for quantifying the Eltonian shortfall, and for improving the design of interaction monitoring networks.