METHODOLOGICAL ADVANCES IN INFERRING ANCIENT FOOD WEBS

Food webs provide quantitative insights into micro- and macro-scale ecological processes. Previous work has shown their utility in understanding community responses to modern and ancient perturbations, including anthropogenic climate change and mass extinctions. However, few paleo-food webs have been constructed due to the difficulty of comprehensively assessing trophic interactions in ancient communities. We present and assess a food web inference model that uses taxon-level functional data (i.e., motility, tiering, feeding, and size) to generate a metaweb of all feasible trophic interactions. The metaweb is then used to generate a series of taxon-level food webs, covering the range of food web structures that may have occurred, given assumptions of how interactions are systematically distributed (“degree distributions”) in communities. We used the inference model to generate food webs for five modern and two ancient communities with existing food webs. We found that it generates realistic food web structures comparable to the classic “Niche Model” based food webs, and that differences between the structures of the original and inferred taxon-level webs are generally small. By using widely available functional data our model can easily be deployed across the fossil record to examine the evolution of food webs through space and time.