Trophic hierarchy in a marine community revealed by network analysis on co-occurrence data

Network analysis is employed in biodiversity studies to explore the possible ecological implications of taxa co-occurrences across time or space, but the conceptual robustness of this approach is still under debate. Herein, by focusing our attention on a marine pelagic community previously described by metabarcoding data gathered over a time-series, we explore the usefulness of co-occurrence networks to derive information about food web structures. In our effort we studied a published co-occurrence network by conducting bibliographic research to identify which links were potentially related to ecological interactions. Although biologically plausible links were the minority, we found that they were enough to construct realistic ecological networks. We therefore assembled trophic and non-trophic networks including both putatively biological links and those not explainable based on the literature; we analyzed these networks separately by estimating several indicators chosen among those resulted as informative of ecological processes in previous research. In the trophic networks, we detected some structural properties and 'trophic hierarchies' commonly found in food webs with approaches different from co-occurrence networks, such as: i) a positive relationship between body size and trophic level (this latter defined in terms of network topology, i.e., the arrangement of nodes and links in a network); and ii) realistic shifts in the network roles played by co-occurring taxa (i.e., hubs and keystone species), which allowed the persistence of hierarchy despite environmental variability. Our results reveal the power of applying co-occurrence networks in capturing the ecological properties of marine communities, plus possible ecological changes occurring under different environmental conditions. However, such application potential is deeply constrained by the acquiring of fine-scale knowledge of the trophic habits of the taxa in the community of reference.